BIOPSYCHOLOGICAL 623 



wanes, the cock rhea becomes a careful and courageous father, not 

 to be separated from his offspring, to whom he is as good as a 

 mother. How limited, and yet how high! 



For obvious reasons there is not much parental care among fishes; 

 yet it is sometimes exhibited in a high degree, as in Sticklebacks and 

 Bubble-fishes, especially when the number of eggs is relatively small 

 and when there are many chances of death in early life. For weeks 

 the fish — usually the male — may guard the developing eggs, fasting 

 all the time; and surprise has been expressed that this should some- 

 times end, in aquarium conditions at least, in the offspring falling 

 victim to the parental appetite. Here, again, there is apt to be 

 misunderstanding, for the instinct to seize a rapidly moving object, 

 after being inhibited for a period by a strong parental instinct, not 

 unnaturally reawakens when the cradle empties and the parental 

 instinct dies away. In natural conditions there is usually a rapid 

 scattering of the progeny. 



Very instructive, again, are the experiments made by Mayer and 

 Soule on the caterpillars of the common milkweed butterfly (Danais 

 plexippus). When the caterpillars have started eating, they may be 

 induced to accept leaves which they would never have begun with. 

 "The momentum of the reaction" — to munch — carries them on, 

 though the proper taste and smell stimulus, without which they will 

 not begin, is absent. The observers note, however, that the tyranny 

 of instinct is not absolutely rigorous. If a "distasteful" leaf is pre- 

 sented at intervals of about thirty seconds, the caterpillar takes 

 fewer and fewer bites, and then refuses altogether. It is beginning 

 to learn. Yet if the distasteful leaf is offered at intervals of a minute 

 and a half, the caterpillar tries it every time and takes about the 

 same number of bites. Its associative memory is very short. Yet 

 there is a hint here that instinctive predisposition may serve as the 

 firm foundation for an improved superstructure of behaviour, the 

 outcome of experience. But the paradox seems to be that the firmer 

 the foundation of instinct, the more difficult it is to build on intelli- 

 gently. Yet some of the instances we have given, such as that of the 

 rhea, and the milkweed caterpillars, seem to us to show that the 

 eking out of instinct by intelligence is undeniable. 



In emphasising the limitedness of instincts — a familiar fact to the 

 field-naturaUst — we must not fall into the error of thinking of them 

 as rigidly stereotyped. Their variability has largely stopped, because 

 so perfectly adaptive, but it has not necessarily stopped. Variations 

 which express themselves in an improvement of the nervous system, 

 such as those which deepened a convolution in the Carnivore's 

 cerebrum, or increased the size of the neopallium in the early 

 Primates, cannot have been without their counterparts in the 

 evolution of instincts. Apart from inferring this from the gradations 

 that occur in different species and genera of social insects, we have 



