The hypothesis that the periods of dormancy previously described in the millipede Polydesmus angustus may be photoperiodically induced diapauses was tested experimentally. In this species, biennial individuals exhibit two successive periods of dormancy: aestivation in the penultimate stadium (stadium VII) and reproductive dormancy in the adults, which emerge in autumn. It was first established that the reproductive dormancy is not a thermally controlled state of quiescence. When adults emerging in autumn were kept at 16 °C under natural photoperiod, their reproduction was delayed for several months in comparison with adults emerging in spring at similar temperatures. This indicates that the reproductive dormancy begins with a period of diapause. Further experiments provided evidence of a photoperiodic induction of the adult diapause. When millipedes were reared under short day length (L:D 12:12 h) throughout their development, they required more time to reproduce than millipedes reared under long day length (L:D 16:8 h) at the same temperatures. Photoperiod influenced reproduction in females, but no significant effects were detected in adult males. On the other hand, stadium VII was markedly longer at L:D 16:8 h than at L:D 12:12 h in both sexes, which strongly suggests that aestivation is also induced by photoperiod. However, the effects on the duration of stadium VII varied among individuals, some of which showed no response to long days. This study is the first to document photoperiodic regulation of the life cycle in the class Diplopoda, a trait common in other classes of terrestrial arthropods.

Annual home-range size indices for 36 male and 52 female adult brown bears Ursus arctos in two study areas in central and northern Scandinavia were estimated to evaluate factors believed to influence home-range size. Male home ranges were larger than home ranges of lone females after controlling for the sexual size dimorphism acting on metabolic needs. Further, home ranges of females with cubs were smaller than home ranges of lone females and females with yearlings. Thus, differences in metabolic need were not able to explain the variation in range size among females of different reproductive categories or between males and females, suggesting roaming behaviour of males in this promiscuous species. Home-range size in both males and females was inversely related to population density along a density gradient that was not linked to food availability. This contradicts the hypothesis that females use the minimum areas that sustain their energy requirements. However, on a large geographical scale a negative relationship between range size and food availability was evident. The annual home ranges in inland boreal environments in Scandinavia are the largest reported for brown bears in Eurasia, and similar to those in inland boreal and montane environments in North America.

The foraging behaviour of captive colonies of the highveld mole-rat Cryptomys hottentotus pretoriae was investigated in an artificial soil-filled burrow system provided with four trays (patches) that varied in geophyte density and mass. An initial trial involving empty trays (only soil) revealed that there was no preference for any specific tray. There were no statistically significant preferences for excavating in any of the patches of different geophyte density. No preferences were evident for excavation in patches containing geophytes of different mass classes. Empty patches seemed to be preferred over patches containing geophytes when combinations of geophyte density/mass were investigated. The duration of handling and the rate of consumption of geophytes were recorded for 23 individuals of two mass classes. Handling time of geophytes was not related to mole-rat sex, but was strongly linked to mole-rat mass class. Handling time of geophytes was related to geophyte mass class. Small geophytes were less profitable to consume. These findings are considered in light of optimal foraging theory and the situation in the field. It was concluded that the mole-rats generally followed the qualitative predictions of optimal foraging theory, although falling short of being energy maximizers.

Mountain hares Lepus timidus L. typify species that occupy a broad geographic range and have flexible foraging and nutritional strategies. Such species may show a range of responses to habitat modification. This study aimed to provide a basis for prediction of the impact of mountain hares on woodland establishment, and of woodland establishment on mountain hare distribution. The selection of and the extent of incorporation of new woodland into the home range of mountain hares was investigated in an area where Scots pine Pinus sylvestris L. woodland was establishing within their usual habitat in Britain, upland heather moorland. Seasonal home, day and night-range sizes of radio-tracked mountain hares were determined using the multinuclear probability polygon technique and analysed using residual maximum likelihood (REML). Habitat selection was analysed using compositional analysis. Three main habitat types were available to hares: heather moorland with trees, heather moorland and grassland-mire. Mean home-range size of mountain hares in summer was 10.3 ha and in winter 9.6 ha. There were no significant seasonal or sex differences in home-range size. Females selected grassland-mire habitat in summer and showed no strong selection for any habitat in winter. Males selected heather moorland in both summer and winter. Heather moorland with trees was not selected preferentially by mountain hares of either sex in summer or winter. The absence of selection for areas of newly establishing-Scots pine woodland suggests that any browsing damage to trees by hares is most likely to be a function of the local abundance of mountain hares, rather than a result of active preference of hares for the modified habitat.

