APRIL 12, 1912] 
raised several of these (uncastrated) males 
each year and they were remarkably uniform 
in horn development. Compare again Wal- 
lace’s figures, which are accurate, being photo- 
graphs. Arkell and Davenport’s statement 
concerning the “variability of the horned 
condition in the males of the Merinos” leads 
me to think they must refer to grade Merinos, 
certainly not to the pure bred ones. 
To my mind the evidence is clear that in 
‘pure Merino sheep castration does prevent de- 
velopment of the horns, and I have no doubt 
that in other breeds also castration has simi- 
lar though perhaps less conspicuous effects. 
In breeds which are horned in both sexes the 
males regularly have better developed horns 
than the females, and castrated males have 
smaller horns than uncastrated ones. See fig- 
ures in Wallace. 
If castration has the effect stated, the as- 
sumed nuclear inhibiting factor of Arkell and 
Davenport is quite superfluous. Their experi- 
mental results are fully in accord with those 
of Wood and are fully covered by the simple 
statement of Bateson that the horned char- 
acter is “ dominant in males and recessive in 
females.” Why this is so I have attempted to 
point out. Presence of the testicle is neces- 
sary for full horn development, in some breeds 
it is necessary for any horn development. 
Reasoning from the experimental results ob- 
tained in poultry it seems probable that injec- 
tions of testicle extract into the female would 
cause increased horn development similar to 
that of the male. It would be interesting to 
know whether the testicle of all breeds would 
behave alike in this experiment. Whether the 
female sex gland acts as an inhibitor of horn 
development would be a wholly different ques- 
tion, yet one capable also of experimental so- 
lution, 
To assume, as Arkell and Davenport do, 
that inhibiting factors present in X-chromo- 
somes affect the horn development seems to 
me unwarranted, for the simple reason that 
neither inhibitors nor X-chromosomes are 
known to exist in sheep. That Guyer has 
recognized the existence of an X-chromosome 
in man has no very direct bearing on the 
SCIENCE 
575 
question, but even Guyer’s result is uncon- 
firmed by Gutherz, who has reinvestigated 
the spermatogenesis of man upon exception- 
ally favorable material. 
Arkell and Davenport reason thus: 
The results of the table [of crosses] accord very 
closely with expectation, so that we are justified in 
concluding that an explanation of the results like 
that we offer is the correct one. 
But Bateson’s explanation accords also; 
wherein lies the superiority of the new one 
offered? To establish the probable correctness 
of a hypothesis it must be shown that no 
other hypothesis accords with observed facts 
equally well. Has this been shown in the case 
before us? 
Consider how one unproved hypothesis has 
been added to another. First it is assumed 
that in hornless animals a gene for horns has 
either been lost or is inhibited. It is equally 
probable that no gene has been lost and that 
nothing is inhibited. Secondly, it is assumed 
that one inhibitor is inferior to one horn- 
gene in power, but that two inhibitors surpass 
one horn-gene, yet two inhibitors are them- 
selves overpowered by two horn-genes; with- 
out all three of these ungrounded assumptions 
of the relative valency of imaginary genes the 
explanation fails altogether. Further, it is 
assumed that the female is capable of carry- 
ing two inhibitors, but the male only one. 
And finally when this colossal structure of 
hypothesis encounters one well-known physio- 
logical fact, the result of castration, that fact 
is calmly brushed aside. Is this a desirable 
extension of Mendelian interpretation? 
W. E. Caste 
March 13, 1912 
THE MOTH OF THE COTTON WORM (ALABAMA 
ARGILLACEA HUBN.) 
To tHE Epitor or Science: In connection 
with the notices appearing in Science (Oc- 
tober 16, November 10 and December 29), 
concerning the moth of the cotton worm and 
the destructive work of the caterpillar on cot- 
ton, a note from Missouri may be of interest. 
During the fall this moth was present in 
great numbers in various parts of the state. 
