TRACHEATES 221 



have been brought about on Lamarckian principles, we 

 must discard them in the case of all instincts. 



But we have already seen, in the second chapter, 

 that instincts present no difficulty to natural selection. 

 Instincts vary, just as the parts of the animal body 

 do, and so can be selected and improved. They are 

 adaptations, and are often only of use in the circum- 

 stances in which their possessors usually live ; they are, 

 therefore, relatively imperfect, as we should be bound 

 to expect in view of their origin by selection. A 

 cricket that saves itself in nature by digging swiftly 

 into the ground repeats the movement even on hard 

 gravel or on a glass plate, when it would do better to 

 run away. A bee stings a human being, whose skin 

 closes over the wound, and retains the sting with its 

 barb, which is fatal to the bee. Its sting is only 

 provided against its chief enemies, the insects, whose 

 coat remains open after the wound is inflicted, and lets 

 the sting out again. The imperfectness of these 

 instincts tells very clearly of their origin by natural 

 selection. 



Hence, though many scientists retain the Lamarckian 

 principle because a number of the characteristics of 

 animals seem to them inexplicable without it, we have 

 seen that these characteristics are also found in circum- 

 stances where the Lamarckian principle cannot be 

 admitted at all. We found this in the case of the 

 skeleton of insects and their co-adaptation. We now 

 find it also in their instincts. 



The Lamarckians, however, retain their theory ex- 

 clusively, though they know it cannot be proved. How 



