3 o8 THE VARIATION OF ANIMALS IN NATURE 



On this basis it can be shown that the chance that variation 

 will lead to an improvement depends on the magnitude of the 

 change and will approach one-half as the latter becomes small. 

 Thus it is always possible for random variation to increase 

 adaptation and, provided the change is small enough, the 

 chances of improvement or the reverse are nearly equal. In 

 a static environment and a stable organism this reasoning 

 would appear to be incontrovertible. But in nature the 

 individual is the only stable unit. Species are complex aggre- 

 gates of numerous strains. The environment is constant only 

 in its tendency to fluctuations and is pulling the organism in 

 different directions in quick succession. The small variations, 

 such as may lead to improvement in a complex organ, must 

 usually confer only a very small advantage on the variant 

 individuals. It is thus highly probable that the new variant 

 will die out before it has had time to spread. We cannot 

 prove that complex organs have not developed by means of 

 Natural Selection, but we can see that the process will be very 

 slow and we may even doubt if geological time has been 

 sufficiently long. In our chapter on Adaptation we discuss 

 the phenomenon of organisation, the most characteristic 

 attribute of living animals. It may be suggested that com- 

 plex organs are only a special instance of that process (cf. 

 Chapter IX). 



It is not quite the same with the problem of what Cuenot 

 (1925) has called co-adaptations. This has been discussed in 

 a very judicious way by Wheeler (1928, pp. 29-33), and Corset 

 (1931) has illustrated a long series of examples in a very 

 thorough monograph. These co-adaptations may be described 

 as complex organs in which the co-ordination between the 

 parts is not physiological but merely mechanical, like the 

 relation between the blade and sheath of a penknife or the 

 button and the button-hole. For example (see discussion, 

 Robson, 1932), a button-like structure is actually known 

 in some of the Cephalopoda, in which the mantle is held 

 closed by a knob on one side fitting tightly into a socket 

 on the other. An interesting example dealt with at some 

 length by Wheeler is the development of ' scrobes ' or grooves 

 for the reception of the antennae in various insects. In 

 ants these grooves are on the head and may run below or 

 above the eyes, and they may have two divisions, one for 



