124 F - B - SUMNER, M. E. MCDANIEL AND R. R. HUESTIS. 



within the next few weeks ; in others it was deferred until the 

 time of marking, when the broods were broken up, those of dif- 

 ferent sex being segregated, and each individual being registered 

 and given its proper serial number and identification mark. 

 There has been no uniformity of practice in regard to the age at 

 which the mice have been thus separated, marked and registered. 

 \Yhenever practicable, this has been done within six or eight 

 weeks after birth, but three months or more have sometimes 

 elapsed before it has been convenient to do so. When, as is com- 

 monly true, no deaths have occurred during the interval which 

 has elapsed since the first count, no harm can have resulted from 

 thus deferring the record of sex. 



A few words are worth while at this point in reference to the 

 possible bearing which such imperfections in our records might 

 be supposed to have upon the results set forth in this paper. As 

 already stated, we do not believe that they seriously affect their 

 value. In the first place, such errors as actually exist are ones 

 of omission. The number recorded for certain broods is doubtless 

 too small, owing to the occasional presence of stillborn young, or 

 of ones which died within the few days following birth. On the 

 other hand, we are confident that mistakes in the identification of 

 the sexes have been so infrequent as to be negligible. 



Regarding this matter of incompleteness of the entries, several 

 things must be said. Firstly, it is doubtful whether any records 

 are possible which are perfectly complete in this respect. It is a 

 practical impossibility to inspect every brood immediately after 

 birth, and in the interval which elapses the mother may eat such 

 stillborn or feeble offspring as are present. 



In the second place, the number of dead or defective members 

 of a litter is small in Pcromyscus, as compared, for example, with 

 ordinary white mice. Stillborn young may be very common 

 among the latter. In our experience they have certainly been 

 rare among the former. As regards very early postnatal mortal- 

 it}-, this is, from the nature of the case, difficult even to estimate 

 from the data at our disposal. Reference to the death rate 

 somewhat later in life may, however, be instructive. Of the 1,567 

 broods, comprising 5,050 young, which are listed in our records, 



