We examined the recruitment rate of nestmates during prey capture in six arboreal ant species: Myrmicaria opaciventris, Platythyrea conradti, Crematogaster sp., Crematogaster clariventris, Tetramorium aculeatum, and Oecophylla longinoda. Termites and grasshoppers were used as prey. Prey were grouped by size: for termites, we presented ants with either workers or soldiers and for grasshoppers, we used small nymphs, mature nymphs and adults. After prey detection by a foraging ant, the total recruits present were counted every 30 s over the course of 15 min, and the mean numbers of workers counted at each time interval was computed and plotted for each prey type. We tested the effect of prey type, prey size and termite caste for each ant species on the number of nestmates recruited using ANOVA (GLM proc). Prey size, prey type and termite caste significantly influenced the mean number of nestmates following initial prey detection. We found that the observed recruitments were well explained by linear models for less arboreal (M. opaciventris) and primitive arboreal species (P. conradti), reflecting stochastic movement by individuals or stochastic collective recruitment. For specialized arboreal ant species, recruitment was well explained by exponential models, reflecting within-group recruitment, either with limited or with unlimited group sizes. Overall, T. aculeatum had the highest level of nestmate recruitment. Surprisingly, O. longinoda, known to be one of the most specialized arboreal ant species, fit a recruitment model more like that of either ground ants with some arboreal foraging habits or primitive arboreal species.