AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 103:69–83 (1997)

Stable Isotope Ratios Indicate Diet and Habitat Use in New World Monkeys

MARGARET J. SCHOENINGER,1* URSZULA T. IWANIEC,1 AND KENNETH E. GLANDER2 1Department of Anthropology, University of Wisconsin, Madison, Wisconsin 53706 2Duke University Primate Center, Durham, North Carolina 27705-5000 KEY WORDS carbon; nitrogen; conservation; Cebus; Ateles; Brachyteles; Alouatta; Costa Rica; Brazil

ABSTRACT This paper demonstrates the use of stable isotope ratios of carbon and nitrogen in animal tissues for indicating aspects of species behavioral strategy. We analyzed hair from individuals representing four species of New World monkeys (Alouatta palliata, the mantled howler; Ateles geoffroyi, the spider monkey; Cebus capucinus, the capuchin; and Brachyteles arachnoides, the woolly-spider monkey or muriqui) for d13C and d15N using previously developed methods. There are no significant differences in either carbon or nitrogen ratios between sexes, sampling year, or year of analysis. Seasonal differences in d13C reached a low level of significance but do not affect general patterns. Variation within species was similar to that recorded previously within single individuals. The d13C data show a bimodal distribu- tion with significant difference between the means. The two monkey popula- tions living in an evergreen forest were similar to each other and different from the other two monkey populations that inhabited dry, deciduous forests. This bimodal distribution is independent of any particular species’ diet and reflects the level of leaf cover in the two types of forest. The d15N data display three significantly different modes. The omnivorous capuchins were most positive reflecting a trophic level offset. The spider monkeys and the muriquis were similar to one another and significantly more positive than the howlers. This distribution among totally herbivorous species correlates with the ingestion of legumes by the howler monkey population. In combination, these data indicate that museum-curated primate material can be analyzed to yield information on forest cover and diet in populations and species lacking behavioral data. Am. J. Phys. Anthropol. 103:69–83, 1997. r 1997 Wiley-Liss, Inc.

