Canadian Journal of Zoology

Transcription

1 CONTRASTS IN BODY SIZE AND GROWTH SUGGEST THAT HIGH POPULATION DENSITY RESULTS IN FASTER PACE-OF- LIFE IN DAMARALAND MOLE-RATS (FUKOMYS DAMARENSIS) Journal: Manuscript ID cjz r1 Manuscript Type: Article Date Submitted by the Author: 20-Dec-2017 Complete List of Authors: Finn, Kyle; Rhodes University Department of Zoology and Entomology, Wildlife and Reserve Management Research Group; Kalahari Research Centre, Kalahari Mole-rat Project Parker, Daniel; Rhodes University, Wildlife and Reserve Management Research Group, Department of Zoology and Entomology; University of Mpumalanga, School of Biology and Environmental Sciences Bennett, Nigel; University of Pretoria, Mammal Research Institute, Department of Zoology and Entomology Zöttl, Markus; University of Cambridge, Department of Zoology; Kalahari Research Centre, Kalahari Mole-rat Project; Linnaeus University Centre for Ecology and Evolution in Microbial model Systems, Department of Biology and Environmental Science Keyword: aridity food distribution hypothesis, Damaraland mole-rat, density dependence, ecological constraints, Fukomys damarensis, subterranean rodents, body size

2 Page 1 of 26 1 CONTRASTS IN BODY SIZE AND GROWTH SUGGEST THAT HIGH POPULATION DENSITY RESULTS IN FASTER PACE-OF-LIFE IN DAMARALAND MOLE-RATS (FUKOMYS DAMARENSIS) K. T. Finn 1,2, D. M. Parker 1,3, N. C. Bennett 4, and M. Zöttl 2,5,6 1 Wildlife and Reserve Management Research Group, Department of Zoology and Entomology, Rhodes University, Grahamstown, South Africa, kyletfinn@gmail.com; 2 Kalahari Mole-rat Project, Kalahari Research Centre, Van Zylsrus, South Africa; 3 School of Biology and Environmental Sciences, University of Mpumalanga, Nelspruit, South Africa, daniel.parker@ump.ac.za; 4 Department of Zoology and Entomology, Mammal Research Institute, University of Pretoria, Pretoria, South Africa, ncbennett@zoology.up.ac.za; 5 Department of Zoology, University of Cambridge, Cambridge, CB23EJ, United Kingdom, mz338@cam.ac.uk; 6 Ecology and Evolution in Microbial Model Systems, EEMiS, Department of Biology and Environmental Science, Linnaeus University, SE Kalmar, Sweden Corresponding author: K. T. Finn, Kalahari Research Centre, Van Zylsrus, Northern Cape 8467, South Africa, , kyletfinn@gmail.com

3 Page 2 of 26 2 CONTRASTS IN BODY SIZE AND GROWTH SUGGEST THAT HIGH POPULATION DENSITY RESULTS IN FASTER PACE-OF-LIFE IN DAMARALAND MOLE-RATS (FUKOMYS DAMARENSIS) K. T. Finn 1,2, D. M. Parker 1,3, N. C. Bennett 4, and M. Zöttl 2,5,6 ABSTRACT We studied the correlates of population density and body size, growth rates, litter size and group size in Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)) at two study sites with contrasting population densities. Group size, litter size and the probability of recapture were independent of study site. However, body size differed between the two study sites, suggesting that population density may affect life-history traits in social mole-rats. At the low-density site (0.13 groups/ha), individuals were significantly larger and subordinate males showed higher growth rates than at the high-density site (0.41 groups/ha), which may indicate that high population density in subterranean rodents enhances pace of life. The larger size of non-reproductive individuals at the lowdensity site could adapt individuals at lower population densities to larger dispersal distances. Keywords: aridity food distribution hypothesis, body size, Damaraland mole-rat, density dependence, ecological constraints, Fukomys damarensis, subterranean rodents

4 Page 3 of 26 3 Introduction Population densities are shaped by the availability and distribution of resources, and ultimately, the energy available to them in their specific habitats (Robinson and Redford 1986). The variation in habitat saturation across habitats may be attributed to food availability and, as such, population density increases due to greater quantity and/or improved quality of food (Banks and Dickman 2000). Often, competition for available resources increases with population density (Dahle and Swenson 2003) resulting in a reduction of body mass (Macdonald et al. 2002), growth rates (Leberg and Smith 1993), home range size (Dahle and Swenson 2003), and survival (Albon et al. 2000). As such, many fundamental aspects of behaviour and ecology are affected by population density (Stearns 1992). Density dependent effects on populations have been studied extensively in large mammals (Saether 1997) and terrestrial rodents (Kokko and Lundberg 2001), but rarely in subterranean rodents. Previous studies on social species of mole-rats have suggested that group size and body mass may vary with geophyte density and distribution (Brett 1991; Spinks et al. 2000a, b). However, it may be difficult to tease apart the effects of density and habitat quality on individual and group level dynamics since population density may not always be positively correlated with food availability (van Horne 1983). To date, the associations of population density and group size, body size, growth rates, and survival in social Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)) are unknown. Sociality in African mole-rats may be an adaptation to limited and clustered resources (Heterocephalus Rüppell, 1842; Cryptomys Gray,1864; Fukomys Kock, 2006; Jarvis et al. 1994). The aridity food distribution hypothesis (AFDH) has been put forward to explain how the social life-style may have arisen in this group (Jarvis et al. 1994). The ecological constraints mole-rats face may constrain offspring to remain in the group due to limited opportunities for both independent foraging and dispersal (Jarvis et al. 1994). However, increased group size also results in enhanced competition, which

