910 
Journal of Agricultural Research 
Vol. XXX, No. 10 
Fig. 12.—Actual (solid line) and theoretical (broken 
line) frequency distributions of the head lengths 
of untreated swine spermatozoa. Recovered 
from the vagina of a sow after natural service. 
The probability that the actual distribution is 
only a chance deviation from the theoretical is 
0.26. This is from the same sample of semen as 
was used in the preparation of Figure 13. (Slide 
No. 522) 
z&v&7-/z? 0/^ zr/tteAwr&zzw /z£wz& 
Fig. 13.—Actual (solid line) and theoretical (broken 
line) frequency distributions of centrifuged 
swine spermatozoa. From near the inside end 
of the centrifuge tube after five minutes of centri¬ 
fuging at 1,600 r. p. m. From the same original 
sample of semen from which Figure 12 was made. 
The probability that the actual distribution was 
only a chance deviation from the theoretical 
distribution is 0.008. (Slide No. 526) 
Fig. 14. —Actual (solid line) and theoretical (broken 
line) frequency distributions of untreated swine 
spermatozoa. The spermatozoa were obtained 
from the testicle of an aged Poland-China boar 
immediately after castration, and the milky fluid 
which was recovered from the epididymis was 
diluted with about six volumes of Ringer’s 
solution and slide 537A was made directly from 
this. The probability that the actual distribu¬ 
tion is only a chance deviation from the theroetical 
distribution is 0.55. (Slide No. 537A) 
The microscopic work on the rabbit 
spermatozoa was directed toward the 
answering of two questions: First, 
whether there really are two different 
sizes of spermatozoa produced in 
approximately equal numbers, and 
second, whether the centrifuging tends 
to separate those two sizes. The 
answer to the first and more funda¬ 
mental question is not certain. One 
of the four curves of normal sperma¬ 
tozoa is distinctly bimodal, one is 
distinctly unimodal, and the other two 
are broader at the top and narrower 
in the middle than the theoretical 
curve but agree with it well at the base. 
That is the result which would be 
given if the curve were a compound 
one composed of two equal normal 
curves whose modes were not very far 
apart. Since the size difference be¬ 
tween the X-containing and Y-con- 
taining spermatozoa is expected to 
be very small (1), these flat-topped 
curves agree very well with the idea 
that there really are two sizes of rabbit 
spermatozoa, but that the means of 
those two sizes are too close together 
for the modes to show separately in 
the curves. All three of the curves of 
centrifuged spermatozoa are unimodal, 
and the two for which there is a con¬ 
trol (figs. 2 and 3) seem to indicate in 
both cases that it is the larger sperma¬ 
tozoa which were separated out both 
to the inside and to the outside of the 
tube. 
Those spermatozoa which were 
plainly distorted or broken were 
not measured, but the distortion of 
many of them may have been so 
slight that it could not be detected 
by inspection. Since the answer to 
the two principal inquiries is con¬ 
cerned only with the average distri¬ 
bution of the two classes of sperma¬ 
tozoa, the extreme variants have 
very little importance in this problem. 
Therefore, in figuring the goodness 
of fit of the actual results to the theo¬ 
retical, the only classes considered 
were those in which the theoretical 
number exceeded six. This arbitrary 
limit includes the great majority of 
the individuals and gives a fair in¬ 
dication of the goodness of fit of the 
actual distribution to the expected 
distribution without giving undue im¬ 
portance to any chance variation 
where the expectation is small. It 
will be noticed that the probability 
is fairly high for all of the centrifuged 
slides as compared to the normal ones, 
there being only one normal one with 
a value as high as either of the three 
centrifuged slides. This indicates that 
the centrifuged samples are more 
homogeneous with respect to head 
length of spermatozoa than untreated 
samples are. 
The general conclusions to be drawn 
from the microscopic work on the 
rabbit spermatozoa are that there are 
indications of a dimorphism in the 
head lengths but no evidence com¬ 
plete enough to establish it beyond 
question. The two sizes differ so 
