1912] HARRIS—STAPHYLEA 213 
bad, especially in 1907.%° These statistical difficulties do not effect 
the fundamental conclusion to be drawn, and it is in perfect accord 
with that from the individual shrubs. There is a sensible corre- 
lation between length of pod and number of seeds, quite independent 
- of the relationship between number of ovules per locule and length. 
TABLE VI 
OvULES LENGTH OF POD, 1906 
FORMED 
PER POD 
3 4 5 6 ‘- 8 9 Io II Totals 
easyer Ae I a os : I 
1 peas we! ee on na ae 
2 te eae I a 2 a . 3 
£8 oes 4 II 9 2 3 w+ mr 29 
0 5c fs 16 se 8 8 2 Ly 58 
20a. I 8 14 24 19 12 5 ne 83 
Bie 2 37 35 24 25 7 4 130 
3 es “ 7 30 63 33 Bs Ne “ 167 
SP ele Xt 14 41 90 69 20 15 : 251 
- | ers ey S38 | x87. .}. 107.01 foe 57 27 7 z 519 
Ae ay ie 4 20 45 43 46 23 12 I 194 
aia he 12 26 38 34 12 9 3 134 
or gachenl ne Io 24 30 29 II 3 5 Sesser’ 
Oe, ate 6 24 24 16 15 6 5 96 
Ee eee 7 19 29 27 14 9 4 sc 
E * eehe ee I 8 26 29 36 13 5 4 a 
St oe s II 3 5 + 27 
<b Re ieee Me 4 2 I 7 
ko Seen I I I me: 4 
i AR ¥ I 2 
Bs be Bear a nee : z 
BO OES I Ir 
Totals 6 | 142 | 437 | Gro | 462 | 248] t32 32 I | 2050 
C. TOTAL SEEDS PER FRUIT AND FRUIT LENGTH 
The normal fruit of Staphylea has three locules. In considering 
the relationship between fertility and length, should we correlate 
between the number of ovules or seeds maturing per locule, or 
should we determine the relationship between the total number 
One must remember that classes 7, 8, and 9 exert a powerful influence on the 
slope of the line, while the extreme classes are represented by so few cases that great 
irregularity is to be expected, even when dealing with as large numbers as we have 
here. While a straight line does not represent our empirical means at all well, I 
eran see that an 5 equation of a higher: order would do any better. Perhaps this 
gularity p the low value of 7,, calculated from the general samples. 
