1006 shifts in intertidal zonation and refuge use by prey after mass mortalities of two predators...

1006

Shifts in intertidal zonation and refuge use by prey after mass mortalities of two predators

SARAH A. GRAVEM 1 AND STEVEN G. MORGAN

Bodega Marine Laboratory, Environmental Science and Policy Department, University of California Davis, 2099 Westside Road, Bodega Bay, California 94923 USA

Abstract. Recent mass mortalities of two predatory sea star species provided an unprece-dented opportunity to test the effect of predators on rocky intertidal prey. Mass mortalities provide insight that manipulative experiments cannot because they alter ecosystems on a larger scale, for longer time periods, and remove both organisms and their cues from the environ-ment. We examined shifts in population size structure, vertical zonation, and use of emersed refuge habitats outside tidepools by the abundant herbivorous black turban snail Tegula fune-bralis, both before and after the successive mortalities of two predatory sea stars. The small cryptic predator Leptasterias spp. suffered a localized but extreme mortality event in November 2010, followed by two mass mortalities of the keystone predator Pisaster ochraceus in August 2011 and autumn 2013. After the local extinction of Leptasterias, the population size of Tegula more than doubled. Also, since Leptasterias primarily inhabited only mid to low intertidal tidepools at this site, small and medium sized snails (which are preferred by Leptasterias) shifted lower in the intertidal and into tidepools after the mortality of Leptasterias. After the mortality of Pisaster in August 2011, large snails did not shift lower in the intertidal zone despite being preferred by Pisaster. Small and medium sized snails became denser in the higher zone and outside tidepools, which was not likely due to Pisaster mortality. Previous studies concluded that Pisaster maintained vertical size gradients of snails, but our data implicate the overlooked predator Leptasterias as the primary cause. This natural experiment indicated that (1) predators exert top- down control over prey population sizes and lower limits, (2) vertical zonation of prey are dynamic and controlled in part by prey behavior, and (3) predators exert the strongest effects on more vulnerable individuals, which typically inhabit stressful habitats higher on the shore or outside tidepools to avoid predation. Because the mass mortalities of two predators drastically reduced both the predation pressure and the chemical cues of preda-tors in the environment, we were able to investigate both the effects of predators on prey populations and the effects on mobile prey behavior.

Key words: intertidal zonation; Leptasterias spp.; mass mortality; natural experiment; nonconsumptive effects; Pisaster ochraceus; sea star wasting disease; size-dependent predation; Tegula funebralis; tidepool; top-down control; vertical size gradient.

INTRODUCTION

Natural experiments are some of the most powerful tools used by ecologists because they can reveal processes occurring in whole ecosystems over space or time. While local extinctions of species by disease, extirpation, or natural disasters are unfortunate events, they provide rare insights into community interactions that are usually impossible or unethical to obtain experimentally (Diamond 1983, Menge et al. 2016). Careful long- term monitoring is essential to capitalize on these rare opportunities. For example, monitoring coral reefs before and after hurri-canes (Connell 1978) and the mass mortality of a sea urchin (Hughes 1994) transformed our understanding of

community stability and alternate stable states, respec-tively. Also, local extirpation or reintroduction of pred-ators in lakes, tropical forests, and coniferous forests led to the reorganization of food webs, illustrating the com-plexity of trophic interactions and importance of top- down control (Zaret and Paine 1973, Terborgh et al. 2001, Ripple and Beschta 2012). More recently, climate change has served as a natural experiment demonstrating that species distributions, phenology, migration patterns, and species interactions are sensitive to warming temperatures (Parmesan 2006, Poloczanska et al. 2013).

Recent mortality events of predatory sea stars on the west coast of North America present a unique oppor-tunity to investigate wellknown concepts in ecology, including keystone predation, trophic cascades, and biotic control of intertidal zonation (Menge et al. 2016). The concept of intertidal zonation suggests that lower limits of species are controlled by biotic interactions while upper limits are controlled by abiotic stresses (Connell 1972, Robles et al. 2009, Mislan et al. 2014). It has also

Ecology, 98(4), 2017, pp. 1006–1015© 2016 by the Ecological Society of America

Manuscript received 15 February 2016; revised 18 Novem-ber 2016; accepted 28 November 2016. Corresponding Editor: Sergio A. Navarrete.

1 Present address: Integrative Biology Department, Oregon State University, 3029 Cordley Hall, Corvallis, Oregon 97331 USA; E-mail: [email protected].

PREY SHIFT ZONES WITH PREDATOR MORTALITYApril 2017 1007

become increasingly apparent that vertical distributions in the intertidal zone are actually dynamic and not per-manent spatial refuges from competition or predation (Robles and Desharnais 2002, Robles et al. 2009, Donahue et al. 2011). Vertical zonation may be especially dynamic for mobile species that may adjust distributions through behavior (Vermeij 1972, Rochette and Dill 2000). Predator control over dynamic lower limits would be supported if mobile prey species respond quickly to mass mortality of predators by moving lower on shore.

Unlike manual removals of predators in most field experiments, mass mortalities of predators may more thoroughly eradicate predators, last longer, and elim-inate chemical cues that can strongly affect behavior of even far- away prey (Kats and Dill 1998, Laundré et al. 2010). When mortalities are extensive, the opportunity arises to assess the subtle, long- term effects of predators on prey species including (1) nonconsumptive effects on prey behavior and distribution and (2) survival and growth of prey released from predation. This thorough removal of predators and especially their chemical cues is much harder to obtain in a manipulative field experiment. On the other hand, mass mortalities of predators are natural experiments that typically lack experimental con-trols that capture influences of extraneous factors, and they lack replication that enables one to test for con-sistency in the responses (Diamond 1983). Thus, the experimenter is less able to attribute the responses by prey to release from predation. While this lack of control and replication is not ideal, extensive predator mortal-ities may enable unprecedented insight into predator–prey interactions (Menge et al. 2016).

One such interaction is the influence of predators on patterns of intertidal zonation; predators often drive vari-ation in prey size among intertidal zones, which are referred to as “vertical size distributions” (Seed 1969, Vermeij 1972, McQuaid et al. 2000). For mobile prey

species, individuals must balance the conflicting demands of (1) avoiding stress near upper limits, (2) avoiding pre-dation or competition near lower limits, and (3) seeking abundant food near lower limits (Paine 1969, Bertness 1977, Rochette and Dill 2000, Pincebourde et al. 2008). Importantly, the balance of these conflicting needs often depends on the size of the organism, resulting in vertical size gradients (Vermeij 1972). For example, large indi-viduals may occur higher on the shore because they can better withstand physiological stress than juveniles (Vermeij 1972) or predators may preferentially consume large individuals low on the shore (Cushman 1989, Rochette and Dill 2000). Alternatively, the opposite pattern may arise when predators preferentially consume smaller individuals low on the shore or large individuals choose to risk predation for increased energetic gain low on the shore (Paine 1969, Vermeij 1972). Thus, mass mor-talities of predators provide an opportunity to determine the effect of predators on vertical size distributions of prey.

In rocky intertidal communities on the west coast of North America, larger individuals of the abundant her-bivorous gastropod, Tegula (formerly Chlorostoma) funebralis (Bouchet and Rosenberg 2015) generally occur lower on the shore than smaller ones (Fig. 1a; Wara and Wright 1964, Paine 1969, Markowitz 1980, Byers and Mitton 1981, Doering and Phillips 1983, Fawcett 1984). This pattern has often been attributed to predation pressure by the original keystone species Pisaster ochraceus, which is most abundance lower on shore (Wara and Wright 1964, Paine 1969, Markowitz 1980). However, Pisaster preferentially consume large snails in the field and laboratory (Markowitz 1980). Thus, it is somewhat perplexing that large snails tend to be lower on shore where Pisaster is most abundant. Hence, it was proposed that Pisaster force all individuals upward but that large, reproductive snails (>12–14 mm diameter) risk venturing low on the shore to forage on

FIG. 1. (a) Diagram depicting the typical trend of decreasing Tegula funebralis size at higher shore levels. Since Leptasterias spp. (six arms) prefers smaller Tegula, we expected small Tegula to increase in the lower zone after (b) the Leptasterias mortality event. Pisaster ochraceus (five arms) prefers larger Tegula, which should increase in the lower zone after (c) the Pisaster mortality events. [Color figure can be viewed at wileyonlinelibrary.com]

Typical pattern Expectation after Leptasterias mortality

Expectation after Pisaster and Leptasterias mortalities

a cb

incr

easi

ng s

hore

leve

l

1008 Ecology, Vol. 98, No. 4SARAH A. GRAVEM AND STEVEN G. MORGAN

abundant food and satisfy high energetic demands (Paine 1969, Markowitz 1980).

One piece of this long- standing puzzle may be missing. The small (1–5 cm diameter) predatory sea star, Leptasterias spp. (see Flowers and Foltz [2001] for species complex) co- occurs with Tegula in low to mid intertidal zones along the west coast of the United States (Morris et al. 1980). Leptasterias nearly exclusively prey on small and medium- sized Tegula (<18 mm) compared to large Tegula (>18 mm) in the field and laboratory (Bartl 1980, Gravem and Morgan 2016). Thus, it may be that Leptasterias, not Pisaster, are actually responsible for the decrease in Tegula size with shore level (Wara and Wright 1964, Paine 1969, Markowitz 1980, Byers and Mitton 1981, Doering and Phillips 1983, Fawcett 1984).

We capitalized on natural mass mortality events of Leptasterias in November 2010 followed shortly thereafter by Pisaster in August 2011 and November 2013 in northern California (Fig. 2). Based on surveys over time at our study site in Horseshoe Cove, Bodega Head, California, USA (38°18′59.4″ N, 123°4′16.3″ W) and qualitative searches of Leptasterias in the surrounding area, the Leptasterias mor-tality event in November 2010 was, as far as we know, localized near Bodega Head but was severe with nearly 100% mortality and no recovery through the present because Leptasterias brood their larvae (Fig. 2; see Appendix S1 and Appendix S1: Fig. S1 for survey details, extent of mortality, and possible causes). The first Pisaster mortality event in August 2011 was caused by a harmful algal bloom that extended along 100 km of coastline, primarily in Sonoma County, California (Rogers- Bennett et al. 2012, Jurgens et al. 2015). The bloom killed nearly 100% of Pisaster at Schoolhouse Beach (Fig. 2, 38°22′28.4″ N, 123°4′44.9″ W), which is 6.4 km north of our study site at Horseshoe Cove.

The second Pisaster mortality was due to sea star wasting disease, which caused severe declines in Pisaster density from Mexico to Alaska (Hewson et al. 2014, Menge et al. 2016), including an 89% decline 6.4 km north of our site at Schoolhouse Beach (Fig. 2).

In addition to affecting vertical size distributions, predators commonly cause prey to increase use of refuges (Lima 1998), which in intertidal communities, may include emersed rock outside tidepools where predators cannot forage at low tide (Feder 1963, Menge and Lubchenco 1981). To first test this effect, we examined whether Tegula responded to Leptasterias presence in tidepools by increasing use of emersed refuges just outside tidepools in spring 2010, before any mortality events. After the Leptasterias mortality event in November 2010, we expected that survival of vulnerable smaller snails would increase and that smaller snails would move lower on shore (Fig. 1). We also expected smaller snails to move into tidepools since Leptasterias is primarily found only inside tidepools and not on emersed rock at this site. When the Pisaster mortalities occurred in August 2011 and November 2013, we expected that snails would descend on the shore and increasingly inhabit tidepools, and this would be most evident for the large snails that are preferred by Pisaster (Fig. 1; Markowitz 1980). Though Pisaster occurs in both tidepools and the emersed matrix, they are unable to forage at low tide when emersed (Menge and Lubchenco 1981). To test these hypotheses, we surveyed tidepools spanning the vertical range of Tegula and measured snail sizes and abundances in tide-pools and emersed refuges. This natural experiment enabled us to investigate the potential top- down effects of two predators on prey population size, size distribu-tions of prey, and prey refuge use at two spatial scales: among intertidal zones and in and out of tidepools.

METHODS

Shifts in prey population after predator mortalities

We determined whether Tegula populations or behavior changed in Horseshoe Cove after the mortalities of both sea star species by surveying the population size structure, vertical zonation, and refuge habitat use of Tegula in 21, 44, and 30 tidepools on 10 April 2010, 1–27 May 2011, and 10–19 April 2014, respectively (arrows in Fig. 2). Each survey was conducted in spring to control for sea-sonal differences in sea star and snail distributions and behaviors (Paine 1969, Menge 1972, Markowitz 1980).

To test size- dependent responses of snails to sea star mortalities, we counted and measured snails separately from inside tidepools and in refuge habitats just outside tidepools termed “halos.” Halos included a 15- cm band of emersed rock surrounding each tidepool where sea stars do not often forage at low tide (Menge and Menge 1974). The percentage of snails in the halo [(snails in the halo/snails in tidepool and halo) × 100] was used to estimate refuge use. Snails were considered to be in the

FIG. 2. Number of Leptasterias spp. (solid circles) removed weekly from eight tidepools in Horseshoe Cove, California, USA and number of Pisaster ochraceus (open circles) removed biweekly from 12 large intertidal boulders 6.4 km north at Schoolhouse Beach. Three mass mortality events occurred at our site (gray boxes): first of Leptasterias spp. in November 2010, second and third of P. ochraceus in late August 2011 and fall 2013, respectively. Tegula funebralis were surveyed in 21, 44, and 30 tidepools in Horseshoe Cove in spring 2010, 2011, and 2014, respectively (black arrows). [Color figure can be viewed at wileyonlinelibrary.com]

20

40

60

80

100

120

2

4

6

8

10

12

14

2009 2010 2011 2012 2013 2014Date

Num

ber

of L

epta

ster

ias

Num

ber

of P

isas

ter


halo if they broke the water’s surface. We restricted surveys to tidepools and halos because this was part of another study that focused on the effect of sea stars on microhabitat choice by snails (Gravem 2015). We divided snails by size into 3- mm increments at their widest shell diameter and defined small snails as <12 mm and not yet mature (Paine 1971), medium snails as 12–18 mm and newly mature, and large snails as >18 mm, mature, rarely eaten by Leptasterias, and preferred by Pisaster (Markowitz 1980, Gravem and Morgan 2016). In 2010, we surveyed tidepools with Leptasterias and tidepools where we had experimentally removed Leptasterias as part of another experiment (see Appendix S1 for removal methods) to determine size- dependent snail responses to Leptasterias. Leptasterias was absent from the site in 2011 and 2014. Pisaster was absent from all tidepools for all surveys though Pisaster occurred in the surrounding area, especially in 2010 and 2011.

We chose tidepools over a large vertical range that encompassed the Tegula zone, and used surveying equipment and United States Geological Survey bench-marks to measure the shore level of each tidepool surface (range 0.95–2.23 m above mean lower low water, MLLW). We categorized tidepools into lower, middle, and higher shore levels for statistical analysis (0.70–1.15, 1.15–1.50, and 1.50–2.30 m above MLLW). These cate-gories do not match classic definitions of low, mid, and high zones, but rather characterize the distribution of Tegula. To estimate snail abundance among different sizes of tidepools, we estimated snail density per liter by quantifying water pumped from tidepools into buckets (range 1.2–85.0 L). This metric was used because tide-pools were rugose and varied in depth, so using the surface area of tidepools would have overestimated den-sities. Further, this was part of a study focused on the effects of waterborne predator chemical cues on prey behavior (Gravem 2015).

Statistical analyses

All data were analyzed using mixed linear models fit using restricted maximum likelihood with Tukey- Kramer honestly significant difference posthoc analyses using JMP 10 software (SAS Institute, Cary, North Carolina, USA). We analyzed whether the presence of Leptasterias and snail size influenced (1) use of halo refuges by snails in 2010 only and whether year, snail size and shore level influenced, (2) snail densities, and (3) use of halo refuges by snails in all years. Tidepool number was included in all models as a random factor (nested in other factors when appropriate) to account from non- independence of snails in separate size classes in the same tidepool (see Appendix S2 for statistical details). To test whether changes in snail densities among shore levels were due to behavioral shifts or simply to overall increases in population density, we used chi- square analyses on the expected vs. observed densities in each zone for each snail size in 2010 vs. 2011 and 2011 vs. 2014 (see Appendix S2 for details).

RESULTS

Predator avoidance

When Leptasterias was present in tidepools (2010 survey only), vulnerable small Tegula tended to be more common in halo refuges, whereas this was not the case for large snails (Fig. 3; Appendix S2: Table S1; Leptasterias presence × Size F6,54 = 2.77, P = 0.019). However, pairwise tests within size classes were not significantly different. When Leptasterias was absent, the percentage of snails in halos was similar among size classes.

Population size structure before and after predator mortalities

The number of snails more than doubled (factor of 2.4) in 2011 after the Leptasterias mortality event (Fig. 4). After the Pisaster mortality events in 2014, the popu-lation again nearly doubled (factor of 1.7). The increase in 2011 was driven primarily by small snails in the 6–9 mm size class (401% average increase; Fig. 4; Appendix S2: Table S2a; Year × Size F12,554 = 7.82, P < 0.001; Year F2,554 = 8.90, P < 0.001), and all other size classes also increased (66%, 112%, 104%, 144%, 95%, and 19% average increases for <6, 9–12, 12–15, 15–18, 18–21, and >21 mm sizes classes, respectively). Small and medium snails in the 9–12 and 12–15 mm size classes drove the population density increase from 2011 to 2014 (Fig. 4; Appendix S2: Table S2b; 108% and 108% average increases, respectively), with the 6–9 and 15–18 mm size classes also tending to increase (54% and 44% average increases, respectively). Snails that were 3–6 mm were

FIG. 3. Percentage (mean ± SE) of total Tegula funebralis by size class (3 mm increments in shell diameter) inhabiting emersed halo refuges around tidepools vs. inside tidepools where Leptasterias spp. were present (solid circles, n = 10 tidepools) or where they had been experimentally removed (open circles, n = 10 tidepools). Data were taken on 10 April 2010 before sea star mortality events in Horseshoe Cove, California, USA. [Color figure can be viewed at wileyonlinelibrary.com]

0

20

40

60

80

100

Sna

ils in

hal

o (%

)

Snail size (mm)

Leptasterias absent Leptasterias present

<6 9–12

18–2

115

–18

>216–

912

–15


rare, likely because snails grow very quickly out of this stage and are not commonly found. We are confident we would have found any of these small snails present in tidepools.

Vertical distribution before and after predator mortalities

Snails shifted vertically after the sea star mortalities, but these responses depended on snail size (Fig. 5; Appendix S2: Table S2a; Year × Shore level × Size class: F24,554 = 1.60, P = 0.037). Small and medium snails tended to be denser in the highest zone in 2010, but shifted lower in 2011 after the Leptasterias mortality event: small snails became denser in the middle and lower than the

higher zone, and medium snails became more evenly dis-tributed, though these differences were not significant in posthoc analyses (Fig. 5; Appendix S2: Table S2c). To further investigate this pattern, we use chi- square analyses and found that more small and medium snails (<15 mm) occurred in the middle zone than expected based on pop-ulation size changes, but fewer than expected small snails (6–12 mm) occurred in the higher zone (χ2 = 1,110.3, df = 20, P < 0.0001), indicating that snails most vul-nerable to Leptasterias shifted to the middle zone when Leptasterias was absent. Fewer than expected snails of nearly all sizes (<18 mm) occurred in the lower zone in 2011, indicating more snails may have moved to the middle rather than lower zone. In 2014 after the Pisaster mortality events, small and medium snails remained dense in the lower and middle zones, but 9–15 mm snails became denser in the higher zone (Fig. 5; Appendix S2: Tables S2a and S2c; Year × Shore level × Size class F24,554 = 1.60, P = 0.037). Further, fewer than expected small snails occurred in the middle zone (6–9 mm) and more than expected small snails occurred in both the higher (9–12 mm) and lower zones (6–9 mm; χ2 = 436.8, df = 20, P < 0.0001). Again, fewer snails than expected of most sizes (9–18 mm) occurred in the lower zone during in 2014, indicating most snails still did not move to the lower zone despite the low densities of both sea stars. Surprisingly, large snails were never denser in the lower than middle and higher zones (Paine 1969, Markowitz 1980, Doering and Phillips 1983). Rather, large snails were evenly distributed among zones during all years and exhibited no clear responses to either Leptasterias or Pisaster mortalities (Fig. 5).

Microhabitat shifts before and after predator mortalities

Overall, halo refuge use was highest in 2010 before any sea star mortalities, lowest in 2011 after Leptasterias

FIG. 4. Population size structure (3 mm increments in shell diameter) calculated as densities (mean ± SE) of Tegula funebralis in tidepools and halo refuges combined in April 2010 before sea star mortality events (circles), May 2011 after a Leptasterias spp. mortality event (squares), and April 2014 after two Pisaster ochraceus mortality events (triangles). For 2010, 2011, and 2014, respectively, n = 21, 44, and 30 tidepools. [Color figure can be viewed at wileyonlinelibrary.com]

0

2

4

6

8

10

Snai

l den

sity

(no.

sna

ils/L

)

Snail size (mm)

2010

2011

2014

FIG. 5. Densities of Tegula funebralis (mean ± SE) in tidepools and surrounding halo refuges by snail size class (3 mm increments in shell diameter) and shore level (a) in April 2010 before sea star mortality events, (b) in May 2011 after a Leptasterias spp. mortality event, and (c) in April 2014 after two Pisaster ochraceus mortality events. For lower, middle, and higher shore levels, respectively, n = 7, 9, and 5 tidepools in 2010 (panel a), n = 10, 16, and 18 tidepools in 2011 (panel b), and n = 9, 11, and 10 tidepools in 2014 (panel c). Shore levels in meters above mean lower low water are 0.70–1.15 m (lower; circles), 1.15–1.50 m (middle; squares), and 1.50–2.30 m (higher; triangles). [Color figure can be viewed at wileyonlinelibrary.com]

0

2

4

6

8

10

12

14

Snai

l den

sity

(no.

sna

ils/L

)

a) 2010 Lower

Middle

Higher

c) 2014

Snail size (mm)

b) 2011

<6 6–9

9–12

12–1

515

–1818

–21

>21 <6 6–

99–

1212

–1515

–1818

–21

>21<6 6–

99–

1212

–1515

–1818

–21

>21


mortality, and intermediate in 2014 when both Pisaster and Leptasterias densities were low (Fig. 6a; Appendix S2: Table S3a; Year: F2,345 = 27.13, P < 0.001). Small and medium snails drove these changes over time; halo use decreased with snail size in 2010, but in 2011 snails <12 mm and 15–18 mm were found in tidepools more often than in 2010, resulting in low halo use among all sizes of snails (Fig. 6a; Appendix S2: Tables S3a and S3b; Year × Size class: F12,345 = 6.86, P < 0.001). In 2014 after the Pisaster mortality events, snails generally shifted from tidepools to halos, except for the smallest (<6 mm) and largest (>21 mm) snails, which remained in tidepools.

Snails generally used halos more often when lower on the shore in 2010 when Leptasterias was abundant, but tidepool use increased in the middle and low zones in 2011 after Leptasterias mortality (Fig. 6b; Appendix S2: Tables S3a and S3c; Year × Shore level F4,345 = 2.41, P = 0.057). More snails occurred in the halos in the lower zone in 2014 than 2011, whereas snails in the higher zone tended to inhabit tidepools during all years. All sizes of snails showed theses trends (Fig. 6b; Appendix S2: Table S3a; Year × Shore level × Size class F24,345 = 0.82, P = 0.705).

DISCUSSION

Similar to other natural experiments that have uncovered previously unrecognized ecological interac-tions (Zaret and Paine 1973, Hughes 1994, Terborgh et al. 2001, Ripple and Beschta 2012), this study suggests biotic control of Tegula populations and zonation by the less- studied sea star predator, Leptasterias spp., but not by the keystone predator, P. ochraceus. The vertical shift of smaller snails from higher to middle shore after Leptasterias died also supports Vermeij’s (1972) generalization that predation pressure causes smaller individuals to flee from

lower to higher on shore. Further, it supports his hypothesis that this trend should be particularly strong when predators prefer smaller prey, as with Leptasterias, but less strong when predators prefer larger prey, as with Pisaster.

Top- down effects of predators on a prey population

Leptasterias apparently regulated Tegula populations prior to their mortality in 2010. After Leptasterias died, the population size of snails increased 2.4- fold, primarily due to a 401% increase in small snails (6–9 mm) less than 2 years old (Paine 1969). Small and medium Tegula are very vulnerable to Leptasterias predation and Tegula comprise between 14% and 24% of Leptasterias diets in their zone of overlap at this site (Bartl 1980, Gravem and Morgan 2016), suggesting that Leptasterias is capable of limiting juvenile survival. Though Pisaster were formerly implicated, Leptasterias predation may have also played a major role in the ~60% decrease in the abundance of Tegula during spring and summer in neighboring Marin County, California (Markowitz 1980).

Conversely, it is possible this surge in juvenile abun-dance was not due to decreased predation but simply to sporadically high Tegula recruitment. A lack of control for other factors, such as recruitment, is a common short-coming of natural experiments. However, multiple lines of evidence suggest that recruitment may be consistently high at this site but juvenile survival may have been pre-viously limited by Leptasterias predation. First, new recruits (6–9 mm) were abundant in both 2011 and 2014 (compared to 2010) with no obvious gaps in recruitment evident in the intervening years, suggesting at least two strong recruitment years occurred after the mortality of Leptasterias. Also, the population size structure both before and after mortalities was not multimodal, as is

FIG. 6. Percentage (mean ± SE) of Tegula funebralis inhabiting halo refuges surrounding tidepools in April 2010 before sea star mortality events (circles), May 2011 after a Leptasterias spp. mortality event (squares), and April 2014 after two Pisaster ochraceus mortality events (triangles) by (a) snail size classes (3 mm increments in shell diameter) and (b) shore levels. For 2010, 2011, and 2014, respectively, n = 7, 10, and 9 tidepools in the lower zone, n = 9, 16, and 11 tidepools in the middle zone, and n = 5, 18, and 10 tidepools in the higher zone. Shore levels in meters above mean lower low water are 0.70–1.15 m (lower), 1.15–1.50 m (middle), and 1.50–2.30 m (higher). [Color figure can be viewed at wileyonlinelibrary.com]

0

20

40

60

80

100

Snai

ls in

hal

o (%

)

Snail size (mm)

a2010

2011

2014

Lower Middle Higher Shore level

b


common for species with sporadic recruitment (Menge et al. 2004). However, the long life span of Tegula (Paine 1969) could diminish the sharpness of recruitment peaks. In addition, the population size structure of Tegula between southern Oregon and Baja California, Mexico is consistently skewed toward juveniles, and reproduction and recruitment in these populations is high year- round (Frank 1975, Fawcett 1984, Cooper and Shanks 2011). Overall, this suggests that Leptasterias had consistently been reducing survival of juvenile Tegula for many years, just as other predators exert top- down control of prey populations by consuming juveniles (Hunt and Scheibling 1997). Since we rarely observed small snails in the emersed rock matrix between tidepools, it is also doubtful that the surge in juvenile abundance was due to juvenile snails shifting out of the matrix into tidepools and halos.

After Pisaster died, survival of juveniles and growth and survival of all snails continued to be high. Increases in abundance of the 9–12 and 12–15 mm snails in 2014 were consistent with the survival and growth of the abundant 6–9 mm cohort of snails from 2011 (Paine 1969), However, this was likely primarily due to continued low predation by Leptasterias, since Pisaster prefers large snails (Markowitz 1980). Snails grow to the size preferred by Pisaster (>17 mm, Markowitz 1980) at around 12 years old (Paine 1969), so it was unsurprising that the densities of large snails did not increase dramatically after Pisaster died. We expect the cohort of snails that appeared in 2010/2011 to grow large enough to substantially increase the densities of large snails within a decade (Paine 1969) unless Pisaster recover first. We also expect continued high survival and growth of small and medium snails because Leptasterias is still absent and, because it broods its larvae, is expected to return to the area slowly.

Effects of predators on prey behavior and distributions

Fairly rapid shifts (<5 months) by small and medium snails to the middle and lower zones and tidepools after the Leptasterias mortality event indicate that zonation is dynamic (Robles and Desharnais 2002, Robles et al. 2009, Donahue et al. 2011) and is partially maintained by antipredator behavior of prey, similar to other mobile intertidal species (Vermeij 1972, Cushman 1989, Rochette and Dill 2000). Further, Leptasterias rather than Pisaster likely relegated small and medium Tegula to the higher zone and outside tidepools. Small and medium snails likely balanced the risks of abiotic stress, predation by Leptasterias, and starvation (Paine 1969, Bertness 1977, Rochette and Dill 2000, Pincebourde et al. 2008). Snails descending lower on the shore and into tidepools in 2011 likely experienced reduced osmotic, desiccation and thermal stress, all of which may be especially harmful to smaller individuals (Marchetti and Geller 1987). Further, they likely benefitted from higher food availability and longer foraging bouts (Underwood 1984, Wright and Nybakken 2007). The fairly rapid behavioral shift after the Leptasterias mortality event is consistent with the

well- documented predation risk allocation hypothesis (Lima and Bednekoff 1999), which posits that at times or locations of lower risk, individuals should be more active and move into preferred habitat.

Since inhabiting lower shore levels or tidepools con-tributes to generally higher growth rates and fecundity (Paine 1971, Underwood and McFadyen 1983, Pardo and Johnson 2005, Perez et al. 2009), these behavioral shifts may have increased survival and growth of snails and contributed to the observed increases in population size over time. Though we cannot separate the noncon-sumptive from consumptive effects of Leptasterias, both may have been important since Leptasterias apparently limited the abundance and altered the behavior of Tegula.

Unlike small snails, large snails remained evenly dis-tributed among intertidal zones and did not shift into tidepools after the mortality of Leptasterias. Though large snails readily respond to contact and waterborne chemical cues of Leptasterias in the laboratory and in short- term field experiments, they are less vulnerable than small snails to Leptasterias predation (Gravem and Morgan 2016). It is likely that large snails are faster and easily outpace Leptasterias when contacted (Gravem and Morgan 2016), so it is unnecessary for large snails to react to sustained exposure to cues in the environment since Leptasterias pose little risk. On the other hand, small snails are likely less able to outpace Leptasterias, so they react to both temporary tactile and sustained waterborne cues with more apparent impacts on their long- term dis-tributions. It is not known whether the lack of predator avoidance behavior by large snails to in response to sus-tained cues from Leptasterias is learned or innate.

We expected large snails to respond to Pisaster mor-tality by descending on the shore and perhaps into tide-pools where Pisaster can continue to forage at low tide, but clear responses were not evident. Similarly, large snails in 2010 were not as abundant in the lower zone as expected (Wara and Wright 1964, Paine 1969, Markowitz 1980, Byers and Mitton 1981, Doering and Phillips 1983). These outcomes were surprising since large snails grow faster and have larger gonads lower on the shore, sug-gesting that this is their preferred habitat (Paine 1969). Perhaps these two trends are linked, and snails at our study site are not as food limited as other sites, allowing larger snails to thrive in the upper intertidal zone. Most prior studies were conducted on exposed rock surfaces, but our site contains abundant tidepools. So large snails may be able to gain enough energy from tidepool algae even at higher shore levels, negating the need to descend lower on the shore. Future analyses of gonad size of large snails would demonstrate whether those inhabiting higher tidepools are food limited. Alternatively, the threat of predation may have been high for large snails in all years causing many to remain high on the shore. Indeed, some Pisaster remained after the mortality events. Further, other predators of large snails likely inhabited the low intertidal zone, including the crab Cancer productus, the octopi Octopus rubescens and Octopus dofleini, and fishes,


such as cabezon (Scorpaenichthys marmoratus). High combined densities of other octopi, crabs, and Pisaster in southern California were proposed to increase large Tegula abundance in the high zone (Fawcett 1984).

Since small and medium snails remain vulnerable to Pisaster, though not preferred over large snails, we expected them to occur even lower on the shore and more abundantly in tidepools after the Pisaster mortality event. Instead, their abundance remained fairly static in the lower zone and in tidepools, but increased in the higher zone and halo refuges. We do not believe this was a response to the mortality of Pisaster. Rather, snails may have been responding to an uncommonly large pulse in recruitment of Pisaster that occurred in summer and early fall 2013 (S. A. Gravem, personal observation). While none of the extremely small (<1 cm diameter in April 2014) Pisaster were observed during our surveys nor were capable of consuming the vast majority of the snails, their trace chemical cues may have been permeating the water column and causing snails to move upward or out of tide-pools. Based on our observations, we have no evidence that snails are able to ascertain the size of their predator visually or through tactile means. Alternatively, some snails may have been forced into less preferred higher and emersed habitats as the population size grew and intraspe-cific competition intensified. This is consistent with our Chi- square analyses showing lower than expected small snails in the middle zone and higher than expected small snails in the lower and higher zones in 2014, suggesting that competition may have forced competitively inferior small snails to expand to all zones rather than concen-trating in the apparently preferred middle zone.

Drivers of vertical size gradients

The next step is to determine the generality of the effect of Leptasterias on the vertical size gradient of Tegula by expanding investigations to include a larger geographic range. Since the range of Leptasterias (Morris et al. 1980, Foltz 1997, Carlton 2007), encompasses all sites previ-ously studied (Wara and Wright 1964, Paine 1969, Markowitz 1980, Byers and Mitton 1981, Doering and Phillips 1983, Fawcett 1984), Leptasterias could have contributed to the patterns that were attributed to Pisaster and other predators. When comparing patterns latitudinally, Fawcett (1984) found that the vertical size gradient (smaller snails higher on shore) was more con-sistent in Northern California and the Pacific Northwest than Southern California. This was attributed to higher combined densities of Pisaster, crabs, and especially octopi in southern California, which may have caused all snails to ascend, thereby precluding a vertical size gra-dient. However, this does not explain why the gradient existed in Northern California and the Pacific Northwest where sea stars and crabs also are extremely common (Morris et al. 1980, Fawcett 1984, Menge et al. 2004). On the other hand, Leptasterias are much more common in Northern California and the Pacific Northwest than in

Southern California (B. Menge, personal communication), and they may be responsible for driving smaller snails higher on the shore at more northern sites.

Though our study suggests Leptasterias influences ver-tical size distributions of Tegula, other factors are likely at play. For example, Pisaster recruitment is also more fre-quent in the Pacific Northwest than in California (Menge et al. 2004), so juvenile Pisaster, which presumably eat smaller snails, may contribute to the stronger vertical gra-dient at more northern sites (Fawcett 1984). Further, experiments showed that snails moved upward in the intertidal in response to Pisaster cues within days (Markowitz 1980). Doering and Phillips (1983) also sug-gested that the vertical distributions of Tegula are main-tained proximally by ontogenetic shifts in responses to light and gravity. Further, some individuals may prefer certain shore levels or tidepool habitats (Frank 1975, Byers and Mitton 1981, Byers 1983), and wave exposure may affect population size structure among sites (Cooper and Shanks 2011); but how these factors drive vertical size distributions is unclear. Thus, Leptasterias as well as other factors may influence vertical size gradients of Tegula on rocky shores.

The oversight of Leptasterias as a strong interactor with Tegula is perhaps due to the notoriety of Pisaster as a keystone species and because it is large, colorful and iconic rather than small, cryptic, and nocturnal like Leptasterias. However, unlike humans, many inverte-brates including Tegula primarily rely primarily on olfactory chemical rather than visual cues (Kosin 1964, Phillips 1978). The observed disparate behavioral responses by snails to these two predators are consistent with laboratory experiments, which suggested that Tegula is able to distinguish the chemical cues of the two sea star species rather than responding to a general chemical cue from sea stars (Bullock 1953, Yarnall 1964).

CONCLUSION

Our natural experiment on the consequences of suc-cessive mass mortality events of two predatory sea star species enabled us to test several key concepts in com-munity ecology. We provided support for biotic control of species lower limits and top- down control of prey pop-ulation size by predators. Combined consumptive and nonconsumptive effects of predators also likely resulted in dynamic zonation and vertical size gradients of mobile prey, because the most vulnerable individuals may typi-cally be eaten or escape to stressful refuges higher on shore and outside tidepools. Strong responses after the Leptasterias mortality event, but not after the Pisaster mortality events, suggested that largely overlooked Leptasterias played a primary role in controlling juvenile survival, population size structure, vertical size gradient, and microhabitat choices of Tegula. Long- term and non-consumptive effects of predators on prey are hard to test with the limited spatial and temporal scales of manipu-lative experiments where predators cannot be removed


completely. Conversely, mass mortalities of predators reduce the concentration of chemical cues from predators, (which can strongly influence prey behavior) and keep predator densities low for many years, allowing more extensive exploration of predator effects. This natural experiment adds new insights on the long- term and non-consumptive effects of predators on intertidal zonation.

ACKNOWLEDGMENTS

Many thanks to the Bodega Marine Laboratory Marine Ecology graduate and undergraduate classes in springs 2010, 2011, and 2014 who collected much of the data, especially Violet Compton, Sarah Wheeler, Max Castorani, Olivia Turnross, Ryanne Ardisana, Jonathan Demmer, Laura Jurgens, Sarah Hameed, Chris Griesemer, and Gabe Ng. Eric Sanford and Brian Gaylord provided valuable feedback on the manuscript. Thanks to Tessa Hill, Matt Robart, and the Bodega Ocean Observing Node for the biophysical data and Gregg Langlois from the California Department of Public Shellfish Protection and Marine Biotoxin Monitoring program for the water sample data. Our funding sources included The Conchologists of America, Mildred E. Mathias Foundation, Henry A. Jastro fel-lowship, UC Davis Graduate Group in Ecology, Bodega Marine Laboratory Fellowship, California SeaGrants R/FISH218, R/MPA14 and R/MPA24 awarded to Steven Morgan, NSF Biological Oceanography grants OCE- 1334448 and OCE- 0927196 awarded to Steven Morgan, and NSF GK- 12 grant 0841297 awarded to Susan Williams. S. A. Gravem and S. G. Morgan conceived the project. S. A. Gravem collected and analyzed the data and wrote the manuscript with significant contributions from S. G. Morgan. This is a contribution of the Bodega Marine Laboratory.

LITERATURE CITED

Bartl, S. 1980. A comparison of the feeding behavior of the six-rayed seastar, Leptasterias hexactis, and from two intertidal habitats. Unpublished student report, Problems in Marine Biology course. Bodega Marine Laboratory Cadet Hand Library, Bodega Bay, California, USA.

Bertness, M. D. 1977. Behavioral and ecological aspects of shore- level size gradients in Thais lamellosa and Thais emar-ginata. Ecology 58:86–97.

Bouchet, P., and G. Rosenberg. 2015. Tegula funebralis. In: MolluscaBase (2015). Accessed through: World Register of Marine Species at http://www.marinespecies.org/aphia.php?p=taxdetails&id=534190 on 2015-09-15

Bullock, T. H. 1953. Predator recognition and escape responses of some intertidal gastropods in presence of starfish. Behaviour 5:130–140.

Byers, B. A. 1983. Enzyme polymorphism associated with habi-tat choice in the intertidal snail Tegula funebralis. Behavior Genetics 13:65–75.

Byers, B., and J. Mitton. 1981. Habitat choice in the intertidal snail Tegula funebralis. Marine Biology 65:149–154.

Carlton, J. T., editor. 2007. The Light and Smith manual: inter-tidal invertebrates from Central California to Oregon. Fourth edition. University of California Press, Berkeley, California, USA.

Connell, J. H. 1972. Community interactions on marine rocky intertidal shores. Annual Review of Ecology and Systematics 3:169–192.

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs—High diversity of trees and corals is maintained only in a non- equilibrium state. Science 199:1302–1310.

Cooper, E. E., and A. L. Shanks. 2011. Latitude and coastline shape correlate with age- structure of Chlorostoma (Tegula) funebralis populations. Marine Ecology Progress Series 424:133–143.

Cushman, J. H. 1989. Vertical size gradients and migratory pat-terns of two Nerita species in the northern Gulf of California. Veliger 32:147–151.

Diamond, J. M. 1983. Ecology—laboratory, field and natural experiments. Nature 304:586–587.

Doering, P. H., and D. W. Phillips. 1983. Maintenance of the shore- level size gradient in the marine snail Tegula funebralis (Adams, A.): importance of behavioral responses to light and sea star predators. Journal of Experimental Marine Biology and Ecology 67:159–173.

Donahue, M. J., R. A. Desharnais, C. D. Robles, and P. Arriola. 2011. Mussel bed boundaries as dynamic equilibria: thresh-olds, phase shifts, and alternative states. American Naturalist 178:612–625.

Fawcett, M. H. 1984. Local and latitudinal variation in preda-tion on an herbivorous marine snail. Ecology 65:1214–1230.

Feder, H. M. 1963. Gastropod defensive responses and their effectiveness in reducing predation by starfishes. Ecology 44:505–512.

Flowers, J. M., and D. W. Foltz. 2001. Reconciling molecular systematics and traditional taxonomy in a species- rich clade of sea stars (Leptasterias subgenus hexasterias). Marine Biology 139:475–483.

Foltz, D. W. 1997. Hybridization frequency is negatively corre-lated with divergence time of mitochondrial DNA haplotypes in a sea star (Leptasterias spp.) species complex. Evolution 51:283–288.

Frank, P. 1975. Latitudinal variation in the life history features of the black turban snail Tegula funebralis (Prosobranchia: Trochidae). Marine Biology 31:181–192.

Gravem, S. A. 2015. Linking antipredator behavior of prey to intertidal zonation and community structure in rocky tide-pools. Dissertation. University of California, Davis, California, USA.

Gravem, S. A., and S. G. Morgan. 2016. Prey state alters trait- mediated indirect interactions in rocky tidepools. Functional Ecology 30:1574–1582.

Hewson, I., et al. 2014. Densovirus associated with seastar wasting disease and mass mortality. Proceedings of the National Academy of Sciences USA 111:17278–17283.

Hughes, T. P. 1994. Catastrophes, phase- shifts, and large- scale degradation of a Caribbean coral reef. Science 265: 1547–1551.

Hunt, H. L., and R. E. Scheibling. 1997. Role of early post- settlement mortality in recruitment of benthic marine inverte-brates. Marine Ecology Progress Series 155:269–301.

Jurgens, L. J., L. Rogers-Bennett, P. T. Raimondi, L. M. Schiebelhut, M. N. Dawson, R. K. Grosberg, and B. Gaylord. 2015. Patterns of mass mortality among rocky shore inverte-brates across 100 km of Northeastern Pacific Coastline. PLoS ONE 10:e0126280.

Kats, L. B., and L. M. Dill. 1998. The scent of death: chemosen-sory assessment of predation risk by prey animals. Ecoscience 5:361–394.

Kosin, D. 1964. The light responses of Tegula funebralis. Veliger 6:46–50.

Laundré, J. W., L. Hernandez, and W. J. Ripple. 2010. The landscape of fear: ecological implications of being afraid. Open Ecology Journal 3:1–7.

Lima, S. L. 1998. Stress and decision making under the risk of predation: recent developments from behavioral, reproduc-tive, and ecological perspectives. Advances in the Study of Behavior 27:215–290.


Lima, S. L., and P. A. Bednekoff. 1999. Temporal variation in danger drives antipredator behavior: the predation risk allo-cation hypothesis. American Naturalist 153:649–659.

Marchetti, K. E., and J. B. Geller. 1987. Effects of aggregation and habitat on desiccation and body temperature of the black turban snail, Tegula funebralis (A. Adams, 1855). Veliger 30:127–133.

Markowitz, D. V. 1980. Predator influence on shore- level size gradients in Tegula funebralis (A Adams). Journal of Experimental Marine Biology and Ecology 45:1–13.

McQuaid, C., J. Lindsay, and T. Lindsay. 2000. Interactive effects of wave exposure and tidal height on population structure of the mussel Perna perna. Marine Biology 137: 925–932.

Menge, B. A. 1972. Foraging strategy of a starfish in relation to actual prey availability and environmental predictability. Ecological Monographs 42:25–50.

Menge, B. A., C. Blanchette, P. Raimondi, T. Freidenburg, S. Gaines, J. Lubchenco, D. Lohse, G. Hudson, M. Foley, and J. Pamplin. 2004. Species interaction strength: testing model predictions along an upwelling gradient. Ecological Monographs 74:663–684.

Menge, B. A., E. B. Cerny-Chipman, A. Johnson, J. Sullivan, S. A. Gravem, and F. T. Chan. 2016. Ecology of the sea star wasting disease epidemic in the keystone predator Pisaster ochraceus: population impact, recovery potential, and preda-tion rate. PLoS ONE 11:e0153994.

Menge, B. A., and J. Lubchenco. 1981. Community organiza-tion in temperate and tropical rocky intertidal habitats: prey refuges in relation to consumer pressure gradients. Ecological Monographs 51:429–450.

Menge, J. L., and B. A. Menge. 1974. Role of resource alloca-tion, aggression and spatial heterogeneity in coexistence of two competing intertidal starfish. Ecological Monographs 44:189–209.

Mislan, K., B. Helmuth, and D. S. Wethey. 2014. Geographical variation in climatic sensitivity of intertidal mussel zonation. Global Ecology and Biogeography 23:744–756.

Morris, R. H., D. P. Abbott, and E. C. Haderlie. 1980. Intertidal invertebrates of California. Stanford University Press, Stanford, California, USA.

Paine, R. T. 1969. The Pisaster- Tegula interaction—prey patches, predator food preference and intertidal community structure. Ecology 50:950–961.

Paine, R. T. 1971. Energy flow in a natural population of herbivorous gastropod Tegula funebralis. Limnology and Oceanography 16:86–98.

Pardo, L. M., and L. E. Johnson. 2005. Explaining variation in life- history traits: growth rate, size, and fecundity in a marine snail across an environmental gradient lacking predators. Marine Ecology Progress Series 296:229–239.

Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual Review of Ecology Evolution and Systematics 37:637–669.

Perez, K. O., R. L. Carlson, M. J. Shulman, and J. C. Ellis. 2009. Why are intertidal snails rare in the subtidal? Predation, growth and the vertical distribution of Littorina littorea (L.)

in the Gulf of Maine. Journal of Experimental Marine Biology and Ecology 369:79–86.

Phillips, D. W. 1978. Chemical mediation of invertebrate defen-sive behaviors and ability to distinguish between foraging and inactive predators. Marine Biology 49:237–243.

Pincebourde, S., E. Sanford, and B. Helmuth. 2008. Body tem-perature during low tide alters the feeding performance of a top intertidal predator. Limnology and Oceanography 53:1562–1573.

Poloczanska, E. S., et al. 2013. Global imprint of climate change on marine life. Nature Climate Change 3:919–925.

Ripple, W. J., and R. L. Beschta. 2012. Trophic cascades in Yellowstone: the first 15 years after wolf reintroduction. Biological Conservation 145:205–213.

Robles, C., and R. Desharnais. 2002. History and current devel-opment of a paradigm of predation in rocky intertidal com-munities. Ecology 83:1521–1536.

Robles, C. D., R. A. Desharnais, C. Garza, M. J. Donahue, and C. A. Martinez. 2009. Complex equilibria in the maintenance of boundaries: experiments with mussel beds. Ecology 90:985–995.

Rochette, R., and L. M. Dill. 2000. Mortality, behavior and the effects of predators on the intertidal distribution of littorinid gastropods. Journal of Experimental Marine Biology and Ecology 253:165–191.

Rogers-Bennett, L., R. Kudela, K. Nielsen, A. Paquin, C. O’Kelly, G. Langlois, D. Crane, and J. Moore. 2012. Dinoflagellate bloom coincides with marine invertebrate mortalities in Northern California. Harmful Algae News 46:10–11.

Seed, R. 1969. Ecology of Mytilus edulis L. (Lamellibranchiata) on exposed rocky shores. II. Growth and mortality. Oecologia 3:317–350.

Terborgh, J., L. Lopez, P. Nunez, M. Rao, G. Shahabuddin, G. Orihuela, M. Riveros, R. Ascanio, G. H. Adler, and T. D. Lambert. 2001. Ecological meltdown in predator- free forest fragments. Science 294:1923–1926.

Underwood, A. J. 1984. Vertical and seasonal patterns in com-petition for microalgae between intertidal gastropods. Oecologia 64:211–222.

Underwood, A. J., and K. E. McFadyen. 1983. Ecology of the intertidal snail Littorina acutispira (Smith). Journal of Experimental Marine Biology and Ecology 66:169–197.

Vermeij, G. J. 1972. Intraspecific shore- level size gradients in intertidal molluscs. Ecology 53:693–700.

Wara, W. M., and B. B. Wright. 1964. The distribution and movement of Tegula funebralis in the intertidal region of Monterey Bay, California. Veliger 6:30–37.

Wright, W. G., and J. W. Nybakken. 2007. Effect of wave action on movement in the owl limpet, Lottia gigantea, in Santa Cruz, California. Bulletin of Marine Science 81:235–244.

Yarnall, J. L. 1964. The responses of Tegula funebralis to star-fishes and predatory snails (Mollusca: Gastropoda). Veliger 6:56–58.

Zaret, T. M., and R. Paine. 1973. Species introduction in a trop-ical lake. Science 182:449–455.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/ecy.1672/suppinfo

OpinionTransformative Research IsNot Easily PredictedSarah A. Gravem,1,* Silke M. Bachhuber,1

Heather K. Fulton-Bennett,1 Zachary H. Randell,1

Alissa J. Rickborn,1 Jenna M. Sullivan,1 and Bruce A. Menge1

Transformative research (TR) statements in scientific grant proposals havebecome mainstream. However, TR is defined as radically changing our under-standing of a concept, causing a paradigm shift, or opening new frontiers. Weargue that it is rarely possible to predict the transformative nature of research.Interviews and surveys of 78 transformative ecologists suggest that most TRbegan with incremental goals, while transformative potential was recognizedlater. Most respondents thought TR is unpredictable and should not be priori-tized over ‘incremental’ research that typically leads to breakthroughs. Impor-tantly, TR directives might encourage scientists to overstate the importance oftheir research. We recommend that granting agencies (i) allocate only a subsetof funds to TR and (ii) solicit more realistic proposal statements.

Transformative Research As an Unrealistic and Potentially Harmful GoalInternationally, granting agencies recently have required proposers and reviewers to identify howproposed research would be transformative. We argue that transformative research (TR; seeGlossary) is inherently unpredictable at the proposal phase and typically becomes recognizablelate in the scientific process. Further, it often arises from ‘incremental’ research (IR), which is[343_TD$DIFF]considered its opposite. This dichotomy was noted by Kuhn and Hawkins [1], who argued thatscientific knowledge proceeds incrementally, occasionally punctuated by paradigm-shiftingdiscoveries. Paradoxically, emphasizing TR actually can hinder scientific discovery. Underzero-sum funding, if breakthroughs result from incremental advances, reapportioning fundingin favor of putative TR actually can decrease breakthroughs. Further, prioritizing TR mightencourage scientists to overstate the importance of proposed research. Overhyping researchexpectationscould underminepublic trust in science, nowmore important thanever touphold [2].

As a caveat, we believe it is important to contemplate the potential of proposed research to begroundbreaking. However, we question whether the lofty definition of TR [318_TD$DIFF]as radically changingour understanding of a concept, causing a paradigm shift, or opening new frontiers ispredictable at the proposal phase, and thus question whether funding decisions shouldstrongly depend on TR statements.

Directives for TR in Grant ProposalsPrioritization of TR has become pervasive among granting agencies including the EuropeanResearch Council (ERC), Canadian Natural Sciences and Engineering Research Council(NSERC), Academy of Finland, US National Institutes of Health (NIH), US Department ofDefense (DoD), and US National Science Foundation (NSF). These agencies often use syn-onyms for TR such as ‘high-risk’, ‘pioneering’, or ‘breakthrough’ research. For example, theERC supports ‘pioneering proposals addressing new and emerging fields of research or

TrendsWe argue that transformative researchis inherently unpredictable.

Emphasizing transformative researchactually may hinder scientificdiscovery.

‘Do you think Bob Paine knew he wasbeing transformative when he startedripping sea stars off rocks? No!’.

Funding agencies should concentrateon the ‘goals’ of the research ratherthan the ‘outcome’.

1Oregon State University, IntegrativeBiology Department, 3029 CordleyHall, Corvallis, OR [314_TD$DIFF]97331, USA

*Correspondence:[email protected](S.A. Gravem).

TREE 2298 No. of Pages 10

Trends in Ecology & Evolution, Month Year, Vol. xx, No. yy http://dx.doi.org/10.1016/j.tree.2017.08.012 1© 2017 Elsevier Ltd. All rights reserved.


proposals introducing unconventional, innovative approaches and scientific inventions’ [3]. Onegoal of NSERC’s Discovery Frontiers grants is to ‘conduct transformative, paradigm-changingresearch’ [4].

The USNSF hasmade the pursuit of TR a top priority by asking for TR [319_TD$DIFF]statements in everymajorresearch proposal solicited. This is one of our own motivations for exploring the issue of TR inproposals, but we believe that our findings and recommendations are relevant internationally.Here, we use the NSF as an example of how TR became a focus of granting agencies. In 20 0 5,the US National Academy of Science [320_TD$DIFF](NAS) reported that the US lagged other countries inscientific advances [5]. This spurred the National Science Board (NSB) to propose ways thatNSF could nurture risky, potentially ‘transformative’ science [6]. NSB argued that ‘high-risk/high-impact’ research should be an important component of any funder portfolio. NSBrecommended ‘a new, distinct, and separate Foundation-wide program to solicit and supporttransformational . . . proposals’. While stressing the importance of TR, the NSB cautionedthat distinguishing TR from IR is often possible only in hindsight, and ‘did not see a need toadjust or to modify the current merit-reviewmechanism at NSF’. In response, NSF, which fundsapproximately 24% of US academic research (www.nsf.gov/about), required all proposers toaddress how each proposal was transformative [7]. The author guidelines were reworded (inbold), ‘How important is the proposed activity to advancing knowledge and understandingwithin its own field or across different fields? . . . To what extent does the proposed activitysuggest and explore creative, original, [321_TD$DIFF]or potentially transformative [322_TD$DIFF] concepts?’ [8].

A Tale of Two Transformative EcologistsThis opinion piece began with a conversation among the coauthors. Sarah Gravem hadsuggested that a planned experiment was insufficiently ‘transformative’. BruceMenge retorted,‘Do you think Bob Paine knew hewas being transformative when he started ripping sea stars offrocks? No! We won’t know if something is important until we test it.’ Menge was mentored bytwo iconic and transformative ecologists, Robert T. Paine and Joseph H. Connell. DuringMenge’s PhD, Paine developed the ‘keystone’ species concept [9,10 ], which experimentallydemonstrated top–down effects by predators in nature and challenged the paradigm thatcompetition drove species coexistence and diversity (e.g., [11–14]). His work on the keystonespecies concept influenced many fields of research throughout the sciences (Figure 1A; seeSupplemental Information 1 online for figure methods; see http://bit.ly/2uJCqA1 for an inter-active version). Nevertheless, the transformative impact of his work grew slowly (Figure 1B).

Did Paine predict the transformativeness of this research when he began his experiments?Withhis 20 16 passing, we will never know. Shortly beforehand, Paine commented to J. Estes(personal communication) that upon seeing the Washington coastal intertidal community, herecognized it was an ‘ecological gold mine’. However, he did not indicate he predicted hisresults. Paine later realized the implications of his result for other systems, but nonetheless thesignificance of his findings was long underappreciated.