Habitat preference is driven by a complex interaction among behavioural patterns, biological requirements, and environmental conditions. These variables are difficult to determine for any species but are further complicated for migratory marine mammals, such as humpback whales Megaptera novaeangliae. Patterns of habitat use in relation to social organization potentially exist for this species on their wintering grounds. Using an integrated GIS approach, we examined the degree to which spatial patterns of habitat stratification are correlated within different humpback whale group types from 6 years of sighting data (1996–2001) collected on the Antongil Bay, Madagascar, wintering ground. Stratification of humpback whale sightings by behavioural classification showed significant variation in depth and distance from shore. Distribution by depth could not be described as a function of group size but could be described as a function of social organization, with mother–calf pairs showing a strong preference for shallower water compared to all other group types. Group size and social organization seem to be factors in distribution by distance from shore. Significant diurnal patterns in distribution by depth and distance from shore also exist, where mother–calf groups maintain a relatively stable distribution and pairs and competitive groups are the most variable. Patterns of habitat preference on this wintering ground appear to be guided by social organization, where distribution by depth and distance from shore highlight areas critical to conservation.

From naked mole rats to elephants, research in wildlife biology increasingly entails interference with individual animals, and there are innumerable reasons to justify this interference for the proper management of populations. As Boyd (2002) has emphasized, however, there is a central question about the extent to which it is reasonable to intervene in a population when the species is endangered. This comment arose as part of a debate published last year in this journal on the possible harmful effects of immobilizing black rhinos and fitting them with radio-collars. Alibhai, Jewell & Towindo (2001) had prompted the debate with a discussion on the possible reduction of fertility in black rhinos caused by immobilization followed by an account of the wounding of a rhino by an ill-fitting radio-collar (Alibhai & Jewell, 2001).

The swarm founders are unusual among the social wasps in having socialized the dispersal stage of the life cycle. Colonies are initiated by groups of workers accompanied by smaller numbers of queens. Thus, swarm founders avoid the colony size bottleneck faced by the independent founders, whose colonies start with one or a few queens. Among the advantages of swarm founding is a reduction of the risk of colony failure due to attrition of the founding adults. Stochasticity in adult mortality is less likely to lead to outright extinction of a large founding group before new workers are produced (pre-emergence period). However, it is not known how important pre-emergence mortality is as a selective force on founding and dispersal behaviour in swarm founders, since colony-wide mortality rates have never been reported for a large-colony social wasp. Sixty-eight swarms of Polybia occidentalis were censused just before colony initiation and again 24 days later. Mortality over this period averaged 0.41±0.12 of the founding swarm population. Including mortality on the day of emigration and extrapolating to day 30, when the first adult offspring eclose, the original absconding swarm would be reduced by 0.52 of its initial size. Rates of loss during the first week, while the colonies engaged in nest construction, did not differ from rates over the full 24 days. Thus, colony founding in P. occidentalis is both costly and highly variable in terms of mortality of the founding adults.

Spiders from the theridiid genus Argyrodes exhibit considerable variation in foraging tactics. However, little is known about the conditions under which Argyrodes spiders switch foraging tactics. Argyrodes flavescens (Pickard-Cambridge) is commonly found in the webs of another spider Nephila pilipes (Fabricius) in Singapore. In this study, a series of prey-choice tests were conducted for A. flavescens, both in the presence and absence of N. pilipes, to investigate the state-dependent prey type preference of A. flavescens. It was found that, in the absence of N. pilipes, well-fed A. flavescens took houseflies more than fruit flies, but starved A. flavescens took more fruit flies than houseflies. Whether N. pilipes spiders were present or absent, both well-fed and starved A. flavescens preferred living prey and rarely took wrapped prey of any kind. When well fed, A. flavescens rarely took mealworms. However, when starved, A. flavescens tended to take freshly captured prey, and also tended to feed together with N. pilipes on a housefly or mealworm captured by N. pilipes. Whether A. flavescens were absent or present, both well-fed and starved N. pilipes took mealworm larvae more often than they took houseflies, and they never attacked fruit flies. This is the first study to show that Argyrodes spiders alter their foraging tactics depending on hunger level, prey type, or the presence of the host. In doing so, Argyrodes spiders may maximize their energy gain and minimize predation risk in different circumstances.

