J. W. H. Harrison 225 



rapid change in average condition the insect has displayed in the 

 pinewood. 



But the change in the period of emergence is a matter much less 

 susceptible of explanation. It cannot be mutational for three reasons : 

 (1) it is universal in a wood of tremendous extent in which free com- 

 munication between the preferred larch zones is barred by belts of pure 

 pines, and, in one case, by a stretch of liardus stricta moor; (2) had it 

 been mutational no selective factor acting in favour of the new character 

 can be at work, for it presents no apparent advantage aiding selection ; 

 (3) it affects both of the germinal strains occurring in the wood. 



The second objection amongst these three also indicates that pure 

 selection, whether by the agency of birds, bats, arachnids or ichneumons, 

 acting on the race as a whole cannot explain the anomaly, even had I 

 not put the matter to test by trying by selection to set up an early 

 emerging strain from birchwood parents. Having secured most excep- 

 tionally a September birch female I bred from her earliest progeny in the 

 succeeding year and continued the procedure for three generations, with 

 the result that the last brood reared emerged on the average a fortnight 

 later than birch broods from wild parents reared alongside them. The 

 experiment was therefore discontinued. Even had it been a success 

 artificially, and it is conceivable that it would have been had I continued 

 for a hundred (!) years, in a state of nature, unlike what occurred in 

 the evolution of 0. filigrammaria, no advantage accrues from early 

 emerging, so nothing presses the balance down in that direction. On 

 the contrary a positive disadvantage manifests itself, for early laid ova 

 of 0. autumnata have been known to hatch the same year ; larvae from 

 these must perish when the trees shed their leaves. 



We are thus forced to cast about for explanation in other directions. 

 That we must look to causes operating within the pinewood, and there- 

 fore environmental, seems certain, for parts of the wood are as far apart 

 as they are from the birchwood, and yet throughout the early emergence 

 obtains. 



Neglecting genuine pine and larch feeding lepidoptera, the most 

 prominent insects in the wood are Phigalia pedaria, Tephrosia histor- 

 tata, Gonodontis bidentata, Hyhernia marginaria, Cidaria suffumata, 

 Melanippe sociata, Larentia muliistrigaHa, Orthosia helvola and Xanthia 

 circellaris. Classifying these on the basis of food plants we find that 

 the first four are not genuine larch feeders but have adopted it as food ; 

 three feed on Galium saxatile, and two are listed as feeding on trees not 

 found in the wood and are presumably in the position of the first four, 



