Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA FROM COSTA RICA AND PANAMA Paul G. Heltne, Dennis C. Turner, and Norman J. Scott, Jr. INTRODUCTION This chapter evaluates the status and trends of two populations of the mantled howler monkey (Alouatta palliata following the usage of Smith, 1970; A. villosa of Hall and Kelson, 1959). Between 1966 and 1971, graduate field teams of the Organization for Tropical Studies (OTS) conducted 12 censuses of howler mon- keys in the tropical dry forest of northwestern Costa Rica. The Costa Rican data are compared to the censuses of the same species of howler living on Barro Colorado Island (BCI), Panama, a nature reserve less harshly seasonal than the Costa Rican sites and with no serious alterations since the completion of the Panama Canal. REVIEW OF PREVIOUS CRITERIA This section will review previous criteria and suggest a new indicator of the stability of a troop or population. Several statistics have been used to signal whether a population is in a depressed, stable, or increasing condition. When an exhaustive census of an isolated area is conducted, the total number of monkeys is obviously a key factor to compare with former tallies. Collias and Southwick (1952) counted a drastically reduced howler population following a yellow fever epidemic on BCI, and statistics derived from their census are regarded as signs of a depressed popula- tion; namely, low average troop size and an increased number of females per male. They reported an average troop size of 7.9 for 1951 (Table 3); later troop size averages were more than twice this figure (Carpenter, 1964). The adult male: adult female (M:F) ratio was 1:3.75. The Collias and Southwick study also suggested that a low proportion of females with infants is indicative of a population in difficulty. All of these criteria are further advanced by Carpenter (1964). In addition, the ratio of adult females to the sum of juveniles plus infants (F:J + I) may be a useful mea- sure of whether a troop or population is maintaining, increasing, or decreasing its numbers. Tables 1 and 2 reveal the reason for attention to the F:J + I ratio. The illustrations presented in these two tables are based on assigning values to certain life history variables, namely: (a) number of safe reproductive years for the adult female, (b) number of young per year per female (all the young born in one year compose a cohort), (c) sex ratio at birth, (d) mortality before completion of breeding potential in each birth cohort, (e) relative effect of mortality on the sexes, (f) times at which mortality is assessed, and (g) number of years to maturity. In Part I of Table 1, the values assigned to the above categories are as follows: (a) the adult female has 4 safe reproductive years, (b) one infant is born per year per female, (c) the sex ratio at birth is 1:1, (d) there is a 50 percent preadult mortality of each new cohort, (e) both sexes are affected equally by the mortality, (f) the mortality is assessed at birth, and (g) 3 years intervene between birth and maturity. These particular assump- tions will subsequently be referred to as the "7-year model." Any troop or population with such charac- teristics should show an F:J + I ratio of around 1:1.5 if 10
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA 11 it is to remain stable. Stability, here, is defined as maintaining a steady supply of adult females. Females with dependent infants (those in the 0-1 age group) would comprise about 50 percent of the total adult female population; the F:I ratio would be about 1:0.50. TABLE 1 Models for Stable or Increasing Populations Models for 4 years of active reproduction for females reaching maturity after 3 preadult years; 1 infant/year/female; 1:1 sex ratio among infants; no sex differential in preadult mortality, t indicates occurrence of a death. Part Iâ7-Year Model Number of females is constant; local population size constant; all preadult mortality assessed at birth. Age in Years I J J F F F F F 0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 Part IIIâMaximum Rate of Growth Increasing number of adult females; local population growing or emigrating; no preadult mortality. Year 1 Mc M, F. F, P Year 2 M F M* K M0 F0 M, F3 M, F2 F, F:J F, F5 F, F7t F3 F, F5 Year 1 M,, M, M2 F3 F, F, Ffi F7f F. F, F2 M,, M,' F.' Year 2 M M0 M, F2 F3 F, F3 F/ F Fo F, M M0t F* F:J + I= 1:1.75 Age in Years I J J F F F F F 0-1 1-2 2-3 3^Â» 4-5 5-6 6-7 7-* I J J F F F F F M M M F F F F Ft F F F M M M F F F F:J + 1 = 1:3.0 1:1.5 Average troop size given four adult females per troop: 21 + 4J + 4F + 1M (or more) = 11 (or more). Part IIâAlternate 7-Year Model Number of females is constant; local population size constant; allo- cation of preadult mortality extended over 2 years. Age in Years I J J F F F F F 0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 Average troop size: 31 + 4J + 4F + IM (or more) = 12 (or more). Average troop size: 41 + 8J + 4F + IM (or more) = 17 (or more) The average troop size generated by the "7-year model" would be (21 + 4J + 4F + 1 or more M) at least 11. If, after averaging data from a number of troops, a