Eviction-driven infanticide and sexually selected adoption and infanticide in a neotropical parrot

Infanticide and adoption have been attributed to sexual selection, where an individual later reproduces with the parent whose offspring it killed or adopted. While sexually-selected infanticide is well known, evidence for sexually-selected adoption is anecdotal. We report on both behaviors at 346 nests over 27 years in green-rumped parrotlets (Forpus passerinus) in Venezuela. Parrotlets are monogamous with long-term pairbonds, exhibit a strongly male-biased adult sex ratio, and nest in cavities that are in short supply, creating intense competition for nest sites and mates. Infanticide attacks occurred at 256 nests in two distinct contexts: (1) Attacks were primarily committed by nonbreeding pairs (69%) attempting to evict parents from the cavity. Infanticide attacks per nest were positively correlated with population size and evicting pairs never adopted abandoned offspring. Competition for limited nest sites was a primary cause of eviction-driven infanticide; and (2) Attacks occurred less frequently at nests where one mate died (31%), was perpetrated primarily by stepparents of both sexes, and was independent of population size. Thus, within a single species and mating system, infanticide occurred in multiple contexts due to multiple drivers. Nevertheless, 48% of stepparents of both sexes adopted offspring, and another 23% of stepfathers exhibited both infanticide and long-term care. Stepfathers were often young males who subsequently nested with widows, reaching earlier ages of first breeding than competitors and demonstrating sexually-selected adoption. Adoption and infanticide conferred similar fitness benefits to stepfathers and appeared to be equivalent strategies driven by limited breeding opportunities, male-biased sex ratios and long-term monogamy.

Methods
We studied green-rumped parrotlets from 1988-2015 at Hato Masaguaral (8º34' N, 67º35' W), a working cattle ranch 45 km south of Calabozo in the state of Guárico, Venezuela. In 1988 and 1989, 106 identical nest boxes made of 1-m deep polyvinyl chloride tubes with hardware cloth interiors were installed along fence lines in a 4 km2 study area. We monitored an additional 20 nest boxes in some years in a third population 2.5 km from the others. Across the 27-year period of this study (1989-2015), a team of 2-5 field researchers annually monitored the fate of all nesting attempts in boxes. We captured 1512 breeding and nonbreeding adults with mist nets. They were banded with individually identifiable color bands and a numbered metal band, as were all nestlings (n = 7346).

Nest boxes were visited daily every 1-3 days throughout the entire nesting season to determine nest contents and the identities of adults in the vicinity. Data collected from these checks, combined with 2,140 hours of nest watches and video recording in nests, provide strong evidence for the occurrence of infanticide and mate loss as well as adoption and nest abandonment. Parrotlets are unable to remove items from their meter-deep nest box because they use their bill and legs to ascend the nest cavity. Thus, nestlings or eggs killed by infanticide, or that died from other causes like starvation, remained in the nest box where we examined them during routine nest checks. Nestlings attacked by parrotlets usually exhibited bruising to the head, but also to the neck, back, wings and/or legs, with visible internal hemorrhaging and external lacerations. 

The obvious commotion at nest boxes from contests among parrotlets alerted us to the death of a parent (which was confirmed by resighting efforts during the current and subsequent years) and attempts to evict breeding pairs from nests. These nests were often watched for periods of 1-3 hours to identify the individuals present, note their social status, and record their behaviors, and thereafter were revisited regularly to determine the fate of the nest and the identity of the attending birds. Social status was classified as actively nesting male-female pairs, nonbreeding male-female pairs, unpaired males, and male-male pairs.

We classified stepparents as adopting unrelated offspring if no offspring mortality or injuries occurred from infanticide attacks, and stepparents exhibited forms of parental care, ranging from feeding of offspring to tolerance. Daily nest checks and periodic nest watches (765 hours at 95 nests) and video monitoring (325 hours at 4 nests) allowed us to observe whether stepparents exhibited adoption behaviors, and to determine whether parents abandoned the remaining eggs and/or nestlings before or after an infanticide or mate loss event. Nest watches and subsequent daily visits to the box determined whether stepparents entered the nest box or were prevented from doing so by the surviving parent, fed their new mates outside of the box, and participated with the widowed parent in defending the box from other parrotlets. Nests were classified as abandoned when multiple visits over several days determined that nestlings had not been fed (i.e., had empty crops and symptoms of dehydration or malnourishment), eggs were not incubated (i.e., were cold, unattended, or buried in the bottom of the box), parents and stepparents were not nearby, and other parrotlets were prospecting at the box or entering it.