selective logging, Aphaenogaster swammerdami, Aphaenogaster, Kirindy Forest, species composition, SL, Madagascar, colony size, colony, Forest Ecology and Management, ant species, ecology, colonies, Cambridge University Press, J. U. Ganzhorn, food intake, Insectes Sociaux, Stable isotope analysis, University of Chicago Press, maximum values, tropical dry forest, dry deciduous forest, Behavioral Ecology, western Madagascar, es, central western Madagascar, Feeding behavior, dry tropical forest, wood ants, Sorg, Hymenoptera, Amazon, D. M. & Ward, Vercauteren Drubbel, Christian Peeters, disturbances, disturbed sites, opportunistic species, size parameters, forest density, tropical rain forest, Joe Burkhardt, German Academic Exchange Service, Aphaenogaster swammerdami meets, Conservation Ecology, trophic levels, References Andersen, The stable isotope analysis, Harvard University Press, tropical rain forests
Investigating behavior and ecology of Aphaenogaster swammerdami (Formicidae) in selectively logged forest: 20 years later a happy ant?
Marie T. Dittmann1,2, Melanie Dammhahn1 & Peter M. Kappeler1,3 1Behavioral Ecology & Sociobiology Unit, German Primate Center (DPZ), Leibniz Institute for Primate Research, Kellnerweg 4, 37077 Gцttingen, Germany E-mail: [email protected]
2ETH Zьrich, Institute of Agricultural Sciences, Universitдtstrasse 2, 8092 Zьrich or Clinic for Zoo Animals, Exotic Pets and Wildlife, Vetsuisse Faculty, University of Zurich, Winterthurerstrasse 260, 8057 Zurich, Switzerland E-mail: [email protected]
3Department of Sociobiology/Anthropology, JohannFriedrich-Blumenbach Institute of Zoology & Anthropology, University of Gцttingen, Kellnerweg 6, 37077 Gцttingen, Germany E-mail: [email protected]
Abstract The impacts of logging have been widely studied at the community level, describing changes in species composition and richness, whereas the small-scale effects on behavior and ecology of single species have received less attention. We investigated whether the Malagasy wood ant Aphaenogaster swammerdami exhibits differences in colony density, colony size, and feeding ecology between three different sites within the Kirindy Forest (CNFEREF), a dry deciduous forest in central western Madagascar. Specifically, we compared undisturbed primary forest, a selectively logged area, and one site exposed to natural disturbance caused by an adjacent river. Transect surveys were used to record colony density as well as diameter of the mound as a predictor of colony size. Focal colonies were selected at all three sites to assess other aspects of colony size, i.e. the number and size of workers and the home range area. The feeding ecology of ants from all three sites was compared by using observational records on food quality and quantity, as well as stable isotope analyses of ant workers. We found that in selectively logged forest, colony density was lower and colony size larger than at the two other sites. Feeding behavior differed slightly, as colonies from selectively logged forest had a higher intake of animal food sources. Thus, we tentatively conclude that A. swammerdami, despite its opportunistic lifestyle, still
exhibits responses to selective logging, which took place 20 years ago. Replicating these findings on a larger scale and determining the specific mechanisms leading to changes in lifestyle along disturbance gradients should be the focus of future studies. Key words: anthropogenic disturbance, Aphaenogaster, Madagascar, Kirindy Forest, selective logging, stable isotope analysis Rйsumй dйtaillй Les impacts de l'exploitation agro-forestiиre sur des communautйs d'espиces ont йtй йtudiйs en dйtail, et on a trouvй que l'exploitation sylvicole peut causer des changements dans la composition ou la richesse en espиces. Cependant, les changements а un niveau plus prйcis comme l'йcologie ou le comportement d'espиces prйcises ont reзu peu d'attention. En consйquence, nous avons йtudiй l'impact de l'exploitation sylvicole semi-mйcanisйe sur la fourmi malgache, Aphaenogaster swammerdami. Pour cela, nous avons choisi trois sites dans la forкt sиche de l'Ouest de Madagascar а Kirindy (CNFEREF) : un site qui йtait exploitй sйlectivement il y a 20 ans, un site exposй aux perturbations d'une riviиre adjacente et un autre site dans la forкt primaire. Un йchantillonnage par transect a йtй utilisй pour dйterminer la densitй des colonies d'A. swammerdami. En mкme temps, le diamиtre de chaque termitiиre a йtй mesurй pour estimer le nombre d'ouvriиres de chaque colonie. Pour les investigations plus approfondies, six colonies focales sont choisis dans chaque localitй. Pour ces colonies focales, nous avons mesurй le nombre d'ouvriиres par la mйthode de capture-marquage-recapture, la taille des ouvriиres, et la taille de l'aire de nourrissage. Le rйgime alimentaire a йtй dйterminй par des observations et par des analyses d'isotopes stables d'azote et de carbone des ouvriиres de chaque colonie. Les rйsultats ont montrй que la densitй des colonies est plus faible dans la forкt exploitйe que dans les autres sites. Pourtant, les termitiиres dans la forкt exploitйe ont des diamиtres plus grands et hйbergent plus d'ouvriиres. Le rйgime alimentaire est lйgиrement diffйrent entre les sites : la proportion de nourriture d'origine animale est plus importante pour les colonies de la forкt exploitйe. Ces rйsultats sont
Dittmann, M. T., Dammhahn, M. & Kappeler, P. M. 2014. Investigating behavior and ecology of Aphaenogaster swammerdami (Formicidae) in selectively logged forest: 20 years later a happy ant? Malagasy Nature, 8: 35-48.
