8 



Another thirty liter mass was prepared as in the 

 above experiment. The bacteria were washed and separated 

 by the same method, but instead of being saponified, they 

 were treated directly with ether. This ether-bacteria 

 mixture was allowed to stand, with frequent shaking, for 

 six days. The ether was then pipetted off. The residue 

 was again treated with ether in the same manner. These 

 two ether extracts were then evaporated. The residue 

 consisted of brownish-colored fats, some of which ap- 

 peared crystalline, other amorphous. This material was 

 washed with water, then taken up with ether, the ether 

 removed and evaporated. The weight of the resultant fats 

 amounted to approximately 0.4 gram. The fats were then 

 saponified with potassium alcoholate at 100C. After 

 acidifying with hydrochloric acid, and cooling, a defin- 

 ite layer of fatty acids was obtained. The alcohol was 

 evaporated to nearly dryness and the fatty acids extract- 

 ed with ether. This ethereal extract was concentrated to 

 dryness and the fatty acids taken up with alcohol. This 

 was treated with sodium carbonate to convert the fatty 

 acids into the sodium salts. The sodium salts were then 

 drystallized out and redissolved in 100 cc. of alcohol. 

 This constitutes the fatty antigen. 



It is desirable, at this point, to examine the 

 antigen to see if any protein material is present. 

 Several different tests were used as follows: 



a. The biuret test. To three cubic centimers of 

 antigen was added an equal amount of strong potassium 

 hydroxide solution. The mixture was well shaken and 

 treated with a few drops of very dilute copper sulphate 

 solution. Absence of color change indicated absence of 

 protein. 



b. The Kjeldahl method, 10 cc. of the antigen were 

 mixed with 20 cc. of concentrated sulphuric acid. 0.2 

 gram of copper sulphate was added and the material was 

 boiled gently for ninety minutes. It was then cooled and 

 diluted to 250 cc. with distilled water, then neutralized 

 with hydroxide, adding a slight excess of the alkali. 

 The material was then distilled, the distillate being 

 collected in 25 cc. of N/10 sulphuric acid. When one- 

 half the liquid had passed over, the process was stopped. 

 The acid solution was then neutralized with tenth-normal 

 sodium hydroxide, using congo red as an indicator. An 

 equal amount of alkali was necessary to neutralise the 

 acid, thus indicating the absence of nitrogen in the 

 antigen. 



