increased pollinator habitat enhances cacao fruit … · increased pollinator habitat enhances...

887

Increased pollinator habitat enhances cacao fruit set and predator conservation

Samantha J. ForbeS1 and tobin d. northField

Centre for Tropical Environmental and Sustainability Science, College of Science and Engineering, James Cook University, P.O. Box 6811, Cairns, Queensland 4870 Australia

Abstract. The unique benefits of wild pollinators to the productivity of agricultural crops have become increasingly recognized in recent decades. However, declines in populations of wild pollinator species, largely driven by the conversion of natural habitat to agricultural land and broad- spectrum pesticide use often lead reductions in the provision of pollination services and crop production. With growing evidence that targeted pollinator conservation improves crop yield and/or quality, particularly for pollination specialist crops, efforts are increasing to substitute agriculturally intensive practices with those that alleviate some of the negative impacts of agriculture on pollinators and the pollination services they provide, in part through the provision of suitable pollinator habitat. Further, similarities between the responses of some pollinators and predators to habitat management suggest that efforts to conserve pollinators may also encourage predator densities. We evaluated the effects of one habitat management practice, the addition of cacao fruit husks to a monoculture cacao farm, on the provision of pollination services and the densities of two groups of entomophagous predators. We also evaluated the impacts of cacao fruit husk addition on pollen limitation, by crossing this habitat manipulation with pollen supplementation treatments. The addition of cacao fruit husks increased the number of fruits per tree and along with hand pollination treatments, increased final yields indicating a promotion of the pollination ecosystem service provided by the special-ist pollinators, midges. We also found that cacao fruit husk addition increased the densities of two predator groups, spiders and skinks. Further, the conservation of these predators did not inhibit pollination through pollinator capture or deterrence. The findings show that, with mod-erate habitat management, both pollinator and predator conservation can be compatible goals within a highly specialized plant–pollinator system. The effectiveness of this habitat manipula-tion may be attributable to the increased availability of alternative habitat and food resources for both pollinators and predators. The results exemplify a win- win relationship between agri-cultural production and biological conservation, whereby agricultural practices to support vital pollinators and pollination services can increase production as well as support species conservation.

Key words: agriculture; biodiversity; ecosystem services; habitat management; natural predators; pollina-tion; Theobroma cacao; wild pollinators.

introduction

Pollinators play an essential role in plant reproduction in most terrestrial ecosystems (Ollerton et al. 2011) and represent an important ecosystem service vital to the maintenance of both wild plant diversity (Ashman et al. 2004, Aguilar et al. 2006) and agricultural production (Klein et al. 2007, Ricketts et al. 2008). In both land-scapes, pollination is strongly dependent on the services provided by wild (i.e., unmanaged) pollinator commu-nities (Biesmeijer et al. 2006, Losey and Vaughan 2006). For agricultural systems in particular, where pollinators are at high risk from pesticides (Brittain et al. 2010) and habitat loss (Kennedy et al. 2013), the importance of wild pollinators has become increasingly recognized in recent

decades (Rader et al. 2016), with wild pollinators shown to provide a unique benefit to agricultural crops that is not provided by commercially managed honey bees (Garibaldi et al. 2013). Despite the importance of wild pollinators in agricultural crop production, evidence of their declines is increasing (Biesmeijer et al. 2006, Cameron et al. 2011, Weiner et al. 2014) and improve-ments in the management and protection of wild polli-nators has become imperative in agricultural landscapes (Buchmann and Nabhan 1996, Allen- Wardell et al. 1998, Potts et al. 2010, Garibaldi et al. 2011).

Wild pollinators require the presence of natural and seminatural habitats for the survival and persistence of their populations (Kearns and Inouye 1997, Kremen et al. 2004). Thus, the development of habitat man-agement practices to promote natural or seminatural habitats in agricultural systems can augment increases in pollinator communities and the pollination services they provide to crops (Tepedino and Stanton 1981, Williams

Ecological Applications, 27(3), 2017, pp. 887–899© 2016 by the Ecological Society of America

Manuscript received 18 February 2016; revised 7 Novem-ber 2016; accepted 22 November 2016. Corresponding Editor: David S. Schimel.

1E-mail: [email protected]

mailto:[email protected]

888 Ecological Applications Vol. 27, No. 3SAMANTHA J. FORBES AND TOBIN D. NORTHFIELD

and Kremen 2007, Potts et al. 2010). For example, the sowing of nectar and pollen- rich wildflower strips can provide alternative food for pollinators (Pywell et al. 2005, Menz et al. 2011, M’Gonigle et al. 2015). For pol-lination specialist crops, where successful pollination depends upon only a small group of pollinating species, the loss of even a few pollinators may lead to complete reproductive failure (Bond 1994). In such cases, habitat management practices can be uniquely attuned to specif-ically promote those important pollinator species (Kleijn et al. 2015, Senapathi et al. 2015). Combining such habitat management practices with the reduction or elim-ination of pesticides can lead to further increases in pol-linator communities, with subsequent positive effects on fruit set and agricultural productivity (Potts et al. 2010).

Some management practices that generally promote pollinator communities (i.e., habitat manipulation and the reduction of pesticide use) may also provide addi-tional benefits to other animals in agroecosystems (Bianchi et al. 2006, Crowder et al. 2012, Wratten et al. 2012). Some pollinators and predators have shown similar responses to increased habitat complexity (Shackelford et al. 2013) and pesticide reduction (Bianchi et al. 2006), with positive effects on fruit set and agricul-tural productivity (Kremen and Miles 2012, Classen et al. 2014). These identified compatibilities have led to calls for the simultaneous implementation of agricultural management practices that benefit both pollinator and predator conservation (Shackelford et al. 2013, Classen et al. 2014). For example, the interspersing of agricultural croplands with natural habitat patches or floral resources, combined with the employment of integrated pest man-agement practices has been shown to promote polli-nators, predators, and the important ecosystem services they provide (Dale and Polasky 2007, Diekotter et al. 2010, Stallman 2011).

These types of examples are most applicable for gener-alist pollinator communities (that utilize a diversity of habitats for their nutritional and reproductive needs) and pollination generalist crops (where pollination can be provided by a range of potential species). However, it remains unclear whether the simultaneous implemen-tation of agricultural management practices that promote pollinator and predator communities can be comple-mentary when utilizing targeted habitat management practices that are specifically tailored for key pollinator species of pollination specialist crops. If possible, adopting such management strategies may be invaluable, particularly for tropical smallholder crops where the pro-motion of naturally occurring ecosystem services is often more economically sustainable than utilizing external inputs (Tittonell 2014).

Cacao (Theobroma cacao L., Malvaceae) is a cauli-florous, neotropical, understory tree native to the northern parts of South America (Motamayor et al. 2002) that produces cocoa, which is among the world’s most economically important fruit tree commodities (Edwin and Masters 2005). The sustained increase in global cocoa

consumption (Rice and Greenberg 2000, Lass 2004) has led to a worldwide push for the intensification of cacao cultivation (Bisseleua et al. 2009, Deheuvels et al. 2014). The transition of cacao cultivation from traditionally diverse agroecosystems to modern, simplified, and chem-ically dependent agricultural models (Moguel and Toledo 1999, Vaast and Somarriba 2014), may inhibit natural biodiversity- driven ecosystem services such as pollination and pest control that often benefit productivity (Schroth and Harvey 2007, Clough et al. 2009, 2011, Gockowski and Sonwa 2011, Wielgoss et al. 2014).

