Why Do Pheidole Oxyops (Forel, 1908) Ants Place Feathers Around Their Nests?

Ecological Entomology (2019), DOI: 10.1111/een.12722

NATURAL HISTORY

Why do Pheidole oxyops (Forel, 1908) ants place feathers around their nests?

INÁCIO J. M. T. GOMES,1 DIOGO F. SANTIAGO,2 RICARDO I. CAMPOS1 and HERALDO L. VASCONCELOS3 1Programa de Pós-Graduação em Ecologia, Departamento de Biologia Geral, Universidade Federal de Viçosa, Viçosa, Brazil, 2Programa de Pós-Graduação em Ecologia e Recursos Naturais, Departamento de Ecologia e Biologia Evolutiva, Universidade Federal de São Carlos, São Carlos, Brazil and 3Instituto de Biologia, Universidade Federal de Uberlândia, Uberlândia, Brazil

Abstract. 1. The ant species Pheidole oxyops (Hymenoptera: Formicidae) has the particular behaviour of placing feathers along the edge of entrances to its nests. Here two hypotheses concerning this behaviour were tested: (i) P. oxyops, due to its essentially carnivore diet and particular nest structure (which acts as pitfall traps), uses feathers to enhance the capture of arthropods; and (ii) P. oxyops uses feathers as a strategy for obtaining water, as feathers could act as retainers of night moisture. 2. To test the first hypothesis, two plots with 20 pitfall traps were established, and inside each plot feathers were placed around half of these traps. It was expected that pitfall traps with feathers would collect a larger number of individuals and species of arthropods. To test the second hypothesis, all feathers were removed from 28 nests and water was supplied (wet cotton balls) to half of these nests. It was expected that colonies with access to an artificial water source would collect fewer feathers than control colonies. 3. Only the first hypothesis was supported. Our results indicate that feathers have the potential to enhance prey capture by P. oxyops providing the colonies with a more diverse and abundant diet. The ants’ habit of placing feathers around the nest, which can be considered as a low-cost foraging strategy, could be particularly important during periods of low food resource availability.

Key words. Dominant species, foraging behaviour, Formicidae, water stress.

Introduction In spite of being conspicuous, one as yet un-investigated behavioural trait is found in the Neotropical ant species Pheidole The life histories of ants comprise an enormous range oxyops (Forel, 1908), which places bird feathers (e.g. from of behavioural attributes that allow them to exploit virtu- doves) around the entrances of its nests (Fig. 1a). Colonies of ally all terrestrial habitats (Hölldobler & Wilson, 1990; Kaspari, P. oxyops are frequently observed collecting feathers in the 2000). However, we know little about how those behavioural vicinity of their nests when these are available within their features influence local habitat adaptations and, consequently, foraging area (I.J.M.T. Gomes, unpublished). This behaviour is ant species distribution (Laskis & Tschinkel, 2009; Rosumek, possibly associated with the ant’s foraging strategy, but the exact 2017). Adaptation to a given habitat includes, among many other mechanisms that trigger it are still unknown. requirements, an efficient foraging strategy for gathering food Pheidole is by far the richest ant genus in the Neotropics, with and water (Ribeiro & Navas, 2008 and references therein; Lanan, 702 described species (AntWeb; http://www.antweb.org). In the 2014). Because of that, ants have developed a myriad of different Neotropical savannas, P. oxyops ants play a highly important foraging strategies that can vary according to species, resources ecological role, because they are essentially predators and are or environmental conditions (Traniello, 1989; Lanan, 2014). regarded as both numerically and behaviourally dominant (Cam- pos et al., 2011; Czaczkes et al., 2011). This predominantly predatory behaviour may be associated with the behaviour of Correspondence: Inácio Gomes, Prédio da Entomologia, sala 210, placing feathers around the nest. Previous studies have shown Campus Universitário, Universidade Federal de Viçosa, 36570-000, that nests of P. oxyops have a peculiar structure, with a wide, Viçosa-MG, Brazil. E-mail: [email protected] deep, and smooth entrance tunnel (Fig. 1b), which is thought to

