Harlequin Frogs Along a Tropical Montane Stream: Aggregation and the Risk of Predation by Frog-Eating Flies

Abstract

We observed larvae of the fly Notochaeta bufonivora (Sarcophagidae) feeding on harlequin frogs, Atelopus varius (Bufonidae), along a gallery forest stream in montane Costa Rica during the dry season. Harlequin frogs tended to aggregate in areas containing waterfall spray, and in these areas an individual's probability of attack by Notochaeta was greater than in areas lacking waterfall spray zones. To ascertain whether this increased risk to a frog resulted from its choosing an area with a waterfall spray zone or from being in an area of high frog density, we manipulated the density of frogs in a series of quadrats. The results of this experiment suggest that the increased risk of attack in the environs of waterfalls was a hazard associated with patch choice by the frogs rather than with aggregation per se. AGGREGATIONS OF ANIMALS have often been viewed in the context of predator avoidance (e.g., Hamilton 1971, Bertram 1978). However, because predation events in nature are usually brief and unpredictable in space and time, one can rarely assess the relationship between an individual's risk of predation and its proximity to conspecifics. We studied an extraordinary predator-prey interaction between the harlequin frog, A. varius (Bufonidae), and a frogeating fly, N. bufonivoraz (Sarcophagidae), along a montane stream in Costa Rica. Harlequin frogs, which are brightly variegated with yellow and black, contain the potent nerve poison tetrodotoxin in their integument (Kim et al. 1975). Although tetrodotoxin probably discourages most predators, or at least most predaceous vertebrates, Notochaeta larvae, which feed internally, appear unaffected by this chemical defense (Crump & Pounds 1985). Because a frog lives several days after attack by Notochaeta, the probability of detecting predation events is much higher than is usually the case for predator-prey systems involving anuran prey. This paper focuses on the relationship between aggregation behavior of harlequin frogs and risk of predation by these frog-eating flies. Male frogs and toads commonly exhibit a clumped pattern of dispersion within the breeding habitat (Whitney & Krebs 1975, Wells 1977), and this pattern may reflect a patchy distribution of resources such as suitable oviposition sites. On the other hand, a clumped pattern of dispersion within the habitat may also be attributable to the advantages of aggregation itself. In theory, being in a group may dilute an individual's probability of being eaten should a predator arrive (Hamilton 1971); this may be especially important for aposematic prey, as a naive predator may avoid the rest of the group after sampling one individual. Also, if females are more attracted to groups of males than to solitary displaying males, the average mating success may be higher for males in groups (Wells