406 
Miscellanea 
Also the combination three birds without crest out of three arises. The chance that this will 
arise is still more remote — 1 in 64. 
Now Davenport rejects the combination 9 : 0 which occurs once in 13 times while he finds 
among his material combinations whose chance of occurring is far more remote, and yet retains 
them ! 
Homozygous and heterozygous birds in such cases can only be distinguished by increasing 
the number of expei-iments, but in cases of doubt crested birds should be mated with non- 
crested birds as then the chance that all of nine for example shall be crested is only 
1 in 2^ = 1 in 512. To get the same probability when mating two crested birds, we must 
obtain 22 ofl'spring. Davenport does not seem to recognise that whenever all the offspring are 
crested, whether there be only two or as many as nine, there is a possibility of one or both 
parents being homozygous. 
Davenport's failui-e to test adequately the gametic constitution of the parents renders any 
appeal to his totals useless and a similar criticism applies to part 2 of the same table. All we 
can say is that in Table II, part 1, there are some matings which are DRxDR and others 
which may be DDxDD, DD x DR, or DRxDR, and that conclusions based on methods such as 
these are of no value. 
His treatment of the inheritance of baldness however is even more faulty, and in Table III, 
one is confronted at the outset with a series of blunders and mis-statements which are absolutely 
unparalleled. In Table III, there are given the results of 26 matings, and in the following 
cases the statements made are contradicted either in the same table or elsewhere in the paper. 
In Table A, Exp. 617, Bird No. 84 should be No. 82 ; and in Exp. G20, Bird No. 83 has a 
" perfect crest," not " imperfect " as stated. Exp. 620 should therefore be included in Table C 
and not in Table A. In Table B, a, Exp. 505, Bird No. 34 is said to have a perfect crest ; 
in Exp. 513, C, /3, it is said to have an imperfect crest, while in Table E, Exp. 403, it is said to 
have a perfect crest. In Table B, a, Exp. 505, four of the offspring are said to be non-crested ; 
only two are given in Table E. In the same experiment (No. 505) Bird No. 7 is said to be with- 
out crest ; in Table C, /3, it is said to have a perfect crest, while in Tables I. and E it is said to 
be without crest. In Table B, a, Exp. 625 should be No. 624. 
In Table B, /3, among the offspring of Exp. 705, there are only three birds with perfect crest 
and not five as stated. In Exp. 714, in the same table, there is said to be one bird with im- 
perfect crest and six in all. In Table E, for this experiment there are four birds with perfect 
crest, one with crest absent, and two which died young. No bird is marked as having an 
imperfect crest. According to his own figures instead of totals of 11 perfect crests and four 
imperfect crests there are nine perfect crests and three imperfect crests together with seven 
birds which are without crests ; such birds are in the same table considered as perfect crested 
birds, i.e. they are not bald. There are thus 16 birds with perfect crests and three with 
imperfect crests. 
Further, no imperfect crested birds are known to have occurred in the ancestry of Birds 
Nos. 157 and 240, and to state as he does that such is "probably" the case is to beg the whole 
question. 
In Table C, a, Exp. 508, there should be one bird without crest. In Exp. 604, Bird No. 62 
has a "perfect crest," not an "imperfect crest" as stated. This mating is therefore wrongly 
included in the Table. In Exps. 608 and 703, Bird No. 67 should have a " perfect crest," not 
an "imperfect crest "as stated. These matings should therefore be included in Table B and 
not in Table C. In Exp. 703, there are only two birds with perfect crest, not three as stated. 
In Exp. 704 there are only three birds with perfect crest, not five as stated. In Exp. 711 
there are two birds with perfect crest and none with imperfect crest, not one of each as stated- 