New World primates exhibit a wide vari- tial of carbon and nitrogen stable isotope ety of subsistence strategies in diverse arbo- ratios in hair for providing data on aspects of real settings. Detailed information on subsis- subsistence. The goal of these investigations tence is available for a limited number of populations of some species but extant popu- lations are remnants of those which existed Contract grant sponsor: National Science Foundation; contract prior to human population expansion and it grant numbers BNS-8819733 and DBC-9118876; Contract grant sponsor: Conservation International and Wildlife Preservation is likely that they do not display the com- Trust; Contract grant sponsor: Wisconsin Alumni Research Foun- plete range of strategies used before forest dation. *Correspondence to: Margaret J. Schoeninger, Department of distributions were altered and restricted Anthropology, University of Wisconsin, Madison, WI 53706. (compare Fedigan et al., 1985, with Lippold, E-mail: [email protected] Received 30 January 1996; revised 11 March 1997; accepted 16 1988). The present study explores the poten- March 1997. r 1997 WILEY-LISS, INC. 70 M.J. SCHOENINGER ET AL. is the establishment of an indicator for ratio in plants is different from the ratio in assessing habitat use and diet from primate the atmosphere. The patterned, predictable populations now represented by museum- change in the ratios permits their use as curated specimens to determine population monitors for the transfer of the elements patterns prior to habitat destruction. We from one part of the biosphere to another focus on extant populations of four species of (Schwarcz and Schoeninger, 1991). Central and South American monkeys for Methods vary for assessing diet, monitor- which relatively extensive information is ing habitat use, and determining preferred available on diet and habitat use, i.e., habitat choice in extant primate populations Alouatta palliata (mantled howler or howl- and within species (e.g., Crockett and Eisen- ing monkey), Ateles geoffroyi (spider mon- berg, 1987). Diet assessments are based on key), Cebus capucinus (capuchin monkey), time spent taking items or foraging, estima- and Brachyteles arachnoides (muriqui or tions of ingested item weight, and frequen- woolly-spider monkey). Two of these species cies of different items taken (Robinson and (Ateles geoffroyi and Brachyteles arachnoi- Janson, 1987). General aspects of diet are des) are highly endangered (Hearn cited in also reflected in various morphological fea- Fedigan et al., 1985; Robinson and Janson, tures. Folivores have absolutely larger and 1987). Successful application of the method relatively larger teeth than frugivores and to well-studied populations of these species insectivores (Rosenberger, 1992) and foli- supports application to extinct populations vores display larger buccal shearing crests of the same species and to extant and extinct relative to lingual shearing crests than frugi- populations of other, lesser known arboreal vores (Kay, 1975). Folivores also have larger monkey species. stomachs and large intestines compared with Although relatively recent, the use of those in frugivores and insectivores (Chiv- stable isotope ratios (13C/12C and 15N/14N) in ers and Hladik, 1980). Habitat use and ecological applications is rapidly proving choice are based on present-day distribu- extremely valuable in monitoring aspects of tions (Lemos de Sa and Strier, 1992) al- plant photosynthesis, animal diet, and cli- though, as discussed previously, these distri- mate (Rundel et al., 1989). Both carbon and butions may not completely represent the nitrogen consist of two stable forms (i.e., totality of habitats used formerly. Within isotopes). Their natural abundance ratios in habitats, there is a positive association be- the geosphere (13C/12C 5 1.1/98.9; 15N/14N 5 tween size of the home range and body size 0.36/99.64) are altered in small, but consis- among primate species although leaf eaters tent, ways during transfer from the geo- (folivores) have smaller home ranges than sphere to the biosphere (from atmosphere do fruit eaters (frugivores) or species that and soil to plants) and subsequently, in the include some animal protein (omnivores) biosphere, from plants to animals (Hoefs, (Milton and May, 1976). The four species in 1987). The alterations occur because while this study were chosen, in part, because they the isotopes of an element like carbon (or display marked differences in their geo- nitrogen) share the same chemical proper- graphic distributions, habitat tolerances, and ties, their reaction rates differ due to their diet. They have also been the subjects of mass differences (Hoefs, 1987). For example, long-term behavioral study as well as large 12C reacts faster than 13C and chemical dental, skeletal, and digestive tract morpho- bonds containing 12C break and form more logical assessments. These studies provide rapidly than those containing 13C. As a the comparative base for our analysis. result the 13C/12C ratio in the product of a The data from the various lines of evi- reaction is different from the ratio in the dence indicate marked variations in diet starting components whenever the product within individual species that are obscurred contains less carbon than that provided in in the generalities stated above. Such varia- the starting components. In other words, tion appears to associate with various as- because there is less carbon in plants (the pects of habitat (available foods and their product) than there is in atmospheric carbon distribution, predators, and competitors, dioxide (the starting material), the 13C/12C among others; Rodman and Cant, 1984). To STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 71 avoid such complicating factors in the pres- folivorous frugivore (Chivers and Hladik, ent study, hair was taken from animals 1980), and their food passage time is the whose populations, with one exception, have slowest, by far, of the four species in our been monitored in long-term studies. Data study (Milton, 1984b). In combination, these from these populations’ habitats and diets traits suggest that folivory is the major are compared with hair stable isotope data. dietary adaptation with fruit taken when There are several advantages of the stable available. They never ‘‘specialize on fruits’’ isotope method. It can be applied to museum- (Milton, 1981:503). curated specimens which usually derive from In contrast to howler monkeys, spider populations where extensive behavioral stud- monkeys (Ateles) are considered by many ies have not been done. Museum-curated ecologists to be among the most sensitive in specimens often consist of skeletal material Central America to habitat disturbance (Ei- and, frequently, pelts but aspects of soft senberg, 1983; Konstant et al., 1985). Ex- tissue morphology (such as the relative di- tant populations of Ateles geoffroyi are highly mensions within the gastrointestinal tract) selective frugivorous feeders consuming can seldom be determined. Although stable merely 20% of their diet as young leaves isotope analysis requires access to equip- with the vast majority of diet coming from ment and expertise with the analytical meth- ripe fruits, on average (data cited in Eisen- ods, it is a more direct and immediate indica- berg, 1983; Strier, 1992). This adaptation is tor of diet and habitat than dental and confirmed by their frugivorous tooth mor- skeletal morphology. This is not to suggest phology (Kay, 1975), digestive tract morphol- that isotope data should replace morphologi- ogy (Chivers and Hladik, 1980), and rapid cal or behavioral studies but rather that food assimilation rates (Milton, 