66 THE PROBLEM OF THE GALLOPING HORSE 



as it were, fused in our visual impression, because each 

 picture lasts on the retina of the eye for one tenth of a 

 second, or (to put it more accurately) because the " im- 

 pression " or condition of the retina produced by each 

 picture persists or endures for the tenth of a second. 



It may, perhaps, be suggested (and, indeed, has been), 

 that it is the " blurred " or " fused " picture produced by 

 the successive poses of the galloping horse's legs in one 

 tenth of a second that the painter ought to imitate on his 

 canvas. In support of this notion we have the fact that 

 the rapidly running wheels of a coach or of a gun-carriage 

 (as in the pictures by Wouwerman) are represented by 

 artists, not with the twelve or fourteen spokes which we 

 know to be there — and would be photographed as separate 

 things in an exposure of the fortieth of a second — but 

 as a blurred haze of some fifty or more indistinct 

 " spokes." In this case it undoubtedly results that the 

 observer of the picture is satisfied and receives the mental 

 impression or illusion of a rapid rotation of the wheel. I 

 have tried the experiment with instantaneous photographs 

 of the galloping horse, and I get three results : first, no 

 combination of successive phases occupying one tenth of 

 a second gives anything resembling the " flying gallop " 

 of the racing plates (the Mycenaean and Stubbsian pose), 

 or any other conventional pose ; second, no combination 

 of successive instantaneous photographs limited to ten 

 seconds gives any pose which satisfies the judgment and 

 suggests a movement like the gallop ; third, the combina- 

 tion which comes nearest to satisfying the judgment as 

 being a natural appearance, but does not quite succeed in 

 doing so, is one formed by the fusion of figs. 2 and 3 01 

 PI. I. This gives all four legs off the ground, drawn up 

 or flexed beneath the horse's body, as in Morot's picture 

 of the sabre-charge at Resonville. 



The fact is that we have to take into consideration two 



