Abstract
As key dispersers of herbaceous seeds, Aphaenogaster ants strongly influence the distribution of woodland plants in eastern North America. Ants within this genus are difficult to distinguish and often are identified by subgroup, but emerging research suggests they occupy species-specific ecological niches. As such, distinct climatic requirements among Aphaenogaster spp. might result in transient plant interactions with climate change. We examine whether there are ecological and distributional differences among Aphaenogaster species that coincide with current taxonomic differentiations. We use occurrence records for six Aphaenogaster spp. that occur in deciduous forests in eastern North America. We associate the geographic patterning of species occurrence with temperature and precipitation data, and we examine whether unique climatic niches characterize each species. We then predict habitat suitability throughout eastern North America using species distribution models. For verification, we test how well the predicted ranges fit observed occurrences using novel data sets for each species. We find that Aphaenogaster species within this cryptic genus demonstrate unique ecological and geographic signatures. Each species within the subgroup generally responds differently to temperature, and somewhat differently to precipitation and seasonal variance, suggesting unique ecological niches for each species. Our results indicate that each ant species may respond uniquely to changes in climate. Such shifts could disrupt current community associations and biotic interactions with ant-dispersed plants.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Ant-mediated seed dispersal (myrmecochory) occurs worldwide (Gorb and Gorb, 2003; Rico-Gray and Oliveira, 2007), and it is a major facilitative interaction in understory woodland communities in eastern North America (N.A.) with direct influence on a substantial part of plant diversity (Beattie and Hughes, 2002; Ness et al., 2009). Myrmecochorous plants benefit from ant-mediated dispersal through seed predator avoidance, decreased intraspecific competition and placement in nutrient-rich environments (see Gorb and Gorb, 2003; Giladi, 2006; Rico-Gray and Oliveira, 2007 and references therein), and the ants benefit by receiving nutrition for larvae from lipid-rich seed appendages (Carroll and Janzen, 1973; Marshall et al., 1979; Morales and Heithaus, 1998; Gammans et al., 2005). The benefits of this interaction are mutual, but the plants depend far more on the ants than vice versa, making the interaction facultative for ants and obligate for plants (Pudlo et al., 1980; Ness et al., 2009; Warren et al., 2010).
Myrmecochory long was considered a diffuse mutualism between a host of scavenging ants and spring-flowering plants (Berg, 1966; Handel, 1976; Beattie et al., 1979; Beattie and Hughes, 2002; Garrido et al., 2002), but researchers increasingly identify specific ant genera as dominant specialists in myrmecochorous interactions worldwide (Anderson, 1988; Espadaler and Gomez, 1996; Garrido et al., 2002; Gorb and Gorb, 2003; Rico-Gray and Oliveira, 2007). Indeed, the timing of propagule release may correspond with peak foraging in certain seed-dispersing ant genera and these genera appear far more effective as seed dispersers than other seed-gathering ants (Giladi, 2006; Boulay et al., 2007; Gove et al., 2007; Zelikova et al., 2008; Bas et al., 2009; Manzaneda and Rey, 2009; Ness et al., 2009). In eastern deciduous forests (USA), members of the ant genus Aphaenogaster dominate these mutualist interactions (Ness et al., 2009).
The Aphaenogaster genus is cosmopolitan (Bolton, 2010) with approximately a dozen species occurring in eastern N.A. (Creighton, 1950; Smith, 1979; Umphrey, 1996). Aphaenogaster spp. colonies typically are found below rocks and logs in mesic deciduous forests in the eastern USA (Talbot, 1934; Beattie, 1978; Lynch et al., 1980; Smallwood, 1982a). Given the high diversity of ant species worldwide, there are many challenges in species identification and taxonomic revisions are common. These considerations make genus-based ant classification a relatively useful way to approach ant research (Fisher and Cover, 2007; Ward, 2007; Bolton, 2010; Guénard et al., 2010). In addition, the taxonomic status of several Aphaenogaster species in the eastern USA remains unsettled, and members of this complex can be difficult to distinguish (Creighton, 1950; Umphrey, 1996). For these reasons, workers studying ant-mediated seed dispersal generally identify Aphaenogaster spp. to genus, particularly members of a species/sub-species known as the fulva-rudis-texana complex (hereafter ‘rudis complex’), which also contains the majority of seed-dispersing species (see Ness et al., 2009 and references therein).