36 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
supportйs par les dates d'analyses d'isotopes stables (le ratio de 15N est plus grand dans les colonies dans la forкt exploitйe). Nous n'avons pas trouvй des diffйrences dans la surface de l'aire d'alimentation et la quantitй de nourriture entre les sites. Mais, des diffйrences entre les colonies occupant les trois sites ont йtй constatйes. Bien que cette йtude manque de rйplicas, nous concluons provisoirement que A. swammerdami montre des rйponses а l'exploitation semi-mйcanisйe de la forкt 20 ans aprиs. Cependant, des йtudes futures devront кtre menйes pour vйrifier la gйnйralitй de ces rйsultats en examinant plus de sites exposйs а diffйrent degrй de perturbation et d'autres espиces de fourmis а Kirindy. Mots clйs : perturbation anthropogйnique, Aphaenogaster, Madagascar, Forкt de Kirindy, exploitation sйlective, analyse des isotopes stables Introduction From an evolutionary perspective, environmental disturbances are a two-edged sword because they not only lead to local ecological instability but also to new adaptations that ultimately contribute to the maintenance of biotic diversity (Darwin, 1859; Connell, 1978). However, in contrast to natural disturbances, anthropogenic interferences increased recently with the expansion of human populations into nearly every natural habitat on our planet. One example of anthropogenic disturbance that is assumed to have relatively low impact on the ecosystem is selective logging. It is considered as a sustainable alternative to clear-cutting as it allows the forest to recover and provides a long-term wood source of timber and fire wood from regrowth. Nonetheless, selective logging does have effects on species composition of animal communities (Magnusson et al., 1999; Sekercioglu, 2002; Peters et al., 2006; Edwards et al., 2011), forest structure (Uhl, 1989; Hall et al., 2003; Okuda et al., 2003), genetic diversity (Jennings et al., 2001) or nutrient cycling (Herbohn & Congdon, 1993). Although several studies have focused on the impact of selective logging on the behavioral ecology of mammals (Ganzhorn, 1995; Laurance & Laurance, 1996; Arnhem et al., 2008) and birds (Lambert, 1992; Chouteau, 2004), ecologically important groups of invertebrates remain poorly studied in this context. The present study contributes insight into the smallscale impacts of disturbances on colony density, size, and feeding behavior of a Malagasy ant. Together with termites, ants are estimated to make up one third of the global animal biomass
(Hцlldobler & Wilson, 2009) and they influence their ecosystems in a variety of ways. Ants occupy different trophic levels as their feeding types range from herbivory to specialized predation (Tillberg et al., 2006). They can serve as pollinators (Beattie, 1985; Hцlldobler & Wilson, 1990), seed dispersers (Bцhning-Gaese et al., 1996) or as food source for other species (McNab, 1984). Ground-dwelling ants play a role in soil turnover and structure (Humphreys, 1981; Lobry de Bruyn & Conacher, 1994) as well as nutrient cycling (Levieux, 1983). Additionally, ants are often considered as bio-indicators (Andersen, 1997; Majer, 1983). Several studies indicated that selective logging does not have an impact on ant species richness (Olson & Andriamiadana, 1996; Vasconcelos et al., 2000) but there is some evidence that it alters species composition (Punttila et al., 1991; Kalif et al., 2001; Dunn, 2004; Widodo et al., 2004). Species composition and richness of a given habitat depend on the ecological flexibility of single species making up the community. Species-specific responses of ants to selective logging can help us understand how species recover from or adapt to particular disturbances, which can be considered agents of natural selection (Sousa, 1984). As mentioned above, although the responses of ant communities to logging have been studied, little is known about how morphological, behavioral, and life history traits of single species are affected by this anthropogenic disturbance. On a species level, logging appears to have a negative impact on population densities in some species (Olson & Andriamiadana, 1996; Vasconcelos et al., 2000; Sorvari & Hakkarainen, 2007), whereas it enhances population growth
in others (Walsh et al., 2004). For example, mounds of Formica rufa were found to be smaller after logging (Domisch et al., 2005), and Sorvari & Hakkarainen (2009) found individual ants to be smaller in logged forest, probably due to nutritional deprivation. There has been some discussion about how ants should adapt to disturbances. Kaspari & O'Donnell (2003) suggest that severe disturbances would select for smaller colonies, which could also result in shorter generation times (Linksvayer & Janssen, 2009). In addition to this `r-strategy', small colonies favor low queen-worker dimorphism, which could result in polygyny instead of investing in a single egg-laying queen that could be killed during frequent disturbances (Linksvayer & Janssen, 2009). In contrast, some species invest in colony growth rather than reproduction to cope with disturbances (i.e. competitors or predators)
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 37
(Linksvayer & Janssen, 2009). However, a small number of unspecialized workers seem to be more advantageous to deal with disturbances than a large number of specialized workers albeit even these can exhibit task redundancy (Wilson, 1984). Obviously, different species vary in their response to disturbances. For example, it was shown that anthropogenic disturbances, such as fire, facilitate the colonization by a dominant species to some degree (Gibb & Hochuli, 2003), indicating that disturbances can indeed be advantageous for opportunistic species that are able to adapt to the altered environments. Madagascar, being one of the global hotspots of biodiversity, shelters vast numbers of endemic species (Myers et al., 2000). About 96% of the 1000 described Malagasy ant species are endemic, but probably only about one fourth of the island's species have been named (Fisher, 2003). The dry forests of western Madagascar with their 86 known ant species are species-rich compared to other localities (Olson & Ward, 1996). Our study took place in Kirindy Forest (CNFEREF), where selective logging (for Commiphora spp.) started in 1987 (Ganzhorn, 1995). Studying the impact of tree extraction on ants 5 - 10 years after selective logging, Burkhardt et al. (1996) detected a slightly higher number of overall ant species in the selectively logged area, whereas the number of ground-dwelling species, in particular, was lower. One of these ground-dwelling species is Aphaenogaster swammerdami (Forel 1886) of the Family Formicidae, for which colonies were significantly less abundant in the selectively logged forest (Burkhardt et al., 1996). This abundant species