areas. We identified P. oxyops using the taxonomic key provided by Wilson (2003), and voucher specimens are deposited at the Zoological Collection of the Federal University of Uberlândia, Uberlândia, Brazil. To verify the occurrence of feathers around the nests of P. oxyops and the possible effect of seasonality on this behaviour, we performed two exploratory samplings. First, in the campus of the Federal University of Uberlândia, we counted the number of feathers around 17 P. oxyops nests in the dry season (in 2010) and around another 19 nests in the wet season (in 2018). Second, Fig. 1. Nests of Pheidole oxyops with (a) or without (b) bird feathers to test if the number of feathers collected by the ants over 24 h around its entrance. The wide, deep and smooth entrance of the nest (b) was dependent on the number of initial feathers, for those same is thought to work as a pitfall trap. Photographs: Ricardo Solar. [Colour nests and also in both seasons (dry and wet), we offered 15 new figure can be viewed at wileyonlinelibrary.com]. feathers positioned 30 cm away from their entrances and counted the number of new feathers that had been collected after 24 h. work as a pitfall trap, enhancing the capture of arthropod prey (Fowler, 1987; Forti et al., 2007; Supporting information, File Hypothesis 1: Feathers increase prey capture S1). Feathers may thus further enhance prey capture, given that they may work as bait (due to their smell, colour or shape) to To test whether feathers increase the number of arthropods some arthropods that end up falling into the nest entrance tun- falling into the nests and thus potentially become prey to P. o x y - nel. In addition, it is possible that feathers directly benefit ants, ops, we established in February 2011 two sampling plots inside as they could obtain food from small residues of bird tissue or the savanna reserve described earlier. The plots were located from prey present in the feathers. 100 m away from each other and inside each plot we installed Pheidole oxyops is a species typical of Brazilian savannas 20 pitfall traps (diameter 5.2 cm, volume 70 ml) along four lines (Wilson, 2003). As such, it often faces water shortage during (5 m apart) with five traps each (5 m apart). The pitfall traps were the dry season (Alvares et al., 2013). The same restriction buried in the ground, so their openings were level with the soil is experienced by Diacamma rugosum (Le Guillou, 1842), a surface. To simulate a P. oxyops nest entrance, we used pitfall common and conspicuous ant species that lives in dry areas traps with approximately the same entrance size as mature nests in Asia. Interestingly, this species was also found to place (diameter 4.3 cm). The openings of the pitfall traps were closed feathers around its nest entrances, and previous studies have for 48 h to minimise the effects of soil disturbance in attracting shown that D. rugosum consumes the water that accumulates on soil arthropods into the traps (dig-in effect). After 48 h, the traps feathers overnight (Moffett, 1985). Therefore, it is reasonable to were opened and partially filled with water and a few drops suppose that, similar to D. rugosum, P. oxyops may use feathers of soap to break the water’s surface tension. Within each plot to obtain water. In addition, we have observed that ants are we randomly selected 10 pitfalls where we placed 15 feathers recruited quickly and in large numbers to moistened cotton balls around them (‘feathers-addition group’). The remaining 10 traps (I.J.M.T. Gomes, unpublished), which indicates that P. oxyops served as controls (‘no-feathers group’). These experimental experiences water shortage during the dry season. If that is the feathers were obtained from our study site and, before placing case, one would expect that feather retrieval rates would increase them around the pitfalls, we carefully checked all feathers for during periods of severe water shortage. the presence of small animals. After 48 h, we collected the pitfall Aiming to identify the factors that trigger this behaviour and traps and took them to the laboratory, where the arthropods were its possible importance in the foraging strategy of P. oxyops,we sorted into morphospecies. If feathers do enhance prey capture, attempted to answer the following question: why do P. oxyops we expected that traps with feathers would collect more arthro- ants place feathers around their nests? In order to answer this pod species and individuals than would traps with no feathers. question, we tested two non-excluding hypotheses: (i) feathers We performed a complementary sampling to compare the increase arthropod prey capture by P. oxyops colonies; and (ii) arthropod species composition and abundance between the feathers act as moisture retainers, and thus increase the supply of ‘pitfall trapping’ and ‘active foraging’ strategies performed by water to colonies, being collected more frequently when water P. oxyops. For this purpose, we selected five nests and, using a is in short supply. forceps, removed from the ants’ mandibles all the food items brought back the nest. This procedure was always done between 09.00 and 10.00 hours in five nests and over three consecutive Materials and methods days. Food items were preserved in 70% ethanol and brought to the laboratory for identification. This study was conducted in two locations: in the campus of the Federal University of Uberlândia and at ‘Reserva Ecológica do Panga’ savanna (cerrado) reserve, both located in the city Hypothesis 2: Feathers as moisture retainers of Uberlândia, Minas Gerais, Brazil (18∘54′S, 48∘15′W). In this region, P. oxyops is commonly found in a variety of habitats, To test whether ants collect more feathers when deprived including open savannas, grasslands, forests, and even urban of water, we conducted a second experiment in June 2010,