1984b). Ob- these lines of investigation complement each served variation in dependence on leaves other. Finally, when the analysis is done on appears to be seasonally related with greater hair, as in the present study, no destruction inclusion in areas with more seasonal rain- of skeletal material is required. fall (data summarized in Strier, 1992). These observations indicate that frugivory is the MONKEY SPECIES preferred and long-term subsistence adapta- Mantled howlers (Alouatta palliata) live tion of Ateles. today in both dry, deciduous and evergreen Capuchins (Cebus capucinus), like howl- forests and appear to be among the least ers, inhabit both moist evergreen as well as affected by human expansion of all the Cen- dry, tropical forests and have been observed tral American nonhuman primates (Glan- foraging in secondary growth as well as der, 1983). Howler diets vary in association primary forests (Fedigan et al., 1985). The with forest composition and across seasons adaptability of Cebus in terms of habitat and (Milton, 1980). Within a dry, deciduous for- diet may contribute to its less threatened est, leaves constitute the majority of diet position relative to Ateles, a situation some- (Glander, 1983) whereas in wetter habitats, what similar to Alouatta. All four Cebus as much time may be devoted to feeding on species are omnivores with diets that in- fruit as on leaves (Crockett and Eisenberg, clude a significant weight percent of insects. 1987; Milton, 1980; Estrada, 1984; Neville Variation between the species appears to et al., 1988). There is an inverse association associate with average species’ body size and between rainfall and folivory within the other morphological distinctions (Ford and genus and a positive association between Davis, 1992; Janson and Boinski, 1992). The rainfall levels and fruit-eating within more robust Cebus apella depends to a mantled howlers in particular, although greater extent on vegetative plant parts leaves consistently constitute 50% or more than is true of the three gracile species (term of the diet (Strier, 1992). The morphological taken from Janson and Boinski, 1992). Esti- evidence suggests that folivory is a long- mates for Cebus capucinus range from 20% term adaptation of the species. Howlers insects on Barro Colorado Island, Panama, have the tooth morphology of a folivore (Kay, to 50% in the early wet season in Santa Rosa 1975), the digestive tract morphology of a National Park in Costa Rica; the rest of the 72 M.J. SCHOENINGER ET AL. diet is largely ripe fruit (data cited in Fedi- undisturbed habitats. Brachyteles live in gan et al., 1985). The various dimensions of both open and closed habitats but is more tooth morphology group capuchins with the similar to Ateles in its reactions to distur- frugivores rather than with the insectivores bance. Dietary strategies, considered as an- (Kay, 1975). Its digestive tract morphology is nual averages of seasonally variant diets, similar to other insectivorous frugivores include omnivory within Cebus, frugivory (Chivers and Hladik, 1980), and its food within Ateles, and frugivory/folivory within passage rates are similar to Ateles (Milton, Alouatta and Brachyteles. 1984b). These features, in combination, sug- The monkey populations from which hair gest that Cebus capucinus, like Ateles geof- samples were taken have different geo- froyi, is a selective frugivore but, in contrast graphic distributions and subsistence strat- to spider monkeys, capuchins consistently egies. With the exception of the muriquis, eat animal protein and so are omnivores the diets of each population are relatively rather than strict herbivores. well characterized. Data on habitats are Muriquis (woolly-spider monkey, Brachyte- available for all four populations. Further, les arachnoides) are restricted to the Atlan- these four populations have been the focus of tic coast of Brazil, where habitat destruc- a morphological study requiring capture dur- tion, high human density, and human ing which hair samples were collected. The hunting all contribute to their highly endan- combination of behavioral and morphologi- gered status (Robinson and Janson, 1987). cal data with well-documented samples for Brachyteles is more folivorous than either analysis provided a fortuitous, yet adequate, Cebus or Ateles, based on the time spent test of the expectations outlined below. feeding on particular items. Their tooth CARBON STABLE ISOTOPE RATIOS morphology shows similarities to folivorous AND EXPECTATIONS OF THE primates (Zingeser, 1973) except that PRESENT STUDY Brachyteles has molars which are relatively and absolutely smaller than Alouatta, and Plants Brachyteles shows an emphasis on lingual Plant material from the particular forests rather than buccal shear (Rosenberger, inhabited by the monkey populations was 1992). The latter traits are normally associ- not available for analysis. Even so, given the ated with frugivory. Observational data re- general nature of plant responses to pro- port a low of 28% (Strier, 1992) and a high of cesses occurring during photosynthesis, we 93% (Milton, 1984a) of foraging time consum- can assume that plant tissues from the ing leaves, although they appear to prefer relatively open canopy of dry, deciduous fruit when available. In forests where both forests have stable isotope ratios (repre- genera exist, Brachyteles is reported to be sented as d13C values1 in per mil—i.e., parts more frugivorous than Alouatta (Strier, per thousand—notation: ‰) that are signifi- 1992). In terms of food passage times, cantly different from the ratios in plant Brachyteles is more similar to Ateles and tissues from closed canopy, evergreen forests. Cebus than it is to the folivorous Alouatta

(Milton, 1984b). These traits may reflect a 1A d value is defined as ‘‘shared characteristic retained from a fru- ] givorous ateline ancestor’’ (Strier, 1992:518) d5 [Rsample [R 1] 3 1000‰ although it has also been suggested that std 2 Brachyteles is the most folivorous New World where R is the isotope ratio (e.g. 13C/12C) and the standard is the primate (Milton, 1984a). internationally recognized standard. A d value that is negative indicates a sample which is depleted in the heavier isotope In summary, the four species in this study relative to the internationally recognized standard and one that is positive indicates a sample which is enriched relative to the show variations in their present-day habitat standard. The international standard for carbon, PDB (Pee Dee distributions and diets. Alouatta and Cebus Belemnite marine carbonate), has more 13C relative to 12C than is true for the vast majority of biological samples, thus most today inhabit open, deciduous forests and biological samples have carbon d values which are negative. The more closed, semievergreen to evergreen opposite is true for nitrogen, where the majority of biological samples contain relatively more 15N than is true of the standard forests in both disturbed and undisturbed (AIR which is atmospheric nitrogen). Thus, most biological conditions whereas Ateles appears only in samples have nitrogen d values which are positive. STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 73

13 TABLE 1. Effects of air CO2,CO2concentration, and light levels on C3 plant d C values Tropical forest1 Experimental plots2 Deciduous Semievergreen Low High Variable measured main canopy main canopy irradiance irradiance

3 3 CO2 concentration in ppm 340–355 350–375 370 370 13 3 3 3 d Cair 28.3, 27.8‰ 29.9, 27.8‰ 27.9‰ 27.9‰ 13 4 4 4 d Cplant 226.9, 224.6‰ 231.1, 226.3‰ 228.0‰ 226.5‰ 1 Data from Broadmeadow et al. (1992). 2 Data from Yakir and Israeli (1995). 3 Diurnal range in values. 4 Range in leaf values.

