24 BRITISH BUTTERFLIES. 



hairs spread out like a circular fan or rosette, is whirled round with 

 great rapidity in a plane parallel to its body, its use being, almost 

 certainly, to frighten away its enemies. Similar eversible glands are 

 described by Hagen as occurring in the larvae of Plebeius argus and 

 Poli/o mm at us corydon. He writes : "You find on the penultimate seg- 

 ment, outside and behind the stigmata, two large white spots, each of 

 which evaginates a white membranous tube, just like the finger of a 

 glove, the top of which is not entirely drawn out." Exactly what 

 measure of protection is afforded by these flagella is not known, nor 

 the manner in which they afford it. Many authors, e.g., Scudder and 

 Dimmock, incline to the opinion that they are of the nature of osma- 

 teria and diffuse odours, but so far the odours do not appea,r to have 

 been detected. The glands are by no means present in all Lycaenid 

 larvae, closely allied species differing in this respect. 



In many lepidopterous larvae (Nat. Hist. Brit. Lep., i., p. 40) there 

 are present two bristles, each standing out backwards from a papilla, 

 placed directly under the anal flap, sometimes looking as if they pro- 

 jected from the base of the anal prolegs, and used by the larvae to throw 

 the pellets of frass to some distance from where they are feeding. 

 These are called " paranal forks" or "paranal tubercles." In 

 butterfly larvae, their place seems to be taken by the " anal comb," which 

 Chapman considers may possibly be homologous with the paranal 

 forks. It is present in Urbicolid larvae. Scudder also figures the 

 structure in Colias (Eurymus) philodice. It will be dealt with at length 

 in some of our detailed larval descriptions. 



One of the most remarkable external features of the Urbicolid larvae 

 is the development of peculiar glandular structures on the venter of 

 the 7th and 8th abdominal segments in their last instars. We have 

 described them somewhat at length in our descriptions of the Urbicolid 

 larvae (in the systematic portion of this volume), and there is no need to 

 redescribe them here. They appear, in the case of the Palaearctic 

 species, to be active only in the final larval stage, when the puparium 

 is thickly sprinkled with the asbestos-looking material secreted by 

 these glands, apparently for the purpose of keeping it watertight, but, 

 in some of the exotic Urbicolids, it would appear that the larval 

 shelter itself may be covered with the secretion for the same purpose, 

 e.g., Niceville notes (Butterflies of Sumatra, p. 538) that the larva of 

 Erinota thrax is covered with a white waxy powder, and that it lives in 

 a shelter made of a portion of one of the enormous leaves of Musa. He 

 adds that the pupa is also covered with the same white powder, which 

 is of the greatest service to the animal, as, in consequence of the heavy 

 showers of rain in the tropics, much water often collects in the rolled-up 

 leaf, and the pupa, if not so protected, would soon be drowned and rot ; 

 as it is, the powder keeps the pupa dry until the water has drained 

 away or dried up. The downy larva of the allied Gangara thyrsi's is 

 similarly covered with a white waxy powder. 



Little is known of the remarkable structures, detailed at length in 

 our accounts of the larvae of the various Urbicolid species, and there 

 called lenticles. Scudder describes them as crateriform chitinous 

 annuli, which are ranged in Longitudinal rows along the abdominal 

 (and sometimes the thoracic) segments. They are found only in 

 certain groups, appear to be an universal characteristic of the earliest 

 stage of the Lycamids, having the appearance of spiracles, only they 



