studies on the ecology and population biology of...

STUDIES ON THE ECOLOGY AND POPULATION BIOLOGY OF LITTLEKNOWN ECUADORIAN ANOLES

KENNETH I. MIYATA1,2,3

CONTENTS

Abstract ------- -- - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - 45I. Ecological Interactions in Three Species of

Ecuadorian Anolis LizardsIntroduction -------- - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - 46Study Sites ------- - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - 47Results ------- - - -- - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - 50

Thermal Biology ------- - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - 50Structural Habitat ------- - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - 50Daily Activity Cycles ------- - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - 54Population Densities ------- - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - 54

Discussion -------- - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - 57Ecological Segregation among the Anoles

of the Rıo Palenque Region -------- - - - - -- - - - - - - - - - 57Co-Adapted Complexes and Recent

Interactions ------- - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - 57Generalizing from Limited Samples ------- - - - - - 58

II. Growth and Population Structure of TwoSpecies of Anolis from the Pacific Slopesand Lowlands of Ecuador

Introduction -------- - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 61Study Sites ------- - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 62Methods ------- - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 62Results ------- - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 65

General Observations and Natural History --- 65Population Densities ------- - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - 66Size and Growth -------- - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - 68Reproduction -------- - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - 71

Discussion -------- - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 72Literature Cited -------- - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - -- - - - - - - - 75

ABSTRACT. Little is known about the ecology andnatural history of South American anoles. This studyreports the results of a variety of different studies onseveral relatively common species of EcuadorianAnolis. In part I, habitat use and population densityare compared among three species of Anolis that occurin sympatry at a number of sites in Ecuador. The threespecies—A. chloris, A. festae, and A. peraccae—areroughly the same body size. These species perchprimarily on tree trunks, and A. chloris perchessubstantially higher than the other two species, whichare similar in perch height. Large differences from oneyear to the next were observed both in mean perchheight and in population densities.

In Part II, natural history, growth rates, andpopulation densities are reported for two little knownAnolis species, A. bitectus and A. gemmosus. Althoughthe two species are from nearby regions and are similarin microhabitat use, they show more differences thansimilarities in most aspects of their biology. The specieshave similar ranges in active body temperatures, but A.bitectus is thermally passive, whereas A. gemmosusappears to thermoregulate. Populations of A. gemmo-sus tend to remain constant through time, whereas A.bitectus undergoes moderate population fluctuations.Both species exhibit little sexual size dimorphism, butin A. bitectus females are larger, and in A. gemmosusmales are larger. Anolis bitectus has a fairly highcharacteristic growth rate, whereas that of A. gemmosusis quite low.

3 Remembrances of Ken Miyata by B Wu and EricLarson; Ray Huey; Greg Mayer; and Jerry Coyne areavailable as online supplementary material.

1 Department of Biology, Harvard University, Cam-bridge, Massachusetts 02138.

2 Ken Miyata died tragically in 1983. This paperincludes two parts of Chapter 2 of his four-chapter,787-page doctoral dissertation, ‘‘Patterns of Diversity inTropical Herpetofaunas,’’ which was completed in1980. Because the premise of the paper, that little isknown about the natural history of South Americananoles, is almost as true now as it was 30 years ago, theeditors of the Bulletin decided that it was worthwhile tomake this material available to a wide audience. Thesetwo sections of the dissertation have been lightly editedto transform it from a doctoral dissertation. A numberof footnotes in the text (but not those in the tables)have been added by the editor, with the help of severalreviewers, to update information. The abstract was alsowritten by the editor because these sections had noabstract in the dissertation. The editors thank MelissaAja and Emily Becker for helping adapt and preparethe manuscript and B. Bock, O. Torres-Carvajal, and L.Vitt for reading and commenting upon a draft of thismanuscript.

Bull. Mus. Comp. Zool., 161(2): 45–78, July, 2013 45

Key words: Anolis chloris, Anolis peraccae, Anolisfestae, Anolis, Ecuador, natural history, habitat use,population density, growth rate, Anolis bitectus, Anolisgemmosus

I. ECOLOGICAL INTERACTIONS INTHREE SPECIES OF ECUADORIAN

ANOLIS LIZARDS

INTRODUCTION

The iguanid lizards of the genus Anolisform a large and complex assemblage ofspecies within three major evolutionarytheatres: the West Indies, Central America,and South America. From the West Indieswe find a large body of work on manyaspects of the biology of the genus (seeWilliams, 1969, 1972, 1976a; Schoener,1977, for summaries of some of this work).1

These studies have led to many refinementsin our understanding of the evolutionaryforces that have shaped this rich and diversefauna. The situation in the two mainlandregions, however, is quite different. For theCentral American anoline fauna, a fewworks deal with the ecology and systematicsof certain species and certain specific faunas(e.g., Taylor, 1956; Fitch, 1972, 1973, 1975;Fitch et a1., 1976),2 but general overviewshave been lacking. Andrews (1976, 1979)has made some attempts to compare islandand mainland anoles in terms of life historystrategies, but no efforts have been made tomake comparisons of entire faunas. ForSouth America, ecological studies have beenvery limited in scope,3 although somesystematically oriented papers have madesubstantial contributions to understandingof major evolutionary patterns (Vanzolini

and Williams, 1970).4 The emphasis ofSouth American anole studies on systematicproblems has two underlying causes: theSouth American anole fauna is very complexand confusion abounds in trying to elucidaterelationships and affinities of the varioustaxa because of extensive convergence andparallelism between and within lineages(Williams, 1976b),5 and South Americananoles are characteristically uncommon inthe field, which discourages most fieldstudies.

The little information available on theecology of South American anoles has beenlargely anecdotal in nature, consisting most-ly of brief comments made by variouscollectors. There have been no detailedpopulation studies of any species of SouthAmerican anoles, although some valuableinformation has been provided in generalfaunal studies of specific regions (Dixon andSoini, 1975; Duellman, 1978).6 A few papersprovide more than anecdotal information onspecific species (Rand and Humphrey,1968; Vanzolini, 1972; Fitch et al., 1976),7

but this information is usually incompleteand covers only a tiny percentage of theSouth American anole fauna. Our knowl-edge of the population ecology of any onespecies of South American anole is scanty atbest, and for most species we have noinformation at all.8

1 This and all other footnotes in the main text havebeen added by the editor: A summary of anole researchconducted since Miyata wrote his thesis can be foundin Losos (2009).

2 More recent studies include Guyer (1986); Andrewsand Nichols (1990); Andrews (1991); Andrews andWright (1994); Parmelee and Guyer (1995); Vitt et al.(1995, 2001, 2003a,b, 2008); Vitt and Zani (2005);Schaepfer (2006).

3 See Avila-Pires (1995) and Vitt et al. (2008) forreferences to more recent publications on SouthAmerican anoles.

4 For a revised view of the situation with the A.chrysolepis species complex, see D’Angiolella et al.(2011).

5 The phylogenetic relationships of South Americananoles are slowly coming into focus Poe, (2004);Castaneda and de Queiroz, (2011, 2013); D’Angiolellaet al., (2011).

6 Avila-Peres (1995); Vitt and Zani (1996); Vitt et al.(1999); Mesquita et al. (2006), (2007); Wernick et al.(2009).

7 Vitt and Zani (1996, 2005); Vitt et al. (2001, 2003a,2003b, 2008); Molina and Gutierrez (2007); Velascoand Herrel (2007); Bock, et al. (2010).

8 Citations in previous footnotes indicate that thesituation has improved in the last 30 years, although westill know little or nothing about most South Americananole species.

46 Bulletin of the Museum of Comparative Zoology, Vol. 161, No. 2

The present paper provides a look intosome of the interactions between severalspecies of Anolis from the Pacific lowlandsof northwestern Ecuador. Although detailedinformation on the population ecology ofthe three major species treated are notavailable, certain patterns and trends can befollowed, and it is possible to make someinferences regarding the way these speciesinteract with each other and the rest of thefauna of the region.

Because the three principal species ofAnolis studied are poorly known, I willbriefly characterize them.

Anolis chloris Boulenger: A small tomedium-sized anole (maximum malesnout-vent length [SVL] ca. 60 mm, femalesca. 58 mm SVL). This species is uniformgreen in color. It ranges through thesouthern Darien Province in Panama andChoco region to northwestern Ecuador.

Anolis festae Peracca: This is a smallspecies of anole, reaching a maximum SVLof 55 mm in the study region. It is endemicto the lowlands of southwestern Ecuador.Fitch et al. (1976) referred to this species asAnolis nigrolineatus.

Anolis peraccae Boulenger: This is anoth-er small species, reaching a maximum SVLof 54 mm. It is found in the lowland forestsof southwestern Colombia and northwest-ern Ecuador.

STUDY SITES

The study sites visited in the course of thefield work were all located within 25 km ofeach other in the lowlands of southernProvincia Pichincha, Ecuador. This arearanges in elevation from about 125 to225 m above sea level. The topography ismostly flat, but with some steep bluffsdropping down into river valleys and lowrolling hills. The entire region was oncecovered by rain forest and was transitionalbetween the tropical moist forest andtropical wet forest formations of Holdridge(1967). Most of this forest has now beencleared, and the area is now rich agriculturalland for the most part. The mean annual

precipitation in the area varies from about3,000 mm in the north to 2,000 mm in thesouth (Dodson and Gentry, 1978), with adistinct dry season running from June toDecember. During the dry season theregion has an almost constant cloud cover,and the humidity remains rather high. Lightmorning rains or drizzles are an almost dailyoccurrence during the so-called dry season,although heavy rains do not occur veryoften. Most of the native forest trees do notdrop their leaves during the dry season, andconditions probably do not often becomesevere.

The primary study site was the CentroCientıfico Rıo Palenque, located 47 kmsouth of Santo Domingo de los Coloradoson the road to Quevedo at an elevation of220 m. A map and general discussion of theclimate and vegetation of the area ispresented in Dodson and Gentry (1978).Although scattered observations were madein the forest there, the major study site wasalong a row of balsa trees (Ochromapyramidale) on both sides of PerimeterRoad between Three Corners Road andDodson Road (Fig. 1). This stretch of roadwas about 900 m long and was bordered onone side by rain forest and on the other byplantations of oil palm (Elaeis guineensis)and banana (Musa paradisiaca) with somesmall patches of secondary forest. The balsatrees were not planted but germinatedshortly after the road was constructed in1971. These trees grew very rapidly, and atthe time the study began in 1974, most wereabout 20 m in height. At the beginning ofthe study they were spaced in a fairlyuniform manner.

This site was first visited in June 1974, atwhich point the high population density ofA. peraccae was noted and preliminarystudies begun. At this time the balsasranged from 15 to 25 cm diameter at breastheight (dbh) and were between 15 and 20 min height. The first planting of oil palm hadbeen made about a month before the firstvisit, and the bananas had been cut down toclear the way for the new crop. The cutbananas sprouted immediately, and by the

ECUADORIAN ANOLES N Miyata 47

end of August 1974, scarcely 2 months afterthey had been cleared, they ranged inheight up to 5 m. This mixture of bananasand oil palm continued through 1975, butby 1976 the bananas were kept cut back topromote growth in the palms.

At the same time, the balsas had contin-ued to grow until most were more than 20 min height and 25 cm dbh, and many began tobreak of their own accord from storms andheavy rains. Certain sections of the studyarea were consequently opened by thesenatural falls, and trees in parts of the sitewere cut and the wood sold. The entirestudy area was affected by this thinning, butthe crowns of the balsas remained more orless continuous. By 1977 nearly all of thebalsas had been removed, and no furtherquantitative studies were possible.