Pollination is recognized as a major extrinsic factor lim-iting cacao production (Bos et al. 2007, Groeneveld et al. 2010). In general, only 10% of the flowers produced by a cacao tree are successfully pollinated by insects, and increases in yields have been obtained when this polli-nation rate is artificially increased (by hand pollination) to 40% (Groeneveld et al. 2010). As a pollination spe-cialist, the successful pollination of cacao is suggested to depend, almost completely, on the cross- pollination services provided by a single dipteran family, the cera-topogonid midges (Ceratopogonidae), whose distinct morphological and behavioral characteristics make them effective pollinators of cacao (Kaufmann 1975b, Young 1982). Low natural midge abundances are reported in cacao plantations (Winder and Silva 1972, Young 1982, Ismail and Ibrahim 1986, Frimpong et al. 2009), which may be driving low natural pollination rates (Frimpong et al. 2011). Low midge abundance and the subsequent inadequate pollination of cacao by midges has been attributed to the fastidious removal of suitable midge habitat (i.e., moist decomposing organic matter for egg oviposition and larval development) (Kaufmann 1975a, Winder 1978, Ismail and Ibrahim 1986) that is associated with the intensification of cacao cultivation (Ismail and Ibrahim 1986), and recent research suggests that pro-viding additional, midge- specific substrates of decom-posing cacao leaf litter, slices of banana pseudostem and cacao fruit husks can increase fruit set (Adjaloo et al. 2013). Thus, it is believed that current cacao production could be increased by enhancing the availability of suitable habitat for midge oviposition within cacao farms, to enhance pollination services through increases in midge densities. Current evidence, however, stems from relatively few field studies (e.g. Young 1982, Adjaloo et al. 2013), where effects on final yields were not measured, and the applicability to other locations is unclear.

The productivity of cacao may also be influenced by predator densities, in two main ways. First, the con-sumption and thus reduction of herbivorous insect pests on cacao by predators, may reduce herbivory and lead to increases in nutrient availability for increased fruit pro-duction (Wielgoss et al. 2014). In contrast, predators could potentially reduce crop productivity by deterring and/or consuming important pollinating species (Way and Khoo 1992). A variety of vertebrate and invertebrate predators are recognized in the literature, for their effi-ciency in the control of major cacao pests (Maas et al.

April 2017 889CONSERVING BENEFICIAL ANIMALS IN CACAO

2013, Wielgoss et al. 2014, Gras et al. 2016). For example, Gras et al. (2016) show that predatory ants provide effi-cient biological control against herbivorous insects in Indonesian cacao agroforestry, thereby leading an indirect increase in cacao yield. In Australian cacao systems, a ubiquitous predatory ant, the weaver ant Oecophylla smaragdina Fab., can play an important role in reducing pest densities (Forbes and Northfield, in press). However, this species is also known to inhibit the pollination services provided by a wide range of species in other crops (Rodriguez- Girones et al. 2013) and their potential influence on the pollination of cacao, remains unclear.

Here, we evaluated a single agricultural management practice designed to improve habitat availability for polli-nators, the addition of cacao fruit husks underneath cacao trees, on natural pollination services (fruit set), harvested fruit counts, and yield (fruit mass) in a commercial cacao farm in Australia. We crossed this habitat manipulation with artificial pollination treatments (hand pollination) to evaluate the effects of cacao fruit husk addition on polli-nation limitation. We also included a treatment that experimentally manipulated O. smaragdina ant presence and monitored each, natural fruit set, harvested fruit counts, and yield per tree in each treatment to measure any potential interactive effects between these predatory ants and other treatments on fruit set, harvested fruit counts, and yield. Finally, we evaluated the potential effect of cacao fruit husk addition on the conservation of two additional native predator groups (spiders and skinks) that may also contribute to the biological control of pests within the system, and evaluated the potential for reduc-tions in pollination services by habitat- mediated increases in the densities of these predator groups.

methodS

Site description

The study was conducted at the 10- yr- old Whyanbeel Valley Estate (145°21′ E, 16°22′ S) located in the Mossman region of North Queensland, Australia. The estate is a 1.8- hs monoculture cacao farm, planted with three self- compatible Trinitario SG2 hybrid cacao varieties of Papua New Guinean origin, organized using a rand-omized planting arrangement. The estate is surrounded predominantly by rainforest and scrubland vegetation of the Wet Tropics World Heritage Area. Fertilization, irri-gation, and herbicide application management practices on the estate are conducted on “as required” basis in response to regular monitoring. Foliar pruning of suckers and insecticide application are typically conducted on a fortnightly basis coinciding with fruit harvest, with the majority of pruned plant material being used as mulch underneath cacao trees. Insecticide application was dis-continued seven weeks prior to experimental initiation to avoid any potential negative impacts on midge polli-nators (Standfast et al. 2003).

Cacao fruit husk addition habitat manipulation

The effects of cacao fruit husk addition on pollination success (fruit set) and harvested fruit counts and yield (fruit mass) were evaluated over a seven- month period from November 2014 to June 2015 to include one of two annual peak flowering periods between December and January (S. J. Forbes, personal observation). Within the estate, 14 randomized experimental plots were estab-lished, each consisting of two adjacent rows of four, con-secutive, flower- bearing trees (Fig. 1A). The two outermost rows of the estate were not included in the experiment to reduce edge effects and no plots were directly adjacent to one another to reduce possible treatment effect spillover. Seven of the 14 plots were randomly selected for a “with cacao husk” treatment (Fig. 1B), where we placed approximately 280 kg (35 kg/ tree) of fresh cacao fruit husks left over from pro-cessing underneath all trees within the selected plots. The remaining seven plots were assigned the “no cacao husk” treatment, where they did not receive the addition of cacao fruit husks (Fig. 1C). The “with cacao husk” treated plots were spaced at a minimum of 15 m away from the “no cacao husk” treated plots to reduce the possibility of treatment effect spillover (Fig. 1A). The fresh cacao fruit husks were applied to the “with cacao husk” plots between 17 and 26 November 2014. A four- week adjustment period was then allowed for decomposition of the applied cacao fruit husks to begin before any surveys were con-ducted on experimental plots. During this adjustment period, all matured cacao fruits were removed from the experimental plots to minimize variation in initial condi-tions in tree physiology. A further 120 kg (15 kg/tree) of fresh cacao fruit husks were supplemented to each “with cacao husk” plot on 14 December 2014 and 15 January 2015, as the existing cacao fruit husks decomposed.

Flower counts

To determine flower production (availability) for sub-sequent evaluations of pollination success (fruit set), we counted the number of newly opened flowers per tree every 5 days between 1 December 2014 and 6 January 2015, coinciding with peak flowering and the decompo-sition of applied cacao fruit husks. All flower counts were conducted in the morning hours between 06:00 and 10:00 and flowers observed in the tree canopy above 2 m in height were inaccessible and were thus excluded from flower counts (as in Groeneveld et al. 2010).

Fruit set counts

To evaluate the influence of cacao fruit husk addition on fruit set, we counted the number of immature cacao fruits (cherelles) set by natural pollination (indicated by cherelle formation in the absence of pollen supplemen-tation described in Pollen supplementation, harvested fruit counts, and mass (yield)) on each experimental tree on 17


February 2015. This date was selected after all the cacao fruit husks had been added, allowing for the progression of cacao fruit husk decomposition that is presumed con-ducive to midge oviposition and larval development. As per flower counts, any cherelles observed developing within the tree canopy above 2 m in height were excluded from fruit set counts and were removed upon detection to minimize competition and compensation effects from non- experimental fruits (as in Groeneveld et al. 2010).

