J. A. Harris 
55 
The technique of counting was quite simple. Variation in the number of 
bractlets is not discontinuous but continuous, if by continuity one understands the 
presence of transitional conditions. These intermediate stages — appearing some- 
times as one or more bracts conspicuously smaller than the others, sometimes as 
large apparently compound or longitudinally fused bractlets — are relatively rare in 
five of the six species considered. In the preparation of the tables for our present 
use, it has seemed sufficient to throw them into the nearest integral class. An 
exception is made in the case of Hibiscus manihot. The bracts in this form are 
not subulate or filiform but broad and very irregular in size and completeness of 
division. Considerable numbers of fruits have been halved between two adjoining 
bract classes*. 
In the species in which the number of aborted ovules could be counted, both 
the number of ovules formed and the number of seeds developing per locule were 
recorded. Only fruits with the same number of locules were included in any 
correlation table j-. 
Analysis op Data. 
Hibiscus militaris, Cav. 
A series of Hibiscus militaris was collected on the margins of a lake in the low 
Illinois flood plain of the Mississippi river across from the Carondelet ferry at 
St Louis. The collection was made at about the middle of September and cannot 
safely be regarded as quite typical of the species in that habitat because the 
greater number of the pods had already dehisced and the collection is there- 
TABLE I. 
Hibiscus militaris. 
Seeds Developing per Locule. 
0 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
Totals 
8 
1 
2 
1 
1 
5 
9 
1 
2 
1 
2 
3 
3 
3 
15 
10 
1 
1 
2 
10 
11 
5 
12 
9 
26 
33 
58 
74 
56 
26 
1 
325 
11 
1 
1 
2 
6 
14 
12 
29 
31 
69 
124 
141 
220 
174 
93 
19 
4 
940 
12 
4 
5 
3 
16 
21 
34 
35 
58 
140 
186 
235 
157 
68 
19 
4 
985 
18 
2 
3 
5 
1 
7 
6 
12 
34 
39 
69 
114 
121 
112 
60 
9 
1 
595 
14 
1 
1 
1 
3 
1 
7 
17 
22 
22 
17 
13 
105 
15 
2 
1 
2 
5 
Totals 
2 
4 
7 
15 
20 
49 
45 
91 
112 
201 
389 
526 
678 
519 
260 
48 
9 
2975 
to 
o 
s-i 
pq 
O 
u 
<o 
a 
* In all the forms except Hibiscus manihot, the number of bracts furnishes a fairly satisfactory 
index to the amount of bract tissue, but in this species the irregularity in size is so great that number 
cannot safely be taken as representing mass. 
t Since we are dealing with integral variates, Sheppard's modification was not applied. 
In calculating the probable error of the constants for the bracts and for correlations involving the 
number of bracts, the actual number of fruits was used — not the number weighted with the locules as 
tabled in many cases. For the probable errors of the means and standard deviations of number of 
ovules or seeds per locule N was taken as the actual number of locules. 