In 1975 a series of isolated tree rowsalong the Santo Domingo de los Colorados-Quevedo road were examined for anoles.Dense populations of anoles were found atseveral sites, and these were monitoredbetween the months of June and Septemberof that year (Fig. 2). These sites usuallyconsisted of single rows of large trees,generally teak (Tectona grandis) or guaya-can (Tebebuia sp.), bordered on one side bythe road and on the other by open fields orbanana plantations. The bases of the treesgenerally had a thick layer of rather dry leaflitter, and the only vegetation was grass. Allof these study sites were destroyed byconstruction work between September1975 and June 1976 when the road waswidened.

Other study sites included a plantation ofrubber (Hevea brasiliensis) at HaciendaCerro Chico and old cacao groves (Theo-broma cacao) near Buena Fe and just east ofthe city of Quevedo. Observations in thesesites were carried out between 1975 and1979 at irregular intervals. Additional ob-servations were made at Tinalandia, a hotellocated some 16 km southeast of SantoDomingo de los Colorados on the road toQuito at an elevation of 700 m.

Much of the information summarized inthis report consists of scattered observationsof lizards made at the above study sites,along with a few observations from else-where. Because of the nature of the anolepopulations, many of these observationswere perforce opportunistic. It should alsobe emphasized that nearly all of theseobservations were made in disturbed habi-tats, for reasons which will be discussedlater.

The daily censuses performed at thePerimeter Road study site at the CentroCientıfico Rıo Palenque consisted of slowlywalking down the road and noting thefollowing for each lizard seen: species, sex(if determinable—this was often not possi-ble with A. chloris), height above groundand diameter of perch at first sighting, andexposure to sun and orientation with respectto the ground. In the censuses performed

Figure 1. Perimeter road at the Centro Cientıfico RıoPalenque balsa study site. The view includes section borderedby forest on left with open banana and oil palm to the right.Photo taken July 1975.


after August 1975, the location of eachlizard seen within the study site was noted,and the study strip was divided into 10-m-long segments. In 1974 and 1975 individualtrees in the section of the study site nearestThree Corners were marked, and individuallizard sightings were associated with partic-ular trees.

An attempt was made to mark andmeasure as many lizards as possible toobtain information on growth and move-ment and spacing patterns. This was latercut back because of several compellingreasons involving increased wariness ofmarked individuals (see below) and thevirtual lack of juveniles, which are necessaryto obtain reliable estimates of growth rates(Andrews, 1976, 1979; Schoener andSchoener, 1978). Nevertheless, a total of65 A. chloris and 88 A. peraccae werecaptured and marked at the Centro Cientı-fico Rıo Palenque. For each animal cap-tured, the snout-vent length (SVL) wasnoted, the reproductive state of adultfemales (obviously gravid or not) wasobserved, and an individual toe-clip markwas made. An initial attempt was made to

color mark each animal individually, butthese marks rarely lasted more than a day ortwo because of wet climatic conditions.Whenever possible, cloacal temperatures(BT) and air temperatures (AT—shadedbulb ca. 1 cm above perch at time of firstsighting) were measured using a Schultheisrapid-registering thermometer.

A similar protocol was followed in all ofthe study areas, except individual positionswere not recorded, and no attempt wasmade to mark the animals individually.

In several places individual behavior wasobserved for periods ranging up to one hourat a time. Most such observations weremade on A. chloris, and none were made onA. festae.

In short, except for the Perimeter Roadbalsa site census area at the Centro Cientı-fico Rıo Palenque, my sampling was oppor-tunistic. These observations did provideconsiderable insight into the biology of theanimals which could have been obtained inno other way. The ephemeral nature of thecharacteristic habitats of the lizards was thejustification for these observational tech-niques, which did prove to be quite useful.

Figure 2. Roadside teak plantation at 41 km S of Santo Domingo de Los Colorados, on road to Quevedo. Photo taken July 1975.


Analysis of perch height data was per-formed using log transformations of perchheights, which were always estimated ininches, to normalize the distributions. Meanswere then transformed back into linearmeasurements. Perch dimensions in the fieldwere always estimated in inches because Iam more familiar with these units and foundit much easier to make estimates. These unitswere retained in the analyses becauseconversions to metric units would haveadded nothing except apparent precision.

RESULTS

Thermal Biology

All three species of Anolis studied werethermally passive (Figs. 3–5). Anolis chloriswas the only one of the three that was everobserved basking, and this basking wasalways observed early in the morning orlate in the afternoon when the lizards were

perched high in the trees and I was unableto sample their body temperature.

The mean body temperatures of all threespecies were similar (Table 1), reflectingtheir thermal passivity and similarity ofhabitat preferences. This similarity heldeven where the species were found in strictsympatry, rather than basing the compari-son on total samples (Tables 2, 3).

Structural Habitat

Ecological studies of anole communitiesin the West Indies have shown the impor-tance of differences in structural habitat inthe reduction of possible niche overlap(Rand, 1964; Schoener, 1968; Schoenerand Gorman, 1968; Rand and Williams,1969; Schoener and Schoener, 1971a,b).Structural habitat for anoles has customarilyreferred to both perch height and perchdiameter, with sympatric species normallysegregating along one or both spatial axes.

Figure 3. Body temperatures of Anolis chloris plotted against air temperature at time of capture. Open circles present multiplerecords. Diagonal line is for body temperature 5 air temperature.


All three species in the present study appearto be obligate tree-trunk perchers (Table 4),and for all practical purposes, perch diam-eter can be ignored. Although a largenumber of A. chloris were perched onsmall-diameter perches, which suggests apossible difference in perch distributionwith respect to the other two species, inevery case, the small-diameter perches wereover 25 ft (7.6 m) above the ground, andtheir presence on small-diameter perches isa function of their perch height.

Mean perch heights for the total data setsof each species (all localities, all times, bothsexes) are presented in Table 5. The threespecies differ significantly from each otherin mean perch height (p y 0.001). Anolischloris clearly chooses perches muchhigher than either A. festae or A. peraccae,and A. peraccae perches slightly higherthan A. festae. Although these data appearto suggest that the three species aresegregated by differences in perch height,

the real pattern is obscured in this pooleddata.

The slight differences in mean perchheight between A. festae and A. peraccaedissolve if the situation is examined morecarefully. Table 6 presents perch heightdata for these two species at two differentlocalities where they occurred syntopically.At the more northern of these localities,41 km south of Santo Domingo de losColorados, there is no significant differencein mean perch height between the twospecies (p 5 0.68). Both species typicallyperch very near the ground. On the otherhand, at Hacienda Cerro Chico there areslight but marginally significant differencesin mean perch height between the twospecies, with A. festae perched higher thanA. peraccae (0.01 , p , 0.05). This isexactly the opposite of what the pooled datashow. The slight differences in mean perchheight between these two species is unlikelyto be of much biological significance despite

Figure 4. Body temperatures of Anolis festae plotted against air temperature at time of capture. Open circles represent multiplerecords. Diagonal line is for body temperature 5 air temperature.


their statistically significant differences incentral tendency.

Perch heights are subject to daily cyclesof change. Figure 6 shows hourly shifts inperch height distribution of A. peraccae atthe Rıo Palenque balsa site. There is a clearupward shift in perch height through theday. Both sexes, however, do not shiftupward at the same time. Females showno difference in mean hourly perch heightbetween 7:00 a.m. and 3:59 p.m. (ANOVA,F8,261 5 1.88, p 5 0.064), whereas malesshow a steady increase in hourly meanperch height (Table 7). In late afternoon

both sexes are typically found on highperches, presumably as a preliminary move-ment to their normal sleeping perches in thetree crowns.

Because A. chloris is more difficult to sexin the field without capture, my data onperch height shifts are less extensive thanfor A. peraccae. Observations were pooledinto bi-hourly samples for analysis, and notall individuals could be sexed (Table 8).Females showed no shift in perch heightsthrough the entire day, whereas malesshowed a strong tendency to perch higheras the day went on. However, because it iseasier to make a positive sex identification ofmales at long distances (particularly whenthey display, which they do often), thiscould bias the data. It does appear likely inany case that males perch higher thanfemales regardless of the time of day.

Because of the much smaller total samplesize, I was unable to analyze the perchheight data for A. festae in terms of daily

Figure 5. Body temperatures of Anolis peraccae plotted against air temperature at time of capture. Open circles representmultiple records. Diagonal line is for body temperature 5 air temperature.

TABLE 1. MEAN BODY TEMPERATURES OF POOLED SAMPLES OF THREE

SPECIES OF ANOLIS FROM WESTERN ECUADOR.*

n MBT (u C) s2

A. chloris 58 24.8 2.36A. festae 28 24.3 0.85A. peraccae 87 24.7 2.28

* F2,170 5 1.42, p 5 0.25.


shifts in perch height. As regards to sexualdifferences in perch height, the answer isunclear. For all of the observations pooled,males perch significantly higher than females(27.7 in [5704 mm] vs. 12.6 in [5320 mm]).However, if only the best-sampled locality isexamined, mean perch height does not differbetween sexes (Table 9).

In addition to these sexual differences inmean perch height and daily shifts in perchheight distribution, species can also exhibitmarked shifts in mean perch heights fromyear to year (Table 10). At the Rıo Palenquebalsa site, both A. chloris and A. peraccaewere perched much higher in 1975 and1976 than in 1974. This was not due tosampling bias because the field protocol wasnot changed, and there is no reason tosuspect that I was any better at spottinglizards perched higher in the trees in thelatter 2 years.

The Rıo Palenque balsa site was dividedinto subsections about midway through the1975 field season, and from that point on, allanimals seen during the censuses wereassociated with a 10-m-long interval of theroadside. Two definite habitat types wereassociated with the balsas along this strip.The first section consisted of trees that wereisolated from forest edge and secondarygrowth by a buffer zone of oil palms, andthe second section consisted of trees thatabutted either forest, secondary growth, or

both (Fig. 1). Table 11 summarizes theperch heights of A. chloris and A. peraccaefrom 1975 and 1976 in these two areas. Thepatterns that emerge are rather peculiar.Anolis chloris in both sections perched atthe same mean height above the ground in1975, but in 1976 the animals in Section 2perched higher than those in Section 1.Anolis peraccae perched at the same meanheight in both sections in each of the 2 years,but showed significant differences in meanperch height from year to year, perchingsomewhat higher in 1976 than in 1975.

The orientation of individual lizards withrespect to the ground may be of possiblesignificance when it comes to assessing howthe lizards utilize their perches. Classical sit-and-wait predation in anoles typically in-volves a lizard perching with its headpointed downward, which allows it to scanthe ground and vegetation below for prey. Ifthe lizards are primarily searching activelyfor their prey, one might expect them to beoriented differently with respect to theground, particularly if they are searchingfor prey on the trunk of a tree or in the

TABLE 2. BODY TEMPERATURE OF ANOLIS FESTAE AND ANOLIS PERACCAE AT TWO LOWLAND LOCALITIES IN WESTERN ECUADOR.

A. festae A. peraccae

pn MBT (u C) s2 n MBT (u C) s2

41 km S of Santo Domingode los Colorados 15 23.8 0.27 4 24.0 0.09 0.61

Hacienda Cerro Chico 5 24.4 1.04 7 24.8 0.52 0.52

TABLE 3. MEAN BODY TEMPERATURES OF ANOLIS CHLORIS AND

ANOLIS PERACCAE IN SYMPATRY AT CENTRO CIENTıFICO RıO PALENQUE.*

n MBT (u C) s2

A. chloris 55 24.9 2.31A. peraccae 54 25.0 2.86

* F1,107 5 0.22, p 5 0.64.

