268 THE entomologist's record. 



proceeded to Scotland, in August, I particularly examined the same 

 plant and succeeded in finding one ovum. Although in captivity, 

 the larv£e of P. astrarcJie feed readily enough upon Geranium samjuineum, 

 repeated searches on that plant have (Avith the exception of one ovum 

 found in July, 1902) always ended in total failure. I cannot help 

 thinking, however, from the large amount of that plant in the insect's 

 favoured localities, that it forms at least one of the foodplants. The 

 ova are laid singly upon the isolated plants of rock-rose. Large 

 masses of the plant should be carefully avoided in searching for ova. 

 Single plants, sheltered by a bush of any sort or by rocks, are the 

 most favoured. In making this statement, I can specially mention 

 three cases. Once, in Fife, I took 15 ova of P. artaxerxea from a 

 plant sheltered by a furze bush, and, on the same day, I took 12 from 

 a small plant placed upon a ledge of rock, about 50 feet up the almost 

 perpendicular face of the basalt. Well do I remember getting down 

 after my climb up. The other case was in Durham, on a small plant 

 sheltered by some brambles. From that plant I took 8 ova. The ova 

 are of the usual Lycaenid type. 



The larv8B emerge in from 6-15 days, depending upon the tempera- 

 ture. They proceed from the upperside of the leaf upon which the 

 egg was laid, to the underside, and feed there. In that stage they are 

 very difficult to see, as the short hairs assimilate closely to the white 

 underside of the rock-rose leaves. In both Scotland and Durham 

 they feed until well on in their second instar, when they retire for the 

 winter and hide under dead leaves at the base of the stem. Even in a 

 hothouse they resolutely refuse to feed up the same year. They are 

 very easy to hibernate in captivity. Early in the spring, toward the 

 end of March and the beginning of April, they commence to feed on 

 the new growth. They still feed on the lower side of the leaves, eating 

 the spongy tissue from the lower side and leaving the epidermis. 

 This turns yellow, and as the larvte rarely eat more than a half of any 

 one leaf before going to another, they are easily discovered by 

 examining plants showing the characteristic yellow colour. Now, my 

 next observation is, I think, absolutely new for the species. I may 

 state with a fair amount of certainty, that I have found as many larvae 

 of this insect wild as anyone, and I have always found them attended 

 by ants. What the precise relationship between them and the ants is 

 I cannot pretend to state. Had I not been too eager I ought to have 

 solved the problem this year. I found an almost full-grown larva, in 

 May, with two ants upon its back. They seemed to be rapidly moving 

 their antennae up and down on the back of the larvae. Leaning too 

 far forward, I disturbed the larva, which dropped, and with it dropped 

 the ants. That this contact with ants is essential for the well-bemg of 

 the larvae up to a certain point, seems certain. We had larvae reared 

 ab ovu, at exactly the same stage as wild larvae discovered soon after 

 leaving their hybernacula in April, 1905. Ants were purposely kept 

 beside the latter, and away from the majority of the former. A few 

 of the former were also introduced near the ants. The wild larvae 

 (from P. ab. saliiiacis) and those from (P. artaxerxes) kept in contact 

 with the ants proceeded very satisfactorily. Those reared away from 

 ants, remained small and always looked unhealthy. In the end they 

 died before pupating. These observations were confirmed by inde- 

 pendent observers in Mr. Johnson and Mr. Carrick, both of Gateshead. 



