142 BRITISH LEPIDOPTERA. 



the first two abdominal segments — elpenor, celerio. (2) With 

 eye-spots repeated on all the segments. Of these, he makes the 

 first the lower. Besides the objection to this order, already noted, 

 it may be added that those with the complete series usually have 

 them of approximately the same value throughout, as though they 

 had all developed together, pari passu, with each other, at least the 

 second spot is usually fairly in line with all the following ones. Had 

 they arisen from an elpenor-form the first two must have receded, whilst 

 the others advanced. This seems a very unlikely thing to happen. 

 Weismann makes no reference to the forms with one eye-spot only 

 on the 1st abdominal. Other difficulties might be raised from 

 Weismann's arrangement, e.g., the varying intensity of the eye-spots 

 when occupying all the segments, from merely faint indications to 

 completely developed eye-spots, nearly identical throughout the 

 length of the larva, going far to show that they were all developed 

 together, others in which the forward eye-spots are best developed, 

 quoted by Weismann ( Lilina bisecia ) as showing the gradual exten- 

 sion of the eye-spots backwards, illustrates the gradual loss of all but 

 the forward spots. There is, apparently, no larva with 3, 4, 5 

 or 6 spots, nothing between 1 or 2 on the one hand, or 7 on the 

 other. It is, however, perhaps, better to insist pointedly 



on an instance that Weismann himself would certainly admit 

 to prove that a complete row of spots preceded their more localised 

 specialisation. In Weismann's Studies, &c, transl. p. 194, Meldola 

 describes the adult larva of Fiorina japonica as possessing the two 

 (and only two) forward eye-spots as in Eumorpha elpenor. In the 

 Tijdschrift, vol. xl, Piepers figures the young larva of this species 

 as possessing a complete row of eye-spots. With a greater know- 

 ledge of exotic forms, doubtless many other such instances would 

 be forthcoming, but one is as good as a dozen, since there is no 

 reason that I can discover to believe we have here any exception 

 to the rule on which Weismann so much insists, and which generally 

 holds good, of the earlier stages representing ancestral conditions. 

 Taking it then as proved by this instance, for Weismann adduces 

 nothing but what is mere plausible matter of inference in support 

 of his own conclusion, that the . initial Eumorphid character was 

 a continuous row of eye-spots, it becomes extremely probable that 

 this state of the evolution of the markings was possessed by the 

 common ancestor of the Phryxids and Eumorphids. It follows 

 from this that the distinction between Phryxids and Eumorphids 

 is not so deep and fundamental as it would be, if Weismann's 

 theory were correct. My own view, founded on pupal characters, 

 was at first very close to Weismann's as regards the position of 

 E. elpenor and The ret ra porcellns, viz., that they were very low in the 

 group. In concluding from the larval evidence that they are, on 

 the contrary, at one of the highest points the pupal indications 

 have to be reviewed. I may say the imaginal evidence also, for 

 these species are certainly much less specialised than such species as 

 Hippotion celerio or Isoples alec to. It is necessary here to repeat that, 

 in dealing with such questions as these, one must keep oneself well 

 in hand and remember that these species are not to be got into a 

 lineal arrangement; a tendency to do this always besets one, if one 

 be not continually on the alert, and one sees the results in many 