lives in large underground nests that have one large entrance hole and conspicuous mounds (BцhningGaese et al., 1999). Outside Madagascar, the genus Aphaenogaster is found on all continents and it has been studied in the Mediterranean (Cerdа et al., 2009; Galarza et al., 2009), North America (Hцlldobler et al., 1978; Sanders & Gordon, 2002; Tschinkel, 2011), Central America (McGlynn et al., 2003, 2004), and Australia (Shattuck, 2008). As its congeners, A. swammerdami seems to be omnivorous. In Madagascar, this species plays an important role in the secondary dispersal of seeds of the Malagasy tree Commiphora guillaumini (Bцhning-Gaese et al., 1996). However, little is known about the lifestyle of A. swammerdami. The aim of this study was to determine differences in the ecology and behavior of A. swammerdami
between logged forest and two control sites consisting of undisturbed primary forest and another unlogged site exposed to natural disturbance by an adjacent river. Specifically, we asked: (1) Do colony density and size parameters (worker number, worker size, mound diameter, and home range) differ between the three sites? (2) Does the feeding behavior of colonies differ between habitat types with respect to dietary composition and quantity? Materials and methods study site
and forest characteristics Data were collected from March to May 2011 in the Kirindy Forest, a dry deciduous forest in western Madagascar, approximately 60 km northeast of Morondava (44°39'E, 20°03'S, 30 - 60 m asl). The study site is located within a 12,500 ha forest concession of the Centre National de Formation, d'Etude et de Recherche en Environnement et Foresterie (CNFEREF) de Morondava. Kirindy Forest is characterized by a dry season of 6 - 9 months and a rainy season between December and February (Sorg & Rohner, 1996). Between 1987 and 1990, selective logging of large trees of Commiphora guillaumini (Burseraceae; local Malagasy name arofy) occurred in selected parts of the forest (Sorg et al., 2003). We collected data at three different sites: (1) partly selectively logged around 1987 (SL) as well as (2) an immediately adjacent unlogged site (adjNL) and (3) another unlogged site about 2 km away adjacent to the Kirindy River (rivNL) (Figure 1). The latter site was chosen as an additional control site for natural disturbances due to the river as A. swammerdami seems to be sensitive to variation in soil humidity (Burkhardt et al., 1996). Unpublished data (F. Koch, C. GroЯheim) of forest density and composition were used to assess forest characteristics of the study sites. This analysis revealed that the number of trees per hectare did not differ among the three study sites (ANOVA, n = 60 randomly chosen squares of 400 mІ, F (2, 57) = 1.88, p = 0.162). Selective logging reduced the number of Commiphora trees per plot in NL as compared to the two unlogged sites (post-hoc tests: SL adjNL: U = 10539, p < 0.001; SL rivNL: U = 36498, p < 0.001; rivNL adjNL: U = 46223, p = 0.122). The number of Commiphora in SL was reduced by 75 - 87% in comparison to the two unlogged sites.
38 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
Figure 1. Overview over the study area in the Kirindy Forest (CNFEREF) and transects at each of the three sites: selectively logged (SL) and unlogged (adjNL) areas in the north and unlogged rivNL close to Kirindy River.
Transect surveys In each area, three transects were conducted on small trails, intersecting both selectively logged and unlogged forest in SL and adjNL (Figure 1). The minimal distance between transects was 120 m in rivNL and 180 m in SL and adjNL. During surveys, colonies within a distance of 15 m on each side of the transect were counted and their perpendicular distance to the transect line was measured. Transects covered a total length of 726 - 836 m. To determine the colony density of A. swammerdami the following formula was used: D = N (2N/ i(di2)) / (2L) (D = density, n = total number of colonies detected, di = perpendicular distance of colony to transect line, L = transect length (Sutherland, 2006, pp. 147)) The mound diameter of each detected colony was measured as mound size is assumed to be a reliable predictor of colony size (Tschinkel et al., 1995). Further, for each of the three plots, six focal colonies were chosen randomly by selecting two colonies from each of the three transects. As a smallscale measurement of colony density, the distance of neighboring A. swammerdami colonies within a radius of 20 m were determined for each focal colony.
Estimation of worker numbers The Jolly-Seber capture-mark-recapture method (Krebs, 1999) was applied for all 18 focal colonies to determine the number of foragers. To capture the ants a wooden stick of approximately 30 cm length with a diameter of 5 mm was held into the entrance hole of a colony for 10 sec after which the individuals climbing up the stick were transferred into a jar. For each capture session, this procedure was repeated 20 times within 5 min. All caught ants were marked with waterproof color on the thorax (Edding 751 Paint Marker, Edding International GmbH, Ahrensburg, Germany). In total, 61 - 151 individuals were marked per colony. Marked individuals were released three hours after the capture at the nest entrance. The recapture procedure was the same as the capture procedure explained above. The entire capture procedure for one colony occurred at the same hour during three consecutive days (three capture sessions) and was carried out on different days than the determination of home ranges and observations. Home ranges The home range of each focal colony was determined by assessing locations of single individuals foraging in the immediate surroundings of the nest. Each randomly encountered individual was presented a
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 39
piece of rice to observe the homing of the animal. As the maximum seed dispersal distance has been recorded to be 10.4 m (Bцhning-Gaese et al., 1999), we presented rice within a radius of 15 m from the nest entrance. In total, we collected 27 to 32 locations for each colony within one or two consecutive days of similar weather conditions (e.g. not during or after strong rain), and evenly split between morning and afternoon. Home range size was determined by calculating the area of the minimum convex polygon of all assessed data points, using the ArcView 3.1 GIS software (ESRI, Redlands, CA, USA). Worker size and weight To estimate body size 40 - 47 workers were collected from each focal colony by the same method used for the capture-mark-recapture procedure. The body length of each individual was measured from the tip of the mandibles to the most posterior point of the abdomen (precision 0.5 mm). To estimate worker weight and thereby biomass, six individuals of each colony were oven-dried and weighed to the nearest µg with an analytical balance. Biomass was approximated by multiplying the site-specific estimated number of workers by the site-specific mean weight of the workers by the colony density for each of the three sites. Observations on feeding ecology Each colony was observed on four different days, each observation period being one hour, and at four different times of the day (07h 30 - 10h, 10h - 12h 30, 12h 30 - 15h, 15h -17h 30) in a randomized order. During observations, category and size of each food item transported into the nest entrance was noted (categories: insect, other invertebrate, vertebrate, fungus, plant, miscellaneous). Stable isotope analysis The ratios of stable isotopes of chemical elements can be used to investigate ecological relationships among organisms such as food webs. Stable nitrogen (15N) can serve as a tracer to estimate the trophic position of organisms because it is enriched by ca. 3 in consumers relative to their diet (Post, 2002). Enrichment in stable carbon (13C) is smaller, ca. 1, and this isotope provides information on the number of basal resources entering a food web (Post, 2002). We applied stable isotope analysis to investigate two aspects of the feeing behavior of A. swammerdami: (1) We aimed at determining the trophic level of this