TABLE 4. PERCH OCCURRENCE OF THREE SPECIES OF ANOLIS IN

WESTERN ECUADOR (% TOTAL OBSERVATIONS).*

A. chloris A. festae A. peraccae

Ground 0.7 1.2 0.6Tree trunks1 89.4 95.2 97.0Branches2 9.2 3.6 1.4Leaves 0.7 0 0.5Banana stems 0 0 0.4Lianas 0 0 0.1n 433 83 1,1441 More than 7.5 cm in diameter, vertical or near vertical.2 Less than 7.5 cm in diameter, vertical or horizontal (includes small treetrunks as well).

* x2 5 93.18, p y 0.001; x2 5 2.92, p 5 0.232 for comparison between A.festae and A. peraccae.


vegetation. Table 12 shows the orientationat first sighting for the three species; thepercentage of lizards facing the ground isnot significantly different between the threespecies.

Daily Activity Cycles

Activity cycles in arboreal lizards aredifficult to estimate because there is alwayssome question as to whether their apparentabsence at a given time is due to lack ofactivity or a shift of activity to an area thatcannot be censused by a ground-basedobserver. Keeping this potential complica-tion in mind, the activity cycles of A. chlorisand A. peraccae at the Rıo Palenque balsasite are presented in Figure 7. Data wereinsufficient to make similar estimates ofactivity patterns in A. festae.

Even though these data on activity cyclesmight represent shifts in activity, ratherthan actual changes in activity, some inter-esting patterns emerge. Anolis chloris is notusually seen until after A. peraccae hasbegun its daily activity. Both species reachan apparent activity peak during the mid-morning hours. Anolis chloris seems to showa definite decrease in activity shortly afternoon, with a slight afternoon peak severalhours later; this pattern was seen in A.peraccae in 1976 but not in 1975. Bothspecies become considerably less conspicu-ous in late afternoon, coinciding with theirstrong upward shift in perch height.

Population Densities

It was not possible to make estimates ofpopulation size with conventional mark-

recapture techniques because the centralassumptions of such methods were not met.All of my comments on density are there-fore based on a crude index of populationdensity—the number of individuals seenper census on the Rıo Palenque balsa site.Although the total number of visible lizardsin any given census could vary considerablydepending on activity patterns of the lizards,these variations should be dampened out inthe course of extensive sampling spreadover a several month interval. In any case,the sampling biases from year to year shouldbe comparable, and the resulting data areprobably adequate to detect major shifts inabundance from year to year.

Because of the variation between differ-ent study sites and habitats, all my compar-isons of population density are restricted tothe populations of A. chloris and A.peraccae at the Rıo Palenque balsa site.These observations spanned the months ofJune to August from 1974 to 1976. The datafrom 1974 were taken only from Section 1 ofthe site. This site was extended in 1975, andfrom about halfway through the samplingperiod, note was taken of which sectioneach individual sighting was made. Thestudy was terminated in 1976 because thebalsa trees in the study area were removedprior to my next visit in April 1977.

The mean number of observations of A.chloris and A. peraccae per census in theentire study area are presented in Table 13.These data suggest that populations of A.chloris remained more or less similar duringthese years, whereas A. peraccae underwenta drastic population decline between 1975

TABLE 5. MEAN PERCH HEIGHTS (IN INCHES) OF THREE SPECIES OF

TRUNK-PERCHING ANOLIS IN WESTERN ECUADOR. LOCALITIES, DATES,

TIMES, AND SEXES LUMPED. ALL MEANS SIGNIFICANTLY DIFFERENT

(P y 0.001).

n Mean Height1

A. chloris 426 120 6 1.1A. peraccae 1,131 26 6 1.0A. festae 79 20 6 1.31 695% confidence limit.

TABLE 6. MEAN PERCH HEIGHTS (IN INCHES) OF ANOLIS FESTAE AND

ANOLIS PERACCAE AT TWO LOWLAND LOCALITIES IN WESTERN ECUADOR.

A. festae A. peraccae

pnMeanHeight n

MeanHeight

41 km S of SantoDomingo de losColorados 40 11.9 8 13.9 0.681

Hacienda Cerro Chico 8 25 38 15.3 0.039


Figure 6. Hourly shifts in perch height distributions of female and male Anolis peraccae at Rıo Palenque balsa study site in 1975and 1976. Sample sizes in parentheses.


and 1976. Although there are no firm data,my observations in the forest at RıoPalenque and elsewhere in the vicinityseemed to affirm this pattern: Anolisperaccae was much more difficult to findanywhere in the region after 1975.

If the two sections in the Rıo Palenquebalsa site are examined separately, severalinteresting trends emerge (Tables 14, 15).In Section 1, A. chloris seemed to haveundergone a decline before the 1975sample, from which it subsequently recov-ered before the 1976 sample. Althoughstatistically significant, this apparent declinein A. chloris might have been an artifact ofthe sampling procedure and is of question-able validity. On the other hand, the

differences seen in A. peraccae (Table 15,top) are undoubtedly real and seem toconstitute a strong downward trend inabundance over the 3 years of the study.This downward trend in populations of A.peraccae was apparently general throughoutthe region; it was considerably more diffi-cult to locate A. peraccae in January 1979than it was to find them in 1974 throughoutthe region.

The downward trend in A. peraccaepopulations was even more pronounced inSection 2 of the Rıo Palenque balsa site.Population densities appeared to be aboutan order of magnitude less in 1976 than in1975, which accounted for the shift inrelative abundances of the two species inthis section (Table 16).

TABLE 7. HOURLY SHIFTS IN MEAN PERCH HEIGHTS (IN INCHES) OF ANOLIS PERACCAE AT RIO PALENQUE BALSA STUDY SITE. DATA ARE FROM 1975

AND 1976 ONLY.

Time

Total1 Females Males

n Mean Height n Mean Height n Mean Height

6:00–6:59 a.m. 5 17.6 — — — —7:00–7:59 a.m. 21 25.8 8 26.6 13 25.38:00–8:59 a.m. 54 27.6 26 18.5 28 39.99:00–9:59 a.m. 87 21.9 41 15.4 46 30.210:00–10:59 a.m. 134 25.1 58 21.2 76 28.611:00–11:59 a.m. 75 40.6 37 29.3 37 54.312:00–12:59 p.m. 39 40.7 15 22.5 24 58.91:00–1:59 p.m. 56 41.4 25 33.1 31 53.32:00–2:59 p.m. 58 39.0 26 24.4 32 57.13:00–3:59 p.m. 79 48.7 35 28.5 43 72.34:00–4:59 p.m. 35 57.8 18 51.5 17 71.15:00–5:59 p.m. 15 86.3 7 68.0 8 106.41 Not all lizards could be positively sexed.

TABLE 8. BI-HOURLY SHIFTS IN MEAN PERCH HEIGHTS (IN INCHES) OF

ANOLIS CHLORIS AT RIO PALENQUE BALSA STUDY SITE. DATA ARE FROM

1975 AND 1976 ONLY.

Time

Females* Males**

nMeanHeight n

MeanHeight

6:00–7:59 a.m. 1 12.0 2 509.18:00–9:59 a.m. 8 59.2 21 125.810:00–11:59 a.m. 19 55.4 55 195.212:00–1:59 p.m. 5 42.0 17 214.62:00–3:59 p.m. 3 160.7 29 279.54:00–5:59 p.m. 2 93.0 16 320.4

* F5,32 5 0.63, p 5 0.680.

** F5,134 5 3.43, p 5 0.006.

TABLE 9. SEXUAL DIFFERENCES IN MEAN PERCH HEIGHTS (IN INCHES)

OF ANOLIS FESTAE IN WESTERN ECUADOR.

n Mean Height

All localities lumped*

Females 30 12.6Males 25 27.5

41 km S of Santo Domingo de los Colorados**

Females 19 9.2Males 8 9.5

* F1,53 5 398.45, p y 0.001.

** F1,25 5 0.009, p 5 0.93.


DISCUSSION

Ecological Segregation among the Anolesof the Rıo Palenque Region

The three species of Anolis studied werethe most conspicuous members of the anolefauna of the region.9 All three species weremost abundant in secondary habitats, andone, A. festae, was restricted to disturbedhabitats in the Rıo Palenque region. At firstconsideration the data on perch height seemto provide a suitable approximation of howthe three species might partition resources.By perching at different heights above theground, the three species may be able toscan different areas of the ground belowand select a different range of prey items.However, a closer examination of theoccurrence of A. festae and A. peraccaeindicates that the rather slight, but statisti-cally significant, differences in mean perchheight could be of no apparent biologicalsignificance.

In sympatry, one or the other of the twospecies is always considerably more abun-dant. In the few places where both speciesoccurred in sufficient number to obtain dataon perch dimensions and body tempera-tures, there was almost complete overlapbetween the two species. Both species are

practically identical in size, and a prelimi-nary examination of the stomach contents ofthe two species reveals no obvious differ-ences in the range of prey taxa or sizestaken. The two species appear to be verysimilar to one another and coexist along arather narrow zone where populations maybe in rather tenuous balance.

Although A. festae was seen on the RıoPalenque balsa site only in 1976, after A.peraccae had undergone a severe decline inpopulation density, only one or two A. festaewere seen. With the removal of the balsatrees after the 1976 field season, it was notpossible to see what this situation might havedeveloped into. It seems clear, however, thatcompetition with A. festae did not cause thedecline in A. peraccae populations.

Competition between A. festae and A.peraccae would appear to be quite severe,given their general similarity in size andmorphology and habitat preferences. Thesetwo species are quite different from A.chloris, which clearly segregates itself eco-logically from the other two species by itsutilization of high perches.

Co-Adapted Complexes andRecent Interactions

One clue to the pattern seen betweenthese three species lies in the historicalpatterns of distribution and the effect ofrecent human disturbance in the region.Both A. chloris and A. peraccae areelements of the Chocoan herpetofauna; A.chloris is found as far north as the DarienProvince in Panama, and A. peraccae (alongwith a vicariant species in a superspeciescomplex) is widespread in the Choco northof central Ecuador. Anolis festae, on theother hand, is a member of a peculiarsouthern Ecuador fauna that might beassociated with the more seasonal and xericlowlands of southwestern Ecuador. Theentire region between Quevedo and SantoDomingo de los Colorados has recentlyundergone a burst of intensive agriculturaldevelopment (Nelson, 1973; Dodson andGentry, 1978), which could have openedpathways through which this southern fauna

TABLE 10. SHIFTS IN MEAN PERCH HEIGHTS (IN INCHES) BETWEEN

1974 AND 1976 IN ANOLIS CHLORIS AND ANOLIS PERACCAE AT CENTRO

CIENTIFICO RIO PALENQUE.

1974 1975 1976

nMeanHeight n

MeanHeight n

MeanHeight

A. chloris 48 35 160 131 160 171A. peraccae 294 15 539 33 123 41

9 Eight species of Anolis have been found on recentvisits to the Rıo Palenque station (A. fasciatus, A. festae,A. gracilipes, A. lyra, A. maculiventris, A.parvauritus, A.peraccae, and A. princeps), although, interestingly, A.chloris has not been found there. Two other species—A. lynchi and A. parilis—likely occur at the station ornearby. Poe et al. have put together a pictorial guide tothese species, available at http://www.oeb.harvard.edu/faculty/losos/jblosos/pdfs/PoeRioPalenque2012.0.pdf.


might have been able to invade what wasformerly a wet Chocoan forest fauna. Theinteractions between A. festae and A.peraccae could thus be of very recent origin,at least in the area encompassed by thestudy, and the two species might not yethave had sufficient time or opportunity tobecome more finely tuned to each otherecologically. The interactions between A.chloris and A. peraccae, on the other hand,are of long duration, and the clear ecologicaldifferences between the two could partiallyreflect this opportunity for ecological inte-gration and extensive sympatry.