In 1970 , Menge took a postdoctoral position with Joseph Connell and William Murdoch atUniversity of California Santa Barbara. Connell’s [15] paper on barnacle competition is along-standing textbook example and was perhaps the first demonstration of the power offield experiments. It is doubtful that Connell expected his research to be transformative.Connell’s interest was the topic of population regulation rather than competition. Whilewalking the intertidal at Millport, Scotland, he noticed that each footfall covered 10 0 [323_TD$DIFF]s ofbarnacles, far more than the 40 rabbits trapped in his entire Master’s research. He alsorealized that the tiny, immobile barnacles allowed manipulative experiments. Like Paine’swork, realization of the importance of Connell’s paper took time, but both studies wereclearly transformative.

GlossaryBreakthrough research: alsoknown as high-risk research. It istransformative, ambitious, and mold-breaking research. Its significancecan be based on tacklingexceptionally wide and complexresearch problems, on challengingexisting theories and scientificparadigms, on radically new ways ofusing methods, as well as onunprejudiced combination andinterdisciplinary integration ofdifferent research perspectives [39].Incremental research (IR):research building upon the results ofprevious studies or testing long-standing hypotheses and theories [6].Transformative research (TR):research driven by ideas that havethe potential to radically change ourunderstanding of an importantexisting scientific or engineeringconcept or leading to the creation ofa new paradigm or field of science orengineering. Such research also ischaracterized by its challenge tocurrent understanding or its pathwayto new frontiers (as defined by NSB[6]).

2 Trends in Ecology & Evolution, Month Year, Vol. xx, No. yy


(See figure legend on the bottom of the next page.)

Trends in Ecology & Evolution, Month Year, Vol. xx, No. yy 3


Our Approach: Investigating the Predictability of TRHere, we focus on TR in ecology, acknowledging that in this field TR can be especially rare, asPaine himself argued [16]. To investigate the usefulness of the transformative proposal criterionfor promoting scientific progress, we interviewed groundbreaking authors, and surveyedopinions of 72 highly cited ecologists. We asked: (i) When do researchers realize their researchis transformative? (ii) Can a researcher predict TR? (iii) Do scientists agree that TR statementsare useful metrics in judging proposals? (iv) Should TR be prioritized over IR? (v) Is prioritizingfunding for TR potentially harmful?

Case StudiesThe authors collectively identified six important ecological advances to be used as case studies.We interviewed the [324_TD$DIFF]authors about the process of doing TR, including Stephen R. Carpenter[17], Camille Parmesan [18], John W. Laundré [19], Joan A. Kleypas [20 ], William J. Ripple [21],and Daniel I. Bolnick [22] (see Supplemental Information 2 online for interview questions). Thesechoices represent studies that have transformed our personal understanding of ecology [325_TD$DIFF]but wedo not suggest that they are the six most transformative ecology studies. In addition, ourexamples are biased toward US ecologists because of our location, but we believe the resultsand our recommendations are relevant worldwide.

SurveyWe identified transformative ecologists using a Web of Science search of all records (1965–20 16) using the keyword ‘ecology’ from the following journals: Journal of Ecology, Ecology,Ecology Letters, Ecological Monographs, American Naturalist, Functional Ecology, Proceed-ings of the Royal Society of London – Biological Sciences, and Oecologia. We selectedpapers with 650 or more citations [326_TD$DIFF], excluding reviews and methods papers [327_TD$DIFF], and contacted thefirst authors. We also contacted first authors of papers in the Western Society of Naturalists’list of Top 10 0 influential papers in Ecology [23] [104_TD$DIFF]. 72 of the 110 authors contacted completedour survey (see Supplemental Information 3 online for more information). We acknowledgethat citation numbers are an imperfect way of indicating TR, but our goal was to contact agroup of influential authors for surveying purposes, not to exhaustively identify TR.

Interview ResultsIn general, the authors said that their transformative studies began as incremental investiga-tions whose importance they only appreciated later. Parmesan’s sentiment that ‘it was more amatter of being in the right place at the right time’ (Box 1) was echoed by many authors. Manystudies had minimal funding. For example, Laundré (see Supplemental Information 4 online)was supported by a small citizen science project, and Ripple (Box 2) used discretionary funds.Of the authors interviewed, only Carpenter had agency funding before his research on trophic

Figure 1. The Reach of Paine's Transformative Research. (A) Network visualization of journal articles from 1991 to20 16 that included the term ‘keystone species’ in the title, abstract, or keywords identified by Web of Science (keywordsand abstracts became available in 1991). Every node is an article and is sized by the number of citations per year. Nodesare linked if they share similar keywords. Groups of documents that are more linked formed 22 clusters of ‘keywordthemes’ indicated by color and labeled by keywords most commonly shared in the group. Clusters that contain 10 or morearticles are labeled. The broad representation from different scientific fields visually emphasize the cross-disciplinary reachof Paine’s transformative research. For example; certain clusters pertain specifically to the genesis of the keystone concept(e.g.; community structure; food web dynamics); while others in different fields of study (e.g.; microbiology; evolution; andgenetics) highlight the breadth of Paine’s work; and the network illustrates the reach of his transformative discovery. Seehttp://bit.ly/2uJCqA1 for an interactive version. (B) The citation rate per year in Web of Science of Paine’s 1966transformative paper [9] that experimentally demonstrated the keystone species effect. This shows that the transformativeimpact of the research began modestly and grew over time.For a Figure360 author presentation of Figure 1, see the figure online at http://dx.doi.org/10 .10 16/j.tree.20 17.0 8.0 12#mmc1



cascades in lakes (see Supplemental Information 5 online). The insightful moments also typicallyoccurred after the study was underway. A chance glance at a poster in the Yellowstone visitorcenter spurred Ripple’s ideas on terrestrial trophic cascades. Laundré noticed [328_TD$DIFF]during datacollection that despite abundant deer, few were eaten by mountain lions, launching the ideaof the ‘ecology of fear’ (see Supplemental Information 4 online). Several were unaware of thetransformative potential of their work until the analysis, or even until the publication stage. Forexample, Kleypas stated that the impact of her ocean acidification research ‘didn’t hit [her] until[she] had put together the results of the study’ (see Supplemental Information 6 online). However,despite publishingwhatmany consider TR, none of the six authors believed that scientists can bedeliberately transformative (Boxes 1 and 2; see Supplemental Information 4–7 online).

Survey ResponsesWhen Do Researchers Typically Realize Their Research Was Transformative?Typically, researchers we surveyed realized the transformative nature of their work during laterresearch stages, including data analysis (18.1% of the time), writing (26.4%), and

Box 1. Case Study: Dr Camille Parmesan’s Work on Species’ Responses to Climate ChangeDr Parmesan is a professor at the University of Plymouth and an Intergovernmental Panel on Climate Change (IPCC)contributor. Her early work focused on the ecology of a butterfly species (Euphydryas editha) in the AmericanSouthwest. Initially, she examined range shifts of a single butterfly species, expecting to find fluctuations due onlyto local extinctions and extensions among specific ecotypes. Instead, she found the now well-known shifts northwardand toward higher elevations in the species’ range [18]. During this time, climate scientists were examining the trend inrising global temperatures, but had not yet made connections to how this would affect organisms. Parmesan’s 1999paper [334_TD$DIFF][40 ], funded by a postdoctoral position at [335_TD$DIFF]the National Center for Ecological Analysis and Synthesis (NCEAS),found that the concerted movement of European butterfly species northward closely correlated with the observednorthward shift in continental isotherms. This paper caught the attention of both the scientific community and policymakers, and finally convinced many that the effects of climate change were nontrivial. A later global meta-analysis [336_TD$DIFF][41]provided themost conclusive evidence for northward species range shifts. These papers were instrumental in spawningthe field of global change biology that has yielded thousands of studies to date. For her part, Dr Parmesan did not believethat researchers can be deliberately transformative; it is simply a matter of being in the right place at the right time.

Box 2. Case Study: Dr William Ripple’s Work on Predators Initiating Trophic CascadesIn the early 1990 s, discussion regarding the role of herbivory and predation in controlling communities was ongoingfollowing the conceptual framework outlined by Hairston and colleagues’ Green World Hypothesis [337_TD$DIFF][42]. DespiteMcLaren and Peterson’s [338_TD$DIFF][43] early work on Isle Royale, trophic cascades were thought to be nonexistent or veryweak in terrestrial compared with aquatic systems [339_TD$DIFF][44]. In their 20 0 0 paper [21], William Ripple and Eric Larsen explicitlydescribed trophic interactions as a trophic cascade in a terrestrial system, and inspired a surge of research acrossdisciplines [340_TD$DIFF][45]. Using university travel funds, they visited Yellowstone in 1998 to investigate whether ongoing recruit-ment failure of aspen since the 1930 s was correlated with climate change, fire suppression, or an increase in browsingelk populations following relaxation of hunting pressure. The potential role of apex predators in regulating aspenrecruitment was considered only after Ripple saw a photograph of a wolf standing among young aspen in the visitor’scenter. Tree core data revealed that the truncated age structure of aspens in the park was tightly correlated with theextirpation of wolves from the park by 1926 [21]. The paper was published inBiological Conservation only after extensiverevisions. The paper received widespread attention among scientists and the public, as Yellowstone is a highly valuedarea with many invested stakeholders.

Follow-up research demonstrated that wolf reintroduction in 1995 led to a decrease in elk browsing and a subsequentincrease in aspen recruitment and canopy cover. Cascading effects of large carnivores on herbivore density and plantrecruitment have now been demonstrated in multiple terrestrial ecosystems [341_TD$DIFF][46]. Research on trophic downgrading (theanthropogenic extirpation of large carnivores) and its effects on overall ecosystem health, function, and diversity is now amajor interdisciplinary research focus [342_TD$DIFF][47].

Ripple said that his initial observations were conducted at a serendipitous time, and that the availability of other data setson predator abundance and tree recruitment in the park dating back to the early 20 th century helped connect hisobservations. During his initial, nearly unfunded, fieldwork in Yellowstone, Ripple did not expect to spur almost twodecades of intensive research on terrestrial trophic cascades. While Ripple believes that the transformative nature ofresearch cannot be predicted, he does believe that it is possible to increase your chances and that ‘asking big questionscan get you big answers [311_TD$DIFF]’.



Figure 2. [307_TD$DIFF][309_TD$DIFF]Responses by 72 Highly Cited Ecologists to an Email Survey Regarding Their Views onTransformative Science and Its Role in Funding Decisions. (A) Survey responses to the question ‘During whichphasedid the transformative potential for your research become solidified?’by highly cited ecologists. Numbers indicate sumof the responses. Only four identified the transformative potential of their research during the proposal stage, whereas 20realized it during data collection and analyses and 34 realized it during writing, publication, or postpublication. (B–H) Likert-scale responses by transformative ecologists to survey questions on transformative and incremental research (TR and IR,respectively). Neutral responses (gray) are centered in the figure,with frequency of strongly disagree anddisagree to the left ofcenter (dark red and red, respectively) and of agree and strongly agree (blue and dark blue, respectively) to the right of center.Ecologists generally disagree with the statements ‘I anticipated transformative research contributions in proposals’ and‘researchers can predict transformative research during the proposal phase’. They generally agreed that ‘transformativeresearch isappropriate for judgingproposals’but therewasnoconsensusonwhether ‘reviewerscandetermine if aproposal ispotentially transformative research’. Ecologistsgenerally strongly disagree that ‘incremental research shouldbe lower fundingpriority than transformative research’ and strongly agree that the ‘Transformative research criterion causes overstatingimportance of research’. Moving forward, ecologists think that the ‘transformative research criterion should be limited totargeted grants’. Full text of questions and counts are included in Supplemental Information 3 online.For a Figure360 author presentation of Figure 2, see the figure online at http://dx.doi.org/10 .10 16/j.tree.20 17.0 8.0 12#mmc2



postpublication (19.4%), but rarely at the proposal stage (5.6%; Figure 2A and SupplementalInformation 3 online). Similarly, the majority (61%) had not anticipated their contributions toecological progress during the proposal phase, though some did (23%; Figure 2B andSupplemental Information 3 online).

Can a Researcher Predict If His or Her Research Will Be Transformative?On average, survey respondents ‘somewhat disagreed’ that TR is predictable at the proposalphase (Figure 2C and Supplemental Information 3 online). One respondent asserted thatpredicting TR was impossible: ‘The most transformative ideas will not be ones predicted inproposals. If you can predict them, they are unlikely to be transformative. The most transfor-mative ideas are the surprises in your data or observations. Surprises are by definition notpredictable.’ This sentiment was actually acknowledged by the NSB and NSF: NSB suggestedthat identifying TR is only possible in hindsight [6], and the then NSF Director Arden Bementnoted ‘Most transformations resulting from research are recognized post hoc, not a priori’ [24].Similarly, Trevors et al. [25] commented ‘Transformative research is often elusive, requiresdifferent approaches and sometimes depends on luck. [It] concerns intangibles such as humanintuition, curiosity, serendipity, unpredictable events, correcting previous knowledge, implau-sible hypotheses, a well-prepared mind and often interpersonal communication.’

Are TR Statements Useful in Reviewing Proposals?Although most respondents disagreed that researchers can predict TR in proposals and mostdid not predict their own TR (Figure 2A–C and Supplemental Information 3 online), respondents‘somewhat agreed’ that the TR statements are appropriate for judging ecological proposals(Figure 2D and Supplemental Information 3 online). One respondent, an active proposer andgrant reviewer, said ‘I agree that ideally we need funding for both IR and TR proposals. But thereality is that research funding is quite limited. Given these constraints, should priority be givento research that has the potential to be transformative? I think a strong case can be made thatthis is one valid criterion.’

Thus, we have a conundrum. Most researchers feel that they cannot predict TR in proposals,but also feel TR statements are useful in the funding process. But can reviewers reliablyevaluate transformative potential? No consensus emerged on this issue (Figure 2E andSupplemental Information 3 online). Concern about reviewer abilities to judge potentiallytransformative proposals was one of the original reasons that TR statements were recom-mended [6]. However, Hossenfelder [26] argued that encouraging reviewers to favor potentiallyTR ‘will not work very well . . . If you random[ly] sample [reviewers], you’re more likely to getconservative opinions just because they’re more common. As a result, [TR] projects are unlikelyto be reviewed favorably. [Reviewers] still have to evaluate if the high risk justifies the potentialhigh payoff. And assessment of tolerable risk is subjective.’

Should TR Funding Be Prioritized over Incremental Research Funding?Transformative requirements can be detrimental to the IR from which many major paradigm-shifting ideas materialize. By placing emphasis on TR, funding agencies might stifle the IRneeded to lay the groundwork required to facilitate proposals that can define a new field orchallenge major paradigms. Our survey revealed broad sentiment among ecologists that IRshould not be a lower priority than TR (Figure 2F and Supplemental Information 3 online). Forexample, one respondent commented ‘thousands of us make incremental steps, . . . thatcollectively move the dial on our science . . . I believe that emphasis on novelty and [TR], whilewell intended, has neutral or more likely negative effects on our path forward. It creates a falsedichotomy between “novel” and “incremental” science.’



[329_TD$DIFF]Incremental research is often viewed as a loaded term. It implies that studies that do not shiftparadigms, redefine a field, or create new technology are less important, or at least less worthyof funding. For example, former NSF director Bement [24] commented ‘if it’s “safe science,”NSF should not fund it.’However, most paradigm shifts would not be possible without a wealthof background knowledge and field-defining research. Kuhn and Hawkins [1] suggested thatmost science was incremental, or what they called ‘normal’. Many agree that incrementaladvances are often prerequisites of TR [6,27–30 ]. The necessity of IR was even featured onFreakonomics Radio [28], where economist Ed Gleaser pointed out that Nobel Prizes are nottypically given for single TR papers. Rather, they are often given for a body of IR, oftenrepresented by dozens of papers, by a particular person. If in fact transformations arise fromIR, then the transformative criterion is actually redundant with the solicitation of IR [31]! This isreflected by [344_TD$DIFF]mixed evidence that soliciting TR led to [345_TD$DIFF]increases in transformative outcomescompared with the typical model [346_TD$DIFF][33].

Is Prioritizing Funding for TR Potentially Harmful?Our results suggest that predicting TR at the proposal stage is not only difficult, but also thatprioritizing it can be potentially harmful. Survey-takers expressed strong consensus thatemphasizing the importance of TR in proposals causes researchers to overstate the potentialimportance of their work (Figure 2G and Supplemental Information 3 online). With limitedfunding, incentivizing ‘promise inflation’ can be harmful to the ‘honest scientists who . . .propos[e] to [make] an important contribution to the solution of an important problem. They riskbeing dismissed as small-timers with no vision’ [27]. One respondent commented that trans-formative statements can be more of an ‘essay contest’ than truthfully representing the likelyimpact of proposed research. Survey results echo this idea. Respondent Dr Lauri Oksanenstated, ‘If folks are requested to emphasize the transformative aspects of a proposal, the likelyresult is self-bragging and a soup of fashionable terms.’ Holbrook [31], an organizer of an NSFworkshop on TR, expressed the opinion that grant writing has turned into a game where‘transformative’ is a buzzword not taken seriously by proposers, reviewers, or authors.

The transformative requirement might also be especially detrimental to ecology relative to otherfields. One respondent remarked, it ‘is certainly a death-knell for long-term research projects.’Unfortunately, successful long-term ecological research requires lengthy periods of datacollection, which can lead to their categorization as ‘incremental’. For example, Hubbell’s[346_TD$DIFF][33] Unified Neutral Theory of Biodiversity and Biogeography was an unanticipated buttransformative consequence of long-termmonitoring and mapping of tropical trees in Panama.This is especially concerning because long-term ecological research contributes relatively moreto ecological knowledge and to public policy [347_TD$DIFF][34].

Looking AheadRecommendation [330_TD$DIFF](i): Allocate Only a Subset of Grant Funds to Potentially TransformativeResearchThis strategy has been adopted by several granting agencies (e.g., ERC, NSERC, Academy ofFinland, US NIH, and DoD). Targeted allocation of funding for TR, in our view, would bothencourage TR and support IR, leading to transformations. Further, survey-takers agreed thatthe TR requirement should be limited to only a subset of funds (Figure 2H, see SupplementalInformation 3 online). Such a strategy would not significantly affect funding for typical incre-mentally focused proposals, would encourage long-term research, and would not incentivizegrant writers to inflate the importance of their research. This last benefit is important; public trustin science and scientists has waned [348_TD$DIFF][35], in part due to antiscience polarization tactics [349_TD$DIFF][36,37].In this atmosphere, overhyping the likely importance of one’s research can be damaging. Thisrecommendation is particularly relevant for the NSF because the TR requirement is ubiquitous



among NSF-funded programs. The NAS, NSB, and others originally recommended that theNSF allocate a fraction of the agency research budget to TR [5,6,27]. Since NSF implementedthe ubiquitous TR requirement with ‘constructive ambiguity’, intending to assess its efficacy inthe future [6,24,29], we believe the time for reassessment has come.

Recommendation [331_TD$DIFF](ii): Solicit Realistic Proposal Statements Regarding Research GoalsWe do not criticize the goal of fostering TR itself, but feel identifying it is unrealistic for mostproposals. As our results suggest, and as granting agencies acknowledge [350_TD$DIFF][6,25,38,39],predicting TR is not usually possible. Thus, we suggest that granting agencies should adopta broader definition of TR or use a different term that is more realistic. Currently, proposalsolicitations of NSF, and to some extent ERC, focus on the potential transformative ‘outcomes’of all proposed research, which seem inherently unknowable (e.g., ‘changing’ our understand-ing, ‘creating’ paradigms or fields, ‘pathways’ to new frontiers, scientific ‘inventions’). Thiscreates a straw man for reviewers to criticize as being either too far-fetched or too uncertain tofund. Rather, we suggest that funding agencies concentrate on the ‘goals’ of the researchrather than the ‘outcome’. For example, the Academy of Finland uses the term breakthroughresearch [351_TD$DIFF][39]. Breakthrough research is rooted in the goals of ‘tackling’ problems, ‘challeng-ing’ theories and paradigms, ‘using’methods, and ‘integrating’ perspectives. We suggest thatjudging proposals by their goals rather than their potential outcomes is both realistic and easierfor reviewers to evaluate.

Concluding RemarksWe suggest (i) TR is usually realized in later stages of research; (ii) researchers cannot predictTR, especially in proposals; (iii) TR predictions are not useful in judging proposals; (iv) TR fundingshould not detract from IR funding; and (v) requiring TR statements incentivizes disingenuous-ness and can stifle certain types of research. Though we wholeheartedly agree that proposalsshould explore the novelty, utility, or advanced contributions the research can provide, werecommend that only a subset of solicitations request TR statements while most should solicitmore realistic statements of research goals. We believe that our findings and recommendationsare relevant for granting agencies worldwide as they determine how to allocate funding amongdifferent research endeavors (see [332_TD$DIFF]Box 3).

AcknowledgmentsWe would like to thank Drs Eric Berlow and Richard Williams for creating the network figure. We thank Daniel Bolnick,

Stephen Carpenter, Joanie Kleypas, John Laundré, Camille Parmesan, and Bill Ripple for insightful interviews. We thank

the 72 authors who responded to our survey, especially those who responded with thoughtful comments. This unfunded

study was performed in compliance with Internal Review Board at Oregon State University.

Supplemental InformationSupplemental information associated with this article can be found, in the online version, at http://dx.doi.org/10 .10 16/j.

tree.20 17.0 8.0 12.

References1. Kuhn, T.S. and Hawkins, D. (1963) The structure of scientific

revolutions. Am. J. Phys. 31, 554–555

2. Lubchenco, J. (20 17) Environmental science in a post-truthworld. Front. Ecol. Environ. 15, 3

3. European Research Council (20 17) ERCWork Programme 2018.

4. Natural Sciences and Engineering Research Council (20 17) Dis-covery Frontiers Call for Proposals: Biodiversity and Adaptation ofBiosystems.

5. National Academies (20 0 7) Rising above the Gathering Storm:Energizing and Employing America for a Brighter EconomicFuture, National Academies Press

Box 3. Recommendations for Granting AgenciesAllocate only a subset of grant funds to potentially transformative research.

Solicit realistic proposal statements regarding research goals[313_TD$DIFF].



6. National Science Board (20 0 7) Enhancing Support of Transfor-mative Research at the National Science Foundation, pp. 1–23,National Science Board

7. National Science Foundation (20 0 7) Important Notice No. 130:Transformative Research, National Science Foundation

8. National Science Foundation (20 16) Proposal and Award Policiesand Procedures Guide: Part 1 – Grant Proposal Guide, NationalScience Foundation

9. Paine, R.T. (1966) Food web complexity and species diversity.Am. Nat. 10 0 , 65–75

10 . Paine, R.T. (1969) A note on trophic complexity and communitystability. Am. Nat. 10 3, 91–93

11. Park, T. (1948) Interspecies competition in populations of Tri-lobium confusum (Duval) and Trilobium castaneum (Herbst). Ecol.Monogr. 18, 265–30 7

12. Hutchinson, G.E. (1959) Homage to Santa Rosalia or why arethere so many kinds of animals? Am. Nat. 93, 145–159

13. MacArthur, R.H. (1972) Geographical Ecology: Patterns in theDistribution of Species, Princeton University Press

14. MacArthur, R.H. (1958) Population ecology of some warblers ofnortheastern coniferous forests. Ecology 39, 599–619

15. Connell, J.H. (1961) The influence of interspecific competition andother factors on the distribution of the barnacle Chthamalusstellatus. Ecology 42, 710 –723

16. Paine, R.T. (20 0 2) Advances in ecological understanding: byKuhnian revolution or conceptual revolution? Ecology 83,1553–1559

17. Carpenter, S.R. et al. (1987) Regulation of lake primary produc-tivity by food web structure. Ecology 68, 1863–1876

18. Parmesan, C. (1996) Climate and species’ range. Nature 382,765–765

19. Laundré, J.W. et al. (20 0 1) Wolves, elk, and bison: reestablishingthe “landscape of fear” in Yellowstone National Park, U.S.A. Can.J. Zool. 79, 140 1–140 9

20 . Kleypas, J.A. (1999) Geochemical consequences of increasedatmospheric carbon dioxide on coral reefs. Science 284, 118–120

21. Ripple, W.J. and Larsen, E.J. (20 0 0 ) Historic aspen recruitment,elk, and wolves in northern Yellowstone National Park, USA. Biol.Conserv. 95, 361–370

22. Bolnick, D.I. et al. (20 0 3) The ecology of individuals: incidence andimplications of individual specialization. Am. Nat. 161, 1–28

23. Stachowicz, J.J. (20 16)WSN 2016 Presidential Address and Top100 Papers.

24. Bement, A.L. (20 0 7) Transformative Research: The Artistry andAlchemy of the 21st Century. In Texas Academy of Medicine,Engineering and Science Fourth Annual Conference. TexasAcademy of Medicine In: https://www.nsf.gov/news/speeches/bement/0 7/alb0 70 10 4_texas.jsp

25. Trevors, J.T. et al. (20 12) Transformative research: definitions,approaches and consequences. Theory Biosci. 131, 117–123

26. Hossenfelder, S. (20 12) What Is Transformative Research andWhy Do We Need it? Back Reaction.

27. Meyer, B. (20 11) Long live incremental research! . In Communi-cations of the Association for Computing Machinery.

28. Dubner, S.J. (20 16) In Praise of Incrementalism, FreakonomicsRadio. Produced by Christopher Werth. Distributed by Freako-nomics, LLC

29. Holbrook, J.B. et al. (20 12) Good transformations: Ambiguity andthe NSF’s experiment with ‘transformative’ research. Sci.Progress

30 . President’s Council of Advisors on Science and Technology(20 12) Transformation and Opportunity: The Future of the USResearch Enterprise, pp. 1–124, President’s Council of Advisorson Science and Technology

31. Holbrook, J.B. (20 15) Transformative research. In Ethics, Sci-ence, Technology, and Engineering: A Global Resource (2nd edn,Vol. 4) Holbrook, J.B. and Mitcham, C., eds, pp. 40 8–410 ,Macmillan

33. Hubbell, S.P. (20 0 1) The Unified Neutral Theory of Biodiversityand Biogeography, Princeton University Press

34. Hughes, B.B. et al. (20 17) Long-term studies contribute dispro-portionately to ecology and policy. Bioscience 67, 271–281

35. Gauchat, G. (20 12) Politicization of science in the public sphere: Astudy of public trust in the United States, 1974 to 20 10 . Am. Soc.Rev. 77, 167–187

36. Hmielowski, J.D. et al. (20 13) An attack on science? Media use,trust in scientists, and perceptions of global warming. PublicUnderst. Sci. 23, 866–883

37. Farrell, J. (20 15) Corporate funding and ideological polarizationabout climate change. Proc. Natl. Acad. Sci. U. S. A. 113, 92–97

38. National Science Foundation (20 17) Challenges of IdentifyingPotentially Transformative Research. In: https://www.nsf.gov/about/transformative_research/challenges.jsp

39. Häyrynen, M. (20 0 7) Breakthrough Research. Funding of High-Risk Research at the Academy of Finland, Academy of Finland

40 . Parmesan, C. et al. (1999) Poleward shifts in geographical rangesof butterfly species associated with regional warming. Nature399, 579–583

41. Parmesan, C. and Yohe, G. (20 0 3) A globally coherent fingerprintof climate change impacts across natural systems. Nature 421,37–42

42. Hairston, N.G. et al. (1960 ) Community structure, populationcontrol, and competition. Am. Nat. 94, 421–425

43. McLaren, B.E. and Peterson, R.O. (1994) Wolves, moose, andtree rings on Isle Royale. Science 266, 1555–1558

44. Strong, D.R. (1992) Are trophic cascades all wet? Differentiationand donor-control in speciose ecosystems. Ecology 73, 747–754

45. Ripple, W.J. et al. (20 16) What is a trophic cascade? Trends Ecol.Evol. 31, 842–849

46. Ripple, W.J. et al. (20 14) Status and ecological effects of theworld’s largest carnivores. Science 343, 151–163

47. Estes, J.A. et al. (20 11) Trophic downgrading of planet earth.Science 333, 30 1–30 6


Marine Ecology. 2017;38:e12418. wileyonlinelibrary.com/journal/maec | 1 of 10https://doi.org/10.1111/maec.12418

© 2017 Blackwell Verlag GmbH

Accepted: 23 January 2017

DOI: 10.1111/maec.12418

O R I G I N A L A R T I C L E

Underwater video reveals decreased activity of rocky intertidal snails during high tides and cooler days

Austin W. Taylor1,2 | Steven G. Morgan1,2 | Sarah A. Gravem1,2,3

1Bodega Marine Laboratory, University of California Davis, Bodega Bay, CA, USA2Department of Environmental Science and Policy, University California Davis, Davis, CA, USA3Integrative Biology Department, Oregon State University, Corvallis, OR, USA

CorrespondenceSarah A. Gravem, Integrative Biology Department, Oregon State University, Corvallis, OR, USA.Email: [email protected]

Funding informationDivision of Ocean Sciences, Grant/Award Number: OCE-0927196 and OCE-1334448; Division of Graduate Education, Grant/Award Number: 0841297; California Sea Grant, University of California, Grant/Award Number: R/FISH218, R/MPA14 and R/MPA24; National Science Foundation Biological Oceanography; National Science Foundation

AbstractNearly all of our understanding of rocky intertidal ecology comes from studies con-ducted at low tide. To study intertidal organisms at high tide, we anchored water-proof digital GoPro® video cameras in waveexposed tidepools and recorded the daytime movements of the black turban snail, Tegula funebralis, over the tidal cycle between May and August 2012 near Bodega Bay, California. Overall, snails moved more quickly and presumably foraged more during low tides and on days with warmer air and perhaps water temperatures. This is similar to other ectotherms that exhibit increased metabolic rates, movement and foraging in warmer conditions. Snails also moved less during flood and high tides, may have moved downward in tidepools at flood tides, and showed evidence of reduced activity on days with larger waves. This inactivity and refuge seeking may have been a strategy to avoid dislodgment by waves. Analysis of snail trajectories showed foraging bouts indicated by alternating zig- zagging and straight movement. There was no effect of temperature, wave height, or tidal phase on distribution of snail turning angles, suggesting that they may have for-aged consistently but moved faster during warm conditions and low tides, thereby grazing a larger area. This is one of few direct recordings of intertidal organisms on waveexposed rocky shores during high tide. The methods used here are easily trans-ferable to other studies, which are needed to increase our understanding of behaviors that structure rocky shore communities during high tide.

K E Y W O R D S

animal movement, GoPro, rocky inter-tidal, tidal cycle, underwater video, wave exposure

1 | INTRODUCTION

Exposed rocky intertidal shores have long been a fruitful system for ecological studies, contributing substantially to ecological theory. However, nearly all observations on these wave- swept shores are conducted during low tide because shores are inaccessible during high tide. Our limited understanding of organisms during high tide has come primarily from indirect rather than direct observations. Such studies include trapping animals during high tide (Hunter & Naylor, 1993; Silva, Hawkins, Boaventura, Brewster, & Thompson, 2010), comparing changes between successive low tides (Chapperon & Seuront, 2009;

Doering & Phillips, 1983; Pardo & Johnson, 2006), observing endoge-nous rhythms and other behaviors during constant conditions or sim-ulated high tides in the laboratory (Morgan, 2007; Vieira et al., 2012), or simulating flow in the laboratory (Denny, 1989; Martone & Denny, 2008). Direct observations in waveexposed environments at high tide are few and have been accomplished by suspending a researcher or video camera above incoming waves (Wara & Wright, 1964; Wright & Nybakken, 2007), by mounting a video camera in the intertidal (Burrows, Kawai, & Hughes, 1999; Miller, 2007; Robles, Alvarado, & Desharnais, 2001), and by listening to and recording rasps of gastro-pod radulae on surfaces (Kitting, 1979).

2 of 10 | TAYLOR eT AL.

Timing of foraging by intertidal animals depends on their responses to environmental stresses at low tide, their resistance to dislodgement by waves at high tide and predation pressures that may change over the diel or tidal cycle. During high tide, incoming waves typically reduce activity of many sedentary invertebrates, such as seastars, sea urchins, some snails and some limpets, (Dayton, 1985; Menge, 1978; Miller, 1974; Pardo & Johnson, 2006; Wright & Nybakken, 2007). Other more mobile or firmly attached species typically increase activity during high tide, such as hermit crabs, brachyuran crabs, lobsters, fishes, mussels, some snails and some lim-pets (Hunter & Naylor, 1993; Morgan, 2007; Palmer, 1974; Robles et al., 2001; Silva et al., 2010). During low tide, intertidal inverte-brates inhabiting exposed rock surfaces generally avoid desiccation and thermal stress by seeking shelter and decreasing activity (Garrity, 1984; Jones & Boulding, 1999; Marchetti & Geller, 1987). These con-ditions are less stressful in tidepools, so animals may actually utilize these wave- free periods to forage. The tidal cycle is superimposed onto the diel cycle, which exposes animals to the sun and warmth during the day and colder conditions at night. Further, daytime low tides expose organisms to visual aerial predators (e.g. birds), day-time high tides expose them to visual aquatic predators (e.g. fish), and nighttime exposes them to nocturnal predators at high and low tides (e.g. crabs and octopi). Thus, intertidal invertebrates should adjust their activities cyclically to minimize physiological stress, dis-lodgement and predation. However nearly all intertidal studies are conducted during daytime low tides, which is just one phase of this complex dual cycle.

To remedy one aspect of this gap in knowledge, we investigated the behavior of the black turban snail, Tegula funebralis (formerly Chlorostoma funebralis Bouchet & Rosenberg, 2015), during the day-time at low, flood and high tide. We conducted our investigations in Horseshoe Cove, which is adjacent to the Bodega Marine Laboratory inNorthernCalifornia,USA(38°19ʹ00.6ʺN,123°04ʹ15.8ʺW).Tegula funebralis occurs in high densities in the intertidal zone of rocky shores along the west of North America from Vancouver Island to Baja California where it grazes on microalgae and macroalgae (Ricketts, Calvin, Hedgpeth, & Phillips, 1985). Tegula funebralis is very tolerant of thermal stress and desiccation at low tide (Tomanek & Somero, 1999). The only direct observations of T. funebralis at high tide revealed that densities were greatest atop rocks at night and they declined in bright moonlight and large swell (Wara & Wright, 1964), and no observations of snails inhabiting tidepools throughout the tidal cycle have been documented.

To assess snail activity throughout the tidal cycle, we designed and fabricated two underwater camera mounts to deploy a digital under-water video camera in tidepools during low, flood and high tides. We used a nearby weather station to assess how temperature (air and seawater) and wave height affected snail activity during our camera deployments. We also deployed a depth and temperature logger in the tidepools (although not during camera deployment) for days at a time to measure the variation in water temperature in the tidepool microhabitats and to assess probable wave exposure over the tidal cycle.

2 | MATERIAL AND METHODS

2.1 | Video recordings

We used a GoPro® Hero 2 (www.GoPro.com) high- definition digital video camera to capture changes in behavior of the common herbivo-rous snail T. funebralis with tidal phase or daily weather conditions. We filmed snails in rocky intertidal tidepools during low, flood and high tides 10 times on nine dates between 25 May and 6 August 2012. Due to limited camera battery life (~4 hr), we were unable to capture ebb tides. We continuously recorded snail movements in three tidepools (6.402, 0.662 and 0.260 m2 surface area) during the daytime as nighttime illumination may have altered snail behavior. Shore levels of the three tidepools were 1.37, 1.37 and 1.14 m above mean lower low water, which we determined with surveying equip-ment and United States Geological Survey geodetic benchmarks. A total of 29.5 hr of recordings was taken, ranging from 73 to 212 min and averaging 180 min. We tracked 198 snails, with 10–29 snails observed during each deployment. An example video at 20× speed is included in Appendix S1.

The GoPro digital video camera is affordable, rugged and versa-tile making it ideal for deployments in the waveexposed rocky intertidal environment. We placed the camera in the stock dive housing and one modified using the Snake River Prototyping BlurFix Lens kit (http://www.snakeriverprototyping.com). Desiccant strips were installed in the housing to prevent condensation from forming while recording. The camera has a 170° wide- angle lens, 11- megapixel res-olution and a small image sensor, which creates a large depth of field, causing all objects to be in focus. We recorded high- definition video (1,920 × 1,080 pixels) and used the optional GoPro battery Bacpac (www.GoPro.com) to extend battery life.

Depending on the topography of the tidepool, we deployed cam-eras in two ways to film snail activity on vertical rock surfaces inside tidepools just below the waterline (see Appendix S2 and Figures S1–S4 therein for design and fabrication of both camera mounts). The camera was deployed above one steep- sided tidepool using an inverted tripod design (Figure 1a). The tripod mount could be raised or lowered and oriented at any angle, making it more versatile for deep or irregular tidepools. The other two tidepools were larger and had flat surfaces, so we used a simpler detachable stainless steel plate mount to fasten the camera to the bottom of each tidepool (Figure 1b). Both camera mounts had to withstand direct impacts by large waves on the windy, waveexposed Northern California coast. The camera was situ-ated 0.25–1 m from the target (Figure 1c), although longer or shorter distances could be used depending on the size of organisms and clarity of the water. An object of known size was placed on the rock surface in the center of the field of view at the start of each video to standardize our distance measurements during video analysis.

2.2 | Image analyses

We used FINAL CUT PRO X software (Apple Inc., http://www.apple.com/final-cut-pro) to increase the video speed 20× and convert videos

| 3 of 10TAYLOR eT AL.

to image sequences. Then we imported images at 20- s intervals into IMAGEJ software as a stack (Rasband, 1997-2017). To track snail movements, we assigned numbers to each individual snail and manu-ally tracked them between frames using the Manual Tracking plugin for IMAGEJ (https://imagej.nih.gov/ij/plugins/track/track.html) to determine their xy co- ordinates in cm. As snails were very abun-dant and we observed that handling them increased their movement (Gravem & Morgan, 2016), we decided not to mark individuals. Thus

it is possible that some snails may have recurred on different days. However, this was likely rare because they were abundant, mobile and videos were taken days to weeks apart. To quantify overall speed and any movement upward or downward on shore, we calculated snail velocity and vertical velocity in cm/min for each 20- s interval as: veloc-ity=3×√[(Yt–Yt-1)2 + (Xt–Xt-1)2]; vertical velocity = 3 × (Yt–Yt-1). To document zig- zagging and other feeding behaviors (Norton, Hawkins, Manley, Williams, & Watson, 1990), we calculated turning angle for each 20- s frame using R software (R Core Team, 2013) and the ltraj function in the ADEHABITATLT package (Calenge, 2006). Snail turn-ing angle may indicate whether snails were foraging as indicated by zig- zagging or in transit as indicated by straight trajectories (Swingland & Greenwood, 1983). Turning angle was measured and graphed as positive for clockwise turns and negative for counter- clockwise turns. For analyses we used the absolute value of the turning angle so that low numbers were straight trajectories and large numbers were sharply angled trajectories. We noted only one occurrence of snail dislodgement. Foraging on drift algae would cause snails to stop mov-ing while feeding and alter our interpretation of movement data, but this was not observed.

2.3 | Stresses over the tidal cycle

We assigned each video frame to low, flood and high tidal phases by first determining what time the tide should cover each tidepool based on its shore level using tide tables at 1- min resolution (http://tbone.biol.sc.edu/tide). This was the “predicted submergence”. From the animal’s point of view, flood tide begins before predicted submer-gence as the first wave peaks hit the tidepool and continues after the predicted submergence as wave troughs expose the tidepool. Thus, we defined flood tide as the 35 min before and after the predicted submergence. We chose 35 min based on our observation that bub-bles from incoming waves always occurred in videos within the period 35 min before and after the predicted submergence. Changing wave heights during the videos likely extended or contracted the length of flood tide, but all videos were taken on days of fairly small significant wave height (mean daily significant wave height: 0.77–3.02 m.)

To determine (i) if ±35 min around predicted submergence was an appropriate time span for flood tide and to (ii) determine the potential wave and temperature stresses the snails may have experienced during the videos, we deployed a temperature and depth logger (HOBO U- 20 Onset corporation) sequentially in each of the three tidepools for 5–7 days between 1 and 20 August 2013 and set recordings to 1- min intervals (Figure 2). Although these logger deployments were 1 year after the snail videos were taken, we deployed them during the same season (summer), in the same tidepools and at the exact locations of the video cameras. In addition, dates of video and logger deployments had similar ranges of mean daily swell heights (0.77–3.02 and 0.63–2.13 m, respectively), air temperatures (8.8–12.3 and 10.8–13.5°C, respectively) and water temperatures (8.6–11.3 and 9.9–12.6°C, respectively; data from http://www.ndbc.noaa.gov/accessed through the UC Davis Bodega Ocean Observing Node http://bml.ucdavis.edu/boon).

F IGURE 1 The (a) tripod mount and (b) bottom mount for a digital video camera used to record movement in tidepools at Horseshoe Cove, California, between 25 May and 6 August 2012. The camera on the tripod mount was submerged in the tidepool using a telescoping arm during deployment to capture the vertical wall of the steep- sided tidepool. See Appendix S2 and Figures S1–S4 therein for details on mount designs, components and fabrication. A still frame of a video (c) shows the underwater view of Tegula funebralis in an experimental tidepool

(a)

(b)

(c)


While we cannot know what the actual temperature and wave conditions were during the videos, we can gain a view of the typi-cal environmental fluctuations the snails experienced in each tide-pool. Admittedly, the magnitudes of the fluctuations were likely different on the dates of video and logger deployments, but the general patterns among tidal phases were likely consistent. Most importantly, the logger deployments provide an estimation of the length of flood tide by showing how long the transition between low and high tide lasted. We predicted that low tide would show increasingly warm temperatures and low wave action (indicated by stable water depth), flood tide would show a sharp drop in water temperature and high wave action (indicated by deepening, fluc-tuating water depth), and high tide would have low stable water temperatures and high wave action (indicated by consistently deep and fluctuating water depth).

2.4 | Weather data

Although we have no direct measures of temperature and wave con-ditions that snails experienced at the scale of the tidepools during the

videos, we used nearby weather stations to measure local weather conditions on the dates the videos were recorded. We obtained data forsignificantwaveheight(meanheightofthehighest⅓ofwaves),from a mooring located 1 km offshore, water temperature from a seawater intake pipe <50 m away at 4 m depth and air temperature from weather instruments <100 m away from the tidepools (Bodega Ocean Observing Node http://bml.ucdavis.edu/boon). The proximity of these instruments renders the majority of the weather data very accurate (i.e. the water temperature during high tide, air temperature, and among- day differences in wave exposure). However, the local scale topography could have somewhat altered the wave heights the tidepools were exposed to, and the size of the tidepools and length of emersion likely affected how warm the tidepools became during low tide. No rainfall occurred during the study. We then ran correlations between mean snail velocity (cm/min) and mean daily (i) significant wave height, (ii) air temperature and (iii) water temperature. As snail velocity was negatively correlated with wave height and positively correlated with both air and water temperatures (Figure 3, see Results for details), we then included these as co- variates in further analyses on tidal phase.

F IGURE 2 Water depths and water temperatures inside three experimental tidepools at the depth of the video cameras for 3 weeks (a) in summer 2013. Panel (b) zooms in on 2 days (Aug 2 and 3 2013) to illustrate the distinction between low, flood, high and ebb tides. The increasingly warm temperatures and shallow stable depth coincide with low tide. The sudden drops in temperature as depth increased indicates flood tide, which we defined as predicted submergence ±35 min using tide tables. Low stable temperatures and deep fluctuating depth indicate high tide. Decreasing depth but stable low temperature indicates ebb tide, which we defined as predicted emergence ±35 min using tide tables. The deployments in the tidepools happened to coincide with some very warm days (left side of (a)) and some days with fairly large waves (right side of (a))Date

8/2/

2013

0:0

0:00

8/2/

2013

6:0

0:00

8/2/

2013

12:

00:0

0

8/2/

2013

18:

00:0

0

8/3/

2013

0: 0

0:00

8/3/

2013

6:0

0:00

8/3/

2013

12:

00:0

0

8/3/

2013

18:

00:0

0

8/4/

2013

0:0

0:00

Dep

th (c

m)

20

30

40

50

60

70

80

90

Tem

pera

ture

(°C

)10

12

14

16

18

20

22

flood

floodebb

ebb

ebb

low

low

high

high

NightNightNight DayDay

8/1/

13

8/2/

13

8/3/

13

8/4/

13

8/5/

13

8/6/

13

8/7/

13

8/8/

13

8/9/

13

8/10

/13

8/11

/13

8/12

/13

8/13

/13

8/14

/13

8/15

/13

8/16

/13

8/17

/13

8/18

/13

8/19

/13

8/20

/13

8/21

/13

Dep

th (c

m)

20

40

60

80

100

120

140

160

180

200

Tem

pera

ture

(°C

)

10

12

14

16

18

20

22nightPool 2 Pool 3 Pool 1(a)

(b)


2.5 | Statistical analyses

We tested the effects of significant wave height, air temperature, water temperature and tidal phase (low, flood, high) on mean snail velocities, mean vertical velocities and mean absolute turning angle conducting analyses of covariance (ANCOVAs) using the general lin-ear model platform in JMP PRO 12 software (SAS Institute) followed by Tukey’s posthoc analyses. We used adjusted sums of squares (Type III) to test the importance of each factor having already consid-ered the others. No transformations were necessary to meet model assumptions of normality and homogeneity of variance. Because the strong effects of weather on snail movement masked the effects of tidal phase in the figures, we ran identical ANCOVA models but removed tidal phase, and then used the model residuals as the response variable for our graphs. To further investigate whether the distributions of snail movements (residual velocity, residual vertical velocity and absolute turning angle) varied among tidal phases, we ran Kolmogorov–Smirnov tests between each of the three phases using the ks.test function in R.

3 | RESULTS

3.1 | Stresses over the tidal cycle

Overlays of temperature and depth fluctuations (Figure 2) in the tide-pools demonstrated that our estimation of flood tide as 35 min around predicted submergence typically captured both the increase in water depth (high wave stress) and sudden drop in temperature as waves first washed over tidepools. Further, the end of flood tide coincided with a stabilization of water temperature indicating full mixing of tide-pool volume with incoming seawater. Low tide typically exhibited an increase in water temperature during the daytime and very little water depth changes indicating little wave action. High tide typically showed fluctuating but high mean water depths, indicating high wave action, and low stable water temperatures.

3.2 | Weather conditions and snail behavior

As expected, mean snail velocity and significant wave height (Figure 3a; r2 = 0.10, F(1,226) = 24.57, p < .001, velocity = 1.891–0.438 × sig-nificant wave height) tended to be negatively related. Snail velocity was positively related to air (Figure 3b; r2 = 0.39, F(1,226) = 147.14, p < .001, velocity = –5.678 + 0.643 × air temperature) and water temperatures (Figure 3c; r2 = 0.30, F(1,226) = 98.95, p < .001, veloc-ity = –3.458 + 0.483 × water temperature).

When testing each weather variable having considered the effects of the other two weather variables and tidal phase, we found a similar strong positive correlation of air temperature with mean snail velocity (Table 1; ANCOVA: r2 = 0.45). However, we did not detect a relationship with wave height, but we detected a negative relationship with water temperature. This was surprising given the trends in Figure 3, and we suspect that tracking behavior over a larger range of wave heights and water temperatures would yield more discernible patterns. Snail vertical velocity was not correlated with any of the weather conditions (Table 1, ANCOVA: r2 = 0.03).

3.3 | Tidal phase and snail behavior

Having already considered the three weather conditions, snails moved faster during low tide than flood or high tide (Figure 4a; Table 1; ANCOVA: r2 = 0.45). The residual snail velocity distribu-tion was also shifted toward faster velocities at low than flood or high tide (Figure 4b; Kolmogorov–Smirnov: Low vs. Flood: D = 0.35, p < .001, Low vs. High: D = 0.33, p < .001). Although tidal phase did not affect snail vertical velocity after having considered the effects of the three weather conditions (Figure 4c; Table 1; ANCOVA: r2 = 0.03), the residual vertical velocity distributions indicated more downward movement during flood tide compared to low tide (Figure 4d; Kolmogorov–Smirnov: Low vs. Flood: D = 0.26, p = .017).

F IGURE 3 Correlations between mean velocities of individual Tegula funebralis from videos and mean (a) wave height, (b) air temperatures and (c) seawater temperatures taken at a nearby mooring during each video deployment

(a) (b) (c)


3.4 | Turning angle and snail trajectories

Mean snail turning angles were not affected by tidal phase (Figure 5; Table 1; Kolmogorov–Smirnov: D < 0.10, p > .771). Wave height was positively correlated with turning angle, but overall model fit was very

low, so this result is not likely ecologically relevant (Table 1; ANCOVA: r2 = 0.04).

By comparing snail movements in xy co- ordinate space (Figure 6a), it was clear that snails generally make clustered sharp turning angles with occasional short bursts of straight movement, indicating foraging,

TABLE 1 Analyses of covariance using adjusted sums of squares (SS) testing the effects of wave height, air temperature, water temperature and tidal phase on mean velocity and vertical velocity of Tegula funebralis

Term

Mean velocity Mean vertical velocity Mean turning angle

df SS F p df SS F p df SS F p

Significant wave height

1 <0.01 <0.01 .992 1 <0.01 0.03 .861 1 3,028 6.76 .010

Air temperature 1 18.56 47.81 <.001 1 0.62 1.98 .161 1 938 2.09 .149

Water temperature 1 1.83 4.70 .031 1 0.23 0.73 .393 1 278 0.62 .431

Tidal phase 2 8.65 11.13 <.001 2 0.83 1.34 .264 2 797 0.89 .412

F IGURE 4 The means (a & c) and distributions (b & d) of individual Tegula funebralis velocities (a & b) and mean vertical velocities (c & d) among the tidal phases having already considered the effects of wave height, air temperature and water temperature on snail velocity. In (a) and (b), values greater (lower) than zero indicate faster (slower) than predicted velocities. In (c) and (d), values greater (lower) than zero indicate more upward (downward) movement than predicted

(a) (b)

(c) (d)


followed by relocation, then foraging (Swingland & Greenwood, 1983). By investigating turning angle (Figure 6b), it is clear that snails were repeatedly alternating between clockwise (positive values) and coun-terclockwise (negative values) sharp turns, again with occasional straight trajectories (near- zero values). Other snails exhibited gener-ally straight paths, which indicate directed relocation or fleeing from a stimulus (Swingland & Greenwood, 1983), although these behaviors were much less common.

4 | DISCUSSION

4.1 | Increased activity at low tide and on warm days

Increased velocity during daytime low tide suggests that T. funebra-lis was active and perhaps foraging (Swingland & Greenwood, 1983). Similarly, we observed increased snail activity on days with warmer air and probably water temperatures. Although we did not measure tidepool water temperature directly during our study, the maximum daily tidepool water temperatures during low tide 1 year later were between 14.2 and 20.7°C compared to maxima between 11.0 and 16.6°C during high tide. Therefore, warmer tidepool water tempera-tures during low tides in our study were likely similar and promoted increased snail activity. Hence, increasingly warm water temperatures as low tide progressed and on warm days likely promoted grazing. In a previous study, feeding rates of T. funebralis were maximal at mod-erately warm water temperatures (15°C) and activity levels increased as water temperature rose to 23°C in the laboratory (Yee & Murray, 2004). Similarly, the thermal optimum, as measured by turning fre-quency in the laboratory, was around 20°C (Tepler, Mach, & Denny, 2011). Furthermore, heart rate increased with temperature until about 31°C in T. funebralis, after which it decreased rapidly (Stenseng, Braby, & Somero, 2005), as is typical when temperature extremes are reached. These studies corroborate our finding that snails are more active during warmer low tides compared to colder flood and high

tides. However, they were not conducted in the same geographic area and local adaptation could result in slightly different physiologi-cal responses by our experimental snails (Gleason & Burton, 2013). Increased activity during warmer temperatures is a typical feature of ectotherms as their metabolism and muscle activity increases with temperature (Huey & Stevenson, 1979). Temperatures during our study apparently did not approach the extreme temperatures that are stressful for T. funebralis (Stenseng et al., 2005; Tomanek & Somero, 1999), because their activity levels did not decrease on the warmest days. We also saw no evidence of stress in the videos, although during heat waves at this site we have observed snails in tidepools upside down with their opercula closed. Our finding of increased activity dur-ing low tide and warm days is somewhat surprising as warm low tides are generally thought to be quite stressful for intertidal organisms. However, tidepools likely are a respite from this stress on most low tides (Garrity, 1984). Indeed, the lack of upward movement at low tide suggests that snails may have remained in tidepools to avoid much higher temperatures and desiccation outside of them. These stresses can be lethal, and T. funebralis and other gastropods outside of tide-pools reduce movement and feeding while seeking refuge during low tide (Jones & Boulding, 1999; Marchetti & Geller, 1987; Wara & Wright, 1964; Yamane & Gilman, 2009). Thus, emersed and immersed T. funebralis may exhibit very different activity patterns over the tidal cycle.

4.2 | Decreased activity at flood and high tide and on wavy days

Snails in waveexposed tidepools decreased activity during flood and high tide compared to low tide, and appeared to move downward at flood tide. Further, snail activity tended to decrease on days with larger waves. These behaviors indicate that snails may mitigate the risk of dislodgement from increasingly frequent and energetic waves during flood tide (Castilla, Steinmiller, & Pacheco, 1998). Strong turbulence

F IGURE 5 The mean (a) and distribution (b) of the absolute turning angles by Tegula funebralis among tidal phases. 0° indicates straight movement, 90° indicates a perpendicular turn and 180° indicates a reversal in direction

(a) (b)


may dislodge and kill T. funebralis, as it does periwinkle snails and limpets (Chapperon & Seuront, 2009; Denny & Blanchette, 2000; Miller, O’Donnell, & Mach, 2007; Pardo & Johnson, 2006; Wright & Nybakken, 2007). Larger waves prevent intertidal snails from return-ing to the upper intertidal zone after dislodgment (Miller et al., 2007), and they tend to seek shelter from waves during high tide (Pardo & Johnson, 2006). Thus, it is not surprising that T. funebralis decreased activity during flood and high tide and wavy days.

In laboratory studies, T. funebralis exhibited a pulse in downward movement when exposed to turbulence (Overholser, 1964), so the snails in our study may have exhibited a similar behavior by moving downward at flood tide. This is in contrast to an other study that has shown that tropical gastropods show a pulse in activity during flood tide on a sheltered rocky intertidal shore, which was attributed to decreased physiological stress coupled with temporarily low fish pre-dation (Garrity, 1984). We did not find a pulse in velocity at flood tide, perhaps because the wave forces at this exposed site are much larger and the temperature stresses are lower than in protected tropical waters.

The observed patterns in T. funebralis behavior may have cascad-ing effects on the intertidal community. Tegula funebralis primarily grazes benthic diatoms and some species of macroalgae (Aquilino, Coulbourne, & Stachowicz, 2012; Thornber, Jones, & Stachowicz, 2008), and it can affect both biomass and community structure of tide-pool macroalgae by changing succession trajectories (Nielsen, 2001). Grazing by T. funebralis in tidepools also slows the growth of both micro- and macroalgae (Gravem, 2015). In other intertidal systems, gastropod density is negatively correlated with wave height, affect-ing the abundance and community structure of algae (Lauzon- Guay & Scheibling, 2009). Further, limpets assess when the risk of dislodge-ment is low enough to graze, affecting the abundance and commu-nity structure of algae (Wright & Nybakken, 2007). Overall, rhythms in foraging and other activities by intertidal species can affect the

abundance and composition of species in tidepools and the exposed areas around them (Morgan, 2007).