5 Page 4 of 26 4 may reduce foraging efficiency (Selman and Goss-Custard 1988) and may lead to reduced growth rates (Young et al. 2015; Zöttl et al. 2016a). Additionally, intra-sexual competition increases with group size, resulting in reduced fecundity and juvenile survival especially in species where dispersal is difficult and philopatry is favoured (Lacey 2004; Young et al. 2006). Our study tested whether population density is associated to different group attributes (i.e. group and litter size) and whether population density predicts individual body size and growth of group members in two populations of wild Damaraland mole-rats with markedly different population densities. We used capture-mark and recapture data from the two sites to test if body mass, body size, and growth rates of individuals differed between two study sites with contrasting population density. A comparison of the relative differences between group and litter sizes was also undertaken to determine if these attributes were contingent on the study site. We predict that growth rates and body size will be decreased in the low-density site as seen in other mole-rats (Brett 1991; Spinks et al. 2000a). Materials and Methods Study Site We captured Damaraland mole-rat groups at two sites in the Northern Cape province of South Africa, the Kalahari Research Centre (KRC, S, E) and Tswalu Kalahari Reserve (Tswalu, S, E), between 2013 and A distance of only 70 km separated the two sites. The combined total size of the KRC study area was approximately 910 hectares (population density of 0.13 groups/ha) and Tswalu was approximately 140 hectares (population density of 0.41 groups/ha). Tswalu was categorized as the high-density site and KRC as the low-density site (Finn 2017). The highdensity site was an experimental site where groups were translocated to a neighboring burrow (within the study site) when we released them and thus the analyses in this paper are restricted to parameters which are unlikely to be affected by different release procedures. The landscape of the low-density site

6 Page 5 of 26 5 comprised arid thornveld with vegetated red arenosol-type sand dunes covered in grasses, bushes, and scattered trees (see Clutton-Brock et al. 1999). While the high-density site lacked the typical rolling sand dunes found further north, the red sand and vegetation was similar to that in the low-density site, except that there were fewer trees. The primary food of the study species, the gemsbok cucumber (Acanthosicyos naudinianus (Sond.)) and eland s bean (Elephantorrhiza elephatina (Burch.)), were present at both sites (Voigt et al. 2014). The summers (October April) in the region are very hot with a mean maximum daily temperature of 34.0 C (range C) and a mean minimum of 15.1 C (range C). During winter (May September) the mean minimum daily temperature is 3.1 C (range C) and a mean maximum of 25.0 C (range C). The majority of the annual rainfall results from sporadic heavy thunderstorms between December and March. The low-density site received an average of 255 mm per year (averaged over , range mm) with 213 mm falling during , 187 mm in , and 188 mm in The high-density site has a higher rainfall pattern compared to the low-density site receiving 357 mm per annum (averaged over , range mm) with 327 mm in , 206 mm in , and 229 mm in Study Species and Capture Methods The Damaraland mole-rat is a highly social species of subterranean rodent (Family Bathyergidae) endemic to the red Kalahari arenosols of southern Africa and inhabits intricate burrow systems in groups of up to 39 individuals, the majority of which are related (Jarvis and Bennett 1993). Reproduction is restricted to a single reproductive female and typically one or two unrelated male consorts, while subordinate individuals exhibit reproductive suppression, but are forced to delay dispersal due to the severe ecological constraints (Bennett and Jarvis 1988; Burland et al. 2004). As a result, subordinates remain within the natal group to assist with cooperative behaviours for the benefit of the entire group (Jarvis and Bennett 1993; Zöttl et al. 2016b).