species, i.e. its position in the food web via stable nitrogen and (2) we tested for differences in isotope ratios between sites to investigate whether the diets of the focal colonies differ in composition between sites. For the stable isotope analysis on average 8 (range 1 - 12) ants were collected from each focal colony, stored in 100% ethanol for four weeks and oven-dried at 60°C for three days prior to stable isotope analysis. Legs and thorax were used for the analysis to avoid artifacts associated with ingested food in the abdomen and head. Further, we determined the habitat baseline of the sampled ants using two common tree species (Diospyros tropophylla, Ebenaceae: nrivNL = 20, nadjNL & SL = 15 and Securinega seyrigii, Euphorbiceae: nrivNL = 19, nadjNL & SL = 20). Leaves were collected, immediately dried in the sun, and stored without any preservative. Leave samples were ground and homogenized with a ball mill before stable isotope analyses. Analyses were carried out by the Center for Stable Isotope Research & Analysis (KOSI), Gцttingen, using an isotope ratio mass spectrometer (Delta Plus, Finnigan MAT, Bremen, Germany) in an online-system after passage through an element analyzer (NA 1110, Carlo Erba, Milan, Italy). Data Analysis
: Observational data All measured variables were tested for normal distribution by performing Shapiro-Wilk tests. To compare mound sizes of colonies among sites we used an ANOVA and post-hoc pair-wise comparisons via Tukey's-HSD tests. Parameter recorded for the 18 focal colonies (number of workers, number of food items brought into the nest, home range area, worker size) were compared among the three sites using Kruskal-Wallis tests, followed by Mann-Whitney-U tests for pair-wise comparisons. We used Spearman correlations to determine relationships between single variables using data from the 18 focal colonies. Significance levels of non-parametric post-hoc tests were adjusted according to the Holm-Bonferroni correction (Holm, 1979). All other significance levels were set at P 0.05. All statistical analyses were carried out in R 2.15.0. Data analyses: Stable isotope analysis We compared stable nitrogen and stable carbon signatures of individuals among sites using ANOVA / Kruskal-Wallis test with corresponding post-hoc tests. Analyses were carried out for both absolute values
40 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
and relative values
corrected for the habitat specific baseline of both tree species. Results Transect surveys and colony density Mound diameter of all colonies detected during the transect surveys differed between the three sites and were largest in logged forest (F = 75.72, p < 0.001; post-hoc tests: all p < 0.001) (Figure 2A). Mean colony density in NL was lowest with 0.028 colonies per meter transect, in contrast to almost ten fold higher densities of 0.201 in adjNL and 0.205 in rivNL (Figure 2A). The distance of the nearest neighbor colony did not differ among sites (22df = 2.15, p = 0.34), though the median distance was highest in SL and significantly higher than in rivNL (post-hoc
test: U = 36, p = 0.005), supporting the findings from the transect surveys, which indicate that SL had the lowest colony density. Comparison among study sites Overall, the estimated number of workers per colony differed among the three sites (22df = 8.51, p = 0.014). Colonies in SL contained more workers compared to adjNL (post-hoc test: U = 34, p = 0.013) and rivNL (U = 33, p = 0.020) (Figure 2B). Home ranges, which were roughly circular around the nest entrance, showed a difference in area close to significance between the three sites (22df = 5.49, p = 0.064). Home ranges were significantly smaller in rivNL than in SL (post-hoc test: U = 0, p = 0.002) (Figure 2C). Mean worker size was highest in SL
Figure 2. A) On the left axis: mound diameters of colonies detected on three ca. 250-m transects per site (rivNL: 96 colonies, SL: 22 colonies, adjNL: 68 colonies). Box plots indicate median, upper and lower quartiles, as well as minimum and maximum values, circles indicate outliers. On the right axis: mean site-specific colony density (in colonies per meter transect) of each three transects per site. B) Estimated number of workers per colony of the focal colonies plotted by site (n = 6 colonies per site). C) The home range area of the focal colonies plotted by site (n = 6 colonies per site). D) Mean size of the workers for all three sites (n = 6 colonies per site). Significant differences (P < 0.05) are indicated by asterisks.
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 41
and differed significantly from rivNL (post-hoc test: U = 32, p = 0.030) (Figure 2D). food intake
and the distance of the nearest neighbor did not differ among sites (p > 0.1). As some variables of the 18 focal colonies were correlated with the estimated number of workers, comparisons between sites were also performed with residuals from these correlations to control for colony size. Residuals of the correlations with the estimated number of workers did not differ among sites for home range area (22df = 4.43, p = 0.11), worker size (22df = 1.82, p = 0.53) and number of food items brought into the nest per hour (food intake) (22df = 1.82, p = 0.40). Correlations between characteristics of focal colonies The estimated number of workers correlated positively with mound size (n = 18, rs = 0.68, p =
0.002). Additionally, mound size correlated positively with home range area, food intake, worker size, and worker weight (rs > 0.63, P < 0.01). By multiplying colony density by the site-specific average number of workers per colony and the average dry weight of the workers, we assessed ant biomass. As the minimal recorded number of workers within a colony was around 200 individuals, this value was used as a proxy for colonies in rivNL as the calculation using their mean mound diameter resulted in negative values. The calculations revealed a biomass of 0.06 g/m transect in rivNL, 0.22 g/m in adjNL, and 0.22 g/m in SL. Feeding ecology: Observational data Colonies in rivNL had a lower intake of insects (22df = 49.2 p < 0.001) and a higher intake of fungi (22df = 7.8, p = 0.02) compared to adjNL and SL (Figures 3 & 4). There were no differences among colonies for
Figure 3. Mean proportions of the food categories of items brought into the nest during 4 hours of observation per colony (n = 6 colonies for each site).
Figure 4. Mean food intake for all 18 focal colonies. The Y-axis corresponds to the number of items brought into the nest entrance during the entire 4 hours of observation per colony, the X-axis corresponds to the recorded food categories.