4.3 | Foraging trajectories

Snails exhibited foraging trajectories by moving in a zig- zagging motion for minutes at a time then exhibiting occasional short linear move-ments followed again by zig- zagging. This type of trajectory is a typi-cal foraging trajectory for grazing animals (Swingland & Greenwood, 1983). However, we did not observe a difference in mean turning angle or change in the turning angle frequency distributions among tidal phases or any strong correlation between turning angle and weather conditions. Therefore, snails may be foraging at all times, but warmer temperatures during warm days or low tide simply increase their velocity, presumably allowing them to graze a larger area.

5 | CONCLUSIONS

Tegula funebralis in tidepools suppressed activity and perhaps moved downward during daytime flood tide, and larger waves may have reduced activity. Increased activity during low tides and warm days suggests they utilize wave- free and warm periods to graze, probably in part because their metabolism increases. We developed an inex-pensive, safe, and effective method for observing organisms dur-ing high tide, which is a gap in our understanding of exposed rocky shore communities. Although we focused on just one of the species recorded, the behaviors of other species were clearly visible, includ-ing abundant hermit crabs, limpets and sculpins. We also focused on velocities and trajectories over the tidal cycle, but many other behav-iors could be observed both inside and outside tidepools using our methods, such as foraging, homing, mating, spawning, hatching and intraspecific and interspecific interactions between predator and

F IGURE 6 An example of an individual Tegula funebralis (a) trajectory and (b) turning angles during a low tide on 25 May 2012. Positive and negative values of turning angles in (b) indicate clockwise and counter- clockwise turns, respectively, and 0° indicates straight paths

(a) (b)


prey and competitors. We restricted our observations to daytime dur-ing spring and summer months, and further exploration of seasonal or diel changes in behavior are logical next steps. Our techniques are easily applied, opening the door to widespread investigations of rocky intertidal communities during the “other half” of the tidal cycle.

ACKNOWLEDGEMENTS

We thank Ned Antell, Danielle Smull, Dawn Barlow, Emily Booth, Megan Miguel, Sam Gordon and Sam Kemiji for assistance with cam-era deployments and video analysis. Thanks to Matt Robart and John Largier for use of the HOBO depth and temperature logger. We also thank Bill, Alice and Ryan Taylor, Junior Guillette, and Mindy, Andrew, Skyler and Taylor Allen for their assistance with camera mount design and construction. Special thanks to Bill and Alice Taylor for provid-ing funding for the cameras and mounts, and Ryan Taylor for making the long drive to Bodega Bay to help deploy the Tripod. The Henry A. Jastro fellowship, UC Davis Graduate Group in Ecology, UC Davis Department of Environmental Science and Policy, and Bodega Marine Laboratory Fellowship all provided research funds or financial sup-port. The University of California Natural Reserve System provided the space to make this research possible. Data on ocean conditions were provided by the UC Davis, Bodega Marine Laboratory. Funding was also in part provided by California SeaGrants R/FISH218, R/MPA14 and R/MPA24 awarded to S.M., National Science Foundation Biological Oceanography grants OCE- 1334448 and OCE- 0927196 awarded to S.M., and National Science Foundation GK- 12 grant 0841297 awarded to Susan Williams. This work complies with the legal and ethical require-ments for research and permits were issued by California Department of Fish and Game (SC- 4688), the University of California Natural Reserve System. This study complies with the current laws of the United States of America. The authors declare that they have no conflict of interest. This is a contribution of the Bodega Marine Laboratory and a project begun during the undergraduate Marine Ecology field course.

REFERENCES

Aquilino, K. M., Coulbourne, M. E., & Stachowicz, J. J. (2012). Mixed species diets enhance the growth of two rocky intertidal herbivores. Marine Ecology Progress Series, 468, 179–189. doi:10.3354/meps09893

Bouchet, P., & Rosenberg, G. (2015). Tegula funebralis. In MolluscaBase. Retrieved from World Register of Marine Species at http://www.ma-rinespecies.org/aphia.php?p=taxdetails&id=534190 on 2015-09-15.

Burrows, M. T., Kawai, K., & Hughes, R. N. (1999). Foraging by mobile pred-ators on a rocky shore: Underwater TV observations of movements of blennies Lipophrys pholis and crabs Carcinus maenas. Marine Ecology- Progress Series, 187, 237–250.

Calenge, C. (2006). The package “adehabitat” for the R software: A tool for the analysis of space and habitat use by animals. Ecological Modelling, 197, 516–519. doi:10.1016/j.ecolmodel.2006.03.017

Castilla, J. C., Steinmiller, D. K., & Pacheco, C. J. (1998). Quantifying wave exposure daily and hourly on the intertidal rocky shore of central Chile. Revista Chilena de Historia Natural, 71, 19–25.

Chapperon, C., & Seuront, L. (2009). Cue synergy in Littorina littorea nav-igation following wave dislodgement. Journal of the Marine Biological

Association of the United Kingdom, 89, 1133–1136. doi:10.1017/s0025315409000150

Dayton, P. K. (1985). The structure and regulation of some South American kelp communities. Ecological Monographs, 55, 447–468. doi:10.2307/2937131

Denny, M. W. (1989). A limpet shell shape that reduces drag: Laboratory demonstration of a hydrodynamic mechanism and an exploration of its effectiveness in nature. Canadian Journal of Zoology, 67, 2098–2106. doi:10.1139/z89- 299

Denny, M. W., & Blanchette, C. A. (2000). Hydrodynamics, shell shape, behavior and survivorship in the owl limpet Lottia gigantea. Journal of Experimental Biology, 203, 2623–2639.

Doering, P. H., & Phillips, D. W. (1983). Maintenance of the shore- level size gradient in the marine snail Tegula funebralis (Adams, A.): Importance of behavioral responses to light and sea star predators. Journal of Experimental Marine Biology and Ecology, 67, 159–173. doi:10.1016/0022- 0981(83)90087- 4

Garrity, S. D. (1984). Some adaptations of gastropods to physical stress on a tropical rocky shore. Ecology, 65, 559–574. doi:10.2307/1941418

Gleason, L. U., & Burton, R. S. (2013). Phenotypic evidence for local adap-tation to heat stress in the marine snail Chlorostoma (formerly Tegula) funebralis. Journal of Experimental Marine Biology and Ecology, 448, 360–366. doi:10.1016/j.jembe.2013.08.008

Gravem, S. A. (2015). Linking antipredator behavior of prey to intertidal zona-tion and community structure in rocky tidepools. PhD dissertation, Davis, CA: University of California.

Gravem, S. A., & Morgan, S. G. (2016). Prey state alters trait- mediated indi-rect interactions in rocky tidepools. Functional Ecology, 30, 1574–1582. doi:10.1111/1365- 2435.12628

Huey, R. B., & Stevenson, R. (1979). Integrating thermal physiology and ecology of ectotherms: A discussion of approaches. American Zoologist, 19, 357–366.

Hunter, E., & Naylor, E. (1993). Intertidal migration by the shore crab Carcinus maenas. Marine Ecology- Progress Series, 101, 131–138. doi:10.3354/meps101131

Jones, K. M. M., & Boulding, E. G. (1999). State- dependent habitat selec-tion by an intertidal snail: The costs of selecting a physically stressful microhabitat. Journal of Experimental Marine Biology and Ecology, 242, 149–177. doi:10.1016/s0022- 0981(99)00090- 8

Kitting, C. L. (1979). The use of feeding noises to determine the algal foods being consumed by individual intertidal molluscs. Oecologia, 40, 1–17.

Lauzon-Guay, J.-S., & Scheibling, R. E. (2009). Food dependent movement of periwinkles (Littorina littorea) associated with feeding fronts. Journal of Shellfish Research, 28, 581–587. doi:10.2983/035.028.0322

Marchetti, K. E., & Geller, J. B. (1987). Effects of aggregation and habitat on desiccation and body temperature of the black turban snail, Tegula funebralis (A. Adams, 1855). Veliger, 30, 127–133.

Martone, P. T., & Denny, M. W. (2008). To break a coralline: Mechanical constraints on the size and survival of a wave- swept seaweed. Journal of Experimental Biology, 211, 3433–3441. doi:10.1242/jeb.020495

Menge, B. A. (1978). Predation intensity in a rocky intertidal community - relation between predator foraging activity and environmental harsh-ness. Oecologia, 34, 1–16. doi:10.1007/bf00346237

Miller, S. L. (1974). Adaptive design of locomotion and foot form in proso-branch gastropods. Journal of Experimental Marine Biology and Ecology, 14, 99–156. doi:10.1016/0022- 0981(74)90021- 5

Miller, L. P. (2007). Feeding in extreme flows: Behavior compensates for mechanical constraints in barnacle cirri. Marine Ecology Progress Series, 349, 227–234.

Miller, L. P., O’Donnell, M. J., & Mach, K. J. (2007). Dislodged but not dead: Survivorship of a high intertidal snail following wave dislodgement. Journal of the Marine Biological Association of the United Kingdom, 87, 735–739. doi:10.1017/s0025315407055221


Morgan, S. G. (2007). Tidal rhythms. In M. Denny & S. D. Gaines (Eds.), Encyclopedia of tidepools and rocky shores (pp. 469–473). Stanford, CA: University of California Press.

Nielsen, K. J. (2001). Bottom- up and top- down forces in tide pools: Test of a food chain model in an intertidal community. Ecological Monographs, 71, 187–217. doi:10.1890/0012- 9615(2001) 071[0187:BUATDF]2.0.CO;2

Norton, T., Hawkins, S., Manley, N., Williams, G., & Watson, D. (1990). Scraping a living: A review of littorinid grazing. Hydrobiologia, 193, 117–138.

Overholser, J. A. (1964). Orientation and response of Tegula funebralis to tidal current and turbulence (Mollusca: Gastropoda). Veliger, 6, 38–41.

Palmer, J. D. (1974). Biological clocks in marine organisms: The control of phys-iological and behavioral tidal rhythms. New York, London, Sydney and Toronto: John Wiley and Sons.

Pardo, L. M., & Johnson, L. E. (2006). Influence of water motion and re-productive attributes on movement and shelter use in the marine snail Littorina saxatilis. Marine Ecology Progress Series, 315, 177–186. doi:10.3354/meps315177

R Core Team (2013). A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.

Rasband, W. (1997-2017). ImageJ. Betzrhesda, MD, USA: US National Institutes of Health. Retrieved from http://imagej.nih.gov/ij/

Ricketts, E., Calvin, J., Hedgpeth, J., & Phillips, D. (1985). Between pacific tides, 5th edn. Stanford, CA: Stanford University Press.

Robles, C. D., Alvarado, M. A., & Desharnais, R. A. (2001). The shifting bal-ance of littoral predator- prey interaction in regimes of hydrodynamic stress. Oecologia, 128, 142–152. doi:10.1007/s004420100638

Silva, A., Hawkins, S., Boaventura, D., Brewster, E., & Thompson, R. (2010). Use of the intertidal zone by mobile predators: Influence of wave expo-sure, tidal phase and elevation on abundance and diet. Marine Ecology Progress Series, 406, 197–210. doi:10.3354/meps08543

Stenseng, E., Braby, C. E., & Somero, G. N. (2005). Evolutionary and acclimation- induced variation in the thermal limits of heart function in congeneric marine snails (Genus Tegula): Implications for vertical zona-tion. Biological Bulletin, 208, 138–144. doi:10.2307/3593122

Swingland, I. R., & Greenwood, P. J. (1983). Ecology of animal movement. Oxford, UK: Clarendon Press.

Tepler, S., Mach, K., & Denny, M. (2011). Preference versus performance: Body temperature of the intertidal snail Chlorostoma funebralis. The Biological Bulletin, 220, 107–117.

Thornber, C. S., Jones, E., & Stachowicz, J. J. (2008). Differences in her-bivore feeding preferences across a vertical rocky intertidal gradient. Marine Ecology- Progress Series, 363, 51–62. doi:10.3354/meps07406

Tomanek, L., & Somero, G. N. (1999). Evolutionary and acclimation- induced variation in the heat- shock responses of congeneric marine snails (genus Tegula) from different thermal habitats: Implications for limits of thermotolerance and biogeography. Journal of Experimental Biology, 202, 2925–2936. doi:10.1016/S1095- 6433(99)90421- X

Vieira, S., Coelho, H., Nolasco, R., Serôdio, J., Barnes, R. S., & Queiroga, H. (2012). Repeated cycles of immersion and emersion amplify the crawling rhythm of the intertidal gastropod Hydrobia ulvae. Journal of the Marine Biological Association of the United Kingdom, 92, 565–570. doi:10.1017/s0025315411000853

Wara, W. M., & Wright, B. B. (1964). The distribution and movement of Tegula funebralis in the intertidal region of Monterey Bay, California. Veliger, 6, 30–37.

Wright, W. G., & Nybakken, J. W. (2007). Effect of wave action on move-ment in the owl limpet, Lottia gigantea, in Santa Cruz, California. Bulletin of Marine Science, 81, 235–244.

Yamane, L., & Gilman, S. E. (2009). Opposite responses by an intertidal predator to increasing aquatic and aerial temperatures. Marine Ecology- Progress Series, 393, 27–36. doi:10.3354/meps08276

Yee, E., & Murray, S. (2004). Effects of temperature on activity, food consumption rates, and gut passage times of seaweed- eating Tegula species (Trochidae) from California. Marine Biology, 145, 895–903. doi:10.1007/s00227- 004- 1379- 6

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the sup-porting information tab for this article.

How to cite this article: Taylor AW, Morgan SG, Gravem SA. Underwater video reveals decreased activity of rocky intertidal snails during high tides and cooler days. Mar Ecol. 2017;38:e12418. https://doi.org/10.1111/maec.12418

Prey state alters trait-mediated indirect interactions inrocky tide poolsSarah A. Gravem*,1,2 and Steven G. Morgan1

1Bodega Marine Laboratory, University of California Davis, PO Box 247, Bodega Bay, 94923 California, USA; and2Department of Integrative Biology, Oregon State University, 3029 Cordley Hall, Corvallis, 97331 Oregon, USA

Summary

1. Several studies on trait-mediated indirect interactions (TMIIs) have shown that predators

can initiate trophic cascades by altering prey behaviour. Although it is well recognized thatindividual prey state alters antipredator and foraging behaviour, few studies explore whether

this state-dependent prey behaviour can alter the strength of the ensuing tritrophic cascade.Here, we link state-dependent individual behaviour to community processes by experimentally

testing whether hunger level and body size of prey altered antipredator behaviour and thuschanged the strength of trophic cascades between predators and primary producers.

2. In rocky intertidal tide pools on the California Coast, waterborne cues from the predatoryseastar Leptasterias spp. (Stimpson) can cause the herbivorous snail Tegula (Chlorostoma)funebralis (A. Adams) to reduce grazing and flee tide pools, resulting in positive indirect effects

on tide pool microalgae.3. However, we show that the strength of this behaviourally-mediated cascade may be contin-

gent on prey hunger level and body size. During short field experiments at low tide, medium-sized snails that were either newly collected from the field or fed for 1 week in the laboratory

mediated strong TMIIs because they grazed less when seastars were present. In contrast, noTMIIs occurred when medium-sized snails had been starved for 1 week because they continued

grazing regardless of seastar presence. Newly collected small snails fled from seastars but didnot mediate cascades because they ate little algae. Despite reaching an apparent size refugefrom predation, many newly collected large snails fled from seastars, but those individuals that

remained tended to graze the algae more quickly, resulting in unexpected negative indirecteffects of seastars on algae cover. The implication of this pattern for the natural system is

unclear.4. Because average hunger level and size of snails vary over time and space in nature, a mosaic

of TMII strength may exist.5. Overall, the strength of tritrophic TMIIs in tide pools depended on individual prey state,

supporting model predictions and adding to sparse empirical evidence. This outcome suggeststhat patterns occurring system-wide over the long term may be influenced by the state-depen-

dent decisions made by the individuals present.

Key-words: adaptive foraging theory, antipredator behaviour, Leptasterias, nonconsumptiveeffect, predator–prey interaction, Tegula funebralis, trait-mediated indirect interaction, trophic

cascade

Introduction

Predator–prey interactions are often expressed as preda-

tion rates on prey, and for simplicity these metrics gener-

ally treat all individuals within a population as

homogenous (Schmitz, Adler & Agrawal 2003; Ohgushi,

Schmitz & Holt 2012). However, adaptive foraging theory

unequivocally demonstrates that predators also exert non-

consumptive effects on prey by changing their foraging

behaviour, and that these behaviours are contingent on

individual variation in prey states, such as body size,

energy reserves, reproductive status or behavioural syn-

dromes (Mangel & Clark 1986; Houston, McNamara &

Hutchinson 1993; Werner & Anholt 1993; Clark 1994;*Correspondence author. E-mail: [email protected]

© 2015 The Authors. Functional Ecology © 2015 British Ecological Society

Functional Ecology 2016, 30 , 1574–1582 doi: 10.1111/1365-2435.12628

Lima 1998; Sih, Bell & Johnson 2004). Trophic cascades

connect predator–prey interactions to a third species,

and the classic mechanism is a form of a density-

mediated indirect interaction (DMIIs, Abrams 1995; Pea-

cor & Werner 1997), whereby predators can benefit pri-

mary producers by reducing grazer population densities.

Similar to foraging theory that focuses on the sublethal

effects of predators on prey, studies on trait-mediated

indirect interactions (TMIIs, Abrams 1995; Peacor &

Werner 1997) demonstrate that predators often exert

sublethal effects on prey behaviour, morphology or phys-

iology that cause indirect cascading effects on primary

producers (Werner & Peacor 2003; Schmitz, Krivan &

Ovadia 2004; Miner et al. 2005). TMII studies weave a

connection between behavioural, population and commu-

nity ecology by linking individuals to emergent commu-

nity patterns and ecosystem processes (Schmitz, Adler &

Agrawal 2003; Schmitz et al. 2008).

Just as predator–prey interactions within foraging the-

ory depend on the state of individuals, it is likely that

TMII strength may be contingent upon the individual

state of the organisms involved. Multiple reviews and

syntheses have identified this concept as underexplored

and called for studies connecting individual state varia-

tion to community-level patterns (Schmitz, Adler &

Agrawal 2003; Schmitz, Krivan & Ovadia 2004; Agra-

wal et al. 2007; Beckerman, Petchey & Morin 2010;

Ohgushi, Schmitz & Holt 2012; Rudolf 2012; Railsback

& Harvey 2013). For example, Beckerman, Petchey &

Morin (2010) point out that there are ‘. . .precious few

advances towards truly synthesizing the connections

between individuals, populations and large intercon-

nected food webs’. In response, theoretical models have

demonstrated that prey body size and hunger level may

alter TMII strength (Schmitz 2000; Luttbeg, Rowe &

Mangel 2003; Persson & De Roos 2003). In addition,

field and laboratory studies have demonstrated that

TMII strengths may be contingent on prey body size,

hunger level and the combination of hunger and risk

frequency (Ovadia & Schmitz 2002; Kotler, Brown &

Bouskila 2004; Freeman 2006; Hawlena & Schmitz

2010; Rudolf 2012; Matassa & Trussell 2014) though

other studies suggest that prey traits may be safely

ignored in certain cases (Ovadia & Schmitz 2002, 2004;

Ovadia et al. 2007). Overall, additional empirical studies

are needed to more fully explore the consequences of

prey state variation for TMIIs.

We investigated whether hunger level and body size of

prey altered antipredator behaviour and the strength of

TMIIs in a tritrophic food chain. Adaptive foraging theory

and the threat-sensitivity hypothesis (Helfman 1989; Wer-

ner & Anholt 1993; Clark 1994; Lima 1998) posit that prey

with high-energy reserves or at higher risk should be wary

of predators and increase refuge use, but prey with low-

energy reserves or at lower risk should forage despite

predator presence. What are the consequences for a third

trophic level, as in a TMII? Predators may exert positive

TMIIs on primary producers when well-fed prey are wary,

whereas they may initiate weak or no TMIIs when hungry

prey forage despite risk (Heithaus et al. 2007; Matassa &

Trussell 2014). In addition, both the consumptive and the

nonconsumptive effects of predators on prey may change

with prey body size, which may then alter effects on prey

resources (Rudolf 2012). For example, predators may

select small size classes of prey due to the ease of capture

or shorter handling time (MacArthur & Pianka 1966). If

all sizes of prey continue to exhibit antipredator responses,

then small prey may mediate both TMIIs and DMIIs while

large prey may mediate only TMIIs if they are rarely cap-

tured. Alternatively, larger prey may stop responding to

predators as they grow and risk abates, and thus cease to

mediate TMIIs or DMIIs (Freeman 2006). Overall, failing

to consider variation in prey traits may lead to erroneous

estimates of the strength and importance of TMIIs (Rudolf

2012).

To determine whether body size or hunger level alters

prey behaviour and changes TMII strength, we examined

a tritrophic food chain in tide pools where the small

(1–5 cm diameter) seastar Leptasterias spp. (L. aequalis

and L. hexactis, considered either sister species or sub-

species, Flowers & Foltz 2001) consumes the common her-

bivorous intertidal snail Tegula (formerly Chlorostoma)

funebralis (Bouchet & Rosenberg 2015), which grazes on

algae (Fig. 1). To provide context for the study, we first

assessed the potential for size-dependent predation by Lep-

tasterias on Tegula by testing whether smaller snails were

eaten more than larger snails in the laboratory and field.

We also examined size-dependent antipredator responses

by comparing evasive behaviour of different sized snails to

both tactile (imminently threatening) and waterborne

(prospectively threatening) predator cues. These experi-

ments provide important supplemental information on

these predator–prey interactions for interpreting the results

of our focal experiments on the effect of prey state on

TMIIs. With this information in hand, we then tested if

snail hunger level (hungry vs. fed) or body size (3 size

classes) changed snail antipredator behaviour, and con-

sequently altered TMIIs on microalgae in the field during

Fig. 1. Leptasterias sp. hunting Tegula funebralis in a rocky inter-tidal pool. Various macroalgae species shown. Photograph bySarah A. Gravem.

© 2015 The Authors. Functional Ecology © 2015 British Ecological Society, Functional Ecology, 30 , 1574–1582

Prey state alters TMIIs in tide pools 1575

low tide. We also discuss the ecological relevance of our

short-term experiments to natural populations of snails.

Materials and methods

EXPER IMENTS ON CONSUMPT IVE AND

NONCONSUMPT IVE EFFECTS ON PREY

To determine the frequency of predation by Leptasterias on Teg-ula relative to other prey species, we first surveyed the diets ofLeptasterias in the intertidal zone in July 2009 (see Appendix S1,Supporting information for details) at the Bodega Marine Reserveat Horseshoe Cove in northern California, USA (38° 180 59!37″ N,123° 40 16!28″ W). To test our expectation that predation riskdecreases with Tegula size, we paired single Leptasterias with sin-gle Tegula in varying size combinations in very small seawatertanks to test whether Tegula reach a size refuge from Leptasteriaspredation (see Appendix S1). Finally, we tested the escaperesponses (fleeing and meandering) of varying sizes of Tegula totactile, waterborne or no Leptasterias cues to test whether all sizesof snails respond to seastars (see Appendix S1).

TMI I F IELD EXPER IMENTS

We tested if Tegula size or hunger level changed their antipredatorbehaviours and thus changed the strength of the TMII betweenLeptasterias spp. and microalgae in tide pools of Horseshoe Cove.We added snails of different hunger levels or size classes to 18small mid-to-high tide pools (1!2–9!3 L, 1!01–1!67 m above meanlower low water). In hunger experiments, we factorially crossedseastar presence with fed, hungry or no snails totalling six treat-ments. In size experiments, we used eight total treatments by fac-torially crossing seastar presence with small, medium, large or nosnails. Small, medium and large size classes were defined as 6–12,12–18 and 18–25 mm at the widest shell width, respectively (sensuPaine 1969). Only medium-sized snails were used in hunger experi-ments. We performed four separate hunger trials (18 July, 2August, 30 August and 27 September, 2012) and six separate sizetrials (24 May, 25 May, 22 June, 26 June, 20 July and 29 August,2012). Though tide pool water temperatures were likely warmer,average daily seawater temperatures (mean " SD) recorded bythe Bodega Marine Laboratory seawater system were 12!3 " 0!3and 12!1 " 0!8 °C for hunger and size trials, respectively. Treat-ments without snails were included to control for any effects onalgae due to handling, seastars, or grazing by other herbivores.During each trial, treatments were randomly assigned to the 18tide pools, resulting in a cumulative total of 9–15 replicate tidepools per treatment. One week prior to hunger experiments, snailswere collected from Horseshoe Cove and held in outdoor, flow-through tanks where they either were starved or allowed to grazead libitum on a thick layer of microalgae (benthic diatoms) thathad been growing naturally in tanks for the ~2 preceding weeks.One day prior to size experiments, snails were collected fromHorseshoe Cove, measured, and held in indoor flow-throughaquaria overnight.

Medium to large seastars (2–5 cm diameter across two longestopposing arms) were added to half of the tide pools and werecontained in small mesh pouches of plastic window screen thatwere affixed to eyebolts drilled into the substrate to preventescape. Because of local Leptasterias mass mortality events inNovember 2010 and August 2011 (Gravem 2015; Jurgens et al.2015), Leptasterias were collected 104 km north of HorseshoeCove at Point Arena (38° 540 47!48″ N, 123° 420 37!83″ W).Leptasterias were maintained in flow-through tanks and fedsmall Tegula and Littorina spp. weekly. We mimicked naturalpredator cue accumulation during low tide by adding seastars to

tide pools the day before the experiment. To keep waterborneseastar cue concentration similar and ecologically relevantamong tide pools, we scaled the number of predators to tidepool volume using natural Leptasterias densities (~0!41 individu-als L#1) that we recorded in tide pools at Horseshoe Cove inJuly 2009, prior to the November 2010 Leptasterias mortalityevent. We haphazardly assigned individual seastars to tide pools,taking care to keep average seastar size similar among tidepools. As a result of the 2010 and 2011 mortality events, Lep-tasterias were naturally absent in the tide pools during theexperiment unless experimentally added. We also scaled snaildensities to tide pool volume and they were added to tide poolsat 50% natural field densities recorded in tide pools in July2009 (13!6 individuals L#1), because 100% density would havemade treatments with large snails overcrowded. We chose to useconstant densities of snails rather than constant biomass becausemore conspecifics can decrease individual responsiveness topredators since the chance of being eaten decreases with groupsize (Dehn 1990).

To assess grazing in TMII experiments, we used unglazedporcelain tiles (2!54 9 2!54 cm) placed in outdoor flow-throughtanks 1–2 weeks before experiments to grow a thin film of benthicdiatoms. Since it was not possible to manipulate algal biomass,algae tiles were haphazardly assigned among tide pools and carewas taken to maintain similar average algal biomass among tidepools. To ensure algal tiles were easily and equally accessible tosnails among tide pools of different sizes, the tiles were denselydistributed with the number of tiles scaled to the tide pool surfacearea (0!010 tiles cm#2). Steep sides in one tide pool limited themaximum density of tiles to 0!006 tiles cm#2. Each tide pool andsurrounding areas were cleared of Tegula, other herbivorous gas-tropods and hermit crabs both the day before and day of eachexperiment. Algal cover on tiles was measured at the end of eachexperiment using a gridded transparent quadrat (2!54 9 2!54 cmwith 25 cells measuring 0!51 9 0!51 cm); tiles always began with100% algae cover.

On the day of experiments, we first deployed tiles in all tidepools, quickly followed by snails, taking care not to place snailson tiles (seastars had been deployed the day before). Snails typi-cally avoid seastar cues by escaping tide pools to refuge habitatsabove the water line (Gravem 2015), which we termed the ‘halo’and defined as substrate 0–15 cm above the waterline in each tidepool. The numbers of snails in the water, in the halo, and grazingwere sampled every 5–10 min in each tide pool for 1 h (size trials)or 45 min (hunger trials). Grazing was identified as visible raspingon the algae-covered tiles, which invariably caused clearing of thethin layer of microalgae. Snails rarely dislodged algae withoutrasping. To determine whether seastars, hunger level or sizecaused those individuals grazing to consume algae faster, snailgrazing rate was estimated as the [(total algal surface area eaten)/(sum snails grazing * time)].

Since trends over time were not always linear, time was groupedinto 15-min increments and treated as a categorical variable dur-ing statistical analyses. We calculated the percentage of total snailsout of water or grazing at each time frame ([snails out of water orgrazing/total snails] * 100) and the percentage algae cover at theend of the experiments. We tested the main and interactive effectsof time frame, seastar treatment and snail treatment on the percent snails in the halos and grazing using restricted maximum like-lihood (REML) mixed models in JMP software (SAS Institute,Cary, NC, USA). The main and interactive effects of seastar andsnail treatments on per cent algal cover and grazing rate were alsoanalysed using REML. In each model, tide pool replicate wasincluded as a random variable (nested within seastar and snailtreatments) to account for non-independence of measures in thesame tide pool over time. All response variables were arcsinesquare root transformed to meet statistical assumptions of nor-mality and equal variances.


1576 S. A. Gravem & S. G. Morgan

Results

EXPER IMENTS ON CONSUMPT IVE AND

NONCONSUMPT IVE EFFECTS ON PREY

In the field, Tegula comprised 24% of Leptasterias diets,

and small snails (<12 mm) constituted 80% of Tegula

eaten (Fig. 2a, Appendix S1). In the laboratory, predation

decreased with snail size with small and medium snails

eaten often and large snails eaten rarely (Fig. 2b,

Appendix S1, 76!6%, 40!0% and 6!4% of small, medium

and large snails, respectively). Overall, snails exhibited the

strongest fleeing responses from tactile, followed by water-

borne, then no cue and meandered more when exposed to

waterborne than tactile cues (Fig. S1, Appendix S1). How-

ever, these differences were strongest for large snails, fol-

lowed by medium snails. Small snails did not change speed

and meandered frequently regardless of cue type (Fig. S1,

Appendix S1).

TMI I F IELD EXPER IMENTS

On average, newly collected medium and large snails

escaped from seastars more quickly than newly collected

small snails (Fig. 3a–c; Time 9 Seastar treatment 9 Snail

size: F8,899 = 2!32, P = 0!018; Time 9 Seastar treatment:

F4,899 = 50!40, P < 0!001). By the end of experiments,

many snails of all sizes had fled tide pools in response to

seastars (Fig. 3a–c; Mean % in halo at end " SE:

15!9 " 2!4, 23!4 " 3!1, and 21!4 " 1!7% for small, med-

ium and large, respectively). In contrast, few snails fled

when seastars were absent (Mean % in halo at end " SE:

3!9 " 1!0, 4!1 " 0!8, and 4!9 " 0!7% for small, medium

and large, respectively). Leptasterias also caused fewer

snails of all sizes to graze, especially between 30 and

60 min (Fig. 3d–e; Seastar treatment: F1,856 = 14!37,P = 0!003; Time 9 Seastar treatment: F4,856 = 5!83, P <0!001; Time 9 Seastar treatment 9 Snail Size: F8,856 =0!79, P = 0!611). More small snails grazed than medium or

large snails after 30 min regardless of seastar presence

(Snail size: F1,856 = 19!36, P < 0!001; Time 9 Snail Size:

F4,856 = 4!05, P < 0!001). This was likely because medium

and large snails quickly consumed the algae and stopped

grazing, which may underestimate the potential effects of

seastars on the grazing activity by medium and large snails

and on TMII strength. Only medium snails mediated posi-

tive TMIIs on algae, and large snails surprisingly mediated

negative TMIIs on algae (Fig. 4a; Seastar treatment:

F1,76 = 0!07, P = 0!790; Seastar treatment 9 Snail size:

F3,76 = 2!86, P = 0!042). This negative TMII may be linked

to an increased grazing rate by individual large but not

medium or small snails when Leptasterias were present,

though this effect was not significant (seastar 9 snail treat-

ment: F2,55 = 1!3, P = 0!29; Mean large snail grazing

rate " SE: 0!41 " 0!07 and 0!29 " 0!05 cm2 graz-

ing snail#1 h#1 with and without seastars, respectively, n =14). Not surprisingly, grazing rates increased with snail

size (snail treatment: F2,55 = 18!8, P < 0!001; Mean grazing

rate " SE: 0!15 " 0!04, 0!22 " 0!03 and 0!35 "0!04 cm2 grazing snail#1 h#1 for small, medium and large,

respectively, n = 28).

Fed snails of medium size mediated positive TMIIs on

algae but medium-sized hungry snails did not (Fig. 4b;

Seastar 9 Snail treatment: F3,44 = 3!65, P = 0!033). On

average, fed snails fled from tide pools more

quickly than hungry snails when seastars were present

(Fig. 5a,b; Time 9 Seastar treatment 9 Snail treatment:

F3,407 = 4!69, P = 0!003). By the end of experiments, more

fed than hungry snails left tide pools with seastars

(Fig. 5a,b; Mean % in halo at end " SE: 31!7 " 2!5%and 10!2 " 1!8% of fed and hungry snails, respectively).

Without seastars, very few fed or hungry snails left tide

pools (Fig. 5a,b; Mean % in halo at end " SE:

3!0 " 0!7% and 0!8 " 0!3% of fed and hungry snails,

respectively). Though on average throughout the experi-

ment the presence of Leptasterias caused fed snails to graze

less, hungry snails only reduced grazing somewhat (59%

and 10% decrease in snails grazing, respectively), there

were no statistical differences between the number of fed

and hungry snails grazing with seastar presence (Fig. 5c,d;

Littorina spp.

Tegulafunebralis Lacuna marmorata Mytilus californianus Limpet

ChitonSmall Medium Large

0

20

40

60

80

100

<9 9–12 12–15 15–18 18–21 >21

Out

com

e (%

)

Snail size (mm diameter)

Eaten Survived attack No attack (a) (b)

Fig 2. (a) Diet of Leptasterias spp. sur-veyed throughout the intertidal zone inHorseshoe Cove, California (n = 21 seast-ars). (b) Percentage of Tegula funebralis indifferent size classes (3 mm increments)that were eaten, survived an attack or notattacked by Leptasterias spp. when snailsand seastars were paired in small tanks for16 days in flowing seawater in the labora-tory.



Time 9 Seastar treatment: F3,407 = 5!25, P = 0!001;Time 9 Seastar treatment 9 Snail treatment: F3,407 = 1!76,P = 0!153).

Discussion

We link state-dependent behaviour to community out-

comes by showing that individual variation in prey hunger

and size may alter the strength of TMII trophic cascades.

Our study confirms model predictions (Schmitz 2000; Lut-

tbeg, Rowe & Mangel 2003; Persson & De Roos 2003) that

individual variation in prey state may change TMII

strength, adding to the growing body of empirical evidence

for this understudied concept (e.g. Ovadia & Schmitz 2002;

Rudolf 2012; Matassa & Trussell 2014). While the very

short-term TMIIs observed here do not necessarily predict

TMII strengths at longer ecological time-scales, we identify

natural circumstances where average size or hunger level in

snail populations may vary over time or space and discuss

the potential ramifications for TMII strength. Further, our

prior experiments suggest that Leptasterias exert positive

long-term TMIIs on algal growth in this system by causing

Tegula to avoid tide pools and reduce grazing for many

months (Gravem 2015), indicating that our short-term

results here may indeed manifest over the long term.

PREY STATE AND TMI IS

Newly collected medium snails escaped from seastars,

reduced grazing and mediated positive TMIIs, suggesting

that they may be important mediators of TMIIs in natural

systems. Newly collected small snails did not graze enough

algae to mediate TMIIs despite strong behavioural

responses. Thus, they may mediate weaker TMIIs than

medium snails in nature, similar to model predictions

where smaller individuals exhibit slower consumption rates

and likely mediate weaker cascading effects (Schmitz 2000;

Persson & De Roos 2003). However at high densities or

over longer periods, small snails probably would have

stronger effects on algae than were observed here. In

addition, over time their energetic demands for growth

may cause them to become less wary and graze more

5

10

15

20

25

30Small Medium Large

Seastars absentSeastars present

0

Snai

ls in

hal

o (%

)

Small

Snai

ls g

razi

ng (%

)

Medium

Time (min)

Large

0 15 30 45 600 15 30 45 600 15 30 45 60

2

4

6

8

10

0

(a) (b) (c)

(d) (e) (f)

Fig 3. Mean ("SE) percentage of Tegulafunebralis in refuge habitats (a–c) and graz-ing (d–f) with and without Leptasteriasspp. present over 1 h in rocky tide pools inHorseshoe Cove, California. Snails weregrouped into size classes (small: <12 mmshell diameter, medium: 12–18 mm, large:>18 mm). Organism densities were scaledto tide pool volume.

None Fed Hungry


0

20

40

60

80

100

None Small Medium Large

Alga

l cov

er (%

)

Snail treatment

(b)(a)

Fig 4. Mean ("SE) percent cover ofmicroalgae remaining on tiles deployed inthe presence or absence of Leptasterias spp.and Tegula funebralis of different (a) sizesafter 1 h or (b) hunger levels after 45 minin rocky tide pools at Horseshoe Cove,California. Snails were grouped into sizeclasses (small: <12 mm shell diameter, med-ium: 12–18 mm, large: >18 mm). Densitiesof organisms and algae tiles were scaled totide pool volume.



frequently regardless of predator presence, similar to small

grasshoppers in other tritrophic TMIIs (Ovadia & Schmitz

2002, 2004). The slower escape response by small snails

was likely due to both slower speed and increased mean-

dering observed in the predator cue experiments in the lab-

oratory (Fig. S1 and Appendix S1).

As in the predator cue experiments, newly collected

large snails in field experiments reacted to seastar presence

by fleeing and grazing less. However, this surprisingly

resulted in negative TMII effects of seastar presence of

algae. This counterintuitive result may have arisen because

individual large snails tended to increase their grazing rates

in the presence of seastars. Thus, when seastars were pre-

sent, fewer large snails grazed but those individuals grazed

faster, likely resulting in lower algal cover when seastars

were present than absent. This may have occurred because

large snails had known low risk of predation and high

opportunity costs of forsaking the available algae, so some

chose to devour the algae quickly before fleeing. However,

we found no other TMII studies indicating that larger or

less vulnerable individuals mediate negative TMIIs because

they increase feeding when predators are present. It is

unclear how our results may translate to TMIIs mediated

by large snails in nature, especially since large snails

appear to be less responsive than medium and small snails

over longer time-scales; they co-occur with Leptasterias

inside tide pools more often than small and medium snails

in field surveys (Gravem 2015) and generally reside

lower in the intertidal zone where Leptasterias and other

predatory seastars are abundant (Paine 1969). Large indi-

viduals in other systems may mediate strong positive

TMIIs because they have strong effects on their resources

but possess the energy reserves to easily stop feeding for

extended periods of time when predators are present (Lut-

tbeg, Rowe & Mangel 2003). Alternatively, they may not

mediate TMIIs because they have reached a size refuge

and continue feeding regardless of predator presence, as

do larger individual sea urchins in the presence of preda-

tory seastars that prefer smaller sea urchin prey (Freeman

2006). Studies on large snail behaviour over varying time-

scales in this system are necessary to determine their role

in the tritrophic cascade.

Hungry snails of medium size responded weakly to

seastar presence and did not mediate TMIIs, suggesting

they may not strongly mediate TMIIs in nature. These

snails apparently risked predation to gain much-needed

energy, similar to predictions of the adaptive foraging the-

ory that suggests prey with lower energy reserves should

forage despite risk (Werner & Anholt 1993; Clark 1994;

Lima 1998). In contrast, fed snails of medium size presum-

ably had higher energetic reserves and they did not risk

foraging when seastars were present. Thus, fed snails did

mediate TMIIs, supporting models predictions that TMIIs

should be stronger when prey have high-energy reserves,

while TMIIs should weaken when low-energy reserves

force prey to continue foraging (Luttbeg, Rowe & Mangel

0

5

10

15

20

25

30

35

40Fed snails

Snai

ls in

hal

o (%

)

0

2

4

6

8

10

Snai

ls g

razi

ng (%

)

Time (min)4530150

Hungry snails

4530150


(a) (b)

(c) (d)

Fig 5. Mean ("SE) percentage of fed andhungry Tegula funebralis in halos (a, b) andgrazing (c, d) with and without Leptasteriasspp. present over 45 min in rocky tidepools at Horseshoe Cove, California. Snailswere either fed microalgae (a, c) or starved(b, d) for 1 week in the laboratory beforeexperiments. Organism densities werescaled to tide pool volume.



2003). Similar results are found in other tritrophic systems;

when tiger sharks were present, sea turtles in better body

condition foraged less often in high-quality seagrass beds

than turtles in worse body condition (Heithaus et al.

2007), and predatory crabs more strongly reduced con-

sumption of mussels by well-fed than hungry intertidal

snails (Matassa & Trussell 2014).

We did not explore the interaction between hunger and

size, but it is possible they are not independent of one

another. Small snails in the field could have lower ener-

getic reserves, and despite fleeing from Leptasterias, they

may eventually re-enter tide pools to graze while large

snails may be able to delay foraging for longer periods

(Peters 1986; Luttbeg, Rowe & Mangel 2003). In contrast,

the energetic demands of reproduction apparently force

medium and large Tegula (>12 mm) to move lower on the

shore despite higher predation risk by Pisaster ochraceus

Brandt (Paine 1969), so the interplay between size and

energetic reserves in the presence of both predators

remains to be determined.

POTENT IAL CAUSES OF LARGE SNA IL BEHAV IOUR

Interestingly, many newly collected large snails responded

to seastars by fleeing and grazing less, even though they

likely were at low risk of predation (Appendix S1). Here,

the prey’s perceived risk of predation may be more impor-

tant than actual risk for determining prey behaviour (Stan-

kowich & Blumstein 2005). Since snails rely more on

chemical than visual cues (Phillips 1978), large snails may

not be able to detect that they are larger than their

attacker, and so may behave suboptimally by fleeing.

Though selection should favour large snails that cease

responding to Leptasterias, strong selective pressure to flee

from Leptasterias early in life could be carried over later in

life with little cost (Yarnall 1964). Snails are probably not

reacting to a general seastar cue, because they appear to

distinguish Leptasterias cues from those of other predatory

seastars, such as Pisaster ochraceus (Yarnall 1964; Gravem

2015). Alternatively, evasive behaviour by large snails may

be advantageous because nonlethal attacks prevented

snails from eating, mating and perhaps respiring and

metabolizing normally from hours to as many as 3!6 days

in the laboratory (Appendix S1). Regardless of their seem-

ingly suboptimal short-term responses, large snails seem to

be less responsive to Leptasterias than small and medium

snails over longer periods of time (Gravem 2015), so some

ontogenetic shifts in behaviour are evident.

EVAS IVE STRATEGIES BY SNA ILS

The evasive responses by snails in the laboratory appeared

to depend on the body size of snails and whether predator

cues were tactile or waterborne (Appendix S1). When

touched by seastars, newly collected medium and large

snails immediately fled in predominantly straight lines, but

when exposed to waterborne cues they meandered more

and fled more slowly. Waterborne cues may be diffuse,

without clear directionality, and may have posed a less

imminent threat than tactile cues that have a clear source

posing an immediate threat. Meandering snails may also

have been casting across waterborne scent plumes to sense

filaments of concentrated cues so they could avoid preda-

tory seastars (Zimmer-Faust et al. 1995; DeBose & Nevitt

2008). Unlike medium and large snails, newly collected

small snails meandered frequently when exposed all cue

types, perhaps because they are less likely to ‘outrun’ seast-

ars. A switch from a straight, directed evasion to erratic

zig-zagging or tacking when facing imminent attack is evi-

dent in diverse prey and it effectively increases the distance

between the predator and the prey (Humphries & Driver

1967; Fitzgibbon 1990).

TMI I STRENGTH IN NATURE

We have shown that state-dependent prey behaviour

potentially alters TMII strength in short experiments, but

further experiments are necessary to determine whether

individual variation in size and hunger level do indeed

alter TMIIs in natural tide pools. Temporal mismatches

are common challenges in experiments linking individuals

to communities because decisions made by organisms

occur nearly instantaneously while ecological outcomes

may manifest on much longer time-scales (Schmitz 2000).

Further, some traits such as hunger level are inherently

fleeting so state-dependent behaviours change on shorter

time-scales than ecological outcomes occur. To isolate the

consequences of energetic state, experiments must be short

or must keep energetic reserves static, whereas dynamic

state variability is easier to incorporate in models (Luttbeg,

Rowe & Mangel 2003; Abrams 2008). In this experiment,

it was impossible to maintain uniformly sized or starved

snails in the field for more than one low tide because snails

easily left tide pools at high tide and grazed on naturally

present algae. Our brief experiments may overestimate

TMII strength for two reasons. First, it is common for

short-term TMII experiments to overestimate the true

strength of TMIIs because prey can temporarily abstain

from feeding with little consequence. Over time, beha-

vioural responses and TMIIs may weaken because ener-

getic demands can cause prey to resume feeding even when

predators are present or prey may become acclimated to

predator cues (Luttbeg, Rowe & Mangel 2003; Okuyama

& Bolker 2007). Secondly, by artificially supplying algae,

we may inaccurately estimate TMIIs since algae in the field

can regrow (Okuyama & Bolker 2007). However, our prior

research in this system suggested that Leptasterias caused

Tegula to reduce grazing and avoid tide pools for at least

10 months, thereby benefitting both microalgal and

macroalgal growth over 1 and 8 months, respectively

(Gravem 2015). These longer-term TMIIs occurred in the

absence of cages, which may artificially concentrate chemi-

cal cues and induce unnatural behaviours, and algae grew

naturally so the effects of snail grazing were much more



realistic. Long-term TMIIs were also apparent in tide

pools containing crabs, snails and algae on the east coast

of the USA (Trussell et al. 2004), further suggesting that

our short-term observations here could result in long-term

community effects.

Though our prior studies demonstrated the potential for

long-term TMIIs in this system, the uniformly sized or

starved populations of snails used in the current experi-

ment are unlikely to occur in natural tide pools. However,

the average size or hunger level of snails can sometimes

vary predictably in nature, which may then change TMII

strengths as suggested by our experiments. For example,

average hunger level of snails may be higher and TMIIs

may be weaker in the fall and winter when algae senesce,

during unproductive years with low upwelling, or at high

shore levels where algae are sparse. Alternatively, when

resources are abundant, TMII strength should increase

and DMII strength decrease because prey become more

wary (Wojdak & Luttbeg 2005). Where small snails are

more common, TMII strength may decrease because they

eat less algae, whereas DMII strength may increase

because smaller snails are more vulnerable to predation.

Tegula tend to be smaller at higher shore levels (Paine

1969; Doering & Phillips 1983), and population size struc-

ture is skewed towards juveniles with decreasing latitude

and wave exposure (Frank 1975; Fawcett 1984; Cooper &

Shanks 2011). On the other hand, Leptasterias tends to

occur lower in the intertidal zone than Tegula, so both

TMIIs and DMIIs may be strongest at low shore levels.

Overall, a mosaic of varying TMII strength may exist as

the density of Leptasterias and the density, population size

distribution, and average energetic state of Tegula vary

over space and time.

Conclusion

This study strengthens the connection between behavioural

and community ecology paradigms by demonstrating that

state-dependent foraging behaviour by prey may alter

TMII trophic cascades. Our data support several theoreti-

cal models suggesting that prey body size and energetic

reserves may alter the indirect cascading effects of preda-

tors on lower trophic levels. We add to a small but grow-

ing body of experiments that aim to fulfil the well-

recognized need to better link individual behaviour to

community processes. Further, we illustrate that including

only consumptive effects (predation rates) and assuming

all individuals are the same in trophic cascades may not

always be sufficient to predict outcomes (Rudolf 2012). In

this case, accurate estimates of trophic cascades require

additional elements, including (i) the nonconsumptive

effects of predators on prey foraging rates (as in all

TMIIs), (ii) variation in these nonconsumptive effects

based on prey state (e.g. size and hunger) and (iii) varia-

tion in the direct consumptive effects including size-depen-

dent predation rate and size-dependent grazing rate.

Our insights resulted from conducting interdisciplinary

experiments on the interplay between foraging theory in

behavioural ecology and TMIIs in community ecology,

and this approach is likely to be a productive avenue of

further investigation.

Acknowledgements

Sarah Traiger performed the Leptasterias field diet surveys and Alex VonBoer performed the Tegula size refuge experiments. Olivia Turnross, NedAntell, Austin Taylor, Danielle Smull, Aiko Michot, Alex Van Boer, LarsClark, Kristin Merala, Jonathan Demmer, Ryanne Ardisana, Sarah Trai-ger, Jenny Bayley Katie Mascovich, Dawn Barlow and Catrin Wendtassisted with experiments. Eric Sanford and Brian Gaylord provided helpfulcommentary on the manuscript. Andy Sih and Ted Grosholz assisted withconceptual development. The Conchologists of America, Mildred E.Mathias Foundation, Henry A. Jastro fellowship, UC Davis GraduateGroup in Ecology, the UC Davis Department of Environmental Scienceand Policy, and the Bodega Marine Laboratory Fellowship all providedresearch funds or financial support. Funding was also in part provided byCalifornia SeaGrants R/FISH218, R/MPA14 and R/MPA24 awarded toSteven Morgan, National Science Foundation Biological Oceanographygrants OCE-1334448 and OCE-0927196 awarded to Steven Morgan, andNational Science Foundation GK-12 grant 0841297 awarded to Susan Wil-liams. The University of California Natural Reserve System provided thespace to make this research possible. This work complies with the legal andethical requirements for research and permits were issued by CaliforniaDepartment of Fish and Game (SC-4688), the University of California Nat-ural Reserve System, and the Bodega Marine Laboratory NonindigenousSpecies Use Committee. The authors declare no conflict of interest. This isa contribution of the Bodega Marine Laboratory.

Data accessibilityData available from the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.89qp4 (Gravem & Morgan 2015).

References

Abrams, P.A. (1995) Implications of dynamically variable traits for identi-fying, classifying and measuring direct and indirect effects in ecologicalcommunities. American Naturalist, 146, 112–134.

Abrams, P.A. (2008) Measuring the impact of dynamic antipredator traitson predator-prey-resource interactions. Ecology, 89, 1640–1649.

Agrawal, A.A., Ackerly, D.D., Adler, F., Arnold, A.E., Caceres, C., Doak,D.F. et al. (2007) Filling key gaps in population and community ecol-ogy. Frontiers in Ecology and the Environment, 5, 145–152.

Beckerman, A.P., Petchey, O.L. & Morin, P.J. (2010) Adaptive foragersand community ecology: linking individuals to communities and ecosys-tems. Functional Ecology, 24, 1–6.

Bouchet, P. & Rosenberg, G. (2015) Tegula funebralis. MolluscaBase(2015). Accessed through: World Register of Marine Species at http://www.marinespecies.org/aphia.php?p=taxdetails&id=534190 (accessed2015-09-15).

Clark, C.W. (1994) Antipredator behavior and the asset-protection princi-ple. Behavioral Ecology, 5, 159–170.

Cooper, E.E. & Shanks, A.L. (2011) Latitude and coastline shape correlatewith age-structure of Chlorostoma (Tegula) funebralis populations. Mar-ine Ecology Progress Series, 424, 133–143.

DeBose, J.L. & Nevitt, G.A. (2008) The use of odors at different spatialscales: comparing birds with fish. Journal of Chemical Ecology, 34, 867–881.

Dehn, M.M. (1990) Vigilance for predators: detection and dilution effects.Behavioral Ecology and Sociobiology, 26, 337–342.

Doering, P.H. & Phillips, D.W. (1983) Maintenance of the shore-level sizegradient in the marine snail Tegula funebralis (Adams, A.): importanceof behavioral responses to light and sea star predators. Journal of Exper-imental Marine Biology and Ecology, 67, 159–173.

Fawcett, M.H. (1984) Local and latitudinal variation in predation on anherbivorous marine snail. Ecology, 65, 1214–1230.

Fitzgibbon, C.D. (1990) Antipredator strategies of immature Thomsongazelles – hiding and the prone response. Animal Behaviour, 40 , 846–855.



Flowers, J.M. & Foltz, D.W. (2001) Reconciling molecular systematics andtraditional taxonomy in a species-rich clade of sea stars (Leptasteriassubgenus hexasterias). Marine Biology, 139, 475–483.

Frank, P. (1975) Latitudinal variation in the life history features of theblack turban snail Tegula funebralis (Prosobranchia: Trochidae). MarineBiology, 31, 181–192.

Freeman, A. (2006) Size-dependent trait-mediated indirect interactionsamong sea urchin herbivores. Behavioral Ecology, 17, 182–187.

Gravem, S.A. (2015) Linking antipredator behavior of prey to intertidalzonation and community structure in rocky tidepools. PhD Dissertation,University of California, Davis, CA, USA.

Gravem, S.A. & Morgan, S.G. (2015) Data from: Prey state alters trait-mediated indirect interactions in rocky tidepools. Dryad Digital Reposi-tory, http://dx.doi.org/10.5061/dryad.89qp4.

Hawlena, D. & Schmitz, O.J. (2010) Herbivore physiological response topredation risk and implications for ecosystem nutrient dynamics. Pro-ceedings of the National Academy of Sciences of the United States ofAmerica, 10 7, 15503–15507.

Heithaus, M.R., Frid, A., Wirsing, A.J., Dill, L.M., Fourqurean, J.W.,Burkholder, D. et al. (2007) State-dependent risk-taking by green seaturtles mediates top-down effects of tiger shark intimidation in a marineecosystem. Journal of Animal Ecology, 76, 837–844.

Helfman, G. (1989) Threat-sensitive predator avoidance in damselfish-trumpetfish interactions. Behavioral Ecology and Sociobiology, 24, 47–58.

Houston, A.I., McNamara, J.M. & Hutchinson, J.M.C. (1993) Generalresults concerning the trade-off between gaining energy and avoidingpredation. Philosophical Transactions of the Royal Society of London Ser-ies B: Biological Sciences, 341, 375–397.

Humphries, D. & Driver, P. (1967) Erratic display as a device againstpredators. Science, 156, 1767–1768.

Jurgens, L.J., Rogers-Bennett, L., Raimondi, P.T., Schiebelhut, L.M., Daw-son, M.N., Grosberg, R.K. et al. (2015) Patterns of mass mortalityamong rocky shore invertebrates across 100 km of Northeastern PacificCoastline. PLoS One, 10 , e0126280.

Kotler, B.P., Brown, J.S. & Bouskila, A. (2004) Apprehension and timeallocation in gerbils: the effects of predatory risk and energetic state.Ecology, 85, 917–922.

Lima, S.L. (1998) Stress and decision making under the risk of predation:Recent developments from behavioral, reproductive, and ecological per-spectives. Advances in the Study of Behavior, 27, 215–290.

Luttbeg, B., Rowe, L. & Mangel, M. (2003) Prey state and experimentaldesign affect relative size of trait- and density-mediated indirect effects.Ecology, 84, 1140–1150.

MacArthur, R.H. & Pianka, E.R. (1966) On optimal use of a patchy envi-ronment. American Naturalist, 10 0 , 603–609.

Mangel, M. & Clark, C.W. (1986) Towards a unified foraging theory. Ecol-ogy, 67, 1127–1138.

Matassa, C.M. & Trussell, G.C. (2014) Prey state shapes the effects of tem-poral variation in predation risk. Proceedings of the Royal Society B-Bio-logical Sciences, 281, 8.

Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K. & Relyea, R.A.(2005) Ecological consequences of phenotypic plasticity. Trends in Ecol-ogy and Evolution, 20 , 685–692.

Ohgushi, T., Schmitz, O. & Holt, R.D. (2012) Introduction. Trait-MediatedIndirect Interactions: Ecological and Evolutionary Perspectives (eds T.Ohgushi, O. Schmitz & R.D. Holt), pp. 1–6. Cambridge UniversityPress, New York, USA.

Okuyama, T. & Bolker, B.M. (2007) On quantitative measures of indirectinteractions. Ecology Letters, 10 , 264–271.

Ovadia, O. & Schmitz, O.J. (2002) Linking individuals with ecosystems:experimentally identifying the relevant organizational scale for predictingtrophic abundances. Proceedings of the National Academy of Sciences ofthe United States of America, 99, 12927–12931.

Ovadia, O. & Schmitz, O.J. (2004) Scaling from individuals to food webs:the role of size-dependent responses of prey to predation risk. IsraelJournal of Zoology, 50 , 273–297.

Ovadia, O., Dohna, H.Z., Booth, G. & Schmitz, O.J. (2007) Consequencesof body size variation among herbivores on the strength of plant-herbi-vore interactions in a seasonal environment. Ecological Modelling, 20 6,119–130.

Paine, R.T. (1969) The Pisaster-Tegula interaction - prey patches, predatorfood preference and intertidal community structure. Ecology, 50 , 950–961.

Peacor, S.D. & Werner, E.E. (1997) Trait-mediated indirect interactions ina simple aquatic food web. Ecology, 78, 1146–1156.

Persson, L. & De Roos, A.M. (2003) Adaptive habitat use in size-structuredpopulations: linking individual behavior to population processes. Ecol-ogy, 84, 1129–1139.

Peters, R.H. (1986) The Ecological Implications of Body Size. CambridgeUniversity Press, New York, NY, USA.

Phillips, D.W. (1978) Chemical mediation of invertebrate defensive behav-iors and ability to distinguish between foraging and inactive predators.Marine Biology, 49, 237–243.

Railsback, S.F. & Harvey, B.C. (2013) Trait-mediated trophic interactions:is foraging theory keeping up? Trends in Ecology & Evolution, 28, 119–125.

Rudolf, V. (2012) Trait-mediated indirect interactions in size-structuredpopulations: causes and consequences for species interactions and com-munity dynamics. Trait-Mediated Indirect Interactions: Ecological andEvolutionary Perspectives (eds T. Ohgushi, O. Schmitz & R.D. Holt), pp.69–88. Cambridge University Press, New York, NY, USA.

Schmitz, O.J. (2000) Combining field experiments and individual-basedmodeling to identify the dynamically relevant organizational scale in afield system. Oikos, 89, 471–484.

Schmitz, O.J., Adler, F.R. & Agrawal, A.A. (2003) Linking individual-scaletrait plasticity to community dynamics. Ecology, 84, 1081–1082.

Schmitz, O.J., Krivan, V. & Ovadia, O. (2004) Trophic cascades: the pri-macy of trait-mediated indirect interactions. Ecology Letters, 7, 153–163.

Schmitz, O.J., Grabowski, J.H., Peckarsky, B.L., Preisser, E.L., Trussell,G.C. & Vonesh, J.R. (2008) From individuals to ecosystem function:toward an integration of evolutionary and ecosystem ecology. Ecology,89, 2436–2445.

Sih, A., Bell, A. & Johnson, J.C. (2004) Behavioral syndromes: an ecologi-cal and evolutionary overview. Trends in Ecology and Evolution, 19, 372–378.

Stankowich, T. & Blumstein, D.T. (2005) Fear in animals: a meta-analysisand review of risk assessment. Proceedings of the Royal Society B: Bio-logical Sciences, 272, 2627–2634.

Trussell, G.C., Ewanchuk, P.J., Bertness, M.D. & Silliman, B.R. (2004)Trophic cascades in rocky shore tide pools: distinguishing lethal andnonlethal effects. Oecologia, 139, 427–432.

Werner, E.E. & Anholt, B.R. (1993) Ecological consequences of the trade-off between growth and mortality-rates mediated by foraging activity.American Naturalist, 142, 242–272.

Werner, E.E. & Peacor, S.D. (2003) A review of trait-mediated indirectinteractions in ecological communities. Ecology, 84, 1083–1100.

Wojdak, J.M. & Luttbeg, B. (2005) Relative strengths of trait-mediated anddensity-mediated indirect effects of a predator vary with resource levelsin a freshwater food chain. Oikos, 111, 592–598.

Yarnall, J.L. (1964) The responses of Tegula funebralis to starfishes andpredatory snails (Mollusca: Gastropoda). Veliger, 6, 56–58.

Zimmer-Faust, R.K., Finelli, C.M., Pentcheff, N.D. & Wethey, D.S. (1995)Odor plumes and animal navigation in turbulent water-flow - a field-study. Biological Bulletin, 188, 111–116.

Received 4 May 2015; accepted 11 November 2015Handling Editor: Maud Ferrari

Supporting Information

Additional Supporting information may be found in the online

version of this article:

Appendix S1. Experiments testing the consumptive and non-con-

sumptive effects of Leptasterias on varying sizes of Tegula.

Fig. S1. Responses of different sized Tegula to tactile and water-

borne cues of Leptasterias in the laboratory.



Abstract

Sea star wasting disease (SSWD) first appeared in Oregon in April 2014, and by June had spread to most of the coast. Althoughdelayed compared to areas to the north and south, SSWD was initially most intense in north and central Oregon and spreadsouthward. Up to 90% of individuals showed signs of disease from JuneAugust 2014. In rocky intertidal habitats, populations of thedominant sea star Pisaster ochraceus were rapidly depleted, with magnitudes of decline in density among sites ranging from 2x to9x (59 to 84%) and of biomass from 2.6x to 15.8x (60 to 90%) by September 2014. The frequency of symptomatic individualsdeclined over winter and persisted at a low rate through the spring and summer 2015 (~5–15%, at most sites) and into fall 2015.Disease expression included six symptoms: initially with twisting arms, then deflation and/or lesions, lost arms, losing grip onsubstrate, and final disintegration. SSWD was disproportionally higher in orange individuals, and higher in tidepools. Althoughhistorically P. ochraceus recruitment has been low, from fall 2014 to spring 2015 an unprecedented surge of sea star recruitmentoccurred at all sites, ranging from ~7x to 300x greater than in 2014. The loss of adult and juvenile individuals in 2014 led to adramatic decline in predation rate on mussels compared to the previous two decades. A proximate cause of wasting was likely the“Sea Star associated Densovirus” (SSaDV), but the ultimate factors triggering the epidemic, if any, remain unclear. Although warmtemperature has been proposed as a possible trigger, SSWD in Oregon populations increased with cool temperatures. Since P.ochraceus is a keystone predator that can strongly influence the biodiversity and community structure of the intertidal community,major communitylevel responses to the disease are expected. However, predicting the specific impacts and time course of changeacross west coast metacommunities is difficult, suggesting the need for detailed coastwide investigation of the effects of thisoutbreak.

Published: May 4, 2016 http://dx.doi.org/10.1371/journal.pone.0153994

Sea Star Wasting Disease in the Keystone Predator Pisaster

ochraceus in Oregon: Insights into Differential Population Impacts,

Recovery, Predation Rate, and Temperature Effects from Long-Term

Research

Bruce A. Menge , Elizabeth B. CernyChipman, Angela Johnson, Jenna Sullivan, Sarah Gravem, Francis Chan

http://dx.doi.org/10.1371/journal.pone.0153994

Citation: Menge BA, CernyChipman EB, Johnson A, Sullivan J, Gravem S, Chan F (2016) Sea Star Wasting Disease in theKeystone Predator Pisaster ochraceus in Oregon: Insights into Differential Population Impacts, Recovery, Predation Rate,and Temperature Effects from LongTerm Research. PLoS ONE 11(5): e0153994. doi:10.1371/journal.pone.0153994

Editor: Erik V. Thuesen, The Evergreen State College, UNITED STATES

Received: March 5, 2016; Accepted: April 6, 2016; Published: May 4, 2016

Copyright: © 2016 Menge et al. This is an open access article distributed under the terms of the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authorand source are credited.

Data Availability: All data files are available from the PISCO/Data One database(https://search.dataone.org/#view/doi:10.6085/AA/publication_data.60.1).

Funding: This research was funded by grants from the David and Lucile Packard Foundation (https://www.packard.org), theKingfisher Foundation (www.kingfisherfoundation.org), NSF (http://www.nsf.gov; OCE1448913), and endowments from theWayne and Gladys Valley Foundation (http://fdnweb.org/wgvalley). This is publication 453 from PISCO, the Partnership forInterdisciplinary Studies of Coastal Oceans, funded in part by the David and Lucile Packard Foundation. The funders had norole in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

The outbreak of Sea Star Wasting Disease (SSWD) that began in summer 2013 along the West Coast of North America was one ofthe largest epidemics in a marine ecosystem in recorded history [1–3], rivaling wellknown disease outbreaks in terrestrial systems[4–6]. It affected 20+ species of sea stars ranging from Baja California, Mexico to Alaska, USA and caused severe declines in seastar populations throughout the affected region (http://www.seastarwasting.org). The grotesque manner in which sea stars “melt”(figure 1 in [3], see also Figs AH in S1 Appendix) as a result of the disease has generated scientific and public concern over thefate of these iconic marine species. In addition to seizing public attention, SSWD provides a unique scientific opportunity toinvestigate several important ecological questions. First, studying the time course and intensity of the outbreak informs ourunderstanding of disease outbreaks in marine systems, which are predicted to increase with climate change and increasinganthropogenic impacts (e.g., [7]). Second, recovery of sea star populations after the disease will provide unprecedented insight intoconnections between metapopulations that exchange planktonic larvae. Finally and most importantly, SSWD has not only impactedsea star populations, but is expected to cause drastic shifts in marine communities. In particular, several foundational ecologicalconcepts, including keystone predation, the intermediate predation hypothesis, trophic cascades, and indirect effects were originally

http://creativecommons.org/licenses/by/4.0/

https://search.dataone.org/#view/doi:10.6085/AA/publication_data.60.1

https://www.packard.org/

http://www.kingfisherfoundation.org/

http://www.nsf.gov/

http://fdnweb.org/wgvalley

http://www.seastarwasting.org/

based on experiments with the intertidal keystone predator Pisaster ochraceus [8–10], one of the species devastated by thedisease. Because SSWD has led to massive declines in P. ochraceus, this event provides an unprecedented opportunity to revisitthe generality of these fundamental concepts over space and time.

We believe that, combined with research reported in other publications ([1–3,11], the study of SSWD exemplifies the approachadvocated by Burge et al. [12]: a “union of the modern and the classic” approaches to understanding marine diseases. For severalreasons, the impact of disease on populations and communities is difficult to investigate [3]. When effects are sublethal, forexample, identifying disease as a cause of reduced performance (e.g., growth, reproduction) or a factor underlying mortality andthus population size is generally more difficult than detecting the effects of other species interactions, such as competition,predation or facilitation. Even when lethal, epidemics can be particularly difficult to discern in aquatic habitats, especially in theabsence of longterm observations [12]. Often, disease epidemics occur so fast that ecologists and microbiologists have limitedopportunities to investigate, making pathogen identification, culturing, and determination of specific impacts under controlledsituations very difficult if not impossible. Since some hypothesize that diversity, expression, and prevalence of diseases willincrease in the future as the climate warms and anthropogenic influences on environments increase ([7,13,14], but see [15]),ecological investigations of diseases and their impacts are increasingly needed if we are to forecast and prepare for possible futureecosystem changes.

Sea stars are diverse and ubiquitous in the world’s oceans [16,17], and some sea star species are known to have been susceptibleto “sea star wasting disease” in the past [3]. These outbreaks, however, have been brief and localized, inhibiting intensiveinvestigation. For example, Dungan et al. [18] reported “scores of disintegrating sun stars” during an outbreak in the rocky intertidalsun star Heliaster kubinjii in the Gulf of California in 1978. Unfortunately, the sun star was virtually extinct within three years, andthe cause of the disease was never identified. Similar SSWD outbreaks in sea stars were reported at about the same time (late1970’s; References and Notes in [18,19]), but these too were usually brief (months) and localized. Later reports of SSWD weremade in southern California coincident with El NiñoSouthern Oscillation (ENSO) events [20,21]. Since these outbreaks werelocalized, affected few species, and were relatively brief, pathogens involved were not investigated. In contrast to these previousepidemics, the 2013–15 SSWD event was unprecedented in scale and duration. Because it also occurred in a system we knowwell, the current outbreak provides an ideal opportunity to identify the largescale ecological impacts of disease.

The 2013–2015 Epidemic

We have been studying sites along 360 km of the Oregon coast for 15–32 years. A main focus of our research has beenexamination of community structure and how it varies spatially and temporally in relation to climatological, oceanographic, andother environmental conditions. Thus, we have a detailed historical understanding of this system and knowledge of typical patternsof variation.

Eisenlord et al. (3) reported results from a study similar to ours in Washington state. On the basis of modeling and comparisons oftemperatures to SSWD, and laboratory experiments monitoring mortality rate of infected sea stars under different temperatures,they suggested that warm temperatures triggered the outbreak, at least in the San Juan Islands. Evidence for a thermal effect fromsouthern Puget Sound and one outer coast site was inconclusive.

In spring 2013, after reports that massive sea star dieoffs caused by SSWD had been observed subtidally off Vancouver BC and insouthern and central California, we intensified our ongoing observations on P. ochraceus at our 12 longterm study sites along theOregon coast (Fig 1). Despite the accelerating reports of SSWD in other regions, we saw no convincing evidence of the disease inP. ochraceus populations through March 2014. In April 2014, however, while sampling P. ochraceus populations for wet mass, armlength, feeding, and condition, we encountered a few possiblydiseased sea stars at one site. This observation triggered theinitiation of intensive quantitative surveys of P. ochraceus populations at all sites. Below, we report the results of surveys assessingthe intensity, symptoms, timecourse, and population impacts of the disease on P. ochraceus, including an investigation ofsusceptibility of individuals by age, subhabitat and sea star color. In addition, we detail an unprecedented influx of small recruits ofP. ochraceus, examine the potential causes, and explore whether these recruits will lead to the recovery of the sea star population.We also show that the demise of this ecologically important predator has led to decreased predation on their mussel prey, andspeculate on the potential ramifications for the ecosystem. Finally, we examine the relationship between wasting frequency andenvironmental conditions, including temperature and ocean acidification, finding that SSWD frequency was associated with cool,not warm temperatures.

Fig 1. Map of study sites along the Oregon coast.

Subtidal bathymetry is shown with gray lines, and scaled in m depth.http://dx.doi.org/10.1371/journal.pone.0153994.g001

Study Sites and Methods

Ethics Statement

http://dx.doi.org/10.1371/journal.pone.0153994.g001

http://journals.plos.org/plosone/article/figure/image?size=medium&id=info:doi/10.1371/journal.pone.0153994.g001

Although no animals were collected for this research, the Oregon Department of Fish and Wildlife requires collecting permits for allcoastal researchers. Ours were permit numbers 18510 (2014) and 19272 (2015). All site access was public except for FogartyCreek, whose owners have granted us indefinite access to the site. No protected species were sampled.

Field Surveys

After finding SSWD at all 12 study sites in April 2014 (Fig 1), we focused our intensive surveys on the 9 core sites (all those in Fig 1except Cape Meares, Seal Rock, and Cape Arago) with the most extensive sets of supplementary observations and experiments.Sites were nested within capes (= regions), with (from north to south; see Fig 1) Fogarty Creek (FC) and Boiler Bay (BB) occurringwithin Cape Foulweather (CF); Yachats Beach (YB), Strawberry Hill (SH), and Tokatee Klootchman (TK) occurring within CapePerpetua (CP), and Cape Blanco North (CBN), Cape Blanco South (CBS), Port Orford Heads (POH), and Rocky Point (RP)occurring within Cape Blanco (CB). All sites had P. ochraceus populations, and typical patterns of community structure (see [22,23]for details). Hereafter we refer to site and capes by the codes listed above.

Wasting Surveys. We conducted two types of surveys to assess disease patterns: wasting (WS) and size and density surveys(SDS). WS involved observing and classifying the status of all sea stars we could find during low tides. Each sea star was classifiedas showing disease symptoms or not (“healthy” vs. “sick” or “wasted”). Individuals were then separated into color groups (purple vs.orange), subhabitats (in or out of tidepools), and size/age groups (adults and juveniles; and in 2015, recruits). Those with diseasesymptoms were categorized in order of increasing severity as: (1) having twisted arms, (2–3) lesions and/or being deflated, (4)having recently lost arms, (5) losing their grip on the substratum, or (6) dissolving or disintegrating or “melting” (see [3], Figs AH inS1 Appendix for photos). These surveys were repeated at the core sites as often as possible, ranging in frequency from every fewdays to two weeks (spring and summer 2014, 2015) to months (fall 2014) apart. We conducted a total of 147 surveys from 18 April2014 to 31 October 2015. We avoided touching animals due to uncertainty about the mechanisms of transmission of the disease.Thus, most observations were of symptoms on the aboral surface (occasional observations revealed no symptoms on oralsurfaces), and animals deep in crevices or other inaccessible habitats could not be categorized. In 2014, we kept track of whetherindividuals were “adults” (>70g or about 15 cm in diameter) or “juveniles” (10 to 70 g or about 5–15 cm in diameter; see [3,24]). In2015 surveys we added “recruits” (individuals <10g and <5 cm in diameter) because of a large pulse of small sea stars appearingat many sites between fall 2014 and spring 2015.

We counted sea stars as orange only when they were a distinctive orange hue. All others, including shades of purple, brown,maroon or blue were categorized as purple. New recruits usually began as white or gray (< 2cm diameter) with aboral radial lines ofbluish tint. We arbitrarily categorized these as “purple” for calculation of wasting frequencies. Color differentiation developed withgrowth; most recruits were bluish at first (1–3 cm diameter) and then gradually became mottled hues of purple, brown and orange(2–5 cm diameter).

Size and Density Surveys. In spring and summer we used replicated “belt” transects (n = 5/site) to quantify the sizes of individuals(“standard” arm length, defined as the distance from the madreporite to the tip of the opposite arm) and density (number per m ).Each transect spanned 510m horizontally and 2m vertically in the low zone just below the mussel bed. In each, we counted andmeasured all sea stars, and before the disease (pre2014) we weighed each sea star. During the 2014 outbreak, we did not directlyweigh individuals to avoid serving as agents of disease transmission. Instead, we used sitespecific surveys from summer 2013 at

2

each site to generate linear regressions between arm length and wet mass, and used these to convert arm lengths measured in2014 to estimated wet mass for each individual. We resumed weighing sea stars in summer 2015 after SSWD frequencies haddeclined to mostly single digits.

Settlement

The life cycle of P. ochraceus’ is as follows. Energy from food consumed during summer is initially stored in pyloric caecae, thenduring winter, is transferred to the gonads [25–27]. Mass spawning occurs in spring (in Oregon, typically in May; [26]), after whichthe cycle begins anew. Each female produces millions of eggs, which are fertilized in the water column, and develop intoplanktotrophic bipinnaria larvae (e.g., [28]). Development to the settling brachiolaria larval stage takes ~4–6 weeks and settlementstarts in July, continuing into the fall. In Oregon, surviving recruits (5–15 mm in diameter) are first reliably detectable starting in Aprilthe following spring, when tides switch to daytime and wave action begins to decrease.

We used two methods to quantify the input of new settlers and recruits (individuals < 10g and < 5 cm in diameter or ~2.5 cm instandard arm length) into populations on the shore. To estimate sea star settlement (individuals ~1 mm diameter and ≤ 1 monthafter settlement), we deployed “turfies,” which were squares of astroturf (turf dimensions = 15 x 15 cm; n = 5 per site; e.g., [29])deployed into the low intertidal at four sites, FC, BB, YB and SH. Sampling began in 2002; in 2013 we added RP and CBN sites.Turfies were held against a PVC basal plate that was fastened to the rock using lag screws with eight wing nuts and machinescrews around the edges of the plate. Turfies were deployed in May and exchanged monthly through September or October.Collected turfies were taken to the lab and frozen at 20°C, then processed by shaking each thawed turfy in a plastic dishpan ofwater until all material had been extracted. The contents of the dishpan were poured through a 417 μm sieve and examined undera microscope to obtain counts of recently settled P. ochraceus, crabs, sea urchins and Leptasterias spp. The latter weredistinguishable from P. ochraceus settlers by having six instead of five arms and usually being larger. Though settlement into turfiesdoes not necessarily reflect settlement onto natural surfaces (algae, rock, mussels, barnacles, etc.), and settlers could either settleor crawl onto turfies, they provide a standardized surface for comparison within and among sites.

Recruitment

Quantification of P. ochraceus recruits used the SDS data, in which larger/older recruits ≥ 10 mm in diameter were reliablydetectable. Based on inspection of spring and summer surveys each year, and judging approximate growth from inspection of sizefrequency distributions, we judged that “recruits’ are animals < 3 cm in diameter and are from the previous year’s cohort of settlers.The presumed age of most of these individuals is about 6–9 months, depending on when the transect survey was done in thespring/summer following settlement in the previous late summer or fall.

Predation Rate Experiments

Predation rates on M. californianus by P. ochraceus have been quantified periodically at two sites, BB and SH, since 1990([23,30]), and annually since 2007 at four additional sites (FC, YB/TK, CBN, and POH). Population predation rates were assessedby transplanting clumps of midintertidal zone mussels of 4–6 cm in shell length to pairs (n = 5) of prepared plots, holding themagainst the rock substratum using Vexar mesh covers fastened to the substrate using lag screws inserted into holes with wall®

anchors, and allowing them to reattach for 4–6 weeks [23,30]. After mesh removal, we fastened either complete (Pisaster) orpartial stainless steel mesh fences (+Pisaster) around plots. Treatments were assigned by coin flip. Thereafter, the number ofmussels surviving in each plot were counted every two (summer, early fall) to four (fall, winter) weeks. Experiments at a site wereended when all mussels had been eaten in +Pisaster plots, or in the following May, when new experiments were begun.

Virtually all predation in these plots was due to P. ochraceus. The mussels used were too large for the whelk Nucella ostrina andthe small sixarmed sea star Leptasterias spp. to consume, the whelk N. canaliculata will not eat M. californianus in Oregon[31–34], and crabs Cancer productus were uncommon at all of our sites.

Temperature

To examine the possible role of temperature, we analyzed 2014 data from a longterm intertidal thermal sensor array, whichincluded 3–6 replicate sensors (a mix of HOBO Temperature/Light Data Logger, Onset Computer Corp. UA00264, and HOBOTidbiT v2 Temp Data Logger, Onset Computer Corp. UTBI001) deployed in the low intertidal zone at each site. Loggers recordeddata every two minutes. To protect them from damage, sensors were housed in stainless steel mesh cages held down by lagscrews. We present data from three sites, one each on CF (FC), CP (SH), and CB (CBN) and compare 2014 data to the longterm“climatology;” i.e., the longterm mean temperature plus 1 SD “envelopes” which enable detection of anomalous means.

pH

Intertidal pH was measured using customdesigned Durafet based (Honeywell Inc.) pH sensors [35]. At SH, sensors were securedto the bedrock during low tides at 0.31 m below Mean Lower Low Water, and on a mooring in 15 m depth (4 m sensor depth) justoffshore. Sensors were serviced at approximately 4week intervals. As with intertidal temperature records, data taken when sensorswere out of water during low tides were separated from inwater temperatures using tidal height timeseries and surveyed height ofthe sensors. Data spikes (outliers) were filtered out using a ±4 standard deviation window. Each unit was calibrated directly againstseawater and/or TRISbased certified reference material (CRM) from A. Dickson’s group (Scripps Institute of Oceanography) orindirectly against seawater CRMcalibrated spectrophotometricallydetermined pH samples. Thus, each deployment record wastraceable to a CRM standard. Temperature was also recorded from each pH sensor and calibrated against a manufacturercalibrated Seabird Electronics (SBE) thermograph in the laboratory. Mooring data were collected 15 April to 22 September 2014and intertidal data were collected 16 May to 10 August 2014.

Data Analysis

Data were analyzed using JMP PRO 12.0.1 (SAS Institute Inc., 2015; http://www.jmp.com). For analysis, all data excepttemperatures were normalized using either arcsine (percent; arcsin[sqrt[x*0.01]]) or ln (counts, densities; ln [x+1]) transformations.In all cases, we examined residuals visually to evaluate the assumption of normality and homoscedasticity. Leverage plots wereinspected to identify outliers. We used analysis of variance and/or regression approaches in most data analyses. Linear contrastswere used for posthoc comparisons. Linear regression analyses evaluated how sea star density changed during the SSWDepidemic. Nested 3way ANOVA analyzed the decline in biomass between seasons (spring, summer) and years (2014, 2015) bycape and site nested within cape, and the effects of cape, zone, year, month nested within year and site nested within cape on P.

®

http://www.jmp.com/

ochraceus settlement. Predation rate experiments were analyzed with nested 2way ANOVA, with factors cape, site nested withincape, and “era” (prewasting vs. wasting). Multiple linear regression was used to determine the relationships among SSWDfrequency, pH and temperature (both averaged over the previous week, the previous two weeks and the previous month) atStrawberry Hill. Akaike Information Criterion corrected for finite sample sizes (AICc) was used to determine the best fit model.

Results

Wasting Surveys (WS)

A total of 70,198 P. ochraceus observations were recorded across all sites from April 2014 to July 2015. By late April 2014, SSWDhad infected populations at all our sites, but at low rates (<1%, Fig 2A–2C). Rates were still low (3–5%) two weeks later, but by theend of May 2014, infection frequencies had accelerated at the two most northerly sites, FC and BB. By the end of June, rates wereall >40% at all but the two most southerly sites (Fig 2A–2C; CF = 52.3 ± 3.7%, CP = 39.9 ± 6.1%, CB = 18.4 ± 7.6%; 1way ANOVAon late May/June data; F = 8.11, p = 0.0017, n = 30; adj R = 0.329). Interestingly, although SSWD frequencies increasedexponentially from late May into June, rates of increase were faster to the north than to the south (Fig 2A–2C).

2,272


Fig 2. Wasting frequency at nine (eight in 2015) sites spread across three capes in 2014 and 2015.

A. and D., CF = Cape Foulweather; B. and E., CP = Cape Perpetua; and C. and F., CB = Cape Blanco. Note that the yaxisscale varies among capes for 2015 data. See Fig 1 for site names, here coded as initials.http://dx.doi.org/10.1371/journal.pone.0153994.g002

By mid/late summer 2014, rates of SSWD were all in the 60% to 80% range and remained high through the end of the year, at leastat those sites where weather conditions allowed observations. By spring 2015, wasting had declined to 25% or less at the 8 sitesstill being monitored (Fig 2D–2F). Frequencies were low in summer 2015 (mostly < 10%) and stayed low into September 2015 withthe exception of TK, where the last sample taken had risen sharply to 45% (Fig 2E). A similar, but unquantified September outbreakwas observed at SH (S. Gravem, pers. obs.).

Through most of the summer, and consistent with (3), juvenile sea stars had lower SSWD frequencies than adults (Fig 3). Althoughmonthly frequencies differed only at CP (July and August) and CB sites (August; Fig 3B and 3C), capescale percent SSWD JuneAugust 2014 averages were consistently lower in juveniles vs. adults at CF and CP, but not at CB (Fig 3; matched pairs ttests). AtCB, juvenile SSWD frequency was lower than adults as well if the September sample was dropped (Fig 3C, see caption). BySeptember, SSWD frequencies in juveniles and recruits approached or exceeded those of adults, perhaps because few adultsremained at those sites.


Fig 3. Frequency of small/juvenile P. ochraceus with wasting compared to frequency of adult P. ochraceus with wasting, averaged across siteswithin capes (± 1 SE).

Asterisks indicate months in 2014 in which juveniles had lower % wasting, and matched pairs statistics indicate overalldifferences across months within capes. Data were sine square roottransformed for analysis. If September data aredropped for CB, t = 3.33, p = 0.004, n = 11.http://dx.doi.org/10.1371/journal.pone.0153994.g003

1



In the early summer, adult sea stars of different colors and subhabitats were differentially susceptible to SSWD (Table 1). Purplesea stars made up ~80% and orange individuals 20% of all populations, and sea stars tended to be found more often outside(~73%) than inside (~27%) tidepools. If sea stars across color and subhabitat combinations were equally affected with wasting, theproportion of each color or subhabitat among asymptomatic (“normal”) and symptomatic (“wasting”) should be similar. Early in theAprilNovember field season, purple sea stars outside tidepool were infected less (43.6 ± 6.2% in May) than normal purple seastars (60.0 ± 5.0%). This difference persisted through July (Table 1). Similarly, in early summer SSWD animals of both colors intidepools were infected more frequently than normal animals (Table 1).

Table 1. Comparison of percentages of total number of P. ochraceus that were “normal” or “wasting” (shown in paired columns) in all

combinations of color and subhabitat by month, summer 2014 .

http://dx.doi.org/10.1371/journal.pone.0153994.t001

Late in the 2014 field season, SSWD frequencies had increased in all categories (Fig 2), and disparities among the frequencies ofSSWD and asymptomatic sea stars of different colors and subhabitats had disappeared (Table 1). As suggested by the changingproportions in successive months in Table 1, these changes seem likely due to the declines in numbers of differentially susceptiblesea stars.

With the exception of RP, lesions were the most frequent symptoms observed across the nine populations (Fig I in S1 Appendix). AtRP, animals with twisting, losing grip and deflation were almost as common as animals with lesions. Twisted animals tended to besecond most common, while the frequencies of other symptoms varied among sites in their occurrence.

Through time across sites and capes, twisting was initially the most common symptom, declining as other symptoms became morecommon (Fig 4). Lesion frequency was generally initially low, but increasing at CF and CP sites. At CB, lesion frequencies weremore variable, being high except for a sharp drop in midJuly with a rebound in August. Other symptoms varied through thesummer, with lost arms tending to increase and deflation and disintegration tending to decrease toward summer’s end (Fig 4).

a


http://journals.plos.org/plosone/article/figure/image?size=medium&id=info:doi/10.1371/journal.pone.0153994.t001

Fig 4. Frequencies of sea star wasting symptoms for P. ochraceus through time, summer 2014, at 9 sites.

Top row = Cape Foulweather sites, middle row = Cape Perpetua sites, and bottom row = Cape Blanco sites. See Fig 1caption for site names.http://dx.doi.org/10.1371/journal.pone.0153994.g004

Size and Density Surveys (SDS)

In summer 2014, P. ochraceus populations declined sharply except at YB and RP (Fig 5, Table 2). Increases in July at YB, SH andRP were likely driven by sand incursions, which typically occur at these sites from July to early fall (e.g., [36,37]). Sand can covermuch of the lower intertidal rock surface, forcing P. ochraceus to move higher on the shore (B. Menge, pers. obs.). This upwardmovement tends to increase P. ochraceus transect numbers, artificially increasing densities. Water temperature can also causevariation in P. ochraceus density [31]. Cold upwelled water tends to inhibit activity and P. ochraceus tend to move downwardwhereas during warmer water events after upwelling cessation, sea stars tend to move upward towards mussel beds.




Fig 5. Decline in adult and juvenile P. ochraceus density, spring and summer 2014, at six sites.

Differences in overall average densities (lntransformed data in a 2way ANOVA testing site and month without interactionterm) are shown by lower case letters in legend, where sites sharing the same letter were not different at p > 0.05.http://dx.doi.org/10.1371/journal.pone.0153994.g005

Table 2. Loglinear regressions of change in P. ochraceus density from April through August 2014 .


P. ochraceus density changes from 2014 to 2015 were driven primarily by decreases in abundance of adults across sites, withcrashes as large as 8x to 9x (FC and SH; Fig 6A). Juvenile declines were smaller, and at CBN and RP juvenile numbersincreased slightly in 2015 compared to 2014 (Fig 6B). Overall, however, total P. ochraceus density actually increased slightly in2015 at all sites (except SH) because of exceptionally high recruitment (Fig 6C and 6D; see below).

a




Fig 6. Magnitude of change from summer 2014 to summer 2015 in density (number/m ) of P. ochraceus at seven sites.Sites are color coded by cape, with blue = Cape Foulweather, red = Cape Perpetua, and green = Cape Blanco. A. Change inadult density, B. Change in juvenile density, C. Change in recruit density, D. Change in total sea star density. Weight rangesfor each life history stage are approximate, and based on Menge 1974, which found that animals up to ~70g were nonreproductive and animals from ~100 g and larger had ripe gonads (before spawning). See Fig 2 caption for site names.http://dx.doi.org/10.1371/journal.pone.0153994.g006

SDS data included actual (or in summer 2015, estimated) measurements of sea star wet mass as well as numbers, enablingestimation of changes in P. ochraceus biomass (Fig J in S1 Appendix). In contrast to the sharp drops in density (Fig 5), biomassdeclined slightly from spring to summer at most sites in 2014 (presumably because most remaining individuals were still adults andjuveniles). However, the largest declines occurred between years (Fig J in S1 Appendix inset; nested 3way ANOVA, season p =0.48, year p < 0.0001). The magnitude of decline from 2014 to 2015, and the proportion of biomass remaining in 2015 variedamong sites (3way ANOVA, p < 0.0001), ranging from a decline of 2.6x with 39% remaining at YB to a decline of 15.8x with 6.3%remaining at RP.

Recruitment

2



An unprecedented surge of P. ochraceus recruits occurred between fall 2014 and spring 2015 (Figs 7 and 8). Historically, theproportion of recruits in SDS was low at all sites except BB, especially in the years leading up to the wasting event (2009–2014; Fig7). The 2014–15 declines in adults and juveniles and huge increases in recruits led to a dramatic shift in population size structure(Fig 8). Data for all sites but YB (begun in 2012) extend back to 2000–2001 at these and the other four sites, and show similarpatterns (B. Menge, unpubl. data). Of the seven populations, only BB showed relatively regular annual recruitment (Fig 7) andrelatively high juvenile frequencies (individuals 10 to ~ 100g; Fig 8A) before 2015. All other populations were similar to YB and CBN(Fig 8B and 8C), consisting mostly of adults (~100g or greater; [24,38]). SSWDcaused mortality and subsequent high recruitmentled to a dramatic shift in size structure of P. ochraceus populations in favor of small, nonreproductive sea stars (Fig 8).

Fig 7. Average (+ 1SE) proportion of recruits of P. ochraceus over time at seven sites.

Sample sizes were ~200 individuals per site per sample date. See Methods for details.http://dx.doi.org/10.1371/journal.pone.0153994.g007



Fig 8. Size frequencies of P. ochraceus from 2012 to 2015 at three representative sites, one per cape.

Samples were taken in spring (black bar) and summer (white bar) in each year except for Boiler Bay, where high wave actionprevented collection of the summer 2015 sample. Sample size was ~200. Mean wet weight (± 1 SE) is shown for eachsample date.http://dx.doi.org/10.1371/journal.pone.0153994.g008

Settlement

In contrast to P. ochraceus recruitment, settlement patterns (i.e., individuals about 1–2 mm in diameter) were normal to low in 2014,at least at the three sites for which we have data (Fig 9; FC, BB and SH), and varied with site nested within cape, zone, and month(Table 3). Settlement in 2014 was lower than the previous four years at all sites combined (2010–13; linear contrasts, p < 0.0001),indicating that the heavy 2015 recruitment did not result from unusually high settlement in fall 2014. These observations indicatethat recruits, as defined here, were ~6 to 9 months old (from the settlement stage; ~11–12 months old from the gamete stage).



Fig 9. Annual average (± 1SE) number of P. ochraceus settlers at four sites.Vertical dashed line indicates the onset of SSWD.http://dx.doi.org/10.1371/journal.pone.0153994.g009

Table 3. Test of effects of cape, zone, year and month on P. ochraceus settlement .


Predation Rate Experiments

a





SSWD sea star decline led to immediate effects on predation rate, with exceptionally low rates recorded in 2014 compared to allprior years in which the experiment was conducted, but with variation among sites (Fig 10; Table 4). Prior to 2014, predation ratewas highest at CP sites, moderate at CB sites, and lowest at CF sites (linear contrasts among capes, p < 0.0001 in allcomparisons). In 2014, however, predation rates did not differ among capes or sites (p = 0.28 or higher). Although summer 2014numbers of P. ochraceus were still relatively high at some sites, particularly YB and SH (Fig 5), predation rate was severelydepressed all summer at all sites except POH (Fig 10). Note that POH had the slowest rate of increase in wasting frequency of allsites (Fig 2C).

Fig 10. Comparison of predation rate on mussels, Mytilus californianus, in summer 2014 compared to rates averaged across 1990–2013 at six

sites.

Numbers above each pair of bars show the magnitude of decrease in predation rate.http://dx.doi.org/10.1371/journal.pone.0153994.g010

Table 4. Nested 2way analysis of variance testing the effects of cape, site nested within cape, and era (prewasting vs wasting) on predation

rates on transplanted mussels .


a





Temperature

Warm water and air temperatures have been linked to both previous wasting outbreaks (e.g., [20,21,39]), and to the currentoutbreak in Washington and southern California [3]. Air and water temperature climatologies (data ranges back to 1999 at FC andCBN, to 1993 at SH) in the Oregon intertidal varied among months and years (e.g., Fig 11). Annual mean air temperatures rangedfrom about 9.5 to 12°C while seawater temperature ranged from about 10 to 11.5°C. Overall, temperatures are higher in spring andearly summer, decline in July and August when upwelling is strongest, and increase again in fall as upwelling ceases (Fig 11). Asthe standard deviations in Fig 11 show, however, cooler and warmer temperatures can occur most months.

Fig 11. Air and water temperatures by month at three sites.

Fogarty Creek and Cape Blanco North climatologies (black lines) were averaged from 1999 to 2014, and at Strawberry Hillfrom 1993 to 2014. The red line shows 2014 data. Climatologies are monthly means ± 1 SD and 2014 data are monthlymeans ± 1 SE.http://dx.doi.org/10.1371/journal.pone.0153994.g011

Although temperatures spiked in May 2014, mean temperatures were not anomalous (i.e., different from past values as indicated bythe 1 SD climatology window). At FC and SH, temperature declined in June, July and August, typical for the upwelling season.Exceptionally high temperatures were reached in OctoberDecember (Fig 11, red lines), but wasting % remained at about the samelevels (Fig 2) even as sea star abundances were declining (Fig 5). Patterns at CBN were similar, with 2014 temperatures remainingin the 1 SD climatology window. Thus, in Oregon, although SSWD began after warm spring temperatures, its frequency increasedduring cold temperatures and was high during both cold and warm water and air temperatures.



Although data in Fig 11 are not consistent with a warm temperature effect on wasting, we examined temperature data in more detailto determine if the initially higher SSWD in the north (see Wasting Surveys) was related to temperature. Because relatively hightemperatures were recorded in May and it was possible that this anomaly might have triggered the outbreak, we focused on AprilJune 2014 temperatures (Fig 11).

At the cape scale, air and water temperatures were partly consistent with a decline in temperature from north to south (Fig K in S1Appendix, Table 5). In April, air temperatures did not differ among the three capes, but water temperatures were cooler at thesouthernmost CB (Table 5, linear contrasts). In May, water (and, more weakly, air) temperature declined from Cape Meares (north)to Cape Mendocino (south). With the onset of upwelling (indicated by summer drops in water temperature; Fig 11), June air andwater temperatures among the three capes of focus were more idiosyncratic, with the highest temperatures at the central CP (Fig Kin S1 Appendix, Table 5, air Cape * Month interaction, linear contrasts). Similar, but weaker, trends were seen at the site scale (datanot shown).

Table 5. Twoway analyses of variance testing effects of cape and month on intertidal air and water temperatures at Capes Foulweather,

Perpetua, and Blanco for AprilJune 2014.


We also examined if extreme daily temperatures in April 2014 might be consistent with (as a “trigger”) a northsouth gradient inwasting. Maximum daily air temperatures recorded at CF (north) were 14.12°C (FC) and 13.15°C (BB); at CP (central) were12.07°C (YB), 14.25°C (SH) and 15.46°C (TK); and at CB (south) were 14.03°C (CBN), 12.25°C (POH), and 18.3°C (RP).Maximum water temperatures ranged from 11.66°C to 13.91°C and were similarly intermingled among capes. Thus, whiletemperatures were approximately consistent with a northsouth gradient at the cape scale, no trend was evident for maximum dailytemperatures.

Since mean temperatures may not capture variability in temperature or exposure to extremes, we also examined the relationshipbetween these factors in climatologies vs. 2014. These comparisons were consistent with results for monthly mean temperatures;neither the patterns of variability of air and water temperature nor the maximumminimum envelopes showed unusual patterns in2014 compared to longterm values (Figs L and M in S1 Appendix). The only exception was that the maximum for May 2014 wasslightly higher than the climatology maximum for May; all other 2014 summer maxima were well within the climatology envelopes.

pH



The coastwide occurrence of SSWD suggests that if some environmental factor triggered the disease outbreak, it was also acoastwide event. A steadily increasing factor along the coast (and in the ocean at large) is ocean acidification (e.g., [40–46]). In2014, pH and temperature data were collected every 10 minutes intertidal and subtidal sensors at SH (Fig N in S1 Appendix).These data show that at 4 m depth offshore, sea water temperature slowly increased from about 10 to about 14°C into midMay,2014, then dropped sharply around the 21 of May. pH changes were similar, remaining around 8.1, then dropping rapidly to about7.8 in midMay. The pH then varied, reaching about 7.6 in early June while temperatures changed concurrently but with lessvariability. Intertidal temperatures and pH varied similarly, with drops in both matching those observed on the mooring sensors.

Because datasets were most complete at SH, we quantified the relationship between pH, temperature and the proportion of thepopulation at this site with wasting disease using stepwise multiple regression. The best fit model (AICc) included only temperatureaveraged over the previous two weeks (Fig 12, wasting frequency = 3.1323–0.228[average water T previous two weeks]; p =0.0004, R = 0.668, AICCc = 4.197). Wasting frequency was inversely related to temperature but unrelated to pH.

Fig 12. Loglinear regression between water temperature averaged over the previous two weeks of the sample date and the frequency (arcsin

transformed) of SSWD on the sample date.


Discussion

Our results provide critical insight into one of the largest marine disease events ever recorded. From spring 2014 to summer 2015in Oregon, SSWD caused drastic population declines between 63 and 84% (adult density) and 80 to 99% (biomass) in P.ochraceus, comparable to those seen elsewhere along the US west coast (3). Further, SSWDcaused declines in predation ratesby sea stars on Mytilus californianus, together with insights from classic experimental studies [8,10], suggests that community

st

2



structure may change dramatically throughout the region. Our detailed and extensive chronology of the disease outbreak was onlypossible because of the ongoing longterm research at these sites, and exemplifies the value of funding long term monitoring. Ourability to document the epidemic as it occurred was facilitated by the earlier reports of wasting from Washington, British Columbiaand southern and central California (e.g., http://www.eeb.ucsc.edu/pacificrockyintertidal/dataproducts/seastarwasting/) and therelatively slow progression and persistence of the disease. Though P. ochraceus populations were severely affected, a surge inrecruitment of young sea stars occurred a few months after the outbreak. It does not appear that the disease outbreak triggeredhigh settlement, but it potentially indirectly facilitated high recruit survival. While the recruitment event may facilitate recovery, thedisease is ongoing and the loss of reproductive adults suggests that the supply of propagules will remain low for years. Hence, thepace of recovery of P. ochraceus populations likely hinges on the survival of these recruits.

Variation in Wasting

We highlight several novel results. SSWD varied with age/size, and subhabitat. For example, animals in tidepools were moresusceptible than those out of tidepools. Hypothetically, greater susceptibility in tidepools could result from the longer exposure towaterborne pathogens. However, individuals out of tidepools are likely to become warmer and more stressed during low tide[45–47], which could decrease immune function [1]. Thus, the causal relationship between susceptibility and microhabitat isunclear.

Counterintuitively, during much of 2014 summer, susceptibility was lower in juveniles compared to adults (Fig 3, adults = 55.3 ±2.6% vs. juveniles 3.9 ± 0.9%). The mechanistic basis for this apparent resistance is unclear and under investigation [1,11].Possibilities could include: (1) higher frequencies of resistant genotypes or phenotypes among juveniles and recruits, (2)physiological differences between adults and small sea stars that make adults more susceptible to SSWD, (3) higher surface areasand larger volumes of water within the water vascular system in adults to “intercept” or retain waterborne pathogens, respectively,and (4) small size may make recruits harder to detect because they may die “faster.”

Finally, wasting varied in timing, first occurring most intensely northward and spreading southward. The rate of increase variedamong capes and sites, but disease frequency persisted at a high rate (~60–80%) into the late fall. SSWD frequency declined overwinter and persisted but at low levels in spring and summer 2015 except at two sites, where a fresh outbreak began in late summer.The mechanistic basis of this trend is unclear.

Ecological Consequences

Unlike some prior wasting events (e.g., [19–21,39]), this event was coast wide, ranging from Alaska to Baja California. It thus ranksamong the most extensive disease events ever recorded for a marine species [2,3]. Its consequences may be among the largestcoastalscale biological disruptions to a large marine ecosystem, joining sea star dieoffs from 1982–83 and 1997–98 El Niñoevents, the dieoff of sea stars (Heliaster kubinjii) in the Gulf of California in the 1970’s, of sea urchins (Diadema antillarum) in theCaribbean in the 1980s, and the historic humancaused extirpation of sea otters (Enhydra lutris) across much of their Pacificcoastal range (e.g., [18,48–51]).

http://www.eeb.ucsc.edu/pacificrockyintertidal/data-products/sea-star-wasting/

A drastic reduction in predation rate was an immediate consequence of SSWD (Fig 10). A hallmark of prior experiments rangingback to 1990 was sitespecific consistency of predation rate up to ~2009 (e.g., [52]). Although predation rates differed among sites,annual rates at each site tended to be similar. In summer 2014, however, even though some sites retained large numbers ofasymptomatic animals into midsummer (Fig 5, SH, YB, and RP), predation rate dropped sharply. This change was not likely tohave been due to cold water temperatures, which occurred during the height of SSWD. Although prior research showed that coldwater temperatures (e.g., those occurring during summer upwelling) can inhibit feeding, P. ochraceus can acclimate and if coldwater events persist, feeding rates will increase ([31,32]). Further, unusually low predation rates persisted into fall 2014 whenwaters had warmed after the cessation of upwelling (B. Menge unpubl. data). We conclude that the steep drop in predation ratewas driven by SSWD, with behavioral effects possibly occurring before external symptoms of the disease were apparent.

Based on prior research [8,10], the longerterm ecological consequences of this SSWD event could include wholesale eliminationof many low zone species and a complete change in the zonation patterns of rocky intertidal communities along the west coast ofNorth America. In other words, P. ochraceus loss may shift the low zone to an alternate community state dominated by mussels,which could potentially persist for years to decades [53]. If mussel abundance change does track the “alternate state” hypothesis, itwould result in losses or large reductions of many species of macrophytes, anemones, limpets, chitons, sea urchins and otherorganisms from the low intertidal zone. For many of these species presence in the low intertidal zone represents upward extensionsof their subtidal ranges (e.g., [54,55]), so unless mussels also invade the subtidal (which is a possibility), they are likely to persistsubtidally. Other species, however, are restricted to the intertidal (e.g., Saccharina sessilus, Katharina tunicata) and may not persistin a musseldominated zone. M. californianus shells typically provide substratum for many other organisms (“epibionts”), includingbarnacles, M. trossulus, limpets, herbivorous gastropods, and some macroalgal species (B. Menge unpubl. data). Further, the“byssal forest” under the mussel bed provides shelter for dozens of species [56,57], so many taxa may actually benefit from theexpected expansion of the mussel bed into the low intertidal zone. Overall, although it is likely that community structure will changeconsiderably, how spatially “universal” (i.e., widespread along the coast) this effect will be is unclear.

An alternative to changes occurring uniformly along the coast is a “mosaic” response. That is, the low intertidal zone will becomedominated by M. californianus at some sites but not at others. For example, mussels might dominate at waveexposed areas, butnot at wave sheltered ones. Many complex interactions occur within P. ochraceusdominated low intertidal communities, and thesecould also influence the rates or direction of change. Thus, mussels may lose dominance at waveexposed areas where alternativepredators become more influential than at present in affecting intertidal community structure. Alternative predators in this systeminclude whelks (Nucella ostrina, N. canaliculata), the small sea star Leptasterias spp., crabs (e.g., Cancer productus), and sea gulls(Larus spp.)[33,58–66]. Facilitation and competition for space, food or light may also occur among various components of the lowzone intertidal causing unexpected community responses [23,37,67–69]. Finally, mussels will only become dominant if they recruitto a location, and recruitment of mussels is highly variable in space [70]. Because wave exposure, compensatory predators, otherspecies interactions, and mussel recruitment may all influence diversity loss and shifts in abundance and species composition, weexpect a mosaic response rather than a wholesale shift to an alternate state along the California Current Large Marine Ecosystem(CCLME).

The near coastwide appearance of unprecedented numbers of recruits in 2015 suggests that P. ochraceus population recoverymay be rapid, but additional considerations could lead to slow recovery. As of summer 2015, we observed that the cohorts ofrecruits appeared to be growing rapidly (perhaps due to a surfeit of small barnacles in the low intertidal zone), and were achieving

“juvenile” status. At this rate, a new cohort of (presumably) reproducing adults may mature within several years on many shores. If,on the other hand, SSWD continues to kill a high proportion of individuals in the coming years, it is possible that many of the currentcohort of recruits will also die off. With the low abundance of adults following SSWD and the slow maturation time in this species,the number of propagules released will likely be low for several years potentially delaying population recovery.

What are the likely implications over the longerterm, or largerscale? If sea star recovery is slow, taking many years, then it seemslikely that mussel invasion of the low zone will be widespread. Precedents for quick recovery, slow recovery, and no recovery existfor sea star mass mortality events. In the New England subtidal, Witman et al. [71] showed that after a huge pulse of musselrecruitment, abundance of predators (mostly the sea star Asterias vulgaris but also crabs) increased and in < 2 years, reducedmussel abundance back to normal, near 0% cover. In southern California, P. ochraceus populations in the Channel Islandsrecovered from severe depletion after the ENSO event of 1997–98 within about 7 years [21]. In this case, however, nearbypopulations of P. ochraceus were unaffected, thus providing a potential source of new recruits for Channel Island populations. The“no recovery” example is the virtual extinction of Heliaster kubinjii in the Gulf of California due to SSWD mentioned earlier [18]. It isunclear which of these examples might apply to the present case. If P. ochraceus adults become scarce along most of their range,then juvenile recolonization may be limited by a lack of reproducing adults due to low numbers of gametes released and/or lowodds of gamete fertilization. Since planktonic larval mortality is typically very high, e.g., 99% or more (e.g., [24]), recovery of P.ochraceus populations may be hindered by a larval abundance bottleneck.

Symptomatology

Twisting or contorting arms and lesions were generally the most common symptom (Fig I in S1 Appendix; see also [1–3]), and ourobservations suggest that arm twisting preceded lesions and other symptoms (Fig 4). Deflation tended to follow arm twisting, butnever was as common as lesions. It is not clear if deflation was an inevitable part of the progression of SSWD, but its mostly lowfrequencies suggest that only some afflicted animals show this symptom. Based on its low frequencies, losing arms may also be asymptom that only some portion of the afflicted animals displayed. Indeed, we observed many disintegrating sea stars that stillpossessed 5 arms.

For animals ultimately dying of SSWD, losing grip and disintegration likely are inevitable symptoms. Both declined in frequencytoward the end of the summer and into the fall (Fig 4M–4R), which is consistent with the major mortality impact of SSWD occurringduring midsummer. These symptoms can progress very fast; we observed one adult animal as “losing grip” early in a tide, and“disintegrating” only four hours later (Fig A in S1 Appendix). Hence, it seems likely that relatively few observations of disintegrationis because this stage is brief, and that wave action during submergence likely washes remains of diseased individuals away.

Changes in Abundance

SSWD caused severe declines in adult and juvenile P. ochraceus populations at nearly all sites. Although density time series goingback to 2001 show that density has fluctuated through time at all sites (B. Menge, unpubl. data), changes as persistent andnegative as those observed during 2014 have not been seen in > 30 years (since 1982–83). Sites where declines were less severe(YB and RP) still had high disease frequencies, and as suggested earlier we believe the apparent lack of declines were driven byfactors driving sea stars upward including warm temperatures (e.g., [31]), and sand incursions. We strongly doubt that alternate

sources of mortality caused abundance changes, because many years of research suggest that P. ochraceus mortality is normallyvery low and is due to the occasional animal being detached from rock and tumbling in the surf or being captured by sea gulls. Wehave observed no other predators on P. ochraceus (sea otters do not occur along the Oregon coast), and have never, in literally10s of thousands of observations, seen cannibalism among or between adults, juveniles, or recruits, either in the field or laboratory.Indeed, P. ochraceus commonly tend to aggregate in large masses, especially around mussel patches. In these masses and inchannels and crevices, adults are often atop juveniles and recruits, with no sign of impending predation (e.g., no everted stomach,no damage). We conclude that the sharp drops in adult and juvenile density and population biomass were driven entirely by SSWD.

Population Replenishment

What was the cause of the unprecedentedly high recruitment in spring 2015? The most likely mechanism is that reduced P.ochraceus abundance led to increases in successful mussel and barnacle recruitment. In Oregon, dense barnacle and musselrecruitment typically occurs each year from July through November [23,52], coincident with the early benthic life history stage of P.ochraceus. Thus, we hypothesize that in previous years, competition with adults for food suppressed survival of settled P.ochraceus, accounting for the usual low sea star recruitment from at least 2001 to 2014 (Fig 7). Of course, the orders of magnitudedifference in size of the recruits and adults means that adults are likely to prefer larger prey [72]. However, adults are oftenobserved with small barnacles and mussels in their everted stomachs, especially when larger prey are scarce (fall and winter), sothey may deplete prey when recruits are most vulnerable to starvation. Such a possibility has a precedent. Along the coast ofMaine, preydriven pulses of mussel recruitment were proposed as the explanation for high recruitment of sea stars in rockysubtidal habitats [71]. More research is needed to clarify the degree of competition between adult and recruit P. ochraceus for smallbarnacle and mussel prey.

An alternative possibility for high survival might be relaxation of predation. This explanation is unlikely given the lack of observedpredation or cannibalism on P. ochraceus juveniles. Possible physical causes of the 2015 recruitment surge also seem implausible.Although ocean temperatures along the US west coast have increased [73,74] and pH is likely decreasing (e.g., F. Chan et al.unpub. data), these changes have been gradual, not sudden like the recruitment pulse. A final possibility is that the recruit pulsewas a chance event. Historically, some echinoderms have been shown to have highly pulsed recruitment [75,76] with no obviouscause. However, the widespread dense P. ochraceus recruitment along the entire coast suggests random chance is unlikely.

Possible Factors Underlying Disease

The proximate factor causing wasting likely includes infection by a densovirus (SSaDV; [2]). SSaDV occurs in high abundance ininfected animals, and inoculation with a viralsized fraction experimentally and exposure to it in sea water (through cohabitationwith infected individuals) both generated wasting in sea stars in lab experiments. However, SSaDV was evidently present in all seastars tested, even asymptomatic individuals, and occurred in the surrounding water column [2]. It was also found in museumspecimens preserved as early as 1942, so has been present in these systems for decades. Thus, whether SSaDV suddenlyincreased in abundance, overwhelming the immunological systems of sea stars, or increased in virulence is unclear. Hewson andcolleagues are pursuing these questions and continuing the search for additional or complementary pathogens in SSWD (I.Hewson, personal communication).

