Evolutionary association between subterranean lifestyle and female sociality in rodents

Subterranean rodents are a good model system to examine adaptive evolution of social organization. Life underground has been proposed either to favor solitariness or, to the contrary, to promote sociality. In concordance with the first idea, most specialized diggers are solitary. However, group-living in several unrelated subterranean rodent species and especially eusociality in two genera of African mole-rats, the Bathyergidae, seem to support the second hypothesis. Thus, none of the two models is fully consistent with empirical data. Here we apply the comparative phylogenetic method to test an evolutionary correlation between fossoriality and female social strategy (solitary breeding vs breeding in group). Both characters show very strong phylogenetic signal, and we found a significant correlation between them. Subterranean lifestyle is readily acquired under female sociality. By contrast, the transition to life underground is extremely unlikely under female solitariness. Thus, not only social behavior may be affected by ecological specialization as it is widely assumed, but it can itself restrain the range of possible specializations. The rates of transition from sociality to solitariness are equal under subterranean and surface-dwelling lifestyle. Sociality loss is irreversible in subterranean lineages, unlike surface-dwelling lineages. Based on the revealed transition rates we suggest that all lineages of subterranean rodents have gone through the stage of cooperation at the beginning of their evolutionary track, whereas group-living is selected against in highly specialized diggers. An odd pattern of distribution of sociality across and within truly subterranean taxa probably derives from the influence of extrinsic factors in combination with phylogenetic inertia.     

Growth and growth substrate levels in spinach under non-steady state conditions of nitrogen nutrition and light

Spinach plants ( Spinacia oleracea L. cv. Subito) were grown in a complete nutrient solution under ample light intensity (14 h day⁻¹ at 660 μmol m⁻² s⁻¹) before being transferred either to a minus-N solution (experiment 1), or to limiting light conditions (6 h day⁻¹ at 220 μmol m⁻² s⁻¹; experiment 2). Shoot growth in experiment 1 decreased significantly from 0.24 day⁻¹ to 0.07 day⁻¹ after the fourth day of transfer. Root relative growth rate increased after 1 day from 0.25 to 0.31 day⁻¹, but decreased on the fifth day after transfer to 0.11 day⁻¹. Shoot growth in experiment 2 decreased significantly from 0.25 to 0.17 day⁻¹ after the fourth day of transfer, while root growth decreased to half of its original level (0.25 day⁻¹) already on the second day. Growth substrate levels in the plants (free sugars, free amino acids) and starch levels depended on the plant age, the moment in the diurnal cycle, and the imposed treatment. Fluctuations in shoot growth or root growth resulting from the light or N limitation could not be explained by a correspondent increase or decrease in the levels of growth substrates. The hypotheses underlying the functional equilibrium theory, assuming shoot and root growth to be controlled by N- and C-containing substrates respectively, and several other growth and partitioning models are therefore questioned. A neglect of the osmotic role of the free sugars in these models might be the explanation for this.     

Natural history and burrow system architecture of the silvery mole-rat from Brachystegia woodland

The silvery mole-rat Heliophobius argenteocinereus (Bathyergidae) is a solitary subterranean rodent, widely distributed throughout eastern and south-eastern Africa in a variety of habitats. Here, we provide the first data on its biology in a typical natural habitat, the Brachystegia woodland. The population density of mole-rats was low (4.6 ha⁻¹) and its distribution across the study site was random. Contrary to subadults and pups, the sex ratio of adult mole-rats was highly female biased (1:5.75), probably due to the higher mortality of males as a consequence of their mating strategy. Reproduction of Heliophobius is seasonal and pups are born at the beginning of the hot dry season. Burrow systems of the silvery mole-rat were long, highly branched, reticulated and comparable in fractal dimension to systems of social bathyergids. Variability in burrow architecture was related to the body mass of the burrow occupants, soil hardness and food supply when tested together. Burrow systems with a higher fractal dimension had inhabitants that had a greater body mass. Longer systems were less branched. Nests were typically deeper than foraging tunnels and experienced negligible temperature fluctuations. The microenvironmental characteristics of the subterranean niche including temperature, humidity and soil characteristics are provided for purpose of comparison with other mole-rat species.     

