Environmental variation and developmental rate among free ranging yellow baboons (Papio cynocephalus).код для вставкиСкачать
American Journal of Primatology 35:15-30 (1995) Environmental Variation and Developmental Rate Among Free Ranging Yellow Baboons (Papio cynocephalus) LAUREN M WASSER’ AND SAMUEL K WASSER’ ’Department of Psychology, and ’Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Washington, and Center for Wildlife Conservation, Seattle, Washington Behavioral development was observed for the first 3 months of life on 18 infants from three troops of yellow balloons (Papio cynocephalus) in Mikumi National Park, Tanzania. Developmental rates were determined for each subject on six major behaviors using the slope from regression equations performed on developmental functions changes in behavior with age. These slopes were examined statistically for associations with social and physical variables that were hypothesized to influence developmental rates. Results showed that faster rates of development were significantly associated with certain environmental and social conditions. Infants born late in their cohort, infants born in a more physically immature state, infants living in a large troop, and female infants of low ranking mothers all had significantly faster rates of development than infants a t the other extreme in these categories. These results are interpreted as evidence for developmental processes that compensate for environmental or physical conditions that negatively influence survivorship. o 1995 WiIey-Liss, Inc. Key words: development, rate, environmental variation, compensatory processes, infant, baboon INTRODUCTION Behavioral development is a process that occurs through a n interaction between a n individual and its environment [Immelman et al., 19811. This interaction evolved in nature where the pressures that shaped i t still exist, and for this reason, it is important to study development in natural settings as well as in controlled laboratory environments [Hines, 1942; Akins et al., 1980; Sackett, 1967; Altmann, 19801. The dynamic nature of development makes its characterization empirically difficult. One approach to this problem is to measure developmental rates [Wasser, 19881. The rate of development, or “velocity,” is a n important characteristic of developmental processes that is measurable. The rate of development of a given trait or behavior is obtained by plotting a quantitative measure of a trait or be- Received for publication May 28, 1993; revision accepted March 9, 1994 Dr. Lauren Wasser’s current address is 11702 Durland Avenue, NE, Seattle, WA 98125. Address reprint requests to Dr. Samuel K. Wasser, Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, XD-44, University of Washington, Seattle, WA 98105. 0 1995 Wiley-Liss, Inc. 16 I Wasser and Wasser havior by age, producing a “developmenbal function.” Then the data is fit with a least squares regression equation to approximate the data with a straight line. The slope of the regression line is an estimate of the rate or velocity of the developmental function. Rasmussen and Tan  conducted a comparative analysis of the rate of behavioral development in primates, using Gompertz models. They found interspecific variation in behavioral development was significantly related to differences in brain and body size. Larger species had slower rates of development. Terrestrial species reached locomotor independence earlier than did arboreal species when controlling for body size, although the actual developmental rates did not differ between species. We have conducted a n intraspecific analysis, examining how environmental and physical variations effect the rate of behavioral development in yellow baboons. The association between environmental variables and the rate of behavioral development was examined in six major infant behaviors on 18 baboons over their first 3 months of life. Baboons provide a n excellent model for studying behavioral development. They are a n altricial primate living in large multi-male, multifemale groups with complex social relationships. Baboon infants spend a large part of their early months on or very close to their mother, relying on breast milk a s a major source of nourishment. Like humans, baboon infants gain independence slowly, using the mother a s a safe base from which to explore. However, baboons mature roughly four times a s fast. METHODS Observations were made on three troops of yellow baboons in Mikumi National Park, Tanzania. The Mikumi baboons have been studied almost continuously since 1974 and are well habituated to the close proximity of observers on foot. These troops are the subjects of a n ongoing, long-term research site initiated in 1974 by Dr. Ramon Rhine and graduate students. Details of the demography, and female reproductive behavior for these troops can be found in Rhine et al. , Wasser and Starling , and Wasser and Norton 119931. The history of the baboon study troops a t Mikumi includes two troop fissions. Three troops were formed when the large original study troop (Viramba troop) fissioned in late 1978 and again in early 1979. The three troops studied are now referred to as V1, V2, and V3. In 1986, when this study was conducted, V1 contained 74 animals (13 adult males, 25 adult females, 12 study