EFFECTS OF THYROID ON PARAMAECIUM 531 



METHODS 



The experiments here reported have been performed on three 

 races of Paramaecia whose hfe histories are given in full below. 

 These races were in each case derived from a single individual and 

 thereafter maintained in four files hj daily isolations following the 

 method described by Calkins ('02). Each individual was placed 

 in a solid shallow watch gl'ass in four drops of the culture medium 

 and at the end of twenty-four hours (or rarely a longer period) 

 the number of Paramaecia present was obtained and the num- 

 ber of divisions calculated. An individual selected at random 

 was then isolated in the same manner, using a pipette reserved 

 for the line in question. Records were also kept of the number 

 of deaths and monstrosities occurring. At the end of every five 

 days the total number of divisions in the four files of each line 

 was obtained and averaged to give the average division rate per 

 day of the line for that period. These five day averages form 

 the units for our discussion. 



The culture media employed for the control lines were in all 

 cases hay infusions. The formulae for these will be cited under 

 the life histories of the three races. The thyroid and other 

 glandular media were prepared in different experiments as emul- 

 sions of fresh glands from rabbits, rats, and cattle, and as sus- 

 pensions of commercial gland preparations. The following 

 preparations of Armour and Company were employed: desic- 

 cated thyroids U. S. P. (sheep), thymus desiccated (calf), pitui- 

 tary body desiccated (ox), desiccated suprarenal capsules (?), 

 pancreas desiccated (pig), spleen desiccated (calf), ovarian sub- 

 stance desiccated (pig). Experiments were also conducted with 

 iodothyrin (combined with sugar of milk), a commercial prepa- 

 ration of Friedr. Bayer and Company. Formulae used are 

 cited under each experiment. 



Experimental lines were in all cases derived by isolating one 

 individual from each file of the control line, thus obtaining two 

 lines of maximum similarity. Where several lines were required 

 as in some experiments, the four files of the control were allowed 

 to multiply until the required number was present in each, and 