42 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
all other food categories (all p > 0.1). During 72 hours of observation, only one dispersal of a Commiphora guillaumini seed was observed. Feeding ecology: Stable isotope analysis The stable isotope habitat baseline differed between rivNL and SL and adjNL in 15N but not in 13C for both tree species (Securinega seyrigii: rivNL: 15N = 6.18 ± 0.85, 13C = -28.7 ± 0.70; SL and adjNL: 15N = 11.5 ± 0.97, 13C = -28.5 ± 0.70; Diospyros tropophylla: rivNL: 15N = 7.2 ± 0.71, 13C = -30.1 ± 0.61; SL and adjNL: 15N = 11.7 ± 1.51, 13C = -30.0 ± 0.62). For A. swammerdami, comparison of absolute 13C values between sites revealed lower values in rivNL compared to SL and adjNL, but no difference within SL and adjNL (F = 14.9, p < 0.001; post-hoc tests: rivNL adjNL and rivNL SL: p < 0.001; SL adjNL: p = 0.84) (Figure 5A). 15N values differed between all sites (2 = 121.5, p < 0.001; post-hoc tests: all p < 0.001) (Figure 5B). When stable carbon values were corrected for the habitat specific baseline of Diospyros tropophylla, difference between rivNL and SL and adjNL remained (F = 5.44, P = 0.005; post-hoc tests: rivNL adjNL: p = 0.007; rivNL SL: p = 0.050; SL adjNL: p = 0.84), however, when corrected for the baseline of S. seyrigii no difference could be found between sites (F = 0.73, P = 0.49). Baseline correction of stable nitrogen data for both tree species revealed significantly higher 15N values in rivNL and SL compared to adjNL (S. seyrigii: F = 57.1, P < 0.001; post-hoc tests: rivNL adjNL and
adjNL SL: p < 0.001; SL rivNL: p = 0.22; D. tropophylla: F = 41.9, P < 0.001; post-hoc tests: rivNL adjNL and adjNL SL: p < 0.001; SL rivNL: p = 0.06) (Figure 6) Discussion The aim of this study was to explore and quantify variation in behavior and ecology of the Malagasy wood ant Aphaenogaster swammerdami among field sites with different sylvicultural histories: one site was selectively logged 20 years earlier (SL), one site consisted of unlogged relatively intact forest (adjNL), and a third site was unlogged but exposed to disturbances due to an adjacent river (rivNL). The two unlogged sites appeared rather similar to one another as compared to the logged site. We found that colonies in SL were larger but less abundant compared to the two unlogged sites (adjNL, rivNL). The estimated biomass in SL and adjNL were four times higher as compared to rivNL, but did not differ within the two sites. Food composition differed slightly between sites in terms of fungi and insects; this factor was additionally highlighted by stable isotope analysis. Do colony density and size parameters differ among the three forest sites? Colony densities in unlogged forest were ten times higher than in selectively logged forest. Though there was no difference in tree density, the logged area by
Figure 5. A) Stable carbon isotope values (13C) of ants of the three study sites (n = 210 individuals). Plotted are median, upper and lower quartiles as well as minimum and maximum values. B) Stable nitrogen values (15N) of ants of the three study sites (n = 145 individuals). Significant differences (P < 0.05) are indicated by asterisks.
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 43
Figure 6. Values of stable nitrogen and carbon values of ants corrected for the site specific habitat baseline assessed from two tree species (Securinega seyrigii and Diospyros tropophylla).
definition lacks Commiphora guillaumini trees, which could explain the lower abundance of colonies at the SL site. Aphaenogaster swammerdami acts as a seed disperser for C. guillaumini and feeds on the arils of the seeds (Voigt et al., 2002). Surprisingly, the dispersal of seeds of C. guillaumini was observed only once even though this study was carried out during the fruiting season of this tree. Thus, A. swammerdami does not seem to depend on C. guillaumini seeds as a major food source and the low colony density might be due to other factors. One possible explanation for the differences in density could be increased pressure due to competition with Pheidole spp., which has been found to influence
negatively A. swammerdami (Burkhardt et al., 1996). However, this explanation remains speculative, as we did not assess the abundance of Pheidole spp. Nonetheless, the finding of decreased ant population densities in selectively logged forest is in agreement with other studies (Olson & Andriamiadana, 1996; Vasconcelos et al., 2000; Sorvari & Hakkarainen, 2007). Colony size was largest in selectively logged forest, where density was lowest. In addition, colonies in the selectively logged forest had the largest home ranges and the highest food intake two parameters highly dependent on colony size (Tschinkel et al., 1995). These findings can be
44 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
explained by two factors: 1) competition between colonies might increase with colony density. Aphaenogaster swammerdami seems to maintain rather exclusive territories and home ranges do not appear to overlap (unpublished data). It is therefore likely that colony size is limited by neighboring nests, which would explain why colonies in SL could grow to a maximum size. This limitation might be due to an effect of saturation in the unlogged sites where the carrying capacity of A. swammerdami might be reached in contrast to the selectively logged area. 2) colony and worker size might depend on nutrient availability and quality. According to Kaspari (2005), positive correlations between food supply, as well as growth in worker Body mass
and number, have been found. However, increased growth due to a higher food supply at the logged site would be contradictory to the lower colony density as in that case one would expect an overall growth of the population, not only within colonies. The estimated biomass of A. swammerdami did not differ between selectively logged and unlogged forest in SL and adjNL. Nevertheless, it was nearly four times lower in rivNL, which might be associated with a higher risk of flooding by the nearby river. Though one has to consider that the biomass of A. swammerdami was only roughly estimated, this finding suggests that the impact of the river as a natural source of disturbance (flooding, higher soil humidity) might actually be higher than the one of selective logging. Another explanation for the lower biomass in rivNL could, again, be the constant competition with Pheidole spp., which were found to be more abundant in this area of the Kirindy Forest in 1995 (Burkhardt et al., 1996) and were observed to force A. swammerdami colonies to leave their nests (J. Burkhardt, unpublished data). This pressure due to inter-specific competition might result in a higher frequency of A. sammerdami colonies moving from one nest to another, perhaps resulting in smaller colonies and individuals, which would decrease their biomass. Does the feeding behavior of colonies differ among sites with respect to composition and quantity? The observational data revealed that all colonies exploit the same food categories. However, these categories were defined at broad taxonomic levels, such as Kingdoms or Phyla and it is possible that the diet between sites differed on a more precise
