Miscellanea 
443 
basis of data already tabled may be stated : Is an individual flower more (or less) likely to 
be abnormal if it is borne on a large * inflorescence than if it is produced on a small one 1 
If the chances of an individual flower being abnormal were in no way influenced by the 
number of flowers per inflorescence, but the chances of belonging to a synanthy were equal for 
all flowers, abnormalities would be distributed among the inflorescences in proportion to their 
number of flowers, and a correlation between the number of flowers and the number of abnprmal 
flowers would result. Therefore this correlation coefticient will not give us the information we 
desire. What we need to know is the correlation between the number of flowers per in- 
florescence and the deviation of the number of abnormal flowers from the probable number 
if abnormalities were distributed among the inflorescences in proportion to the number of 
flowers they produce. 
TABLE IV. 
Spiraea Van Houtii. 
Number of Flowers per Inflorescence. 
5 
6' 
7 
8 
9 
10 
11 
12 
IS 
H 
15 
16 
17 
18 
19 
20 
21 
22 
23 
2^ 
25 
26 
Totals 
0 
2 
3 
5 
1 
5 
1 
5 
1 
8 
2 
6 
1 
9 
9 
1 
13 
5 
1 
16 
3 
2 
22 
3 
3 
17 
5 
7 
17 
7 
3 
12 
9 
3 
1 
18 
10 
6 
1 
21 
7 
10 
12 
10 
2 
14 
6 
1 
17 
6 
1 
9 
4 
3 
9 
1 
3 
1 
1 
1 
1 
234 
91 
45 
8 
2 
3 
1 
1 
Totals 
1 
0 
6 
6 
10 
7 
19 
19 
21 
28 
29 
29 
25 
35 
41 
24 
21 
24 
16 
10 
6 
2 
379 
Table IV. shows the number of flowers per inflorescence and the number of flowers counted 
as abnormal in each inflorescence. Since the abnormality with which we are dealing is the 
fusion t of two flowers there can be no group of "one flower abnormal" ; each abnormality was 
considered as two or more abnormal flowers. Our series of data is not large but it has the 
advantage of being taken all from one clump of plants, probably all arising from the same 
individual, at the Missouri Botanical Garden. 
The correlations are : 
= A11 flowers, 
,= •121 + -034. 
= Abnormal flowers, 
= - -071 + -034. 
With probable errors so large as these any trustworthy conclusion is impossible. But the 
negative sign of the r^.^ coefficient will be sufficient to warn one against concluding that the 
larger (and so presumably the more vigorous, or " better nourished ") inflorescences are more apt 
to produce abnormal flowers. They produce actually more, but this is only to be expected ; the 
question is whether they produce relatively more. 
Finally, I would again warn the reader against any generalization from the above illus- 
trations. They are selected to show the kind of problems to which the biometric method 
described may be pi'ofitably applied. 
* Here the only measure of the size of the inflorescence is the number of flowers which it produces, 
t I wish to express no opinion whatever on the embryological origin of the anomaly, and the term 
fusion is used simply because it is the conventional and convenient term for such cases as these. 
56—2 