population with the above characteristics had an F:J + I ratio of less than 1:1.5, it would be presumed it is under some stress, preventing reproductive success. A troop or population showing such a ratio over a number of years would probably show other evidence of decline (in the absence of immigration). The F:J + I ratio summarizes the status or condition of the whole subadult population, not just one small portion of that population. The F:J + I ratio contains an extended time parameter and may signal a trend in population. Several variations are shown. Part II of Table 1 demonstrates the effect of delaying some of the mortal- ity, i.e., permitting an extra infant to survive during age 0-1. In such a case, more subadults would be required to indicate a stable population. A third possi- bility is an increasing population with a stable number of adult females in a troop. This is illustrated in Part III of Table 1. Here all the infants and juveniles remain alive and half emigrate to form new troops on reaching maturity. The F:J + I ratio becomes 1:3, which is the maximum obtainable under the conditions assigned. A second series of illustrations (Table 2) reveals the usefulness of simulation with such models. The adult female is (a) allowed 8 safe reproductive years and (b) bears one infant every other year, the "A" females reproducing one year and the "B" females the next. All the other values remain as in Part I of Table 1. This model will be known as the "11-year model." A stable population with such characteristics has an F:J + I ratio of 1:0.75 on the average. An F:I ratio of 1:0.25 is generated along with an average troop size of at least 15 (21 + 4J + 8F + 1 or more M). At the maximum rate of growth (no subadult mortality), such a popula-
12 HELTNE, TURNER, and SCOTT TABLE 2 Models for Stable or Increasing Populations Models for 8 years of active reproduction for females reaching maturity after 3 preadult years; 1 infant/2 years/female; 1:1 sex ratio among infants; no sex differential in preadult mortality, t indicates occurrence of a death. Part 1â11-Year Model Number of females is constant; local population size constant; all preadult mortality assessed at birth. Age in Years I J J F F F F F F F F F 0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 M M M A B A B A B A B A* F F F M P F:J = 1:0.75 Average troop size given eight adult females per troop: 21 + 4J + 8F + 1M (or more) = 15 (or more). Part IIâMaximum Rate of Growth Increasing number of adult females: local population growing or emigrating: no preadult mortality. M M MABABABA B A* F F F M M M F F F F:J + I= 1:1.5 Average troop size: 41 + 8J + 8F + 1M (or more) = 21 (or more). tion or troop would have an F:J + I ratio of 1:1.5 (Part II of Table 2). Until computer simulation is performed, we are unable to estimate confidence intervals around single observations of troops or populations. Aberrant fig- ures are generated when sudden drastic gains or losses occur in either portion of the F:J + I ratio, but these aberrations can be recognized by recensusing. A series of censuses will reveal when gains occur (as during immigration) and unusual drops in a particular age category. Therefore, we believe that the F:J + I ratio can be used, together with other indicators, to provide useful estimates of the status of a population. Knowledge of howler biology, however, is spotty concerning the values that might actually be assigned to the categories outlined above. The maximum pub- lished longevity record for a captive howler is 3 years, 9 months (Napier and Napier, 1967). The current longevity record of 14 years (or more) is held by a male kept in captivity on BCI. His female consort is little more than half his age. Thus the age spans of 7 and 11 years used in the above models are perhaps within the range of typical life expectancies for captive howlers. Carpenter (1964) estimates at least 3 years as the duration of infant and juvenile periods of life for A. palliata. There is no reason to assume a sex ratio at birth different from 1:1 or unequal postnatal mortality (except perhaps for males as they are about to attain sexual maturity). It seems reasonable to conclude that a conservative estimate of the F:J + I ratio for a stable population of howlers would be in the vicinity of 1:1. It could be no less than 1:0.75 and might be as high as 1:1.5. Reduction of female life span, increased length of juvenile-infant period, and postponement of mortal- ity all serve to increase the J + I portion of the ratio necessary to maintain a stable population. Southwick and Siddiqi (1968) present data for 41 groups of urban rhesus monkeys (Macaca mulatta) in India. Adult rhesus females are thought to live well beyond 11 years and reproduce each year. Neverthe- less, the average proportions are 4.0 M. 9.1 F, 7.1 J and 7.11, yielding an F:J -1- I ratio of 1:1.56 and an F:I ratio of 1:0.78. Southwick and Cadigan (1972), study- ing habituated longtailed macaques (Macaca fas- cicularis) in Penang, Kuala Lumpur, and Cape Rachado, Malaysia, and in Singapore, report an aver-
