RESPIRATORY DIFFERENCES. 383 



sary to know the exact volume of A plus B and C plus D. The 

 oxygen content at the beginning is determined in the same way 

 by calculation from the water blanks and is brought by ratio 

 to the same volume as the experimental tube. A simple sub- 

 traction then gives the oxygen consumed by the animals in terms 

 of thiosulphatc. To find the oxygen equivalent of the thio- 

 sulphate the latter must of course be standardized. A number 

 of methods are -iven in any text on quantitative analysis. If 

 the M consumption in terms of volume is desired, as is 



u-iially t! . then it is desirable to conduct all of the experi- 



ment- at the ~.une temperature and to determine the volume of 

 the tube- 1'ir this temperature. The oxygen equixalent of the 

 thio-nlph.ite for any given temperature can be determined from 

 data on the \olume of oxygen at different temperatures yixen in 

 hamll ! physical constants. 



The app.ii.iiii- can be modified in various ways to suit the 

 type of animal employed. For Protozoa the tube is closed, alter 

 filling \\\\\\ water and adding the animals, not by means of a 

 lamp but by means of a glass plug. This is inserted deeply 

 into J) and must be firmly wired in. At the end of the experi- 

 ment tin tube is placed in the centrifuge in the inverted po-ition. 

 the bottom of .1 jointing towards the axis of the centrifi 

 I' MI itir^inv:, the Protozoa are driven into the section C phi- 

 /'. \\hiih i- then removed as already <Ie-<-ribed. For l;u 

 animals the -i/e of the apparatus can be increased or in place of 

 the te-t tube .1 a small flask of desired content can be substituted. 

 In -in h cases, the amount of the reagents used and the dilution 

 of the thio-nlphate should be adjusted to the size of the sample 

 obtained for analysis. The apparatus is not suitable for animals 

 whieh fillip iirmly, such as planarians. 



The method is naturally not as accurate as the re-ular \Yinkler 

 method. By the analysis of duplicate samples I have found 

 that the oxygen <<>ntent of 15 cc. of water about 0.07 to 0.08 

 cc. at air -aturation and room temperature can be determined 

 \\ith an error of about O.OO2 to O.OO3 CC. Thi- make- all error 

 of some 3 to 4 per cent, while in the regular YVinkler proce- 

 dure, \\ith -amples of at least 100 cc., the error i- K -- than 0.5 

 percent. The method is thus chiefly of value for comparative 

 \\ork. 