Consumption, growth and evacuation in the Pacific cod, Gadus macrocephalus

Growth of Pacific cod was related to energy consumption (cal g⁻¹ day⁻¹) and was well described by linear equations. Maintenance ration was 11 and 12 cal g⁻¹ day⁻¹ at 4.5 and 6.5° C, respectively. Cod between 200 and 5000 g had similar growth rates when growth was expressed as a function of consumption (cal g⁻¹ day⁻¹). Laboratory consumption of food averaged 0.9 and 1.3% body weight per day at 4.5 and 6.5° C, respectively. At these temperatures growth was 0.34–0.38% body weight day⁻¹.
Maximum stomach volumes equated to approximately 4.7% of body weight with shrimp as prey. At this meal size Pacific cod did not feed the next day. A multiple meal evacuation experiment was used to verify the consumption estimates. A return-to-hunger estimate of the meal size evacuated was 1.5% body weight day⁻¹ at 6.5° C, similar to the 1.3% consumption estimate. For Pacific cod fed a single meal of 1% body weight the estimated instantaneous evacuation rate was 0.63 body weight day⁻¹ at 6.5° C. Meal size markedly affected the evacuation rate.
Measured consumption and growth rates are similar to those of Atlantic cod, Gadus morhua .     

The reproductive biology of the Cape mole-rat, Georychus capensis (Rodentia, Bathyergidae)

Georychus capensis is a seasonal breeder with the reproductive potential of producing two litters during the summer breeding season (August to December). These solitary mole-rats signal to each other through the soil, by drumming with their hind feet. This drumming is probably important in spacing the burrow systems and in triggering the onset of reproductive behaviour. In a captive male the onset of drumming was accompanied by a rise in urinary testosterone concentrations and the enlargement of testes and accessory reproductive glands. The male appeared to drum with a different frequency to the female. Courtship is initiated by the male and copulation involves brief multiple intromissions. The gestation period is about 44 days and the mean litter size is 5.9 with a maximum of 10 pups. Development of the pups is relatively rapid. Inter-sibling aggression begins to develop at 35 days, eventually resulting in the pups dispersing when about 60 days old. Body mass increases exponentially from birth to day 60 and the asymptote is reached around day 260. These features are compared with those of other solitary subterranean rodents and with those of the social Bathyergidae ( Cryptomys hottentotus, Cryptomys damarensis and Heterocephalus glaber ).     

Winter growth of juvenile plaice on the Port Erin Bay (Isle of Man) nursery ground

Juvenile plaice ( Pleuronectes platessa ) were studied in Port Erin Bay, Isle of Man, U.K. between September and March 1989/1990 and 1990/1991. Plaice (>90 mm) were tagged and individual growth rates calculated for the autumn, winter and over-winter time periods. During the study the population of fish >90 mm remained fairly stable both within a study season and between years. Autumn growth rates ranged from 0 to 0.39 mm day⁻¹ (mean₁₉₈₉=0.13 mm day⁻¹, mean₁₉₉₀=0.10 mm day ⁻¹) and winter 0 to 0.5 mm day⁻¹ (mean₁₉₉₀=0.11 mm day⁻¹ and mean₁₉₉₁=0.17 mm day ⁻¹). In general, growth rates were higher at higher mean seasonal temperatures. However, relative growth rate was significantly higher in the colder winter period of 1991 than 1990.     

Energy and ration requirements of juvenile Pacific halibut (Hippoglossus stenolepis) based on energy consumption and growth rates

Growth of captive juvenile Pacific halibut was linearly related to energy consumption (J g⁻¹ day⁻¹) at 4°C by the following equation: growth (% body weight (b.w.) day⁻¹)=0–007 (consumption J g⁻¹ day⁻¹)– 0.192; r² =0.81. Weight gain was independent of size for fish between 9 and 7000 g when growth was expressed as a function of consumption in J g⁻¹ day⁻¹. Maintenance ration determined in feeding–growth experiments averaged 27.4 J g⁻¹ day⁻¹ at 4–0°C. Small halibut ate significantly more food than large fish. Single meals following 2 day fasts averaged 4.1% b.w. for halibut under 100 g, 1.72% b.w. for 1.2 kg fish and 1.1% B.W. for 6.8 kg fish. Both large and small size categories of halibut tended to evacuate their meal in about 3 days even though small fish ate relatively larger meals. Minimum estimates for daily ration to achieve growth rates observed in the Gulf of Alaska were approximately 0.5 to 2.4% b.w. day⁻¹ depending on fish size and whether northern shrimp or yellowfin sole were their prey.     

