Fifteen Thoroughbred racehorses were exercised at the walk, trot and c
anter three times a week for seven weeks on an inclined treadmill. Str
ide rates of each horse were measured at the trot and canter during ea
ch exercise period. At both gaits, stride rates were consistent betwee
n trials for each horse and a general linear relationship between stri
de frequencies at the trot and canter occurred. Stride rates outside o
f the 95% confidence level occurred in only 17 of the 522 trials and,
of these, gait abnormalities could be identified in only five of these
trials. Three horses exhibited signs of dorsal displacement of the so
ft palate during some of the trials. No effects of this condition on s
tride frequencies were observed. The consistency of the stride rates a
t the trot and canter suggests that individual horses have a preferred
stride frequency at a given velocity and gait. Horses that have high
ratios between stride frequencies at the trot and canter would tend to
have longer stride lengths at the canter at a given speed. Since maxi
mum velocity is directly related to the length of stride, horses with
longer strides should have higher peak speeds than those horses with s
horter strides. Many other factors must be considered in any predictio
ns of performance potential.