262 



one pair and two pairs. The evolution of a third 

 pair of balancers would mean an increase in the 

 functional problem of correlation without any 

 corresponding advantage in the way of efficiency. 

 John Burroughs writes somewhere about the gain 

 it would be if we could open one pair of eyes after 

 another and thus see more of the wonder of Nature. 

 The probability is that we should see less, for the 

 difficulty of correlating impressions would be insur- 

 mountable. It is significant that those backboneless 

 animals that have many eyes have little vision, and 

 that the unpaired, median, upward-looking, pineal 

 eye which some backboned animals possessed has 

 not been retained as an eye in higher types. It 

 may be, then, that limits imposed by growth-neces- 

 sities, by the persistence of well-defined structure- 

 units, and by pre-existing organization may throw 

 some additional light on the fact of convergence. 



