MATERIAL AND METHODS. 7 



For distinguishing individuals the system of holes and notches 

 punched in the ears, used by Professor Castle and Dr. Allen, was em- 

 ployed (Allen, 1904). In addition to a book for serial numbers, sex, 

 parentage, color, and date of birth arranged according to the serial 

 numbers, it was found convenient to have another book in which there 

 was devoted to each cage a separate sheet, whereon were set down the 

 serial number, sex, and color of each of the mice in the corresponding 

 cage. When mice were transferred from one cage to another, corre- 

 sponding records were made on each sheet, making it possible, when 

 necessary, to trace a mouse from one cage to another, and to determine 

 its matings. 



Individual records were kept of the breeding females only. These 

 records, the record of litters, etc., were made on paper of uniform size 

 perforated for file-binding. To lessen the possibility of error, the same 

 sheets also served for all subsequent records of insemination, killing, 

 fixing, etc. Finally, a new serial number, corresponding with the numbei 

 on the slides prepared from the killed individuals, was also recorded on 

 these sheets. 



Sobotta (1895) states that under natural conditions mice breed 

 most actively during two periods in the year, one in the spring (April and 

 May), the other in late summer and early autumn (from the middle of 

 August to the end of^September) ; but that if kept warm they breed all 

 winter. Since the mice used in this work have been kept warm and 

 well fed at all times of the year, the conditions have not been favorable 

 for determining the natural breeding seasons. 



As previous investigators have shown, female mice are in heat and 

 ovulate soon after parturition. The eggs for the present work have 

 been obtained, with one exception, from the ovaries and oviducts during 

 the first 40 hours after parturition. 



It has been the custom of the junior writer to look over the stock 

 of breeding mice every 5 days (5 days being the average time before 

 parturition when pregnancy is first easily recognizable) to note pregnant 

 females and to remove from males such as were to be observed and killed. 

 Apparently it has been the habit of former investigators to leave the 

 two sexes together and not to determine with exactness either the time 

 of parturition or of fertilization. It was felt from the first that a fair 

 degree of accuracy in the determination of the times of parturition and 

 insemination would be of great advantage; and, since it was found that 

 parturition may occur at any time during the 24 hours of a day, it was 

 necessary to make frequent observations. 



In order to increase accuracy in observation and to save much time 

 during both day and night, the apparatus illustrated in figs. B, C 

 (plate A), and D was planned and made by the junior writer. Its pur- 

 pose was to serve in recording automatically the approximate time of the 

 birth of litters. In this apparatus advantage has been taken of the 