taxonomic level. Only minor differences could be found in the quantitative composition of the diet. Colonies in the unlogged forest of rivNL fed on more fungi than in SL and adjNL. As rivNL is adjacent to the Kirindy River, this effect can possibly be explained by higher soil humidity and, hence, higher fungus growth. The observational data also revealed that the diet of colonies in SL and adjNL included more insects, a finding supported by the stable isotope analysis (absolute 15N values were 4 5 higher in SL and adjNL). When corrected for the site specific habitat baseline, 15N is significantly lower in adjNL compared to SL and rivNL. This indicates that, within SL and adjNL, the diet of ants in logged forest was composed of more material from higher trophic levels, i.e. animal tissue. This effect could have two possible explanations: (1) the availability of plant-derived food is lower in SL. With regard to selective logging, this would be true for C. guillaumini trees though this study could not find observational evidence for a high dependence of A. swammerdami on C. guillaumini seeds. Alternatively, it is possible that the availability of faunal food sources is higher in SL compared to adjNL, whereas the abundance of plant-derived food sources is equal. (2) It is possible that the deviation in 15N is due to a difference in foraging behavior and competitive ability. One could argue that the larger ants / colonies in SL are superior competitors. Nevertheless, the diet of all colonies was mainly composed of very small items below 1 cm, including insects such as termites, ants, and larvae, which would not be easier to subdue with increasing body or colony size. The storage of specimens in ethanol can lead to shifts in 13C values in stable isotope analysis (Tillberg et al., 2006). Hence, these results have to be interpreted cautiously. However, all specimens underwent the same storage conditions, which would lead to a similar bias in all samples without affecting the relative differences between them. As the main focus of the stable isotope analysis was to determine the trophic level based on 15N values, which are not affected by storage in EtOH. In addition to the results on the differences among forest sites, this study provided new insights into the lifestyle of A. swammerdami. The estimated colony sizes are in agreement with results obtained from other Aphaenogaster spp., whose average number of adult individuals per colony ranges between 100 and 700 (Hцlldobler & Wilson, 1990; Tschinkel, 2011). In general, the ecology of A. swammerdami seems to
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 45
be comparable to it congeners, which live in colonies in the soil with a single large entrance hole, disperse seeds from different plants, and are omnivorous (Sanders & Gordon, 2002; McGlynn et al., 2003; Ness et al., 2009; Boulay et al., 2010). The stable isotope analysis revealed that this species covers approximately two trophic levels (15N values lie 0.6 - 5.4 above the habitat baseline) and exhibits high individual differences
in diet, which can be explained by an omnivore lifestyle a finding supported by the observational data. As found in other ant species that feed on uniformly distributed and continuously renewing resources (Hцlldobler & Wilson, 1990), the home ranges of A. swammerdami extended more or less circularly around the nest entrance and are established by solitary foragers. Linksvayer & Janssen (2009) stated that species that have the least difficulties to cope with disturbances are mainly opportunists, which in ants are characterized by small colony size, unspecialized morphology and behavior, unspecialized diet, and nest sites, as well as polygyny and polydomy. Aphaenogaster swammerdami meets most of these expectations (except for polygyny and polydomy, as we have no information at this time on breeding systems) and, therefore, it is classified as an opportunistic species. Hence, it should be adapted to cope with disturbances. Our results revealed clear divergences among sites exposed to different levels of disturbances, but we cannot conclusively assign these differences to adaptations to selective logging or soil humidity. Without replication of these experiments at additional sites, the mechanisms leading to the differences described herein remain unknown. Differences between logged and unlogged forest should not be reduced to sylvicultural management, as they could simply mirror a geographic effect. Therefore, as our results are descriptive, future research should focus on extending this dataset by investigating other sites exposed to different kinds of disturbances to find a general pattern. This study has provided interesting insights into the life history of A. swammerdami, although information about other ant species inhabiting the Kirindy Forest are lacking. Sensitive ant species that have a less opportunistic life style might show completely different or stronger responses to disturbance. This might in turn be advantageous for A. swammerdami as it could reduce competition from other ant species at disturbed sites. This in turn could explain why colonies in logged forest were larger