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA 13 age troop size of 24 composed as follows: 3.4 M, 8.4 F, 7.8 J, and 4.2 I. The F:J + I ratio is 1:1.43 and the F:I ratio is 1:0.50. STUDY SITES AND PROCEDURES The OTS class projects on the mantled howler monkey (Alouatta pallia ta) were conducted primarily on Finca Taboga, a research facility of the Costa Rican Depart- ment of Agriculture. A second field site was Finca La Pacifica, a private farm owned by Ing. Werner Hag- nauer. Finca Taboga is located 11 km southwest of Canas, Guanacaste Province, Costa Rica. Finca La Pacifica is about 7 km northwest of Canas. Both sites are located in the tropical dry forest region of north- western Costa Rica. On Finca La Pacifica. trees occur in windbreaks between fields, in galleries along rivers, and in larger stands on hillsides too rocky for profitable farming. Finca Taboga offers large patches of continu- ous forest on steep hillsides and in moist alluvial lowlands. During the dry season (January-June), the leaves fall from all but a few species of trees. Riparian galleries and swampy lowland stands are exceptions and are evergreen or only semideciduous throughout the year. The majority of the observations reported were made in the riparian gallery along the Rio Higueron at Taboga. In any given field course, the observations were usually made on a single day over a period of 3-5 hours. The troops, sites, and number of hectares sampled were not identical from year to year. This accounts for some of the variation in total number tallied during each project and places limitations on direct comparisons between the raw figures from the censuses. Immigration from surrounding cutover re- gions into the study area may also have occurred since figures from one census are sometimes insufficient to account for the numbers of juveniles and females seen 3 months later or during the next year. In each project, attention was directed toward troop size and the age and sex composition of each troop. The categories of adult male, adult female with or without infant, juvenile, and infant were generally used or could be determined from the data, which included additional categories of subadult (male, female, or unknown sex), dependent infant, indepen- dent infant, and unsexed adults. In this paper, the definition of infant is limited to dependent infant, the Infant 1 or Infant 2 defined by Carpenter in 1934 (Carpenter, 1964). The definition of juvenile includes all nondependent subadults. How- ever, old juveniles and young females may be inter- changeably misclassified (Carpenter, 1964). We can find no reason to believe that, overall, more females are misclassified as juveniles than juveniles misclas- sified as females. The 1967 and July 1968 census totals allocate "un- sexed adults" (10 and 5, respectively) according to the proportion of adults of known sex. In 1967, one troop of 28 individuals was simply counted. These individu- als were divided proportionately into all four categories. "Unsexed adults and subadults" (18 ani- mals) in the 1966 census were allocated according to the proportion of definitely determined adults and juveniles, and the adults apportioned to sexes as above. Nine "unclassified" individuals from February 1968 were divided between the known females and the juveniles. An isolated pair of juveniles in the February 1968 census were not counted as a troop nor included in the juvenile total. Lone males, encountered in several censuses, were neither counted as troops nor added to the male sum. RESULTS AND COMPARISONS The first portions of Table 3 present the results of nine successive censuses by OTS classes on Taboga and three censuses at La Pacifica. Taboga and La Pacifica average troop sizes are small compared to most BCI census averages except for the crash population cen- sused by Collias and Southwick (1952). The range and standard deviation of the troop size are similar in the three areas (Table 4). The most depressed censuses (BCI, 1951, and Taboga, July 1968) show reduced range and standard deviation of troop size, though later BCI figures (1967 and 1972) are also low in these respects. The average troop size at Taboga shows a declining trend. The correlation coefficient, r, between years elapsed (1966 = 0) and average troop size is r = -0.61 (the null hypothesis has a probability, P, less than 0.09, with 7 degrees of freedom, d.f.). The actual rate of decline in average troop size is -0.68 individuals per year. If the low average of 8.9 from July 1968 is rejected as showing severe influences not part of the general trend (see below), the correlation coefficient becomes r = -0.76 (P <0.05, 6 d.f.), with an actual rate of decline of -0.75 individuals per year. The 1932, 1933, 1935, and 1972 censuses from BCI are very similar to each other in total proportions (M:F:J:I), but differ considerably from those de- veloped from the Taboga counts (Table 3). The 1951, 1959, and 1967 BCI censuses yield proportions that overlap Taboga values from some years. The censuses most similar in average troop size are Taboga (1966) and BCI (1972). However, the proportions of females, juveniles, and infants in these two censuses are mark- edly different. A chi-square test indicates that there is a very low probability that the two samples might have