Generalizing from Limited Samples

Although the data presented here arerather limited, they do cover a fairly longtime span, and strong differences betweenyears and study sites were noted in severalaspects of the biology of the three species ofAnolis. The popular conception of the‘‘unchanging tropics’’ is gradually erodingamong the biological public, but this con-cept finds its way into the general publicwith disturbing frequency. The large fluc-tuations in population density observed inA. peraccae might be extreme (a 10-fold

decrease in a year), but they are well in linewith similar fluctuations in Central Ameri-can anole populations (Sexton et al., 1963;Sexton, 1967; Andrews and Rand, 1982).

Life history information is clearly neededfor South American anoles. The only suchinformation currently available are for twospecies of Ecuadorian anoles, A. bitectusfrom the lowlands and A. gemmosus fromthe cloud forests. I was unable to obtain anyuseful data on growth or longevity of any ofthe three species in the current study,despite the fact that almost 150 individuallizards were marked and released in thecourse of the study. All of the recapturesmade were very short term (none more than70 days), and all involved animals originallymarked as adults.

The complete absence of recapturesbetween years does suggest that the threespecies might be short-lived. This is inkeeping with current ideas concerning thelife history of mainland anoles (Andrews,1979).

The current concept of a typical mainlandanole could be severely biased by the natureof the species studied. Mainland anoles arethought to be generally limited by preda-tion, rather than competition, as in the caseof island anoles. This fundamental differ-ence in selective pressures leads to certainpredictions regarding life history strategiesthat seem to be met by most of the speciesthat have been studied to date (Andrews,1979). However, relatively few mainlandanoles have been studied in any detail, andwhat might be true for them might notnecessarily be true for the majority ofmainland species.

TABLE 11. SHIFTS IN MEAN PERCH HEIGHTS (IN INCHES) OF TWO SPECIES OF ANOLIS AT RIO PALENQUE BALSA STUDY SITE. HORIZONTAL LINES

CONNECT VALUES WHICH ARE NOT SIGNIFICANTLY DIFFERENT (P . 0.05).*

1975 1976

Section 1 Section 2 Section 1 Section 2

n Mean Height n Mean Height n Mean Height n Mean Height

A. chloris 7 73.9 50 82.2 35 80.0 125 161.7A. peraccae 94 23.1 99 25.2 100 38.1 20 57.8

* The differences in mean perch height between species are all significant (p % 0.01).

TABLE 12. ORIENTATION OF THREE SPECIES OF ANOLIS IN WESTERN

ECUADOR WHEN FIRST SIGHTED.*

Head-Up Head-Down

n % n %

A. chloris 60 27.5 158 72.5A. festae 25 32.9 51 67.1A. peraccae 239 31.0 533 69.0

* x2 5 1.19, p 5 0.55.


The relative scarcity of anoles in mainlandhabitats has been repeatedly emphasized(Vanzolini, 1972; Williams, 1976b; Andrews,1979). Andrews (1979) has published a listof mainland and island species for which

estimates of population density are avail-able. From these data, and additional datain Schoener and Schoener (1980), it appearsthat island anoles have densities typically anorder of magnitude or more higher than

Figure 7. Activity cycles of Anolis peraccae (top) and Anolis chloris (lower) at Rio Palenque balsa study site. Data for 1975indicated by dotted line; data for 1976 indicated by solid line.


mainland species. However, even thesefigures might not portray the situationaccurately. The figures presented for main-land anoles are much higher than for thetwo species of South American anolesknown (see below), which suggests thatSouth American anoles might typically bemuch less abundant than Central Americanand Mexican species. These low populationdensities make it extremely difficult tocollect data relevant to life history studies.

Another equally serious problem that isclosely related to the above phenomenon isthe nature of the mainland species that havebeen examined. In most cases, only the mostabundant species of mainland anoles havebeen studied. These species are largelyfound in high densities only in ratherdisturbed habitats, although some rather

specialized species (e.g., the aquatic A.poecilopus and the saxicolous A. gadovii)are also found in high densities in undis-turbed habitats.

Table 17 summarizes information regard-ing the relative abundances and habitatoccurrence of the anoles of Ecuador.Although the vast majority of species occurin forest, relatively few are ever found insufficient peak densities to allow ecologicalstudies, and in nearly all cases these peakdensities are only reached in disturbedhabitats. Although it seems clear that a‘‘typical’’ South American anole is aninhabitant of forest habitats, very few canbe studied in a quantitative fashion, even inhabitats that may not be typical. With this inmind it seems premature to make overly

TABLE 13. NUMBER OF OBSERVATIONS OF ANOLIS CHLORIS AND

ANOLIS PERACCAE PER CENSUS AT RIO PALENQUE BALSA STUDY

SITE, 1975–1976.

A. chloris* A. peraccae**

1975 1976 1975 1976

No. of censuses 57 48 57 48Mean no. of

observationsper census 2.84 3.42 9.35 2.60

s2 3.94 5.96 24.65 3.57

* F1,74 5 1.06, p 5 0.31.

** F1,74 5 89.88, p y 0.001.

TABLE 14. NUMBER OF OBSERVATIONS OF ANOLIS CHLORIS PER CENSUS

AT RIO PALENQUE BALSA STUDY SITE BY SECTION, 1974–1976.

1974 1975 1976

Section 1*

No. of censuses 38 18 48Mean no. of

observations per census 1.13 0.39 0.94s2 2.11 0.35 1.73

Section 2**

No. of censuses 18 48Mean no. of

observations per census 2.83 2.48s2 3.94 5.96

* All three years: F9 5 4.89 (v1 5 2, v2 5 63.8), p 5 0.011; 1974 and 1976:F1,84 5 0.41, p 5 0.52.

** F1,64 5 0.30, p 5 0.59.

TABLE 15. NUMBER OF OBSERVATIONS OF ANOLIS PERACCAE PER CENSUS

AT RIO PALENQUE BALSA STUDY SITE BY SECTION, 1974–1976.

1974 1975 1976

Section 1*

No. of censuses 38 18 48Mean no. of

observations per census 7.79 5.61 2.15s2 14.69 13.68 3.5

Section 2**

No. of censuses 18 48Mean no. of

observations per census 5.44 0.46s2 5.14 0.46

* All three years: F9 5 3558.75 (v1 5 2, v2 547.0), p y 0.0001; 1974 and1975: F1,54 5 3.89, p 5 0.054.

** F ’s 5 75.13 (v1 5 1, v2 5 22.7), p y 0.0001.

TABLE 16. NUMBER OF OBSERVATIONS OF ANOLIS CHLORIS AND ANOLIS

PERACCAE IN TWO SECTIONS OF RIO PALENQUE BALSA STUDY SITE BY

SECTION, 1974–1976.

1974 1975 1976

Section 1*

A. chloris 43 7 48A. peraccae 213 101 98

Section 2**

A. chloris 43 115A. peraccae 98 22

* x2 5 29.80, p y 0.0001.

** x2 5 80.90, p y 0.0001.


inclusive generalizations regarding ecologi-cal patterns in mainland anoles at thepresent time.

II. GROWTH AND POPULATIONSTRUCTURE OF TWO SPECIES OF

ANOLIS FROM THE PACIFIC SLOPESAND LOWLANDS OF ECUADOR

INTRODUCTION

The iguanid lizard genus Anolis has beenthe subject of a large body of ecologicalresearch over the past two decades, con-centrated largely on the islands of the WestIndies and, lately, on selected parts ofCentral America and Mexico. The largeand fundamentally important South Amer-ican radiations have been largely ignored inthis respect, although some importantcontributions have been made regardingthe general natural history of some of the

species (Vanzolini, 1972; Hoogmoed, 1973;Dixon and Soini, 1975; Duellman, 1978).However, there has been a dearth ofquantitative studies regarding even the mostcommon species of South American anoles.This lack of emphasis on quantitativeecological studies has been the result oftwo major factors. First, the systematics ofSouth American anoles are poorly under-stood. As a result, most of the workersconcerned with the biology of these lizardshave directed their efforts largely towardunraveling the complex fabric of theirevolutionary history on the South Americancontinent, rather than focusing on ecologi-cal studies. This approach has been abso-lutely necessary; without some perspectivein which to place observations concerningthe natural history of the various species, itwould be exceedingly difficult to employmany characters of the living animals inanswering broad questions concerning thebiology of the South American forms.

However, there is a second, even morefundamental, reason why some ecologicalstudies have not been carried on concur-rently with the systematic studies, as hap-pened in the West Indies and is happeningin Central America. A basic fact of lifeconcerning the South American anole faunain particular, and the entire forest lizardfauna in general, is that individuals tend tobe rare. Although this fact in itself is of someinterest, it virtually precludes the possibilityof carrying on detailed quantitative studies.Most species of anoles in most places inSouth America are simply too rare to studyin a quantitative manner. In most cases itrequires a considerable effort to secure asample adequate for even basic systematicstudies, and the chances of collectingadequate amounts of ecological data areslim indeed.

The present paper is the first to presentquantitative data concerning South Ameri-can species of Anolis. The two species Istudied are poorly known ecologically. Theextent of the literature on the biology of A.gemmosus consists of a single paper (Fitchet al., 1976) and a short abstract of parts of

TABLE 17. HABIT OCCURRENCE OF ECUADORIAN ANOLIS LIZARDS ON

THE BASIS OF FIELD WORK BETWEEN 1974 AND 1979. (X) INDICATES

PRESENCE IN HABITAT TYPE, (+) INDICATES THAT SPECIES REACHES

MAXIMUM DENSITY IN THIS HABITAT.

Forest Edge Analogue3

A. aequatorialis + +A. bitectus x xA. chloris x +A. chrysolepis scypheus +A. eulaemus group1 xA. fasciatus x +A. festae x +A. fitchi xA. fraseri x xA. fuscoauratus x +A. gemmosus x +A. gracilipes xA. granuliceps xA. lynchi xA. lyra x +A. maculiventris xA. ortonii x +A. parvauritus2 + xA. peraccae x x +A. princeps + xA. punctatus x +A. trachyderma x +1 Undescribed species from Puyo region.2 Editor’s note: considered A. biporcatus parvauritus by many.3 Editor’s note: by ‘‘analogue,’’ the author apparently means forest-like habitatsoccurring in cultivated areas.


the present work (Miyata, 1977). Anolisbitectus remains completely unknown inthe literature outside of a few systematicdiscussions (all rather old) and the briefabstract in Miyata (1977). These two speciesare rather typical South American anoles inmany respects; both are small to medium insize and are associated with forest habitats.One species (A. gemmosus) is a member ofthe autochthonous South American alpharadiation, and the other (A. bitectus) is amember of the more recently arrived betaradiation (Williams, 1976b). Both haverestricted distributions on the Pacific slopesof the Ecuadorian Andes.

STUDY SITES

Anolis gemmosus was studied at RıoFaisanes, a cloud forest locality on the westslope of the Ecuadorian Andes at anelevation of 1,380 m. The Rıo Faisanes is asmall mountain river that crosses the LaPalma–Quito road (Ecuador Highway 28)some 14.4 km northeast of La Palma. Thegeneral area consists of patches of cloudforest, generally confined to narrow streambeds, with considerable cleared agriculturalland on the less precipitous slopes. Theactual study area consisted of a narrow stripof edge vegetation made up primarily offerns and small shrubs located along anarrow and steep road cut. The areacovered by my censuses spanned some150 m of this roadside vegetation on bothsides of the road. Because of the steepnessof the road cuts, perhaps 2 m on either sideof the road were actually sampled in thecensuses.