Pollen supplementation, harvested fruit counts and mass (yield)

To evaluate pollination limitation in the experimental trees, we applied two distinct pollen treatments (pollen supplementation or open pollination control) within each plot, replicated equally within the broader cacao fruit husk addition treatments detailed above, according to a split- plot design (Fig. 1B, C). To manipulate pollination availability, two trees per plot row were randomly selected

for either a “hand pollination” treatment, where pollen was supplemented (flowers not protected from insect pol-lination) to randomly selected, newly opened flowers by manual hand pollination (see Appendix S1 for details) (Fig. 1B, C). The remaining two trees per plot row were assigned an “open pollination” (control) treatment, where newly opened flowers were selected and marked in the same manner, but did not receive the supplemental pollen by manual hand pollination, rather, they remained open to the prospect of natural pollination by insects. Each flower treated was identified with a unique symbol marked with liquid paper (Newell Rubbermaid, Freeport, Illinois, USA) on the tree branch, directly above the base attachment of the flower pedicel. Pollen treatments were conducted once weekly between 1 December 2014 and 6 January 2015, with five flowers per tree randomly selected for the appropriate treatment on each weekly occasion, as well as an additional five flowers per tree selected on each weekly occasion to monitor natural pollination during the hand pollination period. Flowers developing in the tree canopy

Fig. 1. Diagrammatic representations of (A) example experimental plot placement and spacing within the greater Theobroma cacao plantation; (B) example experimental ant and pollen treatment design nested within plots assigned the “with cacao husk” treatment; and (C) example experimental ant and pollen treatment design nested within plots assigned the “No cacao husk” treatment. Individual experimental plots each comprise eight T. cacao trees, organized as two parallel rows of four consecutive trees. Within each plot, regardless of cacao fruit husk treatment (with cacao husk, no cacao husk), four differing ant (ants, no ants) by pollen (hand pollination, open pollination) treatment combinations were randomly assigned and replicated twice within each plot row. [Color figure can be viewed at wileyonlinelibrary.com]

http://wileyonlinelibrary.com


above 2 m in height were inaccessible, so they were not used for pollen treatments or monitoring (as in Groeneveld et al. 2010). All pollen treatments were conducted in the morning hours between 06:00 and 10:00. To reduce the effects of daily weather fluctuations on the success of manual pollen supplementation, under conditions of heavy rainfall or high wind, pollen treatment application was postponed to the next appropriate day. To account for variation in pollen viability and compatibility between the three Trinitario SG2 hybrid cacao varieties, a different cacao variety was used as the pollen donor on each weekly occasion of pollen treatment application. All matured cacao fruits, regardless of pollen treatment, were harvested fortnightly until 26 June 2015. Upon harvest, we measured the mass of each fruit using a hand- held 1 kg spring balance (Super Samson; Salter, Springvale, Victoria, Australia).

Oecophylla smaragdina manipulation

To measure the potential influence of predator presence on pollination success, fruit counts, and yield, we manip-ulated the densities of predatory O. smaragdina ants on cacao trees, whereby two trees per plot row were ran-domly assigned either an “ant presence” (ants) treatment, where ants were allowed to forage on trees, or an “ant exclusion” (no ants) treatment, where ants were physically excluded from trees (Fig. 1B, C). Thus, within each plot, there was a two- by- two factorial manipulation of each pollen treatment and ant manipulation treatment, giving four unique treatment combinations: (1) hand pollination and ant presence; (2) hand pollination and ant exclusion; (3) open pollination and ant presence; (4) open pollination and ant exclusion (Fig. 1B, C). Each of the four unique treatment combinations, nested within the two cacao fruit husk treatments at the plot level (Fig. 1B, C) were repli-cated twice (once on each of the two plot rows) for a total of 14 replicates of each of the four treatment combinations within each of the two cacao fruit husk treatments (7 plots per cacao fruit husk treatment × 2 trees per plot).

Oecophylla smaragdina ants were excluded from selected trees by physically removing the nests of any present ant colonies and by applying ant barriers around the trunks of cacao trees, preventing ant migrations up and down the trunk. Ant barriers involved a base layer of thick cotton batting (height: 15 cm; width: 2 cm) and a layer of black duct tape (50 mm) on top of the batting material. The duct tape was sealed over itself to secure the ant barrier around the trunk with moderate tightness. For trees assigned the “ant exclusion” treatment, a thick layer of Tangle- Trap (Tanglefoot; Contech, Victoria, British Columbia, Canada) was applied on top of the duct tape to completely encircle the tree trunk and inhibit ant move-ments into the tree. We pruned the canopy foliage on all experimental trees to disrupt canopy connectivity and inhibit the movement of ants between canopies of neigh-boring trees. The “ant presence” treatment was identical to the “ant exclusion” treatment in experimental design but did not receive the Tangle- Trap application, allowing

O. smaragdina to move up and down the tree trunk over the ant barriers. For each “ant presence” treated tree where O. smaragdina ants were not observed, we trans-planted an active O. smaragdina nest to the tree (as in Peng and Christian 2005). Ant barriers were applied to trees on 17 November 2014 and the Tangle- Trap appli-cation for ant exclusion treatments was applied on 19 November 2014.

Predator surveys

We examined the effects of cacao fruit husk addition on the densities of two groups of natural predators of arthropods, skinks and spiders, and evaluated the effects of any cacao fruit husk mediated variation in predator densities, on pollination success. Surveys to determine the densities of spiders and skinks were conducted fort-nightly from 11 December 2014 to 5 March 2015. Field surveys were conducted during the morning hours between 07:00 and 12:00. On each observation date, each tree within each experimental plot was visually surveyed for a period of 3 min. During the observation period, all aerial parts of experimental trees (foliage, branches, and trunk) were actively surveyed for spiders and skinks, and the leaf- litter layer underneath each tree was also sur-veyed for skinks. To reduce any possible effects of daily weather fluctuations on surveys (Read and Moseby 2001), under conditions of heavy rainfall or high wind, monitoring was postponed to the next appropriate day.

Statistical analysis

We used a generalized linear mixed effects model to evaluate the effect of cacao fruit husk addition and O. smaragdina presence on flower counts and fruit set (naturally pollinated cherelle counts). For the flower counts, we summed the flower counts across all sample dates prior to the fruit set count on 17 February 2015. As flowers typically abscise within 32 h after anthesis (Aneja et al. 1999), the repeated sampling of trees for flower counts was of no concern. We evaluated fruit set (the number of flowers set to fruit) per tree, using the fruit set counts collected on 17 February 2014. In the models, we included either flower count or fruit set as the response variable, and the independent variables included the treatments of cacao fruit husk (with cacao husks, no cacao husks) and O. smaragdina manipulation (ants, no ants), as well as the interaction between these two treat-ments, as fixed effects. To account for variation in flower counts and potential effects on fruit set, we included flower counts as a covariate in the fruit set (cherelles) analysis. Further, we included a parameter describing the variance between plots (i.e., a random plot effect) to account for greater covariance between trees within the same plot than trees in differing plots (Bolker et al. 2009). The models assumed a negative binomial distribution with a log- link function and were fit using Proc Glimmix in SAS (SAS Institute 2015).


To evaluate the effects of each cacao fruit husk addition, hand pollination, and O. smaragdina presence on the total number of harvested fruits per tree and fruit mass per tree, we used generalized linear mixed models in Proc Glimmix in SAS (SAS Institute 2015). Fixed effects included cacao fruit husk, pollen, and O. smaragdina manipulation treatments, as well as all potential interac-tions between these treatments. As per flower counts and fruit set, we included a parameter to describe the variance between plots (i.e., a random plot effect). Inclusion of flower counts as a covariate did not improve the fit of either model and reduced the pseudo- AICc value (Pseudo-AICc = -2l + 2d + 2dn /(n – d – 1), where l is the log pseudo-likelihood, d is the number of parameters, and n is the sample size) for each model. Thus, flower counts were not included in the final yield analyses. The pseudolikelihood information criteria were based on the pseudolikelihood estimation, and calculated in Proc Glimmix (SAS Institute 2015). The models each assumed a negative binomial distribution to account for overdis-persion in the count data, with a log- link function, and were fit using Proc Glimmix in SAS (SAS Institute 2015).