infants, and 24 juveniles and subadults), V2 contained 68 animals (9 adult males, 19 adult females, 10 study infants, and 30 juveniles and subadults), and V3 contained only 13 animals (1 adult male, 4 adult females, 4 study infants, and 4 juveniles and subadults). Of the 26 infants born during the year of the study, 2 were killed by adult males (eyewitnessed by authors) and five more died during their first year of life. The descriptive statistics (establishment of developmental functions) include data on all 26 infants, but the developmental analyses (correlation of slopes and intercepts of developmental functions with independent variables) are based on data from a subset of 18 infants for whom a minimum of 12 follows each were recorded. One of the independent measures, the neonatal assessment (see below), took several months to develop, and only 7 infants were reliably assessed with that measure. Observations were made on members of one troop each day, 5 days per week. The troop would be found at one of its sleeping sites each morning (-0700 hr) and followed on foot throughout the day until 1630 hr. Each mother-infant pair in the study was followed twice daily, once in the morning and once in the afternoon. Each focal dyad follow [Altmann, 19741 lasted 30 min. There were a n average of 17 Environmental Variation and Rate of Development / 17 TABLE I. The Six Major Infant Behaviors and Proximity Measures Sleep and rest Nipple contact Ventral contact 1-5 meters proximity Explore Play Infants eyes are closed, or infant is inactive and not in proximity to any animal other than the mother. Infant has oral nipple contact with mother. Infant and mother are in ventral-ventrum contact. Infant is between 1 and 5 meters away from mother. Infant engaging in oral or manual manipulation of the environment. Social behavior characterized by freedom of movement, with total body involvement, often accompanied by an open mouth play face. Particularly mouthing and wrestling, chasing, and other rough and tumble behaviors. follows per mother-infant dyad in the study. Only data from ages 0-3 months were used in the developmental analyses. Two types of data were collected: (1) An instantaneous record [Altmann, 19741at 1-min intervals of both mother and infant behavior and their proximity to each other, and (2) a continuous record of the social interactions involving either member of the dyad. The continuous data were used in a n analysis of maternal style (see below). The rest of the analyses used summaries of the instantaneous behavior and proximity data. Both dyad members were almost always within close proximity, (0-5 m) because of the young age of the infants. However, the infant remained the focal subject whenever only one member of the dyad could be reliably observed. The six infant behavior categories used in the analysis are defined below in Table I (details of the entire scoring system can be found in Wasser . These behaviors constituted 86% of the total number of instantaneous observations in the study. Data from the instantaneous samples were summarized to estimate the proportion of time each infant was observed in a given behavior category for each 2-week age block. These summaries produced a set of six developmental functions for each infant (graphs of the estimates of proportion of time spent in a behavior by age). Functions were fitted with a least squares linear regression equation, and the slope of each function was used as a dependent measure of the rate of behavioral development. Multiple significant findings were expected because these dependent measures were intercorrelated. INDEPENDENT MEASURES Organismic Variables The independent measures were of either organismic or environmental origin. The organismic variables included a variety of measures obtained from a neonatal assessment of the condition of infants at birth and the functioning of certain important reflexes (see below). These measures were included to provide information about variability in the physical competency of the neonates. Similar measures have been useful in both short and long term prediction of development in human and non-human primate infants [humans: Prechtl & Beintema, 1964; Brazelton, 19731 (pigtailed macaques: Sackett, unpublished observations). The assessment was performed by rating the infant on a variety of measures using a three-point scale (see Appendix for details on the Neonatal Assessment). Measures included infant sex, the general condition of the infant (such a s size, skin color, presence of facial and body wrinkles, coat condition, and the presence of head molding and facial bruising), early reflexes (including the palmar and plantar grasp, arm and 18 I Wasser and Wasser leg clasp, visual follow, neck tone, rooting and sucking, and flexion of the arms, back, and legs when standing), infant’s balance when standing, and the degree of maternal support provided while clinging. This assessment was performed from the first day of contact with a n infant until the infant had matured beyond all of the measures in the assessment. Mean interobserver reliability on the rating scales was 0.95 (percent agreement) during reliability sessions in which two observers simultaneously but independently scored the same infant. Environmental Variables Environmentally based