252 MUTATIONS OF OENOTHERA SUAVEOLENS DESF. 



and counted in August, at the time of flowering, 8 jastigiata and 

 19 jaculatrix. If we may assume these small experiments as suffi- 

 ciently reliable, we may conclude that there is a coefficient of mu- 

 tation of about 0.5 percent for fastigiata and of about 1 percent 

 for jaculatrix. 



Besides these two mutants other narrow-leaved forms of Oe. 

 suaveolens occur from time to time. One of them I got through the 

 kindness of Mr. Armando Cortezao from Coimbra in Portugal. From 

 these seeds I had a uniform culture in 1914, in which 15 specimens 

 flowered. The leaves were almost as broad as those of the species 

 but more smooth and thin; the whole stature lower. Two other 

 types occurred among seeds sent by Prof. Blaringhem from Fon- 

 tainebleau, but from another station than that which yielded my 

 race. Their leaves were half as broad as those of the species, mea- 

 suring, e.g., 15 x 2.5 cm. as compared with 15 x 5 cm. One of 

 them had large and the other small flowers. It seems that the width 

 of the leaves is as variable in this species as it isinO. Lamar ckiana. 



At the time of protruding its stem, Oe. fastigiata makes broader 

 leaves, and soon they become as broad as those of the species, but 

 remain shorter. The foliage is denser but of a lighter green and smooth, 

 and even at this time the plants cannot be mistaken for Oe. suaveolens. 

 Afterwards the branches are erect, running almost parallel to the 

 main stem, whereas in the species they stand out under an angle 

 of 40° or more. The flowers are also erect, and pressed against the 

 stem, making the spikes narrow instead of loose. The fruits are 

 smaller and thicker, measuring, e.g., 25 x 7 mm. and not turned 

 outward. The whole plant remains much lower than the species, 

 reaching often less than one meter in height. 



I succeeded in saving self-fertilized seeds of 7 specimens of Oe. 

 mut. fastigiata, but they yielded only a small progeny. Among 

 this, no single specimen recurred to the type of Oe. suaveolens, but 

 on the other hand the more common mutations of this species were 

 repeated, and this by almost every parent. All in all I counted 262 

 seedlings, of which 141 were fastigiata, 34 jaculatrix and 87 lutescens, 

 giving percentages of 54, 13 and 33. Almost all of these plants 

 flowered in August and September 1916, making a bright show of 

 the three new types. It is very interesting that this mutant should 

 produce the two others in such high proportions, but the same 

 condition prevails among the offspring of Oe mut. jaculatrix. Thus 

 we see that Oe. mut. fastigiata and Oe. mut. jaculatrix are very 

 mutable derivatives, whereas Oe. mut. lutescens is almost immutable. 



