The diet and foraging behaviour of 15 radio-tagged pangolins were studied in the Sabi Sand Wildtuin for 14 months, together with the community composition and occurrence of epigaeic ants and termites. Fifty-five ant and termite species of 25 genera were trapped in pitfalls of which Pheidole sp. 2 was the most common (27% occurrence). Five termite and 15 ant species were preyed on by pangolins. Six of these species constituted 97% of the diet while ants formed 96% of the diet. Anoplolepis custodiens constituted the major component of the pangolins' diet (77% occurrence) while forming only 5% of the trapped ants. Above-ground ant and termite activity was higher during summer than during winter (an 11-fold difference for A. custodiens), and the above-ground activity was also higher during the day than at night. Pangolins fed for 16% of their foraging time. However, 99% of the observed feeding bouts (mean duration 40 s) were on subterranean prey. The mean dig depth was 3.8 cm. Prey from deeper digs were fed upon for longer periods. A model taking into account various ant characteristics suggests that ant abundance and ant size are the two most important factors determining the number of feeding bouts that pangolins undertake on a particular ant species. Temperature effects on ant activity and their nest characteristics may exclude pangolins from parts of southern Africa.

Parental care, especially as it relates to our understanding of sociality, is a topic that has received considerable recent attention in research on spiders (Aviles, 1997). Predatory strategies, including foraging decisions, prey preferences and predatory versatility, have also been emphasized in the recent literature. The present paper is concerned with both of these areas, parental care and predatory strategies. Scytodes sp. indet. is a versatile predator that captures prey either in its web or away from webs, and prey capture may be either by ambush or by active pursuit regardless of whether in the presence and absence of webs. The web is a sparsely woven sheet across a concave upper leaf surface, with a three-dimensional tubular nest linked to the sheet via a small opening. Leaves may be green and living or brown and dead. In the field, a wide range of insects and spiders is taken, but jumping spiders (Salticidae) are the dominant prey. Predatory attacks are by spitting, with large prey being spat at repeatedly before being subdued. After spitting, the scytodid wraps prey in silk before feeding. Eggs are enclosed in a silk egg sac. The female uses her chelicerae to hold her eggs until hatching. After hatching, there is an extended mother–offspring association, with juveniles tending to remain in their mother's web until after the third moult. Females take prey to their juveniles and either feed alongside the juveniles or leave the juveniles to feed alone.

We examined demographics, dispersal, sex-related behaviour, group structure, and genetic similarities of female feral pigs Sus scrofa on the Chaparral Wildlife Management Area in southern Texas from June 1993 to December 1995. Cumulative and seasonal ranges and core areas were calculated for 18 female pigs representing three distinct sounders. Simultaneous pairs of radio locations were used to assess behavioural associations among pigs, and DNA fingerprinting was used to determine genetic similarity. Behavioural and spatial associations largely corresponded to genetic relationships. Similarity of behavioural dendrograms to genetic dendrograms indicated that genetic relationships of feral pigs played a role in observed population structure. A single discrepancy between genetic and behavioural dendrograms suggested two animals dispersed to an adjacent sounder. Also, one sounder appeared to have been created by fission from a larger, adjacent sounder. Factors that are important keys in understanding the association between genetics and behaviour of feral pigs include dispersal, climate, habitat quality, population densities, and sex-related behaviour.

Brants' whistling rat Parotomys brantsii is a gregarious rodent that undergoes marked population fluctuations every few years. Previous studies have variously described them as solitary or social. In this field study, based on direct observations, the social structure and breeding system of a population that almost quadrupled over 7 months was investigated. At the onset of the breeding season, the animals were solitary, each occupying a separate warren. By the end of the breeding season, there was multiple occupancy of some warrens. Each individual maintained a discrete territory within the warren, and exhibited no apparent social behaviour. Outside the breeding season the behaviour of males and females did not differ and both remained at their resident warrens. During the breeding season male P. brantsii visited more neighbouring warrens than did the females, and their visits were to unoccupied warrens or to those occupied by females. Females tended to remain within their own warren. Both sexes avoided warrens occupied by males and few male–male interactions were observed. Females were solicited at their warrens by males, who were frequently chased away. There was, however, evidence that reproductively receptive females were more tolerant of males, who may guard females following mating. This breeding system, based on competitive mate searching by males, and not overt aggression between males, parallels the scramble competition polygyny described for a number of other species, including North American ground squirrels.

