On Linaria alpina. 169 
suffused with blue so that the flowers appeared entirely self-coloured. 1 
In the F 2 generation obtained by allowing the F x plants to inter¬ 
breed, the numbers indicated the simple Mendelian ratio of three 
without orange to one with orange on the palate. 2 In a case where 
a concolor plant supposed to be pure, but which proved to be a 
crossbred, was crossed back with the type, the offspring were mixed 
in approximately equal numbers. 
These results show that in Linaria alpina the pink colour of the 
variety rosea is recessive to the blue colour occurring in the type and 
in the variety concolor; and that, as regards palate character, presence 
of orange in the type and in var. rosea is recessive to its absence in 
var. concolor. The dominance of blue over red shades is a relation 
with which we are now familiar in several genera, e.g., Stocks, 
Sweet Peas, Salvia Horminum and others, and in the present case 
as in these other genera the effect is no doubt due to a factor which 
when present turns the red colour to blue. The dominance of the 
palate character, absence of orange, was somewhat unexpected in 
view of the fact that in another species— L. vulgaris, where the type 
form is distinguished by the presence of a similar orange-coloured 
area on the palate, and the variety (perlutescens) by its absence, the 
relation between the two has been found by de Vries 3 to be exactly 
the opposite. Crossing L. vulgaris (pale yellow with saffron-yellow 
to orange palate) with the variety perlutescens (palate as well as the 
rest of the flower pale yellow) de Vries obtained plants all having 
the deep yellow palate. A repetition of the experiments in a later 
year gave the same result. In the F 2 generation the proportion of 
the two forms was about 75% with deeply coloured palate to about 
25% self-coloured. We thus have the somewhat unexpected result 
that in two species belonging to the same genus, each having a type 
form and a variety differing from the type apparently in precisely 
the same respect, viz., absence of a particular colour from a corres¬ 
ponding region of the corolla, the presence of this colour behaves 
as recessive to its absence in the one case (alpina), and as dominant 
in the other (vulgaris). 
1 Except in such case as that mentioned above when the end of 
the line of the beard becomes visible and just shows a point 
of orange at the closed lips. 
2 Among the concolor plants of the F 2 generation one individual 
was found to have a single stem bearing flowers showing some 
orange on the palate while all the flowers on the many other 
branches were typically self-coloured. No other instance of 
what seems to have been a bud variation was observed in any 
of the other plants. 
s Die Mutationstheorie II, 1, 1902, p. 152, 