© 2019 The Royal Entomological Society, Ecological Entomology, doi: 10.1111/een.12722 Why do Pheidole oxyops ants use feathers? 3 during the local dry season. We selected 28 P. oxyops nests that were located in the campus of the Federal University of Uber- lândia. Before the beginning of the experiment, we removed all feathers that were found around each of the 28 nests. After 24 h, to simulate a water source for the colonies, we deposited five hydrophilic cotton balls (0.6 g each) around the nest entrances of 14 randomly selected nests (‘water-addition group’). Cotton balls were moistened every 12 h for three consecutive days. It is important to note that during the experiment there was strong evidence that the experimental ant colonies were experiencing water deficit. We can say this because ants were recruited in great numbers to cotton balls within a few minutes of being moistened. For another group of 14 randomly selected nests (no-water group), we also deposited identical hydrophilic cotton balls but this time with no water added. After 3 days, we placed on the ground and near (1 m apart) the entrance of all 28 nests (both treatments) a handful (c.10g)ofnewfeathers (length 3–8 cm). After 24 h, we counted the number of feathers collected by each nest. If feathers act as a water source for ants, we expected that the no-water group would collect more feathers than the water-addition group. Fig. 2. Number of new feathers collected per nest (over 24 h) during the dry and wet seasons. A total of 15 new feathers was offered to each nest. The horizontal line in each box indicates the median, box limits are Statistical analysis first and third distributional quartiles, whiskers extend to most extreme data point, and the asterisk indicates an outlying data point. To compare the number of feathers around the nest entrances between dry and wet seasons, we performed two separate generalised linear models (GLMs). The first used the number errors. For all tests, we performed analyses with the outliers of initial feathers found around the nests and the second removed and the results did not change. We conducted all anal- used the number of new collected feathers after 24 h. Both yses using the software r (R-Core-Team 2016). We performed of these were response variables and the season (wet and dry) LMER tests using the r package ‘lme4’ (Bates et al., 2015). was the explanatory variable for both models. Finally, to test whether the number of new feathers collected by workers after 24 h was dependent on the number of initial feathers, Results independent of the season, we performed another GLM using Feather retrieval during the dry and wet seasons the number of collected feathers as response variable and the number of initial feathers as explanatory variable. We specified The number of ant nests with feathers around the nest entrance a Gaussian error distribution and performed residual analyses did not differ between the dry and wet seasons. During the dry for all models and checked for the distribution of errors. season, 88.2% (N = 17) of the nests had feathers, whereas in the To test Hypothesis 1, we performed linear mixed effects mod- wet season, 94.7% (N = 19) had feathers. Similarly, the mean els (LMER) with the total number of species (richness) and number of feathers per nest did not differ between the dry and individuals (abundance) per pitfall as response variables, the wet seasons (dry, 15.9 ± 17.2; wet, 16.6 ± 14.3; F1,34 = 0.142, treatments (feathers-addition or no-feathers group) as explana- P = 0.888). However, once we offered new feathers, we found tory variables, and grid as a random effect. We first analysed that ants collected significantly more new feathers (after 24 h)in the data using all arthropods found in the traps, and then using the dry than in the wet season (F1,34 = 10.02, P = 0.003; Fig. 2). only those for which we have evidence – either from the litera- Finally, independent of the season, the number of new feathers ture or through direct observations – that they are consumed by collected by ants was not affected by the number of initial P. oxyops as prey. This excluded from our analysis the follow- feathers around their nests (F1,34 = 0.011, P = 0.92). ing groups: Psocoptera, Thysanoptera, Hymenoptera (not considering ants), juveniles and other less representative taxa. We specified a Gaussian error distribution and performed residual Hypothesis 1: Feathers increase prey capture analyses for all models and checked for the distribution of errors. To test Hypothesis 2, we conducted a GLM using the num- On average, pitfall traps with feathers collected 1.32 times ber of feathers deposited by ants around the nest entrance as more arthropod species [𝜒 2(1, N = 40) = 5.39, P = 0.028; response variable and the treatment (water-addition or no-water Fig. 3a] and 1.63 times more arthropod individuals [𝜒 2 (1, group) as the predictive variable. We specified a quasi-Poisson N = 40) = 10.29, P = 0.001; Fig. 3b] than pitfall traps without error distribution for overdispersed data (Crawley, 2012), per- feathers. A total of 617 arthropods (other than P. oxyops indi- formed residual analyses and checked for the distribution of viduals) were collected in traps with feathers and 378 in traps

