Mutations and Evolution. 
145 
The cruciate character. 
Coming now to the cruciata type of mutation, 1 CE. biennis var. 
leptomeres of the Holland sand dunes differs from the species in 
having narrow, linear petals, somewhat crumpled and more or less 
greenish yellow. This constant form Stomps (1913) regards as a 
case of incomplete sepalody, in which the petals have become 
imperfectly transformed into sepals. In Epilobium he describes a 
somewhat different form, E. hirsutum cruciatum de V., which 
originated from a group of a dozen plants (perhaps derived by 
stolons from one) and breeds true. The plants were found by Mr. 
John Rasor in Woolpit, Bury St. Edmunds, and seeds were sent 
to de Vries. They differ from the type in having, in place of the 
large, obcordate, lilac petals, small greenish scales shorter than the 
sepals, and pointed. That this is not a case of arrested development 
is shown by examining the young buds of the species. The petals 
are broad and obcordate even while still small and colourless. The 
buds of the cruciate form show also a tendency to dry up and fall 
when young. This appears to be a more complete case of sepalody. 
Crosses between type and variety gave a Mendelian ratio. Thus 
E. hirsutum cruciatum x E. hirsutum gave an P 2 of 323 plants, of 
which 245 were type and 78 variety. Also the reciprocal cross 
produced 177 plants in F 2 , 130 of which were type and 46 variety, 
while one failed to flower. The sum of the two ratios is 375 : 124, 
showing that the cruciata mutation is a simple Mendelian recessive 
character. 
Not only does the cruciate mutation from (E. biennis differ 
from that of E. hirsutum externally, but also in its hereditary 
behaviour. For in reciprocal crosses of CE. biennis cruciata with the 
parent type, while the Fj is normal and the F 2 splits into the two 
original forms, yet the ratios are not Mendelian. Stomps 
unfortunately gives no figures, but states that whichever form is 
used as the male parent, that form appears in the greater number 
of the F a progeny. The cruciate forms in these two genera, 
while comparable in general, are not therefore strictly parallel 
forms. 
Elsewhere (1913a, p. 25) the writer has shown that, whatever 
the origin of CE. lamarckiana var. cruciata , both normal and 
cruciate flowers may occur on the same plant. This somatic 
1 »See also Mutation Factor , p. 21. In addition to the specimen collected by 
Ernst de Vries, others have been found on the Liineburg Heath, and they 
have been reported from other parts of Holland, probably from independent 
mutations. 