The rudis complex includes six identified species in eastern N.A. (Umphrey, 1996). For Aphaenogaster species, and ants in general, habitat partitioning appears to be driven by temperature (Brian, 1956; Bernstein, 1979; Lynch et al., 1980; Smallwood, 1982a; Cerdà et al., 1997; Retana and Cerda, 2000; Dunn et al., 2007; Sanders et al., 2007), but moisture also plays a role (Cerdà et al., 1997; Warren et al., 2010). Whereas spatial and temporal habitat partitioning is common among ant species (see Parr and Gibb, 2010), little is known about species partitioning within genera (e.g., Brian, 1956). Workers have demonstrated segregation between Aphaenogaster spp. and sympatric ant genera (e.g., Lynch et al., 1980; Fellers, 1987), but only limited, mostly anecdotal, information exists surrounding habitat segregation by species within the rudis complex (Talbot, 1934; Crozier, 1977; Mitchell et al., 2002). Previous taxonomic work on the rudis complex has differentiated species by morphology and genetics (Creighton, 1950; Crozier, 1977; Umphrey, 1996), but the species ranges are not well established and ecological differentiations are relatively unknown. As a result, most ecological research involving the rudis complex, particularly seed-dispersal studies, makes no discrimination among members within this subgroup (also noted by Ness et al., 2009). However, Warren et al. (2011) demonstrated that differences exist in the foraging phenology of two members of the rudis complex, suggesting species-specific ecological roles in seed dispersal. Aphaenogaster picea began foraging early enough in the spring to overlap with seed set by all local myrmecochorous plants, whereas A. rudis began foraging too late in the spring to disperse early flowering species (Warren et al., 2011).
If individual Aphaenogaster spp. respond dissimilarly to rapid climate change, Warren et al. (2011)’s results indicate that the ant–plant mutualism may become disrupted. Myrmecochores comprise a large portion of understory plant diversity, and a potential disruption of the ant mutualism would entail great consequence for forest diversity. It is unfortunate, then, that little is known about how climate impacts the current distributions of individual Aphaenogaster spp. We explore the climatic niches of individual Aphaenogaster ant species to examine ecological and distributional differences among Aphaenogaster species within the rudis complex that coincide with their current taxonomic status. We associate the geographic patterning of species occurrence (from collection localities) with temperature and precipitation data, and we examine whether unique climate associations occur for each species. We then use these climatic niche models to predict habitat suitability throughout eastern N.A. using species distribution models (SDMs). For verification, we test how well the predicted ranges fit observed occurrences using novel occurrence data sets. In doing so, we generate probability distribution maps for each species and, more importantly, assess whether members of the rudis complex are ecologically equivalent or require greater effort for species-specific ecological investigations. These insights will help elucidate the interactions between the ants and the plants, the seeds of which they disperse, and how these relationships may vary under climate change.
Methods
Records of Aphaenogaster species and climate variables were obtained from published literature (Umphrey, 1996) and online databases to investigate the climatic niche and create species distribution models (SDMs) for each member of the rudis complex. Climate layers were obtained from the WorldClim Version 1.4 data set (www.worldclim.org). WorldClim provides gridded high-resolution monthly temperature and precipitation data and derivate variables that are thought to be biologically significant, such as means, seasonality and extremes (Hijmans et al., 2005). The climate data are interpolated from weather stations worldwide and averaged for the years 1950–2000. We downloaded the data in a 2.5 arc-minute (4.65 km2) generic grid format. We used this scale instead of a finer resolution to reflect the probable scale of error in digitizing Umphrey’s (1996) data. Because the targeted study species are not known to occur outside of the eastern portions of the USA and Canada, a mask of eastern North America climate layers limited the spatial data to a target area from Florida to Nova Scotia, North Carolina to Arizona (e.g., Peterson et al., 2007; but see Phillips, 2008; and Van Der Wal et al., 2009). The data were converted to formats conducive for spatial analysis using ArcMap 9.3 (www.esri.com) and DIVI-GIS 7.2.1 9 (www.diva-gis.org).