To evaluate the effects of cacao fruit husk addition on each spider and skink density, we used generalized linear mixed models describing each of the two response vari-ables. To account for autocorrelation in the residuals, we assumed that the residuals taken from consecutive meas-urements in the same plot have autocorrelation coeffi-cient. In other words, we assumed an autoregressive covariance structure describing covariance between sample dates for the same plot (Ives and Zhu 2006). In addition, we assumed that the densities of either spiders or skinks was dependent upon the cacao fruit husk treatment and varied with seasonal change. In order to model these additional factors, we included fixed effects describing cacao fruit husk treatment and sample date. Spider densities were modeled with a negative binomial error distribution to account for overdispersion in the counts, while skink densities were not overdispersed and were therefore modeled with a Poisson error distribution. In each case, we assumed a log- link function in the gen-eralized linear mixed model. Spiders and skinks were often observed utilizing the canopy and understory (ground) habitats of numerous trees for prey capture, and were therefore counted at the plot level. Because counts occurred at the plot level, we ignored within plot treatments, as all treatment combinations were applied to each plot. We conducted Wald F tests to conduct statis-tical inference tests on all fixed effects in all of the gener-alized linear mixed models described above. These analyses were conducted using Proc Glimmix in SAS (SAS Institute 2015). Because spiders often build webs between trees, and skinks were generally found foraging on the ground between trees, we evaluated the effects of predator density for each predator group, on the number of cherelles counted at the plot level. To account for plots with particularly high or low numbers of cherelles and predators (i.e., leverage points), we used robust regression

and mm- estimation in the MASS package in R (R Core Team 2015) to model the effects of each, spider and skink densities and cacao fruit husk treatments on the number of cherelles per plot.

reSultS

Flower and fruit set (cherelles) counts

There were a greater number of flowers per tree in the with- cacao- husk than in the no- cacao- husk plots (F = 4.11, df = 1, 96, P = 0.045; Table 1, Fig. 2A). After accounting for these differences in flower abundance between with- cacao- husk and no- cacao- husk plots, cacao fruit husk addition was found to significantly increase the relative number of cherelles per tree (F = 44.46, df = 1, 95, P < 0.001; Table 2, Fig. 2B). Oecophylla smaragdina presence did not alter the number of flowers, cherelles, or alter the effects of cacao fruit husks on either flowers or cherelles (O. smaragdina × cacao fruit husk interaction; Tables 1 and 2).

Harvested fruit counts and mass (yield)

The hand pollination pollen treatment significantly increased the number of mature fruits harvested per tree (F = 74.50, df = 1, 92, P < 0.001; Table 3, Fig. 3A, B). In addition, cacao fruit husk addition was also found to sig-nificantly increase the number of mature fruits harvested per tree (F = 27.10, df = 1, 92, P < 0.001; Table 3, Fig. 3A, B). There was a significant negative interaction between hand pollination and cacao fruit husk addition on the number of mature fruits per tree (F = 9.48, df = 1, 92, P = 0.003; Table 3, Fig. 3A, B), suggesting that the effects of hand pollination and cacao fruit husk addition were somewhat redundant when applied together. The hand pollination treatment significantly increased the harvested fruit mass per tree (F = 54.69, df = 1, 92, P < 0.001; Table 4, Fig. 3C, D). Further, cacao fruit husk addition also significantly increased the harvested fruit mass per tree (F = 80.67, df = 1, 92, P < 0.001; Table 4, Fig. 3C, D). As in the fruit count data, there was a signif-icant negative interaction between hand pollination and cacao fruit husk addition on harvested fruit mass (F = 9.48, df = 1, 92, P = 0.003; Table 3, Fig. 3C, D), again suggesting that the effects of hand pollination and cacao

table 1. Generalized linear mixed model results for flower counts.

EffectParameter estimate SE F P

Cacao husk 0.3715 0.1833 4.11 0.0454O. smaragdina 0.1022 0.09203 1.23 0.2694O. smaragdina ×

Cacao husk−0.1938 0.1841 1.11 0.2949

Notes: Fixed effects include cacao fruit husk addition and Oecophylla smaragdina presence and we included a random effect of plot. Degrees of freedom were 1, 96 for all F values.


fruit husk addition were somewhat redundant when applied together. Oecophylla smaragdina presence did not alter the number of harvested fruits or fruit mass, nor were there any significant interactions including O. smaragdina (Tables 1 and 2, Figs. 3 and 4).

Predator densities

Cacao fruit husk addition significantly increased both spider (F = 22.87, df = 1, 90, P < 0.001; Fig. 4A) and skink

(F = 19.36, df = 1, 90, P < 0.001; Fig. 4B) densities in the experimental plots. There was no significant change in the densities of spiders (F = 1.37, df = 6, 90, P = 0.234) or skinks (F = 0.89, df = 6, 90, P = 0.509) over time.

The effects of predators on fruit set (cherelles)

After accounting for cacao fruit husk addition, there was no effect of either spiders (t = 0.41, df = 10, P = 0.346; Fig. 4A) or skinks (t = 0.20, df = 10, P = 0.422; Fig. 4B) on the number of cherelles per plot.

diScuSSion

The addition of cacao fruit husks in our experiment significantly increased the number of cherelles per tree as well as the total number and mass of harvested fruits. Hand pollination also improved yields (harvested fruit counts and mass), but there was a significant negative interaction between hand pollination and cacao fruit husk addition. This negative interaction suggests that the two treatments are redundant when applied together and that cacao fruit husk addition serves to increase yields, at least in part, through enhancing pollination services.

Fig. 2. Mean log(x + 1)- transformed (±SE) counts of (A) flowers and (B) cherelles, in plots with (with cacao husk) and without (no cacao husk) cacao fruit husks added as mulch underneath trees, when subject to Oecophylla smaragdina ant presence (with ants) or exclusion (no ants) treatments. Flower counts were summed across nine sample dates between 1 December 2014 and 6 January 2015, and a single count of cherelles present per tree (fruit set) was conducted on 17 February 2015. Cacao fruit husks were applied underneath cacao trees between 17 and 26 November 2014, and supplemented on 15 December 2014 and 15 January 2015. Ant exclusion treatments were initiated on 19 November 2014. Data are log- transformed to reflect the log- link function used in the generalized linear mixed model analysis.

02

46

8

log(

[flow

ers/

tree]

+ 1

)

No cacao husk With cacao husk

A) Flowers

AntsNo ants 0

12

34

log(

[che

relle

s/tre

e] +

1)


B)

AntsNo ants

Fruit set

table 2. Generalized linear mixed model results for fruit set (cherelles).

EffectParameter estimate SE F P

Flowers 0.001972 0.000685 8.28 0.0050Cacao husk 2.0036 0.3005 44.46 <0.0001O. smaragdina 0.3309 0.1829 3.27 0.0736O. smaragdina ×

Cacao husk−0.3576 0.3662 0.95 0.3312

Notes: Fixed effects include cacao fruit husk addition and Oecophylla smaragdina presence. Flower counts were included as a covariate in the fruit set analysis and we included a random effect of plot. Degrees of freedom were 1, 95 for all F values.

table 3. Generalized linear mixed model results for harvested fruit counts.