The distribution of d13C values across all ported as low as 29.9‰ (see Table 1). These plant species is bimodal. The majority of more negative values within forests are due 12 plants (referred to as C3), including herba- to the addition of C-enriched CO2 respired ceous vegetation, trees, and some high lati- from organic detritus in soil (van der Merwe tude grasses, display d13C values with ranges and Medina, 1989; Broadmeadow et al., between 231‰ and 223‰ (O’Leary, 1988). 1992). In open areas such as savannas or In contrast, a significant number of plant deciduous bush/woodlands, therefore, the species (referred to as C4), including most CO2 available to plants is 27.8‰ whereas tropical grasses, have d13C values between that available to plants in forests varies 215‰ and 211‰ (O’Leary, 1988). The C3 according to the amount of soil-respired CO2 and C4 designations refer to the number of and the amount of mixing with atmospheric carbon atoms in the first metabolites formed CO2. This accounts, in part, for the observa- during photosynthesis. A third photosyn- tion that leaves in forests vary in accordance thetic pathway, CAM or crassulacean acid with their position within the canopy (Broad- metabolism (Kluge and Ting, 1978), is com- meadow et al., 1992 and see Table 1). This mon among succulents. The d13C values of variation is often referred to as ‘‘the canopy such plants show a range that overlaps the effect.’’ C3 and C4 distributions. There can be a greater amount of varia- All of the plants eaten by nonhuman tion in plant tissues than can be accounted primates in the forests of Central and South for by the assumption that the canopy effect 13 America follow the C3 pathway but pat- is due only to variations in air CO2 d C 13 terned differences in d C occur within C3 values. In at least one case, air CO2 varied plants (van der Merwe and Medina, 1989; by 2‰ throughout the day over the study Broadmeadow et al., 1992). These differ- period whereas the variation in plant tis- ences, obscurred when only the average for sues was 5‰ (see Table 1). This residual C3 plants is presented, are due to variations variation is due to isotope effects (D in the 13 in: 1) the d C value in the CO2 available to equation above) at particular points during the plant during photosynthesis and 2) a set photosynthesis in C3 plants (Farquhar et al., of isotope effects that occur during the pro- 1982). Of particular impact is the rate of cess of photosynthesis. In other words: diffusion of air CO2 into the plant (Loreto et al., 1992) which is indicated by the ratio of 13C 13C d plant 5d air 2D carbon dioxide concentration inside the leaf 13 (Farquhar et al., 1982) where d Cair refers (Ci) to that in the atmosphere (Ca). When Ci 13 to the d C value in the CO2 available to the equals Ca the fractionation effect of photosyn- plant during photosynthesis and D refers to thesis is at its maximum with the result that the combined set of isotope effects. leaf d13C values are at their most negative In terms of isotopic variation in air CO2, (see Marino and McElroy, 1991, for a good the d13C value in well-mixed atmospheric description of this phenomenon). Increased 13 carbon dioxide today has a d C value of concentrations of air CO2 such as occur in 27.8‰ (Wahlen, 1994) whereas the air closed canopies due to the addition of soil within semievergreen forests has been re- respired CO2 associate with more negative 74 M.J. SCHOENINGER ET AL. leaf d13C values (Broadmeadow et al., 1992) it is covered by tracts of a semideciduous, because Ca approaches Ci. Light affects the tropical dry forest. The muriqui population rate of carbon fixation and stomatal closure, is located at Fazenda Esmeralda, Brazil, which in turn increases Ci relative to Ca, which is also an area of tropical dry forest such that increased light levels associate with seasonal rainfall and a semideciduous with less negative leaf d13C values (Yakir canopy. Based on similarities with other and Israeli, 1995, and see Table 1). regions of the coastal Brazilian forest (Mil- In sum, variation in the d13C values of air ton, 1984a; Strier, 1991), we estimate that CO2 and conditions affecting diffusion into the average annual rainfall is less than the leaf influence the d13C values in leaf 1,500 mm. In contrast, the spider monkey tissue. Plants grown in closed canopies with and the capuchin populations are at La higher concentrations of soil-respired CO2 Selva on the Atlantic side of Costa Rica. This and lower light levels have more negative is an area of tropical wet forest exhibiting d13C values than those in open canopies low seasonality in rainfall with average where the air CO2 is identical isotopically to rainfall around 4,000 mm per year (Janzen, atmospheric CO2 and light levels are higher. 1983). The forest canopy is more closed than the ones at La Pacifica and Fazenda Esmer- Animals alda with evergreen and semievergreen tree Based on the patterned variation docu- stands. We expected that the hairs of mon- mented in plant tissues, we expected varia- keys feeding at La Selva (Ateles and Cebus) 13 tion among the four species in this study. It would have more negative d C values than is well established that animal tissues such those of the other two populations (Alouatta and Brachyteles). as hair and bone reflect the C3/C4 bimodality seen in plants (Vogel, 1978). Hair from ex- NITROGEN ISOTOPE RATIOS perimental animals (DeNiro and Epstein, AND EXPECTATIONS OF THE 1978; Jones et al., 1981; Tieszen et al., 1983) PRESENT STUDY and naturally occurring populations of ani- Plants mals (Minson et al., 1975) and humans (Nakamura et al., 1982; White and Schwarcz, Based on comparison with other tropical 1994) reflects monotonous C3 or C4 diets and forest data and on the general processes intermediate mixes of C3 and C4, and moni- involved in nitrogen uptake by plants, we tors changes in the diet as the hair grows. assume that plants in the forests of Costa Significant for the present study, bone colla- Rica and Brazil have isotope ratios (repre- gen from terrestrial mammals inhabiting sented as d15N values in per mil notation: ‰) East African forests (Ambrose and DeNiro, common to terrestrial plants. Such plants 1986) and an Amazonian forest (van der show a wide range overall but the majority Merwe and Medina, 1991) had d13C values of d15N values fall between 0‰ and 17‰ more negative than expected from a diet of (data compiled in Schoeninger and DeNiro, 13 C3 plant material with average d C values. 1984) with a bimodal distribution (Virginia These more negative values were attributed and Delwiche, 1982). Within single plants, to the canopy effect. We expected that hair the variation in d15N values between plant from arboreal monkeys would vary in d13Cin parts and between years is small (,1‰) association with the level of cover in their during the period of maximum leaf biomass habitats due, ultimately, to differences in (Shearer et al., 1983). The variation between 13 d C air and in isotope effects occurring dur- plant species in single locations and be- ing photosynthesis. tween locations, however, can be quite large The monkey populations in our study live (see Table 2) because the d15N value in in two distinct types of forest. The mantled plants varies with the d15N value of the howler population we sampled is located at nitrogen used by plants (Shearer and Kohl, Hacienda La Pacifica in northwestern Costa 1994). Certain plants, like legumes, grow in Rica. The area has a strongly seasonal rain- symbiotic association with bacteria, located fall pattern with average rainfall of less in nodules on their roots, that fix atmo- than 1,500 mm per year (Glander, 1979) and spheric nitrogen (N2) directly in a process STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 75 which produces no alteration in the 15N/14N TABLE 2. Variability in d15N‰ in legumes ratio from the ratio in air. When such nitro- and nonlegumes by site gen is available to the plant, the plant’s Legume Non-Legume d15N‰ 6 S.E. d15N‰ 6 S.E. tissues have a stable isotope ratio approach- Sites (N) (N) ing that in air, i.e., approaching zero (by Desert1 definition). The amount of N2 fixation by Harpers Well 1.1 6 .50 (9) 3.0 6 .62 (15) leguminous plants is variable and, thus, the Nude Wash 4.0 6 .60 (13) 5.7 6 .30 (19) d15N value of leguminous plants is variable. Clark Dry Lake 4.7 6 .70 (3) 13.0 6 .40 (4) Carrizo Badlands 2.2 6 .20 (2) 6.6 6 .60 (4) With the exception of fertilized soils and Borrego Sink 2.0 6 .20 (3) 6.9 6 .70 (7) extremely young soils (discussed below), Baja California, however, legumes usually have lower 15N Mexico d Paplionaideae 1.8 6 .60 (6) 9.3 6 .50 (13) values than other plants although there are Prosopis 8.9 6 .30 (5) exceptions (e.g., Prosopis and Acacia at Baja Acacia 10.6 6 .80 (3) Tropical forest2 California, Mexico, and Inga at Barro Colo- Barro Colorado 5.1 6 .30 (53) 6.4 6 .30 (52) rado Island, Panama, shown in Table 2). The Swartzia 5.0 6 .30 (11) two plant genera at Baja California presum- Prioria 4.0 6 .30 (10) Tachigalia 4.5 6 .70 (12) ably lack the genetic capacity to achieve Inga 7.4 6 .80 (9) symbiosis with bacteria (Shearer and Kohl, 1 Data from Shearer et al. (1983). 1994). In contrast, Inga, a woody legume, 2 Data from Shearer and Kohl (1986). reportedly shows reduced rates of N2 fixa- tion as a result of soil shading (Van Kessel and Roskoski, 1983; Shearer and Kohl, 1986). degraded plant biomass which is enriched in We expect that leguminous species from the 15N relative to precipitation and parent rock closed canopy forest in Costa Rica have d15N and, also, the soil is leached during precipita- values with a range similar to Inga at Barro tion leaving soil compounds enriched in 15N Colorado Island and that those from the (Shearer and Kohl, 1994). In some areas, open canopies in Costa Rica and Brazil have plants show highly elevated d15N values but d15N values significantly lower than the the majority from older, leached soils show nonlegume species in the same forests. values with a relatively small range of varia- Most plants lack the symbiotic association tion (3‰ to 7‰). The Central and South with bacteria observed in legumes and so American tropical forests which are the fo- are limited to using soil nitrogen. Soil nitro- cus of the present study have well-estab- gen shows a wide range in d15N(24‰ to lished soils that are annually leached by 117‰) depending on its source (parent ma- rainfall. As such, we assumed that the non- terial, biomass degradation, precipitation) leguminous plants in the forests of Costa and various processes affecting soils (leach- Rica and Brazil have d15N values similar to ing, ammonia volatilization, denitrification) those at Barro Colorado Island in Panama. with the result that plants have a wide Animals range of d15N values (Shearer and Kohl, 1994). The most negative leaf d15N values in Based on reported differences in diet the nonlegumes (28‰ to 25‰) are from regions four genera of monkeys should show varia- of young soil with little organic matter, like tion in hair d15N values. The d15N value in Hawaii (,200 years; Vitousek et al., 1989). animal tissues such as hair and bone (DeNiro In such situations, the main nitrogen source and Epstein, 1981; Schwarcz and Schoeninger, is precipitation (217‰ to 14‰) and nitro- 1991) depends on diet (DeNiro and Epstein, gen removal from soil is accomplished largely 1981) and various metabolic factors (Am- by plants (Shearer and Kohl, 1994). This brose and DeNiro, 1986). Within our study, pattern is also common in temperate forests average annual diets are relatively well which recently experienced glaciation. In characterized for three of the four popula- contrast, older, better developed soils sup- tions and are listed in Table 3, the exception port plants that have significantly more being the muriquis from Fazenda Esmer- positive d15N values. In such soils, most alda. In the case of the spider monkeys and nitrogen available to the plant comes from the capuchins, data were taken from re- 76 M.J. SCHOENINGER ET AL.

