On initial evaluation for onset of seizure disorders at nonreferral veterinary practices, 50 previously healthy dogs were enrolled in a study to determine the probability of identifying a specific cause for the seizures. Treatment was not administered prior to entry of dogs in the study. On the basis of antemortem and postmortem test results, 22 dogs (44%) were classified as having primary epileptic seizures (PES; idiopathic or without identifiable cause), 23 (46%) had secondary epileptic seizures (SES; identifiable intracranial cause), and 5 (10%) had reactive epileptic seizures (RES; metabolic or transient noxious cause). Forty-one dogs (82%) had 2 or more seizures before evaluation, with 37 (90%) of these dogs classified as having epilepsy on the basis of an underlying chronic brain disorder. For these 41 dogs, 17 (41%) had PES, 20 (49%) had SES, and 4 (10%) had RES. Among the 9 dogs (18%) with nonrecurring seizures, 5 had PES, 3 had SES, and 1 had RES. Generalized seizures were the most common first-observed seizure type associated with all etiologic classifications in all dogs with recurring and nonrecurring seizures. Diagnosis of SES was statistically more probable when the dog was less than 1 or more than 7 years old at the first seizure, when the first seizure was a partial seizure, or when the first seizure occurred between midnight and 8 AM. A diagnosis of RES was statistically more probable only when the interval between the first and second seizure was brief (< or = 4 weeks).(ABSTRACT TRUNCATED AT 250 WORDS)