independent variables consisted of social factors as well a s experiential effects mediated by characteristics of the mother. These included troop size, maternal dominance rank, maternal parity, birth order within the cohort, maternal style, and infant handling by non-related animals. Altmann [ 19801 found significant intraindividual variation in maternal style in baboons at Amboseli National park in Kenya. In the present study, maternal style was quantified by calculating the percentage of an infant’s social interactions that the mother interrupted by retrieving and withdrawing with her infant. A high score on maternal style was indicative of a more restrictive mother. Maternal dominance rank was calculated from a matrix of directional supplant behaviors (S.K. Wasser, unpublished data). The linear female dominance structure was divided into quartiles to enable comparisons between troops of different sizes. The parity of each mother was determined from a review of the past 12 years of Mikumi records. Infant handling was defined as non-nurturant contact with a n infant by individuals other than its mother. Infants attract much interest from troop members, though not all of the interactions that result are affiliative [Wasser, 19831. Baboons of all ages, and especially females, often handle infants in a manner that is uncharacteristic of mother-infant contacts. This class of behaviors, termed infant handling, was divided into two categories, gentle handling and rough handling. Gentle handling included pulling of limbs, poking, holding away from the ventrum and holding upside down or sideways. Rough handling was defined as dragging, dropping, hitting, or yanking the infant. Both categories involve non-nurturant contact, but the rough handling category was reserved for potentially injurious behaviors. STATISTICAL ANALYSES Data were analyzed with a series of linear and second order polynomial regressions. The six dependent behavioral measures (Table I) were analyzed using data from all of the subjects. Estimates of the proportion of time spent in each behavior were calculated from the 1-min interval instantaneous records of infant behavior taken during the 30-min follows. These proportions were plotted by age and fitted with linear or quadratic regression to produce six descriptive developmental functions. Data from eighteen subjects on whom at least twelve follows existed were included in the analyses of the effects of the independent variables on the rate of development. Again, developmental functions were plotted for each of the six behaviors, but with two differences. Six separate functions (one for each behavior) were plotted for each subject, and the functions only included data from birth to 3 months of age. Again the developmental functions were fit with linear regressions (quadratic regression was unnecessary because of the focus on only the first 3 months of age). The slopes and intercepts of these regression equations became the dependent variables in subsequent analyses, a s estimates of the rate of development and the initial level of the behavior a t birth, respectively. These dependent Environmental Variation a n d Rate of Development I 19 measures of development were examined for their relationship to a number of independent variables including: troop, sex, infant handling, maternal style, maternal dominance, birth order, and neonatal reflexes. Both parametric and nonparametric statistical techniques were employed in these analyses depending on the scaling of the independent variables (ordinal scale used nonparametric, interval scale used parametric). RESULTS Overall Developmental Functions The overall developmental functions are shown in Figure la-f. These functions represent the early behavioral development of the entire 1986 cohort of infants in the Mikumi study troops. Figure 1 shows the change in the amount of nipple contact infants had with their mothers. From a n initial value of 61%, the average percentage of daytime that the infants spent in nipple contact decreased significantly and linearly over the first 200 days, at a rate of 0.25% a day (R = 0.95, P < 0.001.) The amount of time infants slept or rested during the day also decreased linearly with age over the first 6 months of life (Fig. lb). One to five meters proximity from the mother is shown in Figure lc. This proximity measure increased steadily from 0% a t 1 week of age to about 25% a t 6 months of age (r = 0.80, P < 0.001). In contrast, other early behaviors showed curvilinear developmental functions. Mother-Infant Contact (Fig. Id) had a strong quadratic component. This behavior increased slowly a t first and then accelerated after the first 2 months of the infants life (r = 0.96, P < 0.001). Exploration (Fig. l e ) and Play (Fig. I f ) also had quadratic components, but showed more variability (r = 0.74 and 0.86, respectively, P < 0.001 for both). Troop Differences Troop differences in infant development were examined by analysis of variance using the slope of a given developmental function for each infant as the dependent variable. Only one of the six infant behaviors showed a significant troop difference in the overall F test. Nipple contact with the mother declined significantly faster in troop V l (F = 7.46, P = 0.006)-the largest and most competitive of