Worldclim provides 19 climate layers derived from temperature and precipitation records. Because collinearity among climate variables undermines the reliability and interpretation of model output, we examined collinearity between the selected climate variables using a Pearson correlation matrix and a principal component analysis (PCA) using the “prncomp” method and “scale” option (standardizes all variables to unit length) in the “R” statistical package (R Development Core Team 2005). The PCA axes could be used for bioclimatic modeling, but we preferred to use a sensible selection of the original variables for ease of interpretation and prediction. Based on the correlation matrix and PCA, we selected a suite of basic climate variables that is a compromise between capturing most information contained in the 19 original variables, parsimony and ease of interpretation: BIO1—annual mean temperature (Tavg); BIO4—temperature seasonality (standard deviation × 100, Tvar); BIO5—maximum temperature of the warmest month (Tmax); BIO6—minimum temperature of the coolest month (Tmin); BIO12—annual precipitation (Tavg); BIO13—precipitation of the wettest month (Pmax); BIO14—precipitation of the driest month (Pmin); and BIO15—precipitation seasonality (coefficient of variation, Pvar) for bioclimatic analysis. The PCA analysis indicated that Tavg and Tmin were essentially identical (Fig. 1), so Tmin was omitted from the next step in the analysis.
As is often the case with species data (Phillips et al., 2004), we only had access to Aphaenogaster presence data, and the individual species sample sizes were small (Table 1). Therefore, we used the maximum entropy algorithm, Maxent 3.3.3a (Phillips et al., 2006), which has been demonstrated as a relatively robust approach for SDM from small presence-only data sets (Phillips et al., 2004; Elith et al., 2006; Hernandez et al., 2006; Phillips et al., 2006; Pearson et al., 2007; Wisz et al., 2008). Maxent is a general purpose method that estimates probability distributions from incomplete information by finding the probability distribution closest to maximum entropy (closest to uniform) constrained only by associations between species presences and predictor variables (see Phillips et al., 2004; Phillips et al., 2006 for full explanations). The logistic output gives a spatially explicit suitability value ranging from 0 (unsuitable habitat) to 1 (optimal habitat) (Phillips and Dudik, 2008). We used the recommended default settings for Maxent to examine potential species distributions and associated climate drivers.
In the absence of independent validation data, the best method is to build models on one portion of the data (‘training data’) while withholding a portion of the same data for validation (‘testing data’) (Araujo et al., 2005; Phillips et al., 2006; Phillips and Dudik, 2008). A better approach, especially in cases where data are limited, is to train and test the models with independent data sets (Elith et al., 2006; Pearson et al., 2007). We derived training data from Umphrey (1996) for six Aphaenogaster species within the rudis complex: A. carolinensis Wheeler, A. fulva Roger, A. miamiana Wheeler, A. picea Wheeler, A. rudis Enzmann and A. texana Wheeler. Umphrey (1996) collected Aphaenogaster spp. at >76 locations throughout the eastern N.A. to generate the most recent and comprehensive taxonomic revision of the rudis complex. We georeferenced the collection localities by associating spatial directions given in the manuscript with digitized maps to generate GPS coordinates. We then generated an independent testing data set by obtaining Aphaenogaster spp. data reported (along with GPS coordinates) via AntWeb (www.antweb.org), Antbase (www.antbase.org) and Discover Life (www.discoverlife.org). These databases contain records from individual researchers and research institutions identified by species and GPS coordinates for collection location. We evaluated the performance of the species distribution models using independent location data and the area under receiver operator characteristic (ROC) curves (AUC). The AUC gives the probability that a randomly chosen presence site is predicted to be a better fit by the model than a randomly chosen absence site (Phillips and Dudik, 2008). As such, a random fit would have an AUC = 0.5, and a perfect fit would result in AUC = 1.0. When using Maxent, which is based on presence-only data, background data (“pseudo-absence”) are sampled at random locations from the study area. In our models, AUC represents the fit of the testing data to the model generated by the training data, which are independent data sets. The AUC can be inflated, however, by species with narrow distribution ranges relative to the study area (Phillips, 1996). The final models were used to create habitat suitability maps of the predicted geographic distributions.