Effect Parameter estimate SE F P

Cacao husk 0.6885 0.1323 27.10 <0.0001Hand pollination 1.1768 0.1323 74.50 <0.0001O. smaragdina 1.1416 0.1323 1.79 0.1847O. smaragdina × Cacao husk 0.4749 0.2645 3.22 0.0759O. smaragdina × Hand pollination −0.1175 0.2645 0.20 0.6580Cacao husk × Hand pollination −0.8144 0.2645 9.48 0.0027O. smaragdina × Cacao husk × Hand pollination −0.9569 0.5290 3.27 0.0738

Notes: Fixed effects include cacao fruit husk addition, hand pollination, and Oecophylla smaragdina presence and we included a random effect of plot. Degrees of freedom were 1, 92 for all F values.


Moreover, in light of previous studies on cacao polli-nation (Winder 1978, Young 1982), our findings suggest that the decomposing cacao fruit husk substrate mediated increases in the population densities of midges that pol-linate cacao, through increased habitat availability for oviposition and successful larval development, as com-pared to the no- cacao- husk control. Our results support the findings of Frimpong et al. (2011) who demonstrate a positive correlation between the population of ceratopo-gonid midges and fruit set in cacao.

Although adult ceratopogonid midges utilize floral resources (Bystrak and Wirth 1978), their larvae require moist and decomposing organic matter to provide suitable habitat for larval development and adequate food, such as yeast and bacteria present within decomposing matter (Saunders 1959, Bystrak and Wirth 1978, Soria et al. 1979). Indeed, emergence of arthropod fauna from in- laboratory emergence tents revealed an abundance of midges emerging from collected decomposing cacao fruit husks, as compared to the few midges that emerged from

Fig. 3. Mean log(x + 1)- transformed (±SE) harvested (A,B) fruit counts and (C, D) fruit yield per tree, harvested between 2 April 2015 and 26 June 2015, in plots with (with cacao husk) and without (no cacao husk) cacao fruit husks added as mulch, when subject to Oecophylla smaragdina ant presence (A, C) or exclusion (C, D) treatments. Trees either received supplemental pollen by hand pollination (hand pollination, solid line) or no pollen supplementation (open pollination, dashed line). Cacao fruit husks were applied underneath cacao trees between 17 and 26 November 2014, and supplemented on 15 December 2014 and 15 January 2015. Ant exclusion treatments were initiated on 19 November 2014 and weekly pollination treatments were conducted between 1 December 2014 and 6 January 2015. Data are log- transformed to reflect the log- link function used in the generalized linear mixed model analysis.

0.0

0.2

0.4

0.6

0.8

1.0

log(

[frui

ts/tr

ee] +

1)

A)

Hand pollinationOpen pollination

0.0

0.2

0.4

0.6

0.8

1.0

log(

[frui

ts/tr

ee] +

1)

B)


0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

log(

[frui

t bio

mas

s/tre

e] +

1)


C)


0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

log(

[frui

t bio

mas

s/tre

e] +

1)


D)


Ants No ants

No antsAnts

table 4. Generalized linear mixed model results for harvested fruit mass.

Effect Parameter estimate SE F P

Cacao husk 0.9523 0.1323 54.69 <0.0001Hand pollination 0.1607 0.1323 80.67 <0.0001O. smaragdina 1.1566 0.1323 1.56 0.2152O. smaragdina × Cacao husk 0.4392 0.2575 2.91 0.0915O. smaragdina × Hand pollination −0.1087 0.2575 0.18 0.6740Cacao husk × Hand pollination −0.8583 0.2575 11.11 0.0012O. smaragdina × Cacao husk × Hand pollination −0.8679 0.5151 2.84 0.0954

Notes Fixed effects include cacao fruit husk addition, hand pollination, and Oecophylla smaragdina presence and we included a random effect of plot. Degrees of freedom were 1, 92 for all F values.


collected cacao leaf- litter samples (S. J. Forbes, personal observation). The results thus support previous research that cacao pollinator populations depend upon moist and decomposing organic matter for reproduction (Winder 1978), and the removal of these substrates leads yield reductions (Winder 1977). This well documented corre-lation between cacao pollinators and the availability of suitable habitats has prompted recommendations for the maintenance and encouragement of important midge breeding sites for the stability of pollinator populations, particularly during dry conditions when natural abun-dances are low (Winder 1978, Mabbett 1989). However, promoting cacao pollinator oviposition substrates through the addition of decomposing cacao husks nearby cacao trees conflicts with current cultural control of important fungal pathogens of cacao in other cacao producing coun-tries, and thus the potential use of alternative oviposition substrates, such as banana pseudostems (Adjaloo et al. 2012) should be further evaluated.

In addition to changes in pollination rates, unex-pectedly, our study also showed a significant increase in the number of flowers per tree when cacao fruit husks were applied as a mulch underneath cacao trees, indi-cating a possible benefit of this habitat management practice on flower production and/or retention. Two mechanisms that are not mutually exclusive may explain these results: (1) trees with added cacao fruit husks produce more flowers or (2) flowers are retained longer on trees with added cacao fruit husks and were therefore more likely to be observed on our sample dates. For the former mechanism, the addition of cacao fruit husks may have facilitated increased nutrient availability (i.e., decomposing organic matter as a source of tree nutrition) and/or nutrient uptake (i.e., increased water availability for effective nutrient uptake), which generally improves flower production (Vaughton and Ramsey 1995). For the

latter mechanism, increased water availability, from and retained by the decomposing cacao fruit husks, may have caused increased flower retention. Indeed, previous research documents higher levels of flower abscission (drop) and water stress during the drier seasons when rainfall and relative humidity are low (Adjaloo et al. 2012, Frimpong- Anin et al. 2014). Having more flowers during “wetter” conditions when midge populations are higher (Winder 1978) may enhance fruit production, due to better synchrony of flower production and pollinator density (van Schaik et al. 1993). Therefore, there may be strong interactions between abiotic (e.g., seasonal micro-climate conditions and habitat availability) and biotic (synchronization between floral phenology of cacao and midge pollinator population dynamics) factors influ-encing cacao systems that may be manipulated to improve cacao pollination and fruit production.

The densities of two predator groups, spiders and skinks, were significantly higher in the experimental plots where cacao fruit husks were applied beneath trees. Spiders and skinks are generalist predators known to respond to the diversity and complexity of the local habitat (Diaz and Carrascal 1991, Balfour and Rypstra 1998). By enhancing habitat diversity and complexity, the development of robust spider and skink populations can be encouraged through the provision of a greater range of within- habitat features such as microclimates, retreat sites, and alternative prey sources (Rypstra et al. 1999, Pitt and Ritchie 2002). Given this body of liter-ature, the application of cacao fruit husks underneath cacao trees may be providing indirect benefits to native spider and skink predators by increasing the availability of alternative food resources and/or habitats for foraging and shelter. In support of this, a diverse group of insects and other arthropods have been observed emerging from field- collected decomposing cacao fruit husks when

Fig. 4. Mean (A) spider and (B) skink densities (averaged over 17 February and 5 March 2015) and corresponding number of immature fruits (cherelles) present per plot on 17 February 2015 in plots with (white circles) and without (black circles) cacao fruit husks added as mulch. Cacao fruit husks were applied underneath cacao trees between 17 and 26 November 2014, and supplemented on 15 December 2014 and 15 January 2015.

