1918] DEVRIES—MASS MUTATIONS 407 
as the corresponding hybrids between the parent species. No 
brunnea and no contraria were observed. The size of the cultures 
fully warrants these conclusions, but is not large enough to give 
reliable percentage figures. From these facts it is evident that 
among the triple hybrids of O. grandifloraXO. Lamarckiana one 
type, brunnea, is produced only by the non-mutated gametes of 
the first named parent, whereas another type, lutea, is produced 
exactly by the mutated ones. If we assume that one-half of the 
TABLE VI 
Crosses oF O. grandiflora ochracea 
contraria | Total 
lutea | brunnea 
Cross ; | Culture | ovata 
0 grandiflora ochraceaX 
Lamarckiana........... 1915 28 2 fs) ° 3° 
O. Lamarckiana X grandi- 
flora ochracea.......... 1916 s3 16 fo) ° 69 
O. grandiflora ochraceaX 
Lamarckiana nanella... . 1915 25 5 ° = 3° 
SOM a | 106 | 23 | 3 | ° " 
gametes of grandiflora are unchanged and the other half changed 
into ochracea, one-half of: the hybrids must result from the first 
group and the other half from the second. This shows that the pure 
and the ochracea gametes must produce each for one-half ovata and 
for the other their special hybrid. The figures, calculated in 
table VI, indicate 52 per cent ovata, 23 per cent /utea, and 21 per 
cent brunnea, and this corresponds as exactly as might be expected 
to our explanation. Thus we find: | 
50 per cent pure X Lamarckiana = 25 per cent 
ovata+25 per cent brunnea 
50 per cent ochracea X Lamarckiana = 25 per 
cent ovata+25 per cent lutea 
O. grandiflora Lamarckiana = 
This formula may be considered to explain the empirical results 
of our table, since it gives 50 per cent ovata and 25 per cent of each 
of the other hybrids. The empirical figures were 52, 23, and 21 per 
cent, as just mentioned. 
We may go still one step farther and introduce into our con- 
sideration the property of Lamarckiana to produce the twin hybrids, 
