Mutations and Evolution. 
148 
Cat . under the name Lysimachia corniculata non papposa, Virginiana 
major , jiore sulphurea. 1 It must then have been appearing as a 
mutation in the original stock brought from America. Other wild 
American self-pollinated species recently brought into cultivation 
have been shown to be doing the same thing, i.e., producing 
mutations with varying frequency. We must recognize that the 
germ-plasm of a species is a vast aggregation of units, among which 
complicated relationships arise which express themselves in 
throwing off at intervals new forms. 
The frequency with which the mutations in CE. biennis appear 
has been studied by de Vries (1915b). He grew 8500 plants from 
the pure line of Stomps, derived from a single wild plant from 
which they were descended in the third and fourth generation. 
They included 8 dwarfs (mut. nanella) or 0*1%, and dwarfs also 
occurred in later generations. The population of 8500 also 
contained 4 semigigas, or 0*05%, 2 and 27 sulfurea, or 0*3%. The 
latter appeared among the offspring of all seven parent plants. 
This gives a total of 39 mutations, or 0*45%. This is a lower 
percentage than in (E . Lamarckiana, but much higher than in 
Drosophila. Cultures of CE. biennis sulfurea from four of Stomps* 
mutants also gave two dwarfs among over 1000 plants which 
flowered. These give the race CE. biennis mut. (1913) sulfurea mut. 
(1914) nanella. In this connection de Vries also obtained two 
other mutations from crosses. (E. lamarckiana x CE. biennis 
nanella yielded 55 plants, which had the characters of lamarckiana 
X biennis , except one lata mutant. Again, in the cross CE. biennis 
semigigas X CE. biennis, he obtained 19 plants, 1 dwarf mutant, 
10 biennis with 14 chromosomes, and 8 of a new type (described else¬ 
where) with 15 chromosomes. The appearance of two uniform 
types in this cross is significant, for it shows that the meiotic 
processes were taking place so as to reduce the chromosome 
number in the gametes to 7 or 8. Moreover the fact that the 
15-chromosome type was uniform indicates that the same chromo¬ 
some must have been the extra one in every case. 
This brings us to the meaning to be attached to the term 
parallel in connection with mutations. It is obvious that each 
mutation of a given type is parallel to every other of that type. 
1 See Mutation Factor , pp. 66, 158. The fact that the early writers, 
although they recognized sulfurea, found no other variations of CE. biennis , 
shows that it was as uniform then as it is now. 
2 If mut. gigas , appeared only from the union of two diploid gametes, its 
frequency should then be about 1 : 4,000,000. 
