530 MR. L. DONCASTEK ON COLOUR- VARIATION [DeC, 12, 



found that of the males 81 ipev cent, were spotted with dark 

 undersides, 19 per cent, striped only with light undersides. I 

 found rather over 83 per cent, spotted and dark, 16 per cent, 

 striped and light, and about 0*5 per cent, with intermediate under- 

 sides. Of the females, I found 27*5 per cent, with dark under- 

 sides, the same propoi-tion as was observed by Bateson. The 

 occurrence of the different vai'ieties is in every way in close agree- 

 ment with that found by Bateson ; and it may be concluded that 

 their distribution at that season has not changed appreciably in 

 ten years. I did not collect in the Darro valley from which 

 Bateson's second sample was obtained, but I found that on the 

 lower slope of the liill towards the Genii valley the proportions 

 did not differ from those on the top. On the ]3arro slope of the 

 hill, which is very steep and faces north, the beetle did not occur, 

 although S'partiimi was abundant. 



At Gi'anada a considerable proportion of the beetles were in 

 cop., and I collected 119 couples and recorded the characters of 

 each as they were gathered. Care was taken to see that they 

 were really paired, and since 71 pairs remained coupled after 

 they were dead in the killing-bottle, there can be little doubt 

 that all or neai^ly all were really in cop. Table II. gives an 

 analysis of these. Of the 119 pairs, there were 22, or over 18 per 

 cent., in which both male and female were striped green with 

 light undersides. Taking pairing at random among the general 

 population the expectation would be 10"5 per cent. But the pro- 

 portion of males of this type which were paired is considerably 

 higher than in the general population (29 out of 119 or 24'4 per 

 cent.) ; so that random mating out of those paired would give 



17 per cent, of svich pairs, Avhich does not differ greatly from the 



18 per cent, observed. Similarly there were 25 paii-s (21 per 

 ce t.) in which both male and female were red with dark vmder- 

 sides. The expected number on random mating among the whole 

 population is 15"2 per cent., on random mating among those 

 actually found paired about 17 per cent. The numbei-s observed 

 are of course much too small to give reliable conclusions, but they 

 indicate that of the males of the green striped form, and females 

 of the I'ed spotted form, a slightly higher proportion is found 

 paired than in the genei-al population, and that there is possibly a 

 very small tendency towards selective mating between individuals 

 of the same colour type. 



At Ronda Spartiimi bushes were very scarce, and upon many 

 there were no beetles, so that altogether only 106 individuals 

 Avere obtained, 80 of which were males, 26 females. Although 

 these numbers are small, they indicate that the population differs 

 considerably from that of Granada (Table III.). Of the males 

 60 (75 per cent.) were spotted and dark underneath, 9 were 

 green, striped and light, one red, spotted and striped, light, and 

 iO red, spotted and striped, with intermediate undersides. At 

 Granada only 6 intermediates occuri-ed in 1382 specimens; while 



