232 



Tnherltaiice in T/ioroKffhbrcd ITorses 



It will be secn at ouce that the mean re.sult for nien is sub.stantially in agreement with that 

 for horscs. \Vc can hardly supijoso grandparental coiTelation for thc characters in man and 

 horse dealt with to differ much from 'S. The mean value of grandparental correlation in 

 the horse for cases 1 to 4 drawu froni the modern studlx)oks is •3388 and for casea 5 to 8 drawu 

 from the older records •äSSS. I believe thi.s is duc to the closc rclationship of so many of 

 the horses in the carliest period of breeding. The dift'erence may also to some extcnt depend 

 on the somewhat rougher appreciation and de.scription of colour. If there be auy influence 

 of change of sex in the intensity of heredity for coat-colour in the horse, it is not manifested in 

 the abovc .statistics, which thus differ widely from those of eye-colour in man. 



TABLE I. 



Colts and Paternal Grandams. 



Paternal Grandams. 