Food intake and growth in the blue-spotted trevally, Caranx bucculentus Alleyne and Macleay 1877, with reference to predation on penaeid prawns

Blue-spotted trevally, Caranx bucculentus , were fed different rations of pilchard and prawn in order to investigate feeding and growth relationships. Maintenance rations at 25.5° C amounted to 3.7% B.W. day⁻¹ and 2.7% B.W. day⁻¹ for prawns and pilchards, respectively. Additional feeding experiments at 28.9° C yielded a maintenance ration of prawns of 3.8% B.W. day⁻¹, suggesting there is very little if any temperature effect on the feeding-growth relationship over the range studied. Fish fed twice or more each day consumed about 7.3 ± 1.4% B.W. day⁻¹.
Given the biomass of this trevally in Albatross Bay, Gulf of Carpentaria, and the contribution of prawns to its diet, we estimate consumption of commercial prawns at 25 ± 5 g.ha⁻¹ day⁻¹ or 11 g kg⁻¹ day⁻¹.     

Life cycle, growth, mortality and production of Ephemerella major Klapalek (Ephemeroptera) in a trout stream in Belgium

SUMMARY. 1. The life cycle of Ephemerella major Klapalek in a chalk trout stream in Belgium took 1 year. Emergence was highly synchronized with a flight period from mid-May to mid-June. Tiny nymphs occurred from June to late August.
2. The mean instantaneous growth rate was high in autumn (3.6% wet wt day⁻¹), very low from November to February (0.8% wet wt day⁻¹) and high until emergence (2.3% wet wt day⁻¹); short day length seemed to be the major factor reducing growth rate during winter.
3. Mortality was close to zero during winter and 1.4—1.7% day⁻¹during other seasons. Total mortality from egg to adult was 99.6%.
4. The annual production was about 9g wet wt m⁻² year⁻¹ and the annual P/ ratio was 7.5. There was good agreement between the production values estimated by four methods. Production rate was highest in May (13 mg wet wt m⁻² day⁻¹) and zero in February.     

PHOTIC INDUCTION OF Fos IN THE SUPRACHIASMATIC NUCLEUS OF AFRICAN MOLE-RATS: RESPONSES TO INCREASING IRRADIANCE

African mole-rats (family Bathyergidae) are strictly subterranean rodent species that are rarely exposed to environmental light. Morphological and physiological adaptations to the underground environment include a severely reduced eye size and regressed visual system. Responses of the circadian system to light, however, appear to be intact, since mole-rats are able to entrain their circadian activity rhythms to the light-dark cycle and light induces Fos expression in the suprachiasmatic nucleus (SCN). Social organization varies from solitary species to highly elaborated eusocial structures, characterized by a distinct division of labor and in which one reproductive female regulates the behavior and reproductive physiology of other individuals in the colony. The authors studied light-induced Fos expression in the SCN to increasing light intensities in four mole-rat species, ranging from strictly solitary to highly social. In the solitary Cape mole-rat, light induces significant Fos expression in the SCN, and the number of Fos-immunopositive cells increases with increasing light intensity. In contrast, Fos induction in the SCN of social species was slightly greater than, but not statistically different from, the dark-control animals as is typical of most rodents. One species showed a trend for an increase in expression with increased light, whereas a second species showed no trend in expression. In the naked mole-rat, Fos expression appeared higher in the dark-controls than in the animals exposed to light, although the differences in Fos expression were not significant. These results suggest a gradient in the sensitivity of the circadian system to light in mole-rats, with a higher percentage of individuals that are unresponsive to light in correlation with the degree of sociality. In highly social species, such as the naked mole-rat that live in a relatively stable subterranean milieu in terms of food availability, temperature, constant darkness, and devoid of 24-h cyclic environmental cues, the temporal coordination of rest-wake activities may be dependent on social interactions and social status rather than on photic regulation of the circadian timing system. (Author correspondence: moosthuizen@zoology.up.ac.za)     

Comparison of growth rate and formation of otolith increments in age-0 red snapper

