Origin and Development of the Composite. 69 
explained by Mendelism (cp. 11). This difficulty is avoided by 
some Mendelians who claim that little or no progress but only loss 
of characters and diversification have taken place (cp. Bateson, 35, 
p. 89). 
An interesting point is the variability of Mendelian factors or 
genes. That they are invariable is accepted more or less as an 
axiom by many Mendelians and the point was made precise by 
Johannsen (31) in the theory of pure lines or genotypes. Cockerell 
(IV, 19), who finds that a few genes or potentialities can give many 
varieties, concludes that new genes, though rare, do occur in the 
annual sunflowers. The variability of genes is upheld by Castle 
(12), and also by Stout (50) and Harris (23-24) :the last two authors 
criticise Johannsen’s work andthe genotype theory very adversely,and 
this criticism is supported by the details of Jennings’ work on 
Paramaciutn (30 and Gates 35). 
Mutations. 
That mutations or discontinuous variations occur in the 
Compositae is clear from the evidence which has already been men¬ 
tioned (see Chap. IV, A). The truth of the phenomena described 
by De Vries (IV, 87-88) has never been denied, but Mendelians 
claim that all the phenomena can be explained by hybridisation and 
subsequent segregation, with occasional loss of genes (cp. 35). 
White (56) claims priority for his observations (1898-99) of 
mutations in Lycopersicum. Cockayne (X, 15) makes a similar claim 
for Armstrong (1881). Davenport (14) mentions mutations which 
are sterile with the parent species (the accepted proof of specific 
differentiation). Willis throughout his work on Age and Area 
supports the views of De Vries, especially in his earlier papers 
(58-60, X, 74 and IV, 93) and in his special studies of the hill-top 
floras (IX, 43-45). The facts in the latter papers are supported by 
those given by Gibbs (X, 31). 
Two interesting points are mass mutation and tetraploid 
mutation. The former is described by Bartlett (3-4) as the produc¬ 
tion of a large proportion (sometimes 100% of progeny) of mutants 
from certain parents. These mutants are themselves very 
unstable but do not throw off the type form of the species. This 
phenomenon is inexplicable by Mendelian segregation, but the loss 
of a factor giving a 100% mutant progeny is quite in accordance 
with Mendelian ideas. Mass mutation is also another possible 
reason for the occurrence on islands of endemic species of genera 
which are otherwise unrepresented in the same region (cp. 
Differentiation, Chap. X, A). 