Long bones from a taxonomically diverse assemblage of extant terrestrial mammals, spanning more than three orders of magnitude in body mass, have been measured in order to evaluate earlier models proposed for limb allometry as a means of physically coping with increased body size in large species. Linear regression models are unable to explain long bone scaling across a large size range of mammals, as differential scaling is present in large and small species and smaller species tend to approach geometric similarity to a considerably greater extent than larger species. Attempting to explain the morphology of the appendicular skeleton across the large size range of terrestrial Mammalia by means of a standard power function will probably inevitably lead to oversimplification, concealing underlying adaptations for coping with increased size. These include differences in limb posture, decrease in locomotory potential and greater duty factors during fast locomotion in large animals compared to anatomically similar smaller forms, and differential scaling of the appendicular skeleton in small and large mammals.

The distribution of play behaviour among the Australian marsupial families was surveyed. The rate of play in each species observed was categorized on a three-point scale: 1 – no play observed or reported; 2 – play rudimentary; 3 – play common. Common play occurred in the larger-bodied Dasyuridae, in Myrmecobiius, in Vombatidae, and in all Macropodoidea. In all other marsupials, play was rudimentary or absent. Categorical rate of play was not correlated with body mass or with metabolic rate but was correlated with relative brain mass. The distribution of play in Australian marsupials supported the hypothesis that play acts to modify brain development.

Distances at which jumping spiders (Salticidae) use optical cues to distinguish between prey insects and conspecific rivals were investigated using adult males of 37 species. During tests, salticids walked up a ramp toward a mirror or toward an insect enclosed in a transparent petri dish. All species directed threat displays toward their own mirror images and the displays were comparable to each species' typical behaviour during male–male interactions. The salticids never displayed in tests with insects at the top of the ramp. The virtual distances at which the spiders displayed are interpreted as an indication of the distances at which each species can distinguish rivals from prey. Representative species were from the subfamilies Lyssomaninae, Spartaeinae and Salticinae. Discrimination distances relate well to the foveal layer I receptor mosaics of the anterior median eyes for the three subfamilies. Compared with the salticines, the lyssomanines and, except for Portia, the spartaeines tended to have shorter discrimination distances. Portia spp. had discrimination distances comparable to the longest recorded for the salticines. The longest discrimination distances found were for the salticine Mogrus neglectus (max. 320 mm or 42 body lengths) and for the spartaeine Portia fimbriata (280 mm or 47 body lengths).

Foot areas of a wide range of kangaroos, wallabies and rat-kangaroos were measured from fresh material and museum specimens. The areas of the plantar pads scaled according to the predictions of geometric similarity. The terminal claw areas of digits IV and V scaled with a lower than predicted exponent, reflecting functional demands of different species. The pressure that these marsupials exert on the ground during quiet standing is proportional to (body weight)0.35, and is considerably lower than those exerted by ungulates. Hopping kangaroos exert pressures proportional to (body weight)0.37. The pressures exerted by medium and large macropods are comparable to those of galloping ungulates. The findings support the commonly held belief that introduced grazing animals may cause greater mechanical disruption of the soil surface, leading to increased rates of soil erosion, than Australia's indigenous grazing fauna.

Using radio-telemetry, short-term cycles of activity, with a period of about 3 h, were detected in Orkney voles Microtus arvalis orcadensis and in one short-eared owl Asio flammeus, their principal predator. Visual observations showed that owls from four adjacent nests had similar activity patterns. These cycles tended to be synchronous both within and between the two species. Short-eared owls thus appeared to be foraging optimally by timing their hunting to coincide with peaks in vole activity, i.e. at times of maximum potential yield. The extent of diurnal activity in short-eared owls varied seasonally. Daytime activity was conspicuous for only a short period in spring and early summer. Otherwise owls were almost exclusively nocturnal. These seasonal changes in activity were probably a response to variations in vole diurnality, vole population size and daylength. Harassment and kleptoparasitism may have been additional influences that interacted with these seasonal factors to determine the extent of preferred nocturnal hunting. In addition, it is possible that the energetic constraints of breeding may force owls to hunt during daylight. Were males only to hunt at night then they may not be able to provide enough food for their mate and young at a time when nights are short and vole populations are only starting to increase.

