490 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



been the custom, apparently, for most observers to take the cotyledon 

 colour of a Pea from the outer surface only : that is, the surface im- 

 mediately beneath the seed- coat (cf. Weldon, 1902, pi. I.) ; but in my 

 experiments I have not found that to be a true test of cotyledon colour. 

 I prefer to separate the cotyledons and observe the internal colour also. 

 True yellow cotyledons have usually the outer layer yellow, while the 

 larger area within is creamy yellow, shading to white. True green 

 cotyledons have usually the outer layer green, though sometimes this is 

 yellowish or even clear yellow, but the larger area within is always green, 

 shading to a lighter green. In the 1 Eclipse ' race, for instance, I have 

 found odd seeds which, when peeled, had apparently yellow cotyledons 

 instead of the usual green, but when the cotyledons were separated they 

 were found to be quite green in the large area within. 



In dealing with cotyledon colour it therefore seems necessary to 

 observe the inner surface of the cotyledons as well as the outer. 



The sub-races of the 'British Queen,' used as a parent in the ex- 

 periments, had the outer surface of the cotyledons a deep shade of yellow, 

 which, following Professor Weldon's colour scale (1902, pi. I., figs. 1-12), 

 compares best with fig. 5 ; the inner surface of the cotyledons was 

 creamy white. This dark yellow sub-race proved quite constant from 

 seed. The sub-race of 1 Eclipse,' used as a parent in the experiments, had 

 the outer surface of the cotyledons a deep shade of green, which was a 

 full shade darker than the darkest green shown on Professor ^Yeldon's 

 colour scale, fig. 1 ; the inner surface of the cotyledons was a lighter 

 shade of green. This dark green sub-race proved quite constant from 

 seed. 



First Generation (F,). 



Experiment 1. — 1 British Queen ' ? x 'Eclipse' J. 

 (Yellow ? x Green S .) 



As previously described, a single plant of ' British Queen ' was crossed 

 with pollen of a single plant of ' Eclipse,' and produced 19 hybrid seeds, all 

 of which had yellow cotyledons when dried, thus once more demonstrating 

 the dominance of yellow cotyledons over green, and confirming the previous 

 experiments of Mendel (1865), Correns (1900), Tschermak (1900, 1901), 

 and Bateson (1902). Three of these hybrid yellow seeds were peeled and 

 the cotyledons separated ; two were retained whole for reference, and the 

 remainder were sown, a slight portion of the seed-coat having been 

 previously removed from each to compare the actual shade of the outer 

 surface of the cotyledons with those that were peeled and separated. The 

 outer surface of the cotyledons in all these dominant hybrids was found 

 to be an intense shade of yellow, but curiously enough all were uniformly 

 a darker shade of yellow than the pure dominant parent 'British Queen,' 

 thus being a shade darker than Professor Weldon's darkest fig. o (1902, 



P i. i.). 



This increased intensity of tint in the dominant hybrids of the first 

 generation was quite unexpected ; but, as the sequel shows, the results of 

 the second generation provide a clue to the mystery. The inner surface 

 of the cotyledons in those examined was creamy yellow, being also darker 

 than the pure dominant, which was creamy white. 