7 Page 6 of 26 6 Between July 2013 and October 2016, 51 groups of mole-rats (683 individual mole-rats) were captured at the low-density site. Between March 2015 and April 2016, 57 groups (548 individual molerats) were captured at the high-density site. A group of mole-rats was defined as two or more individuals occurring in a single burrow system. We used modified Hickman traps (Hickman 1979) baited with sweet potato (Ipomoea batatas L.) or gemsbok cucumber. We attempted to recapture each group six months from initial capture, but in some cases this schedule was not possible. Consequently, some groups were recaptured up to 12 months from initial capture. Mole-rat burrows were located by digging a trench perpendicular to a line of surface mounds until breaching the tunnel. Traps were placed in the tunnel and checked every two hours (see Finn 2017). Captured individuals were anaesthetized on the day of capture with isoflurane inhalation (Parker et al. 2008), weighed to the nearest gram (Sartorius, Goettingen, Germany) and incisor width (TW; measured across the base of incisors) and head width (HW; from ear to ear) were measured mm using digital calipers (HM Müllner, Eugendorf, Austria; Young and Bennett 2013). Total length (TL; snout to tail tip) was measured dorsally + 1 mm with a tape measure (Young and Bennett 2013). Individuals >20 g were implanted with a subcutaneous transponder tag (Dorset Group, Aalten, The Netherlands) to uniquely identify them. Captured mole-rats were housed up to 10 days at C in artificial tunnel systems made from polyvinyl chloride (PVC) pipes with a plastic box filled with paper towels replicating a nest and provided with sweet potato and gemsbok cucumber ad libitum (Zöttl et al. 2016a). After the last individual was captured we waited up to 48 hours to ensure no other individuals were present then released the group into their burrow. All protocols were approved by the University of Pretoria ethics committee and complied with regulations stipulated in the Guidelines for the Use of Animals in Research (Permit ECO32-13). Statistical Analysis All statistical analyses were performed with R version (R Core Team 2016). Unless otherwise stated, data are reported as median + SD. Group size and within group sex ratios were

8 Page 7 of 26 7 compared using Wilcoxon rank-sum tests. The inter-site difference in original group biomass was tested with a Wilcoxon rank-sum test and intra-site differences in biomass between capture events were tested with a Kruskal-Wallis test. The inter-site differences in group litter size were tested with a Wilcoxon rank-sum test. Litter size was calculated by counting the number of individuals <85g in a cohort of similar sized individuals, then taking the average among cohorts. Additionally, in the rare cases where a reproductive female gave birth to a litter in the lab, that litter was also included. The mean percentage of individuals recaptured within groups between trapping events (i.e. the probability of recapture) was compared between sites using a Fisher Exact test. All inter-site comparisons of body mass and morphometrics [incisor width (TW), head width (HW), and total length (TL)] were performed with Wilcoxon rank-sum tests using the values obtained at first capture. Individuals were classed into age categories based on body mass and reproductive status. Individuals <50 g were considered juveniles and other non-reproductive individuals >50 g were considered subordinates (Young and Bennett 2013). The reproductive female was readily recognized by possessing prominent nipples and a perforate vagina (Bennett and Faulkes 2000; Burland et al. 2004). It is more difficult to identify the reproductive male and thus we assumed the heaviest male in the group to be the breeder male (Bennett and Faulkes 2000). However, multiple males can be breeders in a group (Burland et al. 2004), thus any male group members within 5 g of the heaviest male were also considered reproductive. Intra- and inter-site differences in body size were compared for juvenile, subordinate, and reproductive individuals. There were no significant differences in all measured features (TW, HW, TL, and body mass) between the sexes in juveniles, but subordinates and reproductives exhibited significant differences between the sexes. Thus, sexes were pooled for juveniles, but not for subordinates and reproductives to test for inter-site differences in body sizes. The growth rates in body mass, TW, HW, and TL were compared for the three classes by sex between locations using Wilcoxon rank-sum tests. To calculate the growth rate for each measurement,

9 Page 8 of 26 8 the value at first capture was subtracted from the measured value at second capture, and the difference divided by the number of months between capture. Based on Young and Bennett (2010) and O Riain et al. (2000) we also tested whether reproductive females have a different body shape to non-reproductive females and if body shape differed between locations. This elongation index was calculated by dividing head width by total length (Young and Bennett 2013). A Bonferroni correction was applied to all tests of repeated measurements (Dunn 1961). Results Group Level Analyses The mean group size was similar at the two sites (df = 106, means: the low-density site and the high-density site , p = 0.11) as was the mean biomass (df = 106, means: the low-density site g and the high-density site g, p = 0.48). The group biomass remained similar between capture events at both sites (the low-density site: df = 39, p = 0.47; the high-density site: df = 46, p = 0.47). The sex ratios (number of males divided by females) within groups were similar and slightly male-biased at the two locations (df = 22, means: the low-density site and the highdensity site , p = 0.27) and both sites had comparable litter sizes (df = 106, means: low-density site , the high-density site , p = 0.51). Of all the groups initially captured, and which we attempted to recapture, 85% at both locations were recaptured (df = 1, the low-density site n = 34 of 40, the high-density site n = 40 of 47, p = 1.0) while the remainder were never located and were presumed to be extinct or dispersed (i.e. all individuals died or dispersed). Individual Body Size Analyses Juvenile mole-rats showed no significant differences in body size between the two sites (Table 1). Subordinate females at the low-density site were significantly larger in HW, TL, and body mass than at the high-density site, while males were larger overall, but only significantly so in HW (Figure 1, Table