High densities of sea star populations has been suggested as a possible factor underlying the SSWD outbreak off Vancouver, BC[1]. In Oregon, however, SSWD frequencies were high at all sites during summer, despite the widely different P. ochraceusdensities (Figs 2 and 5).

What are likely environmental changes that could have contributed to the recent outbreak of wasting disease in Oregon? As ourtemperature and pH analyses suggest (Fig 12, Figs KN in S1 Appendix), evidence for climate as a direct trigger of wasting is notsupported by our data. Instead, biological factors, for example changes in the abundance or virulence of SSaDV or changes in hostsusceptibility may be either interacting indirectly with environmental triggers in the outbreak or independently of any environmentalchange. Environmental changes could influence disease transmission or host susceptibility if sea stars were experiencingincreased stress and lowered immune response. Although unusually warm temperatures have been cited as a covarying factorassociated with SSWD outbreaks in Washington (3), British Columbia [39], and southern California [20,21], Oregon temperatureswere not unusually high in 2014, and were cooler than most previous years during the height of the outbreak (Fig 11). In fact,wasting frequency was most strongly related to cooler, not warmer temperatures (Fig 12). Water temperatures in central California(Monterey Bay area) were also cooler than usual in fall 2013 when the outbreak occurred, not warmer (P. Raimondi, pers. comm.).We further note that none of the temperatures we recorded were close to those found to be stressful in lab and field studies[45–47,77–80]. Although the May 2014 air and water temperatures were higher than average, in previous years similar relativelyhigh temperatures have been reached in most summer months (e.g., Fig 11, [31,81]) with no evidence of wasting. Finally, the trulyexceptional high temperatures in autumn 2014 did not appear to lead to a change in SSWD, and in fact, SSWD frequency declinedover the subsequent months.

Ocean acidification is also intensifying ([41], F. Chan et al. unpublished data), but at SH, pH and temperature covaried (linearregression; mean pH previous week = 6.19 + 0.18(mean T previous week); p <0.0001, adj R = 0.797, n = 13). Although ouranalysis suggested only temperature was related to SSWD (and inversely, not positively), it seems likely that the ultimateexplanation of SSWD is multifactorial, involving an as yet unknown combination of environmental and biological factors.

Comparison to Other Results

Eisenlord et al. [3]reported results of their research in Washington, and concluded that, in contrast to our analysis, warmtemperatures were an important driver of SSWD in the sheltered waters of the San Juan Islands, with weaker effects in SouthPuget Sound. Consistent with our results, their analysis suggested large animals were more susceptible than small ones. At onesite on the Washington outer coast (Starfish Point) SSWD prevalence ranged up to 62% but peaked in winter 2015 in contrast tolate summer and early fall in the San Juan Islands. Data from this site, probably most comparable to our Oregon sites, wereinsufficient to analyze a relationship with temperature. Although their conclusion of an impact of warm temperatures is consistentwith prior suggestions, the association of high SSWD with cool conditions in Oregon (and Eisenlord et al.’s observation of highfrequency of SSWD at their open coast site in winter, when ocean and air temperatures were likely low) raises questions about thegenerality of the relationship between temperature and wasting. We can envision three explanations: (1) SSWD is unrelated totemperature, (2) multiple pathogens or pathogen genotypes with differing environmental sensitivities are involved and varygeographically (i.e., different forms occur in outer and inner coasts, latitudinally along the coast, etc.), or (3) some other factor,whether intrinsic or extrinsic, is a key factor. Much remains to be learned.

2

Conclusions

SSWD clearly caused dramatic decreases in population density and biomass along the Oregon coast, with large and immediateeffects on the predation rate of by P. ochraceus on mussels at all sites. Initially, adults in tidepools were most strongly affected.Populations were sharply less abundant by late summer 2014. As observed at many locations along the US west coast, a newcohort of P. ochraceus recruits were detected in spring 2015, dramatically altering the size structure of populations and underlyinga recovery in numbers but not biomass. By spring 2015 had declined to moderate levels, ranging from ~2–15% at all sites but two(SH, TK), where new outbreaks appeared in late summer 2015. The mechanism enabling high recruitment could have been a highabundance of small prey (barnacle and mussel recruits), resulting from depletion of adult and juvenile populations. Given conflictingresults on the role of temperature as a trigger of SSWD, it seems most likely that multiple factors interacted in complex ways tocause the outbreak.

Supporting Information

S1 Appendix. Fourteen supporting figures.

Figures show photos of sea stars with wasting symptoms, the proportion of symptoms at 9 sites, the decline in biomass during thewasting outbreak, capescale mean air and water temperatures for April to June 2014, comparison between the air and waterstandard deviations by month in 2014 to the longterm climatology, comparison between maximumminimum air and watertemperatures in 2014 to the longterm climatology, and a comparison between water temperature and pH onshore and on anoffshore mooring at Strawberry Hill.doi:10.1371/journal.pone.0153994.s001(DOCX)

Acknowledgments

We thank Drew Harvell, Morgan Eisenlord, and two anonymous reviewers for valuable comments on the manuscript. Fieldassistance was provided by Jonathan Robinson, Brittany Poirson, Jerod Sapp, Alexander Carsh, Lacey SchrockPurkey, SilkeBachhuber, Lindsey Sadlou, Courtney Jackson, Melissa Linn, Melissa Britsch, Trevor Baley, Kaelie Sivihok, and Skylar Peven.Michael Frenock and Jerod Sapp processed the temperature data.

Author Contributions

Conceived and designed the experiments: BAM. Performed the experiments: BAM AJ ECC JS SG FC. Analyzed the data: BAM.Wrote the paper: BAM AJ ECC JS SG FC.

http://journals.plos.org/plosone/article/asset?unique&id=info:doi/10.1371/journal.pone.0153994.s001

1.

View Article PubMed/NCBI Google Scholar

2.


3.


4.


5.


6.


7.


8.


9.

References

Fuess LE, Eisenlord ME, Closek CL, Tracy AM, Mauntz R, et al. (2015) Up in Arms: Immune and Nervous System Response to Sea Star WastingDisease. PLoS One 10 e0133053. doi: 10.1371/journal.pone.0133053. pmid:26176852

Hewson I, Button JB, Gudenkauf BM, Miner B, Newton AL, et al. (2014) Densovirus associated with sea star wasting disease and mass mortality.Proceedings of the National Academy of Science, USA 111: 17278–17283. doi: 10.1073/pnas.1416625111

Eisenlord ME, Groner ML, Yoshioka RM, Elliott J, Maynard J, et al. (2016) Ochre star mortality during the 2014 wasting disease epizootic: role ofpopulation size structure and temperature. Philosophical Transactions of the Royal Society B 371: 20150212. doi: 10.1098/rstb.2015.0212

Fisher MC, Garner TW, Walker SF (2009) Global emergence of Batrachochytrium dendrobatidis and amphibian chytridiomycosis in space, time, and host.Annual Review of Microbiology 63: 291–310. doi: 10.1146/annurev.micro.091208.073435. pmid:19575560

Vredenburg VT, Knapp RA, Turnstall TS, Briggs CJ (2010) Dynamics of an emerging disease drives largescale amphibian population extinctions.Proceedings of the National Academy of Science USA 107: 9689–9694. doi: 10.1073/pnas.0914111107

Crawford AJ, Lips KR, Bermingham E (2010) Epidemic disease decimates amphibian abundance, species diversity, and evolutionary history in thehighlands of central Panama. Proceedings of the National Academy of Science USA 107: 13777–13782. doi: 10.1073/pnas.0914115107

Harvell CD, Kim K, Burkholder JM, Colwell RR, Epstein PR, et al. (1999) Emerging Marine Diseases: Climate Links and Anthropogenic Factors. Science285: 1505–1510. pmid:10498537 doi: 10.1126/science.285.5433.1505

Paine RT (1966) Food web complexity and species diversity. American Naturalist 100: 65–75. doi: 10.1086/282400

Paine RT (1969) A note on trophic complexity and community stability. American Naturalist 103: 91–93. doi: 10.1086/282586


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Fuess%5Bauthor%5D+AND+Up+in+Arms%3A+Immune+and+Nervous+System+Response+to+Sea+Star+Wasting+Disease

http://scholar.google.com/scholar_lookup?title=Up+in+Arms%3A+Immune+and+Nervous+System+Response+to+Sea+Star+Wasting+Disease&author=Fuess&publication_year=2015

http://dx.doi.org/10.1073/pnas.1416625111

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Hewson%5Bauthor%5D+AND+Densovirus+associated+with+sea+star+wasting+disease+and+mass+mortality

http://scholar.google.com/scholar_lookup?title=Densovirus+associated+with+sea+star+wasting+disease+and+mass+mortality&author=Hewson&publication_year=2014

http://dx.doi.org/10.1098/rstb.2015.0212

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Eisenlord%5Bauthor%5D+AND+Ochre+star+mortality+during+the+2014+wasting+disease+epizootic%3A+role+of+population+size+structure+and+temperature

http://scholar.google.com/scholar_lookup?title=Ochre+star+mortality+during+the+2014+wasting+disease+epizootic%3A+role+of+population+size+structure+and+temperature&author=Eisenlord&publication_year=2016

http://dx.doi.org/10.1146/annurev.micro.091208.073435

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Fisher%5Bauthor%5D+AND+Global+emergence+of+Batrachochytrium+dendrobatidis+and+amphibian+chytridiomycosis+in+space%2C+time%2C+and+host

http://scholar.google.com/scholar_lookup?title=Global+emergence+of+Batrachochytrium+dendrobatidis+and+amphibian+chytridiomycosis+in+space%2C+time%2C+and+host&author=Fisher&publication_year=2009


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Vredenburg%5Bauthor%5D+AND+Dynamics+of+an+emerging+disease+drives+large-scale+amphibian+population+extinctions

http://scholar.google.com/scholar_lookup?title=Dynamics+of+an+emerging+disease+drives+large-scale+amphibian+population+extinctions&author=Vredenburg&publication_year=2010


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Crawford%5Bauthor%5D+AND+Epidemic+disease+decimates+amphibian+abundance%2C+species+diversity%2C+and+evolutionary+history+in+the+highlands+of+central+Panama

http://scholar.google.com/scholar_lookup?title=Epidemic+disease+decimates+amphibian+abundance%2C+species+diversity%2C+and+evolutionary+history+in+the+highlands+of+central+Panama&author=Crawford&publication_year=2010

http://dx.doi.org/10.1126/science.285.5433.1505

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Harvell%5Bauthor%5D+AND+Emerging+Marine+Diseases%3A+Climate+Links+and+Anthropogenic+Factors

http://scholar.google.com/scholar_lookup?title=Emerging+Marine+Diseases%3A+Climate+Links+and+Anthropogenic+Factors&author=Harvell&publication_year=1999

http://dx.doi.org/10.1086/282400

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Paine%5Bauthor%5D+AND+Food+web+complexity+and+species+diversity

http://scholar.google.com/scholar_lookup?title=Food+web+complexity+and+species+diversity&author=Paine&publication_year=1966


10.


11.


12.


13.


14.


15.


16.

17.

18.


19.

Paine RT (1974) Intertidal community structure: experimental studies on the relationship between a dominant competitor and its principal predator.Oecologia (Berlin) 15: 93–120. doi: 10.1007/bf00345739

Wares J, Schiebelhut L (2015) Is there an association between elongation factor 1α overdominance in the seastar Pisaster ochraceus and "seastarwasting disease?". PeerJ in review: 1–9. doi: 10.7717/peerj.1876

Burge CA, Friedman CS, Getchell R, House M, Lafferty KD, et al. (2016) Complementary approaches to diagnosing marine diseases: a union of themodern and the classic. Philosophical Transactions of the Royal Society B 371: 20150207. doi: 10.1098/rstb.2015.0207

Harvell CD, Mitchell CE, Ward JR, Altizer S, Dobson AP, et al. (2002) Climate warming and disease risks for terrestrial and marine biota. Science 296:2158–2162. pmid:12077394 doi: 10.1126/science.1063699

HoeghGuldberg O, Bruno JF (2010) The impact of climate change on the world's marine ecosystems. Science 328: 1523–1528. doi:10.1126/science.1189930. pmid:20558709

Lafferty K (2009) The ecology of climate change and infectious diseases. Ecology 90: 888–900. pmid:19449681 doi: 10.1890/080079.1

Hyman LH (1955) Echinodermata. The Coelomate Bilateria: International Books and Periodicals Supply Service, 1992. 763 p.

Lawrence JM, editor (2013) Starfish: Biology and Ecology of the Asteroidea. Baltimore, MD: The Johns Hopkins University Press. 267 p.

Dungan ML, Miller TE, Thomson DA (1982) Catastrophic decline of a top carnivore in the Gulf of California rocky intertidal zone. Science 216: 989–991.pmid:17809070 doi: 10.1126/science.216.4549.989

Menge BA (1979) Coexistence between the seastars Asterias vulgaris and A. forbesi in a heterogeneous environment: a nonequilibrium explanation.Oecologia 41: 245–272. doi: 10.1007/bf00377430

http://dx.doi.org/10.1086/282586

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Paine%5Bauthor%5D+AND+A+note+on+trophic+complexity+and+community+stability

http://scholar.google.com/scholar_lookup?title=A+note+on+trophic+complexity+and+community+stability&author=Paine&publication_year=1969

http://dx.doi.org/10.1007/bf00345739

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Paine%5Bauthor%5D+AND+Intertidal+community+structure%3A+experimental+studies+on+the+relationship+between+a+dominant+competitor+and+its+principal+predator

http://scholar.google.com/scholar_lookup?title=Intertidal+community+structure%3A+experimental+studies+on+the+relationship+between+a+dominant+competitor+and+its+principal+predator&author=Paine&publication_year=1974

http://dx.doi.org/10.7717/peerj.1876

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Wares%5Bauthor%5D+AND+Is+there+an+association+between+elongation+factor+1-%CE%B1+overdominance+in+the+seastar+Pisaster+ochraceus+and+%22seastar+wasting+disease%3F%22

http://scholar.google.com/scholar_lookup?title=Is+there+an+association+between+elongation+factor+1-%CE%B1+overdominance+in+the+seastar+Pisaster+ochraceus+and+%22seastar+wasting+disease%3F%22&author=Wares&publication_year=2015

http://dx.doi.org/10.1098/rstb.2015.0207

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Burge%5Bauthor%5D+AND+Complementary+approaches+to+diagnosing+marine+diseases%3A+a+union+of+the+modern+and+the+classic

http://scholar.google.com/scholar_lookup?title=Complementary+approaches+to+diagnosing+marine+diseases%3A+a+union+of+the+modern+and+the+classic&author=Burge&publication_year=2016

http://dx.doi.org/10.1126/science.1063699

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Harvell%5Bauthor%5D+AND+Climate+warming+and+disease+risks+for+terrestrial+and+marine+biota

http://scholar.google.com/scholar_lookup?title=Climate+warming+and+disease+risks+for+terrestrial+and+marine+biota&author=Harvell&publication_year=2002


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Hoegh-Guldberg%5Bauthor%5D+AND+The+impact+of+climate+change+on+the+world%27s+marine+ecosystems

http://scholar.google.com/scholar_lookup?title=The+impact+of+climate+change+on+the+world%27s+marine+ecosystems&author=Hoegh-Guldberg&publication_year=2010

http://dx.doi.org/10.1890/08-0079.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Lafferty%5Bauthor%5D+AND+The+ecology+of+climate+change+and+infectious+diseases

http://scholar.google.com/scholar_lookup?title=The+ecology+of+climate+change+and+infectious+diseases&author=Lafferty&publication_year=2009


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Dungan%5Bauthor%5D+AND+Catastrophic+decline+of+a+top+carnivore+in+the+Gulf+of+California+rocky+intertidal+zone

http://scholar.google.com/scholar_lookup?title=Catastrophic+decline+of+a+top+carnivore+in+the+Gulf+of+California+rocky+intertidal+zone&author=Dungan&publication_year=1982


20.


21.


22.


23.


24.


25.


26.


27.


28.

Eckert GL, Engle JM, Kushner DJ (2000) Sea star disease and population declines at the Channel Islands. Proceedings of the California IslandsSymposium. pp. 390–393.

Blanchette CA, Richards DV, Engle JM, Broitman BR, Gaines SD (2005) Regime shifts, community change, and population booms of keystone predatorsat the Channel Islands. Proceedings of the California Channel Islands Symposium.

Menge BA, Gouhier TC, Hacker SD, Chan F, Nielsen KJ (2015) Are metaecosystems organized hierarchically? A model and test in rocky intertidalhabitats. Ecological Monographs 85: 213–233. doi: 10.1890/140113.1

Menge BA, Hacker SD, Freidenburg T, Lubchenco J, Craig R, et al. (2011) Potential impact of climaterelated changes is buffered by differentialresponses to recruitment and interactions. Ecological Monographs 81: 493–509. doi: 10.1890/101508.1

Menge BA (1975) Brood or broadcast? The adaptive significance of different reproductive strategies in the two intertidal sea stars Leptasterias hexactisand Pisaster ochraceus. Marine Biology 31: 87–100. doi: 10.1007/bf00390651

Mauzey KP (1966) Feeding behavior and reproductive cycles in Pisaster ochraceus. Biological Bulletin 131: 127–144. doi: 10.2307/1539653

Sanford E, Menge BA (2007) Reproductive output and consistency of source populations in the sea star Pisaster ochraceus. Marine Ecology ProgressSeries 349: 1–12. doi: 10.3354/meps07166

Farmanfarmaian A, Giese AC, BR A., Bennett J (1958) Annual reproductive cycles in four species of west coast starfishes. Journal of ExperimentalBiology 138: 355–367. doi: 10.1002/jez.1401380209

Strathmann MF (1987) Reproduction and development of marine invertebrates of the Northern Pacific Coast. Seattle: University of Washington Press.

http://dx.doi.org/10.1007/bf00377430

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Coexistence+between+the+seastars+Asterias+vulgaris+and+A.+forbesi+in+a+heterogeneous+environment%3A+a+non-equilibrium+explanation

http://scholar.google.com/scholar_lookup?title=Coexistence+between+the+seastars+Asterias+vulgaris+and+A.+forbesi+in+a+heterogeneous+environment%3A+a+non-equilibrium+explanation&author=Menge&publication_year=1979

http://www.crossref.org/guestquery?auth2=Eckert&atitle2=Sea+star+disease+and+population+declines+at+the+Channel+Islands&auth=Eckert&atitle=Sea+star+disease+and+population+declines+at+the+Channel+Islands

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Eckert%5Bauthor%5D+AND+Sea+star+disease+and+population+declines+at+the+Channel+Islands

http://scholar.google.com/scholar_lookup?title=Sea+star+disease+and+population+declines+at+the+Channel+Islands&author=Eckert&publication_year=2000

http://www.crossref.org/guestquery?auth2=Blanchette&atitle2=Regime+shifts%2C+community+change%2C+and+population+booms+of+keystone+predators+at+the+Channel+Islands&auth=Blanchette&atitle=Regime+shifts%2C+community+change%2C+and+population+booms+of+keystone+predators+at+the+Channel+Islands

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Blanchette%5Bauthor%5D+AND+Regime+shifts%2C+community+change%2C+and+population+booms+of+keystone+predators+at+the+Channel+Islands

http://scholar.google.com/scholar_lookup?title=Regime+shifts%2C+community+change%2C+and+population+booms+of+keystone+predators+at+the+Channel+Islands&author=Blanchette&publication_year=2005

http://dx.doi.org/10.1890/14-0113.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Are+meta-ecosystems+organized+hierarchically%3F+A+model+and+test+in+rocky+intertidal+habitats

http://scholar.google.com/scholar_lookup?title=Are+meta-ecosystems+organized+hierarchically%3F+A+model+and+test+in+rocky+intertidal+habitats&author=Menge&publication_year=2015

http://dx.doi.org/10.1890/10-1508.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Potential+impact+of+climate-related+changes+is+buffered+by+differential+responses+to+recruitment+and+interactions

http://scholar.google.com/scholar_lookup?title=Potential+impact+of+climate-related+changes+is+buffered+by+differential+responses+to+recruitment+and+interactions&author=Menge&publication_year=2011

http://dx.doi.org/10.1007/bf00390651

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Brood+or+broadcast%3F+The+adaptive+significance+of+different+reproductive+strategies+in+the+two+intertidal+sea+stars+Leptasterias+hexactis+and+Pisaster+ochraceus

http://scholar.google.com/scholar_lookup?title=Brood+or+broadcast%3F+The+adaptive+significance+of+different+reproductive+strategies+in+the+two+intertidal+sea+stars+Leptasterias+hexactis+and+Pisaster+ochraceus&author=Menge&publication_year=1975

http://dx.doi.org/10.2307/1539653

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Mauzey%5Bauthor%5D+AND+Feeding+behavior+and+reproductive+cycles+in+Pisaster+ochraceus

http://scholar.google.com/scholar_lookup?title=Feeding+behavior+and+reproductive+cycles+in+Pisaster+ochraceus&author=Mauzey&publication_year=1966

http://dx.doi.org/10.3354/meps07166

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+Reproductive+output+and+consistency+of+source+populations+in+the+sea+star+Pisaster+ochraceus

http://scholar.google.com/scholar_lookup?title=Reproductive+output+and+consistency+of+source+populations+in+the+sea+star+Pisaster+ochraceus&author=Sanford&publication_year=2007

http://dx.doi.org/10.1002/jez.1401380209

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Farmanfarmaian%5Bauthor%5D+AND+Annual+reproductive+cycles+in+four+species+of+west+coast+starfishes

http://scholar.google.com/scholar_lookup?title=Annual+reproductive+cycles+in+four+species+of+west+coast+starfishes&author=Farmanfarmaian&publication_year=1958

29.


30.


31.


32.


33.


34.


35.


36.


37.

Miller BA, Emlet RB (1997) Influence of nearshore hydrodynamics on larval abundance and settlement of sea urchins Strongylocentrotus franciscanusand S. purpuratus in the Oregon upwelling zone. Marine Ecology Progress Series 148: 83–94. doi: 10.3354/meps148083

Menge BA, Blanchette CA, Raimondi P, Freidenburg TL, Gaines SD, et al. (2004) Species interaction strength: testing model predictions along anupwelling gradient. Ecological Monographs 74: 663–684. doi: 10.1890/034060

Sanford E (1999) Regulation of keystone predation by small changes in ocean temperature. Science 283: 2095–2097. pmid:10092235 doi:10.1126/science.283.5410.2095

Sanford E (2002) The feeding, growth, and energetics of two rocky intertidal predators (Pisaster ochraceus and Nucella canaliculata) under watertemperatures simulating episodic upwelling. Journal of Experimental Marine Biology and Ecology 273: 199–218. doi: 10.1016/s00220981(02)001648

Sanford E, Worth DJ (2009) Genetic differences among populations of a marine snail drive geographic variation in predation. Ecology 90: 3108–3118.pmid:19967866 doi: 10.1890/082055.1

Sanford E, Worth DJ (2010) Local adaptation along a continuous coastline: prey recruitment drives differentiation in a predatory snail. Ecology 91: 891–901. pmid:20426346 doi: 10.1890/090536.1

Evans TG, PadillaGamiño JL, Kelly MW, Pespeni MH, Chan F, et al. (2015) Ocean acidification research in the ‘postgenomic’era: Roadmaps from thepurple sea urchin Strongylocentrotus purpuratus. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 185: 33–42. doi:10.1016/j.cbpa.2015.03.007

D'Antonio CM (1986) Role of sand in the domination of hard substrata by the intertidal alga Rhodomela larix. Marine Ecology Progress Series 27: 263–275. doi: 10.3354/meps027263

Menge BA, Berlow EL, Blanchette CA, Navarrete SA, Yamada SB (1994) The keystone species concept: variation in interaction strength in a rockyintertidal habitat. Ecological Monographs 64: 249–286. doi: 10.2307/2937163


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Miller%5Bauthor%5D+AND+Influence+of+nearshore+hydrodynamics+on+larval+abundance+and+settlement+of+sea+urchins+Strongylocentrotus+franciscanus+and+S.+purpuratus+in+the+Oregon+upwelling+zone

http://scholar.google.com/scholar_lookup?title=Influence+of+nearshore+hydrodynamics+on+larval+abundance+and+settlement+of+sea+urchins+Strongylocentrotus+franciscanus+and+S.+purpuratus+in+the+Oregon+upwelling+zone&author=Miller&publication_year=1997

http://dx.doi.org/10.1890/03-4060

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Species+interaction+strength%3A+testing+model+predictions+along+an+upwelling+gradient

http://scholar.google.com/scholar_lookup?title=Species+interaction+strength%3A+testing+model+predictions+along+an+upwelling+gradient&author=Menge&publication_year=2004


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+Regulation+of+keystone+predation+by+small+changes+in+ocean+temperature

http://scholar.google.com/scholar_lookup?title=Regulation+of+keystone+predation+by+small+changes+in+ocean+temperature&author=Sanford&publication_year=1999

http://dx.doi.org/10.1016/s0022-0981(02)00164-8

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+The+feeding%2C+growth%2C+and+energetics+of+two+rocky+intertidal+predators+%28Pisaster+ochraceus+and+Nucella+canaliculata%29+under+water+temperatures+simulating+episodic+upwelling

http://scholar.google.com/scholar_lookup?title=The+feeding%2C+growth%2C+and+energetics+of+two+rocky+intertidal+predators+%28Pisaster+ochraceus+and+Nucella+canaliculata%29+under+water+temperatures+simulating+episodic+upwelling&author=Sanford&publication_year=2002

http://dx.doi.org/10.1890/08-2055.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+Genetic+differences+among+populations+of+a+marine+snail+drive+geographic+variation+in+predation

http://scholar.google.com/scholar_lookup?title=Genetic+differences+among+populations+of+a+marine+snail+drive+geographic+variation+in+predation&author=Sanford&publication_year=2009

http://dx.doi.org/10.1890/09-0536.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+Local+adaptation+along+a+continuous+coastline%3A+prey+recruitment+drives+differentiation+in+a+predatory+snail

http://scholar.google.com/scholar_lookup?title=Local+adaptation+along+a+continuous+coastline%3A+prey+recruitment+drives+differentiation+in+a+predatory+snail&author=Sanford&publication_year=2010

http://dx.doi.org/10.1016/j.cbpa.2015.03.007

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Evans%5Bauthor%5D+AND+Ocean+acidification+research+in+the+%E2%80%98post-genomic%E2%80%99era%3A+Roadmaps+from+the+purple+sea+urchin+Strongylocentrotus+purpuratus

http://scholar.google.com/scholar_lookup?title=Ocean+acidification+research+in+the+%E2%80%98post-genomic%E2%80%99era%3A+Roadmaps+from+the+purple+sea+urchin+Strongylocentrotus+purpuratus&author=Evans&publication_year=2015


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=D%27Antonio%5Bauthor%5D+AND+Role+of+sand+in+the+domination+of+hard+substrata+by+the+intertidal+alga+Rhodomela+larix

http://scholar.google.com/scholar_lookup?title=Role+of+sand+in+the+domination+of+hard+substrata+by+the+intertidal+alga+Rhodomela+larix&author=D%27Antonio&publication_year=1986


38.


39.


40.


41.


42.


43.


44.


45.


46.

Menge BA (1974) Effect of wave action and competition on brooding and reproductive effort in the sea star, Leptasterias hexactisii. Ecology 55: 84–93.doi: 10.2307/1934620

Bates AE, Hilton BJ, Harley CDG (2009) Effects of temperature, season, and locality on wasting disease in the keystone predatory sea star Pisasterochraceus. Diseases of Aquatic Organisms 86: 245–251. doi: 10.3354/dao02125. pmid:20066959

Feely RA, Doney SC, Cooley SR (2009) Ocean acidification: present conditions and future changes in a highCO world. Oceanography 22: 37–47. doi:10.5670/oceanog.2009.95

Feely RA, Sabine CL, HernandezAyon JM, Ianson D, Hales B (2008) Evidence for upwelling of corrosive "acidified" water onto the continental shelf.Science 320: 1490–1492. doi: 10.1126/science.1155676. pmid:18497259

Hauri C, Gruber N, McDonnell AMP, Vogt M (2013) The intensity, duration, and severity of low aragonite saturation state events on the Californiacontinental shelf. Geophysical Research Letters 40: 3424–3428. doi: 10.1002/grl.50618

Hauri C, Gruber N, Plattner GK, Alin S, Feely RA, et al. (2009) Ocean acidification in the California Current system. Oceanography 22: 61–71. doi:10.5670/oceanog.2009.97

Gruber N, Hauri C, Lachkar Z, Loher D, Frölicher TL, et al. (2012) Rapid Progression of Ocean Acidification in the California Current System. Science337: 220–223. doi: 10.1126/science.1216773. pmid:22700658

Pincebourde S, Sanford E, Helmuth B (2009) An intertidal sea star adjusts thermal inertia to avoid extreme body temperature. American Naturalist 174:890–897. doi: 10.1086/648065. pmid:19827942

Szathmary PL, Helmuth BST, Wethey DS (2009) Climate change in the rocky intertidal zone: predicting and measuring the body temperature of a

2

http://dx.doi.org/10.2307/2937163

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+The+keystone+species+concept%3A+variation+in+interaction+strength+in+a+rocky+intertidal+habitat

http://scholar.google.com/scholar_lookup?title=The+keystone+species+concept%3A+variation+in+interaction+strength+in+a+rocky+intertidal+habitat&author=Menge&publication_year=1994

http://dx.doi.org/10.2307/1934620

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Effect+of+wave+action+and+competition+on+brooding+and+reproductive+effort+in+the+sea+star%2C+Leptasterias+hexactisii

http://scholar.google.com/scholar_lookup?title=Effect+of+wave+action+and+competition+on+brooding+and+reproductive+effort+in+the+sea+star%2C+Leptasterias+hexactisii&author=Menge&publication_year=1974

http://dx.doi.org/10.3354/dao02125

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Bates%5Bauthor%5D+AND+Effects+of+temperature%2C+season%2C+and+locality+on+wasting+disease+in+the+keystone+predatory+sea+star+Pisaster+ochraceus

http://scholar.google.com/scholar_lookup?title=Effects+of+temperature%2C+season%2C+and+locality+on+wasting+disease+in+the+keystone+predatory+sea+star+Pisaster+ochraceus&author=Bates&publication_year=2009

http://dx.doi.org/10.5670/oceanog.2009.95

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Feely%5Bauthor%5D+AND+Ocean+acidification%3A+present+conditions+and+future+changes+in+a+high-CO2+world

http://scholar.google.com/scholar_lookup?title=Ocean+acidification%3A+present+conditions+and+future+changes+in+a+high-CO2+world&author=Feely&publication_year=2009


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Feely%5Bauthor%5D+AND+Evidence+for+upwelling+of+corrosive+%22acidified%22+water+onto+the+continental+shelf

http://scholar.google.com/scholar_lookup?title=Evidence+for+upwelling+of+corrosive+%22acidified%22+water+onto+the+continental+shelf&author=Feely&publication_year=2008

http://dx.doi.org/10.1002/grl.50618

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Hauri%5Bauthor%5D+AND+The+intensity%2C+duration%2C+and+severity+of+low+aragonite+saturation+state+events+on+the+California+continental+shelf

http://scholar.google.com/scholar_lookup?title=The+intensity%2C+duration%2C+and+severity+of+low+aragonite+saturation+state+events+on+the+California+continental+shelf&author=Hauri&publication_year=2013

http://dx.doi.org/10.5670/oceanog.2009.97

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Hauri%5Bauthor%5D+AND+Ocean+acidification+in+the+California+Current+system

http://scholar.google.com/scholar_lookup?title=Ocean+acidification+in+the+California+Current+system&author=Hauri&publication_year=2009


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Gruber%5Bauthor%5D+AND+Rapid+Progression+of+Ocean+Acidification+in+the+California+Current+System

http://scholar.google.com/scholar_lookup?title=Rapid+Progression+of+Ocean+Acidification+in+the+California+Current+System&author=Gruber&publication_year=2012

http://dx.doi.org/10.1086/648065

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Pincebourde%5Bauthor%5D+AND+An+intertidal+sea+star+adjusts+thermal+inertia+to+avoid+extreme+body+temperature

http://scholar.google.com/scholar_lookup?title=An+intertidal+sea+star+adjusts+thermal+inertia+to+avoid+extreme+body+temperature&author=Pincebourde&publication_year=2009


47.


48.


49.


50.


51.


52.


53.


54.


keystone predator. Marine Ecology Progress Series 374: 43–56. doi: 10.3354/meps07682

Pincebourde S, Sanford E, Helmuth BST (2008) Body temperature during low tide alters the feeding performance of a top intertidal predator. Limnologyand Oceanography 53: 1562–1573. doi: 10.4319/lo.2008.53.4.1562

Barber RT, Chavez FP (1983) Biological consequences of El Niño. Science 222: 1203–1210. pmid:17806711 doi: 10.1126/science.222.4629.1203

Chavez FP, Strutton PG, Friederich GE, Feely RA, Feldman GC, et al. (1999) Biological and chemical response of the equatorial Pacific Ocean to the1997–98 El Niño. Science 286: 2126–2131. pmid:10591638 doi: 10.1126/science.286.5447.2126

Estes JA, Palmisano JF (1974) Sea otters: their role in structuring nearshore communities. Science 185: 1058–1060. pmid:17738247 doi:10.1126/science.185.4156.1058

Lessios HA (1988) Mass mortality of Diadema antillarum in the Caribbean: What have we learned? Annual Review of Ecology and Systematics 19: 371–393. doi: 10.1146/annurev.ecolsys.19.1.371

Menge BA, Gouhier TC, Freidenburg TL, Lubchenco J (2011) Linking longterm, largescale climatic and environmental variability to patterns of marineinvertebrate recruitment: toward explaining "unexplained' variation. Journal of Experimental Marine Biology and Ecology 400: 236–249. doi:10.1016/j.jembe.2011.02.003

Paine RT, Trimble AC (2004) Abrupt community change on a rocky shore—biological mechanisms contributing to the potential formation of an alternativestate. Ecology Letters 7: 441–445. doi: 10.1111/j.14610248.2004.00601.x

Dayton PK (1971) Competition, disturbance, and community organization: the provision and subsequent utilization of space in a rocky intertidalcommunity. Ecological Monographs 41: 351–389. doi: 10.2307/1948498


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Szathmary%5Bauthor%5D+AND+Climate+change+in+the+rocky+intertidal+zone%3A+predicting+and+measuring+the+body+temperature+of+a+keystone+predator

http://scholar.google.com/scholar_lookup?title=Climate+change+in+the+rocky+intertidal+zone%3A+predicting+and+measuring+the+body+temperature+of+a+keystone+predator&author=Szathmary&publication_year=2009

http://dx.doi.org/10.4319/lo.2008.53.4.1562

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Pincebourde%5Bauthor%5D+AND+Body+temperature+during+low+tide+alters+the+feeding+performance+of+a+top+intertidal+predator

http://scholar.google.com/scholar_lookup?title=Body+temperature+during+low+tide+alters+the+feeding+performance+of+a+top+intertidal+predator&author=Pincebourde&publication_year=2008


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Barber%5Bauthor%5D+AND+Biological+consequences+of+El+Ni%C3%B1o

http://scholar.google.com/scholar_lookup?title=Biological+consequences+of+El+Ni%C3%B1o&author=Barber&publication_year=1983


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Chavez%5Bauthor%5D+AND+Biological+and+chemical+response+of+the+equatorial+Pacific+Ocean+to+the+1997%E2%80%9398+El+Ni%C3%B1o

http://scholar.google.com/scholar_lookup?title=Biological+and+chemical+response+of+the+equatorial+Pacific+Ocean+to+the+1997%E2%80%9398+El+Ni%C3%B1o&author=Chavez&publication_year=1999


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Estes%5Bauthor%5D+AND+Sea+otters%3A+their+role+in+structuring+nearshore+communities

http://scholar.google.com/scholar_lookup?title=Sea+otters%3A+their+role+in+structuring+nearshore+communities&author=Estes&publication_year=1974

http://dx.doi.org/10.1146/annurev.ecolsys.19.1.371

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Lessios%5Bauthor%5D+AND+Mass+mortality+of+Diadema+antillarum+in+the+Caribbean%3A+What+have+we+learned%3F

http://scholar.google.com/scholar_lookup?title=Mass+mortality+of+Diadema+antillarum+in+the+Caribbean%3A+What+have+we+learned%3F&author=Lessios&publication_year=1988

http://dx.doi.org/10.1016/j.jembe.2011.02.003

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Linking+long-term%2C+large-scale+climatic+and+environmental+variability+to+patterns+of+marine+invertebrate+recruitment%3A+toward+explaining+%22unexplained%27+variation

http://scholar.google.com/scholar_lookup?title=Linking+long-term%2C+large-scale+climatic+and+environmental+variability+to+patterns+of+marine+invertebrate+recruitment%3A+toward+explaining+%22unexplained%27+variation&author=Menge&publication_year=2011

http://dx.doi.org/10.1111/j.1461-0248.2004.00601.x

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Paine%5Bauthor%5D+AND+Abrupt+community+change+on+a+rocky+shore%E2%80%94biological+mechanisms+contributing+to+the+potential+formation+of+an+alternative+state

http://scholar.google.com/scholar_lookup?title=Abrupt+community+change+on+a+rocky+shore%E2%80%94biological+mechanisms+contributing+to+the+potential+formation+of+an+alternative+state&author=Paine&publication_year=2004

http://dx.doi.org/10.2307/1948498

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Dayton%5Bauthor%5D+AND+Competition%2C+disturbance%2C+and+community+organization%3A+the+provision+and+subsequent+utilization+of+space+in+a+rocky+intertidal+community

http://scholar.google.com/scholar_lookup?title=Competition%2C+disturbance%2C+and+community+organization%3A+the+provision+and+subsequent+utilization+of+space+in+a+rocky+intertidal+community&author=Dayton&publication_year=1971

55.


56.


57.


58.


59.


60.


61.


62.


63.


Dayton PK (1975) Experimental evaluation of ecological dominance in a rocky intertidal algal community. Ecological Monographs 45: 137–159. doi:10.2307/1942404

Smith JR, Fong P, Ambrose RF (2006) Dramatic declines in mussel bed community diversity: Response to climate change? Ecology 87: 1153–1161.pmid:16761594 doi: 10.1890/00129658(2006)87[1153:ddimbc]2.0.co;2

Suchanek TH (1992) Extreme biodiversity in the marine environment: mussel bed communities of Mytilus californianus. The Northwest EnvironmentalJournal 8: 150–152.

Berlow EL (1997) From canalization to contingency: historical effects in a successional rocky intertidal community. Ecological Monographs 67: 435–460.doi: 10.2307/2963465

Berlow EL (1999) Strong effects of weak interactors in ecological communities. Nature 398: 330–334. doi: 10.1038/18672

Menge BA (1972) Foraging strategy of a starfish in relation to actual prey availability and environmental predictability. Ecological Monographs 42: 25–50.doi: 10.2307/1942229

Navarrete SA, Menge BA (1996) Keystone predation and interaction strength: interactive effects of predators on their main prey. Ecological Monographs66: 409–429. doi: 10.2307/2963488

Navarrete SA, Menge BA, Daley BA (2000) Species interactions in intertidal food webs: prey or predation regulation of intermediate predators? Ecology81: 2264–2277. doi: 10.1890/00129658(2000)081[2264:siiifw]2.0.co;2

Sanford E, Wood ME, Nielsen KJ (2009) A nonlethal method for estimation of gonad and pyloric caecum indices in sea stars. Invertebrate Biology 128:372–380. doi: 10.1111/j.17447410.2009.00182.x

http://dx.doi.org/10.2307/1942404

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Dayton%5Bauthor%5D+AND+Experimental+evaluation+of+ecological+dominance+in+a+rocky+intertidal+algal+community

http://scholar.google.com/scholar_lookup?title=Experimental+evaluation+of+ecological+dominance+in+a+rocky+intertidal+algal+community&author=Dayton&publication_year=1975

http://dx.doi.org/10.1890/0012-9658(2006)87%5B1153:ddimbc%5D2.0.co;2

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Smith%5Bauthor%5D+AND+Dramatic+declines+in+mussel+bed+community+diversity%3A+Response+to+climate+change%3F

http://scholar.google.com/scholar_lookup?title=Dramatic+declines+in+mussel+bed+community+diversity%3A+Response+to+climate+change%3F&author=Smith&publication_year=2006

http://www.crossref.org/guestquery?auth2=Suchanek&atitle2=Extreme+biodiversity+in+the+marine+environment%3A+mussel+bed+communities+of+Mytilus+californianus&auth=Suchanek&atitle=Extreme+biodiversity+in+the+marine+environment%3A+mussel+bed+communities+of+Mytilus+californianus

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Suchanek%5Bauthor%5D+AND+Extreme+biodiversity+in+the+marine+environment%3A+mussel+bed+communities+of+Mytilus+californianus

http://scholar.google.com/scholar_lookup?title=Extreme+biodiversity+in+the+marine+environment%3A+mussel+bed+communities+of+Mytilus+californianus&author=Suchanek&publication_year=1992

http://dx.doi.org/10.2307/2963465

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Berlow%5Bauthor%5D+AND+From+canalization+to+contingency%3A+historical+effects+in+a+successional+rocky+intertidal+community

http://scholar.google.com/scholar_lookup?title=From+canalization+to+contingency%3A+historical+effects+in+a+successional+rocky+intertidal+community&author=Berlow&publication_year=1997

http://dx.doi.org/10.1038/18672

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Berlow%5Bauthor%5D+AND+Strong+effects+of+weak+interactors+in+ecological+communities

http://scholar.google.com/scholar_lookup?title=Strong+effects+of+weak+interactors+in+ecological+communities&author=Berlow&publication_year=1999

http://dx.doi.org/10.2307/1942229

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Foraging+strategy+of+a+starfish+in+relation+to+actual+prey+availability+and+environmental+predictability

http://scholar.google.com/scholar_lookup?title=Foraging+strategy+of+a+starfish+in+relation+to+actual+prey+availability+and+environmental+predictability&author=Menge&publication_year=1972

http://dx.doi.org/10.2307/2963488

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Navarrete%5Bauthor%5D+AND+Keystone+predation+and+interaction+strength%3A+interactive+effects+of+predators+on+their+main+prey

http://scholar.google.com/scholar_lookup?title=Keystone+predation+and+interaction+strength%3A+interactive+effects+of+predators+on+their+main+prey&author=Navarrete&publication_year=1996

http://dx.doi.org/10.1890/0012-9658(2000)081%5B2264:siiifw%5D2.0.co;2

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Navarrete%5Bauthor%5D+AND+Species+interactions+in+intertidal+food+webs%3A+prey+or+predation+regulation+of+intermediate+predators%3F

http://scholar.google.com/scholar_lookup?title=Species+interactions+in+intertidal+food+webs%3A+prey+or+predation+regulation+of+intermediate+predators%3F&author=Navarrete&publication_year=2000

http://dx.doi.org/10.1111/j.1744-7410.2009.00182.x

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sanford%5Bauthor%5D+AND+A+non-lethal+method+for+estimation+of+gonad+and+pyloric+caecum+indices+in+sea+stars

http://scholar.google.com/scholar_lookup?title=A+non-lethal+method+for+estimation+of+gonad+and+pyloric+caecum+indices+in+sea+stars&author=Sanford&publication_year=2009

64.


65.


66.


67.


68.


69.


70.


71.


72.


Wieters EA, Gaines SD, Navarrete SA, Blanchette CA, Menge BA (2008) Scales of dispersal and the biogeography of marine predatorprey interactions.American Naturalist 171: 405–417. doi: 10.1086/527492. pmid:18220482

Wootton JT (1994) Predicting direct and indirect effects: an integrated approach using experiments and path analysis. Ecology 75: 151–165. doi:10.2307/1939391

Wootton JT (2002) Mechanisms of successional dynamics: consumers and the rise and fall of species dominance. Ecological Research 17: 249–260. doi:10.1046/j.14401703.2002.00484.x

Kavanaugh MT, Nielsen KJ, Chan FT, Menge BA, Letelier RM, et al. (2009) Experimental assessment of the effects of shade on an intertidal kelp: Dophytoplankton blooms inhibit growth of opencoast macroalgae?. Limnology and Oceanography 54: 276–288. doi: 10.4319/lo.2009.54.1.0276

Menge BA (1972) Competition for food between two intertidal starfish species and its effect on body size and feeding. Ecology 53: 635–644. doi:10.2307/1934777

Barner AK, Hacker SD, Menge BA, Nielsen KJ (2015) The complex net effect of reciprocal interactions and recruitment facilitation maintains an intertidalkelp community. Journal of Ecology (in press). doi: 10.1111/13652745.12495

Menge BA, Chan F, Nielsen KJ, Di Lorenzo E, Lubchenco J (2009) Climatic variation alters supplyside ecology: impact of climate patterns onphytoplankton and mussel recruitment. Ecological Monographs 79: 379–395. doi: 10.1890/082086.1

Witman JD, Genovese SJ, Bruno JF, McLaughlin JW, Pavlin BI (2003) Massive prey recruitment and the control of rocky subtidal communities on largespatial scales. Ecological Monographs 73: 441–462. doi: 10.1890/014073

Paine RT (1976) Sizelimited predation: an observational and experimental approach with the MytilusPisaster interaction. Ecology 57: 858–873. doi:10.2307/1941053

http://dx.doi.org/10.1086/527492

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Wieters%5Bauthor%5D+AND+Scales+of+dispersal+and+the+biogeography+of+marine+predator-prey+interactions

http://scholar.google.com/scholar_lookup?title=Scales+of+dispersal+and+the+biogeography+of+marine+predator-prey+interactions&author=Wieters&publication_year=2008

http://dx.doi.org/10.2307/1939391

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Wootton%5Bauthor%5D+AND+Predicting+direct+and+indirect+effects%3A+an+integrated+approach+using+experiments+and+path+analysis

http://scholar.google.com/scholar_lookup?title=Predicting+direct+and+indirect+effects%3A+an+integrated+approach+using+experiments+and+path+analysis&author=Wootton&publication_year=1994

http://dx.doi.org/10.1046/j.1440-1703.2002.00484.x

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Wootton%5Bauthor%5D+AND+Mechanisms+of+successional+dynamics%3A+consumers+and+the+rise+and+fall+of+species+dominance

http://scholar.google.com/scholar_lookup?title=Mechanisms+of+successional+dynamics%3A+consumers+and+the+rise+and+fall+of+species+dominance&author=Wootton&publication_year=2002

http://dx.doi.org/10.4319/lo.2009.54.1.0276

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Kavanaugh%5Bauthor%5D+AND+Experimental+assessment+of+the+effects+of+shade+on+an+intertidal+kelp%3A+Do+phytoplankton+blooms+inhibit+growth+of+open-coast+macroalgae%3F

http://scholar.google.com/scholar_lookup?title=Experimental+assessment+of+the+effects+of+shade+on+an+intertidal+kelp%3A+Do+phytoplankton+blooms+inhibit+growth+of+open-coast+macroalgae%3F&author=Kavanaugh&publication_year=2009

http://dx.doi.org/10.2307/1934777

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Competition+for+food+between+two+intertidal+starfish+species+and+its+effect+on+body+size+and+feeding

http://scholar.google.com/scholar_lookup?title=Competition+for+food+between+two+intertidal+starfish+species+and+its+effect+on+body+size+and+feeding&author=Menge&publication_year=1972

http://dx.doi.org/10.1111/1365-2745.12495

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Barner%5Bauthor%5D+AND+The+complex+net+effect+of+reciprocal+interactions+and+recruitment+facilitation+maintains+an+intertidal+kelp+community

http://scholar.google.com/scholar_lookup?title=The+complex+net+effect+of+reciprocal+interactions+and+recruitment+facilitation+maintains+an+intertidal+kelp+community&author=Barner&publication_year=2015

http://dx.doi.org/10.1890/08-2086.1

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Menge%5Bauthor%5D+AND+Climatic+variation+alters+supply-side+ecology%3A+impact+of+climate+patterns+on+phytoplankton+and+mussel+recruitment

http://scholar.google.com/scholar_lookup?title=Climatic+variation+alters+supply-side+ecology%3A+impact+of+climate+patterns+on+phytoplankton+and+mussel+recruitment&author=Menge&publication_year=2009

http://dx.doi.org/10.1890/01-4073

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Witman%5Bauthor%5D+AND+Massive+prey+recruitment+and+the+control+of+rocky+subtidal+communities+on+large+spatial+scales

http://scholar.google.com/scholar_lookup?title=Massive+prey+recruitment+and+the+control+of+rocky+subtidal+communities+on+large+spatial+scales&author=Witman&publication_year=2003

http://dx.doi.org/10.2307/1941053

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Paine%5Bauthor%5D+AND+Size-limited+predation%3A+an+observational+and+experimental+approach+with+the+Mytilus-Pisaster+interaction

http://scholar.google.com/scholar_lookup?title=Size-limited+predation%3A+an+observational+and+experimental+approach+with+the+Mytilus-Pisaster+interaction&author=Paine&publication_year=1976

73.

74.

75.


76.


77.


78.

79.


80.


81.


Hixon MA, V. GS, Robinson WD, Baker CS, Batchelder HP, et al. (2010) Oregon's fish and wildlife in a changing climate. In: Dello KD, Mote PW, editors.Oregon Climate Assessment Report. Corvallis, Oregon: Oregon Climate Change Research Institute, Oregon State University. pp. 268–360.

Huyer A, Barth JA, Chelton DB, Hales B, Kosro PM, et al., editors (2010) Oceanographic Context: Longterm change in Oregon's marine environment.

Ebert T, Russell MP (1988) Latitudinal variation in size structure of the west coast purple sea urchin: A correlation with headlands. Limnology andOceanography 33: 286–294. doi: 10.4319/lo.1988.33.2.0286

Loosanoff VL (1964) Variations in time and intensity of setting of the starfish, Asterias forbesi, in Long Island Sound during a twentyfive year period.Biological Bulletin 126: 423–439. doi: 10.2307/1539311

Petes LE, Mouchka M.E., MilstonClements RH, Momoda TS, Menge BA (2008) Effects of environmental stress on intertidal mussels and their sea starpredators. Oecologia 156: 1432–1939 doi: 10.1007/s004420081018x

Sanford E (2002) Community responses to climate change: links between temperature and keystone predation in a rocky intertidal systems. In: SchneiderSH, Root TL, editors. Wildlife responses to climate change. Washington, D. C.: Island Press. pp. 165–200.

Monaco CJ, Wethey DS, Helmuth B (2014) A Dynamic Energy Budget (DEB) Model for the Keystone Predator Pisaster ochraceus doi:10.1371/journal.pone.0104658. PLoS ONE 9: e104658. pmid:25166351

Pincebourde S, Sanford E, Helmuth B (2013) Survival and arm abscission are linked to regional heterothermy in an intertidal sea star. Journal ofExperimental Biology 216: 2183–2191. doi: 10.1242/jeb.083881. pmid:23720798

Barth JA, Menge BA, Lubchenco J, Chan F, Bane JM, et al. (2007) Delayed upwelling alters nearshore coastal ocean ecosystems in the northernCalifornia current. Proceedings of the National Academy of Science, USA 104: 3719–3724. doi: 10.1073/pnas.0700462104

http://dx.doi.org/10.4319/lo.1988.33.2.0286

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Ebert%5Bauthor%5D+AND+Latitudinal+variation+in+size+structure+of+the+west+coast+purple+sea+urchin%3A+A+correlation+with+headlands

http://scholar.google.com/scholar_lookup?title=Latitudinal+variation+in+size+structure+of+the+west+coast+purple+sea+urchin%3A+A+correlation+with+headlands&author=Ebert&publication_year=1988

http://dx.doi.org/10.2307/1539311

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Loosanoff%5Bauthor%5D+AND+Variations+in+time+and+intensity+of+setting+of+the+starfish%2C+Asterias+forbesi%2C+in+Long+Island+Sound+during+a+twenty-five+year+period

http://scholar.google.com/scholar_lookup?title=Variations+in+time+and+intensity+of+setting+of+the+starfish%2C+Asterias+forbesi%2C+in+Long+Island+Sound+during+a+twenty-five+year+period&author=Loosanoff&publication_year=1964

http://dx.doi.org/10.1007/s00442-008-1018-x

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Petes%5Bauthor%5D+AND+Effects+of+environmental+stress+on+intertidal+mussels+and+their+sea+star+predators

http://scholar.google.com/scholar_lookup?title=Effects+of+environmental+stress+on+intertidal+mussels+and+their+sea+star+predators&author=Petes&publication_year=2008


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Monaco%5Bauthor%5D+AND+A+Dynamic+Energy+Budget+%28DEB%29+Model+for+the+Keystone+Predator+Pisaster+ochraceus

http://scholar.google.com/scholar_lookup?title=A+Dynamic+Energy+Budget+%28DEB%29+Model+for+the+Keystone+Predator+Pisaster+ochraceus&author=Monaco&publication_year=2014

http://dx.doi.org/10.1242/jeb.083881

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Pincebourde%5Bauthor%5D+AND+Survival+and+arm+abscission+are+linked+to+regional+heterothermy+in+an+intertidal+sea+star

http://scholar.google.com/scholar_lookup?title=Survival+and+arm+abscission+are+linked+to+regional+heterothermy+in+an+intertidal+sea+star&author=Pincebourde&publication_year=2013


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Barth%5Bauthor%5D+AND+Delayed+upwelling+alters+nearshore+coastal+ocean+ecosystems+in+the+northern+California+current

http://scholar.google.com/scholar_lookup?title=Delayed+upwelling+alters+nearshore+coastal+ocean+ecosystems+in+the+northern+California+current&author=Barth&publication_year=2007

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 552: 31–46, 2016doi: 10.3354/meps11766

Published June 23

INTRODUCTION

The concept of trophic cascades has undergonesubstantial expansion in recent years. Originally,predators were shown to decrease populations ofherbivores thereby indirectly increasing populationsof primary producers (Hairston et al. 1960, Paine1980), which was known as density-mediated indi-rect interactions (DMII; Abrams 1995, 2007, Peacor &Werner 1997). In recent decades, investigators have

discovered that predators can cause phenotypicallyplastic behavioral, morphological, physiological orlife history trait changes in prey species, indirectlyincreasing populations of primary producers—a typeof trait-mediated indirect interaction (TMII; Abrams1995, 2007, Peacor & Werner 1997). Indeed, TMIImight be largely responsible for the indirect effectsobserved in trophic cascades (Peacor & Werner 2001,Preisser et al. 2005, Trussell et al. 2006, Peckarsky etal. 2008).

© The authors 2016. Open Access under Creative Commons byAttribution Licence. Use, distribution and reproduction are un -restricted. Authors and original publication must be credited.