compared to colonies in unlogged forest. Several studies show that disturbances can indeed favor (re-) colonization of ecologically dominant species that alter the composition of a community or even replace other, less adaptable species (Fox & Fox, 1982; Gibb & Hochuli, 2003). Despite certain limitations, this study highlights that research on disturbance ecology should not only focus on effects at the community level but also on single species as changes in their lifestyles may impact the entire ecosystem and are the first signals to indicate alterations in species composition. Conclusion This study of the ecology and life history of Aphaenogaster swammerdami revealed significant differences between forest sites with different levels or histories of disturbance. In particular, colonies in parts of the forest that were selectively logged 20 years earlier were larger and fed on slightly more animal derived food, indicating that these ants respond to subtle variation in habitat quality. Our study also contributes new knowledge on the functional ecology of this ant species. Even though a lack of replication renders some conclusions preliminary, our study suggests that the effects of selective logging can be subtle and long lasting for particular species. This can be tested in future studies that examine single species across disturbance gradients, which would ultimately allow identifying the mechanisms leading to alterations in local species composition. Acknowledgements The authors thank Reinhard Langel for help with the stable isotope analysis. Thanks to Flavia Koch and Christian GroЯheim for providing unpublished botanical data. Thanks to Joe Burkhardt and Susanne Foitzik for sharing their broad knowledge on ants. Thanks to Pierre-Yves Henry for sharing his expertise on CMR methods. This project was supported by the German Academic Exchange Service (PROMOS). For comments on an earlier version of this paper, we are grateful to Steve Goodman and Christian Peeters. References Andersen, A. N. 1997. Using ants as bioindicators: Multiscale issues in ant community ecology. Conservation Ecology, Online publication
1, article 8. Arnhem, E., Dupain, J., Vercauteren Drubbel, R., Devos, C. & Vercauteren, M. 2008. Selective logging, habitat quality and home range use by sympatric
46 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
gorillas and chimpanzees: A case study
from an active logging concession in southeast Cameroon. Folia Primatologica, 79: 1-14. Beattie, A. J. 1985. The evolutionary ecology of antplant mutualisms. Cambridge University
Press, Cambridge. Bцhning-Gaese, K., Burkhardt, J. F. & Schmid, J. 1996. Seed dispersal in the tree Commiphora guillaumini: A combination of ornithochory and marmecochory in a dry tropical forest in werstern Madagascar. In Ecology and economy of a tropical dry forest in Madagascar, eds. J. U. Ganzhorn & J.-P. Sorg. Primate Report, 46: 305-310. Bцhning-Gaese, K., Gaese, B. H. &. Rabemanantsoa, S. B. 1999. Importance of primary and secondary seed dispersal in the Malagasy tree Commiphora guillaumini. Ecology, 80: 821-832. Boulay, R., Galarza, J. A., Cheron, B., Hefetz, A., Lenoir, A., van Oudenhove, L. & Cerdа, X. 2010. Intraspecific competition affects population size and resource allocation in an ant dispersing by colony fission. Ecology, 91: 3312-3321. Burkhardt, J., Bцhning-Gaese, K., Hornung, U. & Schmid, J. 1996. Comparison of sampling techniques for the characterization of ant communities in a Malagasy dry forest. Ecology and economy of a tropical dry forest in Madagascar. Primate Report, 46: 165-174. Cerdа, X., Angulo, E., Boulay, R. & Lenoir, A. 2009. Individual and collective foraging decisions: A field study of worker recruitment in the gypsy ant Aphaenogaster senilis. Behavioral Ecology and Sociobiology, 63: 551562. Chouteau, P. 2004. The impacts of logging on the microhabitats used by two species of couas in the western forest of Madagascar. Comptes Rendus Biologies, 327: 1157-1170.Connell, J. H. 1978. Diversity in tropical rain forests and coral reef
s. Science, 199: 1302-1310. Darwin, C. 1859. On the origin of species by means of natural selection. John Murray, London. Domisch, T., Finer, L. & Jurgensen, M. F. 2005. Red wood ant mound densities in managed boreal forests. Annales Zoologici Fennici, 42: 277-282. Dunn, R. R. 2004. Managing the tropical landscape: A comparison of the effects of logging and forest conversion to agriculture on ants, birds, and Lepidoptera. Forest Ecology and Management, 191: 215-224. Edwards, D. P., Larsen, H. T., Docherty, T. D. S., Ansell, F. A., Hsu, W. W., Derhй, M. A., Hamer, K. C. & Wilcove D. S. 2011. Degraded lands worth protecting: The biological importance of Southeast Asia
's repeatedly logged forests. Proceedings of the Royal Society B, 278: 82-90. Fisher, B. L. 2003. Formicidae, Ants. In The natural history of Madagascar, eds. S. M. Goodman & J. P. Benstead, pp. 811-818. University of Chicago Press, Chicago. Fox, M. D. & Fox, B. J. 1982. Evidence for interspecific competition influencing ant species diversity in a regenerating heathland. In Ant-plant interactions in
Australia, ed. R. C. Buckley, pp. 99-100. Junk Press, The Hague. Galarza, J. A., Boulay, R., Cerdа, X., Doums, C., Frederici, P., Magalon, H., Monnin, T. & Rico, C. 2009. Development of single sequence repeat markers for the ant Aphaenogaster senilis and cross-species amplification in A. iberica, A. gibbosa, A. subterranea and Messor maroccanus. Conservational Genetics, 10: 519-521. Ganzhorn, J. U. 1995. Low-level forest disturbance effects on primary production, leaf chemistry, and lemur population. Ecology, 76: 2084-2096. Gibb, H. & Hochuli, D. F. 2003. Colonisation by a dominant ant facilitated by anthropogenic disturbance: Effects on ant assemblage composition, biomass and resource use. Oikos, 103: 469-478 Hall, J., Harris, D. J., Medjibe, V. & Asthon, P. M. 2003. The effects of selective logging on forest structure and tree species composition in a Central African forest: Implications for management of conservation areas. Forest Ecology and Management, 183: 249-264. Herbohn, J. L. & Congdon, R. A. 1993. Ecosystem dynamics of disturbed and undisturbed sites in North Queensland wet tropical rain forest. Journal of Tropical Ecology, 9: 365-380. Hцlldobler, B. & Wilson, E. O. 1990. The ants. The Belknap Press of Harvard University
Press, Cambridge. Hцlldobler, B. & Wilson, E. O. 2009. The super-organism. W. W. Norton & Company, New York. Hцlldobler, B., Stanton, R. C. & Markl, H. 1978. Recruitment and food-retrieving behaviour in Novomessor (Hymenoptera: Formicidae). Behavioral Ecology and Sociobiology, 4: 163-181. Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6: 65-70. Humphreys, G. S. 1981. The rate of ant mounding and earthworm casting near Sydney, New South Wales
. Search, 12: 129-131. Jennings, S. B., Brown, N. D., Boshier, D. H., Whitmore, T. C. & Lopes, J. C. A. 2001. Ecology provides a pragmatic solution to the maintenance of genetic diversity in sustainably managed tropical rain forest. Forest Ecology and Management, 154: 1-10. Kalif, K. A. B., Azevedo-Ramos, C., Moutinho, P. & Malcher, S. A. O. 2001. The effect of logging on the ground-foraging ant community of eastern Amazonia. Studies on Neotropical Fauna and Environment, 36: 215-219. Kaspari, M. 2005. Global energy gradients and size in colonial organisms: Worker mass and worker number in ant colonies. Proceedings of the National Academy of Sciences USA, 102: 5079-5083. Kaspari, M. & O'Donnell, S. 2003. High rates of army ant raids in the Neotropics and implications for ant colony and community structure. Evolutionary Ecology Research, 5: 933-939.
Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami 47
Krebs, C. 1999. Ecological methodology, 2nd edition. Addison Wisley Educational Publishers Inc., Menlo Park, CA. Lambert, F. R. 1992. The consequences of selective logging for Bornean lowland forest birds. Philosophical Transactions of the Royal Society of London B, 335: 443-457. Laurance, W. F. & Laurance, S. G. W. 1996. Responses of five arboreal marsupials to recent selective logging in tropical Australia. Biotropica, 28: 310-322. Levieux, J. 1983. The soil fauna of tropical savannas. IV. The ants. In Tropical savannas, ed. F. Bourliere, pp. 525-540. Elsevier, Amsterdam. Linksvayer, T. A. & Janssen, M. A. 2009. Traits underlying the capacity of ant colonies to adapt to disturbance and stress regimes. Systems Research and Behavioral Science, 26: 315-329. Lobry de Bruyn, L. A. & Conacher, A. J. 1994. The bioturbation activity of ants in agricultural and naturally vegetated habitats in semi-arid environments. Australian Journal of Soil Research, 32: 555-557. Magnusson, W. E., de Lima, O. P., Reis, F. Q., Higuchi, N. & Ramos J. K. 1999. Logging activity and tree regeneration in an Amazonian forest. Forest Ecology and Management, 113: 67-74. Majer, J. D. 1983. Ants: Bio-indicators of minesite rehabilitation, land-use, and land conservation. environmental management
, 7: 375-383. McGlynn, T. P., Shotell, M. D. & Kelly, M. S. 2003. Responding to a variable environment: Home range, foraging behavior, and nest relocation in the Costa Rican rainforest ant Aphaenogaster araneoides. Journal of Insect Behavior, 16: 687-701. McGlynn, T. P., Carr, R., Carson, J. & Buma, J. 2004. Frequent nest relocation in the ant Aphaenogaster araneoides: Resources, competition, and natural enemies. Oikos, 106: 611-621. McNab, B. K. 1984. Physiological convergence amongst ant-eating and termite-eating mammals. Journal of Zoology, 203: 485-510. Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. 2000. Biodiversity hotspots for conservation priorities. Nature, 403: 853858. Ness, J. H., Morin, D. F. & Giladi, I. 2009. Uncommon specialization in a mutualism between a temperate herbaceous plant guild and an ant: Are Aphaenogaster ants keystone mutualists? Oikos, 118: 1793-1804. Okuda, T., Suzuki, M., Adachi, N., Quah, E. S., Hussein, N. & Manokaran, N. 2003. Effect of selective logging on canopy and stand structure and tree species composition in a lowland dipterocarp forest in Peninsular Malaysia. Forest Ecology and Management, 175: 297-320. Olson, D. M. & Andriamiadana, A. 1996. The effects of selective logging on the leaf litter invertebrate community of a tropical dry forest in western Madagascar. In Ecology and economy of a tropical dry
forest in Madagascar, eds. J. U. Ganzhorn & J.-P. Sorg. Primate Report, 175-187. Olson, D. M. & Ward, P. S. 1996. The ant fauna (Hymenoptera: Formicidae) of Kirindy forest (tropical dry forest) in western Madagascar. In Ecology and economy of a tropical dry forest in Madagascar, eds. J. U. Ganzhorn & J.-P. Sorg. Primate Report, 46: 161-164. Peters, S. L., Malcolm, J. R. & Zimmerman, B. L. 2006. Effects of selective logging on bat communities in the southeastern Amazon. Conservation Biology, 20: 14101421. Post, D. M. 2002. Using stable isotopes to estimate trophic position: Models, methods, and assumptions. Ecology, 83: 703-718. Punttila, P., Haila, Y., Pajunen, T. & Tukia, H. 1991. Colonisation of clear-cut forests by ants in the southern Finnish taiga: A quantitative survey. Oikos, 61: 250-262. Sanders, N. J. & Gordon, D. M. 2002. Resources and the flexible allocation of work in the desert ant Aphaenogaster cockerelli. Insectes Sociaux, 49: 371379. Sekercioglu, C. H. 2002. Effects of forestry pracices on vegetation structure and bird community of Kibale National Park
, Uganda. Biological Conservation
, 107: 229-240. Shattuck, S. O. 2008. Australian ants of the genus Aphaenogaster (Hymenoptera: Formicidae). Zootaxa, 1677: 25-45. Sorg, J. P. & Rohner, U. 1996. Climate and tree phenology of the dry deciduous forest of the Kirindy Forest. In Ecology and economy of a tropical dry forest in Madagascar, eds. J. U. Ganzhorn & J.-P. Sorg. Primate Report, 46: 57-80. Sorg, J. P., Ganzhorn. J. U. & Kappeler, P. M. 2003. Forestry and research in the Kirindy Forest/Centre de Formation Professionnelle Forestiиre. In The natural history of Madagascar, eds. S. M. Goodman & J. P. Benstead, pp. 1512-1519. University of Chicago Press, Chicago. Sorvari, J. & Hakkarainen, H. 2007. Wood ants are wood ants: deforestation causes population declines in the polydomous wood ant Formica aquilonia. Ecological Entomology
, 32: 707-711. Sorvari, J. & Hakkarainen, H. 2009. Forest clear-cutting causes small workers in the polydomous wood ant Formica aquilonia. Annales Zoologici Fennici, 46: 431438. Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics, 15: 353-391. Sutherland, W. J. 2006. Ecological census techniques, 2nd edition. Cambridge University Press, Cambridge. Tillberg, C. V., McCarthy, D. P., Dolezal, A. G. & Suarez, A. V. 2006. Measuring the trophic ecology of ants using stable isotopes. Insectes Sociaux, 53: 65-69. Tschinkel, W. R. 2011. The nest architecture of three species of north Florida Aphaenogaster ants. Journal of Insect Science, 11: 1-30.
48 Dittmann et al.: Investigating behavior and ecology of Aphaenogaster swammerdami
Tschinkel, W. R., Adams, E. S. & Macom, T. 1995. Territory area and colony size in the fire ant Solenopsis invicta. Journal of Animal Ecology, 64: 473-480. Uhl, C. 1989. Ecological impacts of selective logging in the Brazilian Amazon: A case study from the Paragominas region of the state of Parб. Biotropica, 21: 98-106. Vasconcelos, H. L., Vilhena, J. M. S. & Caliri, G. J. A. 2000. Responses of ants to selective logging of a central Amazonian forest. Journal of Applied Ecology
, 37: 508-514. Voigt, A. F., Burkhardt J. F., Verhaag M. & BцhningGaese, K. 2002. Regional differences in ant community structure and consequences for secondary dispersal of Commiphora seeds. Ecotropica, 8: 59-66.
Walsh, P. D., Henschel, P., Abernethy, K. A., Tutin, C. E. G., Telfer, P. & Lahm S. A. 2004. Logging speeds little red fire ant invasion of Africa. Biotropica, 36: 637-641. Widodo, E. S., Naito, T., Mohamed, M. & Hashimoto, Y. 2004. Effects of selective logging on the arboreal ants of a Bornean rainforest. Entomological Science, 7: 341349. Wilson, E. O. 1984. The relation between caste ratios and division of labor in the ant genus Pheidole (Hymenoptera, Formicidae). Behavioral Ecology and Sociobiology, 16: 89-98.
MT Dittmann, M Dammhahn, PM Kappeler