The feeding spectrum of the wildcat Felis silvestris Schreber, 1777 was studied in two sites with different ecological characteristics, both situated in the same Mediterranean environment in the high mountain of the Sierra Nevada National Park, south-east Spain, where the rabbit Oryctolagus cuniculus is absent. Scat analysis (n=101 faeces; n=402 prey items) showed that the diet is based on rodents, fundamentally wood mouse Apodemus sylvaticus, Mediterranean pine vole Microtus duodecimcostatus and south-western water vole Arvicola sapidus. Results showed strong differences between the two sites (χ2=74.04, d.f.=5, P<0.001), that is a predominance of voles in the mesic Chico river, whereas mice are predominant in the xeric Tejos ravine. Red-legged partridge Alectoris rufa and carrion also played an important role, especially in biomass terms. The overall diet differed essentially from that of the Mediterranean region, which surrounds the study area, since in these areas rabbits constitute the primary prey. However, the diet of the mountain wildcats is similar to that in the Eurosiberian floral region, despite its distance from the Sierra Nevada. In conclusion, the Iberian wildcat seems to behave as a facultative specialist, since it prefers rabbits whenever they are available, but rodents constitute most of its diet if rabbits are scarce or absent.

Midwife toads present one of the simplest calls in anurans, with the whole energy concentrated in a single band without frequency modulation. The tuning curves of the Iberian midwife toads Alytes cisternasii show the typical bimodal pattern in anurans, with two best excitatory frequencies at 0.412 kHz (corresponding to the amphibian papilla) and at 1.358 kHz (corresponding to the basilar papilla and matching the male call frequency). In this study, the hypothesis that complex calls arose in anurans because they were inherently more attractive to females, since they provided greater acoustic stimulation, was tested. However, our results indicate that splitting the call energy to stimulate both inner ear organs simultaneously, the male call is not more attractive to female midwife toads, but sometimes renders it unattractive. The biological role of the amphibian papilla is discussed in ecological and evolutionary terms.

We are writing to respond to an article published in the Journal of Zoology257: 37–42 by Tuyttens, Macdonald and Roddam. The article describes analyses to investigate the potential effects on European badgers of the fitting and monitoring of radio-collars. The results suggest that there is a transient detrimental effect on body condition for up to 100 days post fitting.

A free-ranging maternity colony of big brown bats Eptesicus fuscus roosting in rock crevices along the South Saskatchewan River in south-eastern Alberta, Canada, was studied to understand better the discrepancy that exists in the literature regarding torpor use by reproductive female bats. Using radio-telemetry, thermoregulatory patterns and roost microclimate were recorded for pregnant, lactating and post-lactating females. Relative torpor use is described in several ways: the proportion of days on which torpor was used, depth, minimum body temperature, time spent in torpor, and a comprehensive torpor unit (degree-min). Pregnant and lactating female E. fuscus used torpor to the same extent overall (degree-min), but pregnant bats used torpor less frequently and with more time in deep torpor. Torpor was used to the greatest extent after weaning (post-lactation). Evidence is presented that the cost:benefit ratio for deep and prolonged periods of torpor may be highest during lactation. Microclimates of rock-crevice roosts mirrored the use of torpor throughout reproduction by bats. Lactation roosts (deeper, larger opening size) were more thermally stable and remained warmer at night compared to the shallow roosts used by pregnant and post-lactating females. It is shown that conclusions about relative use of torpor can differ depending on the units of comparison, necessitating measurement of all aspects of torpor (depth, duration and frequency). Comprehensive measurements, individual-based normothermic temperatures, and a definition of torpor that accounts for all energy savings, allow a more accurate depiction of patterns and facilitates inter-study comparisons.