For wild red snapper Lutjanus campechanus , mean otolith increment deposition rate after marking with oxytetracycline dihydrate (OTC) was daily (0.97 increments day⁻¹) when growth rates were fast (0.63 mm fork length, L _F day⁻¹), but were not daily (0.82 increments day⁻¹) when somatic growth was slow (0.2 mm L _F day⁻¹). For reared larvae ( n =8), increment deposition rates were daily (0.99–1.03 increments day⁻¹), and growth rates ranged from 0.6 to 0.9 mm L _F day⁻¹. Growth rate affected increment deposition rate as a threshold function, i.e. when growth rate was <0.3 mm L _F day⁻¹, deposition was less than daily, but above this level increment deposition did not exceed a daily rate. As growth rates increased increment widths increased. Examination of a sub-sample ( n =8) of the otoliths from the slowest growing wild fish by scanning electron microscopy did not increase increment counts. Because L. campechanus are late spring-early summer spawners, young fish can expect maximum growth due to warm summer temperatures. Thus, daily ageing methods should be well suited to this species.     

Effect of ration size on growth and gross conversion efficiency of young lemon sharks, Negaprion brevirostris

Young lemon sharks, Negaprion brevirostris , were kept under controlled conditions in an aquarium and fed blue runner, Caranx crysos , at different ration levels. The relationship between feeding rate and growth rate was best described by a von Bertalanffy growth curve, which predicted a maximum growth rate of 140 kJ kg⁻¹ day⁻¹ (0·66% b.w. day⁻¹), a maintenance ration of 199 kJ kg⁻¹ day⁻¹ (1·06% b.w. day⁻¹), and losses due to starvation of -236kJ kg⁻¹ day⁻¹ (1·11% b.w. day⁻¹). The relationship between gross conversion efficiency ( K ₁) and feeding rate was also examined. K₁ ranged from - 64 to 25% and did not drop at high ration levels. Activity levels of both starved sharks and sharks fed at maintenance were not significantly different (0·2 body lengths s⁻¹). K ₁ values generated from both laboratory and field data suggest that young lemon sharks can convert food to new tissue as efficiently as teleosts.     

Growth response of pikeperch larvae in relation to body size and zooplankton abundance

The most critical period at onset of feeding in pikeperch Stizostedion lucioperca is short (<5 days at 20° C). The larvae are sensitive to prey density during the first week of exogenous feeding. First-feeding larvae of 6.5 mm total length ( L_T ) needed prey densities of >585 prey l⁻¹ to maintain mass (C_maint), whereas 5 days older larvae of 7 mm L _T C_maint= 55 prey l⁻¹ and for 11 mm L _T larvae C_maint<10 prey 1⁻¹. Changes in specific growth rates for larvae of 7 and 11 mm were similar to a type-II functional response curve reaching a specific growth rate of 26 and 30% day⁻¹, respectively, at 1000 prey l⁻¹, whereas the 6·5 mm larvae showed a linear growth response reaching a specific growth rate of only 9% day ⁻¹ at 1000 prey l⁻¹. The results suggest that prey density is a limiting factor, which might contribute to the high variation between year-class Strengths.     

Determining patterns of use by black bream Acanthopagrus butcheri (Munro, 1949) of re-established habitat in a south-eastern Australian estuary

Acoustic telemetry was used to assess patterns of utilization by black bream Acanthopagrus butcheri of regions with and without re-introduced large woody debris (LWD) in two estuaries of south-eastern Australia, the Mitchell and Tambo Rivers. The fish ( n = 46) were implanted with acoustic transmitters in December 2004 and March 2005 and monitored for c. 12 months. The two principal metrics from the telemetry data, number of visits per day ( N day⁻¹) and residency (amount of time, s) were highly correlated ( r > 0·948), and subsequent analyses were based on N day⁻¹. Rates of N day⁻¹ varied inconsistently among estuarine regions across diel periods and times of year for each river. In the Tambo River during autumn, winter and spring, the N day⁻¹ was greatest in the middle and upper estuarine regions during the day, and often greater in the lower region at dawn and dusk, but varied little among regions in summer. The provision of LWD had little effect on N day⁻¹ in the Tambo River. In the Mitchell River, N day⁻¹ varied little among estuarine regions without LWD, regardless of the time of day, and these patterns were consistent across seasons, but N day⁻¹ was significantly greater to the LWD during the day in winter and spring. Freshwater flows had little effect on monthly patterns of N day⁻¹ among regions in either estuary. The perceived 'benefits' to A. butcheri of re-establishing LWD within estuarine systems of south-eastern Australia depended strongly on the time of year, time of day and river system, but acoustic telemetry was a useful method of evaluating the use by fish of these artificial structures.     