14 HELTNE, TURNER, and SCOTT TABLE 3 Population Parameters of Howler Monkey Troops at Finca Taboga and Finca La Pacifica, Guanacaste Province, Costa Rica, and Barro Colorado Island (BCI), Canal Zone, Panama Location Average Troop Size Number of Troops Adult Males (M) Adult Females Juveniles (F) (J) Infants (I) Total (%) Date of Study (Reference) Total M F J I Taboga Feb. 1 966 (this study) 15.4 7 108 18 54 33 3 16 50 31 3 Feb. 1967 (this study) 13.5 10 135 23 64 38 10 17 48 28 7 Feb. 1968 (this study) 10.0 8 80 13 38 22 7 16 47 28 9 July 1968 (this study) 8.9 8 71 18 30 17 6 25 42 24 8 Feb. 1969 (this study) 13.1 17 223 52 111 37 23 23 50 17 10 Feb. 1970 (this study) 10.4 7 73 22 37 8 6 30 51 11 8 July 1970 (this study) 10.8 8 86 18 52 5 11 21 60 6 13 April 1971 (this study) 11.3 22 248 49 107 65 27 20 43 26 11 June 1971 (this study) 9.9 15 148 33 64 21 30 22 43 14 20 AVERAGES 11.5 â â â â â â 21 48 21 10 La Pacifica Feb. 1967 (this stud.y) 13.2 5 66 9 37 15 5 14 56 23 8 Feb. 1969 (this study) 10.0 6 60 19 22 10 9 32 36 17 15 Feb. 1970 (this study) 12.5 4 5O 8 24 10 8 16 48 20 16 AVERAGES 11.9 â â â â â â 20 47 20 13 BCI 1932 (Carpenter, 1964) 17.3 23 398 63 171 92 72 16 43 23 18 1933 (Carpenter, 1964) 17.5 28 489 82 192 117 98 17 39 24 20 1935 (Carpenter, 1964) 18.2 15 273 49 105 81 38 18 38 30 14 1951 (Collias and Southwick, 1952) 7.9 30 239 36 135 32 36 15 57 13 15 1959 (Carpenter, 1964) 18.5 44 814 146 402 135 131 18 49 17 16 1967 (Chivers, 1969) 14.7 12 176 40 72 34 30 23 41 19 17 1972 (Thorington et al., unpubl.) 15.2 12 182 33 66 40 43 18 36 22 24 been drawn by chance from the same population (chi-square = 23.64, P <0.001, 2 d.f.) It must be noted that the infant and juvenile classes were defined differently in the two studies. Nevertheless, if the two classes, juvenile and infant, are summed together, the composite category covers the same age range in the BCI studies as in the Taboga observations. If the summation of J + I is carried out and the test is performed again on the same two groups (Taboga, 1966, and BCI, 1972), the results yield a chi- square = 5.64 with a probability of less than 0.02 with 1 degree of freedom. Comparison of the 1972 BCI data with the Taboga census next most similar in average troop size (February 1967) gives a chi-square = 4.0 (P <.05; 1 d.f.), even if the juvenile and infant categories are lumped. The values of the ratio F:J + I are presented in Table 4. Values of 1:0.75 and 1:1.5 are taken as minimal indicators of stable populations for the 11-year model and 7-year model, respectively. Even with the 11- year model, we would be more confident about the future of a population with an F:J + I ratio of 1:1, which would allow for some delay in preadult mortal- ity. At Taboga, the ratios skirt the minimal replace- ment value for the 11-year model (except for 1969 and 1970) and never approach that for the 7-year model. The three La Pacifica readings suggest the possibility of a more stable situation than occurs on Taboga. This agrees with our evaluation of the relative degree and rate of environmental alteration at the two sites. In contrast to the Taboga ratios, the F:J + I values from BCI vary around 1:1, except for the 1951 and 1959 values. The 1959 figures may simply indicate a down- ward phase in an oscillation superimposed on what appears to be a continuing upward trend in population size for howlers on BCI.Except for the depression year of 1951, less than 5 percent of the troops censused on BCI are without juveniles or infants. At Taboga, 10-50 percent of the troops counted had no infants or juveniles in six of the nine censuses. Of those five cases at Taboga in which the F:J + I ratio does equal or exceed 1:0.75, females as well as juveniles and infants appear to be experiencing a reduction of num- bers in February and July 1968 and possibly also in
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA 15 TABLE 4 Standard Deviation and Range of Troop Size, and Indicator Ratios Location Number of Mean Troop Range of Troop F:J + I Number of Troops without F:I Number of Troops without M:F Number of One Male Troops Date of Study (Reference) Troops Size Â± SD Sizes (15â) J or I (%) (1:â) Dependent I (%) (1:â) (%) Taboga Feb. 1966 (this study) 7 15.4 Â±12.6 3-39 0.67 0 0.06 5(71) 3.00 2(29) Feb. 1967 (this study) 10 13.5 Â± 7.9 5-28 0.75 1 (10) 0.16 5(50) 2.78 3(30) Feb. 1968 (this study) 8 10.0 Â± 6.2 2-21 0.76 1 (12) 0.18 4(50) 2.92 4(50) July 1968 (this study) 8 8.9 Â± 2.5 5-11 0.77 2(25) 0.20 5(62) 1.88 1(12) Feb. 1969 (this study) 17 13.1Â±NR" NR" 0.54 NR" 0.21 NR" 2.13 NR" Feb. 1970 (this study) 7 10.4 Â± 5.3 5-21 0.38 1 (14) 0.16 2(29) 1.68 1(14) July 1970 (this study) 8 10.8 Â± 8.2 2-29 0.31 4(50) 0.21 5(62) 2.89 2(25) April 1971 (this study) :: 11.3 Â± 5.6 3-26 0.86 3(14) 0.25 5(23) 2.18 8(36) June 1971 (this study) 15 9.9 Â± 4.1 5-19 0.80 0 0.47 1(7) 1.94 4(27) AVERAGES â 11.5Â± 2.1 â 0.65 â 0.21 â 2.38 â La Pacifica Feb 1967 (this study) 5 13.2 Â± 9.1 5-27 0.54 2(40) 0.14 3(60) 4.11 2(40) Feb. 1969 (this study) 6 10.0 Â±NR" NR" 0.86 NR" 0.41 NR" 1.16 NR" Feb. 1970 (this study) 4 12.5 Â± 8.7 5-25 0.75 0 0.33 0 3.00 2(50) AVERAGES â 11.9Â± 1.7 â 0.72 â 0.29 â 2.76 â BC1 1932 (Carpenter, 1964) 23 17.3 7.1 4-35 0.9 6 1 (4) 0.42 2(9) 2.71 4(17) 1933 (Carpenter, 1964) 28 17.5 7.0 4-29 1.12 0 0.51 3(11) 2.34 3(11) 1935 (Carpenter, 1964) 15 18.2 7.1 6-34 1.13 0 0.36 0 2.33 2(13) 1951 (Collias and Southwick. 