Review 
302 
only corroborated the independent results of Miss Lutz and Gates 
regarding the constancy of the fifteen chromosomes in O. nuit. lata, 
hut have found the same number constantly in semilata and also in 
two cases in which lata foliage, occurring as a mutation, is combined 
with other characters inherited from the parents . 1 
Thus 0. biennis mut. lata appeared in a race of normal O. 
biennis , having lata foliage and biennis flowers (15 chromosomes). 
Again. O. mut. lata rubricalyx occurred in the F 2 of 0. rubricalyx 
X O. grandiflora. The great hulk of these plants were blends and 
combinations of the characters of the parents, but lata rubricalyx 
had lata foliage and habit together with the red pigmentation of 
rubricalyx. 'Die possession of fifteen chromosomes by this plant 
also shows that whenever a meiotic irregularity leads to the 
formation of an individual having an extra chromosome, such a 
plant will have the leaves and habit of lata or semilata? It further 
shows the sharp contrast which must be drawn between sporadic 
mutations and the regular processes of inheritance either in pure 
races or in hybrids. 
We may conclude that the work of De Vries and other students 
of (Enothera has resulted in showing that really new characters 
may and do arise by a germinal change, and are not merely 
recombinations of the characters of hybrids. The importance of 
this conclusion is very great at a time when we have been asked to 
suppose that all evolution has been accomplished by the shuffling 
and successive loss of fixed Mendelian unit-characters. 
REVIEW. 
BRITISH VIOLETS: 
A monograph, by Mrs. E. S. Gregory; with an introduction by 
G. Claridge Druce, M.A., F.L.S., Cambridge (W. Heffer & Sons, 
Ltd.), 1912 ; pp. xxiii and 108 ; Price 65 . 6cl. 
I N this work, the twelve British species of violets are described, 
including twenty-seven varieties, and numerous hybrids and 
hybrid-forms. Of the species, three (V. calcarea, V. epipsila, and V. 
montana) are not to be found in the current British floras. V. 
calcarea is closely allied to V. hirta, and was formerly included in 
V. hirta var. calcarea. V. epipsila is closely allied to V. palustris, 
and is by some botanists considered a variety of it; and V. montana 
1 Evidence which shows that Mutation and Mendelian splitting are 
different processes. Section K, British Association, Birmingham. 
s It is possible that one of two other mutants also have an extra 
Chromosome, 