TABLE 3. Forest canopy, diet, and stable isotope ratios in monkey hair

13 1 15 2 d C‰(PDB) d N‰(AIR) Genus species Canopy x Diet x d15N (common name) Locale (mean rainfall)3 N (s.d.) Range(average)4 N (s.d.) range Cebus capucinus La Selva, Closed (4,000 mm) 4 224.6224.8, 65% fruit 4 7.0 6.8–7.0 (capuchin) Costa Rica (0.2) 224.2 20% animal (0.1) prey 15% green plant material Ateles geoffroyi La Selva, Closed (4,000 mm) 5 224.9225.1, 80% fruit 5 5.0 4.8–5.4 (spider monkey) Costa Rica (0.2) 224.7 20% young (0.5) leaves Brachyteles arach- Esmeralda, Open (,2,000 mm) 7 223.5223.7, 60% leaves 7 5.6 5.0–6.1 noides (woolly- Brazil (0.2) 223.5 30% fruita (0.4) spider or muriqui) 80% leaves 20% fruitb Alouatta palliata La Pacifica, Open (1,100 mm) 12 223.4223.8, 63% leaves 10 3.5 2.8–5.3 (mantled howler) Costa Rica (0.2) 223.1 30% fruit 9 (0.7) 2.8–3.8 1 flowers 3.3 .40% legumes (0.4) 1 Student’s t difference of means test significant at 0.001 level comparing closed vs. open canopy species. 2 Student’s t difference of means test significant at 0.001 level comparing Cebus vs. other three species, Ateles and Brachyteles vs. Cebus or vs. Alouatta. 3 Sources: La Selva (Janzen, 1983), Esmeralda (estimate based on comparison with data from similar areas, Milton, 1984; Strier, 1991), La Pacifica (Glander, 1979). 4 Sources: Ateles and Cebus (Janzen, 1983), Brachyteles (aMilton, 1984; bStrier, 1991), Alouatta (Glander, 1979). 5 Data on Alouatta in parentheses refer to calculations without the single outlier (15.3‰). ported compilations from the same region in curred in association with archaeological Costa Rica (Eisenberg, 1983; Freese, 1983). evidence for bean domestication (DeNiro Data on the muriqui diet are estimated and Epstein, 1981; Farnsworth et al., 1985; based on reports from two long-term field Hastorf and DeNiro, 1985). Studies in other studies in other regions of the Brazilian regions failed to confirm this relationship coastal forest (Milton, 1984a; Strier, 1992). although the use of human or other animal The data on the mantled howler diets come wastes (products that are enriched in 15N) to from La Pacifica, which has been the focus of dress fields in nitrogen poor areas could a long-term field study by KG (Glander, account for this lack (Schwarcz and 1979). Schoeninger, 1991). East African baboons Several studies have noted an increase of with low bone collagen d15N values were approximately 3‰ between trophic levels thought to eat legumes although no observa- across a large number of animal species tional data were available for the animals (Schoeninger and DeNiro, 1984; Minagawa analyzed (Ambrose and DeNiro, 1986). Mon- and Wada, 1984; Schoeninger, 1985; Am- keys often eat leguminous plants and the La brose and DeNiro, 1986; among others), Pacifica howler monkeys were observed which is commonly referred to as the ‘‘tro- spending over 40% of their feeding time phic level effect.’’ This pattern suggests that eating leguminous plant species (Glander, the capuchins in our study should have 1979). This provided us an excellent opportu- higher d15N values than the other three nity to analyze hair from animals known to species because they are omnivores while feed on legumes. the others are herbivores. Metabolic factors have been proposed to We also expected variation among the account for a significant amount of within three herbivorous species. Consumption of trophic level variation. Urea excretion af- N2-fixing leguminous plants should be re- fected by water stress and protein intake in flected in hair d15N values in species living ruminants and, possibly, in nonruminants in open canopy environments. Low bone may result in elevated d15N values (Ambrose collagen d15N values in prehistoric Central and DeNiro, 1986; Cormie and Schwarcz, and South American agriculturalists oc- 1996) although some stressed animals do STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 77 not show the expected elevation (Ambrose Hair was cleaned sequentially with water and DeNiro, 1986; Sealy et al., 1987). Fast- and acetone and dried at 90°C. Approxi- ing can result in elevated d15N values in mately 3 mg were weighed into quartz tubes tissues with rapid turnover (e.g. muscle, with excess cupric oxide, copper, and silver. liver, bone collagen in growing animals; The tubes were sealed under vacuum and Hobson et al., 1993). None of the animals in the samples combusted at 900°C in a muffle this study showed any indication of water, furnace for 2 hours. Tubes were allowed to protein, or calorie stress. Thus, we assumed come to room temperature. Carbon dioxide that diet, rather than metabolic factors, and nitrogen gas were purified sequentially, would have the primary influence on hair collected cryogenically on a glass vacuum d15N values. We expected that the howler line, and analyzed on a Finnegan MAT 251. monkeys of La Pacifica would have lower A glycine laboratory standard analyzed re- d15N values than those in the spider monkey petitively produced a standard deviation of population from the closed, shaded canopy of 0.2‰ in d13C and 0.3% in d15N. Separate La Selva. The muriqui monkey population samples of hair, cleaned and prepared from lives in a forest similar to that of the howlers the same animal, had an average difference but because the diet has not been character- in d13C of 0.18‰ (11 animals; 17 replicates) ized for this particular population, specific and in d15N of 0.35‰ (8 animals; 10 repli- predictions could not be made. cates). This replicability of values includes seven duplicates where the two prepara- MATERIALS AND METHODS tions were made in different years. This Hair samples were taken during routine level of replicability is similar to that re- capture as part of a large battery of biologi- ported previously and is within the level of cal specimen collections on animals cap- variation for individuals (DeNiro and tured for marking and morphometrics (Glan- Schoeninger, 1983). der et al., 1991). Hair was pulled as tufts At La Pacifica, mantled howlers were from directly above the eyebrows on the top trapped in 1989, 1990, and again in 1991. of the head and from the middle of the back Hair from three animals captured in two between the shoulder blades. The procedure different years (1989 and 1990 or 1989 and was consistent between animals, years, and 1991) had an average difference in d13Cof seasons. Twenty-eight individuals from the 0.23‰ and in d15N of 0.20‰; neither of these four genera were analyzed. Capture was reaches significance levels using the t-test accomplished (Lemos De Sa and Glander, statistic. Hair from four animals was col- 1993) using the Pneu-Dart system (Pneu- lected in the dry season and from nine in the Dart, Inc., HC 31, Williamsport, PA). This wet season. The average difference was 0.4‰ system employs disposable nonbarbed darts for d13C values and 0.1‰ for d15N values. 3 15 with a ⁄8-in needle that are delivered by a The difference in d N does not reach signifi- carbon dioxide–powered gun. The darts are cance but that for d13C values just reaches loaded with ‘‘Telazol,’’ a nonnarcotic, nonbar- significance at the 0.05 level. biturate, injectable anesthetic. It is a combi- Although significant statistically, this sea- nation of equal parts by weight of tiletamine sonal difference is not large enough to affect hydrochloride (an arylaminocycloalkanone the pattern of results described below. The dissociative anesthetic) and zolazepam hy- data for wet and dry seasons were averaged drochloride (a nonphenothiazine diazepi- for those animals that were captured twice none with tranquilizing properties) (A.H. because use of either the wet season average Robbins Co., Richmond, VA). The dosage is or the dry season average in comparison 25 mg per kg. Under the standard dosage, with the data from the other three species animals are generally immobilized within 1 produces the same pattern. Although data minute, and fall from the trees within 3–5 on hair growth in nonhuman primates is minutes. Animals falling from the trees are limited to vervets (Cercopithecus aethiops, caught in nylon mesh nets (camper’s ham- Isbell, 1995), Japanese macaques (Macaca mock). All procedures have been approved fuscata fuscata, Inagaki and Nigi, 1988), by appropriate animal use committees. and golden lion tamarins (Leontopithecus 78 M.J. SCHOENINGER ET AL.

Fig. 1. The carbon stable isotope data in hair of four in diet items taken by the animals. In contrast, the d15N monkey species from Central and South American clus- values indicate diet. The omnivorous capuchins have ter bimodally. The two from an open canopy forest (the the most positive d15N values in hair while the mantled muriqui or woolly-spider monkey from Brazil and the howlers have the lowest. The spider monkeys, muriquis, mantled howler from Costa Rica) have d13C values that and howlers are all herbivores and the significantly are significantly less negative than the two from a more lower d15N values in the howlers correlate with their closed canopy forest (spider monkeys and capuchins observed legume consumption. Thus, d13C values in the from Costa Rica). This pattern reflects the d13C values hair of arboreal animals indicate the level of forest cover reported in leaves from tropical forests that differ in the in the animals’ habitat whereas the d15N values serve in amount of canopy cover. As such, the values in the monitoring trophic position and ingestion of leguminous animals’ hair associate with factors affecting photosyn- vegetation. thesis in C3 plants rather than with specific differences

rosalia, Dietz et al., 1995) there is consensus ses along individual hair shafts (cf. White, across these species. Molts occur relatively 1993). rapidly (4–6 weeks) and once during the year. This suggests that hair stable isotope RESULTS values will reflect diet from a limited portion Carbon isotope ratios and canopy cover of the year. For this reason, the seasonal difference we report deserves further inves- The results are presented in Table 3 and tigation with larger numbers of individuals plotted in Figure 1. The carbon data cluster analyzed, more species of monkey, and analy- into two groups. Cebus and Ateles are very STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 79