1952) 30 8.0 3.6 2-17 0.50 4(13) 0.27 14 (47) 3.75 24(80) 1959 (Carpenter, 1964) 44 18.5 9.4 3-45 0.66 2(5) 0.33 8(18) 2.75 6(14) 1967 (Chivers, 1969) 12 14.7 Â± 2.5 11-18 0.89 0 0.42 0 1.80 0 1972 (Thorington et al., unpubl.) 12 15.2 Â± 4.2 8-23 1.26 0 0.65 0 2.00 0 AVERAGES â 15.6 Â± 3.7 â 0.93 â 0.42 â 2.53 â "NR = data not recoverable. June 1971. In these situations, the F:J + I ratio stays at or near the level held prior to the beginning of the generalized decline in population, but low values of the ratio follow in later years (see Table 4, Taboga, 1969 and 1970). When the Taboga census figures are compared from year to year (Table 3), the low figures of the 1968 OTS observations are striking. A check with health au- thorities in Guanacaste Province revealed no epidemics or even minor outbreaks of yellow fever or other diseases over the 1968-69 period. Inquiries led to a veterinarian who had been assigned in 1968-69 to Liberia, Costa Rica, 48 miles northwest of Cafias. Several dead howlers that had been brought to him for autopsy showed no evidence of bacterial or viral disease, but he noted that "the gastric, stomach as- sociated glands were swollen, puffy and inflamed." Government laboratories in San Jose reported that the glands were loaded with insecticides, variety not specified. The veterinarian confirmed that 1968 and 1969 were heavy cotton years in the whole area, with large doses of insecticides used to protect the crops. The insecticides were often sprayed from airplanes, increasing the possibility of winds carrying the poisons from the fields to adjacent forested areas. In February 1968, the first categories to decrease were the adult males, adult females, and juveniles. By July 1968, the adult females and juveniles showed a further decline from the 1967 figures. The drop in the juvenile category may have been due to the low number of infants in previous years or to insecticide mortality. Censuses subsequent to 1968 appear to indicate some resurgence, but average troop size never reached the 1967 level (Table 3). Most categories show rises in raw figures in 1969, even though dead monkeys were found at Taboga. It is probable that both an increased area of census and immigration into the census area accounted for the increased total count in
16 HELTNE, TURNER, and SCOTT 1969. Nevertheless, the proportion of juveniles con- tinued to drop, reaching a nadir in July 1970, despite the high number of infants in 1969 both actually and proportionally. During these years, the number of females began to climb again and was accompanied by a larger infant population. With the increase in the number of infants in 1970, the number of juveniles jumped in 1971. However, the female population fell proportionally, perhaps due to the relatively high number of juveniles currently and the low number of juveniles available during the previous years. The adult females of June 1971 appeared to be very active reproductively. If indeed the ratio F:J + I is important, a low proportion of J + I over a period of years should show up as a decrease in the female category and then again in the J -I- I categories because of reduced numbers of females. The Taboga figures (Tables 3 and 4) appear to illustrate such a situation. Carpenter suggests that a low percentage of females with infants may signal a population in difficulty. Carpenter's (1964) figures show that it is usual to have from one-third to one-half or more of the adult females associated with infants. In fact, even during the census of Collias and Southwick shortly after yellow fever apparently decimated the BCI howler population, the F:I ratio dropped only slightly lower than the more usual values. Except for 1971, less than a quarter of the females at Taboga were carrying infants. The 11-year model predicts that an average of 25 percent of the adult females should be with dependent infants, and the 7-year model predicts 50 percent. Up to 70 percent of the troops censused at Taboga showed no dependent infants. In 1967 at La Pacifica, three of the five troops (60 percent) contained no dependent in- fants and two of these contained no juveniles. On BCI only very small troops (e.g., 1 M, 2 F) had no young (Carpenter, 1964). At Taboga troops with 4, 5, and even 9 adult females yielded counts with no infants (OTS files, San Jose, Costa Rica). Carpenter (1964) suggests that a high proportion of troops with only one male and elevated numbers of adult females per adult male (low M:F ratio) might be characteristic of a recent dramatic decrease in popula- tion. These criteria derive from the Collias and South- wick census of BCI howlers in 1951. Combined with an M:F ratio of 1:3.75, 80 percent of their troops showed only one adult male (Table 