Jeremy L. Marshall and Carlos D. Camp 



MATERIALS AND METHODS 



We collected 50 P. ocularis for stomach analysis from two 

 localities during May-July 1993. Both sites were ephemeral wetlands 

 in the lower Coastal Plain of Georgia. The first site, located in Grady 

 County, was dominated by black gum (Nyssa sylvatica) and was situ- 

 ated in a low pine flatwood having a canopy of slash pine (Pinus 

 elliottii) and an understory of saw palmetto (Serenoa repens) and 

 gallberry (Ilex glabra). The second, in Evans County, was a dome of 

 pond cypress (Taxodium ascendens) surrounded by sandhills domi- 

 nated by longleaf pine (P. palustris) and turkey oak (Quercus laevis). 

 Areas similar to both sites were described in detail by Wharton (1978). 



After collection, all specimens were preserved in 10% formalin, 

 and stored in 35% isopropanol. Each frog was measured for snout- 

 vent length (SVL) and dissected for stomach analysis. Individual food 

 items were counted and identified. Because prey items were too small 

 to use volumetric displacement, relative importance of prey was de- 

 termined by comparing each individual prey item to a paper grid and 

 visually estimating the number of grid squares occupied (Camp and 

 Bozeman 1981). 



Twenty of the Grady County frogs were collected between 2200 

 and 2400 EDT on 5 June. This sample consisted of 10 mature males 

 and 10 juveniles that had just completed metamorphosis. We used 

 this sample to make comparisons between feeding of adults and juve- 

 niles. Because we did not independently test for prey availability, 

 other collections were not used for comparisons because of possible 

 complications arising from temporal or between-site differences in 

 available prey items. In addition, although adult females were in- 

 cluded in these samples, small numbers (n = 4) precluded between- 

 sex comparisons. Correlation between the number of prey items eaten 

 and body size was tested using the procedure described by Zar 

 (1984). A comparison of diversity between adult male and juvenile 

 prey species was made using the Shannon-Wiener Index of Diversity 

 (H') (Zar 1984). 



RESULTS 



One hundred-forty individual prey items were identified and 

 consisted entirely of arthropods, mainly insects (Table 1). Springtails 

 (Collembola) were the most numerous group, making up 47% of the 

 food items eaten and found in 56% of the stomachs. Because they are 

 so small, however, they contributed less than 20% of the area occu- 

 pied by all prey items. Hymenopterans, especially ants (Formicidae) 



