452 



Noii-iuirivc Species — Oki Living Resources 



Bullfrogs: 

 Introduced 

 Predators in 

 Southwestern 

 Wetlands 



by 



Philip C. Rosen 



University of Arizona 



Cecil R. Schwalbe 



National Biological Service 



Fig. 1. The worm has turned! In 

 this unstaged photograph taken at 

 Parker Canyon Lake, Cochise 

 County, Arizona, 1964, an intro- 

 duced bullfrog is swallowing a 

 Mexican garter snake, normally a 

 frog-eating species. Such preda- 

 tion appears to be destroying 

 remaining populations of this 

 garter snake in the United States. 



In the American Southwest, much oi the native 

 fish fauna is facing extinction (Minckley and 

 Deacon 1991); frogs in California (Fellers and 

 Drost 1993) and frogs and garter snakes in 

 Arizona (Schwalbe and Rosen 1988) are also in 

 critical decline. Habitat destruction and intro- 

 duced predators appear to be primary causes of 

 native frog declines (Jennings and Hayes 1994), 

 and habitat modification often yields ponds and 

 lakes especially suitable for introduced species. 

 Introduced bullfrogs {Rami catesbeicma) have 

 been blamed for amphibian declines in much of 

 western North America (e.g., Hayes and 

 Jennings 1986: Leonard et al. 1993; Vial and 

 Saylor 1993). E.xtensive cannibalism by bull- 

 frogs renders them especially potent predators 

 at the population level. The tadpoles require 

 only perennial water and grazeable plant mater- 

 ial; hence, transforming young can sustain a 

 dense adult bullfrog population even if alternate 

 prey are depleted. This may increase the proba- 

 bility that native species may be extirpated by 

 bullfrog predation. 



Introduced predatory fishes are apparently an 

 important cause of frog declines (Hayes and 

 Jennings 1986). They have been strongly impli- 

 cated in one important case of decline of native 

 ranid frog (family Ranidae, the "true"" frogs; 

 Bradford 1989). Some introduced crayfish may 

 also be devastating in some areas (Jennings and 

 Hayes 1994). In our study region, however, nei- 

 ther introduced fishes nor crayfish are dominant. 

 We present results that sustain a "bullfrog 

 hypothesis" for some native ranid declines, and 

 we present our study as an example of how evi- 

 dence accumulates to support such a hypothesis. 

 In 1985 we began documenting historical 

 localities for wetland herpetofaunas (reptiles 

 and amphibians), based on museum records and 

 personal interviews, then revisited these and 

 additional areas to detennine current species" 

 status. Results of this process, plus circumstan- 

 Ual evidence, suggested that the bullfrog was a 

 primary cause for declines of leopard frogs and 

 garter snakes in southern Arizona (Schwalbe 

 and Rosen 1988). 



In 1986-89 and 1992-93 we conducted 

 removal censuses of bullfrogs at San 

 Bernardino National Wildlife Refuge 

 (SBNWR), Cochise County. Arizona. We 

 simultaneously monitored native Chiricahua 

 leopard frogs [R. cliiiicalniensis) and Mexican 

 garter snakes (Tluiinuophis eqiies) at the sites of 

 bullfrog removal. A control site, with no bull- 

 frog removal, was established in comparable 

 habitat at Buenos Aires National Wildlife 

 Refuge (BANWR), Pima County, Arizona. 



Evidence for Bullfrog Effects 



Bullfrogs ate garter snakes, including 

 Mexican garter snakes (Fig. 1), as well as 

 numerous frogs, including young bullfrogs and 

 the last observed leopard frogs on our intensive 

 study areas. In addition, these frogs ate other 

 frogs and snakes, lizards, fish, birds, and mam- 

 mals in addition to many invertebrates (see also 

 Bury and Whelan 1984). 



We currently know of no examples of over- 

 lap between populations of the native leopard 

 frogs R. chiricalnieiisis and R. yiirapaiensis and 

 bullfrogs in southern Arizona. Leopard frogs 

 were abundant at both SBNWR and BANWR 

 before bullfrog proliferation, and as recently as 

 1981, bullfrogs and leopard frogs were both still 

 widespread at SBNWR (D. Lanning, The 

 Arizona Nature Conservancy, unpublished 

 data). Leopard frogs apparently were extirpated 

 from our SBNWR^study area by 1989, 



In 1993-94 relict populations of Chiricahua 

 leopard frogs (2-20 adults each) were found 5, 

 10, and 19 km (3.1, 6.2, and 11.8 mi) east of 

 SBNWR. These populations are in areas not 

 occupied by bullfrogs in habitats that may dry 

 too frequently for non-native predators (person- 

 al observations), as seen in native frogs of the 

 central valley of California (Hayes and Jennings 

 1988). These recent findings near SBNWR fur- 

 ther support the bullfrog hypothesis in south- 

 eastern Arizona. 



Checkered garter snakes {Tliamnophis mar- 

 ciainis) are semi-terrestrial and coexist in abun- 

 dance with bullfrogs. The highly aquatic 

 Mexican garter snake, however, has only small, 

 apparently declining populations where its 

 habitat overlaps with that of bullfrogs. Because 

 the bullfrog is also highly aquatic, its effects on 

 the Mexican garter snake have been greater than 

 on the checkered. 



Although Mexican garter snakes do repro- 

 duce where they occur with bullfrogs, few 

 young survive (Fig. 2). Once the young snakes 

 outgrow vulnerability to bullfrog predation, 

 they survive well; young adults marked in 1986- 

 88 have been recovered at ages 7-10 in 1993, 

 equaling and exceeding known ages for garter 