This roadside habitat was chosen for themark-recapture study because of the highpopulation densities of A. gemmosus ob-served there on initial surveys in Junethrough August of 1975. These densitieswere fairly typical of the roadside habitats inthe area, and lizards appeared to be in muchlower densities in the adjacent forest.Furthermore, although the roadside vege-tation was subject to some disturbance inthe course of the study, this activity was

confined largely to just the very edges of thevegetation and was intended primarily tokeep the road shoulders clear. The smallpatches of cloud forest in the same area, incontrast, were being cut back at a steadyrate during the course of the study, anddrastic changes occurred in the structure ofthe forest along the banks of the RıoFaisanes between 1976 and 1979.

Anolis bitectus was studied in an artificialplantation of rubber trees (Hevea brasilien-sis; not indigenous to the region) atHacienda Cerro Chico, a large plantationlocated some 45 km north of Quevedo inProvincia Pichincha at an elevation ofapproximately 125 m. The rubber groveswere planted between 1969 and 1970, thearea previously having been a bananaplantation. The original forest in this areawas probably cleared between 1960 and1965. The rubber groves form a forest-analogue habitat, with a dense overheadcanopy through which little light penetrates.The ground is covered with a thick layer ofherbaceous vegetation between 0.3 and1.5 m in height. The rubber trees wereuniformly spaced and sized, with mostbetween 15 and 25 cm dbh and perhaps10–15 m in height. To facilitate harvest ofthe rubber, small paths were kept clear ofunderstory vegetation, and most of theunderstory could be scanned thoroughlyfrom these walkways.

METHODS

Both species of Anolis were collected byhand, usually as they slept at night. Thisnight sampling has several distinct advan-tages over conventional daytime capture: 1)The lizards do not become increasinglydifficult to catch after they have beenmarked; this factor proved to be a majordifficulty in other mark-recapture studiesattempted in the same region on speciesthat could only be found during the day; 2)the lizards are very easy to capture whenthey are asleep; 3) lizards that sleep onunderstory vegetation are usually very easyto see at night because the pattern of dew or


rain beaded up on their skin is quitedifferent from the surrounding leaves; oneof the species, A. gemmosus has an extremelycryptic pattern (Fig. 8) that makes it verydifficult to find while active. Nighttimesampling appears to be far superior toconventional lizard sampling techniques forpopulation studies such as the present one.

Some daytime sampling was also conduct-ed for A. bitectus. The first sampling of thepopulation at Hacienda Cerro Chico wasdone by means of a common random quadratsampling technique (see Lloyd et al., 1968;Scott, 1976), in which plots of 7.6 by 7.6 mwere laid out on the floor of the rubbergrove, and a group of eight people convergedfrom the perimeter and captured the animalsinside. In this manner, 12 quadrats of 58 m2

each were sampled on 9 July 1976.Subsequent visits to the Hacienda Cerro

Chico study sites (April and November1977, January and April 1978, and January1979) employed a different sampling tech-nique. A marked study plot was laid out inthe rubber grove adjacent to the section

sampled in 1976. This plot consisted of some3,948 m2 and had eight rows of rubber treesspaced approximately 6 m apart. The firsttree in each row was adjacent to a serviceroad, which provided an opening in thecanopy, and the understory vegetation wassomewhat denser along this edge. Each rowcontained 21–23 rubber trees, generallyspaced about 3.5 m apart (range 2–6 m).This plot was worked by following the narrowpaths along the tree rows and scanning thetop of the understory vegetation for sleepinganoles at night. Each census required 1–3 hours to complete, depending on thenumber of people I had to assist.

The sampling procedure for A. gemmosusat the Rıo Faisanes site consisted simply ofwalking slowly along the road while careful-ly scanning for sleeping lizards on ferns andshrubs within reach. I was unable to placeany markers here because of the constantminor disturbances, so individual positionsof each lizard could not be recorded.

For each lizard captured, sex and SVLwere noted. For A. bitectus in the marked

Figure 8. Adult male Anolis gemmosus sleeping on fern at Rıo Faisanes study site.


plot, the exact position was also noted.Measurements were taken with plastic dialcalipers to the nearest 0.1 mm or with aplastic rule to the nearest 1 mm. In bothspecies, sex was easily determined by thepresence of a distinct dewlap on males; eventhe smallest juveniles exhibited this sexualdimorphism. Individuals were marked witha unique toe-clip which removed no morethan a single digit from each limb.

Population estimates for the randomquadrat sample of A. bitectus were calcu-lated in the following manner:

NN~n

ba,

where n is the total number of individualsfrom all b quadrats, and a is the area of eachquadrat in hectares. This gives a densityestimate for number of individuals perhectare. Because the populations cannotbe assumed to be randomly distributed, anunbiased estimate of the variance wascalculated by the method of Cochran(1963):

ss2N~B2 v

b1{

b

B

� �,

where B 5 1/a and

v~Xb

i~1

xi{�xxð Þ2

b{1ð Þ :

The 95% confidence intervals are thengiven by

NN+tb{1 0:05½ �ssN :

The study plot population of A. bitectus wasestimated by the weighted least squaresmethod of Schumacher and Eschmeyer(1943). This estimate of mark-recaptureanalysis has been found to be reliable forlizard population studies by Turner (1977),who recommends its use for the evaluationof lizard population densities because of itsconsistently ‘‘realistic’’ results. This tech-

nique employs multiple recaptures andassumes that the population is closed duringthe sampling period. Each census within agiven sampling period was regarded as asample, and estimates were made indepen-dently for each sampling period (i.e.,marked animals from previous samplingperiods were not included in the calcula-tions unless they were found in the presentsample or a later sample). The populationsize is estimated by

NN~

Psi~2

niM2i

Psi~2

miMi

,

where s is the number of samples, mi is thenumber of marked individuals in the ithsample, Mi is the number of markedindividuals in the population just beforethe ith sample, and ni is the number ofindividuals in the ith sample. The variancesand confidence intervals for the estimateswere calculated in the manner of DeLury(1958):

ss2~1

s{2

Psi~2

mi

ni{Psi~2

miMi

�Psi~2

niM2i

� �26664

37775:

The 95% confidence intervals are thengiven by

Psi~2

niM2i

Psi~2

miMi+ts{2 0:05½ � ss2Psi~2

niM2i

� �1=2:

To obtain density estimates for the studyplot population of A. bitectus, N wasmultiplied by 2.53 to obtain the estimatednumber per hectare of rubber grove habitat.

Population estimates of the A. gemmosuspopulations at the Rıo Faisanes werecalculated using the Jolly–Seber method


(Seber, 1973). This method allows for thedeath, recruitment, immigration, and emi-gration of animals into and out of thepopulation. Turner (1977) has suggestedthat this method could result in unrealisticestimates if its assumptions are not met, butthe data for the A. gemmosus study seem tofit the assumptions fairly well (Seber, 1973).In particular, some individuals persisted inthe population through the entire durationof the study. Using this method, thepopulation size is estimated by

NNi~nizi=rið Þzmi

mi=ni

,

where ni is the number of animals in the ithsample, mi is the number of marked animalsin the ith sample, ri is the number ofmarked animals released after the ithsample that are subsequently recaptured,and zi is the number of animals caughtbefore, but not during, the ith sample thatare subsequently recaptured. The varianceis calculated by

ss2~NNi NNi{ni

� �

1zri

zi

� �1

ri

{1

ni

� �z

1

mi

{1

ni

� �� :

Population densities were estimated invery crude terms because the study area wasnever accurately delineated. The study areaconsisted of both sides of a road over a ca.150-m stretch. Because of the steepness ofthe road cut, only 1–3 m of vegetation couldbe sampled on either side of the road. Forthe purpose of having a crude densityestimate for comparative purposes, I esti-mated the area of the study area as 150 3 55 750 m2. To get an estimate of density perhectare, N was multiplied by 13.3.

Growth rates for the two species of anoleswere calculated separately for each sex. Thecharacteristic growth rate (r) and theasymptotic maximum SVL were calculatedusing a logistic-by-weight growth model andfitting a curve by means of a nonlinear leastsquares regression. The computer program

for this calculation was provided by JonSeger. Choice of the particular growthmodel was based on the results of acomparative study of lizard growth modelsby Schoener and Schoener (1978), whofound that the logistic-by-weight modelconsistently resulted in the best fit.

RESULTS

General Observations and Natural History

Both species of anoles are members ofrelatively simple lizard communities. Themontane A. gemmosus is known from lowercloud forest localities on the western slopeof the Ecuadorian Andes as far north as theColombian border. The known elevationrange is from 1,150 to 2,100 m, with thehighest population densities normally foundbetween 1,200 and 1,600 m. Over this rangeit is sympatric with at least four otherspecies of Anolis, but only two of these areregularly found in sympatry with A. gemmo-sus: the giant, crown-dwelling A. fraseri andthe large trunk-ground A. aequatorialis.Anolis bitectus has a restricted distributionin the lowlands of west-central Ecuadornear the base of the Andes, where it seemsto be associated with tropical moist forestvegetation. It is known from relatively fewlocalities, and before the present study, itwas very poorly represented in collections.The small samples collected in the course ofmy field work easily doubled the number ofspecimens in museum collections. Althoughpossibly found in sympatry with as many as10 congeners at various points over itsrange, it has been recorded in sympatrywith only four: A. fasciatus, A. festae, A.peraccae, and A. princeps. In the study areaat Hacienda Cerro Chico it was found withonly two congeners: A. festae and A.peraccae, both small species of the trunksalong forest edge (see above).

Anolis gemmosus is by far the mostabundant species of lizard in all of thecloud forest localities I have visited innorthwestern Ecuador. Although abundantby South American standards, their pres-ence is often very difficult to detect if the


area is sampled by day. My limited obser-vations on active A. gemmosus agree withthose of Fitch et al. (1976). All of theindividuals they saw were perched abovethe ground, with the majority in foliagerather than on tree trunks or stems. Themajority of my observations of activeanimals were of lizards perched on fernsalong road cuts and in other disturbedhabitats, but I did see several adult malesperched on tree trunks several meters offthe ground in small clearings in otherwiseundisturbed cloud forest. Anolis gemmosuswould appear to be a bush-dwelling anolefor the most part, although individualsmight utilize higher perches when available.

My body temperature data are scant,reflecting the lack of time spent searchingfor these lizards during the day. Themean body temperature of four lizards wasl6.2u C, with a mean air temperature ofl6.lu C. Fitch et al. (1976) gave a mean bodytemperature (MET) of 21.0u C for theirsample of 34 temperatures. The discrepancyin METs occurs because three of my fourtemperatures were of animals found shortlyafter dawn, when air temperatures were stillvery low and the lizards were not obviouslyactive. Interestingly enough, the sleepingperches of A. gemmosus generally seemedto be oriented so that the lizards would beexposed to any early morning sunlight; onthree occasions when I observed sleepinganoles before dawn and watched the sunrise, all of the lizards were met by the firstrays of direct sunlight.

Anolis bitectus was found only in thesingle rubber grove at Hacienda Cerro Chico,despite extensive collecting efforts over a 6-year period in the surrounding region. Thisspecies remains known from a very fewscattered localities in west-central Ecuador.

Anolis bitectus is active in the understoryvegetation at the Hacienda Cerro Chicorubber grove and is very rarely seen on thetrunks of the rubber trees. Active individu-als seem to spend most of their time withinthe dense foliage; nearly all of the undis-turbed individuals seen were perched onsmall stems and vines within 20 cm of the

ground, where they were hidden fromabove by the dense vegetation. It wasexceedingly difficult to find undisturbedindividuals, and most of the lizards foundduring the day were seen after vegetationwas disturbed and they revealed theirpresence by flight movements.

At Hacienda Cerro Chico, A. bitectusseems to be essentially restricted to therubber plantation. Despite some searching,both by day and by night, in adjacentplantations of African oil palm and secondgrowth forest, no specimens were found.Even within the rubber groves it seems tobe restricted to the more deeply shadedinterior sections and avoids the more openedges of the groves.

