SATURNIA. 301 



pyri abandons it almost completely in the 3rd stage and onwards. The succeeding 

 green colour, which is no doubt adaptive, is acquired by S. pyri at an earlier stage 

 and more completely than by S. pavonia. 



2. In the larva of S. spini, the tubercles are not very prominent even in the 

 adult, and the knobs at their summit are not distinctively coloured until the last 

 stage. In S. pavonia the tubercles are conspicuous and the knobs acquire an 

 appearance distinct from the general surface of the body in the 3rd and 4th stages. 

 In S. pyri, on emergence from the egg. the knobs are already coloured and indica- 

 tions exist of the extreme prominence of the tubercles. 



3. The tactile bristles are least developed in S. spini, most in -5". pyri. 



4. The cocoon of S. spini is simpler than that of S. pavonia. That of S. 

 pyri is the best defended of the three. 



5. S. spini is almost sexually monomorphic though the male is somewhat 

 the smaller. The female is very sluggish, and the antennas of the male are pectinate 

 to a high degree. In S. pavonia there is well-marked sexual dimorphism, as regards 

 both size and colouring. The female resembles that of S. spini in aspect ; it is 

 less sluggish though not very active. The antennae of the male are less strongly 

 pectinate. S. pyri again is sexually monomorphic ; the female is a tolerably good 

 flier, and the antennae in the male are less strongly pectinate than in many other 

 species of Satumia. 



Standfuss further concludes from the larval and pupal characters 

 that the three species form a progressive series in protective adapta- 

 tion, S. spini always taking the lowest, and S. pyri the highest, 

 step in the scale, so that S. spini must be considered on this ground 

 also as the phylogenetically oldest, and S. pyri as the phylogenetic- 

 ally youngest, of the three forms. Other interesting details as 

 bearing on the phylogeny are collected from observations on the 

 larvae and imagines of the hybrid forms. These are given at length 

 floe, cit., pp. no et sea. J, and, summarised, read as follows : 



1. The freshly-hatched hybrid larva closely resembles the larval form of the 

 female parent. 



2. With the process of growth a resemblance to the male parent gradually 

 increases. 



3. The final extent of approximation towards the male pai'ent depends on the 

 relative phylogenetic age of the two species, the older being able to transmit its 

 properties^ whether of structure or habit, better than the younger. 



Thus the crossing of ,S. pavonia $ with the phylogenetically 

 younger S. pyri 2 gives rise to a larva in which at first the 

 maternal and afterwards the paternal characters predominate. The 

 resulting perfect insect is by more than two-thirds of its external 

 appearance S. pavonia, and by less than one-third S. pyri. Its 

 habits and functions correspond with its external aspect. It prefers 

 to fly by day, like S. pavonia $ , and pairs easily with the $ of that 

 species, from 43 to 62 per cent, of the eggs being fertile. On the 

 other hand, it does not pair readily with S. pyri, and the resulting 

 eggs on an average of nine cases gave only one larva in 180. 

 Similarly 6". pavonia $ , when paired with the phylogenetically older 

 .S. spini 2 , gives a form of which in the perfect state about two- 

 thirds of the external aspect belong to the type of S. spini. The 

 male flies by night. After crossing with S. pavonia 2 , the resulting 

 eggs were only fertile to the extent of 16 to 22 per cent., while the 

 crossing with S. spini $ , though not easily brought about in conse- 

 quence of their diverse times of appearance, yielded eggs of which 

 from 94 to 98 per cent, were fertile. Thus the male S. pavonia is 

 able to influence the issue of the relatively gigantic £. pyri $ much 

 more than that of S. spini $ . Again, the issue of S. spini $ x 

 •S". pavonia $ is much nearer ,5*. spini than is that of S. pavonia $ 

 X .S". spini $ • Hence — 



