lepidoptc rolls larvte and jmpce. 157 



generalised vertebrate appearance, probably of the 

 serpent type (at any rate in Chcsrocampa), such as would 

 be most efficacious in the case of birds. Mr. H. W. 

 Bates gives an instance, also quoted by Mr. A. R. Wallace 

 (see Linn. Soc. Trans., xxiii., 1862, p. 509, for Mr. 

 Bates' i^aper) of a South American larva which startled 

 him and every one to whom he showed it by its strong 

 resemblance to a venomous serpent, and it is likely that 

 the terrifying appearance of our own larva; in temperate 

 latitudes first arose in the tropics, where the imitated 

 cause of alarm to the enemies of the larva is real and 

 obvious. And it is probable that the success of the same 

 method in countries where the reptilian fauna cannot be 

 said to constitute a source of alarm, is due to inherited 

 memories of a tropical life which live on as that 

 instinctive fear of anything snake-like which is so com- 

 monly exhibited by the higher land vertebrates, including 

 ourselves. At the same*time the fear is less extravagant 

 than that manifested by the few vertebrates (frogs) that 

 still suffer from the attacks of an ancestral enemy, and 

 to whom the cause of alarm is still a reality. 



7. The fluid ejected by the larva of Dicranura 

 viNULA. — This defensive fluid, as is well known, is ejected 

 from a transversely placed aperture on the ventral 

 surface of the prothorax immediately below the head. 

 I have long known that it is powerfully acid, affecting 

 litmus-paper in the most marked manner. During the 

 past summer (1885) I found that the fluid causes violent 

 effervescence when allowed to fall upon sodium bi- 

 carbonate. With the kind assistance of Prof. Wyndham 

 E. Dunstan I was enabled to prove (by the reduction of 

 silver nitrate) that the fluid is formic acid. The smell 

 is also quite characteristic, and affords an indication of 

 the large proportion of acid present in the secretion. It 

 is also an interesting fact that the freshly-made and moist 

 cocoon of D. riiivla is powerfully acid to test-paper. 



In September, 1885, I found a few larvis of D. farciila 

 on willow near Visp, in Switzerland, and 1 examined 

 them to see if they also would eject a fluid when irritated. 

 It was at once obvious that the aperture was present in 

 the same position as in I), viniila, but, as the larvae did 

 not eject any fluid, I applied gentle pressure (for this is 

 often necessary in the case of D. vinulu). To my great 