the troops [Wasser & Starling, 19881-than in either troops V2 or V3. Two other behaviors, ventral contact with the mother, and sleep and rest, showed a similar pattern with respect to troop differences, although neither reached significance in the overall F test. Pairwise comparisons using Fisher’s PLSD test [Welkowitz et al., 19821 showed that V2 vs. V3 were reliably different for ventral contact, and V1 vs. V3 were reliably different for sleep and rest (P < 0.05). Here, too, infants in the larger and more competitive troops (V1 and V2) tended to develop faster than those in the small troop (V3). Sex Differences Sex differences were examined using an unpaired Student’s t test. None of the six behaviors showed a significant main effect for sex differences. However, a significant sex difference was found in a n interaction with maternal dominance rank (see below). Infant Handling The effects of infant handling by troop members other than the mother were assessed with simple and multiple regression analyses, and the Spearman Rank Order Correlation statistic for ordinally scaled variables. Infant handling data 20 I Wasser and Wasser 80 20 10 - a. d. b. e. C. f. - 90 80 70 60 50 40 30 20 401 10 18 - 30 0 . , . , . I . , . , . I . , . , . , . , . I . , . , . I . , O P ~ N W O ~ ~ N W O ~ C O N - - ~ b m k c a r n - ~ b m w m r n . - Age in Days Fig. 1. Developmental functions and simple regression lines for six behaviors averaged over all subjects in the 1986 birth cohort. a: Nipple contact: y = 61.06-0.227 X , r = 0.95; b: Sleep and rest: y = 86.66-0.28 x , r = 0.96; c: 1-5 Meters from mother: y = -2.86 + 0.123 X , r = 0.80; d Mother-infant contact; y = 89.2 + .04 x - .OO x *, r = 0.96; e: Exploration: y = 4.3 + .29x - ,001 x 2 , r = 0.74; f: Play: y = -.48 + .1 x .003x2, r = 0.86. - were collected continuously allowing a frequency measure per 30 min follow. Infant handling frequencies refer to observations made on infants in the first month after birth, when handling was a t its peak. A significant negative relationship was found between the dominance quartile of the mother and the amount of handling W O Environmental Variation a n d Rate of Development / 21 2o 1 0 Q 1 2 3 4 Maternal Dominance Quartile Fig. 2. Correlation between maternal dominance quartiles (1 = highest rank by convention) and the frequency of infant handling by nonrelatives per 30 minute follow during the first month of life (Spearman r = 0.40, P < 0.05, n = 19). her infant received in the first month of life (Fig. 2). (By convention, the highest ranking quartile was encoded as rank 1,and the lowest ranking quartile as rank 4.) Infants of low-ranking mothers received significantly more handling than infants of high-ranking mothers. The least handled infant experienced 25% of its social interactions in a nonnurturant handling context, while other infants experienced up to 70% of all their social interactions as non-nurturant handling. No significant relationship was found between infant handling and any measure of infant development used in this study. However, the frequency of rough handling in the first month of life did show a significant association with survivorship to 1 year of age (see below). Maternal Style A significant relationship was found between maternal restrictiveness (see Methods) and maternal dominance quartile (Fig. 3). High ranking mothers were significantly less restrictive of their infants than were low ranking mothers, consistent with Altmann’s  data from Amboseli. No significant relationship between maternal style and the rate of infant handling was found. Maternal Dominance Quartile Female but not male infant development showed a significant relationship to maternal dominance quartile. Female infants of low ranking mothers developed at a faster rate than did female infants of high ranking mothers. This was true for both ventral contact and play (Fig. 4) (Spearman r = -0.66, P < 0.03 ventral contact; Spearman r = .78, P < 0.007 play). Because ventral contact declines with age, a more negative slope indicates faster development. By contrast, rates of play increase with age. Thus, a more positive slope indicates faster development. Birth Order The birth year a t Mikumi runs from December, the first month of the rainy season, to November [Rhine et al., 1988; Wasser & Norton, 19931. All of the infants born in the same troop in a given birth year comprise a single birth cohort. Birth 22 I Wasser and Wasser 0 1 2 3 4 Maternal Dominance Quartile Fig. 3. Correlation between maternal restrictiveness (defined as the percentage of the infant’s social interactions that the mother interrupted, retrieved, and withdrew with the infant) and maternal dominance quartile (Spearman r = 0.72, P < 0.001, n = 23). order was defined as the numerical order of birth of a n infant in its birth cohort. Spearman r correlation tests revealed a significant relationship between birth order and the rate of infant development (Fig. 5). Infants born later in the birth cohort, developed independence from nipple contact with the mother significantly faster than early born infants (Spearman r = -0.47, P < 0.02). Neonatal Assessments The neonatal assessment scores consisted of two summary variables derived by taking the sum of the scores in two conceptually distinct groups of data: (1)condition of the infant a t birth [i.e. dehydrated, bruised or otherwise traumatized (head molding), noticeably small or large body size1 and ( 2 ) the functioning of the neonatal reflexes (see Appendix). Spearman tests of these ordinally scaled variables showed no significant relationship between infant condition a t birth and any measure of behavioral development. Reflex functioning, however, showed a significant relationship to two behaviors, ventral contact (r = 0.79, P < 0.05, Spearman test) and play (r = 0.85, P < 0.02, Spearman test) (Fig. 6). A high score on reflex functioning indicates a more mature infant, with 18 being the maximum score attainable. Infants with more immature reflexes a t birth reduced ventral contact a t a faster rate over the ensuing 3 months than did infants who scored more mature on the neonatal reflex measures. Play increases with age, and the correlation with neonatal reflexes was in the opposite direction from ventral contact. Thus, a similar interpretation can be made. Infants scoring lower (more immature) on the neonatal assessment measure showed faster rates of change in play behavior toward more mature levels of this behavior. Analysis of Y-Intercepts The value of the function a t the y intercept is a n estimate of the starting point of the individual on that measure, and allows us to distinguish between the two possible interpretations for slope differences. If the intercepts were not significantly different, the presence of a significant difference in slopes of developmental Environmental Variation and Rate of Development I 23 -0.8 I I 0.2. -0.1 0 0 1 2 3 4 Maternal Dominance Quartile Fig. 4. Correlation between maternal dominance quartile and the slope of the ventral contact function in female infants (a)and the slope of the play function in female infants (b)(Spearman r’s = -0.66, P < 0.03 and 0.78, P < 0.007, respectively, n = 9). Since ventral contact is a decreasing function a more negative slope indicates a faster rate of development on this measure. functions would reflect a n acceleration in development for a given age (i.e., reach a certain point earlier). On the other hand, significantly different intercepts would reflect the attainment of age-appropriate developmental abilities from a more immature starting point. The same statistical procedures employed in the analyses of the slopes with the various independent variables were also used to test for significant differences in the y-intercept values. Infants born late in the cohort tended to show higher initial levels of nipple contact than infants born early in the cohort (Spearman, r = 0.49, P < 0.05, n = 18, two-tailed). No other significant relationships were found between the initial level of any behavior and the independent variables. Analysis of D a t a at 3 Months of Age Analyses of behavioral outcomes at 3 months of age were conducted using the behavioral data from 85 -98-day-old infants. Simple regression analyses revealed 24 I Wasser and Wasser -2 0 2 4 6 0 10 12 Birth Order Fig. 5 . Correlation between birth order within the troop and the slope of the nipple contact function (Spearman r = -0.47, P < 0.02, n = 18).Nipple contact is a decreasing function so a more negative slope indicates a faster rate of development on this measure. no significant effects of birth order. However, infants assessed as more immature a t birth on the reflex assessment showed significantly less ventral contact (R = .84, P < .02) a t 3 months of age. Maternal dominance had significant effects on both males and females a t age 3 months. Male infants of low ranking mothers spent less time in nipple and ventral contact with their mothers (R = .77, P < .04 and R = 3 2 , P < .03, respectively) whereas female infants of low ranking mothers slept less and played more than same sex infants of high ranking mothers (R = .76, P < .05 and R = .75, P < .02, respectively). These patterns are consistent with the hypothesis that low ranking infants gain independence faster to compensate for other disadvantages of their mother’s low rank. Survivorship to One Year Both the independent and dependent variables in this study were used as predictors of infant survivorship to one year of age in a multiple regression analysis. Only rough infant handling showed a significant relationship to survival to one year. Infants that experienced relatively high rates of rough handling during the first month of life had significantly lower survivorship to one year of age (r = -0.47, P < 0.05). DISCUSSION Interpretation of the data in this paper is limited by several factors: (1)The number of infants available for study was small. (2) Daily contact with any single troop was not possible because of the need to monitor three different troops. (3) The data could have been influenced by weather and ranging conditions, which vary annually. (4) Repeated statistical tests were performed on the same data set, increasing the probability of a Type I error. However, the consistent pattern of significant results diminishes the possibility that they resulted from chance correlations. In all cases, increased velocity of development was associated with disadvantageous conditions, i.e., conditions that could be associated with higher infant mortality (see below). Such consistency would have been very unlikely