136 DESCRIPTIONS OF AUSTRALIAN MICRO-LEPIDOPTERA, 



attitude of tlie imago in BedelUa (preserved also in Tisclieria) is a 

 rudiment of tlie peculiar position assumed by a Gracilaria. 



The LyonetidcB, thougli a very natural family, have a very wide 

 range of structure in respect of the head and the neuration. 

 Stegommata and Lyonetia are very nearly allied, and might well be 

 considered sections of one genus. Regarding them so, then 

 Bucculatrix, Crohylophora, and Stegommata would form a natural 

 group in the family, distinguished by the roughly tufted head. 

 Opostega and Cemiostoma are also naturally associated, and 

 Phyllocnistis appears by its quite smooth head and apodal larva to 

 be an extreme development of these. Atalopsycha may perhaps 

 be intermediate between the two groups, but it could not yet be 

 safely affirmed. Most of these genera, though small, are of 

 universal range, and the others are known as yet merely as small 

 local developments, so that nothing can yet be predicted from the 

 facts of their geographical distribution. I am disposed to think, 

 hovever, that Buccidatrix, which is the largest of them, and also 

 the most persistent in type, is probably the oldest, and nearest to 

 the original form. It is also the nearest to the NepticulidcB, which 

 we are probably justified in regarding for the present as a very 

 ancient but degenerate development of Bucculatrix, or rather of 

 the progenitors of Buccalatrix. 



In the preceding remarks I have been twice led to assert that 

 a more lowly organised form has been derived from a higher, and 

 I have reason to believe, as in subsequent communications I hope 

 to show, that this has taken place more commonly than is often 

 supposed. Such examples are not, however, as a superficial 

 observer might suppose, a violation of the law of improvement 

 under natural selection. According to the principle of evolution 

 any change may occur if beneficial, and the degradation, and 

 consequent simplification, of an organism must often be as great 

 a benefit, as its complication at the cost of increased requirements. 

 It wiU, I think, be in practice rarely found that the lowest 



