The experiments were performed at Animal Husbandry Experiment Station of the University of Tokyo to investigate the responses of pregnant and non-pregnant thoroughbred mares to visual stimulation with special reference to the changes in behavior and electro cardiogram (ECG), and to evaluate some aspects of the effects of pregnancy on the behavior in mares. Three pregnant and 2 non-pregnant mares were used in this study. Each mare was equipped with a telemeter system for ECG recording,' and ECG at their quiet state was recorded for 10 min in the stall (Quiet Period). Thereafter;, each mare was led to the experi mental field where a blue vinyl sheet covered life-sized silhouette of a horse drawn with black paint was set up, and her behavior was simultaneously recorded with ECG for 10 min using two time-lapse videorecorders (Prestimulation Period). ;After a 10-minute Prestimulation Period, the sheet cover was removed and the same recordings were conducted for 10 min (Stimulation Period). After the visual stimulation, all the mares approached the silhouette and explored it. At the first contact with the silhouette, all pregnant mares explored the head part of the silhouette, while non-pregnant mares began exploration from the parts other than the head. Moreover, the pregnant mares approached the silhouette in a round about way in a variety of gaits with some walking, some trotting and some cantering, while the non-preg nant mares made a beeline for the silhouette at a walking gait. In the unstimulated condition (Quiet Period and Prestimulation Period), the mean heart rates and mean algebraic sum mations of T-wave amplitude in pregnant mares tended to be higher than those in nonpregnant ones, but, in Quiet Period, significant difference was recognized between both group in mean heart rate levels. Immediately after the visual stimulation, heart rate increased in all mares but, the maximal heart rates in pregnant mares, tended to be higher than those in non-pregnant ones.