Table 1. Arthropods collected in pitfall traps with or without feathers and arthropod prey retrieved by foraging workers of Pheidole oxyops (active foraging).

Arthropod Traps with Traps without Active groups feathers feathers foraging

Acari 3.6 ± 0.7 2.5 ± 0.5 0 Araneae 0.5 ± 0.1 0.4 ± 0.1 0.6 ± 0.2 Blattodea 0.3 ± 0.1 0.1 ± 00 Coleoptera 0.3 ± 0.1 0.5 ± 0.1 1 ± 0.4 Collembola 9.1 ± 1.2 7 ± 1.3 0 Diptera 1.2 ± 0.3 1.1 ± 0.2 3 ± 0.5 Formicidae† 13 ± 2.1 5 ± 1.2 1.6 ± 0.5 Hemiptera 0.6 ± 0.1 0.3 ± 0.1 0.4 ± 0.4 Hymenoptera‡ 0.4 ± 0.1 0.4 ± 0.1 0 Isoptera 0.2 ± 0.1 0.1 ± 0 13.6 ± 3.4 (b) Psocoptera 0.4 ± 0.2 0.2 ± 0.1 0.2 ± 0.2 Thysanoptera 0.9 ± 0.3 1.3 ± 0.4 0 Others 0.1 ± 00.1± 0.1 0.6 ± 0.3 Total 617 items 378 items 107 items

†Except P. oxyops. ‡Except ants. Values represent the mean number of individuals per trap/nest ± SEM.

Hypothesis 2: Feathers as moisture retainers

When we experimentally manipulated the availability of water to P. oxyops, we found no difference in the number of feathers collected by nests that received or did not receive a sup-

ply of water (F1, 26 = 0.217, P = 0.645). Ants from nests of the water-addition group collected 8.29 ± 9.05 (mean ± SD) feathers in 24 h, and those from the control (no-water) group collected 6.79 ± 7.37 feathers.