10 Page 9 of ). Reproductive females at the low-density site were significantly larger in all morphometric measurements and body mass, while reproductive males at the low-density site were significantly larger in TL and body mass (Figure 2, Table 1). The elongation index of both reproductive and non-reproductive females was comparable between locations (reproductive females: df = 111, low-density site 7.06, highdensity site 7.01, p = 0.96; non-reproductive females: df = 414, low-density site 6.80, high-density site 6.79, p = 0.45) and reproductive females had a higher elongation index compared to non-reproductive females (df = 525, reproductive females 7.01, non-reproductive females 6.6, p = 0.02). Juvenile males at the low-density site had a significantly greater growth rate in body mass only compared with those at the high-density site (df = 50, the low-density site 7.6 g*month , the high-density site 5.9 g*month , p < 0.01), while there were no significant differences in juvenile female growth rates (Figure 3). Subordinate males exhibited significantly higher growth rates overall at the low-density site, while there were no significant differences in the growth rates of subordinate females between locations (Figure 3). Reproductive males at the low-density site had a significantly higher growth rate in HW, but not in the other measured rates (Figure 3). There were no significant differences between the growth rates in reproductive females from the two locations (Figure 3). Discussion The primary goal of the study was to expand upon the limited data available on density dependent effects on group composition, body size, and growth rates in Damaraland mole-rats by comparing two populations of varying densities. As previously observed in blind mole rats (Spalax ehrenbergi Nehring 1898; Nevo et al. 1986) and common mole-rats (Cryptomys hottentotus Lesson 1826; Spinks et al. 2000b), our data showed that the population density of Damaraland mole-rats was lower at the arid study-site. Our study suggests that group size, biomass, litter size, and group persistence do not appear to be related to fluctuations in density, while body size and individual growth may be related to population density and habitat quality.

11 Page 10 of In naked mole-rats (Heterocephalus glaber Rüppell 1842), group size has been suggested to be density-dependent (Brett 1991), while in Damaraland mole-rats group size is thought to be related to geophyte availability as predicted by the AFDH (Jarvis et al. 1994). In arid habitats, the ecological constraints against dispersal favour philopatry, and a larger group size would increase group foraging efficiency and individual survival under such harsh living conditions (Jarvis et al. 1994; Spinks et al. 2000a). Surprisingly, instead of a larger group size in arid environments, group size was not found to differ either by population density, rainfall, or geophyte availability in the common mole-rat (Spinks et al. 2000a, b). The results of our study were similar, where group size between the arid low-density site and the more mesic high-density site was comparable despite differences in both population density and annual rainfall. It may be that the increased body size of individuals at the lower density site compensates for the expected increase in overall group size, since group biomass was also similar between locations despite the smaller size of individuals at the higher density site. An inverse relationship between juvenile recruitment and population density is expected in rodents with long reproductive tenure where dispersal is difficult and philopatry is favoured (Spinks et al. 2000b; Krebs 2003; Lacey 2004). Although we were unable to directly compare juvenile recruitment between the sites (see Finn 2017), we used litter size as a proxy for juvenile survival by counting cohorts of similar body mass. Our results indicate that population density had no effect on litter size in Damaraland mole-rats, possibly due to the isolated nature of their burrows, where each group is supposedly isolated from its neighbours (Jacobs et al. 1998). Freely passible tunnels have been found to connect neighboring burrow systems in Ansell s mole-rat (Fukomys anselli (Burda, Zima, Scharff, Macholán and Kawalika, 1999); Šklíba et al. 2012; Patzenhauerová et al. 2013) and giant mole-rats (Fukomys mechowii (Peters 1881); Šumbera et al. 2012), two social mole-rat species in Zambia. Due to the short distance between groups at the high-density site (Finn 2017), the presence of a freely passible tunnel between neighboring groups cannot be ruled out. In 47 groups recaptured at the high-density

12 Page 11 of site, no individuals from one group was captured in the adjacent groups. If individuals intrude into neighboring groups they may do so infrequently, or for very short durations. Thus, it appears unlikely that invaders from neighboring groups commit infanticide. At the low-density site, subordinates and reproductive individuals of both sexes were larger and heavier overall, while there was no significant difference in juvenile body size. Additionally, at the lowdensity site, juvenile and subordinate males exhibited significantly greater growth rates and reproductive males exhibited a greater, but not significant, increase in body mass, while there were no significant differences in the growth rates of females between sites. Thus the differences in early life growth rates may explain the resulting differences in adult male body size (sensu Ozgul et al. 2010), but not in the differences observed in subordinate and reproductive female body size at the two sites. These results contrast the results of Spinks et al. (2000a, b) and Nevo et al. (1986) who found that body mass decreased with population density and aridity. Spinks et al. (2000b) attributed this finding to an energy saving adaptation in the presence of a reduced geophyte food density. Perhaps the greater population density constrains groups to forage in a limited area, thus potentially exhausting nearby food resources in order to prevent competition with neighboring groups as predicted by the ideal despotic distribution model (Johnson 2007). As a result, individuals (or small groups) which are weak competitors are displaced by stronger conspecifics (M. Zöttl, unpublished data). We speculate that longevity may be reduced at the high-density site, for example due to increased inter-group competition, and therefore individuals were prevented from obtaining the body size observed in the low-density site due to the high competition for resources. Although we were unable to estimate ages of adult individuals at either site, the similar growth rates of reproductive females, but different body sizes at the two sites supports this hypothesis. Reproductive females at the high-density site must quickly amass a sufficient work force or the entire group risks being ousted from the territory by neighboring groups. Further monitoring of the

