PROTECTIVE COLORATION AND DEFENSIVE STRUCTURES OP LARVAE. 93 



thickened and shortened, so that when the creature throws itself into 

 a sprawling grotesque attitude, with its tail up in the air, as remarked 

 by Hermann Miiller, it resembles a great spider. At the same time, 

 the style of coloration is changed ; it has not the green and red tints 

 of Centra, but is tinted light and dark horn-brown, like the bodies of 

 many large spiders. In the case, then, of Stauropus, variation has gone 

 on in a novel and determinate direction, the process of natural selec- 

 tion ending in a result not to be observed in the case of any other 

 lepidopterous larvae, the initial cause of variation being apparently the 

 result of protection, due to a resemblance to members of another class 

 of arthropods (Bombycine Motlts of America, p. 81). 



It has been pretty clearly proved by recent experiments, that bright 

 colours are readily distinguished by insects, and that the recognition of 

 these, and of moving bodies, are the main features of insect vision. 

 The value of mutant or moving tubercles, therefore, as a means of pro- 

 tection, is evident, and an ichneumon, or carnivorous beetle, or bug, 

 may be frightened away if startled by a moving tubercle like those 

 found on many lame. 



We have previously referred to the protective structures of the 

 larvae of the Schizurae. That of S. leptinoides has, on the 1st abdominal 

 segment, a high dorsal tubercle, that is both mutant and slightly re- 

 tractile, being invaginated when the larva is irritated. Packard says : 

 The movable, terrifying tubercle of the 1st abdominal segment of the 

 larvae of the Schizurae becomes developed shortly before the creatures 

 are half-grown. He connects the development of these tubercles with 

 the larvae changing their hitherto concealed habit, to a more exposed 

 position when feeding. The swollen, coral-red, dorsal hump on the 

 1st abdominal segment of the larva of S. concinna, to which we have 

 already referred, can also be moved by the larva so as to terrify its 

 enemies. 



The presence of moving tubercles is not particularly well illustrated 

 among the larvae of our British lepidoptera, and those of Apatura iris 

 are, perhaps, the best known. They are very prominently developed 

 in the larva of Anoxia arcldppus, which keeps the anterior flexible fila- 

 ments constantly in motion backwards and forwards when eating, and 

 moves them still more rapidly when alarmed. 



Packard has pointed out that the large larval spines of some of the 

 Saturniids are movable in the early larval stages. These spines are 

 marvellously developed in some Saturniid larvae. In that of Cithcnmia 

 rei/alis, for example, there are, on each segment, six well-developed 

 spines, and an additional large median spine on the 8th and 9th 

 abdominal segments. The two large median spines, on the two hinder 

 thoracic segments, are each about 20 mm. in length. The larger 

 spines all end in a swollen, triangular, two-horned, flattened bulb, 

 these appendages being deterrent, and for offensive use in the early, as 

 well as the later, stages of larval life. Packard further states that the 

 four pairs of pro thoracic horns (or spines) in SpMngicampa are not 

 held spread out as in Citheronia reyalis.. but those of each pair are con- 

 stantly held close to each other. The horns and the six silvery, 

 opalescent, shining tubercles (on the 5th to 10th abdominal segments) 

 probably become terrifying by the movements of the larva. The 

 latter are turned on, and throw their light out suddenly, like flashes, 

 and may thus have a deterrent efl'ect on their enemies. It is possible 