0 10 20 30 40

010

020

030

040

050

0

Spiders per plot

Che

relle

s pe

r plo

t

With cacao huskNo cacao husk

A)

0 1 2 3 4

Skinks per plot

B)

With cacao huskNo cacao husk

SkinksSpiders


brought back to the lab (S. J. Forbes, personal obser-vation). It is not clear if spiders and skinks are fulfilling important roles in the biological control of cacao pests, but they at least appear to have no negative impacts on the pollination services provided by midges. Furthermore, the manipulation of a common predator in cacao, O. smaragdina, known to disrupt pollination services in other plants (Rodriguez- Girones et al. 2013), did not influence pollination in our experimental cacao plots. This is in contrast to other ant species in other cacao systems where the presence of ants indirectly improves pollination (Wielgoss et al. 2014). Here, we found that predator conservation and promotion of pollination ser-vices are compatible goals in Australian cacao that can each be enhanced by a simple habitat manipulation: the addition of cacao fruit husks.

Habitat manipulation at the field or farm scale is a common practice that often provides beneficial insects with alternative resources and optimizes their perfor-mance (Landis et al. 2000). In other crops, similar par-allels exist between the management practices used to enhance the efficacy of each, pollinators and natural enemies of pests (Wratten et al. 2012). For example, the establishment of flower- rich habitat within or around intensively farmed agricultural landscapes, is a man-agement tool used to increase the availability of alter-native pollen and nectar resources for pollinators (Steffan- Dewenter and Tscharntke 2001, Kleijn et al. 2004) and parasitoids (Fiedler et al. 2008, Tompkins 2009), as well as providing alternative prey for predators (Landis et al. 2000). Although examples of distinct man-agement practices that simultaneously promote multiple ecosystem services are rare (Shackelford et al. 2013), the current study serves as an important example of the potential to develop simple and effective management practices for dual ecosystem service promotion.

Management practices common to intensified agricul-tural production often promote ecosystem “disservices” by reducing the quantity and quality of resources required to conserve native species (Power 2010). Thus, agricul-tural production and biological conservation have tradi-tionally been viewed as incompatible (Power 2010). However, appropriate, timely, and targeted management practices may alleviate some of the negative impacts of agriculture on native species, and in turn enhance the pro-vision of important ecosystem services (Tscharntke et al. 2005, Power 2010). Indeed, the current study represents one such example of how local, within- farm habitat man-agement employed to ameliorate the negative impacts of habitat simplification, supports the provisioning of polli-nation and predator conservation, leading to a win- win combination of agricultural production and biological conservation (Power 2010). Studies spanning multiple years may be useful in evaluating whether these win- win combinations can be sustained in the long term in cacao. In addition identification of other situations when such synergies are possible will assist the appropriate man-agement of agricultural croplands to benefit biological

conservation, while also profiting from the enhancement of ecosystem functioning through biodiversity (Altieri 1999, Tscharntke et al. 2005). The key to promoting these synergies is identifying the native species that contribute desirable ecosystem services for a given cropping system and developing appropriate and compatible management practices that conserve these species and their habitats.

acknowledgmentS

We thank the Puglisi family and Daintree Estates for access to the study site and necessary facilities and James Cook University for funding through a Cowan Grant Trust Scholarship to S. J. Forbes. We also thank S. Lambert and Mars, Inc. for support during the writing stage of this paper, as well as D. W. Crowder for helpful comments on an earlier draft of the manuscript.

literature cited

Adjaloo, M., B. Banful, and W. Oduro. 2013. Evaluation of breeding substrates for cocoa pollinator, Forcipomyia spp. and subsequent implications for yield in a tropical cocoa pro-duction system. American Journal of Plant Sciences 4:204–211.

Adjaloo, M., W. Oduro, and B. Banful. 2012. Floral phenology of upper Amazon cocoa trees: implications for reproduction and productivity of cocoa. ISRN Agronomy 2012:461674.

Aguilar, R., L. Ashworth, L. Galetto, and M. A. Aizen. 2006. Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta- analysis. Ecology Letters 9:968–980.

Allen-Wardell, G., et al. 1998. The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology 12:8–17.

Altieri, M. A. 1999. The ecological role of biodiversity in agroecosystems. Agriculture Ecosystems & Environment 74: 19–31.

Aneja, M., T. Gianfagna, and E. Ng. 1999. The roles of abscisic acid and ethylene in the abscission and senescence of cocoa flowers. Plant Growth Regulation 27:149–155.

Ashman, T. L., et al. 2004. Pollen limitation of plant reproduc-tion: ecological and evolutionary causes and consequences. Ecology 85:2408–2421.

Balfour, R. A., and A. L. Rypstra. 1998. The influence of habi-tat structure on spider density in a no- till soybean agroecosys-tem. Journal of Arachnology 26:221–226.

Bianchi, F. J. J. A., C. J. H. Booij, and T. Tscharntke. 2006. Sustainable pest regulation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest con-trol. Proceedings of the Royal Society B 273:1715–1727.

Biesmeijer, J. C., et al. 2006. Parallel declines in pollinators and insect- pollinated plants in Britain and The Netherlands. Science 313:351–354.

Bisseleua, D. H. B., A. D. Missoup, and S. Vidal. 2009. Biodiversity conservation, ecosystem functioning, and eco-nomic incentives under cocoa agroforestry intensification. Conservation Biology 23:1176–1184.

Bolker, B. M., M. E. Brooks, C. J. Clark, S. W. Geange, J. R. Poulsen, M. H. H. Stevens, and J. S. S. White. 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends in Ecology & Evolution 24:127–135.

Bond, W. J. 1994. Do mutualisms matter: assessing the impact of pollinator and disperser disruption on plant extinction. Philosophical Transactions of the Royal Society B 344: 83–90.


Bos, M. M., I. Steffan-Dewenter, and T. Tscharntke. 2007. Shade tree management affects fruit abortion, insect pests and pathogens of cacao. Agriculture Ecosystems & Environment 120:201–205.

Brittain, C., R. Bommarco, M. Vighi, S. Barmaz, J. Settele, and S. G. Potts. 2010. The impact of an insecticide on insect flower visitation and pollination in an agricultural landscape. Agricultural and Forest Entomology 12:259–266.

Buchmann, S. L., and G. P. Nabhan. 1996. The pollination cri-sis: the plight of the honey bee and the decline of other polli-nators imperils future harvests. Sciences 36:22–27.

Bystrak, P. G., and W. W. Wirth. 1978. The North American species of Forcipomyia, subgenus Euprojoannisia (Diptera: Ceratopogonidae). Technical Bulletin 1591. Department of Agriculture, Science and Education Administration, Washington, D.C., USA.

Cameron, S. A., J. D. Lozier, J. P. Strange, J. B. Koch, N. Cordes, L. F. Solter, and T. L. Griswold. 2011. Patterns of widespread decline in North American bumble bees. Proceedings of the National Academy of Sciences USA 108:662–667.

Classen, A., M. K. Peters, S. W. Ferger, M. Helbig-Bonitz, J. M. Schmack, G. Maassen, M. Schleuning, E. K. V. Kalko, K. Bohning-Gaese, and I. Steffan-Dewenter. 2014. Complementary ecosystem services provided by pest preda-tors and pollinators increase quantity and quality of coffee yields. Proceedings of the Royal Society B 281:20133148.

Clough, Y., H. Faust, and T. Tscharntke. 2009. Cacao boom and bust: sustainability of agroforests and opportunities for biodiversity conservation. Conservation Letters 2: 197–205.

Clough, Y., et al. 2011. Combining high biodiversity with high yields in tropical agroforests. Proceedings of the National Academy of Sciences USA 108:8311–8316.

Crowder, D. W., T. D. Northfield, R. Gomulkiewicz, and W. E. Snyder. 2012. Conserving and promoting evenness: organic farming and fire- based wildland management as case studies. Ecology 93:2001–2007.