13 Page 12 of population at the high-density site will elucidate the relationship between group size, group displacement, and longevity to determine if it is negatively affected by population density. Recent studies on wild and captive Damaraland mole-rats have shown that males require more time than females to reach maximum body mass (Young et al. 2015; Zöttl et al. 2016a). These results would account for the differences in growth rates between the sexes in our study, where it appears wild males continue gaining body mass at a greater rate than females even after obtaining reproductive tenure. This could be due to the decreased growth rates in males at the mesic, high-density site, indicating that food availability may be reduced at that location (sensu Kjellander et al. 2005). The increased growth rates and larger size in males at the low-density site could benefit individuals during dispersal, increasing their chances for survival while locating a mate in a low-density environment (Jones 1988; Finn 2017). In food supplementation experiments in small rodents, increased food availability lead to increased body size (Banks and Dickman 2000). In both Spalax mole-rats in Israel and African mole-rats, food quality was positively correlated with annual rainfall, and as a result body size and population density increased (Nevo et al. 1986; Jarvis et al. 1998; Spinks et al 2000a). Damaraland mole-rats in Dordabis, Namibia were much larger than either of the populations in this study despite the location having an annual rainfall intermediate between our two sites and lacking gemsbok cucumbers (Jarvis et al. 1998; Young and Bennett 2013). Clearly annual rainfall alone does not explain the variation in body size in Damaraland mole-rats nor the presence of gemsbok cucumbers. The increased body size and growth rates observed at the arid low-density site in this study may indicate that an interaction between population density and habitat characters (annual rainfall, soil conditions, or geophyte availability) affects body size. Since social mole-rat groups exhibit high site fidelity, occur in regions characterized by seasonal sporadic rainfall, patchy distribution of geophytes and burrows, exhibit social dominance in the form of reproductive suppression, and have a high reproductive capacity throughout the life of the

14 Page 13 of reproductive female, it is possible that the typical positive relationship between population density and habitat quality deviates from the norm (van Horne 1983). While a vegetation analysis was not conducted for our study, we found fewer gemsbok cucumbers at the high-density site than the low-density site (K. Finn pers. obs.). If this observation was representative of the overall geophyte abundance and quality at the two study sites, it would corroborate the previous studies (Nevo et al. 1986; Jarvis et al. 1998; Spinks et al. 2000a), indicating that body size in mole-rats may be affected more by geophyte quality than population density. Without a detailed survey of the geophyte density and distribution (sensu Jarvis et al. 1998) at the locations, the relationship between geophyte quality, annual rainfall, population density, and body size in Damaraland mole-rats remains speculation. A limiting factor of this study is the concurrent group translocation experiment at the highdensity site. Thus, we were unable to directly compare juvenile recruitment and survival rates between sites (see Finn 2017), and the effects of density on these rates remain to be determined. The translocation study may also have affected growth rates at the high-density site. However, it is unlikely that different release procedures have caused a specific growth reducing effect, limited to juvenile and subordinate males. The lack of a vegetation survey combined with the differences in both annual rainfall and population density at the two sites is another limitation. The lack of replication at more sites with varying density also limits the scope to our study. We do not know whether our two study populations differ genetically and whether the observed differences in body size are the result of genetic differences arising between the two sites, or whether they may be a plastic response to contrasting environmental conditions. The relatively short distance between the sites suggests that it is more likely that the differences in body size are a result of phenotypic plasticity in response to local conditions. There is interest in the relationship between body mass, population density, reproductive performance, and resource availability because life history theories predict a trade-off between body mass and reproduction when resources are limited (Stearns 1992). Tests of these theories are difficult

