BREEDING POULTRY FOR EGG PRODUCTION. I39 



progeny in 1908 and hence do not appear in the discussion of 

 the fecundity of narrow and broad bred females. On this 

 account they are dropped out of Table D. 



From this table a number of points are to be noted. 



1. The number of females in this experiment which suc- 

 ceeded in producing adult daughters is relatively small with 

 reference to the total number of females bred. That is to say, 

 there is a rather stringent genetic selection here. Examination 

 of the table shows that this is due to the unfavorable environ- 

 mental conditions which existed in house No. i. This was 

 obvious from direct observation during the course of the ex- 

 periment. The way in which it is shown in the table is that 

 whereas but 21.7 per cent and 18.3 percent respectively of the 

 females mated with foreign and Station cockerels in No. i 

 house got adult female progeny, on the other hand 45.5 percent 

 of the females mated i-n house No. 2 got adult daughters. This 

 record of 45.5 per cent for the females in house No. 2 is itself 

 lower than it would have been had not one of the males used 

 in No. 2 house (No. D26) proved to be a rather unsatisfactory 

 bird, which did not make a very good record either in respect to 

 fertility or hatching quality of eggs. 



2. The rather stringent genetic selection brought about by 

 the unfavorable conditions in house No. i does not interfere 

 with the experiment in regard to the effect of inbreeding be- 

 cause the force of this selection was substantially the same in 

 the case of females mated in No. i house with foreign cockerels 

 and females mated 'in the same house with Station cockerels. 

 The difference is only 3.4 percent and cannot be regarded as 

 significant. 



3. With respect to the mean winter egg production of the 

 mated females it is seen that the averages for the females which 

 were originally mated with foreign cockerels, and that for the 

 females in house No, i originally mated with Station cockerels 

 are substantially the same. What advantage there is is in favor 

 of the foreign cockerels, that is, of the broad-breeding side of 

 the experiment. The probable errors of these averages of egg 

 production lie in the neighborhood of 1.2. It is obvious that 

 the difference (1.30) between the mean production of the two 

 sets of females in No. i house is only such as might be expected 

 to arise from random sampling. The females mated with Sta- 



