Behavioral Basis of Second Intermediate Host Specificity among Four Species of <i>Haematoloechus</i> (Digenea: Haematoloechidae)

Cercarial behavior patterns were examined in 4 species of frog lung flukes (Haematoloechus spp.). Cercariae of Haematoloechus complexus, Haematoloechus medioplexus, Haematoloechus longiplexus, and Haematoloechus varioplexus were exposed to 3 species of experimental arthropods and an inanimate control. The number of cercariae attached to an experimental host at 5 min postexposure was recorded. Haematoloechus longiplexus and H. complexus cercariae attached to experimental hosts at higher rates than cercariae of H. medioplexus and H. varioplexus. Cercariae of H. longiplexus attached to experimental hosts in approximately the same numbers as H. complexus, but H. longiplexus penetrated only damselfly naiads, and only at the base of the zygopteran caudal gills. Cercariae of H. complexus, a second intermediate host generalist, were able to penetrate and enter several arthropod species at the intersegmental membranes. Haematoloechus medioplexus and H. varioplexus are restricted to development in dragonfly naiads and cercariae rarely attached to and never penetrated experimental hosts. These behavioral patterns dictate the range of hosts suitable for metacercarial development of H. complexus, H. longiplexus, H. medioplexus, and H. varioplexus. The evolution of disparate patterns of behavior among the cercariae of these 4 congeners has directly affected subsequent patterns of transmission to the definitive host. Host specificity has traditionally been viewed as the result of a complex interaction between parasites and potential hosts. Investigations of this interaction have often focused on the biochemical and environmental requirements of the parasite. For example, urea has been implicated as a component essential to the maintenance of osmotic balance in the cestodes of elasmobranchs (Read et al., 1959). Other investigations concerning the nature of host specificity have examined the degree and effectiveness of immune response by the host. Adema et al. (1994) found that the hemocytes of snails susceptible to infection with larval Echinostoma paraensei suffered impaired function in the presence of this parasite. The hemocytes of resistant snails functioned normally. The role of parasite behavior in directly determining host specificity generally has been ignored, although many studies have examined the role of behavior in increasing the probability that parasite infective stages will come into contact with a potential host. Examples of the increased risk of predation of acanthocephalan-infected amphipods are well known, e.g., Bethel and Holmes (1977). Similarly a myriad of studies have demonstrated that the miracidia and cercariae of many trematode species exhibit behaviors that bring them into the preferred habitat of a potential host. Other cercariae respond directly to stimuli created by potential hosts. For example, Cryptocotyle lingua cercariae significantly increase activity in response to a passing shadow, thus enhancing transmission to excised pectoral fins designed to mimic a fish host (Rea and Irwin, 1991, 1992) and Schistosoma mansoni and Schistosoma haematobium cercariae respond to different chemical and thermal host cues and display different swimming behaviors (Haas et al., 1994). These differences were considered to represent adaptations to the disparate environmental conditions under which these 2 species of schistosome cercariae encounter their hosts. These examples demonstrate the importance of behavior in successful parasite transmission. However, they do not attempt to demonstrate that parasite behavior itself may determine host suitability. The present study examined the influence ofcercarial behavior on the second intermediate hosts available to 4 species of frog lung flukes (Haematoloechus spp.). Haematoloechus spp. are considered to develop typically in odonate hosts (Olsen, 1967). Reports of non-odonate hosts are rare and were limited to dipteran and plecopteran larvae (van Theil, 1930; Dollfus et al., 1960; Combes, 1968). However, Snyder and Janovy (1994) reported the development of Haematoloechus complexus in 2 non-odonate insects and 3 crustacean species as well as in odonates. In the same study, the development of Haematoloechus medioplexus was restricted to anisopteran (dragonfly) naiads. Preliminary investigations implicated cercarial behavior as a causative factor in determining second intermediate host specificity. The present study demonstrates that different cercarial behavioral patterns among 4 species of Haematoloechus directly determine second intermediate host specificity. Thus, host specificity among metacercariae of Haematoloechus species is, at least in part, a function of evolutionary divergence in behavioral patterns rather than present-day biochemical and immunological interactions between parasite and host. MATERIALS AND METHODS Host and parasite maintenance Adult flukes were dissected from wild-caught frogs and identified according to Kennedy (1981). With the exception of H. medioplexus, adult parasites were retrieved from several frog populations in Nebraska. Haematoloechus medioplexus adults were retrieved from frogs of a single population. Snails used were laboratory reared and infected using the techniques of Snyder and Janovy (1994). The planorbid snail Gyraulus parvus served as first intermediate host to Haematoloechus longiplexus, H. medioplexus, and Haematoloechus varioplexus. Haematoloechus complexus developed in the physid snails Physella virgata, Physella gyrina, and Physella heterostropha. Asellus intermedius were reared in the laboratory. All other arthropods used in the host specificity experiments were obtained from natural populations. To ensure that the animals from nature were not naturally infected, subsamples of 10 individuals per species were dissected and examined for metacercariae at the time of experimental exposure (To control). An additional subsample of 10 individuals per species remained unexposed and were held for the duration of the experiment. These ending time (T,) control groups were dissected and examined for Received 26 June 1995; revised 25 September 1995; accepted 25 September 1995.