The first MRP (matrix representation with parsimony) supertree phylogeny of the Lipotyphla is presented, covering all the families that were considered to make up the traditional mammalian order Insectivora. The phylogeny does not examine relationships within the shrew subfamily Crocidurinae, but all other taxa are considered at the species level, drawing upon 41 years of systematic literature and combining information from 47 published sources. The MRP technique is also critically discussed. This study will be of use to comparative biology studies of the Lipotyphla (or of mammals as a whole) and is a rigorous review of past systematic work, as well as clearly demonstrating our current level of knowledge. The supertree clearly details a strong imbalance in phylogenetic understanding across the taxon: a great deal is known about the hedgehogs and gymnures (Erinaceidae), the New World moles (Talpidae), Palaearctic species of Sorex (subgenus Sorex) and the relationships between genera of red-toothed shrews (Soricinae). The supertree, however, clearly shows areas where our knowledge is conflicting or non-existent, and these gaps do not always correspond to obscure species: nothing is known about the systematics of Old World mole genera. Also very little is known about golden moles (Chrysochloridae) and the shrew-tenrec genus Microgale, some of the most threatened mammals on Earth.

The underlying aim of our paper (Tuyttens et al., 2002) was to draw attention to the desirability – for reasons of science and animal welfare – of applying the scientific method to test for effects of techniques in field biology. Therefore, we are grateful to Delahay et al. (2002) for their response which not only advances that aim, but provides us with the opportunity to elaborate on the importance of the issue. In this reply we focus first on the questions they raise about our analysis and then, second, and more importantly, we set the answers in a broader context of what we foresee as a new era of animal welfare science as a component of conservation biology.

The objective of this study was to determine the sexual pattern of the Indian dascyllus Dascyllus carneus. After an initially undifferentiated state, gonads of D. carneus developed an ovarian lumen and primary growth stage oocytes, and subsequently cortical-alveolus stage oocytes. From ovaries with cortical-alveolus stage oocytes and from more developed ovaries, some gonads redifferentiated into testes. From a sample of 163 individuals, two had a gonad containing degenerating vitellogenic oocytes and proliferating spermatogenic tissue, nine had a gonad containing degenerating cortical-alveolus stage oocytes and spermatogenic tissue, and five had a gonad with degenerating primary growth stage oocytes and spermatogenic tissue. The size of these individuals overlapped greatly with the size range of mature females, suggesting that at least in some individuals, redifferentiation toward a testis occurred after spawning as females. This indicates that D. carneus is a functional, diandric protogynous hermaphrodite. Removal of a dominant male(s) did not induce a sex change in any of the ranking females in the laboratory and field groups. There was no difference in the number of chases and signal jumps performed by the ranking female between control and experimental field groups, or before and after removal of the male. However, the sizes of the ranking females were at or beyond the size range of individuals with a mixed-stage gonad, suggesting that the developmental window for female-to-male sex change may not be open ended. In 41 of 43 field groups, in which sex of fish was determined histologically or by the shape of the urogenital papilla, one to several highest size ranks were occupied by males, followed by one to several females. Mature males, however, were not limited to the highest ranks and occurred at various lower size ranks within groups. Individuals with a mixed-stage gonad also occupied various size ranks within groups.

Adult body size and shape were examined in almost 1400 individuals of the tortoises Testudo graeca, T. hermanni and T. marginata from Greece. The size at maturity was greater in females than in males in all three species. Maximum and mean adult sizes were also greater in females than in males in T. graeca and T. hermanni. Males grew to a larger size than females in T. marginata, and mean adult size was similar in the sexes in this species. Sexual dimorphism of shape (adjusted for size covariate) was shown in most of the characters examined, and the degree of this dimorphism differed significantly among the three species. Differences were related to their contrasting courtship behaviours: horizontal head movements and severe biting in T. marginata, vertical head bobs and carapace butting in T. graeca, and mounting and tail thrusting in T. hermanni. There was no difference in the frequency of observations of courtship or fighting among the three species, but courtship was about 10 times more common than combat in males. All species showed greatest courtship activity in autumn; copulation was rarely observed in T. hermanni (only 0.36% of courting males) and not seen in the other species in the field. Observations made throughout the activity season indicated that feeding was equally common in males and females in all three species. Differences in shape were more likely to be the result of sexual selection than of natural selection for fecundity. Detailed predictions are made for sexual dimorphism of other characters in these species.