Results
Principal component analysis
Principal component analysis on the full set of 19 bioclimatic variables indicated that seven principal components can account for 97% of the variability in the 19 variables. Whereas our selected variables are not fully orthogonal and thus cannot account for exactly 97% of the variability in the full climate data set, we are confident that they represent the contained information very well and in a much more parsimonious and less intercorrelated way. Plotting the selected variables on a graph of the first two principal components using the seven selected variables shows that the variables are well selected in terms of maximizing the information (the first two PCA axes explain 88% of the climate variability) covered minimizing correlation (Fig. 1). The ant species, except A. carolinensis and A. fulva, sort apart distinctly along the PC1 axis, which explains 56% of the climate variability and mostly represents temperature and its temporal variance. Most of the species did not sort out as well along the PC2 axis, which explains 32% of the variability and mostly represents annual precipitation and its temporal variance, with only A. texana inhabiting unique precipitation space and A. picea somewhat different from the rest. The most similar species in abiotic associations were A. carolinensis and A. fulva, which were almost identical along both PCA axes. Aphaenogaster texana was the most distinct and was differentiated most from the others by precipitation.
Maxent algorithm
The AUC values for all species models except one were >0.900, with the exception being 0.855, making all models adequate for interpretation (Table 1). The probability of occurrence for the rudis complex species typically peaked around 54–77% across eastern North America (Fig. 2), and the highest probabilities (>90%) only occurred at the southern tip of Florida for A. miamiana (Fig. 2a) and at the highest elevations of the Appalachian Mountains for A. picea (Fig. 2b). Aphaenogaster rudis (Fig. 2c) and A. picea had similar ranges in northeastern N.A. from the Atlantic Coast to the Great Lakes, but A. picea was projected to occur at more northern extremes and at high elevations in the Appalachian Mountains, whereas A. rudis was projected most likely to occur in the lower elevations surrounding the Appalachian Mountains. Aphaenogaster fulva (Fig. 2d) and A. carolinensis (Fig. 2e) were projected to occur throughout eastern N.A. from the Gulf Coast to Nova Scotia, east of the Mississippi River, though the A. carolinensis range did not extend as far northwest in the Great Lakes region or throughout southern Florida. Aphaenogaster texana was projected to only occur with substantial probability in the Texas–Oklahoma region, though this is truncated on the western edge by the projection range (Fig. 2f).
In all models, the temperature variables generally were the most important component of AUC (percent contribution), but some precipitation variables also weighed heavy as predictors of species distributions (Table 1). Mean temperature best predicted A. fulva and A. rudis distributions, although minimum precipitation was almost as important for A. fulva and precipitation variability between seasons was almost as important for A. rudis. Maximum temperature was the best predictor of A. texana distribution, by far, with minimum precipitation also considerable. Temperature variability between seasons and average precipitation contributed relatively little to any of the species models. None of the species distributions corresponded at all with maximum precipitation except A. miamiana, which was best predicted by maximum precipitation. Seasonal variation in precipitation also predicted A. picea and A. rudis distributions.
Discussion
We used PCA analysis to infer large-scale climate niche requirements for six Aphaenogaster ant species and SDMs to validate the climate associations and geographic distributions. At this coarse scale, the climatic variables were capable of accurately predicting the distribution of these species as verified by independent collection data. We use this framework to show that Aphaenogaster species within the cryptic rudis complex demonstrate unique ecological and geographic signatures as all six species generally respond to different features and limits of temperature and precipitation. Aphaenogaster texana and miamiana exhibit very distinct geographic ranges and have little overlap with the other species, whereas A. fulva and A. carolinensis appear to have considerable overlap in ecology and distribution (Figs 1, 2). Aphaenogaster rudis and picea overlap with A. fulva and A. carolinensis in the northeastern portion of N.A., but appear to have inverse distributions with A. picea, which has habitat requirements that appear more northerly and at higher elevations than A. rudis. Taxonomic distance does not appear equivalent to ecologic distance as Umphrey (1996) identified A. fulva as the most morphologically and genetically different from others in the rudis complex, yet it is almost identical in climate associations and projected range with A. carolinensis.