Seasonal variability in growth of larval Japanese anchovy Engraulis japonicus driven by fluctuations in sea temperature in the Kii Channel, Japan

Seasonal variability in the growth of larval Japanese anchovy Engraulis japonicus was examined through otolith microstructure analysis based on the samples collected from the northern side (inner area, IA) and the southern side (outer area, OA) of the Kii Channel from April 2006 to March 2007. Growth trajectories (otolith backcalculated mean standard length of 5 day intervals from 5 days after hatch to 24 days) as well as the most recent 5 day mean growth rate of larvae before capture ( G ₅) differed among months. Growth trajectories showed the same pattern as G ₅. In IA, mean ± s.d. G ₅ ranged from 0·31 ± 0·04 mm day⁻¹ (January) to 0·73 ± 0·06 mm day⁻¹ (October). In OA, mean ± s.d. G ₅ ranged from 0·36 ± 0·05 mm day⁻¹ (January) to 0·79 ± 0·11 mm day⁻¹ (August). G ₅ values declined from November to January and then started to increase. In general, the seasonal patterns of growth were similar between IA and OA, and a clear seasonal pattern in growth was identified. When the relationships among larval growth rate, sea temperature, zooplankton density and larval density were examined, growth rate was positively related with sea temperature in both areas and not related with the other factors. The similar pattern in growth observed between IA and OA was probably due to the low spatial variability in sea temperature compared to its seasonal variability.     

Influence of UV-B radiation on developmental changes, ethylene, CO2 flux and polyamines in cv. Doyenne d'Hiver pear shoots grown in vitro

In vitro shoots of cv. Doyenne ďHiver pear ( Pyrus communis L.) were irradiated under controlled environments for 6 h per day at 5 different levels of biologically effective UV-B radiation (UV-B_BE). UV-B exposure caused a progressive increase in apical necrosis above background levels and stimulated leaf abscission. Shoots grown for 2 weeks at 7. 8 mol m⁻² day ⁻¹ of photosynthetic photon flux (PPF) and treated with 8. 4 or 12. 0 kJ m⁻² day ⁻¹ UV-B_BE produced up to 4 times more ethylene than those given 2. 2 or 5. 1 kJ m⁻² day⁻¹ UV-B_BE or untreated controls. Exposure of shoots to 12 kJ m⁻² day ⁻¹ of UV-B_BE caused an increase in free putreseine content after 4 to 14 days of irradiation. Shoots showed a decrease in CO₂ uptake after 3 days of UV-B: thereafter, they appeared to recover their photosynthetic capacity. Under typical PPF conditions used in micropropagation (90 μmol m⁻² S⁻¹). 8. 4 kJ m⁻² day ⁻¹ of UV-B radiation was injurious to realatively tender tissues of in vitro pear shoots: increasing the level of UV-B_BE to 12 kJ m⁻² day⁻¹ produced even more adverse effects.     

Daily ration estimates for yellowfin sole, Limanda aspera (Pallas), based on laboratory consumption and growth

Several estimates of minimal energy requirements for yellowfin sole were made. Energy expenditures of 1.6, 4.1 and 8.3 cal g⁻¹ day⁻¹ were obtained from starvation weight loss, standard metabolism and maintenance ration procedures, respectively, at 6° C. The temperature effect on energy requirement was reflected in the Q ₁₀ values for starvation weight loss (2.0), standard metabolism (6.3) and maintenance ration (6.5).
Both energy intake and weight of food were linearly related to, and good predictors of, laboratory growth. These relationships were used to estimate the food and energy intake necessary for yellowfin sole to achieve a year's growth in the natural environment. Based on a caloric value of 2.0 kcal g⁻¹ of food (herring fillets), yellowfin would require 0.35 to 0.39% body weight day⁻¹ at 3° C to achieve the mean growth rate exhibited in the Bering Sea. To achieve Gulf of Alaska growth rates at 5 to 6° C, yellowfin would require 0.63% body weight day⁻¹. Based on a caloric value of 0.57 kcal g⁻¹ of food (chopped octopus), yellowfin would require 0.83% body weight per day to achieve the Gulf of Alaska growth rate (6° C). These requirements based on the calorific value of herring fillets, which are three to five times higher than previous estimates of daily ration in this species, are probably conservative estimates since many of their prey species have a lower energy content.     