due to chance. Environmental Variation and Rate of Development I 25 a. 0.0 -0.8 I b. O'*l 0.0 J 12 \ 13 14 15 16 17 18 I 19 Score on Reflex Assessment (age 1 week) Fig. 6. a,b: Correlations between the score on reflex assessment (low score = more immature, maximum obtainable = 18) and the slopes of ventral contact (Spearman r = 0.79, P < 0.05, n = 7 ) and play (Spearman r = -0.85, P < 0.02, n = 7). The results of the descriptive analyses showed substantial developmental changes in the six primary infant behaviors. Three of these behaviors were decreasing functions with age: nipple contact, ventral contact, and sleep and rest; the others were increasing functions with age: 1-5 m proximity to mother, exploration, and play. Troop membership, maternal dominance rank, infant sex, birth order within the cohort, and reflex functioning a t birth each had significant associations with the rate of behavioral development (see Table I1 for summary). Infants born into the large, more competitive troops, infants born late in their cohort, infants assessed as more immature a t birth in terms of reflex functioning, and females born to low ranking mothers, all displayed significantly faster rates of behavioral development on a wide variety of measures, compared to infants at the opposite extreme on these measures. Maternal style was significantly related to maternal dominance rank, the higher the mother's rank the less restrictive she was of her 26 I Wasser and Wasser TABLE 11. Summary of Significant Associations Between Independent Variables and Rates of Behavioral Development (Slopes of Developmental Functions) Indeoendent variable Dependent variable (slope of 1 Stat. P N 10.005 18 10.05 10.05 c0.05 18 18 18 10.02 <0.05 10.02 10.03 10.007 18 7 7 11 11 ANOVA F Troop (overall) Nipple contact Troop (V1 vs. V3) Troop (V2 vs. V3) Troop (V2 vs. V3) Fisher PLSD Sleep and rest PLSD Sleep and rest PLSD Ventral contact PLSD Birth order Neonatal assessment Maternal dominance" = 7.457 = = = 0.366 0.388 0.336 Spearman Rank Order R Nipple contact r = -0.47 Ventral contact r = 0.79 Play r = -0.85 Ventral contact r = -0.66 Play r = 0.78 "Significant for female infants only. infant's social interactions. Finally, rough handling in early infancy had a significant negative association with infant survival to one year of age. One interpretation of these data is that accelerated development occurred in response to disadvantageous developmental conditions. Membership in a large troop may be disadvantageous because of increased social tension and competition [Wasser & Starling, 1988; Wasser & Norton, 19931. Birth order within the annual cohort is a n important socially mediated variable. Infants born early in their cohort have a competitive age advantage over their peers throughout infancy and the juvenile period [Rhine et al., 1988; Wasser & Norton, 19931, and have few age mates in the early months of development. Late-born infants are, on average, competitively disadvantaged because of their younger age. Perhaps the faster rate of development of late born infants enabled them to compete more effectively with their peers up until and beyond weaning. One stimulus for this developmental compensation may have been the presence of numerous, slightly older peers with correspondingly greater physical and behavioral skill levels. The correlation of developmental rates with maternal dominance rank quartile also followed the pattern of developmental acceleration in the face of potential social disadvantage: infants of low ranking mothers showed significantly accelerated behavioral development. However, this was only true for female infants; the rate of behavioral development of male infants showed no relationship to maternal dominance rank. This sex difference may be one of the first signs of the strong but often subtle female competition characteristic of these troops [Wasser & Starling, 19881. Baboons have a matrilocal, matriarchal social system. Females remain in their natal troop and assume the dominance position under their mother; males emigrate to another troop a t sexual maturity, dissociating themselves from the impact of their mother's low rank [Altmann, 19811. Accelerated development in these low ranking female infants may offset their disadvantage to some extent. However, both male and female infants of low ranking mothers showed more mature levels of behavior a t 3 months of age. This suggests that maternal rank is a n important factor for males a t this very young age as well. Infants of low ranking mothers received significantly higher absolute frequencies of infant handling, as Environmental Variation and Rate of Development I 27 well as other social behaviors. This increased social interaction may facilitate the behavioral development of these infants, although other factors may be important a s well. The neonatal assessment also was consistent with the hypothesis of accelerated development accompanying potential disadvantages. However, in this case, the disadvantage was physical rather than social. Infants who were assessed as less immature a t birth showed faster rates of development than infants assessed as more mature. This finding is analogous with the catch-up development demonstrated by premature and small-for-dates human and non-human primate infants