Twinning is rare in anthropoid primates, but it occurs in a variety of species, including New World atelid monkeys (Link et al., 2006). This reproductive strategy is common in some small primates (e.g., Callithrichinae), but it does not seem to be appropriate for larger species, for which the cost of simultaneously rising two infants is quite high (Chapman and Chapman, 1986; Link et al., 2006). Similarly, infanticide (the killing of an infant) is an uncommon primate behavior. This occurs particularly when a mature male wins alpha status and kills unrelated infants. The loss of the infant allows the mother to become receptive sooner and to mate with the new, infanticidal male, likely increasing his reproductive success (i.e. sexual selection hypothesis; van Shaik 2000). Infanticide has been documented in several populations of howler monkeys (Crockett, 2003). In this note we report an unusual case of twinning and infanticide in red howler monkeys (Alouatta seniculus).

The observation took place during a socio-ecological study in Santa Rosa Farm, located in San Martin (Meta, Colombia) (3°36′52.10″N, 73°38′34.20″W, 373 m a.s.l). The study group, composed of two adult males, two adult females, one juvenile male, and one infant male, ranged over 12 ha within a forest fragment of 32 ha. Changes in alpha male status were observed three times during a period of six months (February–August 2004). On July 18th we were following the subordinate adult male (“Tamarindo”) and the alpha male (“Die”) was not observed that day. Both males showed injuries and we suspected a recent change in alpha male status, since Die did not rejoin the group. In the morning we noted that one of the adult females (“Juana”) was giving birth. The infant's tail was first observed and then its hind limbs. Five minutes later (8:15 h), half of the infant's body was outside, but the mother was unable to completely pull it out. The infant's head was still inside after 45 minutes of delivery. More than two hours later (11:28 h) the female finally took out the infant, but it was dead. The mother held it for about 10 min. and, then, dropped it at 11:40 h. Our inspection of the infant suggested that it died asphyxiated by the umbilical cord. Surprisingly, there was another infant attached alive to the cord, but about one third the size of the dead one. We placed the second infant on the forest floor closer to the mother. She went down to the ground after hearing the infant screaming, but probably because of its small size and its attachment to the placenta, it was unable to grab the mother's fur. She took the infant and went to the canopy, where the adult male approached her and took the infant. The male quickly bit the infant's head and dropped it in the creek, where it died.

We suggest that twinning may incur mechanical complications during delivery. Although the resident males fought the day before the infanticide (as expected by the sexual selection hypothesis), we do not know which male had sired the infants. Therefore, it is not possible to confirm the attack completely fits the sexual selection hypothesis.