Publisher: Inter-Research · www.int-res.com

*Corresponding author: [email protected]

Trait-mediated indirect interactions amongresidents of rocky shore tidepools

Steven G. Morgan1,2,*, Sarah A. Gravem1,3, Adam C. Lipus1, Marcos Grabiel1, Benjamin G. Miner4

1Bodega Marine Laboratory, University of California−Davis, Bodega Bay, CA 94923, USA2Environmental Science and Policy, University of California−Davis, Davis, CA 95616, USA

3Department of Integrative Biology, Oregon State University, Corvallis, OR 97331, USA4Department of Biology, Western Washington University, Bellingham, WA 98225, USA

ABSTRACT: Trait-mediated indirect interactions (TMIIs) are an important component of food webstructure and dynamics. We determined whether TMIIs occur in rocky tidepool communities onthe west coast of the USA. In the laboratory, both adults and juveniles of the keystone predatorPisaster ochraceus and adults of a smaller predatory seastar Leptasterias spp. caused the abun-dant herbivorous snail Tegula funebralis to stop foraging and flee the water, inducing a positiveTMII on micro- and macroalgae. Snails preferred 3 common species of macroalgae (Ulva lactuca,Cladophora columbiana and Porphyra spp.) over 4 others, indicating that seastars might providethe strongest benefits to these species in tidepools. In the laboratory, snails responded rapidly toboth species of predatory seastars and many more snails responded than could be eaten; thus,there is a potential for TMIIs to occur in natural populations. Snails responded to waterborne cuesfrom P. ochraceus by reducing grazing and leaving still water, and reducing grazing in laminarflow (0.5 l min−1), resulting in TMII effects at least as far as 75 cm away. Adult P. ochraceus andLeptasterias spp. introduced to tidepools during low tide induced many snails to flee the tidepools.Considerable individual variation occurred in the responses of snails. Medium and large snailsmediated TMIIs and hungry snails were marginally less responsive to seastars potentially alteringTMII strength in nature. Thus, we demonstrated that TMIIs could occur in natural tidepools andshowed how predator and algal identity, predator and prey size, water flow and prey hunger levelmay influence these TMIIs.

KEY WORDS: Trait-mediated indirect interaction · Predator-prey interaction · Chemical cue Community structure · Rocky intertidal tidepools · Nonconsumptive effect

OPENPEN ACCESSCCESS

Mar Ecol Prog Ser 552: 31–46, 2016

More recently, investigators have begun focusingon the context-dependency of TMII and factors alter-ing its strength, such as environmental variation, fea-tures of informational cues, timing of measurementsand individual variation in the traits of mediatingspecies (Luttbeg et al. 2003, Luttbeg & Trussell 2013,Matassa & Trussell 2014, Weissburg et al. 2014,Gravem & Morgan 2016). For example, the transmis-sion of predator cues to prey is strongly influenced bythe environment, and air and water flow are particu-larly important for chemosensory cues (Smee et al.2010, Weissburg et al. 2014). Prey may respond differently depending on predator size or identity,which can influence TMII strength (Peckarsky &McIntosh 1998, Bernot & Turner 2001, Weissburg etal. 2014). Predators will be more likely to benefitsome primary producers more than others becauseprey species usually exhibit preferences for particu-lar forage species (Schmitz 1998). Individual varia-tion in prey species is particularly relevant for behav-iorally mediated TMII because the foraging and theantipredator behavior of prey individuals dependson their traits, including size, hunger and tempera-ment (Werner & Anholt 1993, Lima 1998a, Sih et al.2004). There is mounting evidence that these state-dependent decisions by prey can alter the strength ofTMII (Sih 1992, Freeman 2006). For example, hungryprey individuals are often less responsive to preda-tors, and therefore mediate weaker TMIIs (Heithauset al. 2007, Matassa & Trussell 2014, Gravem & Mor-gan 2016). Large individuals also might have differ-ent antipredator behaviors, vulnerabilities, energeticdemands and feeding rates than smaller prey indi-viduals, which might subsequently affect the strengthof TMII (Luttbeg et al. 2003, Persson & De Roos 2003,Freeman 2006, Gravem & Morgan 2016). Thus, indi-vidual variation in prey state likely alters the strengthof TMII, but empirical tests are necessary to identifywhen individual variation is important, and whethersuch individual variation has community-level con-sequences.

Many TMIIs have been demonstrated in fresh-water communities (Peacor & Werner 2001, Preisseret al. 2005, Trussell et al. 2006), perhaps due to theease of chemoreception of waterborne cues frompredators—the main sensory modality in aquaticinvertebrates (Preisser et al. 2005, Selden et al.2009, Long & Hay 2012). The fewer marine exam-ples of TMII (Dill et al. 2003, Long & Hay 2012) aregenerally from the laboratory (Appleton & Palmer1988, Behrens Yamada et al. 1998, Aschaffenburg2008, Quinn et al. 2012) or from wave-protectedmarine environments such as estuaries (but see Rai-

mondi et al. 2000, Trussell et al. 2004, Heithaus etal. 2007, Wada et al. 2013, Matassa & Trussell 2014).The underrepresentation of examples from open-coast marine envi ronments might be because strongturbulence and advection dilute cues, render chemi-cal cues less detectable and diminish the importanceof TMII (Weissburg & Zimmer-Faust 1993, Weiss-burg et al. 2002, Large et al. 2011). Alternatively,such underrepresentation might simply stem fromthe logistical challenges of working in these envi-ronments during high tide, when cues are beingtransmitted and most behaviors are observable. Tounderstand the importance of TMII in open-coastecosystems, a logical first step is to test their impor-tance in calm, isolated tidepools during low tidewhile chemical cues from predators are being trans-mitted (e.g. Trussell et al. 2004, Gravem & Morgan2016) before conducting experiments during hightides when waves and currents rapidly mix andadvect cues (e.g. Matassa & Trussell 2014). Tide-pools also offer an advantage over other marine sys-tems because replicate communities can be manip-ulated, and habitat shifts out of tidepools can beclearly defined (Trussell et al. 2002, 2004).

We investigated the potential for context-dependency of TMII in tidepool communities onrocky shores of the west coast of the USA. In this sys-tem, the original keystone predator, Pisaster ochra -ceus, and a much smaller predatory seastar, Lepta -sterias spp. (likely a species complex including L.hexactis and L. aequalis; see Flowers & Foltz 2001,Carlton 2007 for details) prey on the abundant her-bivorous snail Tegula funebralis (Yarnall 1964, Paine1969), which in turn grazes micro- and macroalgae,and affects algal biomass and community structure(Nielsen 2001). We began the investigation by deter-mining the feeding preferences of T. funebralis on 7species of common macroalgae, to explore whichspecies might be most strongly influenced by TMIIand to choose the most palatable species for use infeeding trials with seastar predators. We then inves-tigated the context-dependency of TMII in laboratoryand field experiments to determine whether (1) pred-ator identity and life stage, (2) responses by prey towaterborne predator cues in flowing and still water,(3) prey hunger level or (4) prey body size alteredprey responses to predators (grazing and escapebehaviors), and ultimately the strength of TMII onalgae. We hypothesized that both seastar specieswould generate positive TMII, but adult P. ochraceuswould have a more pronounced effect than Leptaste-rias spp. on T. funebralis behavior and indirecteffects on algae, given that it is larger, faster and

32

Morgan et al.: Trait-mediated indirect interactions

likely poses a bigger threat. We also compared simi-lar-sized juvenile P. ochraceus and Leptasterias spp.,expecting that Leptasterias spp. would have strongereffects on snail behavior and algae because it is amore effective predator on mobile prey (Menge1972). Furthermore, we expected that predators farther from prey in the flow would elicit weakerantipredator responses and TMII as the cues becameincreasingly diluted. We also expected smaller snailsto exhibit stronger responses and mediate strongerTMII than larger snails in the presence of Leptaste-rias spp. because Leptasterias spp. is unable to con-sume large snails (Gravem & Morgan 2016). Con-versely, we expected all sizes of snails to respond toand mediate TMII in the presence of juvenile P.ochraceus, since all sizes of snails are vulnerable toadult P. ochraceus and snails might not be able to dis-tinguish P. ochraceus size based on chemical cues(Selden et al. 2009). Finally, we expected that snailresponses and TMIIs would be stronger when snailswere sated rather than hungry because they wouldbe more inclined to suspend feeding when seastarswere present (Werner & Anholt 1993, Clark 1994).Thus, this suite of experiments was intended to re -veal whether TMIIs are likely to occur in nature, andif so, the processes driving them in tidepool commu-nities on the west coast.

MATERIALS AND METHODS

Collections

All experiments were conducted during the springand summer of 2004 to 2014. We collected Tegulafunebralis, Leptasterias spp., Pisaster ochraceus andall macroalgae except Ulva lactuca from HorseshoeCove, adjacent to the Bodega Marine Laboratory innorthern California, USA (38.36° N, 123.1° W). Wecollected sheets of U. lactuca from nearby BodegaHarbor for ease of collection, but it also occurs inHorseshoe Cove. We grew the microalgae for TMIIexperiments in outdoor flow-through seawater tablesfor at least 3 d until it evenly covered the tiles orplates used in the trials. We kept newly collectedsnails and seastars in flow-through seawater tables inthe laboratory for 24 h without food before experi-ments to standardize hunger levels and to acclimateanimals to laboratory conditions. The one exceptionwas for experiments examining the effect of hungerlevel of snails in response to seastars, whereby snailswere starved or fed in the laboratory for 1 wk beforethe experiments began. We measured snails across

the widest diameter of shells, Leptasterias spp. acrossthe widest diameter from the tips of opposing raysand P. ochraceus from the madreporite to the tip ofthe ray.

Palatability of macroalgae

We determined the most palatable species of macro -algae to use in 12 replicate feeding trials by givingseparate T. funebralis (mean diameter: 14.5 ± 0.6 mmSE) a choice of a piece of algae of a standardized size(~2.5 mm2) from 7 species of macroalgae in coveredculture dishes (9 cm diameter × 5 cm tall). Algal spe-cies were Cladophora columbiana, Fucus gardneri,Endocladia muricata, Mastocarpus papillatus, Pelve-tiopsis limitata, Porphyra spp. and Ulva lactuca.Dishes were filled with seawater and bathed in flow-ing seawater in the laboratory. We changed the sea-water and cleaned the dishes twice daily to eliminateany microalgae, and we estimated the percentageof algae eaten compared to the original standard size11 times over 12 d. Because we measured the per-centage of algae eaten on the same pieces of algaeover time, we conducted repeated measures analysisof variance (ANOVA) followed by Tukey’s HSD testto examine the effect of algal species on the arcsinesquare root transformed proportion of algae eaten.Because U. lactuca was the most preferred alga,it was used in subsequent TMII experiments (seeFig. 1).

TMII laboratory experiments

Experimental design and analysis

The basic design of the TMII experiments in thelaboratory was a factorial design of seastar presenceand absence crossed with snail presence andabsence. In all treatments, we placed microalgaethat was covering 1 microscope slide (7.6 × 2.6 cm)or 2 porcelain tiles (2.54 × 2.54 cm), or U. lactuca(~1 g wet weight) in a covered mesocosm half filledwith seawater. In ‘snail present’ treatments, weplaced 1 T. funebralis in each mesocosm. In ‘seastarpresent’ treatments, we added 1 predatory Leptas-terias spp. or P. ochraceus per mesocosm to test forthe effects of seastar presence on snail behavior andTMII. Treatments with a snail but no seastar servedas a con trol for snail behavior without seastars.Treatments without a snail or seastar served as acontrol for the clearing of algae by micrograzers.

33


Treatments without a snail but with a seastar servedas a control for disturbance of algae by the predator.We recorded whether snails were out of the waterevery 5 min for 1 h. Grazing was quantified bymeasuring the cover of microalgae removed using agridded quadrat (1 × 1 cm grid) made of transparentplastic film (the cover on the 2 tiles were averagedwhen present) or weight loss of U. lactuca after blot-ting it dry with tissue paper. For all TMII experi-ments detailed below (ex periments with Leptaste-rias spp., juvenile and adult P. ochraceus, andwaterborne cues in still water), we analyzed themain and interactive effects of the presence of thesnail and seastar on the percentage of algae grazedusing a 2-factor ANOVA. Analyses of the presenceof seastars on the percent time snails spent out ofwater or analyses that did not include the snail-absent controls were analyzed by 1-factor ANOVA.Algae did not change in any of the treatments with-out snails. Proportions were arcsine square roottransformed so they were normally distributed andhomoscedastic, except where noted.

Seastar species and life stage

Though the design was the same, methodologicaldetails differed between trials with adult P. ochraceusand those with juvenile P. ochraceus and Leptasteriasspp. For Leptasterias spp. (mean diameter: 20.0 ±0.05 mm SE) and juvenile P. ochraceus (mean armlength: 33.2 ± 6.2 mm), we used small culture dishes(11 cm diameter × 8 cm depth) as our mesocosms. Wedid not cage seastars because they did not eat snails(11.4 to 21.4 mm diameter), which were similar insize to the seastars. The seastars were selected to besimilar in size to each other to keep the chemical cueconcentrations and encounter rates with snails simi-lar between seastar species. To determine the effectof Leptasterias spp. on grazing by snails, the numberof replicate mesocosms per treatment were 143 forseastar and snail, 141 for snail only, 78 for seastaronly and 73 for neither seastar nor snail. We alsorecorded the percent time spent out of water for 128of the mesocosms for the seastar and snail treatment,126 mesocosms for the snail only treatment and all ofthe seastar only and neither seastar nor snail treat-ments. For juvenile P. ochraceus, we conducted 69replicate mesocosms with a seastar and snail, 63replicates with a snail only, 12 replicates with aseastar only and 12 replicates with neither seastarnor snail. For adult P. ochraceus (6 to 8 cm armlength), we used large, clear plastic bins as our meso-

cosms (40 × 25 × 10 cm) to accommodate their largersize. We added 50 snails to each of 14 replicate meso-cosms for the 2 snail treatments. We caged the sea -stars in plastic mesh pouches to prevent them fromeating snails. We conducted 3 replicate mesocosmsfor controls without seastars or snails, but we did notadd a second control with an empty pouch.

We provided only microalgae to snails in ex -periments with Leptasterias spp. and juvenile P.ochraceus. We provided microalgae or U. lactuca(~1 g wet weight) to snails in experiments with adultP. ochraceus to determine if food type affected theTMII. Experiments with Leptasterias spp. and juve-nile P. ochraceus were conducted for 1 h whilerecording whether snails were out of the water every5 min, and experiments with adult P. ochraceus wereconducted for 8 h (approximate length of a low tide inthe upper intertidal) with mesocosms partially sub-merged in flowing seawater to maintain ambienttemperatures during these longer trials. The numberof snails grazing rather than the percent time out ofwater was recorded at the end of the experiment andanalyzed by ANOVA following arcsine square roottransformation.

Individual variation

During the many replicate trials with Leptasteriasspp. and juvenile P. ochraceus, we were struck by thelarge degree of variation in the responses of snails toseastars and the corresponding effects on algae. Toinvestigate the source of this variation, we specifi-cally examined the effect of snail size (small: ≤13 mmdiameter; large: 14 to 40 mm) for both seastar spe-cies, and hunger (starved or not starved for 24 h) forLeptasterias spp. Unexpectedly, neither size norhunger affected grazing or time spent out of water bysnails (see the Supplement at www.int-res.com/articles/suppl/m552p031_supp.pdf). Consequently,we repeated size experiments with Leptasterias spp.by including more replicates and 3 size classes: small(6 to 12 mm), medium (12 to 18 mm) and large (18 to25 mm), using snails that were collected 1 d beforetrials. The starvation period was increased in hungerexperiments using snails that were collected 1 wkbefore trials, and were starved or allowed to graze onmicroalgae that had been growing naturally in tanksfor ~2 wk. We used 2 unglazed porcelain tiles (2.54 ×2.54 cm) covered with microalgae and medium tolarge Leptasterias spp. (2 to 5 cm diameter with sizerandomized) in both experiments and only medium-sized snails in hunger experiments. Individual snails

34

http://www.int-res.com/articles/suppl/m552p031_supp.pdf

http://www.int-res.com/articles/suppl/m552p031_supp.pdf


were used only once in the experiments. For sizeexperiments, we conducted 26 replicate mesocosmsfor all treatment combinations (size × seastar treat-ment), except the 2 control treatments without snailsfor which there were 9 replicate mesocosms each.For hunger experiments, we conducted 52 replicatemesocosms for the 2 treatments with seastars and 51replicate mesocosms for the 2 treatments withoutseastars, except there were only 12 replicate meso-cosms each for the 2 control treatments withoutsnails. To obtain this many replicates, we conductedthe size experiments on 3 different dates and thehunger experiments on 4 different dates, with treat-ments evenly distributed among dates.

We analyzed the main and interactive effects ofseastar and snail treatment on the proportion of timespent out of water, proportion of time spent grazing,and proportion of algae grazed using ANOVA forsize experiments and restricted maximum likelihood(REML) mixed models for hunger experiments. Foranalyses of hunger experiments, date was includedas a random variable to test for differences amongthe trials due to varying conditions, such as watertemperature and time of year. There were notenough degrees of freedom available to include datewhen analyzing size experiments, but very similarpatterns occurred among all dates. All response variables were arcsine square root transformed tomeet statistical assumptions of normality and equalvariances.

Waterborne cues

We tested whether waterborne cues initiateTMIIs by substituting water saturated with thescent of adult P. ochraceus for the seastar itself inthe same basic design as described above. Wesoaked several adult P. ochraceus (8 to 12 cm armlength) in sea water in aquaria (38 l) for severalhours. We conducted 9 replicate mesocosms of the2 treatments with snails and with or without seastarcues, and 3 replicate mesocosms of the 2 controlswithout snails. Data were analyzed similarly asdescribed above, but using a Poisson general linearmodel (Wald chi-squared statistic) because datatransformations did not result in normality orhomogeneity of variance.

We used a different experimental design to testwhether waterborne cues from adult P. ochraceus initiated TMIIs in flow and whether the effect dimin-ished with increasing distance from the seastar. Wepartitioned plastic raceways (31 × 148 × 25 cm deep)

with unidirectional flow (7.6 cm s−1) into 5 compart-ments (15 cm wide) using perforated plastic dividers(4 mm diameter holes). The first compartment waseither empty (n = 6 trials) or contained 1 adult P.ochraceus (n = 8 trials). The 4 downstream com -partments were located 0−15, 15−30, 60−75 and120−135 cm from the first compartment. We added U.lactuca (68 to 110 mg) and a clear acrylic disc (3 mmthick, 124 mm diameter) covered with microalgae toeach compartment. We then placed 5 T. funebralis(mean 14.5 ± 0.62 mm SE) into each of 3 of the 4 compartments (0−15, 60−75 and 120−135 cm) andallowed them to graze for 16 h; the 15 to 30 cm com-partment served as a control for algal growth or grazing by microinvertebrates (n = 14 total). Wequantified grazing of microalgae and U. lactuca aspreviously described, except U. lactuca was spun in asalad spinner before blotting with tissue paper andweighing it. We analyzed the effects of the presenceof P. ochraceus and distance from seastar on the massof U. lactuca and proportion of microalgae grazedusing a REML mixed model in a split-plot design fol-lowed by Tukey’s test with the presence and absenceof seastars and distance from P. ochraceus as maineffects and trial as a random effect.

TMII field experiments

We investigated the potential for P. ochraceus andLeptasterias spp. to alter snail behavior and initiateTMIIs on tidepool algae in the field by performingbrief experiments designed to be similar to our labo-ratory experiments. We began by surveying smallshallow tidepools without seastars in the mid to highintertidal zone (0.15 to 1.0 m diameter and 1.25 to2.25 m above mean lower low water [MLLW]). Ineach tidepool, we counted T. funebralis in the waterand outside each tidepool in a ‘halo’ encompassingemersed rock <15 cm around the perimeter, whichlikely serves as a refuge from predation during lowtide (Menge & Menge 1974). During experimentswith Leptasterias spp., we added 1 Plexiglas plate(12.7 cm diameter) to all tidepools at the beginning oflow tide and added 2 seastars in mesh pouches to halfof the tidepools. In experiments with P. ochraceus,we added a piece of U. lactuca comparable in size tothe plates to all tidepools and 1 adult seastar in amesh pouch to half of the tidepools. The pouchesallowed the emission of chemical cues, while pre-venting predation or contact with snails. We resur-veyed snails inside and outside tidepools and quanti-fied the proportion of algae grazed using a gridded

35


qua drat ~5 h later, before flood tide. For Leptasteriasspp., we recorded the proportion of snails out ofwater in 23 replicate tidepools at the end of ex -periments with seastars and 17 replicate tidepoolswithout seastars, and we recorded the proportionof microalgae grazed in 8 tidepools per treatment.For P. ochraceus, we recorded the proportion ofsnails out of water in 27 replicate tidepools withseastars and 14 replicate tidepools without seastars,and we recorded the proportion of snails grazing U.lactuca in 15 and 11 of the tidepools, respectively.For each species, we conducted 1-way ANOVAs todetermine the effects of seastar presence on the arcsine square root transformed proportion of algaegrazed by snails and the proportion of snails out ofwater.

RESULTS

Palatability of macroalgae

In descending order, snails ate more Ulva lactuca,Cladophora columbiana and Porphyra spp., followedby similar amounts of Pelvetiopsis limitata and Fucusgardneri and lastly similar amounts of Endocladiamuricata and Mastocarpus papillatus (algae treat-ment: F6,910 = 172.0, p < 0.001; time treatment: F1,910 =596.4, p < 0.001; algae × time: F6,910 = 2.9, p = 0.008;Fig. 1). Based on this feeding trial, we used U. lactucain experiments testing whether Pisaster ochraceusexert TMIIs on macroalgae in the laboratory andfield.

TMII laboratory experiments

Seastar species and life stage

During encounters between Tegula funebralis andboth species of seastars, snails immediately twistedtheir bodies and fled in the opposite direction, out-pacing and escaping seastars. Snails left the watereven when they did not touch the seastars, indicatingthat they were likely responding to dissolved chemi-cal cues. Both Leptasterias spp. (snail treatment:F1,401 = 195.1, p < 0.001; seastar treatment: F1,401 =37.6, p < 0.001; snail × seastar: F1,401 = 43.2, p < 0.001)and juvenile P. ochraceus (snail treatment: F1,152 =52.3, p < 0.001; seastar treatment: F1,152 = 8.5 p =0.004; snail × seastar: F1,152 = 8.5, p = 0.004) exertedpositive TMIIs on algae in the laboratory, decreasinggrazing on microalgae by 3.1 and 2.6 times whensnails were present, respectively (Fig. 2). On aver-age, T. funebralis spent 4.2 fold more time out ofwater when Leptasterias spp. were present thanabsent (t154 = 32.49, p < 0.001; Fig. 2) and 5.1 foldmore time out of the water when P. ochraceus werepresent than absent (t100 = 70.09, p < 0.001).

Adult P. ochraceus initiated TMIIs on U. lactuca:P. ochraceus reduced the number of snails grazing by9.0 fold compared to when they were absent (t11 =4.29, p = 0.002; Fig. 3), resulting in 8.4 times lessalgae grazed (F26 = 11.76, p < 0.001; Fig. 3). The dif-ference in mass of algae was negligible when snailswere absent (mean ± SE change: 0.008 ± 0.006 g),indicating the effects of handling and other grazerswere minimal, and measurement error was slight.

Individual variation

Although Leptasterias spp. and juve-nile P. ochraceus generally caused snailsto forage less and flee the water moreduring TMII experiments in the labora-tory, there was considerable individualvariation among snails (Fig. 4). WhenLeptasterias spp. was absent, 71% ofsnails spent <10% of time out of waterfor the duration of the experiment. Theimpact on microalgae was highly vari-able in the absence of Leptasterias spp.,with 21% of snails grazing nearly allmicroalgae (>90%) and 19% of snailsgrazing little microalgae (<10%). A similar dichotomy in behavior occurredwhen juvenile P. ochraceus was absent,

36

0 2 6 20 27 50 76 100 123 148 194 2690

10

20

30

40

50

60

70

80

90

100

Gra

zed

(%)

Hour

Ulva

Cladophora

Porphyra

Fucus

Pelvetiopsis

Endocladia

Mastocarpus

A

D

D

C

B

E

E

Fig. 1. Palatability of macroalgae. Cumulative mean (±1 SE) percent con-sumption of 7 taxa of macroalgae (Ulva lactuca, Cladophora columbiana,Porphyra spp., Pelvetiopsis limitata, Fucus gardneri, Endocladia muricataand Mastocarpus papillatus) by Tegula funebralis for 269 h during 12 rep -licate trials for each species. Different letters indicate significant differences

(p < 0.001)


with 63% of snails spending <10% of time out of thewater. The tendency to either eat most of the micro-algae or leave it nearly uneaten was even more pro-nounced in the absence of juvenile P. ochraceus, with57% of snails grazing the microalgae nearly entirely(>90%) and 17% of snails leaving the algae nearlyuneaten (<10%).

By contrast, when Leptasterias spp. was present,41% of snails spent >90% of the time out of water,and 63% of snails hardly ate any microalgae (<10%).Similarly, when juvenile P. ochraceus was present,

45% of snails spent >90% of the time out of water,and 48% of snails ate hardly any microalgae (<10%).

Small, medium and large snails responded simi-larly to Leptasterias spp. by spending 3.7, 3.7 and2.3 times more time out of the water, respectively,than when seastars were absent (seastar treatment:F1,150 = 154.31, p < 0.001; size: F2,150 = 1.25, p = 0.290;seastar × size: F2,150 = 1.69, p = 0.188; Fig. 5a). Sea -stars also reduced grazing frequency by snails of allsizes (12.9, 8.3 and 4.8 fold decreases in time spentgrazing for small, medium and large snails, respec-

37

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Mic

roal

gae

graz

ed (%

) No seastar

Seastar

Leptasterias spp.

0

1020

30

4050

6070

8090

100

0

1020

30

4050

6070

8090

100

Tim

e ou

t of w

ater

(%)

Snail

No seastar

Seastar

Pisaster ochraceus

Snail

73 78

141

143

126

128

63

69

12 12

63

63

No snail No snail

69

12 1273 78

Fig. 2. Mean (±1 SE) proportion of microalgae grazed (top) and proportion of time herbivorous Tegula funebralis spent out ofthe water (bottom) in the presence and absence of predatory Leptasterias spp. (left) and juvenile Pisaster ochraceus (right)

after 1 h in the laboratory. Numbers of replicate mesocosms per treatment are shown

0

0.05

0.1

0.15

0.2

0.25

0.3

0

0.05

0.1

0.15

0.2

0.25

0.3

Ulv

a gr

azed

(g)

No seastar

Seastar

Control

Pisaster ochraceus

0

2

4

6

8

10

12

0

2

4

6

8

10

12

Sna

ils g

razi

ng (N

o.)

No seastar

Seastar

Fig. 3. Mean (±1 SE) number of Tegula funebralis grazing (left) and decrease in wet mass of Ulva lactuca after 8 h in the pres-ence and absence of adult Pisaster ochraceus in the laboratory (right); n = 14 replicate mesocosms per treatment (except

n = 3 for controls without snails or seastars)


tively; seastar treatment: F1,150 = 112.91, p < 0.001;size: F2,150 = 1.34, p = 0.265; seastar × size: F2,150 =1.89, p = 0.155; Fig. 5b). Contrary to expectations,only medium and large snails mediated positiveTMIIs on algae (4.9 and 4.3 fold decreases in algaegrazed, respectively) when seastars were present(seastar treatment: F1,165 = 83.21, p < 0.001; seastar ×

size: F3,165 = 8.24, p < 0.001; Fig. 5c). Smallsnails, though responsive to seastars, didnot mediate a detectable TMII (3.4 folddecrease in algae grazed with seastarspresent), because they grazed much lessalgae overall than medium and large snails(size: F3,165 = 24.39, p < 0.001; Fig. 5c).

Hungry snails spent more time in thewater overall (hunger: F1,198 = 8.14, p =0.005; Fig. 5d), and fed and hungry snailsresponded similarly to Leptasterias spp.by spending 3.6 and 4.2 fold more timeout of the water, respectively (seastartreatment: F1,198 = 137.94, p < 0.001;seastar × hunger: F1,198 = 0.22, p = 0.640;Fig. 5d). As expected, hungry snails grazedmore often than fed snails overall(hunger: F1,198 = 60.88, p < 0.001; Fig 5e).While both types of snails grazed lessoften when seastars were present, thisdecrease was marginally nonsignificantlystronger for fed snails than hungry snails(3.8 and 2.7 fold decreases in time spentgrazing, respectively; seastar treatment:F1,198 = 71.39, p < 0.001; seastar × hunger:F1,198 = 3.44, p = 0.065; Fig 5e). While bothsnail types mediated positive TMIIs, fedsnails mediated a marginally nonsignifi-cantly stronger TMII than hungry snails,with 2.9 and 1.6 fold less algae grazedwhen seastars were present for fed andhungry snails, respectively (seastar treat-ment: F1,219 = 33.93, p < 0.001; seastar ×hunger: F2,219 = 2.36, p = 0.097; Fig. 5f).Overall, hungry snails ate more algae thanfed snails, and algae were not grazedwhen snails were absent (hunger: F2,219 =90.70, p < 0.001; Fig. 5f).

Waterborne cues

Waterborne cues from P. ochraceus instill water induced T. funebralis to graze2.6 times less microalgae (Poisson Waldχ2

16 = 11.84, p < 0.001) and resulted in30.6 times more snails leaving the water (PoissonWald χ2

16 = 81.47, p < 0.001; Fig. 6). In flowing water,dissolved chemical cues from P. ochraceus reducedmicroalgae grazed by snails by 9.4 times at 0 to 15cm, and 5.5 times at 60 to 75 cm away (Fig. 7), but didnot significantly reduce grazing (1.1 fold decrease) at120 to 135 cm away (distance treatment: F3,27.58 =

38

0 10 20 30 40 50 60 70

0 10 20 30 40 50 60 70 80

0 10 20 30 40 50 60

0 10 20 30 40 50 60 70

0 5 10 15 20 25 30 35 40 45 50

0 5 10 15 20 25 30 35 40 45 50

0 5 10 15 20 25 30 35 40 45

Percent snails

0-910-1920-2930-3940-4950-5960-6970-7980-89

90-100

0 5 10 15 20 25

Mic

roal

gae

graz

ed (%

)

0-910-1920-2930-3940-4950-5960-6970-7980-89

90-100

Tim

e ou

t of w

ater

(%)

Percent snails

n = 126

n = 100

n = 128

n = 102

No Leptasterias spp. spp.Leptasterias

0-910-1920-2930-3940-4950-5960-6970-7980-89

90-100

Mic

roal

gae

graz

ed (%

)

0-910-1920-2930-3940-4950-5960-6970-7980-89

90-100

Tim

e ou

t of w

ater

(%)

Percent snails Percent snails

n = 63

n = 63

n = 69

n = 65

No Pisaster ochraceus Pisaster ochraceus

Fig. 4. Individual variation in microalgae grazed and time spent out of thewater by Tegula funebralis (%) in the presence and absence of Leptasteriasspp. and juvenile Pisaster ochraceus during 1 h feeding trials in the labora-

tory. Number of replicate mesocosms per treatment are shown


65.2, p < 0.001; seastar treatment: F1,27.12 = 49.4, p <0.001; seastar × distance: F3,27.12 = 13.3, p < 0.001).Similarly, dissolved chemical cues from P. ochraceusreduced U. lactuca grazed by snails by about 33times at 0 to 15 cm, and 16 times at 60 to 75 cm away,whereas the cues did not significantly reduce grazingat 120 to 135 cm away (2.9 times less grazed; distancetreatment: F3,28.29 = 4.3, p = 0.012; seastar treatment:

F1,26.96 = 19.1, p < 0.001; seastar × distance: F3,26.97 =2.9, p = 0.053) after 16 h (Fig. 7). Thus, the cues fromthe seastar reduced grazing on both microalgae andU. lactuca at 0 to 15 and 60 to 75 cm from the seastar,but not significantly at 120 to 135 cm away. Randomeffects of replicate trials ac counted for 4.47% of thevariation in trials with microalgae and 15.41% of thevariation in trials with U. lactuca.

39

0

20

40

60

80

100

Small Med Large

aTi

me

out o

f wat

er (%

)

0

20

40

60

80

100

None Small Med Large

c

Mic

roal

gae

graz

ed (%

)

0

10

20

30

40

50

Small Med Large

b

Tim

e gr

azin

g (%

)

No seastarSeastar

0

20

40

60

80

100

Fed Hungry

e

Snail treatment

0

20

40

60

80

100

Fed Hungry

d

0

60

40

20

80

100

None Fed Hungry

f

Fig. 5. Size and hunger of Tegula funebralis in the laboratory: mean (± SE) percentage of time that individual snails spent outof water (left) or grazing (middle), and mean (± SE) microalgae grazed (right) after 60 min in laboratory tanks with and withoutindividual Leptasterias spp. present. Top panels (a−c) compare snails of different sizes (small: <12 mm; medium: 12−18 mm;large: >18 mm shell diameter). Note the different y-axis scale in (b). Bottom panels (d−f) compare snails that were either fed orstarved for 1 wk in the laboratory before experiments. Replication for size experiments: 26 mesocosms with and without sea stars and 9 mesocosms for the 2 controls without snails (none). Replication for hunger experiments: 52 mesocosms with

sea stars, 51 mesocosms without seastars and 12 mesocosms each for the controls without snails

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Mic

roal

gae

graz

ed (%

)

No cue

Cue

Pisaster ochraceus

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Tim

e ou

t of w

ater

(%)

Fig. 6. Waterborne cues from adult Pisaster ochraceus in the laboratory. Mean (±1 SE) proportion of microalgae grazed (left)and time spent out of the water (right) by Tegula funebralis in the presence and absence of dissolved chemical cues from

P. ochraceus after 1 h in the laboratory (n = 9 mesocosms)


TMII field experiments

Some T. funebralis typically began climbing wallsof tidepools within minutes of adding seastars to tide-

pools and evaded seastars by leaving the water. Onaverage, more snails occurred out of water in thepresence of both Leptasterias spp. (3.3 times more;t38 = 11.24, p = 0.002) and adult P. ochraceus (7.4 times

40

0-15 15-30 60-75 120-1350

10

20

30

40

50

60

70

80

90

100

Mic

roal

gae

graz

ed (%

)

No cue

Cue

0–15 15–30 60–75 120–135–5

0

5

10

15

20

25

30

35

40

45

50

Ulv

a gr

azed

(mg)

Distance (cm) Distance (cm)

**

* *

Pisaster ochraceus

* *

Fig. 7. Waterborne cues from adult Pisaster ochraceus in flow in the laboratory. Mean (±1 SE) change in microalgae (left) andUlva lactuca (right) in flow (7.6 cm s−1) by 5 Tegula funebralis in the presence (n = 8) and absence (n = 6) of dissolved chemicalcues from P. ochraceus that were 0−15, 60−75 and 120−135 cm from the seastar after 16 h in the laboratory. Snails were notplaced at the 15 to 30 cm distance, to serve as a control (see ‘Materials and methods: TMII laboratory experiments: Waterborne

cues’ for details). Significant differences (p < 0.05) between treatments at each distance are indicated by an asterisk

0102030405060708090

100

0102030405060708090

100

Mic

roal

gae

graz

ed (%

) No seastar

Seastar

0102030405060708090

100

0102030405060708090

100

Tim

e ou

t of w

ater

(%)

17

23

Leptasterias spp.

8

8

0102030405060708090

100

0102030405060708090

100

Tim

e ou

t of w

ater

(%)

14

27

0

1

2

3

4

5

6

0

1

2

3

4

5

6

Gra

zing

Ulv

a (%

)

Pisaster ochraceus

11

15

Fig. 8. Adult Pisaster ochraceus and Leptasterias spp. in the field. Mean (±1 SE) proportion of microalgae grazed by Tegula funebralis (top left) or the percentage of snails grazing Ulva lactuca placed in tidepools (top right), and the mean (±1 SE) number of T. funebralis out of the water (bottom) after Leptasterias spp. (p = 0.0017, left) and P. ochraceus (p < 0.001, right)were added or not added to tidepools at Horseshoe Cove, northern California. Number of replicate tidepools are shown


more; t39 = 18.28, p < 0.001; Fig. 8). However, the pro-portion of microalgae grazed did not differ in thepresence and absence of Leptasterias spp. (t14 = 1.81,p = 0.200), and neither did the proportion of snailsgrazing U. lactuca in the presence and absence ofadult P. ochraceus (t24 = 0.47, p = 0.499; Fig. 8).

DISCUSSION

TMII

TMII clearly occurred in the laboratory. Adult andjuvenile Pisaster ochraceus and Leptasterias spp.seastars all caused Tegula funebralis snails to flee thewater and graze less, thereby benefitting microalgaeand macroalgae. However, we documented consid-erable variation in the responses of snails to both species of seastars, which diminished the strength ofTMII in the laboratory. A strong dichotomy in fleeingthe water was evident in the presence of both seastarspecies, in which most individuals spent nearly all oftheir time out of water while others never left thewater. By contrast, snails typically spent most of thetime in the water when these seastars were absent.Thus, snails did not generally spend their time in airunless evading seastars, presumably due to physio-logical stress from limited oxygen exchange, higherair temperature and desiccation. However, whenseastars were present, some snails apparently riskedpredation in lieu of exposing themselves to thesestressful conditions. Another dichotomy occurred forthe amount of algae grazed. When seastars wereabsent, grazing was highly variable, perhaps be -cause some snails were hungrier than others. How-ever when seastars were present, snails grazed verylittle even when they remained in the water, presum-ably due to the imminent threat of predation. Varia-tion in prey traits, such as hunger, size and tempera-ment, are well known to lead prey to exhibit verydifferent antipredator responses, especially whenfood is present (Werner & Anholt 1993, Clark 1994,DeWitt et al. 1999, Sih et al. 2004). Additional exper-iments are needed to demonstrate that the observeddichotomies in the responses of T. funebralis to seast-ars differ consistently among individuals as functionsof one or more prey state variables.

Size did not appear to strongly affect the collectiveresponses of snails to seastars in initial or follow-upexperiments. In the follow-up experiments with Lep-tasterias spp., size did not significantly affect the timeout of water or time spent grazing in the presence ofseastars. The similar flight and grazing responses to

Leptasterias spp. among size classes of snails wasunexpected given that only small T. funebralis arevulnerable to Leptasterias spp. and only small vul-nerable individuals of the sea urchin Strongylocentro -tus franciscanus flee the predatory seastar Pycnopodiahelianthoides (Freeman 2006). This seemingly sub-optimal behavior might be due to the inability ofsnails to distinguish Leptasterias spp. from larger P.ochraceus, which can consume all sizes of snails asthey grow. Alternatively, snails surviving attacks byLeptasterias spp. might not learn that the risk of pre-dation abates as they grow (Yarnall 1964). A thirdexplanation for the similar response of all sizes ofsnails is that seastars starved for 24 h might not haveemitted strong threat cues because recent ingestionof hetero- or conspecifics can increase perception ofrisk cues by prey, causing stronger avoidance ofpredators (Jacobsen & Stabell 2004, Smee & Weiss-burg 2006, Weissburg & Beauvais 2015). This did notappear to be the case in our experiments. All snailsresponded to Leptasterias spp., regardless of size.Moreover, we did not observe size-dependent differ-ences in responses by snails to juvenile P. ochraceus,but large T. funebralis were previously shown to bemore responsive than small ones to P. ochraceus cues(Doering & Phillips 1983), perhaps because adult P.ochraceus prefer large T. funebralis (Markowitz1980).

Though hungry snails spent more time in the watergrazing, and grazed more algae than fed snails over-all, they were only marginally less responsive to Lep-tasterias spp. than fed snails, and TMII strength wasnot significantly weaker when snails were hungry.This was surprising because lower energetic reservesconsistently increase risky foraging behavior andweaken TMIIs in many organisms (Werner & Anholt1993, Clark 1994, Lima 1998b, Heithaus et al. 2007,Matassa & Trussell 2014). Though we did not detect astrong effect of size and hunger of T. funebralis onfleeing and TMII in response to Leptasterias spp. inthe laboratory, we did detect them in companionstudies in the field. In tidepools, small T. funebralismediated weak TMIIs, medium snails mediatedstrong positive TMIIs and large snails (surprisingly)mediated negative TMII in response to Leptasteriasspp. (Gravem & Morgan 2016). Furthermore, smallsnails responded to mass mortality of Leptasteriasspp. by moving into tidepools and lower in the inter-tidal zone where seastars are more prevalent, whereaslarge snails did not (Gravem 2015). In addition, wepreviously found that hungry T. funebralis in rockytidepools continued to forage in the presence of sea -stars, whereas many fed snails began fleeing imme-

41


diately after seastars were placed in tidepools (Gravem& Morgan 2016). Consequently, TMIIs did not occurin tidepools when snails were hungry, but they didoccur when snails were fed. We likely did not detecta stronger effect of the size or hunger of snails onfleeing and TMII in the laboratory due to the closeconfines of predator and prey in the mesocosms.Snails and seastars often touched in mesocosms,evoking strong flight responses, whereas they didnot in tidepools (Gravem & Morgan 2016). Thus,TMII experiments in the laboratory likely overesti-mated the flight responses and TMII operating innature, as with TMII experiments conducted in meso -cosms in other systems (Okuyama & Bolker 2007,Long & Hay 2012).

Individual variation in boldness and habitat prefer-ences of snails might also be responsible for the indi-vidual variation in snail behavior, which may weakenTMII. Individual T. funebralis exhibit consistent be -havioral syndromes and vary in their responsivenessto P. ochraceus, with some individuals being con -sistently bolder than others (Pruitt et al. 2012). In addi-tion, individual T. funebralis possess geneticallybased habitat preferences, including occurring insideor outside of tidepools and at specific shore levels (By-ers & Mitton 1981, Byers 1983), which would affecttheir propensity to leave the water in response topredatory seastars. These innate differences in snailtemperament may have contributed to the high indi-vidual variation that we observed in our experiments.

Waterborne cues

Behavioral observations during our laboratoryexperiments in mesocosms revealed that snails con-tacted by either seastar species immediately turnedaround and fled in the opposite direction, outpacingthe seastars. Even when contact did not occur, snailsevaded capture by fleeing the water, indicating thatthey might have been responding to waterbornechemical cues from seastars. Our subsequent experi-ments, using caged seastars, dissolved chemical cuesof seastars in still water and seastars upstream ofsnails, confirmed that snails respond to waterbornecues of both seastars without physical contact be -tween predator and prey in this system, as has previously been demonstrated (Feder 1963, Yarnall1964, Phillips 1976). T. funebralis has long beenknown to have chemoreceptor organs at the base ofthe respiratory gills that can detect dissolved chemi-cal cues from predators (Burke 1964, Szal 1971, Croll1983).

We demonstrated that TMIIs occur in flowing waterin the laboratory, raising the possibility that water-borne chemical cues from seastars might also initiateTMIIs in the field. In the laboratory, T. funebralis responded to waterborne cues from P. ochra ceus byreducing grazing on both micro- and macroalgae inlow laminar flow (0.5 l min−1) when the seastar was asfar away as 75 cm upstream. Furthermore, we ob-served T. funebralis detect Leptasterias spp. 1 m awayin tidepools; there was however variation among tide-pools, probably due to differences in water circulation.Species in other systems have detected chemical cuesat similar flow rates and distances (Turner & Mont-gomery 2003, Smee et al. 2008, Weissburg et al. 2014,Weissburg & Beauvais 2015), and antipredator behav-iors have also been observed for other prey speciesnear predators in higher flow conditions (Trussell etal. 2004, Freeman 2006, Large et al. 2011). The abilityof T. funebralis to detect waterborne cues from preda-tory seastars remains to be determined on wave-swept rocky shores at high tide; indeed, their behav-iors at high tide are unknown.

T. funebralis responded similarly to Leptasteriasspp., and juvenile and adult P. ochraceus in both thefield and laboratory; the proportion of snails fleeingfrom seastars was consistently high regardless ofspecies. This was surprising because larger adult P.ochraceus pose a greater threat to most snails than dosmaller Leptasterias spp., and Leptasterias spp. aremore efficient predators of mobile prey than arejuvenile P. ochraceus of comparable size (Menge1972). T. funebralis was previously shown to respondmore strongly to adult P. ochraceus than Leptasteriasspp. (Yarnall 1964), perhaps because they were re -acting to the higher concentration of cues from muchlarger adult P. ochraceus. By contrast, we previouslyfound that T. funebralis might respond more stronglyto Leptasterias spp. in the field because small snailsmoved lower on the shore and into tidepools after thelocalized mass mortality of Leptasterias spp. despitethe continued presence of P. ochra ceus (Gravem2015). Molecular identification of the specific cuesemitted by the 2 seastars and experiments that de -termine functional responses of snails to predatorbiomass, density and cue concentration would be re -quired to determine whether T. funebralis can distin-guish between these predators.

Tidepools

Both adult P. ochraceus (this study) and Leptaste-rias spp. (Gravem 2015, Gravem & Morgan 2016)

42


caused snails to flee tidepools in the field, suggestingthe potential exists for TMII to occur in nature.Though we did not explore DMII, not one of the hun-dreds of T. funebralis was eaten by either seastarspecies during laboratory and field trials, and T. fune -bralis is only one of many different prey items in thediets of P. ochraceus and Leptasterias spp. in the field(Feder 1959, Menge 1972, Bartl 1980, Gravem &Morgan 2016). Furthermore, ours and other studiesindicate that this system potentially meets the re -quirements for strong TMII (He et al. 1993, Peacor& Werner 2001, Gravem 2015), including (1) rapidresponses of prey to predators, (2) many more preyresponding to than being eaten by predators and (3)long-term behavioral effects. In our experiments,many dozens of snails rapidly fled from tidepools forphysiologically stressful hot, dry halos where grazingmight be reduced until high tide, similar to otherstudies in this system (Yarnall 1964, Gravem 2015)and to studies demonstrating strong effects of preda-tors on habitat use or grazing by other gastropods(Bernot & Turner 2001, Trussell et al. 2002, 2004,Matassa & Trussell 2011, Wada et al. 2013). Thoughour studies were brief, and halos are only a tempo-rary refuge at low tide, T. funebralis might leave tide-pools for longer than at low tide, thereby extendingthe duration of the TMII on algae inside tidepools.Indeed, long-term manipulations of Leptasterias spp.and T. funebralis demonstrated that Leptasterias spp.caused T. funebralis to continually use halo refugesover 10 mo and caused positive TMIIs on microalgaeover 1 mo and macroalgae over 8 mo (Gravem 2015).Others have also demonstrated strong long-termeffects of seastars on T. funebralis distributions: (1)adding P. ochraceus to the lower intertidal zonecaused snails to flee to the upper intertidal zone(Markowitz 1980); (2) sites with high predator diver-sity in the lower intertidal zone (P. ochraceus, crabsand octopi) apparently caused snails to shift to theupper intertidal zone (Fawcett 1984); and (3) local-ized mortality of Leptasterias spp. enabled snails toshift lower in the intertidal zone and tidepools previ-ously occupied by the seastar (Gravem 2015).Though these patterns indicate strong potential forTMIIs to occur in the field, more extensive and long-term field experiments are needed to determine thestrength of TMIIs and DMIIs mediated by seastars inthese tidepool communities.

Even though snails evaded predatory seastars byfleeing the tidepools, we did not detect significantreductions in grazing in the presence of seastars during field trials. TMIIs might have been maskedby omnivorous hermit crabs which were abundant,

readily grazed microalgae and did not appear to altertheir behavior in response to seastars. TMII on U. lac-tuca might not have been detected because the 1 htrials might not have been long enough; snails oftenscraped the surface of the blade without diminishing it.

If seastars indeed initiate TMIIs in the field, ourpalatability studies indicate that the strongest bene-fits might be conferred on tidepool microalgae andseveral species of macroalgae that were preferred byT. funebralis. These macroalgae included the rela-tively fast-growing U. lactuca, Cladophora columbianaand Porphyra spp. over the other 4 species of algae,which might be slower-growing and chemically orstructurally defended (Watanabe 1984, Steinberg1985). Similar feeding preferences were found in arecent laboratory study, despite minor differences inthe rankings of algae (Aquilino et al. 2012).

CONCLUSIONS

Our small mesocosms and short experiments do notreflect the strength of TMIIs for natural populationsat longer time scales, as with other TMII experimentsconducted in mesocosms in other systems (Okuyama& Bolker 2007, Long & Hay 2012). However, our suiteof experiments do indicate the potential for TMIIs tooccur in tidepools, setting the stage for more chal-lenging field tests to distinguish the interactionstrength and persistence of both TMIIs and DMIIsacross the landscape (Huang & Sih 1990, Schmitz2004, Prasad & Snyder 2006, Mooney & Agrawal2008, Laundré et al. 2010). Moreover, our resultsemphasize the need to incorporate behavior intostudies, to provide comprehensive estimations of thecascading impacts of predation on communities (Peacor & Werner 2001, Schmitz et al. 2003, Preisseret al. 2005, Ohgushi et al. 2012). Finally, by conduct-ing many replicate feeding and behavioral trials inthe laboratory, we gained interesting insights re -garding individual variation in predator−prey re -sponses that warrant further attention. Our studyadds to the growing body of empirical evidence sup-porting model predictions that individual variation inprey state alters TMII strength (Ovadia & Schmitz2002, Luttbeg et al. 2003, Persson & De Roos 2003,Kotler et al. 2004, Freeman 2006, Rudolf 2012,Matassa & Trussell 2014, Gravem & Morgan 2016).

Acknowledgements. We thank the many undergraduateand graduate students who participated in conducting thisresearch over the years. Student support was provided bythe National Science Foundation Research Experience for

43


Undergraduates program (DBI-0753226) awarded to S. L.Williams and E. D. Sanford, National Science FoundationGK-12 (0841297) awarded to S. L. Williams, Mildred E.Mathias Foundation, Henry A. Jastro Fellowship, UC DavisGraduate Group in Ecology, and Bodega Marine Laboratory.This is a contribution of Bodega Marine Laboratory.