Estimation of the growth rates of epiphytic bacteria and Lemna minor in a river

SUMMARY. Growth of Lemna minor fronds in the River Frome during summer was found to be logarithmic with time and the growth rate (log₁₀) was 0.066 day⁻¹. This is equivalent to a doubling time of 4.5 days. The life expectancy of the fronds was 34 days.
The net change in the density of bacteria epiphytic on the lower surface of Lemna fronds in the R. Frome was monitored using a direct microscopic technique. The observed increase in bacterial numbers has been partitioned into the components of attachment and growth, assuming that attachment occurred at a constant rate and that the bacterial population grew logarithmically. The line which fitted the data best gave an attachment rate of 5.7 × 10⁵ bacteria cm⁻² day⁻¹ and a growth rate (log₁₀) for the bacteria of 0.044 day⁻¹ which is equivalent to a doubling time of 164 h.
Estimates of the rate of detachment of bacteria from Lemna plants were obtained from a laboratory experiment which assumed negligible growth of bacteria in 1 h. The number of bacteria which detached per hour and the sizes of the bacterial populations on the plants before and after detachment were estimated using a plating technique. Different detachment rates were monitored. The detachment rate (analogous to growth rate) which is judged to be most similar to an in situ value was 0.0031 h⁻¹ (0.074 day⁻¹). This rate added to the specific growth rate given above resulted in a corrected growth rate of 0.118 day⁻¹ equivalent to a doubling time of 61 h.     

Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coastal environment (NW Mediterranean)

The impact of viruses and protists on bacterioplankton mortality was examined monthly during 2 years (May 2005–April 2007) in an oligotrophic coastal environment (NW Mediterranean Sea). We expected that in such type of system, (i) bacterial losses would be caused mainly by protists, and (ii) lysogeny would be an important type of virus–host interaction. During the study period, viruses and grazers together were responsible for 50.6 ± 40.1% day⁻¹ of bacterial standing stock losses (BSS) and 59.7 ± 44.0% day⁻¹ of bacterial production losses (BP). Over the first year (May 2005–April 2006), protists were the principal cause of bacterial mortality, removing 29.9 ± 20.4% day⁻¹ of BSS and 33.9 ± 24.3% day⁻¹ of BP, whereas viral lysis removed 13.5 ± 17.0% day⁻¹ of BSS and 12.3 ± 12.3% day⁻¹ of BP. During the second year (May 2006–April 2007), viruses caused comparable bacterial losses (29.2 ± 14.8% day⁻¹ of BSS and 40.9 ± 20.7% day⁻¹ of BP) to protists (28.6 ± 25.5% day⁻¹ of BSS and 32.4 ± 20.0% day⁻¹ of BP). In 37% of cases higher losses of BP due to viruses than due to protists were found. Lysogenic infection was detected in 11 of 24 samplings. Contrary to our expectations, lytic infections dominated over the two years, and viruses resulted to be a significant source of bacterial mortality in this oligotrophic site.     

Ecological constraints drive social evolution in the African mole-rats.

The African mole-rats (family Bathyergidae) are subterranean hystricomorph rodents occurring in a variety of habitats and displaying levels of sociality which range from solitary to eusocial, making them a unique mammalian taxonomic group to test ecological influences on sociality. Here, we use an extensive DNA-based phylogeny and comparative analysis to investigate the relationship between ecology, sociality and evolution within the family. Mitochondrial cytochrome-b and 12s rRNA trees reveal that the solitary species are monophyletic when compared to the social species. The naked mole-rat (Heterocephalus glaber) is ancestral and divergent from the Damaraland mole-rat (Cryptomys damarensis), supporting previous findings that have suggested the multiple evolution of eusociality within the family. The Cryptomys genus is species-rich and contains taxa exhibiting different levels of sociality, which can be divided into two distinct clades. A total of seven independent comparisons were generated within the phylogeny, and three ecological variables were significantly correlated with social group size: geophyte density (p < 0.05), mean months per year of rainfall greater than 25 mm (p < 0.001), and the coefficient of rainfall variation (p = 0.001). These results support the food-aridity hypothesis for the evolution of highly social cooperative behaviour in the Bathyergidae, and are consistent with the current theoretical framework for skew theory.