HARRIS: OVULES AND SEEDS IN CERCIS 530 
TABLE III 
DIFFERENCES AND PROBABLE ERRORS OF DIFFERENCES IN THE STANDARD DEVIATION 
OVULES AND SEEDS AND IN THE CORRELATION OF OVULES AND SEEDS 
OR SELECTED INDIVIDUALS OF CERCI 
Ovules | Seeds ; Ovules and Seeds 
Individuals wo sani Garren a ea LF 
Compared | erences in Differences in | ; Differences in | é 
| Sta ie Devadis! | Ratio | | Standard Deviation | Ratio | Correlation | Ratio 
Tand 5 -193 £.019 | 10.28 | -189 4.024 7.93 sO7T E025 || 2.77 
7 and 9 187 -b.021 9.04 -291.028 | 10.47 | .053.028 | 1.93 
and 1 | 219 +.037 5.87 .9706,053° =| 6,00 -089+.036 | 2.44 
4 and 118 146 +.02 | 8642) ioe OSS a. Os4 1202 .073 4.027 | 2.72 
and 9 -292+.016 | 18.04, .441+.022 20.32 | .048-.025 —-T.91 
6 and 10 -TOO+.051 | £.98 |'\- .TET+.073 1.52 -302+.065 | 4.66 
TABLE [V 
RELATIONSHIP BETWEEN NUMBER OF OVULES PER Pop AND NUMBER OF SEEDS PER 
| 
Tree | spa CoP. “os You ro2/Er,, 
- 695 | -7807 | 5734-017 —.074+.025 2.93 
a 274 | .6709 | .488 +.031 —.060 +.058 1.03 
3 242 | | {8189 | .679 +.023 +.112 +.063 ae te, 
a: | 318 | .8323 | 648 +.022 +.047 +.056 82 
S| 698 | .8304 | 502 4.019 —.1522.025 6.08 
6 IIr | .7452 | 342 4.056 —.178 +.092 1.94 
7| 418 | 8062 | 497.025 —,.128 +.032 3.94 
8 | 672 | 8311 -556-+.018 —.043 4.026 1.64 
9| 1,942 7724 | 550+4.012 —.016 +.015 1.03 
to | 152 | .£8243 | .645 4.032 +.081 —.080 1.01 
115 | 400 | .8420 | 07 4.021 —.028 +.034 82 
116 | 880 | .7987 | 531 4.016 | —.062 +.023 2.73 
ies <a at Renee -720.015 | —.070 +.032 2.19 
4-76, 5.13, 5.54 and 6.15 with probable errors always in the 
second place of decimals and ranging from .016 to .057 for ovules 
are so clearly significantly different that it is needless to calculate 
probable errors. The differences for mean number of seeds are 
also clearly significant. 
The standard deviations for both ovules and seeds present a 
problem more difficult of solution by mere inspection. Values 
like .63, .68, .71, .74, .78, .85 and .93 with probable errors not 
exceeding .04 in any case appear to be significant. Differences 
exceeding their probable error as widely as those for the few 
random pairs given in Table III certainly indicate that the trees 
must be regarded as individual in variability as well as type 
For both ovules and seeds the coefficients of variation differ 