Whereas we only used seven distal (sensu Austin, 2002) variables, namely temperature, precipitation and their variability over time and space, to evaluate and predict Aphaenogaster spp. habitat suitability, we believe this approach was useful for two reasons: (1) a large body of literature on ant distributions and assemblies indicates that temperature and, secondarily, precipitation are the key environmental factors (Lynch et al., 1980; Smallwood, 1982a, b; Herbers, 1985; Fellers, 1987; Fenner, 1987; Hölldobler and Wilson, 1990; Mitchell et al., 2002; Sanders et al., 2007; Warren et al., 2010) and (2) employing only a few factors in niche-type investigations is most useful for hypothesis formation and testing and, if the factors are appropriate, simplifies the analysis and interpretation to the core limiting requirements (Chase and Leibold, 2003). The sample sizes are large enough to examine ecological and geographic differences across the target species (see Pearson et al., 2007; Wisz et al., 2008), but we feel they are not large enough to choose a climate projection and make predictions about potential Aphaenogaster spp. distributions under warming scenarios.
These models should be considered as projecting geographic regions with habitat similar to regions where the species occur rather than as delineating the actual range limits for each species (see Pearson et al., 2007). The collection and assembly of more data linking the distribution and ecology of these species with relevant habitat parameters would be useful to advance such projections. As such, there is high similarity between our projections and previous range estimates for members of the rudis complex (Creighton, 1950; Umphrey, 1996), though our projections place A. carolinensis further inland than previous reports and A. texana is projected, with very low probability, further eastward (Fig. 2). It is interesting to note that these two species once were considered the same species with a range that included both current distributions. An additional consideration is that, because Aphaenogaster is a cryptic genus, collection location could influence collectors in species identification. Whereas this bias would reduce the independence of the test data sets, it is likely most relevant to A. carolinensis and A. texana. We suggest searching Antweb (www.antweb.org), Antbase (www.antbase.org) and Discover Life (www.discoverlife.org) for further information on collecting sites and collectors responsible for species identification.
Crozier (1977) found shifts between sympatric rudis complex species—though these were not identified by accepted species names or those delineated by Umphrey (1996)—within 1 km along elevation gradients in the Appalachian Mountain foothills of north Georgia, USA. In such mountainous terrain, shifts in temperature and moisture equivalent to 100 km of latitude can occur across 1 km of elevation (Warren, 2008; Warren, 2010a; Warren, 2010b). Microhabitat ecological differences among species in the rudis complex cannot be captured at the scale measured here, but the demonstrated associations with unique portions of broad-scale temperature and precipitation gradients also may occur at the microscale, particularly where the species distributions overlap. Such climate segregation likely influences their role and effectiveness as seed dispersers and, ultimately, as facilitators of woodland plants. At macroscales, our data suggest that each species (except possibly A. fulva and A. carolinensis) may respond uniquely to changes in climate. Seed set phenology differs among ant-dispersed plant species (e.g., Radford et al., 1968) and foraging phenology differs between at least two species in the rudis complex (Warren et al., 2011). Species-specific climate responses could result in the disruption of current community associations (Root et al., 2003; Williams and Jackson, 2007) and co-evolved mutualisms (Leathwick and Austin, 2001; Parmesan, 2007; Brook, 2009; Cavender-Bares, 2009). Warren et al. (2011) found that A. picea and A. rudis exhibited differing foraging phenologies, so that A. rudis activity did not overlap with early fruiting plant species. We demonstrate here that A. picea and A. rudis are not ecologically equivalent so that shifts in their distributions could result in asynchronous phenology with early blooming plant species if, for example, A. rudis replaces A. picea.
