In addition to being frugivorous, Cebus and Saimiri stand out among the New World primates of similar body size in being heavily dependent on animal matter for protein (faunivory). A detailed description of the morphology and behavior of the two genera is presented with the object of evaluating the interaction and respective contributions of morphological and behavioral adaptations to foraging patterns. Our conclusions include the following: First, body size is extremely important in explaining the observed variation in diet. Second, the emphasis on faunivory is facilitated more by behavioral than by morphological specialization. Third, whatever morphological specializations are present, particularly in Cebus, are probably favored by diet at the most food-depauperate time of year. Fourth, although morphology may well reveal what a primate may potentially eat, to map this potential onto actual diet requires a detailed knowledge of its natural ecosystem. Finally, we consider whether the behavioral data support the tenuous morphological evidence for grouping Cebus and Saimiri within the clade Cebinae.

Coordinated travel by social groups is well documented, often with evidence that cognitive spatial maps are employed. Yet the mechanisms by which movement decisions are made and implemented within social groups are poorly known. In a field study of white-faced capuchin monkeys in Costa Rica it was demonstrated that a specific call, the "trill," is used by adults in the initiation and directing of troop movement. The trills of subadults were restricted to vocal exchanges with other subadults. Continuous vocal recordings were collected of the vocalizations of the 14 members of the study troop. A cumulative 33.7 h of continuous samples and 1,892 sonagrams were analyzed. In addition to vocalizations clearly associated with alarm, distress, or agonistic contexts, two distinct call types were identified, trills and huhs. Age-sex classes differed in the rate at which both types of calls were produced in different spatial positions within the troop. Adult females and males produced higher rates of trills when in the leading edge compared to all other spatial positions in a traveling troop. Trills at the edge of a stationary troop represented 36 "successful" and 3 "unsuccessful" start attempts; the troop usually moved in the trajectory predicted by a trilling adult's location on the troop periphery within 10 min of the initiation of trilling. Adults also altered the trajectory of traveling troops by trilling at the side and back of the troop (10 "successful" and 4 "unsuccessful" attempts). Huh vocalizations were most predictably produced when a capuchin is in a dense fruit patch. These results emphasize the role vocalizations serve in the coordination and trajectory of group movement in nonhuman primates, especially those populations that are arboreal or in which visual contact is otherwise impeded. o 1993 WiIey-Liss, hc.*

Ecological and behavioral data from longterm field studies of known individuals in two closely related squirrel monkey species (Saimiri oerstedi and S. seiureus) were used to examine hypotheses about the source of variation in female bonding among groupliving primates. Social relationships in species which live in cohesive groups are thought to depend on the nature of competition for resources. S. oerstedi and S. sciureus both live in large groups and are subject to intense predation. Direct feeding competition both between and within groups is extremely low in S. oerstedi; in this species female relationships are undifferentiated, no female dominance hierarchy is evident and females disperse from their natal group. S. sciureus also experiences very low levels of between-group competition, but within-group direct competition for resources is frequent; this species demonstrates differentiated female relationships, a female dominance hierarchy, and female philopatry. The correlated ecological and social variables found in these two congeners further minimize the minor effects of phylogenetic differences and emphasize the importance of food distribution in determining social characteristics.

Formerly applied to studies of responsivity in children, in more recent years the concept of temperament has been applied to nonhuman primates at the individual, species, and now population levels. While the concepts of temperament and personality have been less distinguished in nonhuman primate studies than in the human literature, temperamental and personality differences have now been identified among individual primates and among primate species in a number of studies. At the individual level, certain temperamental characteristics have been associated with age, sex, and most frequently rank. At the species level, temperamental profiles have been linked to intraspecific differences in social systems, sociodemographics, and features of life history and ecology. In this report we discuss the application of the temperament concept to nonhuman primates and review findings from studies of primate temperament at the individual, population, and species level. We also cite evidence for genetic and experiential influences on temperament in primates, outline concepts related to possible evolutionary influences on temperament, and discuss the possible relation of temperamental characteristics to social behavior and ecology in selected species. o 1995 Wiley-Liss, Inc.Key words: temperament, personality, behavioral traits, reactivity, individual differences, species differences INTRODUCTION: THE CONCEPT OF TEMPERAMENTThe concept of temperament originated from studies of infants and young children, in which consistent response styles in novel situations could be identified in a certain proportion of children at an early age. Such terms as bold vs. shy, inhibited vs. uninhibited, highly reactive vs. unreactive, highly fearful vs. unfearful, and emotional vs. unemotional have been used to characterize the behavioral styles of young children in response to novel and/or challenging situations. The similarity among terms indicates the primary dimension by which children are characterized-that is, whether they tend to approach or avoid novel stimuli and to what extent they show negative emotional and heightened physiological responses to 104 I Clarke and Boinski novelty. While the precise definitions vary in the literature, the term temperament refers to behavioral styles or tendencies (rather than discrete behavioral acts) that show continuity over time and can be identified in early infancy. Such characteristic styles are reflected in the degree and nature (i.e., approach vs. avoidance [Kagan et al., 19921) of responsivity to novel or stressful stimuli, and nearly all temperamental frameworks include a dimension of intensity, distress, or emotionality [Lewis et al., 19891. In addition to behavioral responsivity, physiological reactivity to challenge is considered an important component of temperament, since it is believed that temperamental response styles are based on biological underpinnings that are reflected in physiological responsiveness [Fox, 1989; Goldsmith et al., 1987; Gunnar, 1988; Healy, 1989; Kagan, 1989; Kagan ...