SCIENTTFK! NOTES AND OBSRUVATIONS. 295 



June 2Gtli, and if this is anywhere near the normal period at Sandown 

 in a state of nature, the good specimens which I met with a month or 

 two later (this year I took one absolutely perfect on August 3Uth), must 

 belong to this class of " forwards," and as they are certainly entirely 

 fertile, we shall have a complete second cycle, or a genuine, though 

 partial, " double brood." At any rate, whether the August rioata be- 

 long to a second brood, or are retarded emergences, I determined this 

 year to make them serve my purpose of oljtainiug a side-))y-side com- 

 parison of the early stages, and though 1 did not arrive at SandoAvn 

 until August 4th, too late to meet with many rivata, I managed to 

 capture a worn $ on August 13th, and took a J soclala the same day 

 for my comparison. In order not to lose time, as I feared the vivata 

 would not last long, I slipped a small shoot of Galium niolJugn into each 

 of the boxes where I confined my $ $ , and on my arrival home, I found 

 that both had commenced to lay ; the vivata, however, only laid 

 six eggs, and was dead l)y the morning of the loth; the sociata had 

 then laid 70, and I let it go. Both species will lay freely in chip boxes 

 on the little bits of Galiuin introduced, but very rarely if ever, on the box 

 itself, when they can get the food-plant. Both laid the eggs singly, on the 

 underside of the leaves, at the edge, and generally near the tip, one only 

 (sociata), on the stem. The eggs, superficially viewed (I was not able 

 to subject them to microscopic examination), are similar in form and 

 consistency, of the ordinary ovoid form of the group, smooth and shining ; 

 but rivata, besides being of course the larger, is distinctly paler, so that 

 I should describe it as almost cream colour, while sociata is decidedly 

 tinged with yellow ; it also appeared that rivata, was perhaps, slightly 

 the narrower proportionately, at the narrower end. No change takes 

 })lace till very shoi'tly before hatching (I have no precise observations 

 to record, but certainly well within twenty-four hours), when the usual 

 darkening, through an opacpie but not dark gTe3ash to quite a deep 

 leaden tint, occurs. A fcAv sociata hatched on August 23rd, and the 

 rest very shortly after ; one rivata on August 24th, four more within a day, 

 the sixth proving infertile. The duration of the egg stage with rivata, 

 may thus be taken as one day longer than with its ally, atmospheric 

 conditions being identical. This observation is supported by one made 

 upon the two sj^ecies last year, within ten days of one another, when (in 

 July), each hatched a day more rapidly than this year (in August), but 

 the relative period was the same. The larva3 when first hatched are 

 very similar, and, rather curiously, the size difference is less observable 

 than in tlie eggs ; indeed 1 wrote that they were " apparentl_v of practically 

 the same size." They are of a unicolorous greyish-yellow, tlie head 

 deeper, and more of an orange tint. Both would occasionally drop by a 

 web when touched, a hal)it which most of the " carpet " larvas seem to 

 liave in their first skin, though, as far as I recollect, only ijah'ata retains 

 it into the second, and even in this case the maturer larva entirely abandons 

 it. Most of the sociata reached their first moult on August 28th. tlie 

 more backward ones being just the size of their contemporary rivata ; 

 on that day I again compared the non-moulting larva?, and remarked 

 that rivata was decidedly brighter in coloui', almost apple-gi-een, while 

 sociata was of a duller, more glaucous green, and also that rivata was 

 smoother, sociata being a little rugose laterally. Rivata reached its first 

 moult the next day (August 29th), and after this, another com]iarison 

 of the contemporaries showed similar differences to those just noted ; 



