COCOA. 



5 



Mr. Charles Heiscli, F.C.S., in a paper " On the Composition of 

 Various Kinds of Cocoa," * observes : " It is well known that different 

 varieties of cocoa fetch very different prices ; but as far as I am 

 aware, no careful examination has been made to ascertain if these 

 variations are caused by any difference in their composition regarded 

 as articles of food, or if they be due solely to differences in flavour, 

 which, after all, may be only matters of taste. In none of the pub- 

 lished analyses of cocoa which I have seen is any mention made of the 

 kind of bean analysed, it is therefore not surprising that the results 

 published vary very considerably. Thus, while in Dr. Hassall's book 

 we are told that cocoa contains albuminoid matter 16*7 per cent. ; in 

 Dr. Parkes' Practical Hygiene it is stated to contain from 13 to 18 

 per cent, of protein substance. In neither case is it mentioned 

 whether the bean was examined raw or after roasting. Having 

 through the kindness of a friend obtained samples of various cocoa 

 beans, both raw and roasted, which he assured me were unmixed, I 

 made a number of analyses of the roasted beans. The results are 

 shown in the following table. They are not so complete as I had 





Per- 

 centage 

 of Husk. 



Result op Examination of Roasted Bean after Removal of Husk. 



Fat. 



Nitro- 

 gen. 



Albumi 

 noid 

 sub- 

 stances. 



Ash. 



Ash 

 soluble 



in 

 water. 



Ash 

 soluble 

 in HCl. 



Phosphoric 

 Acid in 

 Ash, cal- 

 culated as 

 H3 PO^. 



Mois- 

 ture. 



Starch. 

 Gum, 

 CelIu-5 

 lose, &c. 



Caracas 



13-8 



48-4 



1-76 



11-14 



3-95 



2-15 



1-80 



1-54 



4-32 



32-19 



Trinidadt .. 



15-5 



49-4 



1-76 



11-14 



2-80 



-9 



1-90 



-93 



3-84 



32-82 



Surinam , . 



15-5 



54-4 



1-76 



11-14 



2-35 



-80 



1-55 



1-23 



3-76 



28-35 



Guayaquil . . 



11-5 



49-8 



2-06 



13-03 



3-50 



1-75 



1-75 



1-87 



4-14 



30-47 



Grenada , . 



14-6 



45-6 



1-96 



12-40 



2-40 



•60 



1-80 



1-35 



3-90 



35-70 



Bahia 



9-6 



50-3 



1-17 



7-40 



2-60 



•90 



1-70 



1-26 



4-40 



35-30 



Cuba .. 



12-0 



45-3 



1-37 



8-67 



2-90 



-95 



1-95 



1-13 



3-72 



39-41 



Para .. 



8-5 



54-0 



2-00 



12-66 



3-05 



1^40 



1-65 



1-00 



3-96 



26-33 



f 1 am inclined to think that the Trinidad sample was not of the finest quality. 



hoped to make them, but they comprise the more important con- 

 stituents. In the first column of the table is noted the proportion of 

 husk in the several varieties. This difference appears to be mainly 

 due to the husk in some kinds being much thicker than in others ; in 

 all cases these thick husks separate more readily from the bean in the 

 process of roasting, and can be taken off with much greater facility. 

 The other estimations are made on the roasted bean after removal of 

 the husk. The albuminoids are calculated from the total nitrogen 

 found by combustion with soda lime ; the nitrogen contained in the 

 theobromine is thus included, but in the roasted bean this is so small 

 that the difference is hardly worth consideration ; hereafter I hope to 

 estimate the theobromine in the different varieties, as well as the 

 starch, gum, cellulose, &c. It will be observed that in none of the 

 above samples do the albuminoid substances reach the amount men- 

 tioned by Hassall or Parkes, but as neither of them give the method 

 by which the albuminoids were ascertained, no attempt can be made 

 to account for the difference. The amount of these substances in 

 * The 'Analyst' for Oct. 31, 1876. 