LITERATURE CITED

Abrams PA (1995) Implications of dynamically variable traitsfor identifying, classifying and measuring direct andindirect effects in ecological communities. Am Nat 146: 112−134

Abrams PA (2007) Defining and measuring the impact ofdynamic traits on interspecific interactions. Ecology 88: 2555−2562

Appleton RD, Palmer AR (1988) Water-borne stimuli re -leased by predatory crabs and damaged prey inducemore predator-resistant shells in a marine gastropod.Proc Natl Acad Sci USA 85: 4387−4391

Aquilino KM, Coulbourne ME, Stachowicz JJ (2012) Mixedspecies diets enhance the growth of two rocky intertidalherbivores. Mar Ecol Prog Ser 468: 179−189

Aschaffenburg MD (2008) Different crab species influencefeeding of the snail Nucella lapillus through trait-mediated indirect interactions. Mar Ecol 29: 348−353

Bartl S (1980) A comparison of the feeding behavior of thesix-rayed seastar, Leptasterias hexactis, and from twointertidal habitats. BML student report. Bodega MarineLaboratory Cadet Hand Library, University of CaliforniaDavis, Bodega Bay, CA

Behrens Yamada S, Navarrete SA, Needham C (1998) Pre-dation induced changes in behavior and growth rate inthree populations of the intertidal snail, Littorina sitkana(Philippi). J Exp Mar Biol Ecol 220: 213−226

Bernot RJ, Turner AM (2001) Predator identity and trait-mediated indirect effects in a littoral food web. Oecolo-gia 129: 139−146

Burke WR (1964) Chemoreception by Tegula funebralis.Veliger 6: 17−20

Byers BA (1983) Enzyme polymorphism associated withhabitat choice in the intertidal snail Tegula funebralis.Behav Genet 13: 65−75

Byers B, Mitton J (1981) Habitat choice in the intertidal snailTegula funebralis. Mar Biol 65: 149−154

Carlton JT (ed) (2007) The Light and Smith manual: inter-tidal invertebrates from central California to Oregon,4th edn. University of California Press, Berkeley, CA

Clark CW (1994) Antipredator behavior and the asset-protection principle. Behav Ecol 5: 159−170

Croll RP (1983) Gastropod chemoreception. Biol Rev CambPhilos Soc 58: 293−319

DeWitt TJ, Sih A, Hucko JA (1999) Trait compensation andcospecialization in a freshwater snail: size, shape andantipredator behaviour. Anim Behav 58: 397−407

Dill LM, Heithaus MR, Walters CJ (2003) Behaviorally medi-ated indirect interactions in marine communities andtheir conservation implications. Ecology 84: 1151−1157

Doering PH, Phillips DW (1983) Maintenance of the shore-level size gradient in the marine snail Tegula funebralis(A. Adams): importance of behavioral responses to lightand sea star predators. J Exp Mar Biol Ecol 67: 159−173

Fawcett MH (1984) Local and latitudinal variation in predation on an herbivorous marine snail. Ecology 65:

1214−1230Feder HM (1959) The food of the starfish, Pisaster ochraceus

along the California coast. Ecology 40: 721−724Feder HM (1963) Gastropod defensive responses and their

effectiveness in reducing predation by starfishes. Ecol-ogy 44: 505−512

Flowers JM, Foltz DW (2001) Reconciling molecular system-atics and traditional taxonomy in a species-rich clade ofsea stars (Leptasterias subgenus Hexasterias). Mar Biol139: 475−483

Freeman A (2006) Size-dependent trait-mediated indirectinteractions among sea urchin herbivores. Behav Ecol 17: 182−187

Gravem SA (2015) Linking antipredator behavior of prey tointertidal zonation and community structure in rockytidepools. PhD thesis, University of California, Davis, CA

Gravem SA, Morgan SG (2016) Prey state alters trait-mediated indirect interactions in rocky tide pools. FunctEcol, doi: 10.1111/1365-2435.12628

Hairston NG, Smith FE, Slobodkin LB (1960) Communitystructure, population control and competition. Am Nat94: 421−425

He X, Wright R, Kitchell JF (1993) Fish behavioral and com-munity responses to manipulation. In: Carpenter S,Kitchell J (eds) The trophic cascade in lakes. CambridgeUniversity Press, Cambridge, p 69−84

Heithaus MR, Frid A, Wirsing AJ, Dill LM and others (2007)State-dependent risk-taking by green sea turtles medi-ates top-down effects of tiger shark intimidation in amarine ecosystem. J Anim Ecol 76: 837−844

Huang CF, Sih A (1990) Experimental studies on behav-iorally mediated, indirect interactions through a sharedpredator. Ecology 71: 1515−1522

Jacobsen HP, Stabell OB (2004) Antipredator behaviourmediated by chemical cues: the role of conspecific alarmsignalling and predator labelling in the avoidanceresponse of a marine gastropod. Oikos 104:43–50

Kotler BP, Brown JS, Bouskila A (2004) Apprehension andtime allocation in gerbils: the effects of predatory riskand energetic state. Ecology 85: 917−922

Large SI, Smee DL, Trussell GC (2011) Environmental con-ditions influence the frequency of prey responses to predation risk. Mar Ecol Prog Ser 422: 41−49

Laundré JW, Hernandez L, Ripple WJ (2010) The landscapeof fear: ecological implications of being afraid. Open EcolJ 3: 1−7

Lima SL (1998a) Nonlethal effects in the ecology of pre -dator-prey interactions: What are the ecological effects ofanti-predator decision-making? Bioscience 48: 25−34

Lima SL (1998b) Stress and decision making under the riskof predation: recent developments from behavioral,reproductive, and ecological perspectives. Adv StudBehav 27: 215−290

Long JD, Hay ME (2012) The impact of trait-mediated indi-rect interactions in marine communities. In: Ohgushi T,Schmitz OJ, Holt RD (eds) Trait-mediated indirect inter-actions: ecological and evolutionary perspectives. Cam-bridge University Press, New York, NY, p 47−68

Luttbeg B, Trussell GC (2013) How the informational envi-ronment shapes how prey estimate predation risk andthe resulting indirect effects of predators. Am Nat 181: 182−194

Luttbeg B, Rowe L, Mangel M (2003) Prey state and experi-mental design affect relative size of trait- and density-mediated indirect effects. Ecology 84: 1140−1150

44

http://dx.doi.org/10.1890/0012-9658(2003)084%5b1140%3APSAEDA%5d2.0.CO%3B2

http://dx.doi.org/10.1086/668823

http://dx.doi.org/10.1016/S0065-3454(08)60366-6

http://dx.doi.org/10.2307/1313225

http://dx.doi.org/10.2174/1874213001003030001


http://dx.doi.org/10.1890/03-3002

http://dx.doi.org/10.1111/j.0030-1299.2004.12369.x

http://dx.doi.org/10.2307/1938288

http://dx.doi.org/10.1111/j.1365-2656.2007.01260.x

http://dx.doi.org/10.1086/282146

http://dx.doi.org/10.1111/1365-2435.12628

http://dx.doi.org/10.1093/beheco/arj014

http://dx.doi.org/10.2307/1932529

http://dx.doi.org/10.2307/1929828

http://dx.doi.org/10.2307/1938329

http://dx.doi.org/10.1016/0022-0981(83)90087-4

http://dx.doi.org/10.1890/0012-9658(2003)084%5b1151%3ABMIIIM%5d2.0.CO%3B2

http://dx.doi.org/10.1006/anbe.1999.1158

http://dx.doi.org/10.1111/j.1469-185X.1983.tb00391.x

http://dx.doi.org/10.1093/beheco/5.2.159

http://dx.doi.org/10.1007/BF00397079

http://dx.doi.org/10.1007/BF01071744

http://dx.doi.org/10.1007/s004420100705

http://dx.doi.org/10.1016/S0022-0981(97)00106-8

http://dx.doi.org/10.1111/j.1439-0485.2008.00227.x


http://dx.doi.org/10.1073/pnas.85.12.4387

http://dx.doi.org/10.1890/06-1381.1

http://dx.doi.org/10.1086/285789


Markowitz DV (1980) Predator influence on shore-level sizegradients in Tegula funebralis (A. Adams). J Exp MarBiol Ecol 45: 1−13

Matassa CM, Trussell GC (2011) Landscape of fear influ-ences the relative importance of consumptive and non-consumptive predator effects. Ecology 92: 2258−2266

Matassa CM, Trussell GC (2014) Prey state shapes theeffects of temporal variation in predation risk. Proc R SocB 281: 20141952, doi: 10.1098/rspb.2014.1952

Menge BA (1972) Competition for food between two inter-tidal starfish species and its effect on body size and feed-ing. Ecology 53: 635−644

Menge JL, Menge BA (1974) Role of resource allocation,aggression and spatial heterogeneity in coexistenceof two competing intertidal starfish. Ecol Monogr 44: 189−209

Mooney KA, Agrawal AA (2008) Phenotypic plasticity. In: Tilmon KJ (ed) Specialization, speciation, and radiation: the evolutionary biology of herbivorous insects. Univer-sity of California Press, Berkeley, CA, p 43−57

Nielsen KJ (2001) Bottom-up and top-down forces in tidepools: test of a food chain model in an intertidal com -munity. Ecol Monogr 71: 187−217

Ohgushi T, Schmitz O, Holt RD (2012) Introduction. In: Ohgushi T, Schmitz O, Holt RD (eds) Trait-mediatedindirect interactions: ecological and evolutionary per-spectives. Cambridge University Press, New York, NY,p 1−6

Okuyama T, Bolker BM (2007) On quantitative measures ofindirect interactions. Ecol Lett 10: 264−271

Ovadia O, Schmitz OJ (2002) Linking individuals with eco-systems: experimentally identifying the relevant organi-zational scale for predicting trophic abundances. ProcNatl Acad Sci USA 99: 12927−12931

Paine RT (1969) The Pisaster-Tegula interaction: preypatches, predator food preference and intertidal commu-nity structure. Ecology 50: 950−961

Paine RT (1980) Food webs: linkage, interaction strengthand community infrastructure. The third Tansley lecture.J Anim Ecol 49: 667−685

Peacor SD, Werner EE (1997) Trait-mediated indirect interactions in a simple aquatic food web. Ecology 78: 1146−1156

Peacor SD, Werner EE (2001) The contribution of trait-mediated indirect effects to the net effects of a predator.Proc Natl Acad Sci USA 98: 3904−3908

Peckarsky BL, McIntosh AR (1998) Fitness and communityconsequences of avoiding multiple predators. Oecologia113: 565−576

Peckarsky BL, Abrams PA, Bolnick DI, Dill LM, and others(2008) Revisiting the classics: considering nonconsump-tive effects in textbook examples of predator-prey inter-actions. Ecology 89: 2416−2425

Persson L, De Roos AM (2003) Adaptive habitat use in size-structured populations: linking individual behavior topopulation processes. Ecology 84: 1129−1139

Phillips DW (1976) Effect of a species-specific avoidanceresponse to predatory starfish on the intertidal distribu-tion of two gastropods. Oecologia 23: 83−94

Prasad RP, Snyder WE (2006) Diverse trait-mediated interac-tions in a multi-predator, multi-prey community. Ecology87: 1131−1137

Preisser EL, Bolnick DI, Benard MF (2005) Scared to death?The effects of intimidation and consumption in predator-prey interactions. Ecology 86: 501−509

Pruitt JN, Stachowicz JJ, Sih A (2012) Behavioral types ofpredator and prey jointly determine prey survival: poten-tial implications for the maintenance of within-speciesbehavioral variation. Am Nat 179: 217−227

Quinn BK, Boudreau MR, Hamilton DJ (2012) Inter- andintraspecific interactions among green crabs (Carcinusmaenas) and whelks (Nucella lapillus) foraging on bluemussels (Mytilus edulis). J Exp Mar Biol Ecol 412: 117−125

Raimondi PT, Forde SE, Delph LF, Lively CM (2000) Pro-cesses structuring communities: evidence for trait-mediated indirect effects through induced polymor-phisms. Oikos 91: 353−361

Rudolf V (2012) Trait-mediated indirect interactions insize-structured populations: causes and consequencesfor species interactions and community dynamics. In: Ohgushi T, Schmitz O, Holt RD (eds) Trait-mediatedindirect interactions: ecological and evolutionary per-spectives. Cambridge University Press, New York, NY,p 69−88

Schmitz OJ (1998) Direct and indirect effects of predationand predation risk in old-field interaction webs. Am Nat151: 327−342

Schmitz OJ (2004) Perturbation and abrupt shift in trophiccontrol of biodiversity and productivity. Ecol Lett 7: 403−409

Schmitz OJ, Adler FR, Agrawal AA (2003) Linking individ-ual-scale trait plasticity to community dynamics. Ecology84: 1081−1082

Selden R, Johnson AS, Ellers O (2009) Waterborne cues fromcrabs induce thicker skeletons, smaller gonads and size-specific changes in growth rate in sea urchins. Mar Biol156: 1057−1071

Sih A (1992) Prey uncertainty and the balancing of anti -predator and feeding needs. Am Nat 139: 1052−1069

Sih A, Bell A, Johnson JC (2004) Behavioral syndromes: anecological and evolutionary overview. Trends Ecol Evol19: 372−378

Smee DL, Weissburg MJ (2006) Hard clams (Mercenariamercenaria) evaluate predation risk using chemical sig-nals from predators and injured conspecifics. J ChemEcol 32:605–615

Smee DL, Ferner MC, Weissburg MJ (2008) Alteration ofsensory abilities regulates the spatial scale of nonlethalpredator effects. Oecologia 156:399–409

Smee DL, Ferner MC, Weissburg MJ (2010) Hydrodynamicsensory stressors produce nonlinear predation patterns.Ecology 91: 1391−1400

Steinberg PD (1985) Feeding preferences of Tegula fune-bralis and chemical defenses of marine brown algae.Ecol Monogr 55: 333−349

Szal R (1971) ‘New’ sense organ of primitive gastropods.Nature 229: 490−492

Trussell GC, Ewanchuk PJ, Bertness MD (2002) Field evi-dence of trait-mediated indirect interactions in a rockyintertidal food web. Ecol Lett 5: 241−245

Trussell GC, Ewanchuk PJ, Bertness MD, Silliman BR (2004)Trophic cascades in rocky shore tide pools: distinguish-ing lethal and nonlethal effects. Oecologia 139: 427−432

Trussell GC, Ewanchuk PJ, Matassa CM (2006) Habitateffects on the relative importance of trait- and density-mediated indirect interactions. Ecol Lett 9: 1245−1252

Turner AM, Montgomery SL (2003) Spatial and temporalscales of predator avoidance: experiments with fish andsnails. Ecology 84:616–622

45

http://dx.doi.org/10.1111/j.1461-0248.2006.00981.x

http://dx.doi.org/10.1007/s00442-004-1512-8

http://dx.doi.org/10.1046/j.1461-0248.2002.00304.x

http://dx.doi.org/10.1038/229490a0

http://dx.doi.org/10.2307/1942581

http://dx.doi.org/10.1890/09-0017.1

http://dx.doi.org/10.1007/s00442-008-0995-0

http://dx.doi.org/10.1007/s10886-005-9021-8

http://dx.doi.org/10.1016/j.tree.2004.04.009

http://dx.doi.org/10.1086/285372

http://dx.doi.org/10.1007/s00227-009-1150-0

http://dx.doi.org/10.1890/1051-0761(2003)084%5b1081%3ALITPTC%5d2.0.CO%3B2

http://dx.doi.org/10.1111/j.1461-0248.2004.00592.x

http://dx.doi.org/10.1086/286122

http://dx.doi.org/10.1034/j.1600-0706.2000.910215.x


http://dx.doi.org/10.1086/663680

http://dx.doi.org/10.1890/04-0719

http://dx.doi.org/10.1890/0012-9658(2006)87%5b1131%3ADTIIAM%5d2.0.CO%3B2

http://dx.doi.org/10.1007/BF00557847

http://dx.doi.org/10.1890/0012-9658(2003)084%5b1129%3AAHUISP%5d2.0.CO%3B2

http://dx.doi.org/10.1890/07-1131.1

http://dx.doi.org/10.1007/s004420050410


http://dx.doi.org/10.1890/0012-9658(1997)078%5b1146%3ATMIIIA%5d2.0.CO%3B2

http://dx.doi.org/10.2307/4220

http://dx.doi.org/10.2307/1936888


http://dx.doi.org/10.1111/j.1461-0248.2007.01019.x

http://dx.doi.org/10.1890/0012-9615(2001)071%5b0187%3ABUATDF%5d2.0.CO%3B2

http://dx.doi.org/10.2307/1942311

http://dx.doi.org/10.2307/1934777

http://dx.doi.org/10.1098/rspb.2014.1952

http://dx.doi.org/10.1890/11-0424.1

http://dx.doi.org/10.1016/0022-0981(80)90065-9


Wada Y, Iwasaki K, Yusa Y (2013) Changes in algal com munity structure via density- and trait-mediatedindirect interactions in a marine ecosystem. Ecology 94: 2567−2574

Watanabe JM (1984) Food preference, food quality and dietsof three herbivorous gastropods (Trochidae: Tegula) in atemperate kelp forest habitat. Oecologia 62: 47−52

Weissburg M, Beauvais J (2015) The smell of success: theamount of prey consumed by predators determines thestrength and range of cascading non-consumptiveeffects. PeerJ 3:e1426

Weissburg MJ, Zimmer-Faust RK (1993) Life and death inmoving fluids: hydrodynamic effects on chemosensory-

mediated predation. Ecology 74: 1428−1443Weissburg MJ, Ferner MC, Pisut DP, Smee DL (2002) Eco-

logical consequences of chemically mediated prey per-ception. J Chem Ecol 28: 1953−1970

Weissburg M, Smee DL, Ferner MC (2014) The sensory ecology of nonconsumptive predator effects. Am Nat184: 141−157

Werner EE, Anholt BR (1993) Ecological consequences ofthe trade-off between growth and mortality rates medi-ated by foraging activity. Am Nat 142: 242−272

Yarnall JL (1964) The responses of Tegula funebralis tostarfishes and predatory snails (Mollusca: Gastropoda).Veliger 6: 56−58

46

Editorial responsibility: Lisandro Benedetti-Cecchi, Pisa, Italy

Submitted: October 9, 2015; Accepted: May 10, 2016Proofs received from author(s): June 21, 2016

http://dx.doi.org/10.1086/285537

http://dx.doi.org/10.1086/676644

http://dx.doi.org/10.1023/A%3A1020741710060

http://dx.doi.org/10.2307/1940072

http://dx.doi.org/10.7717/peerj.1426

http://dx.doi.org/10.1007/BF00377371

http://dx.doi.org/10.1890/12-0725.1

ORIGINAL PAPER

Sex and microhabitat influence the uptake and allocationof mycosporine-like amino acids to tissues in the purple seaurchin, Strongylocentrotus purpuratus

Sarah A. Gravem • Nikki L. Adams

Received: 9 September 2011 / Accepted: 14 August 2012! Springer-Verlag 2012

Abstract UVR-absorbing mycosporine-like amino acids(MAAs) were detected in tissues of Strongylocentrotuspurpuratus and in ten species of Rhodophyte macroalgae

(eight previously untested) collected from intertidalmicrohabitats in November and January 2006–2007 in

Central California (35"090N, 120"450W). In sea urchins,

MAA concentrations were higher in ovaries than testes,while epidermal concentrations were similar between

sexes. Ovaries and epidermal tissues had similar MAA

signatures and broadband UVA/UVB absorbance, whiletestes had a narrower absorption ranges shifted toward

higher energy wavelengths. Sea urchins occupying pits in

the substrate exhibited lower MAA concentrations thanthose outside pits, suggesting adult microhabitat may

impact UV protection. Light levels did not influence

gonadal MAA concentrations, but correlated with elevatedepidermal MAA concentrations for males in the sunniest

microhabitat. This study suggests sex and habitat strongly

influence MAA concentrations among individual S. pur-puratus and that allocation of MAA sunscreens to tissues in

response to UVR is sex-dependent.

Introduction

Despite the efforts to reduce emissions of ozone-depleting

substances, the penetration of harmful solar ultravioletradiation (UVR) through the thinned ozone layer, is pre-

dicted to continue above pre-1970s levels for several decades

(Madronich et al. 1998; Forster et al. 2007) and may increasewith global climate change (McKenzie et al. 2007; Zepp

et al. 2011). UVR is detrimental to many marine organisms,

including macroalgae and marine invertebrates (Franklinand Forster 1997; Lesser et al. 2003; Campanale et al. 2011;

Hader et al. 2011; Lamare et al. 2011). In temperate coastal

waters, UVB (280–320 nm) and UVA (320–400 nm) pene-trate to several meters depth (Tedetti and Sempere 2006).

Therefore, marine organisms inhabiting shallow waters or

exposed during low tides are at risk for UVR damage. Manyorganisms have defenses that mitigate UVR damage

including photorepair and antioxidants (van de Poll et al.

2002; Lamare et al. 2006, 2011; Lesser 2010). However, it islikely advantageous to prevent UVR-induced damage by

utilizing UVR-absorbing sunscreens.One preventative defense against UVR in marine

organisms is the presence of mycosporine-like amino acids

(MAAs). MAAs are a suite of approximately 21 water-soluble compounds (Rastogi et al. 2010) that maximally

absorb light in the UVR range (309–360 nm, Dunlap and

Shick 1998). They are characterized by a cyclohexenone orcyclohexenimine chromophore conjugated with the nitro-

gen substituent of an amino acid or its imino alcohol (Sinha

et al. 1998) and are stable over long periods of time in vivo(Adams and Shick 2001; Adams et al. 2001). MAAs are

synthesized by macroalgae (especially Rhodophytes) and

microorganisms including microalgae and cyanobacteria(Sinha et al. 1998; Oren and Gunde-Cimerman 2007).

MAAs are also present in many marine invertebrates and

Communicated by H.-O. Portner.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s00227-012-2045-z) contains supplementarymaterial, which is available to authorized users.

S. A. Gravem ! N. L. AdamsBiological Sciences Department, California Polytechnic StateUniversity, San Luis Obispo, CA 93407, USA

Present Address:S. A. Gravem (&)Bodega Marine Laboratory, University of California Davis,P.O. Box 247, Bodega Bay, CA 94923, USAe-mail: [email protected]; [email protected]

123

Mar Biol

DOI 10.1007/s00227-012-2045-z

http://dx.doi.org/10.1007/s00227-012-2045-z

fishes (Shick and Dunlap 2002; Sinha et al. 2007; Rastogi

et al. 2010), and are acquired through diet or symbioses(Carroll and Shick 1996; Shick et al. 1999; Carefoot et al.

2000; Newman et al. 2000). Exposure to UVR can stimu-

late MAA production in algae and uptake in other inver-tebrates (Franklin et al. 1999; Karsten et al. 1999; Shick

et al. 1999; Hoyer et al. 2002). Further, MAAs protect sea

urchin embryos from UVR-induced damage (Adams andShick 1996, 2001) and phytoplankton against photoinhi-

bition caused by UVR (Neale et al. 1998; Karsten et al.1999), implicating MAAs as protective UVR sunscreens.

Sea urchins, especially their embryos, have served as

model organisms for examining the negative effects ofUVR on development and the protective role of MAAs

(Adams and Shick 1996, 2001; Lesser et al. 2004, 2006;

Lamare et al. 2011). Though much of this work has beenconducted using the green sea urchin, Strongylocentrotusdroebachiensis, little work has been performed on the

California purple sea urchin Strongylocentrotus purpura-tus, which is widely studied and was the first invertebrate

deuterostome to have its genome sequenced and annotated

(Sea Urchin Genome Sequencing Consortium et al. 2006).There is only one published account of MAAs in a single

S. purpuratus (Lamare and Hoffman 2004), which reported

extremely low concentrations. Nevertheless, the ecology ofpurple sea urchins in California indicates they are likely to

contain MAAs. Adults eat MAA-producing Rhodophyte

macroalgae (Ebert 1968; Dayton 1975) and are frequentlyexposed to UVR when inhabiting the intertidal and shallow

subtidal. Also, developing S. purpuratus larvae are vul-

nerable to UVR (Adams and Shick 2001; Campanale et al.2011) during the weeks to months they develop in the

water column (Strathmann 1987).

The traits of individual sea urchins may strongly affectMAA concentrations in their tissues, with potentially far-

reaching fitness effects. Intertidal S. purpuratus are rela-

tively sedentary (Grupe 2006; personal observation), so itis likely that their microhabitat will determine both their

exposure to the sun (e.g. shady or sunny microhabitat) and

their intake of algae, which may affect MAA uptake orallocation. Further, sexes may differ in their uptake or

allocation of MAAs, especially to gonadal tissues.

To further understand how MAA concentrations inintertidal S. purpuratus tissues are affected by sex, season,

microhabitat, and algal availability, we sampled sea urchin

tissues and ten common species of red macroalgae forMAAs. Sea urchins and algae were collected from four

tidepool microhabitats with differing sunlight exposure.

We hypothesized that sea urchin ovaries would have higherconcentrations of MAAs than testes and that both tissues

would increase in MAAs as the spawning season approa-

ched, but that epidermal concentrations of MAAs would besimilar between sexes and collection months. We

hypothesized that algal tissue and sea urchin epidermal

tissue would show higher concentrations of MAAs in highlight microhabitats, but that gonadal tissues would not, as

has been found for ovaries of S. droebachiensis exposed to

UVR in the laboratory (Adams et al. 2001).

Materials and methods

Ten common species of Rhodophyte macroalgae and 119adult Strongylocentrotus purpuratus were collected in

October 2006 (algae) and in November 2006 and January

2007 (sea urchins) from four rocky intertidal microhabitatson the jetty at Port San Luis, California (35"090N,

120"450W). The microhabitats included horizontally ori-

ented surfaces in tide pools where sea urchins were (1) notburrowed into pits (‘‘Non-pit’’ microhabitat), (2) burrowed

into pits (‘‘Pit’’ microhabitat) which sea urchins excavate

from the substratum, and vertical walls in tide pools facingeither (3) south or (4) north with sea urchins outside of pits

(‘‘South-facing’’ and ‘‘North-facing’’ microhabitats, respec-

tively). For each microhabitat, three representative tide poolstations were chosen; all stations were in the mid-intertidal

zone within 50 meters distance of each other, were between

0.91 and 6.51 m2 in area, and had similar sea urchin densities.We hypothesized that sun exposure would be highest for sea

urchins in horizontally oriented Non-pit microhabitats, fol-

lowed by South facing then North-facing microhabitats,and lowest in the Pit microhabitat.

Station measurements

To compare the relative solar exposure of sea urchins

inhabiting the four microhabitats, we used three comple-mentary approaches. First, in November 2006 three Solar

Pathfinder (Perusion) measurements were taken and aver-

aged for each station, except for the three-Pit microhabitatstations because the instrument did not fit inside the pits.

The Solar Pathfinder is commonly used in terrestrial and

stream ecology to measure sun exposure (Li et al. 1994;Naumburg and DeWald 1999; Arkle and Pilliod 2010) and

consists of a spherical dome that reflects a panorama of the

location, including any shade-casting objects. Solar Path-finder Assistant (Perusion) imaging software was used

calculate daily solar energy from a photograph for each

month of the year based on shading as the sun anglechanges predictably daily and seasonally. We compared

August through January light measurements because solar

exposure during the months prior to collections potentiallyinfluenced MAA uptake; sea urchins can retain MAAs in

their tissues for months (Adams and Shick 1996), and

S. purpuratus gametes were likely developing during thisperiod (Giese et al. 1991). Second, to obtain sunlight

Mar Biol

123

measurements inside pits and to supplement data from the

Solar Pathfinder, HOBO Pendant# light loggers (Onset)were deployed in each station for 16 days in December

2011. Though these data were taken years after sea urchin

collections, the angle of the sun follows a fixed paththrough the sky each year, so these measurements provide

good approximations of the relative sun exposure of sea

urchins collected during the same season in past years.

Although these loggers measure illuminance (lux), which isnot easily converted to a standard irradiance metric, they

were helpful for this study because they are small enough

to fit inside pits and measure relative differences in sunlightamong microhabitats. Finally, to further validate the Solar

Pathfinder data and to specifically compare UVA and UVB

levels among microhabitats (except the Pit microhabitat,because the instruments were too large), we used an IL

1400A radiometer (International Light, Newburyport, MA,USA) coupled with a UVA (model SEL033) or UVB

sensor (model SEL240) with maximal peak sensitivities at

350 nm and 295 nm, respectively, on two sunny afternoonsin December 2011. We averaged measures from two or

three locations in each station held at *15 cm below the

water’s surface. Admittedly, the Solar Pathfinder, HOBOlight logger and UV sensor measurements have limitations

(e.g., too large for pits, mismatch in timing, units in lux),

but these three strategies complement one another andwhen combined obtain our objective of statistically testing

differences in relative sunlight exposure among micro-

habitats. Shore level (vertical height above mean lower lowwater, MLLW) was also measured for each station,

because it potentially affected sun exposure and algal

availability to sea urchins.To compare differences in sea urchin dietary MAA

availability among microhabitats, both attached and drift

algal availability were surveyed four times. Two surveyswere performed simultaneous with sea urchin collections in

November 2006 and January 2007, and two surveys were

performed in December 2006 to assess algae available tograzing sea urchins prior to the January sampling period.

Multiple survey dates allowed us to test whether trends

were consistent despite inherent variation in algal abun-dance over time. Because it was not possible to quantify

algal intake by sea urchins, our goal was to estimate

relative algal availability in the different microhabitats and

contrast this with concentrations of MAAs in sea urchins.Algal cover (cm2 m-2) of each species attached to the

substrate was sampled using a 0.25 m2 quadrat in each

station. Because MAAs vary among algal species, weconverted algal cover values to ‘‘attached algal MAA

availability’’ using the following equation:

To determine the mean dry wt cm-2 for each species, we

scraped algae from five plots (25 or 6.25 cm2 depending onthe species) near the stations for each of 10 common species

and lyophilized, weighed and averaged the scrapes. The

mean concentration of MAAs (nmol mg-1 dry wt) for eachspecies was calculated using 12 algal specimens collected

and analyzed for MAAs (Table 1, methods described

below). Rarely occurring species were not included in thecalculations. Drift algal availability was measured in each

station by collecting and weighing Rhodophyte macroalgaeheld by 50 random sea urchins. A conversion to MAA

availability was not performed for drift algae because

samples were not consistently identifiable to species.

Collection and preparation of specimens

Twelve specimens of each of ten common algae species

were collected in October 2006 and included Calliarthrontuberculosum, Corallina officianalis var. chilensis, Coral-lina vancouveriensis, Endocladia muricata, Mastocarpusjardinii, Mastocarpus papillatus, Mazzaella flaccida, Os-mundea spectabilis, Prionitis lanceolata, and Pterocladiacapillacea. One healthy, whole algal specimen from each

species was collected from each station within 30 cm of

sea urchins when available (algae in the Pit microhabitatwere collected adjacent to pits). They were immediately

transported to Cal Poly, cleaned of epibiota, frozen in

liquid nitrogen and pulverized. Samples were then lyoph-ilized and stored at -80 "C for later MAA extraction and

analysis. Rhodophytes from the November 5, 2006, drift

algal samples in each station were also prepared for MAAsanalysis as described above.

Adult S. purpuratus were collected on November 5,

2006, before they were fully gravid and on January 15,2007, just before the spawning season. Sea urchins were at

X

algal species

species cover ðcm2Þm2

$ species meandryweight ðmgÞcm2

$ species meantotal½MAA& ðnmolÞdry weight ðmgÞ

! "

¼ MAAs available to sea urchins ðnmolÞm2

Mar Biol

123

least 30 mm in diameter to ensure sexual maturity (Giese

et al. 1991), and were held in enclosed seawater aquaria for

less than 4 days. Whole wet wt, test diameter and epider-mis and gonad wet wt were recorded for five female and

male sea urchins from each station. Sea urchins were

rinsed, trimmed of their spines, gonads tissues were dis-sected, and epidermal tissues from the five ambulacral

sections of the test (those including tube feet) were sam-pled (*15–20 % of whole wet wt, between 4.3 and

20.8 g). Gonadal and epidermal tissues were frozen,

lyophilized, and stored at -80 "C for later MAA extractionand analysis. Gonadal index (GI), a measure of fecundity,

was calculated for each sea urchin as: (wet wt of gonad/wet

wt of whole urchin) 9 100.

Mycosporine-like amino acid analysis

A fixed dry weight of each tissue type (200 mg gonads,

1.000 g epidermis, and 200 mg algae) was extracted for

MAAs using three serial 60-min. extractions in 80 % HPLC-grade methanol at room temperature. At the start of the first

extraction, the samples were sonicated (Branson Sonifier 250)

for 20 s to lyse cells. The three extracts were pooled, and algaeextracts were filtered through a Waters Sep Pack Plus C-18

column to remove pigments as per Adams et al. (2001).

Extracts were analyzed for MAAs using a Hewlett Packard1100 Series reverse-phase high performance liquid chroma-

tography (HPLC) with a Phenomenex Phenosphere C-8 col-

umn at a flow rate of 0.8 mL min-1 (as per Adams et al.2001). MAAs were identified by comparing peak absorbance

and retention times with known MAAs, and identification of

representative peaks was confirmed by co-chromatography

using standards provided by Dr. W. C. Dunlap. The peak areas

of the MAAs mycosporine glycine, shinorine, porphyra-334,and mycosporine 2-glycine were calculated using a 55 %

methanol and 0.1 % acetic acid mobile phase, and the MAAs

palythine, asterina-330, palythinol and usujirene were calcu-lated using a 25 % methanol and 0.1 % acetic acid mobile

phase. MAA concentrations in nmol mg-1 dry wt were cal-culated from HPLC peak areas using peak area integration of

MAA standards calibrated in this system, then adjusted for

extraction efficiency for each tissue as per Dunlap and Chalker(1986). No standard was available for the MAA usujirene, so

its identity was inferred by elution time, peak absorption at

357 nm, and co-chromatography with a methanolic extract ofPalmaria palmata, an algal species known to contain usuji-

rene (Sekikawa et al. 1986). The concentration and extraction

efficiency of usujirene was calculated using the standard forpalythene, its trans isomer. Unidentified MAAs were rare and

not quantified. Absorption spectra (290–700 nm) were mea-

sured using a Jasco Model V550 UV/Vis spectrophotometerfor one representative of each of the four tissue types (ovaries,

testes, female, and males epidermis) containing equal con-

centrations of total MAAs.

Statistical analyses

All data were analyzed on JMP software using Fit Model

and sequential sums of squares. Models including random

variables were tested using Restricted Maximum Likeli-hood (REML) tests and those without were analyzed using

ANOVA. Significant terms (P \ 0.05) in all models were

Table 1 Maximal UV absorption wavelengths (kmax) and mean con-centrations of individual and total MAAs (nmol mg-1 dry wt ± SD) forthe ten species of red algae collected among the stations. Multiplying

concentrations of individual MAAs below by their molecular weightsyields mg MAA g-1 dry wt. Molecular wts (g mol-1) are Shin: 332, PT:224, P-334: 346, A-330: 288, PL: 302, Usu: 284

Species N kmax

(nm)Concentration of MAAs (nmol mg-1 dry weight)

Shin PT P-334 A-330 PL Usu Total

Calliarthrontuberculosuma

12 329 0.50 ± 0.30 0.24 ± 0.18 tr. 0.02 ± 0.02 n.d. 0.03 ± 0.03 0.79 ± 0.52

Corallina officianalisvar. chilensisa

12 320 0.28 ± 0.12 0.24 ± 0.21 tr. 0.02 ± 0.03 n.d. 0.03 ± 0.02 0.57 ± 0.31

Corallinavancouveriensisa

12 321 0.35 ± 0.18 0.19 ± 0.10 tr. 0.02 ± 0.01 n.d. 0.01 ± 0.01 0.57 ± 0.30

Endocladia muricata 12 335 8.16 ± 2.17 0.24 ± 0.83 0.01 ± 0.02 n.d. n.d. n.d. 8.41 ± 2.63

Mastocarpus jardinii 12 335 9.60 ± 3.01 n.d. 0.09 ± 0.08 n.d. n.d. n.d. 9.69 ± 3.01

Mastocarpus papillatus 12 335 11.41 ± 3.32 n.d. 0.05 ± 0.05 n.d. n.d. n.d. 11.46 ± 3.34

Mazzaella flaccida 12 327 1.38 ± 0.63 3.63 ± 1.30 tr. 0.32 ± 0.14 n.d. 0.47 ± 0.26 5.81 ± 2.16

Osmundea spectabilis 11 327 0.12 ± 0.06 0.60 ± 0.17. 0.38 ± 0.13 1.87 ± 0.63 0.23 ± 0.07 n.d. 3.20 ± 0.99

Prionitis lanceolata 12 326 4.78 ± 1.84 6.21 ± 2.56 0.01 ± 0.02 0.41 ± 0.18 n.d. 0.28 ± 0.18 11.69 ± 4.34

Pterocladia capillacea 12 330 3.09 ± 1.85 1.77 ± 0.99 0.02 ± 0.02 0.25 ± 0.16 n.d. 0.22 ± 0.12 5.34 ± 3.02

Shin shinorine, PT palythine, P-334 porphyra 334, A-330 asterina 330, PL palythinol, Usu usujirene, tr. trace MAAs detected, n.d. MAAs not detecteda Coralline algae species with calcium carbonate skeletons

Mar Biol

123

further analyzed using Student Newman–Keuls post hoc

analyses. Transformations were applied where appropriateto best achieve linearity, homogeneity of variance, and

normal distribution of residuals. Select model outputs and

supplemental figures with statistical results are provided inOnline Resource 1.

To analyze trends among microhabitats in sun exposure

(Solar Pathfinder), drift and attached algal MAA avail-ability, and wet wt of sea urchins, we used mixed REML

models with collection date, microhabitat and their inter-actions as categorical predictors, and station nested within

microhabitat as a random variable to control for among-

station variation within microhabitats. Maximum and meandaily illuminance (HOBO light loggers) and average solar

UVA and UVB irradiance (IL UV sensors) in each station

were compared among microhabitats with date included asa categorical random predictor variable. Variation in shore

level and sea urchin density among microhabitats was

analyzed using a one-way ANOVA.Microhabitat variation in total concentration of MAAs

and relative concentration of shinorine in algae was ana-

lyzed using REML analyses with species and station(nested within microhabitat) included as random variables

to control for species or stations that may have masked any

microhabitat trends.Models analyzing concentrations of MAAs in sea urchin

gonadal and epidermal tissues were performed separately

because epidermal samples contained spine and test tissues,so standardizing concentrations by dry wt rendered the two

tissue measurements incomparable. Nevertheless, we were

able to examine allocation of MAAs to epidermal versusgonadal tissues (concentration of epidermal MAAs/[con-

centration of epidermal ? gonadal MAAs]). For the three

response variables outlined above (gonadal and epidermalconcentrations of MAAs and the proportion of MAAs in

the epidermis) and for gonadal index, we ran models using

microhabitat, month, sex, and their interactions as predic-tors and included station nested within microhabitat as a

random variable. We also compared relative concentrations

of individual MAAs (concentration of individual MAA/concentration of total MAAs) among microhabitats, sexes

and tissues (month was removed because no terms were

significant) and sea urchin number and station (nestedwithin microhabitat) were included as random variables.

Results

Microhabitat characteristics

All three methods of light measurement document differ-

ences in solar exposure among microhabitats. First, theSolar Pathfinder analysis performed in 2006 showed higher

solar exposure levels in the Non-pit microhabitat compared

to the South-facing microhabitat (though this was not sig-nificant), and lower levels in the North-facing microhabitat

(Fig. 1a, REML, F3,186 = 10.3, P = 0.004). As expected,

solar exposure decreased approaching winter months, butthe pattern among microhabitats was consistent each month

(REML, Month x Microhabitat: F15,186 = 1.13, P = 0.335).

Second, measurements from HOBO light loggers taken inDecember 2011 corroborate this pattern and highlight low

solar exposure in Pits, with average daily maximum illumi-nance highest in the Non-pit (57,920 ± 4,470 lux) followed

by the South-facing microhabitat (33,450 ± 3,040 lux),

with North-facing and Pit microhabitats (3,110 ± 250 and7,450 ± 3,270 lux, respectively) having similar low solar

exposure (Fig. 1b; REML, F3,149 = 128.24, P \ 0.001).

Average daily mean illuminance from HOBO light loggersfollowed a similar trend, being highest in the Non-pit

(6,760 ± 540 lux), followed by the South-facing and North-

facing pit microhabitats (2,960 ± 300, 490 ± 140, 420 ±40 lux; Fig. 1b, REML, F3,149 = 118.4, P \0.001). The

difference between Non-pit and North-facing microhabitats

was about 55,000 lux at maximum each day (Fig. 1b), whichis approximately equivalent to direct versus indirect sunlight.

Finally, instantaneous UV measures further supported this

trend, showing higher UVA in the Non-pit compared toNorth-facing microhabitats, with South facing as interme-

diate (REML, F3,17 = 5.14, P = 0.010; X ± SE = 360.3 ±

49.5, 239.2 ± 40.1, and 143.3 ± 16.3 lW UVA cm-2 andN = 6 for Non-pit, South facing and North facing, respec-

tively). For UVB measurements, North-facing microhabitats

showed low levels, but no significant differences weredetected, possibly due to small sample size and low UVB levels

during winter afternoon measurements (X ± SE = 1.0 ± 0.2,

1.1 ± 0.4 and 0.5 ± 0.1 lW UVB cm-2 and N = 6 for Non-pit, South facing and North facing, respectively). Together

these three approaches test differences in relative sun exposure

of sea urchins and algae in the separate microhabitats.Though there was a small range in shore level (0.8–1.4 m

above MLLW) among the 12 stations, MLLW in the Non-pit

microhabitat station was higher (X ± SE = 1.34 ± 0.04 m,N = 3) than those in the other three microhabitats, which

did not differ from one another (ANOVA, F3,8 = 8.77,

P = 0.007; X ± SE = 0.91 ± 0.11, 0.96 ± 0.03, and0.84 ± 0.09 m above MLLW and N = 3 for South-facing,

North-facing and Pit microhabitats, respectively). Any effect

this shore level difference may have had on sunlight levels wascaptured by the HOBO light loggers (Fig. 1b), and likely only

increased exposure in the already sunny Non-pit microhabitat.

Algal availability

The availability of MAAs to Strongylocentrotus purpura-tus from algae growing nearby (see calculation in methods)

Mar Biol

123

was generally lower in the Non-pit microhabitat than in the

other three microhabitats, (Fig. 2a; REML, F3,30 = 5.41,

P = 0.025). This varied among surveys but 3 of the 4surveys exhibited this trend (Fig. 2a; REML, Date x

Microhabitat: F9,30 = 4.06, P = 0.003). The amount of

Rhodophyte drift algae available to sea urchins was similaramong microhabitats on most sampling dates, except in

November, when there was low availability in the Non-pit

microhabitat (Fig. 2b; REML, Date x Microhabitat:

F9,32 = 4.03, P = 0.003).

Concentration of MAAs in algae

MAAs were present in every algal specimen tested, and

species varied in concentration of total and individual

MAAs (Table 1). Having statistically controlled for inter-specific differences, algal specimens in the Non-pit

microhabitat showed higher mean concentrations of MAAsthan algae from the Pit microhabitat, with the South-facing

and North-facing microhabitats being intermediate

(REML, F3,105 = 13.20, P = 0.001). The relative con-centration of the MAA shinorine tended to be highest in

algae from Non-pit microhabitats and lowest in those from

Pit microhabitats, but this was not significant (Fig. S1).MAAs were detected in all samples of Rhododphyte

drift algae. Drift algal samples were composed primarily of

fleshy algae, but the mean concentration of MAAs in driftalgal samples (X ± SE = 1.54 ± 0.37 nmol mg-1 dry wt,

N = 12) was much lower than those of the fresh fleshy

attached algal species tested (X ± SE = 7.34 ± 0.50nmol mg-1 dry wt, N = 71).

Gonadal indices

Gonadal indices (GIs) of adult S. purpuratus in the

North-facing microhabitat were significantly greater thanthose in the South-facing and Pit microhabitats, with

Non-pit urchins having intermediate GIs (REML, F3,221 =

3.62, P = 0.032; X ± SE = 11.76 ± 0.4, 10.0 ± 0.5,12.8 ± 0.6 and 11.9 ± 0.5 and N = 20 for Non-pit, South-

facing, North-facing and Pit microhabitats, respectively).

These results were consistent among sexes and months(microhabitat 9 sex: F3,221 = 0.72, P = 0.544, microhab-

itat 9 month: F3,221 = 0.36, P = 0.780). GIs decreased

from November to January for females but not males,(REML, sex 9 month: F1,221 = 7.83, P = 0.006), suggest-

ing females may have begun spawning before the January

collection, though ovaries were still ripe upon dissection. Nocorrelation between GI and concentration of MAAs was

found for the 119 individual S. purpuratus sampled at this

study site (Linear correlation: r2 \ 0.01, F1,235 = 1.97,P = 0.160).

Concentration of total MAAs in sea urchins

Overall, concentrations of MAAs in sea urchin gonads

increased as spawning season approached (Fig. 3a, b;Table S1, P = 0.005), while the concentrations of MAAs

in epidermal tissues did not (Fig. 3c, d; Table S2,

P = 0.569). All results outlined below were consistentbetween months unless otherwise noted (Tables S1 and S2,

0

5000

10000

15000

20000

25000

Aug Sept Oct Nov Dec Jan

Non-pit

South facing

North facing

Dai

ly S

olar

Ene

rgy

(kJ

m-2

day

-1)

Month

Sol

ar Il

lum

inat

ion

(lux)

20000

60000

100000 Non-pit

20000

60000

100000 North facing

20000

60000

100000 South facing

20000

60000

100000

12/10/11

12/12/11

12/14/11

12/16/11

12/18/11

12/20/11

12/22/11

12/24/11

12/26/11

Time

Pit

a

b

Fig. 1 Solar exposure among the four microhabitats. a Average dailysolar energy (kJ m-2 day-1 ± SE) each month in the microhabitatsas estimated by Solar Pathfinder measurements in each station(N = 3) in November 2006. Measurements inside the Pit microhabitatare not included, because the instrument did not fit inside the pits.b Time series over 16 days in December 2011 of average solarillumination in stations representing the four microhabitats recordedby HOBO light loggers. These data were acquired as a comparison tosupplement Solar Pathfinder data and assess low light levels insidepits. (N = 3, except for the pit microhabitat where N = 2). Tickmarks indicate noon on each day

Mar Biol

123

month interaction terms) so the November and January

data are presented together below, but are separate in fig-ures to illustrate consistency in trends over time (Fig. 3).

Females had generally higher concentrations of MAAs

in their ovaries than males had in their testes (Fig. 3a, b;Table 2; Table S1, P = 0.006) though this difference was

only significant in the Non-pit and South-facing micro-

habitats (Fig. 3a, b; Table S1, P = 0.024). Surprisingly,females had overall higher epidermal MAA concentrations

than males (Fig. 3c, d; Table 2; Table S2, P = 0.01).

However, this relationship was variable with microhabitat

and month (Fig. 3c, d; Table S2, P = 0.005 and

P = 0.026) and the difference in concentrations of MAAswas extremely small (Table 2), so this trend is not likely

ecologically relevant. Variation in concentration of total

MAAs (nmol mg-1 dry wt) in all tissues was high andranged from 0.02 to 9.5 in ovaries, 0.05–1.65 in testes,

\0.01–0.66 in female epidermis, and \0.01–1.28 in male

epidermis.MAA concentrations tended to be lower in sea urchins

from Pit microhabitats compared to those from othermicrohabitats, though this varied among the sexes and

tissues considered (Fig. 3; Tables S1 and S2: P = 0.024

and P = 0.005 for gonads and epidermis, respectively).The lower concentration in Pit urchins was found for

ovaries (Fig. 3a, b; Student’s, P \ 0.02 for all compari-

sons), for female epidermal tissues (Fig. 3c, d; Student’s,P \ 0.02 for all comparisons) and for male epidermal tis-

sues (Fig. 3c, d) though only when compared to the Non-

pit and North-facing microhabitats (Student’s, t57 =-4.17, P \ 0.01 and t58 = -2.53, P = 0.02, respectively).

Conversely, concentrations of MAAs in testes showed little

variation among microhabitats, perhaps because concen-trations were low (Fig. 3a, b; Student’s, P [ 0.07 for all

comparisons).

Males collected from the sunniest Non-pit microhabitat(Fig. 1) in January had much higher concentrations of

epidermal MAAs than males from the other microhabi-

tats (Fig. 3d; Table S2: REML, P = 0.026; Student’s,P \ 0.01 for all pairwise comparisons). We then compared

the proportion of MAAs in the epidermal versus gonadal

tissues from both months for each sea urchin and found thatmales allocated nearly twice as many MAAs to their epi-

dermal tissues in the sunny Non-pit microhabitat compared

to males in the other microhabitats (Fig. 4; Table S3:P = 0.034; Student’s, P \ 0.01 for all pairwise compari-

sons), which did not differ significantly from one

another (Student’s, P [ 0.29 for all comparisons). Con-versely, female sea urchins showed no differences in

the mean proportion of MAAs in epidermal tissues among

the microhabitats, (Fig. 4; Student’s, P [ 0.59 for allcomparisons).

Concentrations of individual MAAs in sea urchins

We identified seven MAAs in S. purpuratus tissues

including shinorine, palythine, porphyra-334, mycosporineglycine, asterina-330, mycosporine 2-glycine and usujirene

(Table 2) which was similar to the results of Adams et al.

(2001) and Carroll and Shick (1996) for ovaries ofS. droebachiensis. Relative concentrations of individual

MAAs (concentration of individual MAA/concentration of

total MAAs) showed typical ‘‘MAA signatures,’’ whichwere similar in ovaries and epidermis of both sexes and

0

500

1000

1500

2000

Nov 5 Dec 19 Jan 1 Jan 15

Non-pitSouth facingNorth facingPit

Atta

ched

Alg

al M

AA

Ava

ilabi

lity

(nm

ol M

AA

m-2

)

Sample Date

0

80

160

240

320

400

480

560

Nov 3 Dec 18 Dec 31 Jan 14

Non-pitSouth facingNorth facing

Pit

Drif

t Alg

al A

vaila

bilit

y (m

g se

a ur

chin

-1)

Sample Date

a

b

Fig. 2 Measurements among four microhabitats of mean monthly(±SE) a attached algal MAA availability to sea urchins (seecalculation in methods, N = 3) and b drift algal availability to seaurchins (N = 3) in winter 2006–2007. Sea urchins were collected forMAA analysis on November 5, 2006, and January 15, 2007

Mar Biol

123

contained mostly shinorine, (X ± SE: 53.4 ± 2.7,

72.8 ± 2.0 and 62.3 ± 2.7 % and N = 118, 119 and 116for ovaries, female and male epidermis, respectively) some

palythine (X ± SE: 38.6 ± 2.9, 19.8 ± 2.1 and

30.3 ± 2.7 % and N = 118, 119, and 116 for ovaries,female and male epidermis, respectively), and small

0

0.5

1

1.5

2

2.5

3

Non-pit South North Pit

OvariesTestes

Con

cent

ratio

n of

MA

As

(nm

ol m

g-1 d

ry w

t)

Microhabitatfacing facing

0

0.5

1

1.5

2

2.5

3


OvariesTestes

Con

cent

ratio

n of

MA

As

(nm

ol m

g-1 d

ry w

t)


0

0.1

0.2

0.3

0.4

0.5

0.6


Female EpidermisMale Epidermis

Con

cent

ratio

n of

MA

As

(nm

ol m

g-1 d

ry w

t)


0

0.1

0.2

0.3

0.4

0.5

0.6


Female EpidermisMale Epidermis

Con

cent

ratio

n of

MA

As

(nm

ol m

g-1 d

ry w

t)


a b

c d

Fig. 3 The mean concentrationof MAAs(nmol mg-1 dry wt. ± SE) inS. purpuratus tissues amongmicrohabitats for gonadaltissues in a November andb January and for epidermaltissues in c November andd January. White bars representfemale tissues and gray barsrepresent male tissues (N = 15except for females from theSouth-facing microhabitat inNovember where N = 14)

Table 2 UV absorption maxima (kmax) and mean concentrations ofindividual and total MAAs (nmol mg-1 dry wt ± SE) for each seaurchin tissue type averaged among microhabitats for the twocollection months. The total concentrations of MAAs by dry wt ingonadal and epidermal tissues are not directly comparable, due toinclusion of pieces of test in epidermal samples. An asterisk (*)

indicates the individual MAA with the highest relative concentrationcompared to other MAAs in the tissue type ([individual MAA]/[totalMAAs]). Multiplying concentrations of individual MAAs below bytheir molecular weights yields mg MAA g-1 dry weight. Molecularwts (g mol-1) are Shin: 332, PT: 224, P-334: 346, Myc-gly: 245,A-330: 288, Myc-2-gly: 288, Usu: 284

Tissue type N kmax

(nm)Concentration of MAAs (nmol mg-1 dry wt)

Shin PT P-334 Myc-gly A-330 Myc-2-gly Usu Total

Ovaries 118 331 *0.82 ± 0.11 0.21 ± 0.06 0.05 ± 0.01 0.05 ± 0.02 0.02 ± \0.01 0.07 ± 0.07 tr. 1.22 ± 0.17

Testes 119 322 0.02 ± \0.01 *0.28 ± 0.02 tr. tr. tr. n.d. tr. 0.31 ± 0.02

Femaleepidermis

119 330 *0.16 ± 0.01 0.02 ± \0.01 0.01 ± \0.01 tr. tr. tr. n.d. 0.19 ± 0.06

Maleepidermis

116 331 *0.13 ± 0.02 0.02 ± \0.01 0.01 ± 0.001 tr. tr. tr. n.d. 0.16 ± 0.02

Shin shinorine, PT palythine, P-334 porphyra 334, myc-gly mycosporine glycine, A-330 asterina 330, myc-2-gly mycosporine 2-glycine, Usuusujirene, tr. trace MAAs detected, n.d. MAAs not detected

Mar Biol

123

amounts of other MAAs. Conversely, testes had low rela-

tive concentrations of shinorine compared to the othertissues (X ± SE: 10.7 ± 1.2 %, N = 119; Student’s,

P \ 0.01 for all comparisons) but over twice the relative

concentration of palythine (X ± SE: 88.5 ± 1.3 %,N = 119; Student’s, P \ 0.001 for all comparisons).

Ovaries also had higher relative concentrations of my-

cosporine glycine than the other three tissue types(X ± SE: 3.3 ± 0.9 % N = 118; Student’s, P \ 0.01 for

all comparisons), which only contained trace amounts

(Table 2). The relative concentration of shinorine alsotended to be higher in sea urchins from Non-pit micro-

habitats and lower in those from Pit microhabitats (Fig.

S2a), while palythine showed the opposite trend (Fig. S2b).

MAA absorption spectra in sea urchins

Absorption maxima of ovaries and female and male epi-

dermal tissues were similar and in the low UVA range

(Fig. 5; Table 2; kmax = 331, 330 and 331 nm, respec-tively), while the absorption maximum for testes was at a

shorter, higher energy wavelength near the UVA/UVBcusp (Fig. 5; Table 2: kmax = 322 nm). Ovaries had the

broadest absorption spectrum, spanning the UVA and some

of the UVB range, while testes showed a constricted

spectrum. Within the UVB range (280–320 nm), ovaries

and testes showed similar absorbencies, while both epi-dermal tissues absorbed less UVB (Fig. 5).

Discussion

Our study establishes the presence of MAAs in the tissuesof the widely studied purple sea urchin Strongylocentrotuspurpuratus, as well as ten species of Rhodophyte macro-

algae on California’s Central Coast. Multiple patterns inconcentrations of MAAs were identified that may affect

fitness or survival of sea urchins. Most importantly, we

found that ovaries exhibit higher MAA concentrations thantestes, that different MAA signatures among sexes and

tissues influence UVR absorption, that microhabitat affects

concentrations of MAAs in sea urchins, and that malesappear to more readily allocate MAAs to epidermal tissues

in response to UVR than females. These findings are

important for understanding how MAAs compare amongdifferent species of sea urchins, how sea urchins allocate

MAAs among their tissues, and how sea urchins may

respond increased UVR stress with climate change(McKenzie et al. 2007; Zepp et al. 2011).

Total MAAs in algae

This study is the first to demonstrate the presence of MAAsin eight species of red algae including Calliarthron

0

0.1

0.2

0.3

0.4

0.5

0.6

Female Male

Non-pit

South facing

North facing

Pit

Pro

port

ion

of M

AA

s in

Epi

derm

al T

issu

e

Sex

*

Fig. 4 The mean (±SE) proportion of the total concentration ofMAAs detected in S. purpuratus epidermal tissues ([epidermalMAAs]/[gonadal ? epidermal MAAs]) in the microhabitats for eachsex for both November and January (N = 30 except for females in theSouth-facing microhabitat where N = 29). The asterisk (*) denotes asignificantly higher mean proportion of MAAs in the epidermis forthe males from the Non-pit microhabitat compared to males from theother three microhabitats

0

0.2

0.4

0.6

0.8

1

1.2

300 320 340 360 380 400

Ovaries

TestesFemale EpidermisMale Epidermis

Abs

orpt

ion

Wavelength (nm)UVAUVB

Fig. 5 Typical absorption spectra of MAA extracts for gonadal andepidermal tissues for male and female S. purpuratus. Each spectrumshown is from a single representative tissue extract, each with thesame concentration of total MAAs

Mar Biol

123

tuberculosum, Corallina vancouveriensis, Endocladia mu-ricata, Mastocarpus jardinii, Mastocarpus papillatus,Mazzaella flaccida, Osmundea spectabilis, and Prionitislanceolata (reviewed by Sinha et al. 2007). MAAs have been

previously detected in only two species in our study; extractsfrom Corallina officianalis from the North Sea (Karsten et al.

1998) contained somewhat higher concentrations of shino-

rine (X ± SD: 0.88 ± 0.27 nmol mg-1 dry wt) than oursamples (Table 1, X ± SD: 0.28 ± 0.12 nmol mg-1 dry wt,

N = 12), but contained no palythine, asterina 330, or usuji-rene. Specimens of Pterocladia (Ptercladiella) capillaceafrom Japan (Lee and Shiu 2009) showed similar concentra-

tions of total MAAs (maximum *9 nmol mg-1 dry wt) toour samples (Table 1, X ± SD: 5.34 ± 3.02 nmol mg-1

dry wt, N = 12), but while they found porphyra-334 to be

dominant and shinorine and palythinol to be present, wefound very little porphyra-334 and no palythinol (Table 1).

Our results demonstrate the ubiquity of MAAs in Rhodo-

phyte macroalgae and establish that MAAs are available toconsumers, such as purple sea urchins, in Central Califor-

nia’s coastal marine environment.

Sex- and tissue-specific differences in concentrations

of MAAs in sea urchins

The existence of higher concentrations of MAAs in ovaries

compared to testes is not surprising (Fig. 3; Table 2) and is

consistent with other studies in corals (Michalek-Wagner2001), holothuroids (Karentz et al. 1991) and sea urchins

(Bosch et al. 1994; Carroll and Shick 1996; Karentz et al.

1997; McClintock and Karentz 1997). MAAs confer pro-tection from cleavage delay and fatal developmental

abnormalities in irradiated S. droebachiensis embryos

(Adams and Shick 1996, 2001), so MAAs may protect S.purpuratus embryos similarly. In comparison, MAAs can

only protect sperm during the *20 min (Pennington 1985)

fertilization period. Further, MAA uptake in sperm islimited by sperm’s small size, while eggs, which are

34,000 times larger (Levitan 1993), can more easily utilize

MAAs. Adams and Shick (1996) calculated that MAAs inS. droebachiensis sperm may not protect the nucleus due to

sperm’s small optical radius. However, it is possible that

MAAs present the fluid surrounding the sperm could pro-vide protection, similar to MAA-containing cells sur-

rounding tunicate embryos (Epel et al. 1999).

Concentrations of individual MAAs and absorption

spectra of sea urchin tissues

The differences in MAA signatures and absorption spectra

of testes versus the other tissues (Table 2; Fig. 5) appear to

represent a strategy by males where the breadth of UVR

absorption by MAAs is decreased, while the relative

absorption at higher energy UVR wavelengths is increased.The small size of sperm limit the number of MAA mole-

cules they can contain (Adams and Shick 1996), but the

over twofold increase in the relative concentration of UVB-absorbing palythine in testes (kmax = 320 nm, X ± SE:

88.5 ± 1.3 %, N = 119) may protect sperm DNA by

maximally absorbing UVR in the shorter, higher energywavelengths (Fig. 5, kmax = 322 nm). The concentration

of palythine in the ovaries was comparable to that in testes(Table 2), which correlates with elevated absorption of

these tissues in the UVB range compared to epidermal

tissues (Fig. 5). Shinorine (kmax = 334 nm) dominatedovary and epidermal tissues and likely caused the maximal

absorption of UVR to occur at longer wavelengths com-

pared to testes (Table 2; Fig. 5). Finally, the combinationof MAAs found in ovaries and epidermal tissues (Table 2)

caused UV absorption spanning a broad range of wave-

lengths (Fig. 5), which could protect the more diversecellular components of developing larvae and epidermis.