4). Other BCI counts yield no more than 2.75 adult females for each adult male, with less than a quarter of the troops having only one adult male. The M:F ratio is not markedly nor consist- ently depressed at Taboga (Table 4), and the July 1968 value differs greatly from that of Collias and South- wick. La Pacifica censuses yield both the highest and lowest ratios reported to this date. The troops that have only one male are more frequent at Taboga and La Pacifica than on BCI (except for the 1951 depres- sion). Again the lowest proportion of troops with only one male was found in July 1968; but this census and the April 1971 count both yielded a troop in which no males were observed (OTS files). Two La Pacifica troops were censused in both 1969 and 1970 (Table 5). La Pacifica II showed an F:J + I ratio changing from 1:0.80 to 1:0.33 and became a single male troop. La Pacifica III showed an opposite change in the F:J + I ratio, but the value attained just equals 1:0.75. The ratio of adult males to adult females changed from 1:2.8 to 1:3 in La Pacifica III and from 1:2.5 to 1:6 in La Pacifica II. Sixty percent or more of the adult females were without dependent young in both years. In the dry-season censuses of 1969 and 1970, the troop counts from supposedly less-favorable areas were separated from counts made in more-favorable habitats (OTS files). The "less-favorable" habitats in 1969 were dry, deciduous, hillside forests with a minimum of riparian trees, where an average troop size was 9 compared to 15.4 in the "more-favorable" moist forest on the alluvial lowlands. By contrast, in the riparian-dry hillside forest, 6 of 23 adult females (26 percent) were with infants and F:J + I was 23:21 (1:0.91). In the moist forests only 17 of 88 females (19 percent) were with infants, and the F:J + I ratio was 88:39 (1:0.44). The M:F ratio was between 1:2 and 1:2.5 in both habitats. The La Pacifica counts from 1969, from riparian strips and strips of dry forest remaining between cultivated fields, were similar to the riparian-dry hillside data from Taboga. The poorer statistics from the "more-favorable" Taboga lowland forests are probably related to the rapidly progressing clearing in that area. The February 1970 census considered differences between dry areas and wet areas (including riparian forests). From the Taboga area, only one dry hillside troop was studied. It yielded 4 M, 7 F, and only 1 I, suggesting harsh pressures preventing successful re- TABLE 5 Organization for Tropical Studies Com- parison of the Same Troop Over Two Years Troop Year Total M F J I La Pacifica 11 1969 11 2 5 2 2 1970 9 1 6 I 1 La Pacifica III 1969 27 5 14 4 4 1970 25 4 12 5 4
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA 17 production. The two dry La Pacffica areas yielded 5 M. 18 F. 6 J, and 5 1. The two counts from the "wet areas" of La Pacifica were, by contrast, 3 M, 6 F, 4 J, and 3 I. The La Pacifica figures were markedly better than the Taboga values and the "wet" areas better than the "dry," even at Taboga, though one of the Taboga "wet area" troops was composed of 4 M, 5 F, and no J or I (OTS files). The differences between Taboga and La Pacifica probably reflect the greater intensity of ongoing environmental disturbance at Taboga. CONCLUSIONS The censuses of howler monkeys at two sites in the tropical dry forests of Costa Rica present a picture of a distressed and declining population. In some respects, the howlers of Taboga and La Pacffica are even more seriously depressed than the 1951 population on BCI, which had just been decimated by yellow fever. In other characteristics, the howler population of Guanacaste Province is similar to that stressed popula- tion or intermediate between it and the subsequently censused BCI population. Of particular concern are the relatively low numbers of juveniles and infants. Theoretically, a female howler should be impregnated shortly after she becomes physiologically able to con- ceive (either after rearing an infant into the juvenile period or losing an infant prior to that time). Without ruling out the possibility of a birth season, we deduce that births and impregnations may occur at any time of year. This agrees with the conclusion of Carpenter (1964) from his BCI studies. However, the low numbers of infants relative to adult females at Taboga and La Pacifica suggest that fertilization or gestation are often unsuccessful or that early infancy is a period of heavy mortality. Even if an infant survives through its period of dependency, it apparently has a poor chance of reaching adulthood in most years. It is of interest to calculate, given the suggested models, how many infants and juveniles would be expected in the various years. For instance, in 1966 the 7-year model would call for at least 27 infants as compared to the 3 observed and 54 juveniles as com- pared to the 33 observed. The 11-year model suggests 14 infants and 27 juveniles. These calculations allot all preadult mortality at birth. If the assessment of preadult mortality were extended in a more natural fashion, more than 81 young animals would be ex- pected given the 7-year