Discussion

As anticipated by Hypothesis 1, pitfall traps baited with feathers collected significantly more arthropods than did those with no feathers. The primary reason for this could be that the smell, colour or shape of feathers enhances the capture of arthropod Fig. 3. (a, b) Differences in the number of morphospecies (richness) (a) prey. Considering that P. oxyops ants build nests that can act as and individuals (abundance) (b) of arthropods collected between pitfall pitfall traps (Fowler, 1987), it is reasonable to expect that feath- traps with or without bird feathers around their openings. The horizontal ers will present the same capture-enhance effect on real nests. line in each box indicates the median, box limits are first and third Pheidole oxyops seemstorelyontwobasicstrategiesfor distributional quartiles, whiskers extend to most extreme data point, and capturing prey. First, the ant workers forage many meters away asterisks indicate outlying data points. from their nests, actively seeking prey and, whenever necessary (e.g. when locating a large prey), they recruit a massive number without feathers. Analyses in which we excluded the arthropods of nestmates (Fowler, 1984; Czaczkes & Ratnieks, 2012). that did not represent food items to P. oxyops produced the Second, they can also use a passive strategy for gathering food same results; pitfall traps with feathers collected 1.44 times by constructing nests that work as pitfall traps (Fowler, 1987). more species [𝜒 2(1, N = 40) = 12.27, P < 0.001] and 1.74 times Our results indicate that by using feathers to enhance prey more individuals [𝜒 2(1, N = 40) = 12.72, P < 0.001] than traps capture, P. oxyops seems to improve this passive and therefore without feathers. less expensive strategy of gathering food. Ants (other than P. oxyops), Acari and Collembola were the Classic reviews of ant-foraging behaviours have never listed most common arthropods found in the pitfall traps (with or traps as a foraging strategy (e.g. Traniello, 1989; Lanan, without feathers). By contrast, the most common groups of 2014). However, here we give additional support to the idea arthropods captured by P. oxyops through active foraging were that nests of P. oxyops act as pitfall traps (Fowler, 1987). To termites and Diptera (Table 1). our knowledge, the use of traps by ants was only previously