Dale, V. H., and S. Polasky. 2007. Measures of the effects of agricultural practices on ecosystem services. Ecological Economics 64:286–296.

Deheuvels, O., G. X. Rousseau, G. S. Quiroga, M. D. Franco, R. Cerda, S. J. V. Mendoza, and E. Somarriba. 2014. Biodiversity is affected by changes in management intensity of cocoa- based agroforests. Agroforestry Systems 88: 1081–1099.

Diaz, J. A., and L. M. Carrascal. 1991. Regional distribution of a Mediterranean lizard: influence of habitat cues and prey abundance. Journal of Biogeography 18:291–297.

Diekotter, T., S. Wamser, V. Wolters, and K. Birkhofer. 2010. Landscape and management effects on structure and function of soil arthropod communities in winter wheat. Agriculture Ecosystems & Environment 137:108–112.

Edwin, J., and W. A. Masters. 2005. Genetic improvement and cocoa yields in Ghana. Experimental Agriculture 41: 491–503.

Fiedler, A. K., D. A. Landis, and S. D. Wratten. 2008. Maximizing ecosystem services from conservation biological control: the role of habitat management. Biological Control 45:254–271.

Forbes, S. J., and T. D. Northfield. in press. Oecophylla smarag-dina ants provide pest control in Australian cacao. Biotropica.

Frimpong, E. A., B. Gemmill-Herren, I. Gordon, and P. K. Kwapong. 2011. Dynamics of insect pollinators as influenced by cocoa production systems in Ghana. Journal of Pollination Ecology 5:74–80.

Frimpong, E. A., I. Gordon, P. K. Kwapong, and B. Gemmill-Herren. 2009. Dynamics of cocoa pollination: tools and

applications for surveying and monitoring cocoa pollinators. International Journal of Tropical Insect Science 29:62–69.

Frimpong-Anin, K., M. K. Adjaloo, P. K. Kwapong, and W. Oduro. 2014. Structure and stability of cocoa flowers and their response to pollination. Journal of Botany 2014:1–6.

Garibaldi, L. A., et al. 2011. Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecology Letters 14:1062–1072.

Garibaldi, L. A., et al. 2013. Wild pollinators enhance fruit set of crops regardless of honey bee abundance. Science 339: 1608–1611.

Gockowski, J., and D. Sonwa. 2011. Cocoa intensification sce-narios and their predicted impact on CO2 emissions, biodiver-sity conservation, and rural livelihoods in the guinea rain forest of West Africa. Environmental Management 48:307–321.

Gras, P., T. Tscharntke, B. Maas, A. Tjoa, A. Hafsah, and Y. Clough. 2016. How ants, birds and bats affect crop yield along shade gradients in tropical cacao agroforestry. Journal of Applied Ecology 53:953–963.

Groeneveld, J. H., T. Tscharntke, G. Moser, and Y. Clough. 2010. Experimental evidence for stronger cacao yield limita-tion by pollination than by plant resources. Perspectives in Plant Ecology Evolution and Systematics 12:183–191.

Ismail, A., and A. G. Ibrahim. 1986. The potential of ceratopo-gonid midges as insect pollinators of cocoa in Malaysia. Pages 471–484 in M. Y. Hussein and A. G. Ibrahim, editors. Biological control in the Tropics. proceedings of the First Regional Symposium on Biological Control, held at Universiti Pertanian Malaysia, Serdang from 4-6 September 1985, Universiti Pertanian, Malaysia.

Ives, A. R., and J. Zhu. 2006. Statistics for correlated data: phylogenies, space, and time. Ecological Applications 16: 20–32.

Kaufmann, T. 1975a. Ecology and behavior of cocoa pollinat-ing Ceratopogonidae in Ghana, West Africa. Environmental Entomology 4:347–351.

Kaufmann, T. 1975b. Studies on the ecology and biology of a cocoa pollinator, Forcipomyia squamipennis I. & M. (Diptera, Ceratopogonidae), in Ghana. Bulletin of Entomological Research 65:263–268.

Kearns, C. A., and D. S. Inouye. 1997. Pollinators, flowering plants, and conservation biology: much remains to be learned about pollinators and plants. BioScience 47:297–307.

Kennedy, C. M., et al. 2013. A global quantitative synthesis of local and landscape effects on wild bee pollinators in agroeco-systems. Ecology Letters 16:584–599.

Kleijn, D., F. Berendse, R. Smit, N. Gilissen, J. Smit, B. Brak, and R. Groeneveld. 2004. Ecological effectiveness of agri- environment schemes in different agricultural landscapes in the Netherlands. Conservation Biology 18:775–786.

Kleijn, D., et al. 2015. Delivery of crop pollination services is an insufficient argument for wild pollinator conservation. Nature Communications 6:7414.

Klein, A. M., B. E. Vaissiere, J. H. Cane, I. Steffan-Dewenter, S. A. Cunningham, C. Kremen, and T. Tscharntke. 2007. Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B 274:303–313.

Kremen, C., and A. Miles. 2012. Ecosystem services in biologi-cally diversified versus conventional farming systems: bene-fits, externalities, and tradeoffs. Ecology and Society 17:40.

Kremen, C., N. M. Williams, R. L. Bugg, J. P. Fay, and R. W. Thorp. 2004. The area requirements of an ecosystem service: crop pollination by native bee communities in California. Ecology Letters 7:1109–1119.

Landis, D. A., S. D. Wratten, and G. M. Gurr. 2000. Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology 45:175–201.


Lass, T. 2004. Balancing cocoa production and consumption. Pages 8–15 in J. Flood and R. Murphy, editors. Cocoa futures: a source book on some important issues facing the cocoa industry. CABI-FEDERACAFÉ, USDA, Chinchiná, Colombia.

Losey, J. E., and M. Vaughan. 2006. The economic value of ecological services provided by insects. BioScience 56: 311–323.

Maas, B., Y. Clough, and T. Tscharntke. 2013. Bats and birds increase crop yield in tropical agroforestry landscapes. Ecology Letters 16:1480–1487.

Mabbett, T. 1989. Midges the insect key to cocoa pollination. Cocoa and Coffee International 4:56.

Menz, M. H. M., R. D. Phillips, R. Winfree, C. Kremen, M. A. Aizen, S. D. Johnson, and K. W. Dixon. 2011. Reconnecting plants and pollinators: challenges in the restoration of polli-nation mutualisms. Trends in Plant Science 16:4–12.

M’Gonigle, L. K., L. C. Ponisio, K. Cutler, and C. Kremen. 2015. Habitat restoration promotes pollinator persistence and colonization in intensively managed agriculture. Ecological Applications 25:1557–1565.

Moguel, P., and V. M. Toledo. 1999. Biodiversity conservation in traditional coffee systems of Mexico. Conservation Biology 13:11–21.

Motamayor, J. C., A. M. Risterucci, P. A. Lopez, C. F. Ortiz, A. Moreno, and C. Lanaud. 2002. Cacao domestication I: the origin of the cacao cultivated by the Mayas. Heredity 89: 380–386.

Ollerton, J., R. Winfree, and S. Tarrant. 2011. How many flowering plants are pollinated by animals? Oikos 120: 321–326.

Peng, R. K., and K. Christian. 2005. Integrated pest manage-ment in mango orchards in the Northern Territory Australia, using the weaver ant, Oecophylla smaragdina, (Hymenoptera: Formicidae) as a key element. International Journal of Pest Management 51:149–155.

Pitt, W. C., and M. E. Ritchie. 2002. Influence of prey distribu-tion on the functional response of lizards. Oikos 96:157–163.

