Habitat selection behavior in three species of anuran larvae : environmental cues, ontogeny, and adaptive significance

Habitat selection behavior was investigated as a proximate mechanism influencing spatial distributions of three species of anuran larvae. Rana cascadae, Hyla regilla, and Bufo boreas overlap in distribution in lakes and ponds of the Cascade Mts. and were chosen to provide a comparative basis for analyzing habitat selection behavior. Major objectives of the study were to define the proximate cues used by tadpoles in selecting a habitat, describe the ontogenetic basis of cue preference, and delineate the possible selective constraints acting on the habitat selection process. At one study site (Little Three Creek Lake), tadpole distributions were quantitatively assessed relative to habitat variab1es (shore distance, temperature, dissolved oxygen, depth, substrate type, and macrovegetation) throughout a summer. These data were subjected to correlation analysis and multiple linear regression analysis. Significant associations for each species were with temperature, depth, substrate, vegetational patterns, and shore distance. Tadpoles were most frequent1y found in protected areas along shallow gently sloping shorelines, over fine as opposed to coarse substrate types, in high temperature areas, and often amongst vegetation (Carex and Isoetes) (observations consistent at over 25 other sites surveyed). Larvae of each species selected similar habitat types (overlap was high) but B. boreas was more of habitat generalist due to high dispersal tendencies, schooling behavior, and midwater feeding habits. R. cascadae and H. regilla larvae tended to remain in the vicinity of the oviposition sites which were sometimes patchily distributed. Mark and census experiments showed that dispersal was limited in these species, certain features (e.g. substrate type, deep or cold water) represented barriers to dispersal, tadpoles tended to disperse more from nonpreferred habitats than preferred habitats, and disappearance rates were greater in nonpreferred habitats. Benthic sample analyses showed preferred substrates (fine particulates) contained greater amounts of total organics and chlorophyll (i.e., potential food sources) than coarse substrates which were avoided by tadpoles. Because such field studies do not distinguish correlated variables from those actually perceived and responded to by tadpoles, laboratory choice experiments were used. Tadpoles were reared over one of four substrates (smooth substrate, sand, gravel, or rock) and periodically tested for preferences during development. Larvae of each species (lab reared and wild tadpoles) preferred finer grained substrates regardless of rearing experience, age, or population source tested. Tadpoles preferred an open habitat to one with artificial vegetation that resembled Carex and all species exhibited a strong tendency to aggregate with conspecifics when tested in the laboratory. R. cascadae tadpoles preferred water depths at which they were reared and wild B. boreas larvae tended to prefer shallow depths. Stereotypic responses of R. cascadae were modified following rearing in a complex environment where a positive stimulus (high temperature or food) was associated with a nonpreferred substrate, and a "negative" stimulus (lower temperature or no food) was associated with a preferred substrate. Tadpoles came to prefer the previously nonpreferred substrate when later tested but these modified preferences were gradually lost when tadpoles were removed from the rearing regime. With the exception of vegetation preference tests, lab choice experiments corroborated field observations. Habitat selection seems to be a complex process in these tadpoles. More than one or two key cues are involved, cues responded to are of a structural, microclimatic, and biologic nature, responses to some cues were very rigid and to others modifiable by experience, and some adaptability in behavioral responses was evident. In the ephemeral breeding habitats of the high Cascades, traits that influence growth are likely to be closely linked to overall fitness. The habitat selection behaviors observed in these larvae (i.e., habitats and factors conducive to rapid growth) reflect an overall strategy to minimize the time needed for successful completion or metamorphosis by maximizing components of growth and development.