TABLE 4. d13C in tissue differs by forest type species. Capuchins which are partially insec- 13 Leaves1 Monkey hair2 tivorous have hair d C values similar to Tropical d13C‰ 6 S.D. d13C‰ 6 S.D. those in the almost completely frugivorous forest type (N) (N) spider monkeys. It has been suggested that Deciduous 228.0‰ 6 1.1 (5) 223.5‰ 6 0.3 (19) fruit d13C values are less negative than leaf Semievergreen 228.8‰ 6 1.2 (4) Evergreen 230.4‰ 6 1.5 (13) 224.8‰ 6 0.3 (9) values (van der Merwe and Medina, 1991); yet it is the frugivorous monkey that shows 1 Data from Broadmeadow et al. (1992) and van der Merwe and Medina (1989). Leaves were taken from main canopy; data that more negative average value. A small shown relative to PDB. 13 2 Data shown relative to PDB. difference in d C values between trophic The difference of means in the monkey hair and in the deciduous levels (,1‰) has been documented in both vs. evergreen leaves is significant at the .001 level, two-tailed test. the marine and terrestrial environments when large numbers of individuals were analyzed (Schoeninger and DeNiro, 1984). similar to each other; the ranges overlap and This effect may contribute to the average the means differ by only 0.1‰ more than difference between these two monkey spe- analytical precision. These monkeys are the cies; but a larger number of samples must be ones captured in the closed canopy forest of analyzed in order to confirm this possibility. La Selva, Costa Rica. The Alouatta and Muriqui monkeys and mantled howlers are Brachyteles data form a second cluster; the both frugivorous folivores and it is possible means for these two genera are within 0.1‰ that the similarity of their hair d13C values of each other and the ranges of the data is recording specific dietary items. The simi- overlap almost completely even though the larity of the other two species with markedly Alouatta individuals were captured in Costa different diets, however, argues against this Rica and the Brachyteles individuals are interpretation and supports our interpreta- from Brazil. The two forests, however, are tion that the monkey tissue is recording the very similar to each other in being deciduous photosynthetic response of plants to varying and in having an open canopy. levels of soil respired CO2 and light levels in The difference in d13C values between the the two types of forest. These data indicate two data clusters is of the same magnitude that monkey hair d13C values faithfully rep- and in the same direction as that in leaves resent the type of forest cover of their particu- from open canopies compared with closed lar habitats. canopies. Table 4 shows average data on leaves from main canopies in three types of Nitrogen isotope ratios and diet tropical forests with the average d13C values in the monkey hair. The average leaf values The nitrogen stable isotope ratios in the differ significantly between the evergreen monkey hair show three groups of d15N and deciduous forests (P 5 .001, two-tailed values (Table 3) with Cebus having the most test, df 5 16) as do the two sets of monkey positive values as expected based on its data (P 5 .001, two-tailed test, df 5 26). Al- omnivorous diet. The ingestion of only 20% though we did not have hair samples from animal matter reported for these animals monkeys living in a strictly semievergreen seems low to account completely for the high forest, we expect that the hair from animals d15N value in the capuchins’ hair and sug- living in this type of forest would be interme- gests that a majority of their dietary protein diate between the two extremes. It should be is coming from insects. noted that the standard deviations in the Alouatta palliata displays the lowest aver- monkey hair are approximately one-third age d15N values of all the species studied. that reported for the leaves. This is due to With the exception of one outlier, the range averaging by the monkeys as they eat forest of values falls well below the other species. foods with slightly different d13C values (see The average value (with and without the Bada et al., 1990, for another example of this outlier) is significantly different from the process). two other herbivorous species at the 0.001 The carbon isotope patterns are indepen- level. Since howlers, spider monkeys, and dent of the specific diets of the individual muriquis all obtain their nitrogen from 80 M.J. SCHOENINGER ET AL. plants, the trophic level effect cannot pro- information on the habitats of individual vide an explanation. populations which may, or may not, approxi- The pattern supports our prediction, based mate their present-day habitat distribution. on their dependence on leguminous plants, In contrast, the nitrogen stable isotope of lower d15N values in the howlers at La data appear to be tracking species’ diet. Pacifica (Glander, 1979). As highly selective Omnivorous capuchins have the most posi- feeders (Glander, 1983) the majority of these tive d15N values in hair as expected from individuals are focusing on legumes, yet a their trophic position. The mantled howler single outlier has a hair d15N value similar population, which consumes significant to those in the spider monkeys and in the quantities of leguminous plants, has the muriquis. This individual howler (sampled least positive average values. Our study is 15 in both 1989 and 1990 with the same d N the first demonstration of a correlation be- value in both years) appears to have been tween this dietary signal and observational selecting a diet different from that of the rest data reporting legume consumption. These of the group. This is intriguing as howlers data support the use of d15N values in mu- are reported to use the same routes for seum-curated specimens to indicate trophic travel, often following each other in single position and dependence on legumes. file (Milton, 1980). The level of variation in Endemic species, in general, are vulner- our data suggests a different behavioral able to extinction with increasing human pattern in this particular individual. The activity (Dobson et al., 1997), but it is appar- other two herbivorous species, Ateles and ent that all endemic primate species have Brachyteles, are more similar to each other not reacted in the same manner to distur- than either is to the other two monkey bance. Extant populations of howler mon- populations even though their average d15N values are not identical to each other. The keys, capuchins, and muriquis exist in a data fall between the other two species as variety of forest habitats with variable re- predicted based on their designation as fru- sponses to human intervention; spider mon- givorous folivores lacking observational data keys are more restricted in habitat and indicating dependence on leguminous plants. response. Stable isotope analysis of protein- aceous tissues (hair, nails, skin, muscle, DISCUSSION AND CONCLUSIONS bone collagen, etc.) can provide data on habitat use and diet which should further The carbon stable isotope data in hair from four species of Central and South our understanding of these species differ- ences. If carbon isotope studies of animals American monkeys correlate with the d13C values reported in leaves from forests with taken prior to habitat alteration reveal that similar amounts of canopy cover. Even earlier distributions of howlers and capu- though each species selects its own unique chins were more restricted than today, it mix of food items within the arboreal envi- would suggest that preferred habitat differs ronment, the carbon stable isotope data in markedly from today’s general pattern, pro- the animals’ hair reflect factors affecting viding further support of the contentions that these species have highly adaptable photosynthesis in C3 plants rather than specific differences in diet items taken by behavioral repertoires. Further studies of each monkey species. This pattern strongly other populations of howler monkeys, such indicates that hair samples from animals as ones where legumes are not a major part taken early in this century carry within of the diet (e.g., Estrada, 1984), could reveal their carbon isotope signal information re- aspects of their subsistence strategies that garding the forest cover of the areas at the have permitted them to acclimate to human time the animals lived. Hair from museum- expansion. It would also be useful to analyze curated pelts of monkeys taken in the past other species which have significantly differ- could be analyzed in order to determine the ent foraging strategies (e.g. solitary vs. group various types of forests inhabited by these foraging) to evaluate the apparent recorda- and other species prior to extensive habitat tion of an alternative strategy in one of our destruction. Such a study would provide howler monkeys. Comparison of the four STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 81 Cebus species should indicate the level of Brazil; and OTS for permission to capture omnivory in the single robust species and in animals at La Selva, Costa Rica. We are also the other two gracile species as well as any extremely grateful to Norlin Benevenga for variation within single species. Perhaps his advice on protein metabolism, Leanne there are patterns here which associate with Nash who pointed out the howlers’ emphasis the success of this genus in response to on legumes, and Tom Sharkey for his exper- human disturbance. In contrast, extant tise on photosynthesis. Discussions with Jef- populations of spider monkeys are thought frey Bada, Melinda Carter, Tom Crenshaw, to have limited distribution due to their Jim Moore, Michele Morgan, Donald Pate, disappearance from former habitats (Lip- Holly Reeser, and Karen Strier on various pold, 1988) and the possible decline in num- versions of the manuscript were extremely bers in other areas (Fedigan et al., 1985). It helpful. The constructive comments of sev- has been suggested that they ‘‘may be among eral anonymous reviewers, especially one the first primate species to decline with stable isotope expert who carefully read severe disturbance’’ (Eisenberg, 1983:452). more than one version, resulted in a greatly If animals taken prior to human population improved manuscript. We appreciate the expansion lived in habitats similar to those time and effort of all the above. used today, it would indicate that loss of specific habitat is the main causal variable LITERATURE CITED contributing to this disturbance. Alterna- tively, if a variety of habitats were used by Ambrose SH and DeNiro MJ (1986) The isotopic ecology of East African mammals. Oecologia 69:395–406. earlier populations, then the cause of popu- Bada JL, Peterson RO, Schimmelmann A, and Hedges lation decline could be due to some other REM (1990) Moose teeth as monitors of environmen- factor (e.g., human hunting) rather than tal isotopic parameters. Oecologia 82:102–106. Broadmeadow MSJ, Griffiths H, Maxwell C, and Bor- habitat loss per se. Additional analyses of land AM (1992) The carbon isotope ratio of plant muriqui monkeys should advance our under- organic material reflects temporal and spatial varia- standing of this species’ apparently flexible tions in CO2 within tropical forest formations in Trinidad. Oecologia 89:435–441. ecological adaptation (e.g. Milton, 1984a vs. Chivers DJ and Hladik CM (1980) Morphology of the Strier, 1991) which does not appear to have gastrointestinal tract in primates: Comparisons with permitted it to withstand human expansion other mammals in relation to diet. J. Morphol. 166: 337–386. as successfully as the howlers or capuchins. Cormie AB and Schwarcz HB (1996) Effects of climate Studies such as these, as well as others on on deer bone d15N and d13C: Lack of precipitation effects on d15N for animals consuming low amounts of additional populations of the species we C4 plants. Geochim. Cosmochim. Acta 60:4161–4166. selected and on other species, promise to Crockett CM and Eisenberg JF (1987) Howlers: Varia- provide data useful in delimiting the ‘‘kinds tions in group size and demography. In BB Smuts, DL Cheney, RM Seyfarth, RW Wrangham, and TT Struh- of human activities [which] most threaten saker (eds): Primate Societies. Chicago: The Univer- them’’ (Pulliam and Babbitt, 1997). sity of Chicago Press, pp. 54–68. DeNiro MJ and Epstein S (1978) Influence of diet on the ACKNOWLEDGMENTS distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42:495–506. Supported in part by grants BNS-8819733 DeNiro MJ and Epstein S (1981) Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. to KEG and DBC-9118876 to M.F. Teaford Cosmochim. Acta 45:341–351. and KEG. The hair sample collecting portion DeNiro MJ and Schoeninger MJ (1983) Stable carbon of the project was further supported by and nitrogen isotope ratios of bone collagen: Varia- tions within individuals, between sexes, and within Conservation International and Wildlife populations raised on monotonous diets. J. Arch. Sci. Preservation Trust International. Funding 10:199–203. Dietz JM, Baker AJ, and Allendorf TD (1995) Correlates for the stable isotope analysis was provided of molt in golden lion tamarins (Leontopithecus rosa- by the Wisconsin Alumni Research Founda- lia). Am. J. Primatol. 36:277–284. tion. Thanks are due to Dr. Stephan Schmid- Dobson AP, Rodriguez JP, Roberts WM, and Wilcove DS (1997) Geographic distribution of endangered species heiny and the Board of Directors of Haci- in the United States. Science 275:550–553. enda La Pacifica for permission to capture Eisenberg JF (1983) Ateles geoffroyi (Mono Arana, Mono animals on La Pacifica in Costa Rica; Sr. Colorado, spider monkey). In DH Janzen (ed.): Costa Rican Natural History. Chicago: University of Chicago Jose Carlos Reis de Magalhaes for permis- Press, pp. 451–453. sion to capture animals on his farm in Estrada A (1984) Resource use by howler monkeys 82 M.J. SCHOENINGER ET AL.