© 2019 The Royal Entomological Society, Ecological Entomology, doi: 10.1111/een.12722 Why do Pheidole oxyops ants use feathers? 5 described in Allomerus decemarticulatus (Myrmicinae), a deserve further investigation. For instance, feathers might act plant-ant that lives in mutualistic association with Hirtella as a camouflage, thereby protecting the nests against predators. physophora (Chrysobalanaceae). In this system, ants construct Another possibility is that the deposition of feathers around the a spongy and gooey platform that can grab large insects that land nests might be a response to environmental harsh conditions, on it (Dejean et al., 2005). In both cases, the use of traps could such as flooding (Kolay & Annagiri, 2015) or heating. Never- represent an energy saving and is also related to diversification theless, in our study region P. oxyops only occurs in non-flooded of the ant diet (Dejean et al., 2005). areas. Furthermore, it is unlikely that feathers provide signifi- Here, we showed that P. oxyops could also be using feathers cant thermal insulation, given that P. oxyops nests are very deep to increase diet diversity. When foraging actively, P. oxyops ants (reaching 5 m in depth) (Forti et al., 2007) and nests with feath- prey especially on termites and flies, whereas the most common ers are found both in open and closed habitats (e.g. semidecidu- arthropod groups sampled in the feather-addition pitfall traps ous forests). were Collembola, Acari and ants (Table 1), which are known In short, the results of our study indicate that placing feathers to be consumed by generalist Pheidole species such as P. around nest entrances is a common behaviour in P. oxyops and, oxyops (Byrne, 1994; Wilson, 2003, 2005). This indicates that although this strategy seems to enhance prey capture, we found feathers not only increase but also diversify the food resources little evidence that it is influenced by nest water shortage. The available to P. oxyops colonies. It has already been found use of feathers to increase arthropod capture seems to be less that diet diversification plays an important role in ant colony costly than active prey hunting. Moreover, the use of feathers survivorship, as it increases the chances of finding and capturing seems to enhance diet diversification, which is an important prey during periods of resource scarcity (Traniello, 1989; Pirk strategy to deal with food shortage and competition (Pirk et al., et al., 2009). In addition, it is expected that ants adapt their 2009). In fact, the optimal foraging theory predicts that the foraging strategy according to resource availability (Gordon, ecological success of a given species depends on its ability 2002). Finally, it is reasonable to suppose that strategies that to adjust its foraging strategies to maximise food intake with increase diet diversity (such as P. oxyops feathers) would be the least amount of energy possible (Stephens & Krebs, 1986). positively selected along evolutionary time, especially bearing in Despite the fact that we have shown that feathers enhanced mind that the cerrado is a savanna that experiences a strong dry prey capture by P. oxyops, we do not yet know whether such season, during which there is a marked decline in the availability an increment in the food supply affects colony fitness. Future of arthropod prey (Vasconcelos et al., 2009). However, further studies should thus focus on how feathers and other ‘predatory studies that evaluate the fitness consequences of the feather traps’ could benefit the reproductive success of P. oxyops and retrieval behaviour displayed by P.oxyops are needed to evaluate other predatory ant species. this hypothesis. In our exploratory samplings, we found that P. oxyops col- Acknowledgements lected more feathers in the dry than in the wet season. How- ever, seasonal differences in feather retrieval rates could also be We are grateful for the logistical support provided by the Fed- explained by seasonal differences in prey availability. In fact, eral University of Uberlândia. Financial support was provided a previous study in our study area showed that the abundance by Coordenação de Aperfeiçoamento de Pessoal de Nível Supe- of litter arthropods is much smaller in the dry than in the wet rior (undergrad fellowship to IJMTG), Fundação de Amparo à season (Vasconcelos et al., 2009). Therefore, it could well be Pesquisa do Estado de Minas Gerais and Conselho Nacional possible that P. oxyops collects more feathers in the dry season de Desenvolvimento Científico e Tecnológico (CNPq grants to enhance the rate of ‘passive’ prey capture during a period of 302588/2015-9 and 441225/2016-0 to HLV). Finally, we are prey shortage. Furthermore, and contrary to our expectations, grateful to Ricardo Solar for kindly providing some photos. our experiment showed no difference in the number of collected All parts of this work comply with the current research laws feathers between the water-addition and no-water nest groups. of Brazil. The authors have no conflict of interest to declare. This lack of difference might be explained by two main factors. First, it is important to note that this experiment was conducted Author contributions over a short period of time; it is possible that 3 days were not enough to produce significant changes in water availability to the IJMTG and HLV conceived the initial study and carried out nest. Second, it is possible that collection of feathers is an innate the water stress survey experiment. DFS conducted the pitfall behaviour related solely to increasing prey capture, which means trapping experiments. IJMT performed the analyses and wrote that ants will collect feathers regardless of the amount of water the manuscript with support from HLV, RIC and DFS. available for the colonies. In accordance with this last hypothesis, we found that, independent of the season, the number of new collected feathers is not influenced by the number of initial Supporting Information feathers already placed around the nests. Finally, due to these Additional supporting information may be found online in the contradictory results and methodological issues, we cannot dis- Supporting Information section at the end of the article. regard the possibility that feather retrieval behaviour displayed by P. oxyops represents a strategy for obtaining water. File S1. Video of Pheidole oxyops nests acting as a pitfall trap. It Apart from the two hypotheses tested here, there are other shows a Camponotus sp. ant unsuccessfully trying to climb and hypotheses concerning feather-collection behaviour by ants that leave the nest.