model. Under the 11-year model, more than 41 preadults would be expected, producing an F:J + I ratio more like that found on BCI and among macaques (Southwick and Cadigan, 1972; Southwick and Siddiqi, 1968). If the F:J + I composi- tion of undisturbed stable howler populations is in the realm of those generated by the 7- and 11-year models, then it is clear that the Taboga population is in continuing difficulty. Comparisons between the two Costa Rican sites and among habitats within sites present suggestive differ- ences. The moist alluvial forests in the Taboga low- lands would appear to be very favorable habitats for howlers, while the riparian strips and windbreaks at La Pacifica appear to be only marginal. However, La Pacifica supports populations that seem to be more stable than at Taboga. The key may be that the remaining forested areas on La Pacifica are not being altered further by human intervention, while the pro- gressive cutting of the Taboga forest causes or necessi- tates reduction of the population size. Deforestation may produce many pressures on the howler population. The continual cutting progressively reduces the total resource area and probably leaves some habitable areas without monkeys, because the howlers are unable to reach these patches by tree pathways. Emigration from cutover areas may pro- duce a transiently high population density in adjacent forest. This high density could be misleading unless one knows the history of the local forest and the numbers in the various age-sex classes of the popula- tion over the previous several years (density estimates are not available for the Costa Rican sites). Chivers (1969) suggests that there is an adjustment mechanism available to howler monkeys when their population density increases but does not yet actually exceed the capacity of their environment. He hypothesizes that under the influence of social pressures (mainly the avoidance of aggressive interactions), the number of troops becomes slightly more numerous and troop size and range diminish slightly. Yet, Chivers' hypothetical population would be increasing, which means that status indicators such as the F:J + I ratio would remain well above minimal replacement levels. Social factors may be operative in population read- justment at Taboga, where howlers were seen fighting, and fresh wounds and scars were observed (OTS files, 1971). However, the indicators are often so depressed that additional advantages or pressures must dictate small troop size. For example, the folivorous howlers normally graze very lightly in any one tree. Given this norm, we suggest that smaller groups can forage more efficiently. With increased troop size, more trees must be visited and more time and energy spent in obtaining food. This leads to decreased feeding efficiency, which eventually may limit troop size in any environment. The limitations may come into effect earlier in an altered setting, where several troops may be forced to visit the same food sources. Where the arrangement of
18 HELTNE, TURNER, and SCOTT trees is in strips, as in the windbreaks and narrow riparian stands at La Pacific;;, foraging is mostly linear. Efficiency of foraging in this restricted topog- raphy may also call for a lower average troop size than would be found in a continuously forested area permit- ting travel in any direction. Part of the precarious condition of howlers in this study is almost certainly due to disturbance of the habitat by deforestation. However, even in the ab- sence of human disruptions, it is possible that during most years the Costa Rican tropical dry forest is a marginal habitat for the howler monkey. In a series of "normally bad" years, the population total and the average troop size may decrease steadily, and the signal ratios may indicate reproductive activity slightly above or below maintenance levels. It is easy to visualize a continually decreasing average troop size with F:J + I ratios considerably more favorable than those that we have designated as minimal replacement values. In a marginal environment, a population may depend on near-maximum reproductive success in the sporadic "good" years to recoup losses. Deforesta- tion, insecticide toxicity, and epidemics of disease sharply accentuate the marginal character of the habitat. During or subsequent to such assault, the J + I portion of the F:J + I ratio may indicate drastic reduction in population size and status. Clearly, the population parameters can only di- minish to certain levels. Below these levels, the tem- poral patchiness of the environment will take on the nature of a random factor, because small troops may no longer be able to survive until a good year. The population as a whole may be unable to stabilize its numbers or prevent further decline. Perhaps the Taboga data illustrate the descending limb of a cyclic phenomenon related to marginal habitat: but if so, it is not apparent that the depth of the oscillation has yet been reached. Even very unsophisticated modeling suggests that some parameters of the Taboga popula- tion may be approaching threshold levels. In general, to make the most meaningful estimates of the status of a particular population or species, and the conditions under which it can survive, major advances are required in theoretical and field investigations. Table 6 presents a compilation (Napier and Napier. 