Objective To quantitatively and qualitatively compare electroretinography (ERG) recordings in awake, sedated, and anesthetized dogs. Animals Six 6-month-old Beagles. Procedures-A brief ERG protocol for dogs was used. Following 1-minute and subsequent 5-minute dark adaptation, mixed rod-cone responses were recorded bilaterally with a handheld multispecies ERG device with dogs in each of 3 states of consciousness: awake, sedated (dexmedetomidine and butorphanol), and anesthetized (atropine and hydromor-phone, followed by propofol and midazolam and anesthetic maintenance with isoflurane). Low- and high-frequency noise levels were quantified via Fourier analysis, and the effect of consciousness state on signal amplitude, implicit time, and noise was analyzed via repeated-measures ANOVA. In addition, 13 veterinary ophthalmologists who were unaware of the dogs' consciousness states subjectively graded the ERG recording quality, and scores for each tracing were compared. Results-ERG amplitudes were highest in awake dogs and lowest in anesthetized dogs. Implicit times were shortest in awake dogs and longest in anesthetized dogs. Differences in b-wave amplitudes and a-wave implicit times were significant. Neither low- nor high-frequency noise levels differed significantly among consciousness states. Furthermore, no significant differences were identified among observers' scores assigned to ERG tracings. Conclusions and Clinical Relevance-Anesthesia and sedation resulted in significant attenuation and delay of ERG responses in dogs. Chemical restraint of dogs had no consistently significant effect on low- or high-frequency noise levels or on observer perception of signal quality.