relative to their normal peers [Astbury et al., 1983; Siegel, 1983; Ungerer et al., 1983; Parkinson et al., 19863. It is important to distinguish between catch-up and compensation. Catch-up refers to a n increased rate of change in a parameter initially assessed as immature (e.g., weight gain in premature human infants). Compensation refers to faster development of behaviors that were not necessarily initially below the norm, but still developed more rapidly to compensate for disadvantages experienced postnatally. In our study, the birth order and maternal dominance effects on rate of development are viewed as examples of compensatory as opposed to catch-up development. Smart [ 19651 reported a n example of what we are calling a compensatory process in development in ducks. He found that the primary feathers of late hatched Redhead ducklings (Aythya americana) emerged a week earlier than those of early hatchlings, and the late ducklings fledged at a younger age. The need to fly by the end of the nesting season presumably puts considerable selection pressure on this type of developmental plasticity in late nesting birds. Compensatory processes have been considered a fundamental rule of behavioral development [Bateson, 19761 because they engender tolerance for wide environmental fluctuations. The above arguments support the hypothesis of accelerated development to compensate for social or physical disadvantages in baboon infants. This interpretation is not without precedent, and has considerable theoretical validity from a n evolutionary perspective. Organisms possesses a myriad of regulatory processes that adjust to and compensate for disruptions, from the genetic control of morphogenesis in fruit flies [Waddington, 19421 to the regulation of the water and mineral balance in mammalian cells [Ganong, 19811, to the control of physical growth in children [Tanner, 19781. It is reasonable to extend this principle to the realm of behavioral development a s well. Indeed, it would be surprising if behavioral development did not show some compensatory processes. However, if acceleration in the rate of development provides individuals with a competitive edge, then why hasn’t “run-away’’ selection continually operated to make faster rates the norm? It probably has to some degree. However, acceleration of development may eventually become prohibitively costly. Further acceleration beyond the norm may be beneficial in the long run only for individuals who are initially at a physical or environmental disadvantage. There is also the need to coordinate the development of many different processes in complex organisms like primates. Accelerated development in one area may be limited by the need to integrate i t into other areas to maintain a functional whole. An alternative hypothesis for the results in this study places the focus of causality on the mother. Mothers who find themselves in more disadvantageous conditions (e.g., low ranking, giving birth late in the season, having a n infant with underdeveloped reflexes) may choose to cover their losses by investing less in their infants. This would force their infants to develop independence more rapidly by virtue of their mothers greater neglect, not by any intrinsic developmental process. This hypothesis could only account for the effects of these variables on infant 28 I Wasser and Wasser behaviors mediated by the mother, such as nipple contact, ventral contact, and proximity, but not for the infant initiated behaviors of play and exploration. Also, from a n adaptive standpoint, mothers might be better off in terms of lifetime reproductive success by abandoning such high risk infants, reducing their interbirth interval. The alternative of reducing investments in these infants would further reduce their offsprings chances of survival, resulting in continued costs with even less chance of accruing a net benefit. CONCLUSIONS 1.The rate of behavioral development provides a quantifiable and informative dependent measure. 2. Significant positive associations exist between the rate of behavioral development and physical and social disadvantages a t birth. 3. These rate changes are interpreted as to compensatory processes operating to minimize the effects of the social and physical disadvantages. And 4.The rate of behavioral development should be a useful measure to investigators in a variety of fields, especially to studies on the effects of social, ecological, and physical variables. ACKNOWLEDGMENTS This work was part of the doctoral dissertation research of the author. Warmest thanks go to Dr. Gene P. Sackett, advisor on the project, Mary Liebermann for data collection, and Dr. Jeanne Altmann for comments on the manuscript. We also thank Dr. Ramon Rhine, Mr. Guy Norton, Mr. C. Kibasa, and Tanzanian National Parks, the Serengeti Wildlife Research Institute, and the Tanzanian Commission for Science and Technology for permission to work a t Mikumi. REFERENCES Akins, F.R.; Mace, G.S.; Hubbard, J.W. &. Akins, D. BEHAVIORAL DEVELOPMENT OF NONHUMAN PRIMATES: AN ABSTRACTED BIBLIOGRAPHY. New York, Plenum Press, 1980. Altmann, J. Observational study of behavior: samuling methods. BEHAVIOUR 49: 227-267: 19y4. Altmann, J. BABOON MOTHERS AND INFANTS. Cambridge. - , MA. Harvard University Press, 