THE ROLE OF REVERSED SELECTION 123 



what appears to be clear evidence that reversed selection has 

 caused real retrogression. Thus certain blind crabs inhabiting 

 caverns have lost their eyes, but not the stalks on which the eyes 

 were carried ; since the eye is a more ancient organ than the stalk, 

 it should, on cessation of selection, have been more persistent ; 

 yet it disappeared first Evidently something more than mere 

 cessation of selection was at work. In utter darkness the eye, a 

 prominent and vulnerable organ, would be not only useless, but 

 worse than useless ; and therefore reverted individuals would be 

 actively selected for survival, and the organ would be more rapidly 

 eliminated than could happen under cessation of selection. In 

 other cave-dwelling animals, which, as we suppose, are not 

 descended from stalk-eyed ancestors, the rudimentary eye is 

 covered by skin. Presumably in this case the partial retrogression 

 has resulted from mere cessation of selection, the eye being pro- 

 tected from injury and therefore from reversed selection by the 

 progression which resulted in the growth of the skin. Here, as 

 ever, nature has followed the line of least resistance. 



204. But though reversed selection is seldom a cause of retro- 

 gression, it has nevertheless a very important function. Reversed 

 selection implies reversed progression of the race, which, in turn, 

 implies reversed development of the individual. When, therefore, 

 we see a structure better developed in the immature individual 

 than in the adult, we may be sure that the diminution in the adult 

 is due to reversed selection. Thus the remote ancestor of the 

 modern horse, the Hipparion, had three functional toes. The 

 embryo of the horse also has three toes of considerable size. But 

 the two outer toes in each limb, which in the embryo are nearly as 

 well developed as in the Hipparion, degenerate partially during 

 development, so that the horse is born with only one functional 

 toe. " Occasionally a foal is born with two hoofs on one or more 

 of its limbs ; at very long intervals a foal appears with three hoofs 

 on one or more of its limbs." Now, when the outer toe persists in 

 the individual, we have plainly, in a real sense, an arrest of 

 development. The toes remain in the embryonic, the ancestral 

 condition ; the whole life-history is not recapitulated. But in the 

 normal horse there is no arrest of development ; the recapitula- 

 tion is carried out to its fullest extent, though in the later stages 

 in a reversed direction. The toes, which were useful to the 

 ancestors, are harmless to the embryo, but would be harmful to 

 the adult, and, there .re. ' . e been eliminated by reversed selec- 

 tion in the latter before retrogression has eliminated them in the 