In this study we explored quantitatively the relationships between the size of bats, the frequencies in their echolocation calls, and the incidence of moths and beetles in their diets. We focused on the predictions of the allotonic frequency hypothesis which states that some insectivorous bats increase their access to moths that can hear echolocation calls by shifting to frequencies to which the ears of these insects are less sensitive. The hypothesis predicts that the frequencies dominating the echolocation calls of bats may be correlated with the incidence of moths in their diets. We collected data for 62 species of bats that take airborne prey, usually flying insects, 25 species of high duty cycle echolocating bats (Rhinolophidae and Hipposideridae) and 37 species that are low duty cycle echolocators (Vespertilionidae and Molossidae). For bats whose echolocation calls are dominated by frequencies < 100 kHz our regression analyses showed a parabolic dependency between moth consumption (% volume or % frequency) and echolocation call frequency (kHz), supporting the allotonic frequency hypothesis. The use of echolocation calls dominated by frequencies outside the range of best hearing by moths may indeed increase the availability of these insects to the bats. However, when the same analysis was performed with only the bat species using echolocation calls dominated by sounds > 100 kHz, the relationship was not statistically significant, suggesting that morphological characteristics rather than echolocation call frequency may limit the range of potential prey items. Our analyses also demonstrate the importance of jaw morphology as a predictor of the incidence of beetles or moths in the diets of bats, and reveal that generally bigger species (as defined by forearm length) use echolocation calls dominated by lower frequencies than smaller species. In both high duty and low duty cycle echolocating bats the relationship between body size and dominant call frequency was best described by a linear model. We also propose that perch hunting was central in the development of the high duty cycle approach to echolocation.

Small mammal communities (shrews and rodents) were studied using mark-and-recapture techniques from November 1991 to June 1992 at six sites on the Accra Plains of Ghana. The sites were two traditionally protected ‘sacred groves’, two plots in the legally protected Shai Hills Resource Reserve, and two fire-protected sites: secondary forest and a Gmelina tree plantation. A total of 241 individuals representing four species of shrews and nine species of rodents was caught. Abundance, diversity, and composition of small mammals varied greatly between sites. Praomys tullbergi was the most abundant species in the sacred groves, whereas Mastomys erythroleucus dominated in the fire-protected site and in the plantation. The more open grassland areas in the Shai Hills were dominated by Lemniscomys striatus and Uranomys ruddi. Shrews were caught only in forests having dense leaf litter. Legally protected reserve sites had the highest diversity of small mammals but population densities were low. In contrast, the sacred groves had lower diversity levels but much higher population densities, resulting in 1.9 to 14.6 times the biomass of any of the other sites. The rodents Mus minutoides and Hylomyscus alleni and the shrews Crocidura buettikoferi and C. lamottei were found only in the sacred groves. Sacred groves play an important socio-cultural role. Their local ecological and economic importance as reservoirs for rare and useful plants is well known. This study shows that they are also important refuges for certain small mammal species and support characteristic mammal communities. An attempt to measure the impact of the invading exotic tree Azadirachta indica (neem tree) on small mammals was inconclusive, although trapping success was low in areas with dense stands of neem.

Diets of red Cervus elaphus and roe Capreolus capreolus deer in Scottish plantation forests were assessed by botanical and chemical analyses of 144 rumen samples. Samples were taken for summer and winter from four forests with a range of deer densities and ratios. Red deer ate mainly grasses in summer, and a mixture of sedges, rushes, grasses and heaths (ericoid plants) in winter. Roe deer ate predominantly forbs (herbaceous plants other than graminoids) in summer, and a mixture of forbs, heaths and tree and shrub browse in winter. The two species ate significantly different quantities of forbs and grasses (P < 0.001). Both species ate more grasses and forbs in summer, and more heaths in winter (P < 0.05). There were few differences between forests for either red or roe deer in diet composition or its seasonal variation. Washed rumen contents of roe deer contained more nitrogen than did those of red deer (P < 0.05) in both seasons, and both species contained more nitrogen in summer than in winter (P < 0.05). Proportions of the main plant categories within rumens were compared with their abundance recorded in the ground flora. These comparisons showed roe deer to be highly selective, whilst red deer ate plants in proportions similar to their availability. Niche overlap increased and dietary niche breadths widened during the winter. The main overlap was for heaths Calluna vulgaris and Vaccinium myrtillus in winter. However, C. vulgaris is widespread and common and competition was considered to be more likely for forbs and possibly V. myrtillus: the use of some forb species by red deer in summer may limit their availability to roe deer in winter, when they remain an important dietary item. Consideration of the different digestive specializations of the two species suggested that any dietary competition is likely to affect roe deer more than red deer.