152 Half-lioariness in Matthiola 



by the presence of a basal area or spot of a different colour from the 

 rest of the petal, being thus distinguished from the nearly allied 

 E. cicuturium in which all five petals are alike and unspotted. If 

 however the flowers of E. pimpinellif'oiiirni are recorded during a season, 

 it will be found that occasionally the earliest flower has all five petals 

 spotted, though I do not think I have ever observed more than one 

 flower of this type on an axis, and even one is not very common. One 

 or two succeeding flowers may show four petals spotted, occasionally 

 one may follow with three spots, and then the plant settles down to the 

 2-spot pattern until the close of the flowering season when a downgrade 

 series begins once more, the spots diminish in size until in the last 

 flowers they often cannot be detected at all. The point of interest is 

 that the earlier lateral axes also often produce at first a few flowers with 

 more than two petals spotted, although the first flowering stem may 

 have already passed on to the stage of producing only the 2-spot 

 flowers. We do not as yet understand the causes operating in these 

 cases, and a statement of the facts is all that is possible. 



The inter-relations existing between the half-hoary Stock and the 

 glabrous or fully hoary types are especially interesting as they reveal to 

 us for the first time a difference in the relations of the two factors H and 

 iT to a third factor. Hitherto in absence of evidence to the contrary 

 these two factors have appeared to stand in a complementary but 

 equivalent causal relation to surface character. Results obtained from 

 any series of matings involving the presence of H but not K could 

 equally be obtained by using the corresponding forms containing K but 

 not H, provided that the substitution of K for H in these operations 

 was made throughout. Such is not the case when the half-hoary form 

 is employed in these matings, and it is from the use of this type that 

 we obtain the first piece of evidence that the i-ole played by one factor 

 of the HK pair is different from and not merely complementary to that 

 of the other. 



The results of various matings in which this type was employed are 

 set forth below (p. 153). 



The condition here designated as quarter-hoary is quite definite, 

 and the individual exhibiting this character has a definite constitution 

 and behaviour. As regards appearance the quarter-hoary plant is at 

 first completely glabrous, but in the later stages of development leaves 

 are produced with marginal hairs, or with even a few on the surface. 

 Scattered hairs may be present on the ab-axial side of the pedicels and 

 a fair number occur on the basal region of the sepals, although not 



