138 THE COTTON PLANT IN EGYPT chap. 



On the other hand, far more complex analyses of flower 

 colour have been proved indisputably for such plants as 

 Antirrhinum, Matthiola, Lathyrus, &c.,* so that the 

 indications of our fragmentary evidence may be considered 

 as quite probable, even in a simple sap-colour. 



Petal . spot. — The Egyptian flower is characterised 

 by a rich crimson spot at the base of the petal. This spot 

 is not so large as in G. herhaceum, where it occupies the 

 whole of the petal claw, but it is conspicuous. The 

 typical Upland cottons and Hindi have no such spot, the 

 petal being self-coloured. The intensity of the spot may 

 difi'er within commercial Egyptian varieties, like the petal 

 colour. Similarly, commercial stocks of, e.g., "King" 

 Upland contain a notable proportion of plants with spots 

 on their petals. Such difi"erences are probably due to 

 doubtful pedigree. 



A cross between "full spot" Egyptian and "spotless" 

 Upland gives an intermediate Fi. The Fi spot is smaller, 

 more vague in outline, and less noticeable than the spot of 

 the Egyptian parent. In cases where the latter had a spot 

 which was relatively small, the Fi is proportionately 

 inconspicuous. 



In F2 the ratios are very erratic. Taking them in 

 the order "full: intermediate : noae," we find two families 

 from Fi sister plants giving 11 : 22 : 18, and 23 : 42 : 31. 

 In this case the ratio approximates to 1 : 2 : 1, but it is 

 closer to 3:9:4, and the divergence from even this is 

 more than can be explained by errors of observation. 

 Probably, as in the case of petal colour, there are more 

 classes than we have admitted. This view is substantiated 

 by the fact that we have been unable to discriminate in 

 many cases between "Full" and "Full?". A.s in the 

 case of petal colour we find that " spotless " breeds true, 

 while " full " may either breed true or break, and inter- 



* Seu Bateson, Saunders, Baur, Wheldale, &c. 