Plethodontid salamanders of the North American subfamily Desmognathinae offer excellent opportunities for an integrative understanding of the form, functions and phylogeny of sexual behaviour patterns. The first description of courtship in the Black Mountain dusky salamander Desmognathus welteri is presented. Once oriented toward their partners, males produce multiple tactile, visual and chemical stimuli that probably function to stimulate, or ‘persuade’, females to mate. Persuasion is accomplished by behaviour patterns that include head rubbing, butterfly, pulling and snapping, and perhaps nudging and tail undulation. Sperm transfer is indirect by a spermatophore that is deposited on the substrate during a sequence of behaviour patterns involving both sexes known as tail-straddling walk. Fine-scale synchronization and orientation of partners at this time is crucial for successful insemination. The evolutionary histories of eight sexual behaviour patterns for 19 taxa of desmognathine salamanders are then reconstructed. Our phyloethological analyses suggest that the courtship of desmognathines consists of a mosaic of ancestral (plesiomorphic) and derived (apomorphic) behavioural traits. The greatest variation among taxa involves behaviour patterns that are exhibited early in sexual encounters as, or shortly after, males orient toward females. Behaviour patterns that occur at closer range and most obviously are persuasive in function show less variation among taxa, and those that lead to sperm transfer (tail-straddling walk) are invariant. While our phyloethological analyses reveal aspects of the evolutionary history of behavioural form in the Desmognathinae, further work is needed before thorough analyses of behavioural functions can be conducted.

The anuran lower jaw is composed of three pairs of bones: dentaries, angulosplenials and mentomeckelians. Although the lower jaw is toothless, except in Gastrotheca guentheri, enlarged fangs or odontoids have evolved at least four times independently in some myobatrachids, hylids, ranids and leptodactylids through both parallel and convergent evolutionary events. Fangs seem to represent the single best design solution to enable an anuran to inflict a bite-like wound, but the biological role of biting varies among species. Fangs are projections of the dentaries in ranids, but in the hylid frog Hemiphractus and in ceratophryine leptodactylids, they form a sinosteotic unit with the dentaries and mentomeckelians. Comparisons of morphology, behaviour and diet among frog taxa with enlarged fangs reveal that the fangs may be the result of either sexual or natural selection. Those fangs that evolved in response to sexual selection seem to be relatively larger than those that resulted of natural selection.

The predaceous crab Eriphia smithii (Xanthidae) has one larger claw with molar teeth on either the right or the left cheliped, which it uses to crush the shell of prey. Whether the handedness of crabs affected successful predation on two snail species, Nerita albicilla (Neritidae) and Planaxis sulcatus (Planaxidae) was experimentally investigated. The fate of snails of each species was analysed by multiple logistic regression with three explanatory variables: handedness, shell-size index and individuality of crabs. No effect of handedness was detected in attacks on N. albicilla, probably as a result of the spherical and more symmetrical shell morphology of this species. In contrast, right-handedness contributed to greater attack success on P. sulcatus, which has more conical shells. Further investigation of how snail shells were broken revealed that left-handed crabs had more difficulty breaking the aperture of larger P. sulcatus, which was thought to cause the difference in attack success between right- and left-handed crabs. The advantages conferred by handedness are discussed.

Flat lizards Platysaurus intermedius wilhelmi occur on small discrete rock outcrops in Mpumalanga Province, South Africa. These rock outcrops are structurally simple and this, combined with the lizard's behaviour (ambush foraging in the open), make them ideal for field studies of anti-predatory behaviour. Lizards were approached in the field and how escape behaviour was influenced by habitat and age-sex class was recorded. Juveniles (c. 4 months of age) responded quite differently to an approaching human ‘predator’ compared with adult males and females (which responded similarly). Compared to adults, juveniles allowed a closer approach by the investigator; took longer to find a refuge and therefore fled further; were more likely to remain visible in the open and maintain visual contact with the investigator; and more likely to flee into vegetation when given the opportunity to take refuge in a crevice. We suggest that because a greater suite of predators (including arthropods living in rock crevices) feed on small juvenile lizards, this may affect their choice of refuge and result in the avoidance of crevices when chased. Finally, because juveniles were frequently found on small rock outcrops, the influence of rock outcrop area on anti-predatory behaviour was tested. Escape behaviour (time to refuge) was independent of rock outcrop area.