The Significance of llic Atavists. 519 
Six plants had a terminal spike at the top of the main 
stem with a small comb; in six others this comb was ab- 
sent, although small fasciations occurred here and there 
amongst the lateral branches. These six plants were 
selected as seed-parents, and from their seeds I obtained 
the second generation (1894), in which about half the 
individuals bore terminal spikes without a comb. There 
were 41 of these. The remaining 53 had combs ranging 
from 1-3 centimeters broad and were thrown away. Of 
the selected atavists, several produced lateral combs later, 
and each one of these plants was destroyed as soon as 
I discovered the anomalv, until at the end of the harvest 
J * 
period there were left only five plants which, though 
profusely branched, manifested no trace of fasciation. 
In the following year there were 29 plants with small 
combs and 6 without any at all ; these latter were again 
isolated as seed-parents. In 1896 I had 38 individuals, 
all of which, without exception, produced combs, the 
length of which ranged between 2 and 8.5 centimeters 
with a mean of 4.5. Only one plant had no comb at the 
apex of the stem ; but had a much flattened lateral branch 
instead. This, and the plants with the smallest terminal 
combs, were selected as seed-parents. Again, in the 
following year (1897), no progress was made, inasmuch 
as only a single weakly individual failed to produce a 
comb. Therefore I gave up the experiment and I con- 
clude that complete atavists are very rare in Celosia 
cristata and that even under repeated selection in the 
minus direction they will only be obtained in very small 
quantities. At present at least there seems not to be any 
prospect of obtaining a pure atavistic strain. 
The experiments I have described show that, as a rule, 
there is no sharp limit between the fasciated individuals 