Some MAAs have variable antioxidant capabilities,

which may mitigate negative effects of oxidation in seaurchin tissues. For example, mycosporine glycine effec-

tively scavenges free radicals (Dunlap and Yamamoto

1995; de la Coba et al. 2009) and was primarily found inovaries (Table 2) so may provide further protection to

developing embryos. Elevated relative concentrations of

palythine in testes compared to other MAAs (Table 2) mayalso confer increased scavenging of superoxide radicals

and inhibition of lipid peroxidation as found for asterina-

330 plus palythine in algal extracts (de la Coba et al. 2009).These sex- and tissue-specific differences in MAA sig-

natures may be due to sex-specific dietary preferences, to

physiological differences in MAA uptake or transport todifferent tissues, or to interconversion of MAAs by sea

urchins or by bacteria. No studies have shown sex-specific

feeding preferences for any sea urchin species, probablybecause sex is difficult to determine externally. Alterna-

tively, uptake of MAAs may vary between sexes with

differing rates of interconversion of shinorine to otherMAAs by the bacteria Vibrio harveyii, which is found in

the guts of other sea urchins and holothuroids (Dunlap and

Shick 1998; Bandaranayake and Des Rocher 1999; Shickand Dunlap 2002). Tissue-specific differences in concen-

trations of individual MAAs may also be caused by inter-

conversion of MAAs within tissues (Whitehead et al. 2001;Conde et al. 2003; Shick 2004; reveiwed by Carreto and

Carignan 2011) or by differences in the transfer of MAAs

from the gut to tissues (Mason et al. 1998; Carefoot et al.2000; Shick and Dunlap 2002). For instance, increased

palythine in S. purpuratus testes may be a result of con-

version of dietary shinorine or porphyra-334 to mycospo-rine glycine then palythine in testes (Whitehead et al. 2001;

Mar Biol

123

Shick 2004). Alternatively, different carrier proteins may

transfer palythine to testes (Dunlap et al. 1989; Mason et al.1998), or differences in chemical properties, such as the

more non-polar character of palythine (Furusaki et al.

1980) or the acidic and polar nature of shinorine, mayallow them to be more easily transferred to or intercon-

verted within different tissues.

Microhabitat variation in concentrations of MAAs

in sea urchins

Our data suggest that microhabitat has long-term fitness or

survival consequences for intertidal S. purpuratus. Spe-cifically, intertidal S. purpuratus living in pits may produce

larvae with lower survival due to decreased MAAs

(Fig. 3a, b; Adams and Shick 1996, 2001). This potentialdisadvantage of inhabiting a pit may be offset by higher

adult survival due to reduced predation (Grupe 2006),

desiccation, or dislodgement by waves (Denny and Gay-lord 1996). Thus, inhabiting a pit likely increases adult

survival but reduces offspring viability.

Though we do not have data on the specific UV expo-sure nor the actual dietary MAA intake of each sea urchin

studied, we used microhabitat as a proxy for both solar

exposure (Fig. 1) and dietary MAA availability (Fig. 2).Thus, our inferences regarding these factors on concen-

trations of MAAs in sea urchins are speculative. However,

many studies have detected effects of UV exposure anddietary MAAs on concentrations of MAAs in animal tis-

sues (Karentz et al. 1997; Newman et al. 2000; Adams

et al. 2001; Lamare et al. 2004).Males appear to respond more strongly to UVR than

females; a nearly twofold increase in the proportional con-

centrations of MAAs in epidermal tissues of males (Fig. 4) inthe sunniest (Fig. 1) and highest Non-pit microhabitat was

observed. This could mitigate negative effects of UVR,

potentially maintaining growth or survival in these highstress environments. Interestingly, this response did not

occur in females (Fig. 4). MAAs are important for embryo

and larval health and survival (Adams and Shick 1996,2001), so females may allocate a higher proportion of MAAs

to their ovaries regardless of their own exposure to UVR.

These findings suggest the sexes have differing abilities toprotect themselves from increased UVR with climate change

(McKenzie et al. 2007; Zepp et al. 2011).

We observed no clear trend of increased concentrationsof MAAs in gonadal tissue (Fig. 3a, b) with sunlight

exposure across microhabitats (Fig. 1), consistent with the

results for S. droebachiensis under different UVR treat-ments (Adams et al. 2001). This is not surprising because

very little ambient light penetrates the test to expose the

gonads to UV (Walker et al. 2007; Adams, unpublisheddata for S. purpuratus). Thus, it would not be advantageous

for females to alter concentrations of MAAs in their ova-

ries based on their immediate UVR conditions beforereleasing embryos into the plankton.

Surveys of attached algal MAA and drift algal avail-

ability show no association with the concentrations ofMAAs in sea urchin tissues (Figs. 2, 3). These results are

surprising because strong effects of dietary MAAs has been

documented in other sea urchins (Adams et al. 2001; La-mare et al. 2004), seahares (Carefoot et al. 2000), krill

(Newman et al. 2000), and marine snails (Przeslawski et al.2005). Unfortunately, it was impossible to sample actual

intake of algae in the field, and the lack of correlation

between these factors suggest algal availability does notnecessarily translate to algal consumption. For example, Pit

urchins had low concentrations of MAAs (Fig. 3) but had

moderate levels of attached and drift algal availability(Fig. 2). Conversely, Non-pit urchins had moderate to high

concentrations of MAAs (Fig. 3) but low attached and

perhaps drift algal availability (Fig. 2).The lack of correlation among the microhabitats between

gonadal index (an indicator of dietary biomass, Vadas 1968;

Himmelman 1978) and concentrations of MAAs (an indi-cator of dietary MAAs, Adams et al. 2001; Lamare et al.

2004) suggests sea urchins in different microhabitats had

different diets, and may explain the lack of associationbetween algal availability and concentrations of MAAs in

sea urchins. For example, both Pit and Non-pit urchins had

moderate to high gonadal indices, but Pit urchins had lowconcentrations of MAAs and Non-pit urchins had moderate

to high concentrations (Fig. 3). Pit urchins likely stayed in

their pits (Grupe 2006) and relied mostly on caught driftalgae, which showed lower concentrations of MAAs by

weight than attached fleshy algae. Further, brown algae

(Phaeophyceae) were common in drift but do not containhigh amounts of MAAs (Shick and Dunlap 2002). Non-pit

urchins may have been obtaining MAAs by consuming algae

faster, supplementing their diets with microalgae, or con-suming available algae high in MAAs. Individual algae from

the Non-pit microhabitat contained higher concentrations of

MAAs than conspecifics in other microhabitats (see results),possibly due to both high sunlight levels (Fig. 1) and high

shore levels. Further, high relative concentrations of shino-

rine, the primary MAA accumulated by sea urchin tissues(Table 2), were found in both sea urchins and algae in the

Non-pit microhabitat (Figs. S1 and S2) suggesting a potential

transfer from algae to sea urchins, as has been noted betweentrophic levels for MAAs in several other organisms (Carroll

and Shick 1996; Newman et al. 2000; Carefoot et al. 2000).

Future research

Our study highlights correlations of UV exposure andMAAs from algal diets with concentrations of MAAs in

Mar Biol

123

S. purpuratus tissues. However, specific relationships

between these factors remain unclear and further research isnecessary and underway in our laboratory. Our current

research examines synergistic effects of natural UV exposure

and controlled MAA diets on the concentrations of MAAs ingonads and epidermis tissues of S. purpuratus over time.

These experiments will explicitly test the correlative patterns

in concentrations of MAAs in sea urchins with UV exposureand algal availability noted in this field survey.

Conclusion

This study is the first to identify multiple MAAs in the purple

sea urchin, S. purpuratus, and in eight species of Rhodophyte

macroalgae. High concentrations of MAAs in ovaries andbroadband UV protection from many types of MAAs suggest

they are important for protecting sea urchin embryos from

developmental delay and abnormalities (Adams and Shick1996, 2001). The narrower absorption spectrum of testes

peaked at higher energy UV wavelengths and suggests males

were maximizing the concentration the MAA palythine toprotect sperm DNA. Inhabiting a pit reduced overall con-

centrations of MAAs, with potentially far-reaching fitness

consequences, especially for offspring survival. Our datasuggest that male S. purpuratus increase concentrations of

epidermal MAAs in sunny environments, but that females do

not, perhaps because allocation of MAAs to ovaries byfemales is prioritized and crucial for embryo survival. How-

ever, the ability to phenotypically adjust concentrations of

MAAs depending on sunlight levels may be important forS. purpuratus populations in the near future, as the penetration

of UVR into the atmosphere is predicted to continue to be above

natural levels due to ozone thinning and ozone interactionswith climate change (McKenzie et al. 2007; Zepp et al. 2011).

Acknowledgments We thank many volunteers on this projectespecially L. Chang, M. Daugherty, J. Campanale, A. Lydon,J. Kershner, B. Brown and T. Moylan. Thanks to M. Moline,S. Morgan and S. Miller for guidance on the project and manuscriptand J. W. White, J. Sklar, and S. Rein for help with statistical anal-yses. We appreciate the generosity of W. Dunlap for providing MAAstandards, D. Pilliod and R. Arkle for the Solar Pathfinder,K. A. Miller for algal identification, and the Cal Poly BiologyDepartment for supplies and advice. Funding sources include CalPoly College-Based Fees to S. Gravem, Office of Naval Researchaward # N00014-04-1-0436 to N. Adams, and the National ScienceFoundation Grant IBN—0417003 awarded to N. Adams. All experi-ments and collections complied with the current laws of the UnitedStates. The authors declare that they have no conflict of interest.

References

Adams NL, Shick JM (1996) Mycosporine-like amino acids provideprotection against ultraviolet radiation in eggs of the green sea

urchin Strongylocentrotus droebachiensis. Photochem Photobiol64(1):149–158. doi:10.1111/j.1751-1097.1996.tb02435

Adams NL, Shick JM (2001) Mycosporine-like amino acids preventUVb-induced abnormalities during early development of thegreen sea urchin Strongylocentrotus droebachiensis. Mar Biol138(2):267–280. doi:10.1007/s002270000463

Adams NL, Shick JM, Dunlap WC (2001) Selective accumulation ofmycosporine-like amino acids in ovaries of the green sea urchinStrongylocentrotus droebachiensis is not affected by ultravioletradiation. Mar Biol 138(2):281–294. doi:10.1007/s002270000464

Arkle RS, Pilliod DS (2010) Prescribed fires as ecological surrogatesfor wildfires: a stream and riparian perspective. Forest EcolManag 259(5):893–903. doi:10.1016/j.foreco.2009.11.029

Bandaranayake WM, Des Rocher A (1999) Role of secondarymetabolites and pigments in the epidermal tissues, ripe ovaries,viscera, gut contents and diet of the sea cucumber Holothuriaatra. Mar Biol 133(1):163–169. doi:10.1007/s002270050455

Bosch I, Janes P, Schack R, Steves B, Karentz D (1994) Survey ofUV-absorbing compounds in sub-tropical sea urchins fromFlorida and the Bahamas. Am Zool 34:102A

Campanale JP, Tomanek L, Adams NL (2011) Exposure to ultravioletradiation causes proteomic changes in embryos of the purple seaurchin, Strongylocentrotus purpuratus. J Exp Mar Biol Ecol397:106–120. doi:10.1016/j.jembe.2010.11.022

Carefoot TH, Karentz D, Pennings SC, Young CL (2000) Distributionof mycosporine-like amino acids in the sea hare Aplysiadactylomela: effect of diet on amounts and types sequesteredover time in tissues and spawn. Comp Biochem Physiol C126(1):91–104. doi:10.1016/S0742-8413(00)00098-0

Carreto JI, Carignan MO (2011) Mycosporine-like amino acids:relevant secondary metabolites. Chemical and ecologicalaspects. Mar Drugs 9(3):387–446. doi:10.3390/md9030387

Carroll AK, Shick JM (1996) Dietary accumulation of UV-absorbingmycosporine-like amino acids (MAAs) by the green sea urchin(Strongylocentrotus droebachiensis). Mar Biol 124(4):561–569.doi:10.1007/BF00351037

Conde FR, Carignan MO, Churio MS, Carreto JI (2003) In vitro cis-transphotoisomerization of palythene and usujirene: implications on thein vivo transformation of mycosporine-like amino acids. Photo-chem Photobiol 77(2):146–150. doi:10.1562/0031-8655(2003)0770146IVCTPO2.0.CO2

Dayton PK (1975) Experimental evaluation of ecological dominancein a rocky intertidal algal community. Ecol Monogr 45(2):137–159. doi:10.2307/1942404

de la Coba F, Aguilera J, Figueroa FL, de Galvez MV, Herrera E(2009) Antioxidant activity of mycosporine-like amino acidsisolated from three red macroalgae and one marine lichen.J Appl Phycol 21(2):161–169. doi:10.1007/s10811-008-9345-1

Denny M, Gaylord B (1996) Why the urchin lost its spines:hydrodynamic forces and survivorship in three echinoids.J Exp Biol 199(3):717–729

Dunlap WC, Chalker BE (1986) Identification and quantitation ofnear-UV absorbing compounds S-320 in a hermatypic sclerac-tinian. Coral Reefs 5(3):155–160. doi:10.1007/BF00298182

Dunlap WC, Shick JM (1998) Ultraviolet radiation-absorbingmycosporine-like amino acids in coral reef organisms: abiochemical and environmental perspective. J Phycol 34(3):418–430. doi:10.1046/j.1529-8817.1998.340418

Dunlap WC, Yamamoto Y (1995) Small-molecule antioxidants inmarine organisms: antioxidant activity of mycosporine-glycine.Comp Biochem Physiol B 112(1):105–114. doi:10.1016/0305-0491(95)00086-N

Dunlap WC, Williams DM, Chalker BE, Banaszak AT (1989)Biochemical photoadaptation in vision: UV-absorbing pigmentsin fish eye tissues. Comp Biochem Physiol B 93(3):601–608.doi:10.1016/0305-0491(89)90383-0

Mar Biol

123

http://dx.doi.org/10.1111/j.1751-1097.1996.tb02435

http://dx.doi.org/10.1007/s002270000463

http://dx.doi.org/10.1007/s002270000464

http://dx.doi.org/10.1016/j.foreco.2009.11.029

http://dx.doi.org/10.1007/s002270050455


http://dx.doi.org/10.1016/S0742-8413(00)00098-0

http://dx.doi.org/10.3390/md9030387

http://dx.doi.org/10.1007/BF00351037

http://dx.doi.org/10.1562/0031-8655(2003)0770146IVCTPO2.0.CO2

http://dx.doi.org/10.1562/0031-8655(2003)0770146IVCTPO2.0.CO2

http://dx.doi.org/10.2307/1942404

http://dx.doi.org/10.1007/s10811-008-9345-1

http://dx.doi.org/10.1007/BF00298182

http://dx.doi.org/10.1046/j.1529-8817.1998.340418

http://dx.doi.org/10.1016/0305-0491(95)00086-N

http://dx.doi.org/10.1016/0305-0491(95)00086-N

http://dx.doi.org/10.1016/0305-0491(89)90383-0

Ebert TA (1968) Growth rates of the sea urchin Strongylocentrotuspurpuratus related to food availability and spine abrasion.Ecology 49(6):1075–1901. doi:10.2307/1934491

Epel D, Hemela K, Shick M, Patton C (1999) Development in thefloating world: defenses of eggs and embryos against damagefrom UV radiation. Am Zool 39(2):271–278. doi:10.1093/icb/39.2.271

Forster P, Ramaswamy V, Artaxo P, Berntsen T, Betts R, Fahey DW,Haywood J, Lean J, Lowe DC, Myhre G, Nganga J, Prinn R,Raga G, Schulz M, Van Dorland R (2007) Changes inatmospheric constituents and in radiative forcing. In: SolomonS, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, TignorM, Miller HL (eds) Climate change 2007: the physical sciencebasis. Contribution of working group I to the fourth assessmentreport of the intergovernmental panel on climate change.Cambridge University Press, Cambridge

Franklin LA, Forster RM (1997) The changing irradiance environ-ment: consequences for marine macrophyte physiology, produc-tivity and ecology. Eur J Phycol 32(3):207–232. doi:10.1080/09670269710001737149

Franklin LA, Yakovleva I, Karsten U, Luning K (1999) Synthesis ofmycosporine-like amino acids in Chondrus crispus (Florideo-phyceae) and the consequences for sensitivity to ultraviolet Bradiation. J Phycol 35(4):682–693. doi:10.1080/09670269710001737149

Furusaki A, Matsumoto T, Tsujino I, Sekikawa I (1980) Crystal andmolecular structure of palythine trihydrate. Bull Chem Soc Jpn53(2):319–323. doi:10.1246/bcsj.53.319

Giese AC, Pearse JS, Pearse VB (1991) Reproduction of marineinvertebrates: echinoderms and lophophorates, vol 6. Blackwell,New York

Grupe BM (2006) Purple sea urchins (Stronglyocentrotus purpuratus)in and out of pits: the effects of microhabitat on populationstructure, morphology, growth and mortality. Master Thesis.University of Oregon, Eugene

Hader DP, Helbling EW, Williamson CE, Worrest RC (2011) Effectsof UV radiation on aquatic ecosystems and interactions withclimate change. Photochem Photobiol 10(2):242–260. doi:10.1039/c0pp90036b

Himmelman JH (1978) Reproductive cycle of the green sea urchin,Strongylocentrotus droebachiensis. Can J Zool 56(8):1828–1836. doi:10.1139/z78-249

Hoyer K, Karsten U, Wiencke C (2002) Induction of sunscreencompounds in Antarctic macroalgae by different radiationconditions. Mar Biol 141(4):619–627. doi:10.1007/s00227-002-0871-0

Karentz D, McEuen FS, Land MC, Dunlap WC (1991) Survey ofmycosporine-like amino acid compounds in Antarctic marineorganisms: potential protection from UV exposure. Mar Biol108(1):157–166. doi:10.1007/BF01313484

Karentz D, Dunlap WC, Bosch I (1997) Temporal and spatialoccurrence of UV-absorbing mycosporine-like amino acids intissues of the antarctic sea urchin Sterechinus neumayeri duringspringtime ozone-depletion. Mar Biol 129(2):343–353. doi:10.1007/s002270050174

Karsten U, Sawall T, Hanelt D, Bischof K, Figueroa FL, Flores-MoyaA, Wiencke C (1998) An inventory of UV-absorbing mycosp-orine-like amino acids in macroalgae from polar to warm-temperate regions. Bot Mar 41(5):443–453. doi:10.1515/botm.1998.41.1-6.443

Karsten U, Bischof K, Hanelt D, Tug H, Wiencke C (1999) The effectof ultraviolet radiation on photosynthesis and ultraviolet-absorb-ing substances in the endemic Arctic macroalga Devaleraearamentacea (Rhodophyta). Physiol Plant 105(1):58–66. doi:10.1034/j.1399-3054.1999.105110

Lamare MD, Hoffman J (2004) Natural variation of carotenoids in theeggs and gonads of the echinoid genus, Strongylocentrotus:implications for their role in ultraviolet radiation photoprotec-tion. J Exp Mar Biol Ecol 312(2):215–233. doi:10.1016/j.jembe.2004.02.016

Lamare MD, Lesser MP, Barker MF, Barry TM, Schimanski KB(2004) Variation in sunscreen compounds (mycosporine-likeamino acids) for marine species along a gradient of ultravioletradiation transmission within Doubtful Sound, New Zealand. NZJ Mar Freshwat Res 38(5):775–793

Lamare MD, Barker MF, Lesser MP, Marshall C (2006) DNAphotorepair in echinoid embryos: effects of temperature onrepair rate in Antarctic and non-Antarctic species. J Exp Biol209(24):5017–5028. doi:10.1242/jeb.02598

Lamare M, Burritt D, Lister K (2011) Ultraviolet radiation andechinoderms: past, present and future perspectives. Adv MarBiol 59:145–187. doi:10.1016/B978-0-12-385536-7.00004-2

Lee TM, Shiu CT (2009) Implications of mycosporine-like aminoacid and antioxidant defenses in UV-B radiation tolerance for thealgae species Pterocladiella capillacea and Gelidium amansii.Mar Environ Res 67(1):8–16. doi:10.1016/j.marenvres.2008.09.006

Lesser MP (2010) Depth-dependent effects of ultraviolet radiation onsurvivorship, oxidative stress and DNA damage in sea urchin(Strongylocentrotus droebachiensis) embryos from the Gulf ofMaine. Photochem Photobiol 86(2):382–388. doi:10.1111/j.1751-1097.2009.00671

Lesser MP, Kruse VA, Barry TM (2003) Exposure to ultravioletradiation causes apoptosis in developing sea urchin embryos.J Exp Biol 206(22):4097–4103. doi:10.1242/jeb.00621

Lesser MP, Lamare MD, Barker MF (2004) Transmission ofultraviolet radiation through the Antarctic annual sea ice andits biological effects on sea urchin embryos. Limnol Oceanogr49(6):1957–1963. doi:10.4319/lo.2004.49.6.1957

Lesser MP, Barry TM, Lamare MD, Barker MF (2006) Biologicalweighting functions for DNA damage in sea urchin embryosexposed to ultraviolet radiation. J Exp Mar Biol Ecol328(1):10–21. doi:10.1016/j.jembe.2005.06.010

Levitan DR (1993) The importance of sperm limitation to the evolutionof egg size in marine invertebrates. Am Nat 141(4):517–536

Li HW, Lamberti GA, Pearsons TN, Tait CK, Li JL, Buckhouse JC(1994) Cumulative effects of riparian disturbances along highdesert trout streams of the John Day Basin, Oregon. T Am FishSoc 123(4):627–640. doi:10.1577/1548-8659(1994)123\0627:ceorda[2.3.co;2

Madronich S, McKenzie RL, Bjorn LO, Caldwell MM (1998)Changes in biologically active ultraviolet radiation reaching theearth’s surface. J Photochem Photobiol B Biol 46(1–3):5–19.doi:10.1016/S1011-1344(98)00182-1

Mason DS, Schafer F, Shick JM, Dunlap WC (1998) Ultravioletradiation-absorbing mycosporine-like amino acids (MAAs) areacquired from their diet by medaka fish (Oryzias latipes) but notby SKH-1 hairless mice. Comp Biochem Physiol A120(4):587–598. doi:10.1016/S1095-6433(98)10069-7

McClintock JB, Karentz D (1997) Mycosporine-like amino acids in38 species of subtidal marine organisms from McMurdo Sound,Antarctica. Antarct Sci 9(4):392–398. doi:10.1017/S0954102097000503

McKenzie RL, Aucamp PJ, Bais AF, Bjorn LO, Ilyas M (2007)Changes in biologically-active ultraviolet radiation reaching theearth’s surface. Photochem Photobiol Sci 6(3):218–231. doi:10.1016/S1011-1344(98)00182-1

Michalek-Wagner K (2001) Seasonal and sex-specific variations in levelsof photo-protecting mycosporine-like amino acids (MAAs) in softcorals. Mar Biol 139(4):651–660. doi:10.1007/s002270100625

Mar Biol

123

http://dx.doi.org/10.2307/1934491

http://dx.doi.org/10.1093/icb/39.2.271

http://dx.doi.org/10.1093/icb/39.2.271

http://dx.doi.org/10.1080/09670269710001737149

http://dx.doi.org/10.1080/09670269710001737149

http://dx.doi.org/10.1080/09670269710001737149

http://dx.doi.org/10.1080/09670269710001737149

http://dx.doi.org/10.1246/bcsj.53.319

http://dx.doi.org/10.1039/c0pp90036b

http://dx.doi.org/10.1039/c0pp90036b

http://dx.doi.org/10.1139/z78-249

http://dx.doi.org/10.1007/s00227-002-0871-0

http://dx.doi.org/10.1007/s00227-002-0871-0

http://dx.doi.org/10.1007/BF01313484

http://dx.doi.org/10.1007/s002270050174

http://dx.doi.org/10.1515/botm.1998.41.1-6.443

http://dx.doi.org/10.1515/botm.1998.41.1-6.443

http://dx.doi.org/10.1034/j.1399-3054.1999.105110




http://dx.doi.org/10.1016/B978-0-12-385536-7.00004-2

http://dx.doi.org/10.1016/j.marenvres.2008.09.006

http://dx.doi.org/10.1016/j.marenvres.2008.09.006

http://dx.doi.org/10.1111/j.1751-1097.2009.00671

http://dx.doi.org/10.1111/j.1751-1097.2009.00671


http://dx.doi.org/10.4319/lo.2004.49.6.1957


http://dx.doi.org/10.1577/1548-8659(1994)123%3c0627:ceorda%3e2.3.co;2

http://dx.doi.org/10.1577/1548-8659(1994)123%3c0627:ceorda%3e2.3.co;2

http://dx.doi.org/10.1016/S1011-1344(98)00182-1

http://dx.doi.org/10.1016/S1095-6433(98)10069-7

http://dx.doi.org/10.1017/S0954102097000503

http://dx.doi.org/10.1017/S0954102097000503

http://dx.doi.org/10.1016/S1011-1344(98)00182-1

http://dx.doi.org/10.1007/s002270100625

Naumburg E, DeWald LE (1999) Relationships between Pinusponderosa forest structure, light characteristics, and understorygraminoid species presence and abundance. Forest Ecol Manag124(2–3):205–215. doi:10.1016/s0378-1127(99)00067-5

Neale PJ, Banaszak AT, Jarriel CR (1998) Ultraviolet sunscreens inGymnodinium sanguineum (Dinophyceae): mycosporine-likeamino acids protect against inhibition of photosynthesis. J Phycol34(6):928–938. doi:10.1046/j.1529-8817.1998.340928

Newman SJ, Dunlap WC, Nicol S, Ritz D (2000) Antarctic krill(Euphausia superba) acquire a UV-absorbing mycosporine-likeamino acid from dietary algae. J Exp Mar Biol Ecol255(1):93–110. doi:10.1016/S0022-0981(00)00293-8

Oren A, Gunde-Cimerman N (2007) Mycosporines and mycosporine-like amino acids: UV protectants or multipurpose secondarymetabolites? FEMS Microbiol Lett 269(1):1–10. doi:10.1111/j.1574-6968.2007.00650

Pennington JT (1985) The ecology of fertilization of echinoid eggs—the consequences of sperm dilution, adult aggregation, andsynchronous spawning. Biol Bull 169(2):417–430

Przeslawski R, Benkendorff K, Davis AR (2005) A quantitativesurvey of mycosporine-like amino acids (MAAs) in intertidalegg masses from temperate rocky shores. J Chem Ecol31(10):2417–2438. doi:10.1007/s10886-005-7110-3

Rastogi RP, Richa SinhaRP, Singh SP, Hader DP (2010) Photopro-tective compounds from marine organisms. J Ind MicrobiolBiotechnol 37(6):537–558. doi:10.1007/s10295-010-0718-5

Sea Urchin Genome Sequencing Consortium, Sodergren E, Wein-stock GM, Davidson EH, Cameron RA, Gibbs RA, Angerer RC,Angerer LM, Arnone MI, Burgess DR, Burke RD, Coffman JA,Dean M, Elphick MR, Ettensohn CA, Foltz KR, Hamdoun A,Hynes RO, Klein WH, Marzluff W, McClay DR, Morris RL,Mushegian A, Rast JP, Smith LC, Thorndyke MC, Vacquier VD,Wessel GM, Wray G, Zhang L, Elsik CG, Ermolaeva O, HlavinaW, Hofmann G, Kitts P, Landrum MJ, Mackey AJ, Maglott D,Panopoulou G, Poustka AJ, Pruitt K, Sapojnikov V, Song X,Souvorov A, Solovyev V, Wei Z, Whittaker CA, Worley K,Durbin KJ, Shen Y, Fedrigo O, Garfield D, Haygood R, PrimusA, Satija R, Severson T, Gonzalez-Garay ML, Jackson AR,Milosavljevic A, Tong M, Killian CE, Livingston BT, Wilt FH,Adams N, Belle R, Carbonneau S, Cheung R, Cormier P, CossonB, Croce J, Fernandez-Guerra A, Geneviere AM, Goel M, KelkarH, Morales J, Mulner-Lorillon O, Robertson AJ, Goldstone JV,Cole B, Epel D, Gold B, Hahn ME, Howard-Ashby M, Scally M,Stegeman JJ, Allgood EL, Cool J, Judkins KM, McCafferty SS,Musante AM, Obar RA, Rawson AP, Rossetti BJ, Gibbons IR,Hoffman MP, Leone A, Istrail S, Materna SC, Samanta MP,Stolc V, Tongprasit W, Tu Q, Bergeron KF, Brandhorst BP,Whittle J, Berney K, Bottjer DJ, Calestani C, Peterson K, ChowE, Yuan QA, Elhaik E, Graur D, Reese JT, Bosdet I, Heesun S,Marra MA, Schein J, Anderson MK, Brockton V, Buckley KM,Cohen AH, Fugmann SD, Hibino T, Loza-Coll M, Majeske AJ,Messier C, Nair SV, Pancer Z, Terwilliger DP, Agca C,Arboleda E, Chen N, Churcher AM, Hallbook F, HumphreyGW, Idris MM, Kiyama T, Liang S, Mellott D, Mu X, Murray G,Olinski RP, Raible F, Rowe M, Taylor JS, Tessmar-Raible K,Wang D, Wilson KH, Yaguchi S, Gaasterland T, Galindo BE,Gunaratne HJ, Juliano C, Kinukawa M, Moy GW, Neill AT,Nomura M, Raisch M, Reade A, Roux MM, Song JL, Su YH,Townley IK, Voronina E, Wong JL, Amore G, Branno M, BrownER, Cavalieri V, Duboc V, Duloquin L, Flytzanis C, Gache C,Lapraz F, Lepage T, Locascio A, Martinez P, Matassi G,Matranga V, Range R, Rizzo F, Rottinger E, Beane W, Bradham

C, Byrum C, Glenn T, Hussain S, Manning G, Miranda E,Thomason R, Walton K, Wikramanayke A, Wu SY, Xu R,Brown CT, Chen L, Gray RF, Lee PY, Nam J, Oliveri P, Smith J,Muzny D, Bell S, Chacko J, Cree A, Curry S, Davis C, Dinh H,Dugan-Rocha S, Fowler J, Gill R, Hamilton C, Hernandez J,Hines S, Hume J, Jackson L, Jolivet A, Kovar C, Lee S, Lewis L,Miner G, Morgan M, Nazareth LV, Okwuonu G, Parker D,Pu LL, Thorn R, Wright R (2006) The genome of the sea urchinStrongylocentrotus purpuratus. Science 314(5801):941–952.doi:10.1126/science.1133609

Sekikawa I, Kubota C, Hiraoki T, Tsujino I (1986) Isolation andstructure of a 357 nm UV-absorbing substance usujirene fromthe red alga Palmaria palmata (l. O. Kuntze). Jpn J Phycol34(3):185–188

Shick JM (2004) The continuity and intensity of ultraviolet irradiationaffect the kinetics of biosynthesis, accumulation, and conversionof mycosporine-like amino acids (MAAs) in the coral Stylophorapistillata. Limnol Oceanogr 49(2):442–458

Shick JM, Dunlap WC (2002) Mycosporine-like amino acids andrelated gadusols: biosynthesis, accumulation, and UV-protectivefunctions in aquatic organisms. Annu Rev Physiol 64:223–262.doi:10.1146/annurev.physiol.64.081501.155802

Shick JM, Romaine-Lioud S, Ferrier-Pages C, Gattuso JP (1999)Ultraviolet-B radiation stimulates shikimate pathway-dependentaccumulation of mycosporine-like amino acids in the coralStylophora pistillata despite decreases in its population ofsymbiotic dinoflagellates. Limnol Oceanogr 44(7):1667–1682

Sinha RP, Klisch M, Groniger A, Hader DP (1998) Ultraviolet-absorbing/screening substances in cyanobacteria, phytoplanktonand macroalgae. J Photochem Photobiol B Biol 47(2–3):83–94.doi:10.1016/s1011-1344(98)00198-5

Sinha RP, Singh SP, Hader DP (2007) Database on mycosporines andmycosporine-like amino acids (MAAs) in fungi, cyanobacteria,macroalgae, phytoplankton and animals. J Photochem PhotobiolB Biol 89(1):29–35. doi:10.1016/j.jphotobiol.2007.07.006

Strathmann MF (1987) Reproduction and development of marineinvertebrates of the northern Pacific Coast: data and methods forthe study of eggs, embryos, and larvae. University of Washing-ton Press, Seattle

Tedetti M, Sempere R (2006) Penetration of ultraviolet radiation inthe marine environment, a review. Photochem Photobiol 82(2):389–397. doi:10.1562/2005-11-09-IR-733

Vadas RL (1968) The ecology of Agarum and the kelp bedcommunity. Dissertation. Univeristy of Washington, Seattle

van de Poll WH, Hanelt D, Hoyer K, Buma AGJ, Breeman AM(2002) Ultraviolet-B-induced cyclobutane-pyrimidine dimerformation and repair in arctic marine macrophytes. Photo-chem Photobiol 76(5):493–500. doi:10.1562/0031-8655(2002)0760493UBICPD2.0.CO2

Walker C, Tatsuya U, Lesser M (2007) Gametogenesis andreproduction of sea urchins. In: Lawrence J (ed) Edible seaurchins: biology and ecology, vol 1. Elsevier, Amsterdam, p 24

Whitehead K, Karentz D, Hedges JI (2001) Mycosporine-like aminoacids (MAAs) in phytoplankton, a herbivorous pteropod (Lima-cina helicina), and its pteropod predator (Clione antarctica) inMcMurdo Bay, Antarctica. Mar Biol 139(5):1013–1019. doi:10.1007/s002270100654

Zepp RG, Erickson DJ, Paul ND, Sulzberger B (2011) Effects of solar uvradiation and climate change on biogeochemical cycling: interac-tions and feedbacks. Photochem Photobiol Sci 10(2):261–279. doi:10.1039/c0pp90037k

Mar Biol

123

http://dx.doi.org/10.1016/s0378-1127(99)00067-5

http://dx.doi.org/10.1046/j.1529-8817.1998.340928

http://dx.doi.org/10.1016/S0022-0981(00)00293-8

http://dx.doi.org/10.1111/j.1574-6968.2007.00650

http://dx.doi.org/10.1111/j.1574-6968.2007.00650

http://dx.doi.org/10.1007/s10886-005-7110-3

http://dx.doi.org/10.1007/s10295-010-0718-5


http://dx.doi.org/10.1146/annurev.physiol.64.081501.155802

http://dx.doi.org/10.1016/s1011-1344(98)00198-5

http://dx.doi.org/10.1016/j.jphotobiol.2007.07.006

http://dx.doi.org/10.1562/2005-11-09-IR-733

http://dx.doi.org/10.1562/0031-8655(2002)0760493UBICPD2.0.CO2

http://dx.doi.org/10.1562/0031-8655(2002)0760493UBICPD2.0.CO2

http://dx.doi.org/10.1007/s002270100654

http://dx.doi.org/10.1039/c0pp90037k

474 THE AMERICAN BIOLOGY TEACHER VOLUME 73, NO. 8, OCTOBER 2011

ABSTRACT

The accumulation of plastic in the oceans is an ever-growing environmental con-cern. Plastic debris is a choking and entanglement hazard for wildlife; plastics also leach toxic compounds into organisms and ecosystems. Educating students about the marine debris problem introduces fundamental concepts in toxicology, ecology, and oceanography. Students will learn about the toxicity of plastics, col-lect and analyze data on plastic debris, and put their new knowledge to work by writing a congressional bill that addresses the problem of marine debris.

Key Words: Ocean pollution; toxicology; ecosystem health; ocean literacy; envi-ronmental science.

The National Oceanic and Atmospheric Administration (NOAA) defines marine debris as “any persistent solid material that is manu-factured or processed and directly or indirectly, intentionally or unintentionally, disposed of or abandoned into the marine environment…” (http://marinedebris.noaa.gov/). This issue is but one example of numerous human impacts on the ocean. Unfor-tunately, most Americans are ocean-illiterate (Ocean Project, 2009) – that is, they do not understand “the ocean’s influence on you, and your influence on the ocean” (National Geographic Society, 2006). The activities described below will assist students in making connec-tions to the ocean.

The ever-growing problem of marine debris begins on land, where streams and rivers carry debris to the coast. Ocean currents then transport debris to remote areas, where it may take centuries to break down (Goldberg, 1994). Awareness of the marine debris problem has increased because of recent reports on the Eastern Pacific Garbage Patch (EPGP). The EPGP is an area between California and Hawaii that contains a large quantity of small “microplastic” pieces derived from the breakdown of larger plastic items (Marks & Howden, 2008). The plastic debris both near the coast and in the EPGP can endanger the health of marine organisms (Derraik, 2002).

Plastic debris poses a danger to all forms of aquatic life. Many marine organisms can become entangled or can ingest plastic debris. Certain marine species, such as sea turtles and seabirds, mistake plastic for prey items (Nevins et al., 2005; Hyrenbach et al., 2009;

Mrosovsky et al., 2009). For example, adult albatrosses inadvertently feed their chicks plastics instead of natural food items, which affects chick growth and may cause mortality (Ryan & Jackson, 1987; Pierce et al., 2004). Furthermore, an insidious hazard lurks within plastics. The toxic chemicals added to make plastics more flexible, known as plasticizers, can leach out into the environment and into organ-isms that ingest plastic (Rahman & Brazel, 2004). Other dangerous chemicals can concentrate on plastic surfaces (Mato et al., 2001), increasing the toxicity of plastics.

A major concern about the toxic compounds associated with plastics is that they can disrupt hormone regulation in the cells of

organisms (Oberdörster & Cheek, 2001). Hormone disruption occurs when a chemical acts as a natural hormone in a cell (Figure 1); it can change reproductive ability and mating behavior, contribute to tumor development, and negatively affect offspring (van de Merwe et al., 2010; Wuttke et al., 2010). For example, male fish exposed to hormone-disrupting com-

pounds can develop ovaries (Gray & Metcalfe, 1997). Additionally, certain plastics, such as styrene (Styrofoam), are carcinogenic (Vod-icka et al., 2006).

The lessons below (objectives listed in Figure 2, materials and equipment in Table 1) couple process skills with the underlying science of plastic pollution. They can be modified to be appro-priate for grades 6–10. Oikonos Ecosystem Knowledge in partner-ship with many collaborators (see Acknowledgments) created the overall marine debris curriculum. The plastic-pollution curriculum described here was further developed and modified by CAMEOS, a National Science Foundation Graduate K–12 program, at the Univer-sity of California Davis Bodega Marine Laboratory.

Lesson 1: Introduction to Marine Debris J

& Toxicology of PlasticsLesson Time:1 hour

This lecture familiarizes students with key concepts regarding the origin and transport of debris, the chemical structure of plastics,

The American Biology Teacher, Vol. 73, No. 8, pages 474–478. ISSN 0002-7685, electronic ISSN 1938-4211. ©2011 by National Association of Biology Teachers. All rights reserved. Request permission to photocopy or reproduce article content at the University of California Press’s Rights and Permissions Web site at www.ucpressjournals.com/reprintinfo.asp. DOI: 10.1525/abt.2011.73.8.9

H O W T O D O I T The Ecotoxicology of Plastic Marine Debris

S U S A N N E M . B R A N D E R , R A C H E L E . F O N TA N A , TAW N Y M . M ATA , S A R A H A . G R AV E M , A N N A L I E S E H E T T I N G E R , J E S S I C A R . B E A N , A M B E R I . S Z O B O S Z L A I , C A R O L A . K E I P E R , M E G H A N E . M A R R E R O

Plastic debris poses a

danger to all forms of

aquatic life.

THE AMERICAN BIOLOGY TEACHER ECOTOXICOLOGY 475

and the toxicology of plastic-associated chemicals and how they can affect marine organisms. The presentation, found on the CAMEOS website (http://bml.ucdavis.edu/education/cameos/resources/ ecotoxicology/) includes extensive teacher notes. Following Lesson 1, students will be able to define terms (e.g., plasticizer, bioindicator species) and understand concepts (e.g., how certain compounds disrupt hormones).

Lesson 2: Campus Debris Survey J

& Plastic AnalysisLesson Time: 90–120 minutes

Lesson 2 involves a campus debris survey and analysis of debris found. Students will formulate a scientific protocol; collect, analyze, and interpret data; and compare results.

Campus Debris Survey J

Assemble debris collection teams of 4 or 5 students. (Alternatively, UÊstudents can bring in a variety of plastic debris from home).Give students the list of materials in Table 1 and discuss the UÊimportance of each debris-collection team using the same pro-tocol. For example, in order to have comparable data between groups, students should be collecting in areas of similar size and over similar time intervals. Assign each team to a distinct area, such as a sports field (equiv-UÊalent areas approximately 20 × 20 m are best). Teams should measure and mark their collection area. One stu-UÊdent will record data (using clipboard with data sheet) while others wear protective gloves and place trash collected into two bags, one for plastics and one for nonplastics. All debris should be recorded and students should collect as much debris as possible in their designated area during the collection time (~15-minute collection period is best).

Remind students to record any observations as they are con-UÊducting the survey. Students should save plastic trash for part 2 of this lesson.

Figure 1. The cellular pathway of hormone disruption.

Figure 2. Ecotoxicology of Marine Debris lesson objectives.

Table 1. Materials and equipment needed.Lesson 1 Lesson 2 Lesson 3

Computer Transect tape “Making a Bill” worksheeta

Projector Oikonos data sheet for campus debris surveya Computer lab with Internet accessb

Introduction to Marine Plastics Presentationa

Clipboards

TrashbagsPlastic glovesStopwatchesb

Tables of plastic types and toxic compoundsData sheet for plastic analysisa

aAvailable on CAMEOS website, http://bml.ucdavis.edu/education/cameos/resources/ecotoxicology/bOptional, lesson can be done without


Plastic Data Analysis J

In the second part of this lesson, students will categorize and analyze the plastic debris to understand the kinds of plastics and associated contaminants found on their school campus.

Groups will sort collected plastics into categories described in Table 2 and note the contaminants associated with each category (Table 3), following the example in Figure 3. Upon completion, data can be graphed using Microsoft Excel or by hand. Data sheets and Excel lessons with step-by-step instructions can be found on the CAMEOS website. Students can summarize Lessons 1 and 2 in a lab-oratory report assignment, using the background information from Lesson 1 and the methods and results developed in Lesson 2.

Lesson 3: Putting Science into Action J

Lesson 3 requires students to combine new knowledge of marine debris and plastic pollution with concepts from civics and government classes. Students formulate and justify legislation on the basis of their scientific knowledge, reinforcing the importance of science to society.

Give students the “Making a Bill” worksheet found on the UÊCAMEOS website. The introduction describes how science influences policy and the actions a bill can mandate, such as pollutant monitoring.

Have students go to the Library of Congress website UÊ(http://thomas.loc.gov/bss/111search.html) to search for

Table 2. Types of plastics.Plastic Type Full Name Code Examples Recyclable?

PETE Polyethylene terephthalate 1 Soda bottles Yes

HDPE High density polyethylene 2 Milk jugs, shampoo bottles, yogurt containers Yes

PVC Polyvinyl chloride 3 Clear food packaging, candy wrappers, some bottles

Sometimes

LDPE Low density polyethylene 4 Squeezable bottles, shopping bags Yes

PP Polypropylene 5 Caps, straws, some bottles Yes

PS Polystyrene 6 Disposable plates and cups, CD cases Sometimes

PC, other Polycarbonate 7 Water jugs, sunglasses, DVDs Not usually

Table 3. Toxic compounds in or associated with plastics: their uses and effects.

Toxic Compound Use Effect(s)Plastic Type(s) Concentration

Bisphenol A (BPA) Plasticizer, can liner Mimics estrogen PVC, PC 43–483 mg/kg in PVC food wrappersa (López-Cervantes & Paseiro-Losada, 2003)

Phthalates Plasticizer, artificial fragrances

Interferes with testosterone, sperm motility

PS, PVC 0.5–30.8 mg/kg in food wrappersb (Castle et al., 1988)

Persistent organic pollutants (POPs)

Pesticides, flame retardants, etc.

Possible neurologi-cal and reproductive damage

All plastics

Dioxins Produced in manufacture of PVC, during waste incineration

Carcinogen, interferes with testosterone

All plastics

Nonylphenol Antistatic, antifog, surfactant (in detergents)

Mimics estrogen PVC 10–3300 μg/gc (Inoue et al., 2001)

Polyaromatic hydrocarbons (PAHs)

Produced when fossil fuels are burned

Developmental and reproductive toxicity

All plastics

Polychlorinated biphenyls (PCBs)

Electronics manufacture Interferes with thyroid function

All plastics

Styrene monomer Structure of polystyrene Forms DNA adducts PS <0.001–0.071 mg/kgd (Tawfik & Huyghebaert, 1998)

a Total amount detected in food wrappers.b Amount leached into various food products (i.e., candy bars, sandwiches).c Total amount migrated from PVC food wrap to various food products.d Amount leached into beverages from styrene cups (more styrene leached into beverages with higher fat content).

THE AMERICAN BIOLOGY TEACHER ECOTOXICOLOGY 477

environmental terms related to this topic (e.g., plastic, pol-lution, conservation). This process will introduce students to the content and format of environmental bills. If you do not have computer access at school, example bills are available for printing on the CAMEOS website.

Students will write a formal bill that includes a title, list of spon-UÊsors, background information, proposed action, and potential funding methods.

Good bills should incorporate knowledge from the marine UÊdebris lessons and be creative. Once bills are complete, have each group outline their bill and present it to the class for debate, discussion, and possibly a vote.

Websites with Supplemental Materials & InformationLesson materials on the CAMEOS website: http://bml.ucdavis.edu/education/cameos/resources/ecotoxicology/

Oikonos Ocean Stewardship’s website contains additional ocean edu-cation activities and resources not covered in this article: http://www.oikonos.org/projects/oceanstewardship.htm

Acknowledgments J

This material is produced in conjunction with the Bodega Marine Laboratory, University of California Davis, NSF GK-12 CAMEOS program. The National Science Foundation (grant no. 0841297) supported this work. Development of marine debris and watershed activities by Oikonos Ecosystem Knowledge were supported initially

by the California Coast Commission with support from Jennifer Stock of the Cordell Bank National Marine Sanctuary followed by support from the City of Benicia Water Education Program. Addi-tional contributions were made in partnership with ACES Signals of Spring, Hawaii Pacific University, U.S. Fish and Wildlife, and Richardson Bay Audubon. Additionally, the authors thank CAMEOS principal investigators Susan Williams and Vic Chow; along with Michelle Chow, Ocean Discovery!, and the participating teachers and students.

ReferencesCastle, L., Mercer, A.J., Startin, J.R. & Gilbert, J. (1988). Migration from plas-

ticized films into foods 3. Migration of phthalate, sebacate, citrate and phosphate esters from films used for retail food packaging. Food Addi-tives and Contaminants, 5, 9–20.

Derraik, J.G.B. (2002). The pollution of the marine environment by plastic debris: a review. Marine Pollution Bulletin, 44, 842–852.

Goldberg, E.D. (1994). Diamonds and plastics are forever? Marine Pollution Bulletin, 28 , 466.

Gray, M.A. & Metcalfe, C.D. (1997). Induction of testis-ova in Japanese medaka (Oryzias latipes) exposed to p-nonylphenol. Environmental Toxicology and Chemistry, 16 , 1082–1086.

Hyrenbach, D., Nevins, H., Hester, M., Keiper, C., Webb, S. & Harvey, J. (2009). Seabirds indicate plastic pollution in the marine environment: quantifying spatial patterns and trends in Alaska. In M. Williams & E. Ammann (Eds.), Marine Debris in Alaska: Coordinating Our Efforts, pp. 57–62. Fairbanks, AK: Alaska Sea Grant College Program, University of Alaska Fairbanks.

Figure 3. Sample data sheet for Lesson 2, which shows how to record data and add comments about contaminants found in plastics collected.


Inoue, K., Kondo, S., Yoshie, Y., Yoshimura, Y., Horie, M. & Nakazawa, H. (2001). Migration of 4-nonylphenol from polyvinyl chloride food packaging films into food simulants and foods. Food Additives and Contaminants, 18 , 157–164.

López-Cervantes, J. & Paseiro-Losada, P. (2003). Determination of bisphenol A in, and its migration from, PVC stretch film used for food packaging. Food Additives and Contaminants, 20, 596–606.

Marks, K. & Howden, D. (2008). The world’s rubbish dump: a garbage tip that stretches from Hawaii to Japan. Independent, 5 February.

Mato, Y., Isobe, T., Takada, H., Kanehiro, H., Ohtake, C. & Kaminuma, T. (2001). Plastic resin pellets as a transport medium of toxic chemicals in the marine environment. Environmental Science and Technology, 35, 318–324.

Mrosovsky, N., Ryan, G.D. & James, M.C. (2009). Leatherback turtles: the menace of plastic. Marine Pollution Bulletin, 58 , 287–289.

National Geographic Society. (2006). Ocean literacy: the essential principles of ocean sciences. Available online at http://www.coexploration.org/oceanliteracy/documents/OceanLitChart.pdf.

Nevins, H., Hyrenbach, D., Keiper, C., Stock, J., Hester, M. & Harvey, J. (2005). Seabirds as indicators of plastic pollution in the North Pacific. In Plastic Debris Rivers to the Sea Conference Proceedings, 2005. Available online at http://conference.plasticdebris.org/whitepapers.shtml.

Oberdörster, E. & Cheek, A.O. (2001). Gender benders at the beach: endo-crine disruption in marine and estuarine organisms. Environmental Toxicology and Chemistry, 20, 23–36.

Ocean Project. (2009). America, the ocean, and climate change: Key findings. Available online at http://theoceanproject.org/resources/America_the_Ocean_and_Climate_Change_KeyFindings_1Jun09final.pdf.

Pierce, K.E., Harris, R.J., Larned, L.S. & Pokras, M.A. (2004). Obstruction and starvation associated with plastic ingestion in a northern gannet Morus bassanus and a greater shearwater Puffinus gravis. Marine Ornithology, 32, 187–189.

Rahman, M. & Brazel, C.S. (2004). The plasticizer market: an assessment of traditional plasticizers and research trends to meet new challenges. Progress in Polymer Science, 29 , 1223–1248.

Ryan, P.G. & Jackson, S. (1987). The lifespan of ingested plastic particles in seabirds and their effect on digestive efficiency. Marine Pollution Bulletin, 18 , 217–219.

Tawfik, M.S. & Huyghebaert, A. (1998). Polystyrene cups and containers: styrene migration. Food Additives and Contaminants, 15, 592–599.

van de Merwe, J.P., Hodge, M., Whittier, J.M., Ibrahim, K. & Lee, S.Y. (2010). Persistent organic pollutants in the green sea turtle Chelonia mydas: nesting population variation, maternal transfer, and effects on development. Marine Ecology Progress Series, 403, 269–277.

Vodicka, P., Koskinen, M., Naccarati, A., Oesch-Bartlomowicz, B., Vodickova, L., Hemminki, K. & Oesch, F. (2006). Styrene metabolism, genotoxicity, and potential carcinogenicity. Drug Metabolism Reviews, 38 , 805–853.

Wuttke, W., Jarry, H. & Seidlova-Wuttke, D. (2010). Definition, classification and mechanism of action of endocrine disrupting chemicals. Hormones, 9 , 9–15.

SUSANNE M. BRANDER ([email protected]), RACHEL E. FONTANA, TAWNY M. MATA, SARAH A. GRAVEM, ANNALIESE HETTINGER, JESSICA R. BEAN, and AMBER I. SZOBOSZLAI performed this work while they were graduate stu-dents at Bodega Marine Laboratory, University of California–Davis. BRANDER is now adjunct faculty in the Department of Biology and Marine Biology at the University of North Carolina, Wilmington, 601 S. College Road, Wilmington, NC 28403. CAROL A. KEIPER is a founding board member and researcher with Oikonos Ecosystem Knowledge, a nonprofit organization working to increase understanding of human impacts on marine ecosystems. MEGHAN E. MARRERO is an Associate Professor of Secondary Education at Mercy College in Dobbs Ferry, NY, and President of the New York State Marine Education Association.

Ada High School, Ada, OK Alcott High School for the Humanities, Chicago, ILAppleby College, Oakville, Ontario, CanadaArchbishop Curley High School, Baltimore, MDArroyo High School, San Lorenzo, CA Belleville East High School, Belleville, ILBorah High School, Boise, ID Cardinal Gibbons High School, Raleigh, NCCentral Magnet School, Murfreesboro, TNCharleston High School, Charleston, ILColonia High School, Colonia, NJ Cuyahoga Community College, Parma, OHDesert Vista High School, Phoenix, AZDouglas High School, Douglas, ALDryden High School, Dryden, Ontario, Canada East Chapel Hill High School, Chapel Hill, NC Edison State College, Naples, FL Freedom High School, Freedom, WIGeorge Mason High School, Falls Church, VAGrace King High School, Metairie, LAGreat Plains High School, Watertown, SDGreensburg Salem High School, Greensburg, PAGreenville Technical College, Greenville, SC Helena High School, Helena, MT Hillcrest High School, Dallas, TXHinsdale Central High School, Hinsdale, IL Incarnate Word Academy, Houston, TXIncarnate Word Academy, St. Louis, MO International School of Minnesota, Eden Prairie, MNIowa City West High, Iowa City, IA John F. Kennedy High School, Cedar Rapids, IA KC Distance Learning, Bloomsburg, PA Kent County High School, Worton, MDLake Highland Preparatory School, Orlando, FL

Laurens District 55 High School, West Laurens, SC Lincoln High School, Esko, MNMartin Luther King Science Magnet, Nashville, TN Marysville High School, Marysville, KSMidland Park High School, Midland Park, NJMinnetonka High School, Minnetonka, MNMonte Vista Christian High School, Watsonville, CA Mount Saint Mary Academy, Watchung, NJNassau Community College, Garden City, NYNathan Hale-Ray High School, Moodus CTNaugatuck Valley Community College, Waterbury , CTNewport High School, Bellevue, WANorthridge High School, Middlebury, IN Osburn Park High School, Manassas, VA Plainwell High School, Plainwell, MIPike County High School, Zebulon, GAPikeview High School, Princeton, WV Prince of Peace Schools, Clinton, IA Rickover Naval Academy, Chicago, ILSalesianum School, Wilmington, DESaltsburg High School, Saltsburg, PA Skyline High School, Sammamish, WA Southern Wells High School, Poneto, INSycamore High School, Cincinnati, OHTiffin Columbian High School, Tiffin, OHTrinity High School, Trinity, NCUnionville High School, Kennett Square, PA University Laboratory School-LSU, Baton Rouge, LA University of Chicago Laboratory Schools, Chicago, ILVincennes University, Vincennes, IN Ware Shoals High School, Ware Shoals, SC Whiting High School, Laramie, WY Wooster High School, Reno, NV York Community High School, Elmhurst, IL

Sponsored by The mission of the NABT BioClub is to recruit, support, nurture, and promote students who have an interest in biological sciences for personal reasons, academic preparation, the bet-terment of society, and possible career opportunities by provid-ing guidance, resources, and activities to meet these goals.

Look for the BioClub logo to indicate recommended articles for NABT BioClub members. If you are interested in forming a chapter of the NABT BioClub, contact George Sellers at [email protected].

M E M B E R S

1006 shifts in intertidal zonation and refuge use by prey after mass mortalities of two predators...

Documents