Parasites face a trade-off if the highest quality hosts are also most resistant to exploitation. For brood parasites, well-defended host nests may be both harder to parasitize and harder to predate, leading to better survival of parasitic chicks. This trade-off could be accentuated if brood-parasitic adaptations to reduce front-line defences of hosts, such as mimicry of hawks by Cuculus cuckoos, do not deter hosts which aggressively mob raptors. Here we investigate the costs and benefits to the African cuckoo (Cuculus gularis) of specializing on a highly aggressive host species, the fork-tailed drongo (Dicrurus adsimilis). Field experiments showed that drongos strongly attacked and mobbed both cuckoo and hawk models, implying that hawk mimicry does not deter front-line defences against African cuckoos. Attacks on cuckoo and hawk models generally declined after the egg stage but attacks on snake models sharply increased, suggesting drongos may treat hawks more like cuckoos than predators. We suggest that the cost to cuckoos of parasitizing an aggressive host may be alleviated by subsequent benefits to their offspring, since drongo nests survived better than nests of other species with similar nesting ecology. These results are indicative of a trade-off between host quality and susceptibility for a brood parasite.

We presented models of three species at drongo nests in 2019: an African cuckoo, a little sparrowhawk (Accipiter minullus), and a ring-necked dove (Streptopelia capicola). For each species we used two different models, which were 3D-printed in polylactic acid, and painted. In 2021, we repeated presentations of the hawk and dove models, alongside a rubber snake snake (Supkeyer 130 cm). The models were presented in a random order with an hour between trials. Prior to each trial, we set up a video recorder (Sony CX280 Full HD Handycam) in a blind 10 m from the nest, and tied the model to a branch approximately 2 m from the nest. Trials were filmed and lasted for 5 minutes, beginning when a drongo came within 5 m of the model.

We quantified mobbing dives towards each model from video recordings. Mobs were defined as close swoops accompanied by a short sound, and sometimes involved direct contact with the model . As it was not possible to count blind to treatment type, mobs in a subset of videos were independently counted by another observer without prior knowledge of the experiment.

We also checked nests of fork-tailed drongos, dark-capped bulbuls (Pycnonotus tricolor) and tchagra bush-shrikes (Tchagra australis and Tchagra senegalus) for predation every three days, +/- 1 day.

Methods for processing the data: All analyses were carried out in R 4.1.3, see RMarkdown document for details

Excel csv files and a html document.