We operationally defined adoption as any relationship between an adult and orphan infant or juvenile in which the adult shows species-specific maternal behavior towards the orphan for at least a two month period. Following a standard definition of adoption 
, we required that the adult be permanently associated with the orphan, as well as, at the very least, wait during travel for, provide protection in conflicts to, and share food with the orphan. These behaviors are altruistic in the sense that they are costly for the adopting individuals and do not bring any visible benefit to them, while being beneficial to the orphans. Since juvenile chimpanzees remain associated with their mothers for over ¾ of their time and become clearly independent of their mothers only when they have reached adolescence, we included this period in our study. It is important to underline that adult chimpanzees at Taï do not wait for juveniles or infants, or react to their whimpering at being left behind until they are their mothers. In chimpanzees, orphans suffer tremendous costs in terms of reduced survivorship (orphans less than 5 years of age normally do not survive 
) or retardation in physical development (up to 6 years delay 
). However, if adopted, such orphans may present almost normal physical development 
In 3 communities of Taï chimpanzees that have been studied for 27 years, we observed 36 cases of individuals being orphaned and surviving this traumatic event for over 2 months. In 18 of these cases, an adoption was observed to occur (). In addition, during that same time interval, 22 small unweaned infants (average age of 1.85 years) lost their mothers and disappeared with them, while a handful died within a couple weeks, before an adoption was possible. Eight adoptions were performed by females, including 1 by an older sibling and 3 by possible unrelated ‘friends’ of the dead mother (). However, in general, the presence of living close relatives (e.g., full or half siblings) did not increase the likelihood of adoption in our sample (see ). Two adoption cases by adult females who breast fed unrelated young infants (less than 20 months of age) for many years are particularly noteworthy (Nabu and Totem, ). These two examples are illustrative of the huge potential benefit adoption has on unweaned orphans, however, in our sample, adoption did not increase the likelihood of the orphans surviving two years following the death of their mother (see ). An equal number of male (N
10) and female (N
8) orphans were adopted. Orphans adopted by male rather than female adults did not differ significantly in either sex or age ().
Number of orphans and adoptions seen in the 27 years of observation of three study groups of the Taï chimpanzee project.
Successful adoptions of orphaned infants in three communities of Taï chimpanzees.
Figure 1 The presence of a close kin does not increase the likelihood of adoption in Taï chimpanzees (Fisher exact test: p=0.463).
Of special interest are the adoptions by males. These cases include 3 older siblings and 6 adults performing 10 adoptions; 1 of the adoptive adult males proved to be the father of the orphan, while 3 were unrelated to the orphan, and 2 were of an unknown relationship to the orphan (). Male chimpanzees are considered to be adults when they are 15 years old. At this age, they become very social with other males of the group, spend a lot of time grooming one another, and compete aggressively for access to females. They are also the main hunters in the group and actively defend the territory from intruders 
. Male chimpanzees, like males of polygynous human societies, have not been observed to develop long-term bonds with specific females, nor to invest much in their own offspring 
. Male chimpanzees at Taï have not been observed to show obvious paternal behavior, except in terms of playing more often and, to a lesser extent, grooming more with their own offspring than with other youngsters of the same age 
. We were able to test for genetic relatedness in 4 out of the 6 cases of orphans adopted by adult males, and in 3 of them, the males were not related to the youngsters (exclusion at 2 to 4 loci were found) 
As can been seen in , adoption of orphans by adult males represented an important investment in the youngsters, as, minimally, males were seen to share food with them as well as wait for them and support them during social conflicts (this is the operational definition of adoption we used). The two infants adopted by the alpha male of the North Group, Brutus, matured after adoption without showing any of the physical retardation so typically seen in orphans. Therefore, they benefited dramatically from this investment. Five years into the adoption, Ali was becoming a healthy adolescent male when he died during an Ebola outbreak. Another cost of adopting orphans for males is that rivals of the adopter may use the orphan to harass him, and Ulysse, a middle-ranking male, seemed unable to cope with this and was seen to actively interrupt both the adoptions of his presumed younger sister, Bonnie, and of the unrelated male, Brando, following the increasing harassment exerted against the orphans by his rival males ().
Paternal-like behavior observed during an adoption by adult males (with the maternal investment as reference).
Remarkably, all adult males of the East Group that adopted young orphans went a step further by investing in unweaned small infants and carrying them dorsally during travel for many months (see and of Porthos with Gia) (). Since, Taï chimpanzees walk about 8 km per day on average, this represents a notable investment. Porthos' adoption of Gia lasted for 17 months, until his death due to Anthrax, and he was seen to carry her even in extremely risky situations, such as during encounters with neighboring communities 
. Furthermore, some males were seen to share their night nest with their adopted infant (). Fredy, the 3rd
ranking male of the East Group, adopted Victor, the son of Vanessa, who died from Anthrax in late December 2008, and shared his nest with him every night, carried him on his back for all long travels, and shared the Coula nuts he opened from December 2008 to July 2009. For example, on February 17th, Fredy cracked 196 Coula nuts for 2h05mn and shared pieces of 79% of them. This gives a measure of the altruistic investment made in an unrelated infant.
The adult male Porthos with his adopted female infant Gia.
The adult male Porthos carrying on his back the adopted female infant Gia.
These adoptions by adult males of orphans that are often not their own offspring plainly show that, contrary to earlier sweeping conclusions 
, chimpanzees are sensitive to the welfare of unrelated group members. Adoption, which is not uncommon in the animal kingdom, including in humans, is normally explained by close genetic relatedness 
, so the adoption of unrelated orphans by adult males is notable. Nevertheless, might chimpanzee males gain some indirect benefit from investing so heavily in unrelated infants? One potential long-term benefit of adoption by adult males is that once an orphan becomes an adult, 10 years later, he could become an ally of the aging male. This might have happened in the case of Ali, had he survived longer. However, this certainly does not apply to adoption of female orphans. Economists have proposed that, in humans, either the improved reputation of the altruistic individual might compensate for the costs or that a social normative rule, in the form of punishment by one's peers, imposes such behaviors 
. Punishment seems unlikely in the case of adoption of small orphans and we have not seen this to occur. On the other hand, improved reputation might play a role, but we have not yet determined that males who adopt orphans are more reproductively successful, as should have been the case for Brutus, the male that adopted two orphans, and Ulysse 
The observations of adoptions in Taï chimpanzees support our proposition, emerging from Hamilton's rule, that altruism is not hard wired and will be directed specifically towards individuals that profit significantly from this act. The fact that we could not show that adopted orphans have higher survival rates than non-adopted orphans () is most likely due to the high mortality rates in the Taï chimpanzee study groups for over 2 decades, which may easily overshadow potential beneficial effects such as this 
. By guaranteeing that all individuals have a safe environment and access to food, captive situations might not mimic situations in which the welfare of others is an issue. Altruism in the case of adoption in forest chimpanzees seems to be the outcome of the specific socio-ecological conditions faced by the individuals. The high level of adoption observed in Taï chimpanzees compared to other well-studied East African populations might result from the fact that the Taï population coexists with a large population of leopards and the resulting high predation pressure exerted by these cats seems to have promoted strong within-group solidarity in the form of care for all injured individuals as well as joint coalition defense against the leopards 
. Once established, this care for the welfare of others seems to have been generalized to new social contexts, including adoption 
. Any discussions about the evolution of altruism must include the caveat that dissimilar socio-ecological conditions will lead to important population differences in both chimpanzees and humans and we need to remain very careful before making any claims about species differences.