Coprophagy, or the behavior of eating feces, is classified as autocoprophagy when the individual eats its own feces, or allocoprophagy when it eats the feces of others (Hirakawa, 2001; Graczyk and Cranfield, 2003). This habit is observed in lagomorphs, rodents, marsupials, and primates. Among leporids, coprophagy occurs in the form of caecotrophy (the reingestion of soft feces or caecotrophs) and serves to improve the absorption of vitamins and microbial proteins (Hirakawa, 2001). Caecotrophy has also been observed in a prosimian, the sportive lemur Lepilemur leucopus (Hladik, 1978). Among anthropoid primates, coprophagy has been observed in captive and wild apes (chimpanzees, gorillas, orangutans and gibbons: Hill, 1966; Gilloux et al., 1992; Warniment and Brent, 1997; Nash et al., 1999; Faraldo and Taylor, 2003; Graczyk and Cranfield, 2003; Krief et al., 2004), Old World monkeys (baboons and rhesus macaques: Brent et al., 2002; see also Graczyk and Cranfield, 2003), and New World monkeys (marmosets, tamarins and capuchin monkeys: Anderson et al., 1991; Clark, 1994; Wissman, 1999; Taylor, 2002).
Krief et al. (2004) discussed a number of hypotheses to explain coprophagy by captive primates: (a) food deficiency, (b) boredom, (c) social stress, and (d) medical problems. The only report of coprophagy in capuchin monkeys (Anderson et al., 1991) was of an occurrence during the integration of a tame adult female into a captive group. Anderson et al. argued that this abnormal behavior may have been related to food deficiency and/or social stress, since the human-raised female used to eat feces before adjusting to the standard primate food pellets offered in captivity. She was also frequently involved in agonistic interactions with other group members soon after her integration.
In this paper we report cases of coprophagy by captive brown capuchin monkeys and examine whether they fit the “food deficiency” and “social stress” hypotheses cited above. We predict that if food limitation is the primary stimulus for this behavior, coprophagy will be more frequent during those periods in which the monkeys have no food available in the cage. On the other hand, if social pressure can explain this behavior, we can expect to find an inverse relationship between social rank and individual frequency of coprophagy. In addition, we would expect that a given individual will be more likely to eat feces after being harassed by other groupmates.
The social behavior of a group of 10 brown capuchin monkeys (Table 1) was studied over 219.5 observation hours, from 8 April to 24 October 2003 at the Parque Farroupilha in Porto Alegre, Rio Grande do Sul, Brazil. This urban park has many visitors and is close to busy city streets. The study group was kept in a wired hexagonal cage with a cement floor, measuring approximately 5.5 m on each side with a height of 3.5 m. For behavioral enrichment the cage was equipped with a wheel, two tires and one movable ladder. The monkeys were fed only once a day, at about 09:00 h, with fruits, vegetables and sometimes peanuts. The cage was washed with water once a week on Thursday afternoons, so the monkeys had no food available until Friday morning.
Age-sex composition of the study group, social rank, and recorded cases of autocoprophagy and allocoprophagy by each individual.
We recorded all occurrences of coprophagy following the behavior sampling rule with continuous recording (Martin and Bateson, 1993). We determined the dominance status of each individual based on the frequency and distribution of agonistic interactions within dyads (see Janson, 1985). An individual was considered to have high social rank if it was the initiator of aggression more frequently than a recipient of aggression. In contrast, low social rank was indicated by a higher frequency of aggression received than performed.
We recorded a total of 73 events of coprophagy (0.33 events per hour of observation). Six individuals (Amelia, Fabi, Barba, Dali, Guri and Fiba) were observed to eat feces. The remaining four (Roger, Chico, Tiburcio and Assis) never demonstrated this behavior (Table 1), but were occasionally seen to drink urine from the floor. Coprophagy was more prevalent than expected in females than in males (41 events vs. 32 events; χ2 = 59.671, d.f. = 1, p < 0.001). Allocoprophagy was more frequent than autocoprophagy.
