A study was undertaken to evaluate the response of different test systems to preparations of heat-stable enterotoxin (ST) derived from Eschericihia coli strains recovered from diarrheal disease of humans, pigs and calves. Sterile broth culture supernatants of enterotoxigenic strains of E. coli were heated at 65 degrees C for 30 minutes and tested for the presence of heat-stable enterotoxin. Three test systems, namely, ligated intestine of weaned pigs, ligated intestine of rabbits and the infant mouse test were used in attempts to detect ST in the culture supernatants. Two patterns of reaction were observed in response to ST-containing preparations: either the preparation elicited a response in the three tests or the preparation elicited a reaction only in the ligated pig intestine. A response in all three tests were observed for 5/5 human ST-producing E. coli, 5/5 bovine enterotoxigenic E. coli, 5/5 "atypical" porcine enterotoxigenic E. coli, 3/3 St(+)LT(-) porcine E. coli of serogroup O138:K81 and 4/24 LT(+)ST(+) porcine E. coli. A response only in the ligated pig intestine was obtained with 5/5 ST(+)LT(-) porcine E. coli belonging to serogroups other than O138:K81 and to 20/24 ST(+)LT(+)E. coli from pigs. The results are consistent with the view that there are two kinds of ST, one of which (ST1) reacts in all three tests and the other (ST2) which reacts only in the ligated pig intestine. The findings underscore the limitations of the infant mouse test as a means of detecting ST in porcine isolates of E. coli, since the test fails to detect ST produced by a large number of these E. coli strains. There appeared to be a relationship between kind(s) of ST produced and the animal species from which the producing organism was recovered.