1918] DEVRIES—MASS MUTATIONS 413 
described, in the beginning of the flowering period. The percentage 
figures for the appearance of types R and T correspond to those 
derived from our table for ovata, which were 24 and 4.5. The 
figure for lorea is a low one, but in the cross O. grandiflora X biennis 
we have seen that 43 per cent lorea were split off in the second 
TABLE XI 
-_ Secon ¢ GENERATION OF CROSSES OF QO. I desiierides lorea; CULTURES OF fee 
| Second 
Cross | genera- ovata | lutea R T  |nanella} lorea | Total 
tion 
O. grandiflora lorea X<La-| 
marckiana 50 630 ys Lovatal: S44 .ic0: 25 7 ° 6 gz 
oO. ico rs loreaXLa-| ° 
marckiana .;........-°}Intea |. i... 51 4 ° ° 28 83 
O. player lg grandi-| 
Hora fovea oe, ovate S805. 26 | 10 re) 2 96 
O. Lamarckiana X grandi-| | 
Cte teria CER Noe ac 68 2 ° 2 9 81 
O. grandiflora lorea X nan- 
Pee eG MA ie eae ovata} 51 a 15 3 ° 2 71 
0. “grandifor lorea X nan- 
RPG IS ne CP ae etek too | 6 ° I 9 66 
TOA ico ee 163 | 169 | 78 20 3 56 | 489 
Percentage: 205 1, 6 | 16 4 I aE fett> es 
generation. The question whether this phenomenon conforms to 
the formula of Mendel for monohybrids remains to be answered by 
more extensive cultures. 
F. MASS MUTATIONS, CONSIDERED AS SECONDARY MUTATIONS 
After describing the facts observed in my cultures and experi- 
ments, we may now proceed to the discussion of the principle of 
BartLett, already quoted. He assumed that a fundamental 
mutation occurred in one of the two gametes in a generation pre- 
ceding that in which the mass mutation appeared. We are not 
concerned, however, with the question whether all instances of 
mass mutation are due to the same internal processes, but only 
with the problem of explaining the production of mut. ochracea 
irom O. grandiflora by means of BARTLETT’S suggestion. 
In order to proceed in an empirical way, and to rely as much 
as possible on analogy with well ascertained facts, I shall start from 
