364 THE BIOLOGY OF INSECTS 



in the food of the caterpillars. Breeding experiments by 

 pairing a melanic male with a normal female and feeding 

 their caterpillars on untreated hawthorn resulted in a 

 generation of moths all normal in wing-colour. Three 

 pairs of these, inbred, gave rise to caterpillars which, fed 

 like their parents on untreated leaves, developed into ninety- 

 three moths of which seventy were normal and twenty- 

 three melanic. From these results it seems clear that the 

 moths' germ-cells had become so affected through the lead 

 salt eaten by the caterpillars that the melanism could be 

 inherited through two generations without repetition of the 

 inducing cause, and that the factor for this condition behaves 

 as a typical Mendelian recessive. Here we see, under the 

 stimulus of abnormal food, the origin and transmission of 

 a mutational character, which may be regarded as a case 

 of partial inheritance of an " acquired change," though the 

 sceptic may protest that it is not definitely an example of 

 " use-inheritance," because the induced body-characters 

 result from chemical action on the germ-plasm. The 

 extreme abnormaUty of the food warns us not to attach 

 undue importance to this interesting demonstration of 

 germinal susceptibility. The effect is comparable to the 

 heritable degeneration induced in some mammals after 

 alcoholic poisoning of the parents, and must be regarded 

 to some extent as pathological. In order to prove that 

 use-inheritance is an important factor in evolution it is 

 necessary to obtain inherited modifications by means of 

 influences that are unquestionably part of the natural 

 environment of normal living creatures. 



A nearer approach to this desired result may be found 

 in Harrison's interesting experiments on the inheritance 

 of the egg-laying instinct in sawflies as described in his 

 latest paper (1927). In strains of Pontania salicis, a 

 common species whose larvae live in galls on willows, it 

 was demonstrated that a certain species of plant is definitely 

 chosen by the female fly and " that the instincts guiding 

 the female in the choice of food-plant are germinally fixed." 

 A race of P. salicis thus attached to Salix anderso?iia?ia was 