(Alouatta palliata) in the rain forest of Los Tuxtlas, Lemos de Sa RM and Strier KB (1992) A preliminary Veracruz, Mexico. Int. J. Primatol. 5:105–131. comparison of forest structure and use by two isolated Farnsworth P, Brady JE, DeNiro MJ, and MacNeish RS groups of woolly spider monkeys, Brachyteles arachnoi- (1985) A re-evaluation of the isotopic and archaeologi- des. Biotropica 24:455–459. cal reconstructions of diet in the Tehuacan Valley.Am. Lippold LK (1988) A census of primates in Cabo Blanco Antiq. 50:102–116. Absolute Nature Reserve, Costa Rica. Brenesia 29: Farquhar GD, O’Leary MH, and Berry JA (1982) On the 101–105. relationship between carbon isotope discrimination Loreto F, Harley PC, Di Marco G, and Sharkey TD and the intercellular carbon dioxide concentration in (1992) Estimation of mesophyll conductance to CO2 leaves. Aust. J. Plant Physiol. 9:121–137. flux by three different methods. Plant Physiol. 98: Fedigan LM, Fedigan L, and Chapman C (1985) Census 1437–1443. of Alouatta palliata and Cebus capucinus monkeys in Marino BD and McElroy MB (1991) Isotopic composition Santa Rosa National Park, Costa Rica. Brenesia of atmospheric CO2 inferred from carbon in C4 plant 23:309–322. cellulose. Nature 349:127–131. Ford SM and Davis LC (1992) Systematics and body Milton K (1980) The Foraging Strategy of Howler Mon- size: Implications for feeding adaptations in New keys: A Study in Primate Economics. New York: World monkeys. Am. J. Phys. Anthropol. 88:415–468. Columbia University Press, 165 pp. Freese CH (1983) Cebus capucinus (Mono Cara Blanca, Milton K (1981) Food choice and digestive strategies of white-faced capuchin). In DH Janzen (eds): Costa two sympatric primate species. Am. Nat. 117:476–495. Rican Natural History. Chicago: University of Chicago Milton K (1984a) Habitat, diet, and activity patterns of Press, pp. 458–460. free-ranging woolly spider monkeys (Brachyteles Glander KE (1979) Howling monkey feeding behavior arachnoides E. Geoffroy 1806). Int. J. Primatol. 5:491– and plant secondary compounds: a study of strategies. 514. In GG Montgomery (ed.): The Ecology of Arboreal Milton K (1984b) The role of food-processing factors in Folivores. Washington, D.C.: Smithsonian Institution primate food choice. In PS Rodman and JGH Cant Press, pp. 561–574. (eds): Adaptations for Foraging in Nonhuman Pri- Glander KE (1983) Alouatta palliata (Congo, howling mates: Contributions to an Organismal Biology of monkey, howler monkey). In DH Janzen (ed.): Costa Prosimians, Monkeys, and Apes. New York: Columbia Rican Natural History. Chicago: University of Chicago University Press, pp. 249–279. Press, pp. 448–449. Milton K and May ML (1976) Body weight, diet and Glander KE, Fedigan LM, Fedigan L, and Chapman C home range area in primates. Nature 259:459–462. (1991) Capture techniques and measurements of three Minagawa M and Wada E (1984) Stepwise enrichment monkey species in Costa Rica. Folia Primatol. 57: of 15N along food chains: Further evidence and the 70–82. relation between d15N and animal age. Geochim. Hastorf CA and DeNiro MJ (1985) New isotopic method Cosmochim. Acta 48:1135–1140. used to reconstruct prehistoric plant production and Minson DJ, Ludlow MM, and Troughton JH (1975) cooking processes. Nature 315:489–491. Differences in natural carbon isotope ratios of milk Hobson KA, Alisauskas RT, and Clark RG (1993) Stable- and hair from cattle grazing tropical and temperate nitrogen isotope enrichment in avian tissues due to pastures. Nature 256:602. fasting and nutritional stress: Implications for isoto- Nakamura K, Schoeller DA, Winkler FJ, and Schmidt pic analyses of diet. Condor 95:388–394. HL (1982) Geographical variations in the carbon Hoefs J (1987) Stable Isotope Geochemistry. New York: isotope composition of the diet and hair in contempo- Springer-Verlag, 241 pp. rary man. Biomed. Mass Spec. 9:390–394. Inagaki H and Nigi H (1988) Annual changes in hair Neville MK, Glander KE, Braza F, and Rylands AB length of the Japanese monkey (Macaca fuscata fus- (1988) The howling monkeys, genus Alouatta. In RA cata). Primates 29:81–89. Mittermeier, AB Rylands, AF Coimbra-Filho, and Isbell LA (1995) Seasonal and social correlates of changes GAB de Fonseca (eds.): Ecology and Behavior of in hair, skin, and scrotal condition in vervet monkeys Neotropical Primates. Washington, D.C.: World Wild- (Cercopithecus aethiops) of Amboseli National Park, life Fund, pp. 349–453. Kenya. Am. J. Primatol. 36:61–70. O’Leary MH (1988) Carbon isotopes in photosynthesis. Janson CH and Boinski S (1992) Morphological and BioSci. 38:328–336. behavioral adaptations for foraging in generalist pri- Pulliam HR and Babbitt B (1997) Science and the mates: The case of the cebines. Am. J. Phys. Anthro- protection of endangered species. Science 275:499– pol. 88:483–498. 500. Janzen DH (1983) Costa Rican Natural History. Chi- Robinson JG and Janson CH (1987) Capuchins, squirrel cago: University of Chicago Press, 816 pp. monkeys, and atelines: Socioecological convergence Jones RJ, Ludlow MM, Troughton JH, and Blunt CG with Old World primates. In BB Smuts, DL Cheney, (1981) Changes in the natural carbon isotope ratios of RM Seyfarth, RW Wrangham, and TT Struhsaker the hair from steers fed diets of C4,C3 and C4 species (eds.): Primate Societies. Chicago: The University of in sequence. Search 12:85–87. Chicago Press, pp. 69–82. Kay RF (1975) The functional adaptations of Primate Rodman PS and Cant JGH (1984) Adaptations for molar teeth. Am. J. Phys. Anthropol. 43:195–216. Foraging in Nonhuman Primates. New York: Colum- Kluge M and Ting IP (1978) Crassulacean Acid Metabo- bia University Press, 351 pp. lism: Analysis of an Ecological Adaptation. Berlin: Rosenberger AL (1992) Evolution of feeding niches in Springer-Verlag, 201 pp. New World monkeys. Am. J. Phys. Anthropol. 88:525– Konstant W, Mittermeier RA, and Nash S (1985) Spider 562. monkeys in captivity and in the wild. Primate. Con- Rundel PW, Ehleringer JR, and Nagy KA (1989) Stable serv. 5:82–109. Isotopes in Ecological Research. New York: Springer- Lemos de Sa RM and Glander KE (1993) Capture Verlag, 525 pp. techniques and morphometrics for the woolly spider Schoeninger MJ (1985) Trophic level effects on 15N/14N monkey, or muriqui (Brachyteles arachnoides, E. Geof- and 13C/12C ratios in bone collagen and strontium froy 1806). Am. J. Primatol. 29:145–153. levels in bone mineral. J. Hum. Evol. 14:515–525. STABLE ISOTOPES MONITOR NEW WORLD MONKEY HABITS 83

