In a survey of 900 Dalmatian dogs, brainstem auditory-evoked responses (BAER) and clinical observations were used to determine the incidence and sex distribution of bilateral and unilateral BAER abnormalities and their association with heterochromia iridis (HI). To assess the efficacy of BAER testing in guiding breeding programs, data from 749 dogs (subgroup A), considered to be a sample of the population at large, were compared with data from a subgroup (subgroup B; n = 151) in which selection of breeding stock had been based on BAER testing from the beginning of the 4-year survey. Brainstem auditory-evoked responses were elicited by applying click stimuli unilaterally, while applying a white noise masking sound to the contralateral ear. Under these conditions, BAER were either normal, unilaterally absent, or bilaterally absent. Dogs with bilaterally absent BAER were clinically deaf; dogs with unilaterally absent BAER were not clinically deaf but appeared dependent on their BAER-normal ears for their auditory-cued behavior. Dogs with unilaterally absent BAER often were misidentified as normal by uninformed observers. Among the 900 dogs, 648 (72.0%) were normal, 189 (21.0%) had unilateral absence of BAER, and 63 (7.0%) had bilateral absence of BAER or were clinically deaf and assumed to have bilaterally absent BAER (n = 4). Total incidence in the population sampled was assumed to be higher, because some bilaterally affected dogs that would have been members of subgroup A undoubtedly did not come to our attention. Among females, 24.0% were unilaterally abnormal and 8.2% were bilaterally abnormal whereas, among males, 17.8% were unilaterally abnormal and 5.7% were bilaterally abnormal.(ABSTRACT TRUNCATED AT 250 WORDS)