References
Anderson A. N. 1988. Dispersal distance as a benefit of mymecochory. Oecologia 75: 507-511
Araujo M. B., Pearson R. G., Thuiller W. and Erhard M. 2005. Validation of species–climate envelope models under climate change. Global Change Biol. 11: 1504-1513
Austin M. 2002. Case studies in the use of environmental gradients in vegetation and fauna modeling: theory and practice in Australia and New Zealand. In: Predicting Species Occurrences: Issues of Accuracy and Scale (Scott J.M., Heglund P.J. and Morrison M.L., Eds), Island Press, Washington, D.C. pp 64-72
Bas J. M., Oliveras J. and Gomez C. 2009. Myrmecochory and short-term seed fate in Rhamnus alaternus: Ant species and seed characteristics. Acta Oecol. 35: 380-384
Beattie A. J. 1978. Plant-animal interactions affecting gene flow in Viola. In: The Pollination of Flowers by Insects (Richards A.J., Ed), Academic Press, London. pp 151-164
Beattie A. J., Culver D. C. and Pudlo R. J. 1979. Interactions between ants and the diaspores of some common spring flowering herbs in West Virginia. Castanea 3: 177-186
Beattie A. J. and Hughes L. 2002. Ant-plant interactions. In: Plant-Animal Interactions: an Evolutionary Approach (Herrera C.M. and Pellmyr O., Eds), Blackwell Science, Oxford. pp 211-235
Berg R. Y. 1966. Seed dispersal of Dendromecon: its ecologic, evolutionary, and taxonomic significance. Am. J. Bot. 53: 61-73
Bernstein R. A. 1979. Schedules of foraging activity in species of ants. Oecologia 48: 921-930
Bolton B. 2010. Identification Guide to the Ant Genera of the World. Harvard University Press, Cambridge, MA. 232 pp
Boulay R., Carro F., Soriguer R. C. and Cerda X. 2007. Synchrony between fruit maturation and effective dispersers’ foraging activity increases seed protection against seed predators. Proc. R. Soc. B-Biol. Sci. 274: 2515-2522
Brian M. V. 1956. Segregation of species of the ant genus Myrmica. J. Anim. Ecol. 25: 319-337
Brook B. W. 2009. Global warming tugs at trophic interactions. J. Anim. Ecol. 78: 1-3
Carroll C. R. and Janzen D. H. 1973. The ecology of foraging by ants. Annu. Rev. Ecol. Syst. 4: 231-258
Cavender-Bares J. 2009. The merging of community ecology and phylogenetic biology. Ecol. Lett. 12: 693-715
Cerdà X., Retana J. and Cros S. 1997. Thermal disruption of transitive hierarchies in Mediterranean ant communities. J. Anim. Ecol. 66: 363-374
Chase J. M. and Leibold M. A. 2003. Ecological Niches: Linking Classical and Contemporary Approaches. University of Chicago, Chicago. 180 pp
Creighton W. S. 1950. The Ants of North America, vol 104. Bulletin of the Museum of Comparative Zoology. The Cosmos Press, Inc., Cambridge, MA, USA. 585 pp
Crozier R. H. 1977. Genetic differentiation between populations of the ant Aphaenogaster ‘rudis’ in the southeastern United States. Genetica 47: 17-36
Dunn R. R., Parker C. R. and Sanders N. J. 2007. Temporal patterns of diversity: Assessing the biotic and abiotic controls on ant assemblages. Biol. J. Linn. Soc. 91: 191-201
Elith J., Graham C. H., Anderson R. P., Dudik M., Ferrier S., Guisan A., Hijmans R. J., Huettmann F., Leathwick J. R., Lehman A., Lucia J. L., Lohmann L. G., Loisell B. A., Manion G., Moritz C., Nakamura M., Nakazawa Y., Overton J. M., Peterson A. T., Phillips S. J., Richardson K., Scachetti-Pereira R., Schapire R. E., Soberon J., Williams S., Wisz M. S. and Zimmermann N. E. 2006. Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29: 129-151
Espadaler X. and Gomez C. 1996. Seed production, predation and dispersal in the mediterranean myrmecochore Euphorbia characias. Ecography 19: 7-15
Fellers J. H. 1987. Interference and exploitations in a guild of woodland ants. Ecology 68: 1466-1478
Fenner M. 1987. Seedlings. New Phytol. 106: 35-47
Fisher B. L. and Cover S. P. 2007. Ants of North America. University of California Press, Berkeley and Los Angeles. 216 pp
Gammans N., Bullock J. J. and Schonrogge K. 2005. Ant benefits in a seed dispersal mutualism. Oecologia 146: 43-49
