232 
PACIFIC SCIENCE, Vol. 1, October, 1947 
Digit I of the wing is riot separable in the 
two series; the ranges overlap, but the means 
of total length are statistically different. 
Length of phalanx 1 of digit II may or may 
not be a good criterion to separate the spe¬ 
cies, for tests (Table 15) show that the 
means and variances could have come from 
a single stock, although the tests for differ¬ 
ences in the means only show them to be 
significantly different. Phalanx 2 of digit II 
shows a highly significant difference in 
mean length between the two series. The 
length of digit III is in no way significantly 
different in the series under study (Tables 
5 and 15). In all measurements of length 
of digits the coefficient of variation is greater 
in am plus, but this fact results at least in part 
from the few specimens of californianus that 
were available. In the latter series the range 
of the sample is perhaps equivalent to about 
70 per cent of the actual range, but in 
am plus the sample is representative of more 
than 90 per cent of the whole population. 
Relatively few specimens of the sternum 
were available for study (Table 6), but 
statistically the series of calif ornianus may be 
representative of about 85 per cent of the 
actual stock. The series for am plus exhibits 
about 85 to 95 per cent of the probable 
range of variation of the total population. 
In every measurement of the sternum the 
ranges overlap. Thus, raw measurements 
are of no value in distinguishing the two spe- 
TABLE 8 
Spread of the Furculum (mm.) 
calij ornianus ampins 
Number of specimens.. 8 10 
Range. 81-102 92-114 
Mean . 96.0±2.5 100±2.3 
Standard deviation ... 7.1 ±1.8 7.3 ±1.6 
Coefficient of variation. 7.5± 1.9 7.3±1.6 
cies, although the fossil is larger in each 
instance. Significance tests (Table 16) on 
the means of sternal measurements show sig¬ 
nificant differences. Ratios showing intra- 
membral proportions of the sternum do not 
vary significantly between the two species 
(Tables 6 and 16). 
To judge from the extent of the standard 
deviation, the samples of the coracoids avail¬ 
able for the two series are adequate to repre¬ 
sent about 95 per cent of the total popula¬ 
tion. In no raw measurement can individual 
coracoids of the two forms be distinguished. 
In the fossil the coracoid is largest in every 
dimension, but comparable ranges overlap. 
Results of the significance tests show that 
the two series differ significantly in their 
means. The coefficient of variation for cora- 
coidal length is greater in am plus, but the V 
for depth of shaft is larger in calif ornianus. 
Although the mean spread of the furcu¬ 
lum is greatest (Table 7) in am plus, the 
means do not differ significantly (Tables 15 
and 16). 
The series of scapulae for californianus 
probably are representative of 95 per cent of 
TABLE 9 
Measurements (mm.) and Proportions of the Scapula* 
NO. 
SPEC¬ 
IMENS 
RANGE 
MEAN 
STANDARD 
DEVIATION 
COEFFICIENT 
VARIATION 
Length . 
Width middle of blade. 
{ 23 
\ 
105-120 
119-132 
9.1-11.9 
10.7-12.6 
115 ± .78 
124± 1.10 
10.7± .13 
3.74± .55 
3.78± .81 
.60 ± .09 
3.25± .48 
3.05± .65 
5.6l± .83 
l 11 
11.3± .18 
.58± .12 
5.13±1.09 
PROPORTION 
Width : length.. 
7.6-10.3 
9.3 
( u 
8.2-10.2 
9.2 
* First row of figures under each category pertains to californianus, the second row to amplus. 
