68 



MISCELLANEOUS PAPERS. 



mentecl, but some of them were iinfermented. All results are cal- 

 culated on the water-free samples. 



75-154,242. 

 75-154,218. 



13-2-3 



3-20-2 



125b-l-t . . 

 123a-l-12 . 

 124b-l-l . . 

 123a-l-l 1 . 

 125a-l-2 . . 

 123a-l-l . . 

 125b-l-l . . 

 123a-l-8 . . 

 123a-l-3 . . 

 123a-l-2 . . 

 123a-l-9 . . 

 125a-l-4 .. 

 3-20-.5-12. . 

 124b-l-6 . . 

 3-20-5-6... 

 109-3-4.... 



yield number of tobacco samples. 



Differ- 

 ence. 



Per 



cent. 

 -0.01 



- .09 



- .09 



- .09 



- .01 



- .10 



- .08 



- .07 



- .16 



- .08 

 + .03 



- .09 



- .09 

 + .04 

 + .13 



- .07 

 + .01 



- .09 

 + .08 

 + .05 



Average difference, 0.04 per cent. 



CONCLUSION. 



It will be observed that the results shown in the above table indicate 

 that the new method gives figures averaging a few hundredths per 

 cent lower than the Kissling method. This difference is due to the 

 water added to the sample along with the caustic soda, which is not 

 taken into account in the calculations. From 3 to 5 c. c. of water are 

 thus added, depending upon the character of the tobacco. As is well 

 known, tobacco is markedly hygroscopic, and a considerable portion of 

 the added water is doubtless held in loose chemical combination with 

 the salts contained in the tobacco, and hence does not act as a solvent 

 for the nicotine. The portion of the water not thus combined, how- 

 ever, really forms a part of the nicotine extract. It was found by 

 experiment that the relative solubility of nicotine in ligroin and in 

 water is in the ratio of 11 to 9. The mutual solubilities of ligroin 

 and water are so slight as to be negligible, and there is no appreciable 

 contraction in volume when the two are mixed. Taking all the facts 

 into consideration, it is clear that the actual error resulting from the 

 addition of the water to the sample to be analyzed is too small and 

 variable to attempt any correction therefor. 



It should be observed in this connection that those types which con- 

 tain the highest percentage of nicotine belong to the class of heavy 

 filler tobaccos, and hence require the addition of only the smallest 

 quantity of water. This, of course, tends to equalize this source of 

 error. The analyses by the Kissling method were not made in dupli- 

 cate,- and differences in the results obtained by the two methods 

 amounting to one-tenth of 1 per cent or more are doubtless due to 

 slight experimental errors. 



102 — VII 



