PROTECTIVE COLORATION AND DEFENSIVE STRUCTURES OF LARVJ2. 77 



twigs and other objects upon which they normally rest. We have 

 already seen (ante, p. 42) that the specialisation of hairs and spines makes 

 them unsuitable for the food of insectivorous birds. Other lines of 

 specialisation by means of which they are protected are by eversible 

 glands (sometimes taking the form of flagella), acid excretions, 

 obnoxious odours, dangerous-looking spines, and horns, and spots ; 

 even remarkable attitudes help to swell the sum tdtal of the defensive 

 possibilities of larva?. 



That larva? are protected by having a habitation into which to 

 retire, and that they thus gain an advantage in the struggle for 

 existence, appears certain. Niceville mentions (Butterflies of Sumatra, 

 p. 394) that the larva of a large Skipper butterfly (H'ulari irava) 

 and that of a Nymphalid butterfly (Amathiisid phidippm) live, at 

 the same time, on the leaves of Cocos nucifera. He says that, 

 owing to their general abundance, the two species often have a severe 

 struggle to live together, in which the more robust Hesperid, which 

 secures a shelter for itself by spinning the leaves together, is generally 

 victorious. 



The various means by which larvae are protected, owing to their 

 similarity to some part of their food-plant, or by their resemblance to 

 some object common upon it, is well-known.. Some larva? resemble 

 structures on the leaves ; thus, whilst the larva of Apoda avellana 

 assimilates to the surface of an oak leaf, that of Heterot/enea cruciata 

 has been compared with a gall. Packard also says that the larva of 

 Lithacodes fasciola and those of Packard ia are entirely green, oval in 

 form, and might easily be mistaken for a fold or bend in a leaf. 

 The greater part of the Geometrid larva? resemble twigs, whilst 

 arboreal Noctuid larvae are either coloured so as to suit their envi- 

 ronment, or otherwise resemble some portion of the tree sufficiently 

 well to escape detection, whilst ground-feeding larva? resemble, in tint, 

 the ground on or under which they rest by day. 



Elliott says that the larvae of the American Heterogenea flexuosa and 

 H. testacea&ve wonderfully similar to the red dipterous or aphidid galls on 

 oak and other leaves. Packard, too, notes the resemblance between 

 these larvae and the small reddish-green galls, which appear late in 

 summer on the leaves at the time when the larva? themselves become fully 

 grown. He then adds : These forms being thus protected from observa- 

 tion and harm, do not need the armature of the larvae of the other group 

 (of this superfamily), and the tubercles and spines have disappeared 

 through simple disuse ; while being without poison-bearing spines, 

 they have also lost by disuse the bright colours and conspicuous spots 

 of the armed genera. On the other hand, the larvae of Odoneta, 

 Einpretia, Euclea, and allied forms, with their remarkably bright 

 colours and markings, and poison-bearing (? urticating) tubercles, feed 

 conspicuously, the warning colours and showy ornamentation repelling 

 the attacks of birds. We are inclined to the belief that the armed 

 slug- worms were the earlier, from the probability that, in the Coleop- 

 tera, the earliest and most generalised groups were the Staphylinidae 

 and the carnivorous Carabidae, and their allies ; while the later, most 

 extremely modified forms were the weevils and Scolytidae, in which the 

 larvae are footless. In the Diptera, also, it is not improbable that those 

 families with the most perfectly developed larvae, such as the Cidicidae 

 and Tipulidae, were the earliest and most generalised types, while the 



