ANALYSIS OP COFFEE. 25 



(4.) The solution is allowed to remain ou the saud-bath, with the condeuser in place, 

 for exactly fifteen minutes after the removal of the flame. 



(5.) The whole of the liquid and more or less of the solid matter in the flask is now 

 thrown upon a dry filter (heavy, white German paper) of 13 cm. diameter, and about 75 

 c.c. will be found to pass through in from ten to fifteen minutes. This filtrate is used 

 for the determination of (1) the specific gravity of the extract, (2) caffeine, (3) sugar, (4) 

 tannin. I always make the extract in duplicate, and thus have enough material for all 

 these estimations. 



The degree of constancy to be attained by this method of making extracts, may be 

 judged from the following synopsis of results given in detail in Tables III and V. 



Of the fourteen duplicates given in Table III, the maximum difference for any pair is 

 000024, and the average diff'ereuce is 000009. Of the thirteen duplicate extracts in Table 

 V, Nos. 1 and 10 are altogether abnormal, difieriug to the extent of 000102 and O'OOIOI. 

 The highest difference for the remainder is 0"00040, and the mean 0.00012. 



Even if we include the abnormal diff"erences referred to, the average difference for 

 the twenty-seven samples is only 00001*7,'_and by omitting these from consideration, the 

 average difference is 000011. 



I find, upon referring to my laboratory note-book, that the exceptional extracts 

 referred to were made at a time when my mode of operating was not so well defined as I 

 have described ; and, strictly, the analyses in question should not have appeared in the 

 table, and would not have done so had the illustration of this mode of making extracts 

 been the only reason for its construction. 



As will be shewn in the sequel, a difference of 10 per cent, in the amount of chicory 

 in a mixture of coffee and chicory, corresponds to a difference of 0.00184 in the density of 

 the extract ; so that the constancy of result obtainable by this method is such as to enable 

 us to determine well within one per. cent, of the amount of chicory in a mixture of 

 coffee and chicory. 



I may explain here that the specific gravity of the extract has been taken strictly at 

 62" Fah., and the comparison is with water at the same temperature, as unity. The bottle 

 used in this work was, in the first place, most accurately adjusted by myself, and in order 

 further to check its accuracy, I have occasionally confirmed the result by weighing a 

 loaded glass bulb of known specific gravity in the liquid. After each weighing, the 

 bottle has been emptied, the extract again brought to the proper temperature, the bottle 

 refilled and reweighed until consecutive weighings did not differ by more than two 

 milligrammes. 



In his " Manual of Food Analysis," A. W. Blyth gives a table for the estimation of 

 the percentage of chicory present in a mixture of coffee and chicory from the specific 

 gravity of the ten per cent, extract. It does not follow, however, that this table is applic- 

 able to extracts made in the way I have described, and as a matter of observation I have 

 learned that it leads to very incorrect interpretations when applied to extracts so made. 

 (See Table IV.) 



From the accompanying table (see Table I) — the results of work upon sixteen samples 

 of pure coffee — it will be seen that the mean specific gravity of a 10 per cent, extract of 

 coffee, made in the way described, is 1'00986 : — 



Sec. iii, 1887. 4. 



