16 BULLETIN 1413, IT. S. DEPARTMENT OF AGRICULTURE 
The four samples of press cake numbered 105, 109, 110, and 111 
(Tables 2 and 3) represent three distinct commercial types of pressed 
cocoa residues. The last three were produced by one of the largest 
manufacturers of cacao butter, and are examples of dry-pressed 
cakes. Sample No. 105 is typical of ordinary by-product cocoa of 
average fat content, and represents a lot of cake that had been 
purchased for the recovery of its fat by solvent extraction. Num- 
ber 109 is a composite of dry-pressed cakes resulting from the press- 
ing of accumulated fines containing cacao germs. The high con- 
tent of ash and of crude fiber mark this sample as including some 
shell material and probably cocoa dust and sweepings. Inclusion 
of the germs tended to increase the nitrogen content of this sample 
of low-grade material. Cacao germs separated from roasted beans 
and analyzed by Richards (57), had a nitrogen content of 5.5 per 
cent, equivalent to 6.7 per cent of ammonia. Samples 110 and 111 
are excellent examples of the dry-pressed, straight nibs, by-product 
cocoa. No. 110 is stated to have been made from Sanchez (San 
Domingo) beans, whereas No. Ill was made from " almost straight 
run" Accra (Gold Coast, West Africa) beans. Both of these 
varieties of cacao are classed by van Hall (19) as among the 
lowest priced and most ordinary sorts, considered from the point 
of view of the flavor and quality of the chocolate produced from 
them. 
Accra beans particularly, however, are well suited to the needs 
of the manufacturer primarily after a large yield of cacao butter, 
as both the proportion of nibs and the fat content of the nibs are 
relatively high (58, p. 112). The imports of beans from British 
West Africa, largely Accra, amounted to more than one-third of all 
cacao brought into the United States in 1923 (53, Table 3). 
Samples 113 and 114 were submitted by a manufacturer as rep- 
resenting, respectively, the nibs cake and the whole bean cake pro- 
duced in the factory test pressing, mentioned on page 9. This 
experiment, which was designed to compare the yield of cacao butter 
from ordinary nibs with that obtained by pressing all parts of the 
bean together, showed an increased production amounting to 4 pounds 
of butter (per 100 pounds of roasted nibs fraction), when the 
ground nibs, shells, and germs were pressed together. Both the 
separated nibs, from which sample 113 was produced, and the 
manipulated whole beans, of which sample 114 represents the press 
cake, were from the same lot of " fair fermented " and roasted 
Accra cacao. 
Unfortunately, however, these samples can not be considered rep- 
resentative as to fat content of the respective cakes. No. 113 con- 
sisted of two chunks from the outside edge of the nibs press cake, 
and No. 114 was a single piece from the interior of the whole-bean 
cake. As the periphery of any hydraulically pressed oil cake is 
richer in oil or fat than the interior portions of the cake, and as 
cocoa press cakes are disk-shaped, only a sample consisting of a 
perfect sector could have been representative as to fat content of 
either cocoa cake. 20 Nevertheless, the decidedly higher percentage 
20 Sectors taken from top, middle, and bottom cakes in the press would be required to 
give a representative sample of an entire pressing'. 