1967) of the current knowledge regarding age at matu- rity, longevity, and the duration of gestation and lacta- tion for New World primate genera. The sparse data are almost exclusively from captive situations and include no indication of the variation that might be expected between species or between sexes. Adequate information on the life history variables distinctly depends on extended studies of local populations in which animals are individually identifiable. Scott and co-workers are beginning to collect such information by recensusing howler troops containing marked indi- viduals at La Pacifica. Freese, working throughout the year in the tropical dry forest of northwestern Costa Rica, is developing statistics on within-year variation in troops (see pp. 4-9). However, every genus must be examined and, in many genera, individual species deserve study; for among sympatric species the varia- TABLE 6 Gestation, Lactation, Maturity, and Longevity Data for New World Primates* Genus Gestation Lactation â¢ SOURCE: Napier and Napier, I967. 1 Data from personal breeding colony (Heltne. unpublished). Age to Maturity Record Longevity (yr, mo) Callithrix 140 days 6 mo 14 mo 12 yr Cebuella â 3 mo â 4 yr, 11 mo Saguinus 140- 145 days â â 9 yr, 10 mo Leontopithecus 132- 134 days 3 mo â 10 yr, 4 mo Callimico 150- 170 days" â ~1 1/2-2 yr" 4 yr, 9 mo Cebus ca. 180 days â â 40 yr Saimiri 168- 182 days â â 10-20 yr Aotus â â â 11 yr, 7 mo Callicebus â â â â Pithecia â â â 13 yr, 8 mo CMropotes â â â 15 yr Cacqjao â â â 8 yr, 9 mo Alouatta â 18-24 mo 3 1/2-4 yr 3 yr, 9 mo Atdes ca. 139 days â â 20 yr Lagothrix ca. 225 days 12 mo+ 4 yr 13 yr Brachyteles â â â
COMPARISON OF CENSUS DATA ON ALOUATTA PALLIATA 19 tions in the life history parameters and behavioral patterns strongly relate to the exquisite division of resources in the tropical forest ecosystems. Computer simulation can produce estimates of sta- ble populations given any particular model. Even at such an unsophisticated level, simulation may interact in an extremely useful fashion with field checks of the status of local populations. In order to attain their true potential, however, simulations must incorporate feedback from long-term field investigations. Only with detailed statistical information on the life history parameters of undisturbed populations can computer simulation demarcate the evolutionary strategies avail- able to primate species and the conservation strategies available to us. Deforestation and other human assaults on tropical forests are producing dire conditions for many primate populations. It is urgent that the governments of Central and South American countries, environmen- talists, conservationists, and the biomedical commu- nity and its animal suppliers express their concern in this matter in terms of increased attention to and support of the kinds of primatological studies suggested in this volume. ACKNOWLEDGMENTS We wish to express our gratitude to OTS field leaders Gordon Orians and Christopher C. Smith and to several dozen OTS student re- searchers, especially Darcy B. Kelley. We thank Richard W. Thorington, Jr. (U.S. Museum of Natural History) and Miguel A. Schon (Johns Hopkins University) for reading the manuscript and for allowing us to use unpublished data collected with the support of the Smithsonian Environmental Sciences Program. Charles South- wick, Jill Wolhandler, and Tom Sayvetz (all of Johns Hopkins University) ably criticized drafts of this paper. For the information on possible insecticide mortality, we are indebted to Dr. Roman Miguel, Veterinario, Centre Agricola Veterinario Regional de Liberia, Costa Rica, currently head of the rabies project, Zoonosis Department, Ministerio Salubridad Piiblica, Republica de Costa Rica. The fieldwork was supported by National Science Foundation grants to the educational program of the Organization for Tropical Studies. REFERENCES Carpenter, C. R. 1964. Naturalistic behavior of nonhuman pri- mates. The Pennsylvania State University Press, University Park, x + 454 pp. Chivers, D. J. 1969. On the daily behavior and spacing of howling monkey groups. Folia Primatol. 10:48-102. Collias, N. E., and C. Southwick. 1952. A field study of population density and social organization in howling monkeys. Proc. Am. Philos. Soc. 96:143-156. Hall, E. R., and K. R. Kelson. 1959. The mammals of North America, vol. 1. The Ronald Press Company, New York. Napier, J. R., and P. H. Napier. 1967. A handbook of living primates. Academic Press, New York, xiv + 456 pp. Smith, J. D. 1970. The systematic status of the black howler monkey, Alouatta pigra Lawrence. J. Mammal. 51:358-369. Southwick, C. H., and F. C. Cadigan, Jr. 1972. Population studies of Malaysian primates. Primates 13:1-18. Southwick, C. H., and M. R. Siddiqi. 1968. Population trends of rhesus monkeys in villages and towns of northern India, 1959- 1965. J. Anim. Ecol. 37:199-204.