References Kolay, S. & Annagiri, S. (2015) Dual response to nest flooding during monsoon in an Indian ant. Scientific Reports, 5,1–9. Alvares, C.A., Stape, J.L., Sentelhas, P.C., De Moraes Gonçalves, Lanan, M.C. (2014) Spatiotemporal resource distribution and forag- J.L. et al. (2013) Köppen’s climate classification map for Brazil. ing strategies of ants (Hymenoptera: Formicidae). Myrmecologi- Meteorologische Zeitschrift., 22, 711–728. cal news / Osterreichische Gesellschaft fur Entomofaunistik., 20, Bates, D., Mächler, M., Bolker, B.M. & Walker, S.C. (2015) Fitting lin- 53–70. ear mixed-effects models using lme4. Journal of Statistical Software, Laskis, K.O. & Tschinkel, W.R. (2009) The seasonal natural history of 67, 1–48. the ant, Dolichoderus mariae , in northern Florida. Journal of Insect Byrne, M.M. (1994) Ecology of twig-dwelling ants in a wet lowland Science., 9, 1–26. tropical Forest. Biotropica, 26, 61. Moffett, M.W. (1985) An Indian ant novel method for obtaining water. Campos, R.I., Vasconcelos, H.L., Andersen, A.N., Frizzo, T.L.M. National Geographic Research., 1, 146–149. et al. (2011) Multi-scale ant diversity in savanna woodlands: an Pirk, G.I., De Casenave, J.L., Pol, R.G., Marone, L. et al. (2009) intercontinental comparison. Austral Ecology., 36, 983–992. Influence of temporal fluctuations in seed abundance on thediet Crawley, M.J. (2012) The R Book. Wiley Publishing, Chichester, UK. of harvester ants (Pogonomyrmex spp.) in the central Monte desert, Czaczkes, T.J., Nouvellet, P. & Ratnieks, F.L.W. (2011) Cooperative Argentina. Austral Ecology., 34, 908–919. food transport in the Neotropical ant, Pheidole oxyops. Insectes R-Core-Team (2016) –R: A Language and Environment for Statisti- Sociaux., 58, 153–161. cal Computing. [WWW document]. URL http://www.r-project.org/. Czaczkes, T.J. & Ratnieks, F.L.W. (2012) Pheromone trails in the Brazil- [accessed on 27 September 2016] ian ant Pheidole oxyops: extreme properties and dual recruitment Ribeiro, P.L. & Navas, C.A. (2008) Colony dehydration and water action. Behavioural Ecology and Sociobiology., 66, 1149–1156. collection by specialized caste in the leaf-cutting ant Atta sexdens Dejean, A., Solano, P.J., Ayroles, J., Corbara, B. & Orivel, J. (2005) rubropilosa. Journal of Insect Behavior., 21, 549–558. Insect behaviour: arboreal ants build traps to capture prey. Nature, Rosumek, F.B. (2017) Natural history of ants: what we (do not) 434, 973–973. know about trophic and temporal niches of neotropical species. Forti, L.C., Camargo, R.S., Fujihara, R.T. & Lopes, J.F.S. (2007) Sociobiology, 64, 244–255. The nest architecture of the ant, Pheidole oxyops Forel, 1908 Stephens, D.W. & Krebs, J.R. (1986) Foraging Theory. Princeton (Hymenoptera: Formicidae). Insect Sci., 14, 437–442. University Press, Princeton, New Jersey. Fowler, H.G. (1984) Recruitment, group retrieval and major worker Traniello, J.F.A. (1989) Foraging strategies of ants. Annual Review of behavior in Pheidole oxyops Forel (Hymenoptera: Formicidae). Entomology., 34, 191–210. Revista Brasileira de Biologia., 44, 21–24. Vasconcelos, H.L., Pacheco, R., Silva, R.C., Vasconcelos, P.B. et al. Fowler, H.G. (1987) The nest of Pheidole oxyops as a tipfall-trap (2009) Dynamics of the leaf-litter arthropod fauna following fire in (Hymenoptera: Formicidae). Pedobiologia, 30, 373–376. a neotropical woodland savanna. PLoS One, 4,1–9. Gordon, D.M. (2002) The regulation of foraging activity in red harvester Wilson, E.O. (2005) Oribatid mite predation by small ants of the genus ant colonies. The American Naturalist., 159, 509–518. Pheidole. Insectes Sociaux., 52, 263–265. Hölldobler, B. & Wilson, E.O. (1990) The Ants. Massachussets, Harvard Wilson, E.O. (2003) Pheidole in the New World. Harvard University University Press, Cambridge, Massachusetts. Press, Cambridge, Massachusetts. Kaspari, M. (2000) A primer on ant ecology. Ants: Standard Methods for Measuring and Monitoring Biodiversity (ed. by D. Agosti, J. D. Majer, L. E. Alonso and T. R. Schultz), pp. 9–24. Smithsonian Accepted 13 January 2019 Institution Press, Washington, DC. Associate Editor: Sophie Evison