12 Harris, Observations on the Physiology of Seed Development in Staphylea. 
danger line of chance errors to attribute mucb biological significance 
to our results. 
Approaching the same problem from another siele, we examine 
the correlation constants for position and number of ovules failing 
to develop into seeds for the twenty individuals of the 1906 series. 
They appear in Table 2. 
These constants for r pf are eciually divided between positive 
and negative. Only three deviate frome 0 more than — . 100, and these 
only slightly. The average for the series is — . 0183 4-. 0109. 
Obviously no significance whatever is to be attached to this value, 
and we must conclude that there is no evidence that the position 
of a fruit on its inflorescence has anything to do with the failure 
of ovules to develop into seeds. 
Table 2. 
Shrub. 
• 
Tps 
Position of Fruit 
and Number of 
Seeds. 
r pf 
Position of Fruit 
and Number of 
Ovules Failing. 
11 
-.084 
.026 
12 
—.088 
—.040 
13 
-.128 
— .082 
14 
—.048 
.041 
15 
-.006 
.093 
16 
—.058 
—.108 
17 
-.018 
.013 
18 
.091 
—.090 
19 
.064 
.079 
20 
.063 
—.165 
21 
—.097 
—.000 
22 
.005 
—.046 
23 
.018 
.013 
24 
-.006 
—.027 
25 
—.009 
.026 
26 
—.070 
—.092 
27 
-.126 
-.146 
28 
—.025 
.068 
29 
—.017 
.036 
30 
—.079 
.033 
For r vs for the three general samples, we have: 
1906, Table X, r = ~. 0148 + . 0086. 
1908, Table XI, r = — . 0077 + . 0061. 
1909, Table XII, r = +. 0128 ± . 0085. 
How slender these relationships really are, may be best seen 
by expressing them in terms of regression. We have for the three 
straight line equations for position and number of seeds: 
1906, s — . 7426 —.0167 p. 
1908, s = . 8866 — . 0098 p. 
1909, 6- = . 7384 + . 0159 p. 