15 Page 14 of since they require long-term data sets on an individual s reproductive success and body mass under varying levels of resource availability (Johnson 2007). While the longitudinal data presented here is still in its infancy, there is great potential for long term comparative work on this relationship using these two populations. The longevity of mole-rats and the varying density and body size of the two populations studied presents a useful combination to further study density-dependent reproductive performance over time and how it relates to habitat quality. Future work should quantify the differences in geophyte biomass, distribution, and density at the two study sites to disentangle the relationship between population density, habitat quality, and individual body size. Additionally, determining whether there is sufficient genetic exchange between the two areas would determine if the resulting differences are from an actual genetic divergence, or phenotypic plasticity resulting from differing habitat conditions. It would also be of interest to compare the reproductive performance of females to determine if density or habitat quality affects lifetime reproductive success as well as juvenile survival and recruitment. Acknowledgements Thanks to Jack Thorley and Tim Clutton-Brock for the use of their data. Thanks to E. Oppenheimer & Son, Duncan McFadyen, Gus van Dyk, and Dylan Smith for permission to conduct research at Tswalu Kalahari Reserve. Thanks to Tim Clutton-Brock, Marta Manser, David Gaynor, and the Kalahari Research Trust for permission to use the facilities and vehicles to conduct research at the KRC. Thanks to Cobus Lamprecht for permission to capture mole-rats on his farm. Discussions with David Gaynor provided insight on analytical methods, statistical tests, and interpretation of results. Thanks to Jack Thorley and Katy Goddard for managing captures at the KRC, and Sally Bournbush, Kyle Flesness, Johan Kjellberg Jensen, Simon Kershenbaum, Thomas Manning, Sean McGregor, Adam Mitchell, Jakov Musafija, Francesco Santi, Tash Waite, and Adrya Webb for assistance capturing. Two anonymous reviewers provided helpful comments and suggestions to improve the manuscript. We are grateful to the Northern Cape

16 Page 15 of Department of Environment and Nature Conservation for permission to conduct research in the Northern Cape. The field study was funded by a British Ecological Society Grant to M. Z. and by European Research Council grant to Tim Clutton-Brock. We thank the University of Pretoria for ethics clearance to conduct the research (Permit EC032-13). References Albon, S.D., Coulson, T.N., Brown, D., Guinness, F.E., Pemberton, J.M., et al Temporal changes in the key factor and the key age group influencing population dynamics. J. Anim. Ecol. 69(6): doi: /j x. Banks, P.B., and Dickman, C.R Effects of winter food supplementation on reproduction, body mass, and numbers of small mammals in montane Australia. Can. J. Zool. 78(10): doi: /z Bennett, N.C., and Jarvis, J.U.M The social structure and reproductive biology of colonies of the mole-rat Cryptomys damarensis (Rodentia, Bathyergidae). J. Mammal. 69(2): doi: / Bennett, N.C., and Faulkes, C.G African mole-rats: Ecology and Eusociality. Cambridge University Press, Cambridge, U.K. Brett, R.A The ecology of naked mole-rat groups: burrowing, food and limiting factors. In The Biology of the Naked Mole-rat. Edited by P.W. Sherman, J.U.M. Jarvis, and R.D. Alexander. Princeton University Press, Princeton, N.J. pp Burland, T.M., Bennett, N.C., Jarvis, J.U.M., and Faulkes, C.G Group structure and parentage in wild groups of cooperatively breeding Damaraland mole-rats suggests incest avoidance alone may not maintain reproductive skew. Mol. Ecol. 13: doi:10.111/j x x.

17 Page 16 of Caughley, G Eruption of ungulates populations, with emphasis on Himalayan Thar in New Zealand. Ecology 51(1): doi: / Clutton-Brock, T.H., Gaynor, D., McIlrath, G.M., MacColl, A.D.C., Kansky, R., et al Predation, group size and mortality in a cooperative mongoose, Suricata suricatta. J. Anim. Ecol. 68(4): doi: /j x. Dahle, B., and Swenson, J.E Home ranges in adult Scandinavian brown bears (Urus arctos): effect of mass, sex, reproductive category, population density and habitat type. J. Zool. (Lond.) 260(4): doi: /s Dunn, O.J Multiple comparisons among means. J. Am. Stat. Assoc. 56(293): doi: / Finn, K.T Density-dependent effects on body size, philopatry, and dispersal in the Damaraland mole-rat (Fukomys damarensis). M.Sc. thesis, Department of Zoology and Entomology, Rhodes University, Grahamstown, South Africa. Available from [accessed 3 February 2018] Hickman, G.C A live trap and trapping technique for fossorial mammals. S. Afr. J. Zool. 14(1): doi: / Hoogland, J.L The Black-Tailed Prairie Dog: Social Life of a Burrowing Mammal. University of Chicago Press, Chicago, I.L. Jacobs, D.S., Reid, S., and Kuiper, S Out-breeding behaviour and xenophobia in the Damaraland mole-rat, Cryptomys damarensis. S. Afr. J. Zool. 33(3): doi: / Jarvis, J.U.M., and Bennett, N.C Eusociality has evolved independently in two genera of bathyergid mole-rats but occurs in no other subterranean mammal. Behav. Ecol. Sociobiol. 33(4): doi: /bf