Potts, S. G., J. C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, and W. E. Kunin. 2010. Global pollinator declines: trends, impacts and drivers. Trends in Ecology & Evolution 25:345–353.

Power, A. G. 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philosophical Transactions of the Royal Society B 365:2959–2971.

Pywell, R. F., E. A. Warman, C. Carvell, T. H. Sparks, L. V. Dicks, D. Bennett, A. Wright, C. N. R. Critchley, and A. Sherwood. 2005. Providing foraging resources for bumble-bees in intensively farmed landscapes. Biological Conservation 121:479–494.

R Core Team. 2015. R: a language and environment for statisti-cal computing. R Foundation for Statistical Computing, Vienna, Austria. www.r-project.org

Rader, R., et al. 2016. Non- bee insects are important contribu-tors to global crop pollination. Proceedings of the National Academy of Sciences USA 113:146–151.

Read, J. L., and K. E. Moseby. 2001. Factors affecting pitfall capture rates of small ground vertebrates in arid South Australia. I. The influence of weather and moon phase on capture rates of reptiles. Wildlife Research 28:53–60.

Rice, R. A., and R. Greenberg. 2000. Cacao cultivation and the conservation of biological diversity. Ambio 29:167–173.

Ricketts, T. H., et al. 2008. Landscape effects on crop pollina-tion services: Are there general patterns? Ecology Letters 11:499–515.

Rodriguez-Girones, M. A., F. G. Gonzalvez, A. L. Llandres, R. T. Corlett, and L. Santamaria. 2013. Possible role of

weaver ants, Oecophylla smaragdina, in shaping plant- pollinator interactions in South- East Asia. Journal of Ecology 101:1000–1006.

Rypstra, A. L., P. E. Carter, R. A. Balfour, and S. D. Marshall. 1999. Architectural features of agricultural habitats and their impact on the spider inhabitants. Journal of Arachnology 27:371–377.

SAS. 2015. Version 9.4. SAS Institute, Cary, North Carolina, USA.

Saunders, L. G. 1959. Methods for studying Forcipomyia midges, with special reference to cacao- pollinating species (Diptera, Ceratopogonidae). Canadian Journal of Zoology 37:33–51.

van Schaik, C. P., J. W. Terborgh, and S. J. Wright. 1993. The phenology of tropical forests: adaptive significance and con-sequences for primary consumers. Annual Review of Ecology and Systematics 24:353–377.

Schroth, G., and C. A. Harvey. 2007. Biodiversity conservation in cocoa production landscapes: an overview. Biodiversity and Conservation 16:2237–2244.

Senapathi, D., J. C. Biesmeijer, T. D. Breeze, D. Kleijn, S. G. Potts, and L. G. Carvalheiro. 2015. Pollinator conservation: the difference between managing for pollination services and preserving pollinator diversity. Current Opinion in Insect Science 12:93–101.

Shackelford, G., P. R. Steward, T. G. Benton, W. E. Kunin, S. G. Potts, J. C. Biesmeijer, and S. M. Sait. 2013. Comparison of pollinators and natural enemies: a meta- analysis of land-scape and local effects on abundance and richness in crops. Biological Reviews 88:1002–1021.

Soria, S. D. J., W. W. Wirth, and H. A. Besemer. 1979. Breeding places and sites of collection of adults of Forcipomyia spp. midges (Diptera, Ceratopogonidae) in cacao plantations in Bahia, Brazil: a progress report. Revista Theobroma 8:21–29.

Stallman, H. R. 2011. Ecosystem services in agriculture: deter-mining suitability for provision by collective management. Ecological Economics 71:131–139.

Standfast, H., I. Fanning, L. Maloney, D. Purdie, and M. Brown. 2003. Field evaluation of Bistar 80SC as an effec-tive insecticide harbourage treatment for biting midges (Culicoides) and mosquitoes infesting peridomestic situations in an urban environment. Bulletin of the Mosquito Control Association of Australia 15:19–33.

Steffan-Dewenter, I., and T. Tscharntke. 2001. Succession of bee communities on fallows. Ecography 24:83–93.

Tepedino, V. J., and N. L. Stanton. 1981. Diversity and compe-tition in bee- plant communities on short- grass prairie. Oikos 36:35–44.

Tittonell, P. 2014. Ecological intensification of agriculture: sus-tainable by nature. Current Opinion in Environmental Sustainability 8:53–61.

Tompkins, J. M. L. 2009. Endemic New Zealand plants for pest management in vineyards. Pages 234–245 in Proceedings of the 3rd international symposium on biological control of arthropods, Christchurch, New Zealand. USDA Forest Service, Washington, D.C., USA.

Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan-Dewenter, and C. Thies. 2005. Landscape perspectives on agricultural intensification and biodiversity: ecosystem service manage-ment. Ecology Letters 8:857–874.

Vaast, P., and E. Somarriba. 2014. Trade- offs between crop intensification and ecosystem services: the role of agrofor-estry in cocoa cultivation. Agroforestry Systems 88:947–956.

Vaughton, G., M. Ramsey 1995. Pollinators and seed produc-tion. Pages 475–490 in J. Kigel and G. Galili, editors. Seed development and germination. Marcel Dekker, New York, New York, USA.

http://www.r-project.org


Way, M. J., and K. C. Khoo. 1992. Role of ants in pest- management. Annual Review of Entomology 37:479–503.

Weiner, C. N., M. Werner, K. E. Linsenmair, and N. Bluthgen. 2014. Land- use impacts on plant- pollinator networks: inter-action strength and specialization predict pollinator declines. Ecology 95:466–474.

Wielgoss, A., T. Tscharntke, A. Rumede, B. Fiala, H. Seidel, S. Shahabuddin, and Y. Clough. 2014. Interaction complex-ity matters: disentangling services and disservices of ant com-munities driving yield in tropical agroecosystems. Proceedings of the Royal Society B 281:20132144.

Williams, N. M., and C. Kremen. 2007. Resource distribu-tions among habitats determine solitary bee offspring pro-duction in a mosaic landscape. Ecological Applications 17: 910–921.

Winder, J. A. 1977. Field observations on Ceratopogonidae and other Diptera: Nematocera associated with cocoa flowers in Brazil. Bulletin of Entomological Research 67:57–63.

Winder, J. A. 1978. Cocoa flower Diptera: their identity, polli-nating activity and breeding sites. Pans 24:5–18.

Winder, J. A., and P. Silva. 1972. Cacao pollination: microdip-tera of cacao plantations and some of their breeding places. Bulletin of Entomological Research 61:651–655.

Wratten, S. D., M. Gillespie, A. Decourtye, E. Mader, and N. Desneux. 2012. Pollinator habitat enhancement: benefits to other ecosystem services. Agriculture Ecosystems & Environment 159:112–122.

Young, A. M. 1982. Effects of shade cover and availability of midge breeding sites on pollinating midge populations and fruit- set in 2 cocoa farms. Journal of Applied Ecology 19:47–63.

Supporting inFormation

Additional supporting information may be found in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/eap.1491/full

data availability

Data associated with this paper have been deposited in the James Cook University Tropical Data Hub: https://doi.org/10.4225/28/58460666db00c and https://doi.org/10.4225/28/58460fb90ea19

http://onlinelibrary.wiley.com/doi/10.1002/eap.1491/full

http://onlinelibrary.wiley.com/doi/10.1002/eap.1491/full

https://doi.org/10.4225/28/58460666db00c

https://doi.org/10.4225/28/58460666db00c

https://doi.org/10.4225/28/58460fb90ea19

increased pollinator habitat enhances cacao fruit … · increased pollinator habitat enhances...

Documents