I04 



THE ALUMNI JOURNAL. 



universally adopted, is that devised by 

 Paul and Cownley, which consists in 

 treating the powdered leaves with MgO, 

 and exhausting the dried mixture with 

 alcohol. Two years ago, Allen, in an 

 exhaustive paper on caffeine (Pharma- 

 ceutical Journal lyll. , p. 213), hinted that 

 this process was open to suspicion, the 

 complete exhaustion with alcohol being 

 in some cases impossible. Having occa- 

 sion recently to determine the caffeine 

 content of a large number of samples of 

 tea, I have made a thorough study of the 

 above and other processes, and the results 

 are embodied in the present paper. 



At the outset I would say that a large 

 portion of my work has been merely con- 

 firmatory of the work published in Allen's 

 paper previously referred to, a paper 

 which I venture to say has not in this 

 connection received the attention it mer- 

 ited. This work may be briefly summa- 

 rised as follows ; — 



1 . Aqueous solutions of caffeine can be 

 concentrated by boiling, and subsequent 

 ly evaporated to dryness in a water oven 

 without any loss of alkaloid. 



2. Boiling caffeine with lime causes 

 decomposition, the loss varying under 

 different conditions from 20 to 50 per 

 cent. Boiling with magnesia causes 

 practically no change. 



3. Admixtures of powdered tea leaves 

 with lime or magnesia cannot be com- 

 pletely exhausted by percolating with 

 solvent, such as chloroform, ether, etc. 

 This fact was first pointed out by Paul. 



4. Determinations of caffeine, based 

 on treatment of the leaves with lime and 

 subsequent boiling with water are value- 

 less owing to decomposition of the alka- 

 loid. 



As the above facts have been previous- 

 ly pointed out, a detailed account of con- 

 firmatory experiments is not necessary. 

 Finally, as the result of my own observa- 

 tions, based on the examination of a very 



large number of samples of tea, I have 

 come to the conclusion that the process 

 ot determining caffeine in tea by treating 

 with magnesia and exhausting with alco- 

 hol, is open to considerable error and 

 should be abandoned. 



The difiiculty met with in extracting 

 caffeine from the leaves is probably due 

 to its association in the leaf with other 

 bodies, possibly, as suggested by Allen, 

 in the form of a glucoside similar to that 

 occuring in kola nuts ; or possibly owing 

 to combination with some tannoid body. 

 Experiments with a view of elucidating 

 this point have met hitherto with but 

 little success. 



Zoller (Zeitschr. Anal. Chem. XII., p. 

 106) supposes that the cellular structure 

 of the tea leaf causes the obstinate reten- 

 tion of part of the caffeine, and has sug- 

 gested a process based on the destruction 

 of the cellulose with sulphuric acid and 

 subsequent exhaustion with alcohol after 

 neutralisation with lead oxide. This 

 process gives accurate results, but owing 

 to the retention of the caffeine by the 

 charred tissue, prolonged treatment with 

 large quantities of the solvent is necessary. 

 This process is too tedious for every- day 

 work. 



Herlant's process, based on the solu- 

 bility of caffeine in sodium benzoate is 

 open to objection that complete extract- 

 ion cannot be obtained under two weeks* 

 time. 



Squibb's process, consisting in boiling 

 the tea leaves for a few minutes with 

 magnesia, is one of the most satisfactory 

 yet devised, the only objection being that 

 the boiling is not continued long enough. 

 Several hours boiling is necessary to en- 

 sure complete extraction. 



The method finally adopted as the most 

 satisfactory, and as yielding the best and 

 most concordant results, was that sug- 

 gested by Allen. The process is as fol- 

 lows : Six grms, of finely powdered tea 



