J. W. H. Harrison 235 



choose our own Oporabia dilutata as an example ; it produces its 

 silveriest forms in those very districts. Whatever melanism exists 

 there does not affect woodland moths; it manifests itself in insects 

 found on spray-drenched and open spaces, whether rocky or sandy in 

 nature. 



(7) Lastly, to demonstrate that the effect of natural selection is quite 

 negligible as a factor in progressive melanism, I carefully studied the 

 case of Polia chi, which in the Team Valley produces about 50 °/^ each 

 of typical and of the melanic forms grouped under the name olivacea, 

 and near Middlesbrough about 10 °/^ of dark and 90 °/^ pale forms. For 

 several years and on every day during their season — rain or fine — either 

 my wife or my brother carefully noted the positions of the insects resting 

 on three walls: (1) old and dark in parts, proceeding from Birtley to 

 Newcastle ; (2) old light yellow sandstone, proceeding to Burnmoor ; 

 (3) mixed new greyer sandstone and old reddish ones leading to Chester- 

 le-Street. On these three walls I have seen up to three hundred 

 examples daily, so that the present test is not confined to few insects ; 

 in the evening full particulars would be given to me, and sometimes alone 

 and sometimes accompanied by my brother I would go over the ground 

 to investigate the fate of the insects observed earlier in the day. Never 

 was there any diminution of numbers in which more olivacea vanished 

 than type chi ; as a matter of fact we used to consider it a marvellous 

 thing if even a single one had disappeared. 



Now to proceed. The melanism both of the continuous type shown 

 by Larentia multistrigaria, 0. dilutata and 0. autumnata as well as the 

 discontinuous form encountered in Amphidasys betularia var. douhle- 

 dayaria and Boarmia consonaria is germinally fixed, so that any possi- 

 bility of its being ontogenetic is excluded ; besides, except in the very 

 unlikely case of its being possible to affect mature ova and spermatozoa 

 by the same agency and in the same direction as the somata are affected, 

 the impossibility of its being other than germinal is capable of absolute 

 proof in many insects like Spilosoma lubricipeda, Orgyia antiqua and 

 Lycia hirtaria. In all of these cases gametogenesis is completed long 

 before scale and pigment formation, so that unless the germ plasm were 

 in itself capable of giving rise to melanic forms the phenomenon would 

 not be observed in succeeding generations. And if it did occur merely as 

 a somatic aberration, then, not being germinal, selection either artificial 

 or natural would not be capable of fixing melanism of the degree already 

 attained, much less of urging it on to its climax in such totally melanic 

 forms as Oporabia dilutata var. melana or Gonodontis bidentata var. 



