xvi] Practical Hints 295 



F z those whites which are quite devoid of colour, are pure 

 to whiteness. The crimson-reds are pure to redness. The 

 magenta types may be pure to magenta, or may throw the 

 hypostatic reds. The darker colours are recessive to the 

 lighter, in the magenta classes, though the mutual relations 

 of the light to the dark among the reds are not so certain. 



The large eye, combined with a homostyle structure, is 

 recessive to the small or long-styled type. As mentioned 

 (p. 139) the dark blotches round the eye can appear only 

 in those plants which have coloured stigmas. Another 

 curious point is to be noticed in this respect : that when the 

 dark blotches are developed in a large-eyed type, the blotch 

 extends over the whole area included in the "eye." If the 

 stigma be green the eye is therefore yellow (Plate VI, fig. 20). 

 The stigma may be coloured, though the eye is yellow ; 

 but if the stigma be coloured and the factor for the blotches 

 is present, then the appearance of the deep red and deep 

 magenta flowers figured in the bottom row (Plate VI, figs. 19 

 and 21) is produced. Such types are exceedingly distinct, 

 and might most naturally pass for different species. More- 

 over, of these singular varieties the red would always breed 

 true in both shape and colour, so that the illusion in that 

 case would be complete. The deep magenta type might also 

 be homozygous and breed true, but some individuals would 

 throw the red variety also. All the offspring of these 

 dark-eyed types would of course have the large dark eye. 



As regards the relations of the various types all that 

 can be positively asserted is that the chief classes are 

 (i) pure white, (2) tinged white, (3) crimson-reds, and 

 (4) magenta-reds. There is some grading between the two 

 red classes, and between the pure and the tinged whites. 

 The distinction between the two kinds of red also cannot 

 be followed among the tinged whites. The numerical pro- 

 portions are therefore not quite certain, but presumably 

 the various whites collectively are 12, magenta-reds 3, and 

 the crimson-reds i. 



In the plate no attempt is made to indicate these numerical 

 proportions and the flowers there shown are simply chosen 

 as representing the most distinct types which occur. The 

 polymorphism of such an /% family is greatly increased by 

 the existence of extreme diversity in the shapes of the 



