Current U.S. and Japanese regulatory guidelines mandate the use of at least six animals for dermal and ocular irritation testing, while EEC/OECD guidelines require only three animals. An analysis of data from this laboratory in general supports the use of as few as three animals for acute irritancy testing. Data from a series of 140 dermal and 118 ocular irritation tests were examined to determine injury rates associated with testing, and to generate six-, five-, four-, three-, and two-rabbit primary dermal irritation indices for dermal tests and to categorize ocular tests similarly. Results were then examined to determine whether fewer animals could have been used to obtain findings similar to those originally noted with six rabbits. These results suggest that the number of animals used in dermal and ocular irritation testing can be reduced, especially if a sequential approach involving a preliminary in vitro/SAR test is taken. In addition, low rates of severe injury were evident for dermal and ocular irritation tests, as well as low discomfort rates for ocular irritation tests using a topical anesthetic. Using a low level of dermal irritation as a screen for ocular testing could reduce injury and discomfort rates further.