Garrido J. L., Rey P. J., Cerda X. and Herrera C. M. 2002. Geographical variation in diaspore traits of an ant-dispersed plant (Helleborus foetidus): are ant community composition and diaspore traits correlated? J. Ecol. 90: 446-455
Giladi I. 2006. Choosing benefits or partners: a review of the evidence for the evolution of myrmecochory. Oikos 112: 481-492
Gorb E. V. and Gorb S. N. 2003. Seed Dispersal by Ants in a Deciduous Forest Ecosystem. Kluwer Academic Publishers, Dordrecht, The Netherlands. 242 pp
Gove A. D., Majer J. D. and Dunn B. 2007. A keystone ant species promotes seed dispersal in “diffuse” mutualism. Oecologia 153: 687-697
Guénard B., Weiser M. D. and Dunn R. R. 2010. Ant genera of the world. http://www.antmacroecology.org/ant_genera/index.html
Handel S. N. 1976. Ecology of Carex pedunculata (Cyperaceae), a new North American myrmecochore. Am. J. Bot. 63: 1071-1079
Herbers J. M. 1985. Seasonal structuring of a north temperate ant community. Insect. Soc 32: 224-240
Hernandez P. A., Graham C. H., Master L. L. and Albert D. L. 2006. The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography 29: 773-785
Hijmans R. J., Cameron S. E., Parra J. L., Jones P. G. and Jarvis A. 2005. Very high resolution interpolated climate surfaces for global land areas. Int. J. of Climatol. 25: 1965-1978
Hölldobler B. and Wilson E. O. 1990. The Ants. Belknap, Cambridge, Mass. 732 pp
Leathwick J. R. and Austin M. P. 2001. Competitive interactions between tree species in New Zealand’s old-growth indigenous forests. Ecology 82: 2560-2573
Lynch J. F., Balinsky E. C. and Vail S. G. 1980. Foraging patterns in three sympatric forest ant species, Prenolepis imparis, Paratrechina melanderi and Aphaenogaster rudis (Hymenoptera, Formicidae). Ecol. Entomol. 5: 353-371
Manzaneda A. J. and Rey P. J. 2009. Assessing ecological specialization of an ant-seed dispersal mutualism through a wide geographic range. Ecology 90: 3009-3022
Marshall D. L., Beattie A. J. and Bollenbacher W. E. 1979. Evidence for diglycerides as attractants in an ant–seed interaction. J. Chem. Ecol. 5: 335-344
Mitchell C. E., Turner M. G. and Pearson S. M. 2002. Effects of historical land use and forest patch size on myrmecochores and ant communities. Ecol. Appl. 12: 1364-1377
Morales M. A. and Heithaus E. R. 1998. Food from seed-dispersal mutualism shifts sex ratios in colonies of the ant Aphaenogaster rudis. Ecology 79: 734-739
Ness J. H., Morin D. F. and Giladi I. 2009. Uncommon specialization in a mutualism between a temperate herbaceous plant guild and an ant: are Aphaenogaster ants keystone mutualists? Oikos 12: 1793-1804
Parmesan C. 2007. Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Global Change Biol. 13: 1860-1872
Parr C. L. and Gibb H. 2010. Competition and the role of dominant ants. In: Ant Ecology (Lach L., Parr C.L. and Abbott K.L., Eds), Oxford University Press, Oxford, UK. pp 77-96
Pearson R. G., Raxworth C. J., Nakamura M. and Peterson A. T. 2007. Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. J. Biogeogr. 34: 102-117
Peterson A. T., Papes M. and Eaton M. 2007. Transferability and model evaluation in ecological niche modeling: a comparison of GARP and Maxent. Ecography 30: 556-560
Phillips D. L. and Dudik M. 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31: 161-175
Phillips S. 1996. A brief Maxent tutorial. www-2.cs.cmu.edu/afs/cs/user/aberger/www/html/tutorial/tutorial.html
Phillips S. J. 2008. Transferability, sample selection bias and background data in presence-only modelling: a response to Peterson et al. (2007). Ecography 31: 272-278
Phillips S. J., Anderson R. P. and Schapire R. E. 2006. Maximum entropy modeling of species geographic distribution. Ecol. Model. 190: 231-259
Phillips S. J., Dudik M. and Shapire R. E. 2004. A maximum entropy approach to species distribution modeling. Proc. 21st Int. Conf. Machine Learning, Banff, Canada
Pudlo R. J., Beattie A. J. and Culver D. C. 1980. Population consequences of changes in ant–seed mutualism in Sanguinaria canadensis. Oecologia 146: 32-37
R 2005. R Development Core Team. R Foundation for Statistical Computing. http://cran.r-project.org. Accessed May 2009