Body temperature data for A. bitectus aresummarized in Figure 9. Individuals werenever observed basking. The intensity ofthermoregulation, k (Huey and Slatkin,1976), is 0.89.

Population Densities

The estimated population sizes for the RıoFaisanes population of A. gemmosus arepresented in Table 18. The density estimatespresented are based on the rough approxi-mation of the sampled area and are bestregarded as maximum density estimates fortwo reasons. First, this study site wasoriginally selected because it appeared tohave the highest population density of A.gemmosus along a 15-km stretch of roadleading through appropriate cloud foresthabitat. I roughly estimated it to have fromtwo to five times the density of other similarareas along the road when the initial surveyswere made in 1975 and 1976. Second, theentire habitat sampled was ecotonal incharacter, lying at the edge of a large patchof closed-canopy cloud forest. Within theadjacent forest, population densities weremuch lower; whereas sampling periods alongthe road resulted in from 13 to 31 individualsobserved, sampling in the adjacent forest onthe same night would generally result infewer than five individuals being found, eventhough much more time was spent and


several times the area was covered. It seemsdoubtful that this discrepancy was due todifferences in sleeping perch preferencesbecause the forest lizards were found onsimilar perches to those along the edge. Theestimated population densities in Table 18,therefore, are estimates of maximum densitiesof A. gemmosus observed in edge habitats,and areal estimates of density for largepatches that include considerable cloud forestwould be much lower, perhaps as much as anorder of magnitude or more.

The populations of A. gemmosus remainedmore or less constant over the course of the

study and from season to season within eachyear. The rather large variances obtained inthe estimates are fairly typical of the Jolly–Seber technique (Seber, 1973), but the closeagreement of the estimates makes thisconstancy seem realistic.

The estimates of population size anddensity of A. bitectus at Hacienda CerroChico are presented in Table 19. Popula-tions of A. bitectus appeared to be spread

TABLE 18. POPULATION AND DENSITY ESTIMATES OF ANOLIS GEMMOSUS

AT RIO FAISANES, ECUADOR. SEE TEXT FOR DISCUSSION OF ESTIMATION

TECHNIQUE.

n Variance Density/ha

August 1976 98 65 1,304April 1977 75 38 995November 1977 80 34 1,065

Figure 9. Body temperatures of Anolis bitectus plotted against air temperature at time of capture. Open circles represent multiplerecords. Diagonal line is for body temperature 5 air temperature.

TABLE 19. POPULATION AND DENSITY ESTIMATES OF ANOLIS BITECTUS

AT RUBBER GROVE STUDY PLOT AT HACIENDA CERRO CHICO, ECUADOR.

SEE TEXT FOR DISCUSSION OF ESTIMATION TECHNIQUES.

n 95% Range Density/ha

July 19761 — — 4452

April 1977 139 105–203 352November 1977 308 219–516 780January 1978 298 105–618 755April 1978 183 107–630 464January 1979 113 43–178 2861 Different sampling technique used.2 The 95% confidence interval for this density estimate is 256–644.


out more or less uniformly throughout therubber plantation, although densities ap-peared to be somewhat lower near themargins of the groves, so the densityestimates provide a reliable picture of theactual densities of the population.

The populations of A. bitectus variedsomewhat from one sampling period to thenext, but with no clear pattern related toseasonality. The lowest population densitieswere recorded in January 1979 and April1977, corresponding roughly to the begin-ning and end of the wet season in this area.However, population densities in Januaryand April 1978 were higher, and the 95%confidence limits overlapped with the high-est recorded density in November 1977. In1977 it appeared as though recruitment intothe population occurred during the dryseason, with the estimated population den-sity perhaps doubling in a 7-month period.In 1978, however, the population appearedto decrease during the dry season.

Size and Growth

Anolis gemmosus reaches a maximumrecorded SVL of 66 mm (Fitch, 1976),although the largest individual marked atthe Rıo Faisanes locality was only 62.5 mmSVL. Fitch also recorded a maximumfemale SVL of 63 mm, whereas the largestfemale at the Rıo Faisanes was 56.8 mm. Themean SVLs that Fitch recorded for hissample of A. gemmosus, taken from Tandapi,a locality some 20 km due south of the RıoFaisanes at 1,460 m elevation, were as high as(males) or higher than (females) the maximaI recorded. Fitch found significant sexualdimorphism in his sample, with the sexualdimorphism ratio (SDR 5 female SVL/maleSVL) equal to 0.94. The SDR for the RıoFaisanes population was 0.92, also highlysignificant. Only adults were used to calcu-late the mean SVLs; adults were consideredall those individuals with SVL . 50 mm (sizeof the smallest females with oviducal eggs).

There were no apparent shifts in the size-frequency distribution of the Rıo Faisanespopulation (Table 20) during the course of

the study. Adults made up a large propor-tion of the population (70%), with adultscomprising 62% of the males and 75% ofthe females. This difference in the propor-tion of adults is not significant (x2 5 2.16,p 5 0.14).

Growth rates using the logistic-by-weightgrowth model and unfixed data (Schoenerand Schoener, 1978) were calculated foreach sex (Figs. 10, 11; Table 21). Thecalculated asymptotic SVL is somewhatsmaller than the recorded maximum SVLs,10

but the SDR calculated from them is quiteclose to that actually found in the popula-tion (0.95). Anolis gemmosus appears to bequite long-lived for a small lizard; adultfemales first marked in July 1976 remainedin the population until at least January 1978(data for April 1978 were lost, but at leastone adult female from the original markedcohort was still present at that time). Growthappears to cease once sexual maturity isreached, at least in the case of females.Because it was not possible to determine atwhat size males attained maturity, it is notclear whether they continue to grow afterthat point. If the assumption made above thatthey reach maturity at the same size asfemales (and the same age, since the

TABLE 20. SIZE FREQUENCY DISTRIBUTION OF ANOLIS GEMMOSUS AT

RIO FAISANES DURING FOUR SAMPLING PERIODS, 1976–1978.

July1976

April1977

November1977

January1978

Female SVL (mm)*

21–30 0 1 0 031–40 1 2 0 041–50 4 3 3 151–60 17 7 7 15

Male SVL (mm)**

21–30 3 0 0 031–40 5 0 2 341–50 1 1 0 151–60 6 4 4 661, 3 0 0 3

* x2 5 9.91, p 5 0.36.

** x2 5 11.88, p 5 0.46.

10 A common finding in studies on anoles (Stampsand Andrews, 1992).


Figure 10. Absolute growth rate of Anolis gemmosus females plotted against initial size. Open circles represent multiple records.

Figure 11. Absolute growth rate of Anolis gemmosus males plotted against initial size. Open circles represent multiple records.


characteristic growth rates are essentially thesame) holds, then males would continue togrow somewhat past this point.

Anolis bitectus reaches a maximum maleSVL of 56.0 mm, and females reach56.5 mm. Females reach sexual maturity atabout 45 mm SVL, and this was taken as theminimum adult size for the Hacienda CerroChico population. Adults show significantsexual dimorphism in size, with the femalesslightly larger than the males (SDR 5 1.02,p 5 0.03). Although the lumped data showsignificant size dimorphism, three of the sixsampling periods show no significant di-morphism (Table 22). In all cases, however,females are slightly larger than males, andthe slight dimorphism in adult size isprobably real.

There were no significant shifts in thesize-frequency distribution of A. bitectusfemales (Table 23). Males may show somedegree of shifting in the size distribution,however, but larger samples would probably

be necessary to demonstrate this. Adultmales compose a smaller proportion of allmales (5l%) than adult females do of allfemales (62%; x2 5 4.88, p 5 0.03).

Growth curves for the A. bitectus popu-lation at Hacienda Cerro Chico are shownin Figures 12 and 13, and the growth rateparameters for the logistic-by-weight modelare presented in Table 21. The asymptoticmaximum SVLs are again lower than thoseactually recorded in the population, and theSDR calculated from these asymptotes(1.07) is slightly higher than the onerecorded for the lumped sample, althoughit is similar to some of the individualsampling period SDRs (Table 22). UnlikeA. gemmosus, female A. bitectus appear tocontinue growing after sexual maturity isreached (Fig. 12).

By substituting the calculated values ofthe characteristic growth rate r and theasymptotic maximum SVL into the logistic-by-weight growth equation, it is possible tocalculate the time needed to reach sexualmaturity provided the size at hatching(SVL0) and at sexual maturity (SVLsm) areknown. The equation is

Tsm~{1

rIn

SVL30 SVL3

max{SVL3sm

� �SVL3

sm SVL3max{SVL3

0

� �

Although I have no data on the hatchlingsize of either species, I have used thesmallest individual recorded of each speciesas an indication of hatchling size, assuming

TABLE 22. SEXUAL SIZE DIMORPHISM IN ANOLIS BITECTUS, ADULTS ONLY (SVL $ 44 MM).

Males Females

SDR1n SVL s n SVL s

July 1976 4 48.4 0.522 10 51.3 8.800 1.06*April 1997 19 50.0 3.222 27 50.1 8.165 1.00November 1977 20 48.8 5.735 24 49.4 7.986 1.01January 1978 26 48.1 5.969 27 50.3 2.612 1.05**April 1978 18 48.2 4.916 18 50.8 3.838 1.05**January 1979 21 50.7 8.776 17 51.1 7.232 1.011 Female SVL: Male SVL.

* Dimorphism significant at p , 0.01.

** Dimorphism significant at p , 0.0001.

TABLE 21. GROWTH RATES OF ANOLIS BITECTUS AND A. GEMMOSUS IN

WESTERN ECUADOR.

n

MaximumEstimated

SVL r SSQR

A. bitectus

Male 16 48.8 0.0248 39.783Female 31 52.0 0.0211 69.919

A. gemmosus

Male 13 57.7 0.0104 24.662Female 41 54.5 0.0094 31.799


that hatchling males are the same size ashatchling females. These values are 21.6 mmfor A. bitectus and 26.8 mm for A.gemmosus. Both sexes were assumed toreach sexual maturity at the same SVL (seeabove), and the minimum SVL of gravidfemales was used as this parameter; for A.bitectus this was 45 mm, and for A.gemmosus this was 50 mm. Substitutingthese values and the appropriate values ofSVLmax and r from Table 21 into the above

equation, the following results were ob-tained. Both male and female A. bitectusreach sexual maturity in about 150 days(males 147, females 150); male A. gemmosustake 271 days to reach maturity, and femaleA. gemmosus require almost a full year(343 days).

Reproduction

In both species the same basic pattern ofreproductive activity was noted. In nearly all

Figure 12. Absolute growth rate of Anolis bitectus females plotted against initial size. Open circles represent multiple records.

TABLE 23. SIZE FREQUENCY DISTRIBUTION OF ANOLIS BITECTUS AT HACIENDA CERRO CHICO DURING SIX SAMPLING PERIODS, 1976–1979.

July 1976 April 1977 November 1977 January 1978 April 1978 January 1979

Female SVL (mm)*

21–32 2 8 5 3 6 833–44 5 5 17 10 3 545. 10 26 24 28 18 17

Male SVL (mm)**

21–32 6 9 4 16 7 433–44 4 9 13 13 16 445. 4 19 20 26 18 17

* x2 5 13.44, p 5 0.20.

** x2 5 17.47, p 5 0.06.


cases, adult females sampled at all seasonswere gravid. At least one large oviducal eggcould be felt or seen in almost every femalelizard more than 2 mm larger than theminimum adult female size mentionedabove. The few cases of non-gravid femalescould very well have been of lizards whichhad just laid their egg. I have no informationon clutches for the marked populations, butanoles are known to generally have single-egg clutches (Fitch, 1970).