18 Page 17 of Jarvis, J.U.M., O Riain, M.J., Bennett, N.C., and Sherman, P.W Mammalian eusociality: a family affair. Trends Ecol. Evol. 9(2): doi: / (94) Jarvis, J.U.M., Bennett, N.C., and Spinks, A.C Food availability and foraging in wild groups of Damaraland mole-rat Cryptomys damarensis: implications for sociality. Oecologia, 113(2): doi: /s Johnson, M.D Measuring habitat quality: a review. Condor, 109(3): doi: / Jones, W.T Density-related changes in survival of philopatric and dispersing kangaroo rats. Ecology, 69(5): doi: / Kjellander, P., Gaillard, J.M., and Hewison, A.J.M Density-dependent responses of fawn cohorts body mass in two contrasting roe deer populations. Oecologia, 146(4): doi: /s z. Kokko, H., and Lundberg, P Dispersal, migration, and offspring retention in saturated habitats. Am. Nat. 157(2): doi: / Krebs, C.J How does rodent behaviour impact on population dynamics? In Rats, Mice and People: Rodent Biology and Management. Edited by G.R Singleton, L.A. Hinds, C. J. Krebs, and D.M Spratt. Australian Centre for International Agricultural Research, Canberra. pp Lacey, E.A Sociality reduces individual direct fitness in a communally breeding rodent, the colonial tuco-tuco (Ctenomys sociabilis). Behav. Ecol. Sociobiol. 56(5): doi: /s Leberg, P.L., and Smith, M.H Influence of density on growth of whilte-tailed deer. J. Mammal. 74(3): doi: / Macdonald, D.W., Newman, C., Stewart, P.D., Domingo-Roura, X., and Johnson, P.J Densitydependent regulation of body mass and condition in badgers (Meles meles) from Wytham woods. Ecology, 83(7): doi: /

19 Page 18 of Nevo, E., Beiles, A., Heth, G., and Simson, S Adaptive differentiation of body size in speciating mole rats. Oecologia, 69(3): doi: /bf O Riain, M.J., Jarvis, J.U.M., Alexander, R., Buffenstein, R., and Peters, C Morphological castes in a vertebrate. Proc. Natl. Acad. Sci. U. S. A. 97(24): doi: /pnas Ozgul, A., Childs, D.Z., Oli, M.K., Armitage, K.B., Blumstein, D.T., et al Coupled dynamics of body mass and population growth in response to environmental change. Nature, 466: doi: /nature Parker, W.T., Muller, L.I., Gerhardt, R.R., O Rourke, D.P., and Ramsay, E.C Field use of isolfurane for safe squirrel and woodrat anesthesia. J. Wildl. Manage. 72(5): Available from [accessed 3 February 2018] Patzenhauerová, H., Šklíba, J., Bryja, J., and Šumbera, R Parentage analysis of Ansell s mole-rat family groups indicates a high reproductive skew despite relatively relaxed ecological constraints on dispersal. Mol. Ecol. 22(19): doi: /mec R Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria. [ Robinson, J.G., and Redford, K.H Body size, diet, and population density of neotropical forest mammals. Am. Nat. 128(5): Available from [accessed 3 February 2018] Saether, B.E Environmental stochasticity and population dynamics of large herbivores: a search for mechanisms. Trends Ecol. Evol. 12(4): doi: /s (96) Selman, J., and Goss-Custard, J.D Interference between foraging redshank Tringa tetanus. Anim. Behav. 36(5): doi: /s (88)

20 Page 19 of Šklíba, J., Mazoch, V., Patzenhauerová, H. Hrouzková, E., Lövy, M., et al A maze-lover s dream: burrow architecture, natural history and habitat characteristics of Ansell s mole-rat (Fukomys anselli). Mamm. Biol. 77(6): doi: /j.mambio Solomon, N.G., and Getz, L.L Examination of alternative hypotheses for cooperative breeding in rodents. In Cooperative Breeding in Mammals. Edited by N.G. Solomon and J.A. French. University of Cambridge Press, Cambridge, U.K. pp Spinks, A.C., Bennett, N.C., and Jarvis, J.U.M. 2000a. A comparison of the ecology of two populations of the common mole-rat, Cryptomys hottentotus hottentotus: the effect of aridity on food, foraging and body mass. Oecologia, 125(3): doi: /s Spinks, A.C., Jarvis, J.U.M., and Bennett, N.C. 2000b. Comparative patterns of philopatry and dispersal in two common mole rat populations: implications for the evolution of mole rat sociality. J. Anim. Ecol. 69(2): doi: /j x. Stearns, S.C The Evolution of Life Histories. Oxford University Press, Oxford, U.K. Šumbera, R., Mazoch, V., Patzenhauerová, H., Lövy, M., Šklíba, J., Bryja, J., and Burda, H Burrow architecture, family composition and habitat characteristics of the largest social African molerat, the giant mole-rat constructs really giant burrow systems. Acta Theriol. 57(2): doi: /s van Horne, B Density as a misleading indicator of habitat quality. J. Wildl. Manage. 47(4): doi: / Voigt, C An example of burrow system architecture of dispersing Damaraland mole-rats. Afr. Zoo. 49(1): doi: / Young, A.J., and Bennett, N.C Morphological divergence of breeders and helpers in wild Damaraland mole-rat societies. Evolution, 64(11): doi: /j x.

