PREDICTING THE PERSISTENCE OF AMPHIBIAN POPULATIONS WITH THE HELP OF A SPATIAL MODEL

1. We have used a stochastic population model with immigration to calc ulate extinction rates for two amphibian species: the common toad, Buf o bufo, and the crested newt, Triturus cristatus. 2. Population models for amphibians are difficult to parameterize. It was not possible to measure all the relevant parameters for the model. The other parameter s were obtained from other studies or estimated on the basis of simila r species. 3. Contour maps for persistence were generated using the mo del. The persistence of populations associated with a pond is affected both by the pond's internal dynamics and its proximity to a source po nd. On a 20-generation time scale, toad ponds that were initially occu pied were relatively unaffected by the proximity of a source. Persiste nce was almost certain (>95%) when the average carrying capacity of a pond was greater than 30 adult females. Initially unoccupied toad-pond s were equally likely to persist if they lay within 4 km of a source a nd could support more than 50 adult females. Initially occupied newt p onds were likely to persist if they supported more than 40 females or lay within 0.5 km of a typical source pond. Initially unoccupied newt ponds were mainly dependent upon immigration rates. Very small ponds ( N-k <10) could persist if they lay within 0.75 km of a source pond, wh ile large ponds (N-k much greater than 10) only needed to be within 1. 5 km of a source pond. 4. This model allows us to use the data availab le, to make predictions about the criteria which must be met by a land scape to ensure the survival of amphibian populations. It offers the p ossibility of better predictions when the data is refined. The model a lso suggests directions for further research such as the statistical s tudy of environmental variation, the nature of minimum viable populati ons for toads and the statistics of long-range dispersal.