Hatchling-sized juveniles were not pres-ent in every sample. This size class waspoorly represented in A. gemmosus, com-prising no more than two individuals in anysampling period, and I am unable to makeany statements regarding the seasonality oftheir appearance. For A. bitectus, thepercentage of hatchlings (20–26 mm SVL)in the population varied from 2% inNovember to 3% in January to 11% in Aprilto 19% in July (x2 5 7.67, p 5 0.05). Thissuggests a peak in hatching some time nearthe end of the rainy season and a consider-

able decrease in hatching toward the end ofthe dry season. Because females continuedto be almost uniformly gravid throughoutthe year, this suggests either a decrease inhatching success during the dry season or anincrease in egg retention and longer inter-vals between clutches. I have no data thatcan distinguish between these alternatives.

DISCUSSION

Although the two species of Anolisconsidered in this paper are both fromnearby regions and are basically similar inperch preferences, generally perching onvegetation rather than on tree trunks, theyshow more differences than similarities inmost aspects of their biology. Although bothspecies exhibit similar ranges in active bodytemperatures (see Fitch et al. [1976] formore temperature data on A. gemmosus), A.bitectus is thermally passive in the strictestsense (Huey and Slatkin, 1976), whereas A.gemmosus appears to exhibit some thermo-regulatory behaviors. The populations of A.

Figure 13. Absolute growth rate of Anolis bitectus males plotted against initial size. Open circles represent multiple records.


gemmosus also tend to remain constantthrough time, with a stable size-frequencydistribution, whereas A. bitectus undergoesmoderate population fluctuations and mayshow some shifts in size-frequency distribu-tion (at least in the case of males). Bothspecies show similar slight degrees of sexualsize dimorphism, but in A. bitectus thefemales are larger, and in A. gemmosus themales are larger. Anolis bitectus has a fairlyhigh characteristic growth rate, whereas thatof A. gemmosus is quite low. To what extentare these differences due to differences inthe local environment and to what extentmight they be due to different competitivemilieus?

The apparent difference in thermal strat-egies between the two species probablyreflects differences in the climates to whichthe species are exposed. Anolis bitectus livesin an area where temperatures tend to bewarm, rarely dropping below 20u C even atnight. These lizards are still quite mobile atthat temperature and probably can begintheir daily activity early in the morning. Thecloud forests that A. gemmosus inhabits, onthe other hand, are characteristically cool,and air temperatures before dawn can be aslow as 10u C or even lower. At thistemperature the lizards are virtually immo-bile, and it might take a considerableamount of time for the air to reachtemperatures commensurate with activitytemperatures of the lizards. By selectingsleeping perches exposed to any availablemorning sun, they are able to reach activitytemperatures more quickly, and they can doso behaviorally rather than relying on morecomplex physiological adaptations to lowtemperatures. Furthermore, because earlymornings are the only times of reliablesunlight in the cloud forest habitat, it mightbe important for these lizards to beginactivity early so that when the almostinevitable cloud cover settles in an hour ortwo after dawn they can already be goingabout their business. Behavioral adaptationsto thermoregulation appear to be muchmore easily evolved in the genus Anolis thanphysiological responses, and a trend to

increasing thermoregulation with altitudehas been noted in the West Indies (Hertzand Huey, 1981).

The relative constancy of the A. gemmo-sus populations is also probably a reflectionof the cloud forest environment. Althoughprecipitation in the cloud forests of westernEcuador show some degree of seasonality,more or less paralleling that of the adjacentPacific lowlands, cloud cover is almostconstant in this region, and afternoon mistsand drizzles occur virtually daily. In terms ofproductivity, the cloud forest environmentis probably quite constant as periods ofwater shortage are scarce and of shortduration. Given a relatively constant levelof productivity, it is not surprising thatpopulations of at least some animals mightbe relatively constant as well. The lowlandareas that A. bitectus inhabits, on the otherhand, have a definite dry season, andalthough this period is not generally severeenough to result in deciduous vegetation,there is no question but that seasonalvariation in productivity exists. Further-more, this area is subject to considerableannual variation in precipitation, both withrespect to total amount and to the exacttiming and duration of the wet season. Thisadds more possible variation to productivity,which is difficult to assess over short-termstudies such as the present one but whichcould easily result in the moderate popula-tion fluctuations observed in the A. bitectuspopulation.

Andrews (1979) provided compellingevidence to suggest that mainland speciesof Anolis differ fundamentally from islandspecies in a number of important aspects oflife history, which are related to the basicroles the lizards play in their respectivecommunities. Her basic thesis is that islandspecies, which tend to be dominant mem-bers of their terrestrial vertebrate faunas,are basically food-limited, and intraspecificcompetition has resulted in life historystrategies that tend to be K-selected (Mac-Arthur and Wilson, 1967; Pianka, 1970).Mainland species, on the other hand, aregenerally not dominant members of their


terrestrial vertebrate faunas and are subjectto intensive predation pressure from a widevariety of potential predators that are absentfrom the West Indies (certain families ofbirds in particular); intraspecific competitionis therefore not as severe as on the islands,and the mainland species are more r-selected. These conclusions were reachedafter a consideration of a variety of studiesdealing with the ecology of West Indian,Mexican, and Central American anoles. Howdo the two species of South American anolefit into this general picture?

One of Andrews’ predictions concerningmainland species is met by both A. bitectusand A. gemmosus. Both species exhibit verysmall degrees of sexual size dimorphism.This slight difference in mean adult size isprobably not of any consequence in terms ofresource partitioning. Slight, but statisticallysignificant, differences in adult size could bedue to a number of factors. The smaller sexmight be more susceptible to predation thanthe larger because of differences in behav-ior. If males tend to live shorter lives,perhaps because of more conspicuousperches subjecting them to increased dan-ger from predation, this could account for aslight difference in mean adult SVL. Thismight be the case for A. bitectus. Thelongest interval between recaptures formales was 197 days, whereas 11 femaleswere recaptured after intervals of more than200 days, and the longest recapture intervalwas 349 days. No similar pattern wasobserved in the A. gemmosus population,as individuals of both sexes persisted in thepopulation for more than 600 days. Thelarger size of males could be due to moregrowth after sexual maturity is reachedbecause they do not have to allocate partof their energy intake to production of eggs.

Both species of South American anolesdiffer in one major respect from the CentralAmerican anoles. Both A. bitectus and A.gemmosus require long periods to reachsexual maturity. For the small, lowland A.bitectus, approximately 5 months is requiredto reach the minimum SVL at whichoviducal eggs are noted. This is considerably

longer than most small Central Americananoles require and is comparable to theamount of time needed by a large specieslike A. frenatus (Andrews, 1976). This ismore or less in keeping with the maturationtime required by the West Indian species.Anolis gemmosus requires even longer toreach sexual maturity, with females needingalmost a year to reach minimum adult SVL.However, the maturation time of male A.gemmosus is fairly similar to that of aCentral American montane species, A.tropidolepis (Andrews, 1976), which sug-gests that the cloud forest environmentmight play a major role in determining theslow growth rate.

It is not immediately clear why SouthAmerican anoles might require longer toreach sexual maturity than the CentralAmerican species. The lowland communi-ties that A. bitectus is from are certainlyquite similar in most respects to lowlandhabitats in Central America in terms ofpossible competitors and predators, andthere is no reason to believe that the basicstructure of the communities differs in anysignificant and obvious way. The answer tothis problem could in fact be quite simple;A. bitectus might not be ecologically similarto the Central American species studied.Unlike most of the Central Americanspecies, which tend to be associated withedge and disturbed habitats in general, theSouth American species allied with A.bitectus are typically forest-restricted spe-cies. Because no populations of A. bitectusaside from the study population werediscovered in six years of field work, itmight be reasonable to assume that theirpresence in high densities there is the resultof a historical accident. I strongly suspectthat A. bitectus is primarily a forest speciesand that its presence in the rubber grove atHacienda Cerro Chico might be mostunusual.

Most Central American anoles that havereceived detailed study are characteristicallyspecies of the forest edge, and most of thestudies have been conducted in disturbedhabitats. It is generally only in these habitats


that mainland anoles reach population den-sities high enough to warrant study. Howev-er, by restricting the species studied to thosenormally associated with edge situations, thepossibility of mistaken generalization couldset in. Although anoles are abundant in edgesituations, and many species inhabit thesehabitats, a large percentage of species,particularly in South America, are essentiallyrestricted to forest interiors. Unfortunately,these species are nearly always quite rare,and there exists little possibility of conduct-ing quantitative field studies on a species thatmight only be seen once every 2 weeks. If, asI believe, A. bitectus is normally such aspecies, then the Hacienda Cerro Chicorubber grove might have provided an oppor-tunity to obtain some insight into the biologyof one of these species normally difficult tostudy, and this could account for some of thedifferences in life history.

It is in precisely those species that arecharacteristic of edge situations that onemight expect to encounter relatively r-selected life history strategies. Because mostWest Indian anoles abound in such habitats,the comparisons that Andrews (1979) madebetween Central American and West Indiananoles are legitimate, and the patterns shefound are of real and general significance.However, I do not believe that it is possibleas yet to determine whether these patternsare typical of mainland anoles as a whole.Given the implausibility of much of the fieldwork necessary to establish whether thesepatterns are typical, a thorough understand-ing of the ecology of mainland anolesprobably lies far in the distant future.

LITERATURE CITED

ANDREWS, R. M. 1976. Growth rate in island andmainland anoline lizards. Copeia 1976: 477–482.

ANDREWS, R. M. 1979. Evolution of life histories: acomparison of Anolis lizards from matched islandand mainland habitats. Breviora 454: 1–51.

ANDREWS, R. M. 1991. Population stability of a tropicallizard. Ecology 72: 1204–1217.

ANDREWS, R. M., AND J. D. NICHOLS. 1990. Temporaland spatial variation in survival rates of the tropicallizard Anolis limifrons. Oikos 57: 215–221.

ANDREWS, R. M., AND A. S. RAND. 1982. Seasonalbreeding and long-term population fluctuations inthe lizard Anolis limifrons, pp. 405–412. In E. B.Leigh, A. S. Rand, and D. M. Windsor (eds.), TheEcology of a Tropical Forest: Seasonal Rhythmsand Long-Term Changes. Washington, D.C.:Smithsonian Institution Press.

ANDREWS, R. M., AND S. J. WRIGHT. 1994. Long-termpopulation fluctuations of a tropical lizard. A test ofcausality, pp. 267–284. In L. J. Vitt and E. R.Pianka (eds.), Lizard Ecology. Princeton, NewJersey: Princeton Univ. Press.

AVILA-PIRES, T. C. S. 1995. Lizards of BrazilianAmazonia (Reptilia: Squamata). Zoologische Ver-handelingen 299: 1–706.

BOCK, B. C., A. M. ZAPATA, AND V. P. PAEZ. 2010.Survivorship rates of adult Anolis mariarum(Squamata: Polychrotidae) in two populations withdiffering mean and asymptotic body sizes. PapeisAvulsos de Zoologia 50: 43–50.

CASTANEDA, M. d. R., AND K. DE QUEIROZ. 2011.Phylogenetic relationships of the Dactyloa cladeof Anolis lizards based on nuclear and mitochon-drial DNA sequence data. Molecular Phylogeneticsand Evolution 61: 784–800.

CASTANEDA M. d. R. and K. de Queiroz. 2013.Phylogeny of the Dactyloa Clade of Anolis Lizards:New Insights from Combining Morphological andMolecular Data. Bulletin of the Museum ofComparative Zoology 160(7): 345–398.

COCHRAN, W. G. 1963. Sampling Techniques. NewYork: John Wiley & Sons.

