Analysis of a 35 year data base on frogs from species assemblages in southeastern Brazil, suggests three kinds of species changes: the appearance of Hyla albopunctata at Boraceia; the possible extinction of Fritziana ohausi at Boraceia when the bamboo species in which Fritziana reproduces flowered and died; and population reductions and extinctions of a number of species at several localities in southeastern Brazil in 1979. We argue that the cause of this latter event was most likely due to unusually heavy frosts. The unusual event in 1979 had differential effects on the species which resulted in both short-term ecological changes within the assemblages effected and potential long-term evolutionary effects on the assemblages and individual species involved. SHORT-TERM CLIMATIC EVENTS ARE NOT usually considered to have strong evolutionary consequences on biological populations. However, through study of frog collections made over the last 35 years at Boraceia, Brazil, together with our field observations, we are able to document dramatic short-term changes in the structure of certain herpetofaunal assemblages in Southeast Brazil that may have evolutionary consequences. We postulate that a dramatic change took place in 1979, which due to an unusual climatic event, was regional in scope. Boraceia is a field station run by the Museu de Zoologia da Universidade de Sao Paulo. The station is located in the forested watershed of the city of Sao Paulo at approximately 2 3?38'S, 45?50'W on the edge of the coastal escarpment at an elevation of about 850 m (Fig. 1). It is in the Atlantic Forest Morphoclimatic Domain; morphoclimatic domains are defined geomorphologically by superposition of features of relief, soils, climate, and vegetation (Ab'Saber 1977). The first collections of frogs from Boraceia date from the late 1940s. A description of this frog fauna is being published (Heyer et al., in press). Long term changes of species assemblages have never been reported for a diverse herpetofauna, and such data on single species are limited to a few studies, e.g., Beebee's (1983) work on Bufo calamita, Sweet's report on Rana boylii (pers. comm.-paper presented at the joint meeting of the Herpetologists' League and Society of Amphibians and Reptiles, 7-12 August 1983), and Tyler and Davies' (1985) report on Rheobatrachus silus. As these kinds of data are difficult to document, it is important to indicate the nature of the data base. The frog fauna of Boraceia consists of at least 63 species. Some of these species are encountered rarely, and although we expect that the effects we report pertain, as well, to these rarely encountered species at Boraceia, we cannot document them. We report only on common, conspicuous species, for which we are certain that we can ascertain a change in abundance. One possible sampling problem might be collecting techniques, by which different collectors visiting in different years capture some taxa more efficiently than others. This potential problem is avoided in this case, as one of us (Rand) collected extensively at Boraceia in 1963 and again in 1983 and in 1984. Another possible sampling problem relates to weather conditions. We have found frog activity at Boraceia limited below 15iC and normal above 15iC during moist or wet conditions. The data used for this paper were gathered when weather conditions were good to excellent for frog activity. The data base consists of field observations and data associated with museum specimens. At Boraceia, we have drawn on the field notes of James P. Bogart, Gene Miller, Charles W. Myers, Craig E. Nelson, and G. Bruce Williamson (representing almost 7 months of field time from 1969-1979) as well as our own field notes (representing about 4 months of field work from 1962-1985). We have examined data for 2970 museum specimens from Boraceia. We also include some observations from other localities in southeastern Brazil (Fig. 1). The herpetology ' Received 29July 1986, revised and accepted 17 February 1987. 230 BIOTROPICA 20(3): 230-235 1988 This content downloaded from 207.46.13.128 on Wed, 07 Sep 2016 05:13:07 UTC All use subject to http://about.jstor.org/terms FIGURE 1. South America with Atlantic Forest Morphoclimatic Domain in black. Insert (bounded by 150 on North, 500 on West) showing Brazilian localities mentioned in text: (1) Santa Teresa, Espirito Santo; (2) Teres6polis, Rio de Janeiro; (3) Itatiaia National Park, Rio de Janeiro; (4) Brejo da Lapa, limits of Rio de Janeiro and Minas Gerais; (5) Sao Sebastiao, Sao Paulo; (6) Boraceia, Sao Paulo; (7) Paranapiacaba, Sao Paulo; (8) Santos, Sao Paulo; (9) type locality of Cycloramphus duseni, Parana. group from the Universidade Federal Rural de Rio de Janeiro (including Cruz and Peixoto) has made frequent field trips annually to Teresopolis (State of Rio de Janeiro) from 1960 to 1985 and Santa Teresa (State of Espirito Santo) from 1974 to 1986.