SUPERFETATION AMONG MICE. 273 



My stock comprises four geographical races of Peromyscus 

 maniculatus , viz., rubidus, sonoriensis, and two quite distinguish- 

 able races within the assemblage generally known as " gambeli." 



In breeding these mice, it is my general practice to place one 

 male with three to five females. Each male is left with the 

 respective group of females for a varying period, which ordinarily 

 does not exceed 20 days. Record is in every case kept of the 

 day when the male is admitted, and of the day when he is 

 removed. Each female, when obviously pregnant, is isolated in 

 a separate cage. The young are ordinarily left with the mother 

 until they are 6 to 8 weeks old. At this time, they are commonly 

 separated according to sex, the males and females being thence- 

 forth kept apart until they are mated, according to the require- 

 ments of the work. Occasionally, this separating is not done 

 early enough, however, and matings between brother and sister, 

 or between mother and son, occur. The youngest mice in which 

 I have known fertilization to take place were both 42 days old 

 (possibly a few days younger), the next youngest pair were 55 

 days old or less. Because of these precocious, and therefore, 

 unexpected conceptions, the exact date of birth of the resulting 

 young cannot always be recorded. They were sometimes several 

 days old when first found. Now it is among very young parents 

 that the phenomena to be discussed seem to be commonest, 

 as will be pointed out presently. For this reason, it happens 

 that in all but two among the eight cases comprised in my first 

 table the father was still present in the cage, and had access to 

 the mother up to the time of the isolation of the latter, when her 

 brood was first discovered. This uncertainty as to the date of 

 the last coition somewhat complicates the interpretation of 

 these cases, though, as will be shown, the evidence, even here, 

 for a deferred fertilization of the ova is not thereby affected. 



Before considering the phenomena presented by the tables, it 

 will be well to make inquiry as to the normal period of gestation 

 in Peromyscus. Since I have been interested only incidentally 

 in determining this period, I have in no case restricted the mating 

 of the parents to a single observed copulation, as has Daniel 

 (1910), nor have I determined the exact hour of the birth of the 

 brood, as have Long .and Mark (1911). I have recorded, how- 



