108 BRITISH LEPIDOPTERA. 



There was sufficient material here for the basis on which to con- 

 struct the broad lines of a natural genealogical tree, if used in con- 

 junction with the tables given us by Chapman and Hampsonf . But 

 the desiderated clue as to the actual details of such was not obtained 

 until the publication of Chapman's valuable paper, " The phylogeny 

 and evolution of the Lepidoptera from a pupal and oval standpoint." 

 In this we had a factor that could be applied in the way desired, and 

 that showed us, not which were specialised and which generalised 

 superfamilies, but which of the specialised and which generalised 

 superfamilies of the various stirpes were related to each other. 



This paper showed that the form of egg found in each different 

 superfamily is very constant, and that there appears to be no rapid tran- 

 sition from one form to the other among the Lepidoptera. There are, 

 broadly, among the higher Obtect families, two forms of egg, the flat 

 and the upright egg, the former being divisible into the Geometrid 

 and the Bombycid. The Geometrid egg is generally marked by a greater 

 roughness and by coarser ribbing or network; the Bombycid is 

 smoother and more polished, although there are many striking excep- 

 tions to this otherwise pretty general rule. Chapman is inclined to 

 derive these two forms of fiat eggs from distinct origins, very low down 

 in the evolutionary scale, but thinks it probable that the various forms 

 of the upright egg (Noctuid, Papilionid, etc.) had a common origin, 

 though very low down. He is supported in this conclusion by the pre- 

 sence of the chin-gland (ante, p. 94), which is found only in Papilionids, 

 Noctuids, Notodonts and other superfamilies with upright eggs, but 

 nowhere among those with flat eggs, and we may accept Chapman's 

 conclusion that, however widely the butterflies are separated from the 

 Noctuids, and the evidence of the Hesperid pupa shows that the 

 butterfly stirps separated from the Noctuid stirps a very considerable 

 way below any Noctua-like form usually placed with the Macros, 

 the evidence of the egg and the presence of the larval chin-gland, 

 suffice to show that they jointly separated from the Geometrids and 

 Bombycids still lower down. The evidence of the egg, too, shows 

 that the Noctuids and Papilionids were not derived, as Meyrick 

 suggests, from any Pyralid form, as the Pyralids are, in some respects, 

 of a higher type than the Hesperids, and yet the former still belong- 

 very markedly to one of the flat-egged stirpes. No very clear indica- 

 tion has yet been obtained to show where the upright egg branched 

 from the flat egg. The most probable point is between the Cossids 

 and Zeuzerids. These superfamilies are, in many respects, somewhat 

 closely allied. The former has an upright, the latter a flat, egg, and 

 Chapman considers that we have here, probably, the point where the 

 two forms are still unfixed and capable of easy variation. The alli- 

 ance (by pupa) of Castnia with Cossus, would perhaps point to this 

 also as being somewhat near the origin of the butterfly stirps. 



Accepting the principles here laid down, there can be no doubt 

 that the flat egg is the ancestral form, and the upright egg a more 

 specialised structure. Examination of a large number of eggs of 

 species belonging to several superfamilies, shows that the upright 

 eggs which characterise the Notodonts, Noctuids, Lithosiids, Euchro- 



* Trans. Ent. Soc. Lond., 1893, pp. 118-119. 

 f Annals and Mayazine of Natural History, 1894, pp. 258-259. 



