J. W. H. Harrison 249 



(b) Inheritance within the Limits of the Local Races. 



When I first conceived the notion that the autumnata races attached 

 to the Wilton pinewood and the Lonsdale alderwood had been derived 

 from that found in open birchwoods or alders in Cleveland and elsewhere, 

 it became of paramount importance to determine whether the main 

 birch types linked up by the apparently continuous fluctuating variation 

 were inherited or not ; whether any given female either gave or tended 

 to give the whole range of variation ; or, to state the problem once 

 more, whether there were genetically diverse strains existing within the 

 birch wood population. To answer my own question I captured three 

 wild females, one a beautiful, clearly lined specimen with a pure white 

 central area, a second of the strikingly barred sandbergi form and the 

 third a dull expressionless grey with obliterated markings, and reared 

 their progeny. In all cases the offspring leant most unmistakably to 

 their mother although some little approach toward other forms existed 

 — a transiticm slightly more noticeable in the brood from the barred 

 female. From pairs selected to be as nearly like their parents as possible 

 further broods were obtained with a repetition of the previous year's 

 results save that fluctuating variation seemed more steadied. After a 

 further generation producing evidence of the same value the experiments 

 were discontinued save in one instance. Reciprocal crosses between the 

 insect with the white ground and melanochroic forms of the species 

 0. dilutata were made ; in both crosses the hereditary nature of the 

 parental condition was perfectly obvious, the peculiar autumnata character 

 proving almost dominant. Further evidence to the same effect was 

 obtained in the experiment in sex-linked inheritance described below, 

 where the male employed was of the weaker sandbergi form ; that its 

 characteristics are germinally impressed the condition of the F2 genera- 

 tion emphatically proves. 



Thus we have abundant proof that the population of the birchwood 

 is mixed, being composed of strains genetically diverse and therefore 

 capable of isolation when submitted either to artificial or to natural 

 selection. 



(c) The Inheritance of Melanism. 



My experiments in this direction were very uninspiring as regards 

 their indications. I paired a melanic Middlesbrough male with a white 

 Irish female and reared the result of the cross to the F3 generation. 