D’ANGIOLELLA, D., T. GAMBLE, T. C. AVILA-PIRES, G. R.COLLI, B. P. NOONAN, AND L. J. VITT. 2011. Anolischrysolepis Dumeril & Bibron, 1837 (Squamata:Polychrotidae), revisited: Molecular phylogeny andtaxonomy of the Anolis chrysolepis species group.Bulletin of the Museum of Vertebrate Zoology 160:35–63.

DELURY, B. 1958. The estimation of population sizeby marking and recapture procedure. Journalof the Fisheries Research Board of Canada 15:19–25.

DIXON, J. R., AND P. SOINI. 1975. The reptiles of theupper Amazon, Iquitos region, Peru. I: lizards andamphisbaenians. Milwaukee Public Museum Con-tributions in Biology and Geology 4: 1–58.

DODSON, C. H., AND A. H. GENTRY. 1978. Flora of theRıo Palenque Science Center, Los Rıos Province,Ecuador. Selbyana 4: 1–628.

DUELLMAN, W. E., 1978. The biology of an equatorialherpetofauna in Amazonian Ecuador. University ofKansas Museum of Natural History, MiscellaneousPublication 65: 1–352.

FITCH, H. S. 1970. Reproductive cycles in lizards andsnakes. University of Kansas Museum of NaturalHistory, Miscellaneous Publication 52: 1–247.

FITCH, H. S. 1972. Ecology of Anolis tropidolepis inCosta Rican cloud forest. Herpetologica 28: 10–21.

FITCH, H. S. 1973. Population structure and survivor-ship in some Costa Rican lizards. Occasional


Papers of the Museum of Natural History, Univer-sity of Kansas 18: 1–41.

FITCH, H. S. 1975. Sympatry and interrelationships inCosta Rican anoles. Occasional Papers of theMuseum of Natural History, University of Kansas40: 1–60.

FITCH, H. S. 1976. Sexual size differences in themainland anoles. Occasional Papers of the Museumof Natural History, University of Kansas 50: 1–21.

FITCH, H. S., A. F. ECHELLE, AND A. A. ECHELLE. 1976.Field observations on rare or little known mainlandanoles. University of Kansas Science Bulletin 51:91–128.

GUYER, C. 1986. Seasonal patterns of reproduction ofNorops humilis (Sauria: Iguanidae) in Costa Rica.Revista de Biologıa Tropical 34: 247–251.

HERTZ, P. E., AND R. B. HUEY. 1981. Compensation foraltitudinal changes in the thermal environment bysome Anolis lizards on Hispaniola. Ecology 62:515–521.

HOLDRIDGE, L. R. 1967. Life Zone Ecology. San Jose,Costa Rica: Tropical Science Center.

HOOGMOED, M. S. 1973. Notes on the Herpetofauna ofSurinam. IV: the lizards and amphisbaenians ofSurinam. The Hague: W. Junk.

HUEY, R. B., AND M. L. SLATKIN. 1976. Cost and benefitsof lizard thermoregulation. Quarterly Review ofBiology 51: 363–384.

LLOYD, M. R., R. F. INGER, AND F. W. KING. 1968. Onthe diversity of reptile and amphibian speciesdiversity in a Bornean rain forest. AmericanNaturalist 102: 497–515.

LOSOS, J. B. 2009. Lizards in an Evolutionary Tree:Ecology and Adaptive Radiation of Anoles. Berke-ley, California: Univ. of California Press.

MACARTHUR, R. H., AND E. O. WILSON. 1967. TheTheory of Island Biogeography. Princeton, NewJersey: Princeton Univ. Press.

MESQUITA, D. O., G. R. COLLI, F. G. R. FRANCA, AND

L. J. VITT. 2006. Ecology of a Cerrado lizardassemblage in the Jalapao region of Brazil. Copeia2006: 460–471.

MESQUITA, , D. O., G. R. COLLI, AND L. J. VITT. 2007.Ecological release in lizard assemblages in Neo-tropical savannas. Oecologia 153: 185–195.

MIYATA, K. 1977. Report, pp. 61–73. In E. E. Williams(ed.), Anolis Newsletter III. Cambridge, Massa-chusetts: Museum of Comparative Zoology.

MOLINA-ZULUAGA, C., AND P. D. A. GUTIERREZ-CARDE-

NAS. 2007. Uso nocturno de perchas en dosespecies de Anolis (Squamata: Polychrotidae) enun bosque Andino de Colombia. Papeis Avulsos deZoologia 47: 273–281.

NELSON, M. 1973. The Development of Tropical Lands:Policy Issues in Latin America. Baltimore: JohnsHopkins Univ. Press.

PARMELEE, T. C., AND C. GUYER. 1995. Sexual differenc-es in foraging behavior of an anoline lizard, Noropshumilis. Journal of Herpetology 29: 619–621.

PIANKA, E. R. 1970. On r- and K-selection. AmericanNaturalist 104: 592–597.

POE, S. 2004. Phylogeny of anoles. HerpetologicalMonographs 18: 37–89.

RAND, A. S. 1964. Ecological distribution in anolinelizards of Puerto Rico. Ecology 45: 745–752.

RAND, A. S., AND S. S. HUMPHREY. 1968. Interspecificcompetition in the tropical rain forest: ecologicaldistribution among lizards at Belem, Para. Pro-ceedings of the United States National Museum125: 1–17.

RAND, A. S., AND E. E. WILLIAMS. 1969. The anoles of LaPalma: aspects of their ecological relationships.Breviora 327: 1–19.

SCHAEPFER, M. A. 2006. Growth rates and bodycondition in Norops polylepis (Polychrotidae) varywith respect to sex but not mite load. Biotropica38: 414–418.

SCHOENER, T. W. 1968. The Anolis lizards of Bimini:resource partitioning in a complex fauna. Ecology49: 704–726.

SCHOENER, T. W. 1977. Competition and the niche, pp.35–136. In C. Gans and D. W. Tinkle (eds.).Biology of the Reptilia. Vol. 7. Ecology andBehavior A. New York: Academic Press.

SCHOENER, T. W., AND G. C. GORMAN. 1968. Some nichedifferences in three Lesser Antillean lizards of thegenus Anolis. Ecology 49: 819–830.

SCHOENER, T. W., AND A. SCHOENER. 1971a. Structuralhabitats of West Indian Anolis lizards. I: LowlandJamaica. Breviora 368: 1–53.

SCHOENER, T. W., AND A. SCHOENER. 1971b. Structuralhabitats of West Indian Anolis lizards. II. PuertoRican uplands. Breviora 375: 1–39.

SCHOENER, T. W., AND A. SCHOENER. 1978. Estimatingand interpreting body-size growth in some Anolislizards. Copeia 1978: 390–405.

SCHOENER, T. W., AND A. SCHOENER. 1980. Densities,sex ratios, and population structure in four speciesof Bahamian Anolis lizards. Journal of AnimalEcology 49: 19–53.

SCHUMACHER, F. X., AND R. W. ESCHMEYER. 1943. Theestimation of fish populations in lakes and ponds.Journal of the Tennessee Academy of Science 18:228–249.

SCOTT, N. J., JR., 1976. The abundance and diversity ofthe herpetofaunas of tropical forest litter. Biotro-pica 8: 41–58.

SEBER, G. A. F. 1973. The Estimation of AnimalAbundance and Related Parameters. London: C.Griffin.

SEXTON, O. J. 1967. Population changes in a tropicallizard Anolis limifrons on Barra ColoradoIsland, Panama Canal Zone. Copeia 1967: 219–222.

SEXTON, O. J., H. F. HEATWOLE, AND E. H. MESETH.1963. Seasonal population changes in the lizard,Anolis limifrons, in Panama. American MidlandNaturalist 69: 482–491.

STAMPS, J. A., AND R. M. ANDREWS. 1992. Estimatingasymptotic size using the largest individuals persample. Oecologia 92: 503–512.


TAYLOR, E. H. 1956. A review of the lizards of CostaRica. University of Kansas Science Bulletin 38: 3–322.

TURNER, F. B. 1977. The dynamics of populations ofsquamates, crocodilians and Rhynchocephalians,pp. 157–264. In C. Gans and D. W. Tinkle (eds.),Biology of the Reptilia. Vol. I. Ecology andBehavior A. New York: Academic Press.

VANZOLINI, P. E. 1972. Miscellaneous notes on theecology of some Brazilian lizards. Papeis Avulsos deZoologia (Sao Paulo) 26: 83–115.

VANZOLINI, P. E., AND E. E. WILLIAMS. 1970. SouthAmerican anoles: the geographic differentiationand evolution of the Anolis chrysolepis speciesgroup (Sauria, Iguanidae). Arquivos de Zoologia(Sao Paulo) 19: 1–298.

VELASCO, J. A., AND A. HERREL. 2007. Ecomorphology ofAnolis lizards of the Choco region in Colombia andcomparisons with Greater Antillean ecomorphs.Biological Journal of the Linnean Society 92: 29–39.

VITT, L. J., T. C. S. AVILA-PIRES, M. C. ESPOSITO, S. S.SARTORIUS, AND P. A. ZANI. 2003a. Sharing Amazonrainforest trees: ecology of Anolis punctatus and A.transversalis (Squamata: Polychrotidae). Journal ofHerpetology 37: 48–57.

VITT, L. J., T. C. S. AVILA-PIRES, P. A. ZANI, S. SARTORIUS,AND M. C. ESPOSITO. 2003b. Life above ground:ecology of Anolis fuscoauratus in the Amazonrainforest with comparisons to its nearest relatives.Canadian Journal of Zoology 81: 142–156.

VITT, L. J., S. S. SARTORIUS, T. C. S. AVILA-PIRES, AND

M. C. ESPOSITO. 2001. Life on the leaf litter: theecology of Anolis nitens tandai in the BrazilianAmazon. Copeia 2001: 401–412.

VITT, L. J., D. B. SHEPARD, G. H. C. VIEIRA, J. P.CALDWELL, G. R. COLLI, AND D. O. MESQUITA. 2008.Ecology of Anolis nitens brasiliensis in CerradoWoodlands of Cantao. Copeia 2008: 144–153.

VITT, L. J., AND P. A. ZANI. 1996. Organization of ataxonomically diverse lizard assemblage in Amazo-nian Ecuador. Canadian Journal of Zoology 74:1313–1335.

VITT, L. J., AND P. A. ZANI. 2005. Ecology andreproduction of Anolis capito in rainforest ofsoutheastern Nicaragua. Journal of Herpetology39: 36–42.

VITT, L. J., P. A. ZANI, AND R. D. DURTSCHE. 1995.Ecology of the lizard Norops oxylophus (Polychro-tidae) in lowland forest of southeastern Nicaragua.Canadian Journal of Zoology 73: 1918–1927.

VITT, L. J., P. A. ZANI, AND M. C. ESPOSITO. 1999.Historical ecology of Amazonian lizards: implica-tions for community ecology. Oikos 87: 286–294.

WERNECK, F. P., G. R. COLLI, AND L. J. VITT. 2009.Determinants of assemblage structure in Neotrop-ical dry forest lizards. Austral Ecology 34: 97–115.

WILLIAMS, E. E. 1969. The ecology of colonization asseen in the zoogeography of anoline lizards onsmall islands. Quarterly Review of Biology 44:345–389.

WILLIAMS, E. E. 1972. The origin of faunas. Evolution oflizard congeners in a complex island fauna: a trialanalysis. Evolutionary Biology 6: 47–89.

WILLIAMS, E. E. 1976a. West Indian anoles: a taxonomicand evolutionary summary. I: introduction and aspecies list. Breviora 440: 1–21.

WILLIAMS, E. E. 1976b. South American anoles: thespecies groups. Papeis Avulsos de Zoologia (SaoPaulo) 29: 259–268.


studies on the ecology and population biology of...

Documents