56 



NA TURE 



[December 19, 1895 



How insects first acquired wings, and from what struc- 

 ture they were derived, is still a profound mystery. A 

 favourite conjecture is that they arose by the modification 

 of tracheal gills. In favour of that supposition is the 

 fact that in some dipterous larvae (Chironomus, Simulium, 

 Culex, Muscida;), three pairs of rudiments (imaginal 

 folds) form on the dorsal surface of the thoracic segments, 

 and as many on the ventral surface. The ventral rudi- 

 ments ultimately become the legs of the fly. Of the dorsal 

 rudiments the second becomes a wing, the third a rudi- 

 mentar>' wing, and the first the respiratory organ of the 

 pupa, which may take the form of a trumpet or a bunch 

 of branching tubes. In the larvcC of the Muscidce this 

 anterior dorsal appendage is said by Weismann to form 

 much later than the rest, only during the last larval stage, 

 and to be greatly inferior in size. In Chironomus and 

 Simulium it forms at the same time, and is quite similar 

 to them. Where an anterior spiracle exists, as in the 

 Muscidae, the anterior dorsal appendage forms close 

 behind it, and ultimately replaces it. Now, if it could 

 be clearly shown that all three dorsal appendages form 

 one series, and that one or two are converted into wings, 

 while the foremost becomes a breathing-organ, we should 

 at least know that a wing and a breathing-organ may 

 have a common origin. So rnuch seems probable, but 

 we must still wait for conclusive facts before we can 

 .explain how insects first got their wings. 



In many Orthoptera, and other hemimetabolic insects, 

 effective wings are developed without any resting-stage. 

 I think that we may safely infer that historically wings 

 were acquired before complete metamorphosis set in, and 

 before insect-larvas underwent degeneration. 



Though the fact must, of course, be well known, I do 

 not remember that any zoologist has expressly mentioned 

 that the resting-stage of insects is unique among animals. 

 The nearest phenomenon of the same kind is the encyst- 

 ment of certain Protozoa and parasites ; but cessation of 

 all the functions of active life, when these have once 

 been assumed, for the purpose of effecting a definite 

 advance in organisation, is a thing peculiar to insects. 

 Hence we cannot check our interpretations by examples 

 taken from outside the class. I3ut all insects do not 

 pupate, and we may learn something by studying the 

 hemimetabolic insects in which there is no quiescent pupa. 



Nearly all these insects (some few Orthoptera and some 

 Rhynchota are exceptions) ultimately acquire wings, the 

 rudiments of which may be externally visible long before 

 the imago stage is reached. In hemimetabolic insects it 

 is unusual to find any marked change in the food, the 

 mouth-parts, the general form of the body, the texture of 

 the skin, the length of the legs, the muscles, or the 

 nervous system. It is often difficult to distinguish a full- 

 sized larva from the imago without close examination. 

 In Ephemeridas, Perlidie, and Odonata the larva is 

 aquatic and the imago aerial, so that notable changes 

 appear in the mode of respiration, but still the transition 

 is effected without a resting-stage. 



We have seen that wings may be acquired by insects 

 which pass through no pupa-stage. There is, however, 

 hardly a case (the parasitic flea is one) of an insect 

 which pupates without acquiring at least the rudiments 

 of wings in one or other sex. The main purpose of 

 pupation would in general be defeated if the adult were 

 wingless. 



Change in the form of the mouth-parts, if so consider- 

 able as to involve a new method of feeding, e.g. the 

 substitution of sucking for biting, cannot be effected 

 without a quiescent pupa-stage. It seems inevitably to 

 involve a rather prolonged interval during which the old 

 mouth-parts are not in working order, while the new 

 ones are not yet complete. But though considerable 

 changes in the mouth-parts imply a quiescent pupa, they 

 are not the sole reason for pupation. The larva and 

 imago of many Hymenopteraand Coleoptera have similar 



NO. 1364 VOL. 53] 



mouth-parts, yet they all pupate. Neither acquisition of 

 wings nor great changes in the mouth-parts can be the 

 sole reason for pupation, for pupation is not indispensable 

 to acquisition of wings, and it is not always followed by 

 material changes in the mouth-parts. 



Lubbock quotes from his earlier memoir on Chloeon 

 {Linn. Trans.., vol. xxv. p. 486, 1865), the remark 

 "that the occurrence of metamorphoses arises from 

 the immaturity of the condition in which some animals 

 leave the egg." Insects are specially referred to, 

 for in the memoir quoted we are told that the necessity 

 for change depends on the fact that most insects 

 leave the &%g in a very early condition, and that this 

 again is probably owing to the fact that the amount 

 of nourishment in the ^<g% is insufficient to carry the 

 insect to maturity. Brauer adds that the eggs of insects 

 with complete metamorphosis are regularly smaller in 

 proportion to the parent than those with incomplete 

 metamorphosis or none. 



I believe that Lubbock's explanation is true of many 

 marine invertebrates, but not of insects. The vast number 

 of eggs laid by a crab or a sea-urchin is very likely one 

 cause of the smallness of the eggs and of the unlikeness 

 of the larva to the parent. The transformation which 

 assimilates the larva to the parent is in such cases effected 

 as soon as possible after the migatory larva has settled 

 down and begun to feed. But in the insect there is, as 

 a rule, no important advance in structure during almost 

 the whole of the larval period. If deficiency of nourish- 

 ment had occasioned a temporary arrest of develop- 

 ment, abundance of nourishment would surely have made 

 up the loss sooner. The time of pupation seems to me 

 too distant from the time of hatching, especially as the 

 whole period of active feeding and rapid growth inter- 

 venes, to be entirely due to the conditions of nourish- 

 ment of the larva in the &%%. Special facts, of which 

 many could be cited, tell against Lubbock's explanation. 

 The Staphylinidae commonly lay relatively large eggs, 

 and yet undergo complete metamorphosis. Some insects 

 actually retrograde before hatching out, and lose legs 

 which they had already acquired, a pretty clear proof 

 that want of nourishment was not the cause of what is 

 called their immaturity at birth. The state of the insect 

 at hatching seems to me to depend far more upon the 

 conditions of larval life than upon the supposed privation 

 of nourishment during embryonic development. 



It is plain that insects have gained very much by 

 complete metamorphosis. The extraordinary numbers 

 and range of the holometabolic insects settle that fact 

 decisively. If further proof were required, we might point 

 out that the resting-stage or quiescent pupa seems never 

 to have been lost in any insect which once possessed 

 it. It is hard to prove a negative, but I cannot call to 

 mind a single clear instance. So powerfully has adapta- 

 tion acted upon insects that almost every organ and 

 almost every stage is known to disappear at times. 

 Wings, legs, eyes, mouth-organs, head, are known to be 

 deficient in the larva, and a very few adult insects have 

 no functional wings, legs, eyes, or mouth-org^ans. The 

 single order Diptera furnishes us with examples of sup- 

 pression during the larval stage of all these organs. The 

 ^^^f the larva, the winged imago may disappear as 

 independent stages in the pupiparous Diptera. But the 

 quiescent pupa remains in every case where it can be 

 shown to have once existed. At most the pupa (in holo- 

 metabolic orders) becomes in some degree capable of 

 locomotion ; it never feeds. 



We cannot reckon among the advantages secured by 

 complete metamorphosis the acquisition of wings, for 

 many insects which acquire wings have passed through 

 no resting-stage. Among these hemimetabolic insects 

 are the dragonflies, which take their prey on the wing, 

 but in general the hemimetabolic insects gain nothing 

 by flight, except facilities for dispersal and egg-laying. 



