90 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9 
in the unresolved groups, which might be resolved further by more careful 
work. 
Turning now to an analysis of the data presented in the tables, it appears 
that there are two strains, nos. 13 and 33, each of which cannot be like any 
other strain in the series. Two others, nos. i and 2, fall together and 
cannot be thrown with any other group. No. 12 can be associated only 
with no. 35. However, nos. 9, 10, and 34 form a natural group, with which 
no. 35 can also be associated. The latter, falling between no. 12 and the 
group composed of nos. 9, 10, and 34, can be placed with either, and it is of 
course a matter of judgment where it should go. No. 12 has the appendages 
consistently longer than the spores, whereas in nos. 9, 10, 34, and 35 the 
appendages are consistently shorter than the spores. A group consisting 
of nos. 9, 10, 34, and 35 is therefore adopted, leaving no. 12 to stand alone. 
Applying our criteria to nos. 34 and 35, we find that they might have 
been associated with nos. 31 and 32 if they had not been placed with nos. 
9 and 10. We have therefore left nos. 31 and 32 to be associated with other 
strains, which prove to be nos. 8, 28, 29, and 30. We might therefore 
segregate a group consisting of these six strains. Comparing the generation 
means for spore length and appendage length, however, we find that there is 
decided justification for considering that such a group would constitute an 
unnatural assemblage. In nos. 31 and 32 the spores are consistently, 
generation after generation, longer than the appendages; in no. 8 the exact 
contrary is the case; in nos. 28, 29, and 30 the spores and appendages are 
of the same mean length. Such consistent correlations cannot be looked 
upon as due to mere chance, and we are therefore forced to consider nos. 
31 and 32 as constituting one group, and no. 8 another. Nos. 28, 29, and 
30 come next to no. 7 in spore length, and this form proves distinct on the 
basis of appendage length. The next shorter strain, no. 27, proves to be 
distinct from no. 32 as to spore length. We have no choice, then, other 
than to consider nos. 28, 29, and 30 as a group. 
The remaining twenty strains offer more difficult problems. Probably 
no two persons would group them in exactly the same way, but the results 
would be very similar. Tabulating the possible likenesses brings out the 
fact that there are three outstanding strains which are difficult to associate 
with any considerable number of others. These are nos. 7,11, and 14. No. 
7 falls just on the border of an otherwise relatively uniform group consisting 
of nos. 3, 4, 19, 21, 22, 23, and 24, and may be thrown with them or kept 
apart, as seems preferable. We consider it a distinct strain. Nos. 26 and 
27 fall together, and cannot be combined with the large group last mentioned 
because of the significant difference between nos. 26 and 24. Nos. 11 and 
25 are not particularly alike, but the data do not show a significant difference 
between them. Nos. 5, 6, 15, 16, 17, 18, and 20 form a relatively uniform 
group. No. 14 is best kept by itself. If it is thrown into one of the other 
groups it necessitates the removal of some other strain or strains from that 