21 Page 20 of Young, A.J., and Bennett, N.C Intra-sexual selection in cooperative mammals and birds: why are females not bigger and better armed? Philos. Trans. R. Soc. Lond. B Biol. Sci. No. 368(1631): doi:1098/rstb Young, A.J., Carlson, A.A., Monfort, S.L., Russell, A.F., Bennett, N.C., et al Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proc. Natl. Acad. Sci. U. S. A. 103(32): doi: /pnas Young, A.J., Jarvis, J.U.M., Barnaville, J., and Bennett, N.C Workforce effects and the evolution of complex sociality in wild Damaraland mole-rats. Am. Nat. 186(2): doi: / Zöttl, M., Thorley, J., Gaynor, D., Bennett, N.C., and Clutton-Brock, T. 2016a. Variation in growth of Damaraland mole-rats is explained by competition rather than by functional specialization for different tasks. Biol. Lett. (London, U. K.) 12(12): doi: /rsbl Zöttl, M., Vullioud, P., Mendonça, R., Torrents Ticó, M., Gaynor, D., et al. 2016b. Differences in cooperative behavior among Damaraland mole rats are consequences of an age-related polyethism. Proc. Natl. Acad. Sci. U. S. A. 113(37): doi: /pnas

22 Page 21 of Figure 1: Body size differences in subordinate Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)) at a low-density (Low) and a high-density site (High). Subordinate individuals are all nonreproductive individuals >50 g. Individuals were measured and weighed while under anaesthesia to ensure accuracy of measurements. (a) Tooth width was measured across the upper incisors (males p = 0.55, females p = 0.62). (b) Head width was the greatest distance between the ears (males p < 0.01, females p < 0.01). (c) Total length was measured from snout to tail tip (males p = 0.10, females p = 0.04). (d) Body mass was the mean between the body mass at capture and at release (males p = 0.05, females p = 0.03). Figure 2: Body size differences in reproductive Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)) at a low-density (Low) and high-density (High) site. Reproductive females were readily recognized by having prominent nipples and a perforate vagina. The reproductive male was assumed to be the heaviest individual in the group as well as any male with a body mass within 5 g. Individuals were measured and weighed while under anaesthesia to ensure accuracy of measurements. (a) Tooth width was measured across the upper incisors (males p = 0.89, females p < 0.01). (b) Head width was the greatest distance between the ears (males p = 0.41, females p < 0.01). (c) Total length was measured from snout to tail tip (males p < 0.01, females p < 0.01). (d) Body mass was the mean between the body mass at capture and at release (males p < 0.01, females p < 0.01).

23 Page 22 of Figure 3: Mean growth rates for a) male and b) female Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)) at the low-density site (white bars) and high-density site (black bars). Juveniles are individuals 50 g. Subordinates are non-reproductive individuals greater than 50 g. Reproductive females were readily recognized by having prominent nipples and a perforate vagina. The reproductive male was assumed to be the heaviest individual in the group as well as any male with a body mass within 5 g. The growth rate was calculated by dividing the body mass difference between capture events by the time (in months) between captures.

24 Page 23 of 26 Table 1: Body size attributes for juvenile, subordinate, and reproductive Damaraland mole-rats (Fukomys damarensis (Ogilby, 1838)). n df TW* HW* TL* Mass Juveniles Low Density mm mm cm g High Density mm mm cm g P-value Subordinate Males Low Density mm mm cm g High Density mm mm cm g P-value < 0.01 ** Females Low Density mm mm cm g High Density mm mm cm g P-value 0.16 < 0.01 ** 0.04 * 0.03 * Reproductive Males Low Density mm mm cm g High Density mm mm cm g P-value < 0.01 ** < 0.01 ** Females Low Density mm mm cm g High Density mm mm cm g P-value < 0.01 ** < 0.01 ** < 0.01 ** < 0.01 ** * Tooth Width (TW), Head Width (HW), Total Length (TL) sexes were combined p < 0.05, ** p < 0.01

25 a) b) Page 24 of Tooth Width (mm) 6 5 Head Width (mm) c) d) 20 Total Lenght (cm) Body Mass (g) Low High Low High Males Females Low High Low High Males Females

26 a) Page 25 of 26 Canadian Journal b) of Zoology 35 7 Tooth Width (mm) 6 Head Width (mm) c) 21 d) Total Lenght (cm) Body Mass (g) Low High Low High Males Females 80 Low High Low High Males Females

27 a) b) Page 26 of Weight Rate (g * month 1 ) 5 0 Juvenile Subordinate Reproductive Juvenile Subordinate Reproductive