1980. Astbury, J.A.; Orgill, A.A.; Bajuk, B.; Yu, V.Y. Determinants of developmental performance of very-low-birth-weight survivors a t one and two years of age. DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY 255'09-716, 1983. Bateson, P.P.G. Rules and Reciprocity in Development. Pp. 401-421 in GROWING POINTS IN ETHOLOGY. P.P.G. Bateson; R.A. Hinde, eds. Cambridge, Cambridge University Press, 1976. Brazelton, T.B. Neonatal Behavioral Assessment Scale. CLINICS IN DEVELOPMENTAL MEDICINE (No. 50). London, Heinemann, 1973. Ganong, W.F. REVIEW OF MEDICAL I PHYSIOLOGY, 10th edition. Los Altos, CA, Lange Medical Pub., 1981. Hines. M. CONTRIBUTIONS TO EMBRYOLOGY. Carnegie Institute, 196:155-209, 1942. Immelmann, K.; Barlow, G.W.; Petrinovich, L.; Main, M. eds. BEHAVIORAL DEVELOPMENT, THE BIELEFELD INTERDISCIPLINARY PROJECT. Cambridge, Cambridge University Press, 1981. Parkinson, C.E.; Scrivener, R.; Graves, L.; Bunton, J.; Harvey, D. Behavioral differences of school age children who were small for dates babies. DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY 28:498-505, 1986. Prechtl, H.; Beintema, D. THE NEUROLOGIC EXAMINATION OF THE FULL TERM NEWBORN INFANT. London, Heinemann, 1964. Rasmussen, D.T.; Tan, C.L. The allometry of behavioral development: Fitting sigmoid curves to ontogenetic data for use in interspecific allometric analyses. JOURNAL OF HUMAN EVOLUTION 23:159-181, 1992. Rhine, R.; Wasser, S.K.; Norton, G. Eightyear study of social and ecological corre- Environmental Variation and Rate of Development I 29 lates of mortality among immature baboons of Mikumi National Park, Tanzania. AMERICAN JOURNAL O F PRIMATOLOGY 16:199-212, 1988. Sackett, G.P. Some persistent effects of different rearing conditions on preadult social behavior of monkeys. JOURNAL OF COMPARATIVE AND PHYSIOLOGICAL PSYCHOLOGY 64(2):363-365, 1967. Siegel, L.S. Correction for prematurity and its consequences for the assessment of the very low birth weight infant. CHILD DEVELOPMENT 54:1176-1188, 1983. Smart, G. Development and maturation of primary feathers of Redhead ducklings. JOURNAL OF WILDLIFE MANAGEMENT 29:533-536, 1965. Tanner. J . FOETUS INTO MAN: PHYSICAL GROWTH FROM CONCEPTION TO MATURITY. Cambridge, MA, Harvard University Press, 1978. Ungerer, J.A.; Sigman, M.; Beckwith, L.; Cohen, S.E.; Parmelee, A.H. Sleep behaviors of preterm children at three years of age. DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY 25:297-304, 1983. Waddington, C.H. Canalization of develop- ment and the inheritance of acquired characters. NATURE, Nov. 14:3811, 1942. Wasser, L. Marra compensatory processes in the behavioral development of wild vellow baboons (Papio cynocephalus). DISSERTATION ABSTRACTS INTERNATIONAL. Ann Arbor, MI, 1988. Wasser, S.K. Reproductive competition and cooperation among female yellow baboons. Pp. 349-390 in SOCIAL BEHAVIOR OF FEMALE VERTEBRATES. S.K. Wasser, ed. New York, Academic, 1983. Wasser, S.K.; Norton, G. Baboons adjust secondary sex ratio in response to predictors of sex-specific offspring survival. BEHAVIORAL ECOLOGY AND SOCIOBIOLOGY 32:273-281, 1993. Wasser, S.K.; Starling, A.K. Proximate and ultimate causes of reproductive suppression among female yellow baboons at Mikumi National Park, Tanzania. AMERICAN JOURNAL OF PRIMATOLOGY 16: 97-121, 1988. Welkowitz, J.; Ewen, R.B.; Cohen, J. INTRODUCTORY STATISTICS FOR THE BEHAVIORAL SCIENCES, 3 ed. New York, Academic Press, 1982. 30 I W a s s e r and Wasser APPENDIX. Details on the Neonatal Assessment in the Field ~ The neonatal assessment was completed at the end of the day after a t least two follows were conducted on the mother-infant pair. Each of the following categories was given a numerical score of either 1-3, 4, or 5 depending on the measure. Size of infant: small, average, large Skin: bright pink, pink, darkening Face wrinkles: many, some, none Body wrinkles: many, some, none Coat condition: wet, sticky, flaking, fluffy Face and head: face bruised, head flattened, normal Grasp: observed the infant’s ability to hold on to the mother’s coat with its hand and feet while the mother is walking.Scored as none if the infant did not grip a t all (often with the feet in very young infants); scored a s weak if the hand or foot did not grasp continuously and new holds were established repeatedly; scored as strong if the grasp appeared continuously (palmar = hands, plantar = feet). Clasp: observed the amount of space between the infant’s ventral surface and the mother’s ventrum. Scored clasp a s normal if surfaces touched or were very close. Scored as weak if 1-2 inches of space were observable between the two animals’ ventral surfaces. Scored a s absent if more than 2 inches of space are observable in this position. Note that if there was no grasp reflex scored, then there was no clasp reflex scored for that set of limbs (arm or legs) Vision: eyes closed, squinted, fixated stare, slow visual follow, full visual follow Neck tone: flacid, tremor, toned Rooting: none, weak, strong Sucking: none, weak, strong Locomotory postures: scored for arms, legs, and back as either flexed or extended Balance: observed while infant was standing alone, scored as shaky or steady Maternal support: observed the mother locomoting with the infant. Noted if the mother used one arm to support the infant. Scored as: always, sometimes, never.