THE LlFE-HISTORY OF A LErlDOl'TfiKOUS INSECT. 67 



fly), Trichopteua (Caddis-flies), Coleoptera (Beetles), Siphonaptera 

 (Flea), Drpi-ERA (House-fly), Lei'iduptera (Moths and Butterflies), 

 Hymenoi'tera (Bees, Ants, Icbneunion-flies, &c.), are the Orders into 

 which the Holo-metahola are usually divided, but the number of these 

 will vary, according as the weight given to certain characters by different 

 authors varies. 



3. On the relationship of the various stages in an insect's 

 life. —If we consider the characters of the various stages in each of 

 the three groups, we are at once struck by the fact that in the Rolo- 

 metabola, to which the Lepidoptera belong and in which we are 

 therefore more particularly interested, there is an immense gap between 

 the larva and the pupa, much greater than that between the pupa 

 and the imago. Now we may fairly assume that the original tendency 

 of all insects was to have, not widely separated changes, but rather a 

 sequence of comparatively closely related ones, and that the features 

 which characterise the metamorphoses of the Lepidoptera point to the 

 probability of differentiation in very opposite directions between the 

 adult larva and the pupa, resulting in the quiescent condition now 

 characteristic of the latter. As a matter of fact, we find that the earliest 

 stages of the lepidopterous larva show a development often termed 

 lower, but in my opinion simply more divergent, than that of the larva 

 of the Hemi-metahola, so that several stages are apparently missed be- 

 tween larva and pupa ; at the same time, the imago has undergone so 

 much gTcater a progressive development, than have those of the other 

 group, that the gap becomes still more striking. 



The study of the metamorphoses of the Lepidoptera has led Mr. 

 Poulton to conclude, that " the suppression of intervening stages has 

 left the first or larval stage in an extremely ancestral condition, so that 

 the larva in Lepidoptera is far more ancient than the first stage of those 

 insects (Orthoptera, etc.) which still retain the more ancestral method 

 of metamorjihosis. These, therefore, have lost the early stages, whilst, 

 Lepidoptera, etc. have lost all the stages intervening between the 

 ancient and a very late stage" (Trans. Linn. Soc. Lond., 1889, p. 190). 

 I do not agree with this. My own impression is, that the whole 

 of the metamorphoses of the earliest modifications of the ancestral type, 

 were confined within very narrow limits, and that the slight changes 

 characteristic of the A-metahola at the present time represent this con- 

 dition much more correctly than any other, and that their larvas are 

 the more ancestral, whilst at the same time the needs, habits, etc. of 

 those insects which are supposed to have attained to the most advanced 

 development in the imago state, and which differ profoundly from those 

 more ancestral forms, have also undergone great modification in the form 

 of their larvae, such modification tending towards a condition of inactive 

 helplessness in that stage. But this does not necessarily show a more 

 ancestral form, but rather a modification in response to environment. 

 That is to say, if these larvaa are all essentially the outcome of the 

 ancestral form, those of Lepidoptera (and the assumed higher groups), 

 must be distinctly more specialised and farther removed from the 

 assumed primitive type, and instead of having reverted towards such, 

 they are, in reality, much more specialised, when compared with the 

 primitive Thysanuran standard which we set up. Instead of approaching 

 the primitive tyjje, then the lejiidopterous larva undoubtedly differs 

 very greatly from it, and shows, in reality, a very high standard of 

 specialisation. 