Radford A. E., Ahles H. E. and Bell C. R. 1968. Manual of the Vascular Flora of the Carolinas. The University of North Carolina Press, Chapel Hill, USA. 1245 pp
Retana J. and Cerda X. 2000. Patterns of diversity and composition of Mediterranean ground ant communities tracking spatial and temporal variability in the thermal environment. Oecologia 123: 436-444
Rico-Gray V. and Oliveira P. 2007. The Ecology and Evolution of Ant–Plant Interactions. The University of Chicago Press, Chicago. 320 pp
Root T. L., Price J. T., Hall K. R., Schneider S. H., Rosenzweig C. and Pounds J. A. 2003. Fingerprints of global warming on wild animals and plants. Nature 421: 57-60
Sanders N. J., Lessard J. P., Fitzpatrick M. C. and Dunn R. R. 2007. Temperature, but not productivity or geometry, predicts elevational diversity gradients in ants across spatial grains. Glob. Ecol. Biogeogr. 16: 640-649
Smallwood J. 1982a. The effect of shade and competition on emigration rate in the ant Aphaenogaster rudis. Ecology 63: 124-134
Smallwood J. 1982b. Nest relocation in ants. Insect. Soc. 29: 138-147
Smith D. R. 1979. Catalog of Hymenoptera in America North of Mexico, vol 2. Smithsonian Institution Press, Washington, D.C. 1198 pp
Talbot M. 1934. Distribution of ant species in the Chicago region with reference to ecological factors and physiological toleration. Ecology 15: 416-439
Umphrey G. J. 1996. Morphometric discrimination among sibling species in the fulva–rudis–texana complex of the ant genus Aphaenogaster (Hymenoptera: Formicidae). Can. J. Zool. 74: 528-559
Van Der Wal J. L., Shoo P., Graham C. and Williams S. E. 2009. Selecting pseudo-absence data for presence-only distribution modeling: how far should you stray from what you know? Ecol. Model. 220: 589-594
Ward P. 2007. Phylogeny, classification, and species-level taxonomy of ants (Hymenoptera: Formicidae). Zootaxa 1668: 549-563
Warren R. 2010a. An experimental test of well-described vegetation patterns across slope aspects using woodland herb transplants and manipulated abiotic drivers. New Phytol. 185: 1038-1049
Warren R., Giladi I. and Bradford M. A. 2010. Ant-mediated seed dispersal does not facilitate niche expansion. J. Ecol. 98: 1178-1185
Warren R. J. 2008. Mechanisms driving understory evergreen herb distributions across slope aspects: as derived from landscape position. Plant Ecol. 198: 297-308
Warren R. J. 2010b. A test of temperature estimation from solar irradiation and a simple statistical method to integrate elevation into prediction models. Castanea 75: 67-77
Warren R. J., Bahn V. and Bradford M. A. 2011. Temperature cues phenological synchrony in ant-mediated seed dispersal. Global Change Biol. 17: doi:10.1111/j.1365-2486.2010.02386.x
Williams J. W. and Jackson T. J. 2007. Novel climates, no-analog communities, and ecological surprises. Front. in Ecol. Envir. 5: 475-482
Wisz M. S., Hijmans R. J., Li J., Peterson A. T., Graham C. H. and Guisan A. 2008. Effects of sample size on the performance of species distribution models. Divers. Distrib. 14: 763-773
Zelikova T. J., Dunn R. R. and Sanders N. J. 2008. Variation in seed dispersal along an elevational gradient in Great Smoky Mountains National Park. Acta Oecol. 34: 155-162
Acknowledgments
We would like to thank Gary J. Umphrey for publishing Aphaenogaster collection location data, which we used for training our models; AntWeb, Antbase and Discover Life for making available the Aphaenogaster spp. collection data we used for testing our models; and WorldClim for making available the climate data we used for determining climate niche differences among species. We thank the two anonymous reviewers for helpful manuscript suggestions.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Warren, R.J., McAfee, P. & Bahn, V. Ecological differentiation among key plant mutualists from a cryptic ant guild. Insect. Soc. 58, 505–512 (2011). https://doi.org/10.1007/s00040-011-0174-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00040-011-0174-x