Schoeninger MJ and DeNiro MJ (1984) Nitrogen and and 13C/12C ratios in Amazonian rain forests. Geo- carbon isotopic composition of bone collagen from chim. Cosmochim. Acta 53:1091–1094. marine and terrestrial animals. Geochim. Cosmo- van der Merwe NJ and Medina E (1991) The canopy chim. Acta 48:625–639. effect, carbon isotope ratios and foodwebs in Amazo- Schwarcz H and Schoeninger MJ (1991) Stable isotope nia. J. Arch. Sci. 18:249–259. analyses in human nutritional ecology. Yrbk. Phys. Van Kessel C and Roskoski JP (1983) Nodulation and Anthropol. 34:283–321. N2-fixation by Inga jinicuil, a woody legume in coffee Sealy JC, van der Merwe NJ, Lee Thorp JA, and plantations. III. Effect of fertilizers and soil shading Lanham JL (1987) Nitrogen isotopic ecology in south- on nodulation and nitrogen fixation (acetylene reduc- ern Africa: Implications for environmental and di- tion) in Inga jinicuil seedlings. Plant Soil 72:95–105. etary tracing. Geochim. Cosmochim. Acta 51:2707– Virginia RA and Delwiche CC (1982) Natural 15N 2717. abundance of presumed N2-fixing and non-N2-fixing Shearer G and Kohl DH (1986) N2-fixation in field plants from selected ecosystems. Oecologia 54:317– settings: Estimations based on natural 15N abun- 325. dance. Aust. J. Plant Physiol. 13:699–756. Vitousek PM, Shearer G, and Kohl DH (1989) Foliar Shearer G and Kohl DH (1994) Information derived 15N abundance in Hawaiian rainforest: Patterns and from variation in the natural abundance of 15Nin possible mechanisms. Oecologia 78:383–388. complex biological systems. In E Buncel and WHJ Vogel JC (1978) Isotopic assessment of the dietary Saunders (eds.): Isotopes in Organic Chemistry: Heavy habits of ungulates. S. Afr. J. Sci. 74:298–301. Atom Isotope Effects. Amsterdam: Elsevier, pp. 191– Wahlen M (1994) Carbon dioxide, carbon monoxide and 237. methane in the atmosphere: abundance and isotopic Shearer GB, Kohl DH, Virginia RA, Bryan BA, Skeeters composition. In K Lajtha and RH Mitchener (eds): JL, Nilsen ET, R. SM, and Rundel PW (1983) Esti- Stable Isotopes in Ecology and Environmental Sci- mates of N2-fixation from variation in the natural ence. Oxford: Blackwell, pp. 93–113. abundance of 15N in Sonoran Desert ecosystem. Oeco- White C and Schwarcz H (1994) Temporal trends in logia 56:365–373. stable isotopes for Nubian mummy tissues. Am. J. Strier KB (1991) Diet in one group of woolly spider Phys. Anthropol. 93:165–188. monkeys, or muriquis (Brachyteles arachnoides). Am. White CD (1993) Isotopic determination of seasonality J. Primatol. 23:113–126. in diet and death from Nubian mummy hair. J. Arch. Strier KB (1992) Atelinae adaptations: Behavioral strat- Sci. 20:657–666. egies and ecological constraints. Am. J. Phys. Anthro- Yakir D and Israeli Y (1995) Reduced solar irradiance pol. 88:515–524. effects on net primary productivity (NPP) and the d13C Tieszen LL, Boutton TW, Tesdahl KG, and Slade NA and d18O values in plantations of Musa sp., Musaceae. (1983) Fractionation and turnover of stable carbon Geochim. Cosmochim. Acta 59:2149–2151. isotopes in animal tissues: Implications for d13C analy- Zingeser MR (1973) Dentition of Brachyteles arachnoi- sis of diet. Oecologia (Berlin) 57:32–37. des with reference to Alouattine and Atelinine affini- van der Merwe NJ and Medina E (1989) Photosynthesis ties. Folia Primatol. 20:351–390.