Coprophagy was more common in the afternoon than in the morning (47 events or 0.47 events per hour of observation vs. 26 events or 0.22 events per hour; χ2 = 11.000, d.f. = 1, p < 0.001), suggesting that food availability may play an important role in the occurrence of this behavior. However, we recorded only nine instances of coprophagy (12%) in which there was no food on the cage floor. The frequency of coprophagy on Thursdays (when the cage was cleaned and devoid of food) was 0.21 events per hour, not above the expected level based on sampling effort (14 events; χ2 = 1.785, d.f. = 1, NS).
We observed a total of 326 cases of agonistic interactions during the study (1.49 events per hour of observation). Social rank did not explain inter-individual differences in coprophagy (r2 = 0.002, n = 10, F-ratio = 0.016, p = 0.901). For example, the two highest-ranking individuals (Roger and Chico) and the lowest-ranking individual (Assis) never ate feces, whereas Amelia and Fabi (ranked third and ninth, respectively) showed the highest frequencies of coprophagy (Table 1). In addition, only rarely had the individual observed eating feces been harassed earlier in the day (Barba: two cases; Fabi: three cases; and Guri: two cases). Therefore, the “social stress” hypothesis was not supported.
We observed a high frequency of coprophagy in this group of captive brown capuchin monkeys, a behavior reported only once before in this species (Anderson et al., 1991). In a similar study on the social behavior of a captive group of nine capuchin monkeys at the Parque Zoológico de Sapucaia do Sul, Rio Grande do Sul, Brazil, Daniel B. Montano did not observe a single case of coprophagy during 208 hours of observation over 16 months (pers. comm.).
Its prevalence in adult females and immature individuals may be related to a diet insufficient in protein, as described for marmosets by Flurer and Zucker (1988). The tendency of females and immatures to exploit diets richer in protein has been described for a number of primates, and is related to the nutritional demands of gestation, lactation, and growth (see Bicca-Marques and Calegaro-Marques, 1994). This tendency, coupled with the fact that coprophagy was more frequent in the afternoon, supports the “food deficiency” hypothesis. However, most cases of coprophagy were observed when there was still food available in the cage, and its frequency on Thursdays (when the cage is cleaned) was not greater than that expected by chance.
Although social rank may interfere with food access, and may contribute to within-group differences in food deficiency, we found no relationship between this variable and the frequency of coprophagy. Thus, if the individual's degree of social pressure (reflected here by frequency of harrassment) is inversely related to social rank, the “social stress” hypothesis may also be rejected to explain the occurrence and distribution of coprophagy in this group.
We did not test the “boredom” and “medical problem” hypotheses for coprophagy, both of which would have required detailed information on individual differences in personality and health status. Therefore, it is not possible to evaluate whether the observed distribution of coprophagy among group members was caused by a single factor or by an interplay between food deficiency, boredom, and medical problems. Finally, it is possible that this uncommon behavior has been transmitted culturally among group members, as proposed for captive chimpanzees by Nash et al. (1999). The observations of Roger, Chico, Tiburcio and Assis (the only four individuals not involved in coprophagy) drinking urine is compatible with Stemmler-Morath's (1937, apud Hill, 1966) findings that apes began coprophagy by drinking their urine. To better understand the causes and significance of coprophagy for capuchin monkeys, it is important that researchers report all observations of this behavior both in captivity and in the wild.
We thank the personnel of the Parque Farroupilha, especially the zookeepers, for their support during this study.
 Helissandra Mattjie Prates and Júlio César Bicca-Marques, Pontifícia Universidade Católica do Rio Grande do Sul, Faculdade de Biociências, Av. Ipiranga 6681, Pd 12A, Porto Alegre 90619-900, Rio Grande do Sul, Brazil. E-mail: <email@example.com>.