48 Dr. Stenhouse cm, Divi-divi. 



it contains with solution of gelatine, evaporating to the consistence of 

 an extract, and then treating it with alcohol After allowing it to 

 subside, the alcoholic solution is to be drawn off, the greater portion 

 of the spirit recovered bj distillation, and the residue is to bo evapor- 

 ated to dryness on the water bath. It is then to be introduced into a 

 stoppered bottle, and repeatedly agitated with ether; almost the whole 

 of the ether is to be distilled off, and the residue left to spontaneous 

 evaporation. Abundance of reddish coloured crystals soon appear. 

 They are to be purified by being repeatedly crystallized out of alcohol 

 and water, and by digestion with animal charcoal. Lastly, by uniting 

 them to oxide of lead, and decomposing the insoluble precipitate with 

 sulphuretted hydrogen, they are rendered beautifully white. The 

 crystals then exhibit the silky lustre of gallic acid, with which acid 

 their re-actions with salts of iron and other re-agents completely corres- 

 pond. When distilled, they yield abundance of pyrogallic acid. 

 When dried at 212''F. and subjected to analysis, 



I. 0.3034 gramme acid gave 0.550 carbonic acid and 0.1018 water. 



II. 0.3052 gave 0.5505 carbonic acid and 0.1012 water. 



I. II. Calculated per cent. 



C 50.12 49.87 7 C = 49.89 



H 3.72 3.71 3 H = 3.49 



O 46.16 46.42 5 = 46.62 



100.00 100.00 100.00 



These results approach very closely the calculated numbers of hydra- 

 ted gallic acid given above. 



In order to determine the atomic weight of the acid, I formed the 

 basic gallate of lead by adding a solution of the acid obtained from 

 the Divi-divi to an excess of boiling acetate of lead. It precipitated 

 as a yellow, slightly crystalline powder, and was also dried at 212"?. 



L 0.706 of this salt gave 0.3325 lead, and 0.1802 oxide = 76.25 per 

 cent, oxide of lead. 



IL 0.887 salt gave 0.315 lead, and 0.340 oxide of lead = 76.58 per 

 cent, oxide lead. 



Nowthebibasic gallate of lead C H 0' + 2Pb O contains 76.69 per 

 cent, oxide of lead; there can therefore b(^ no doubt that it was the 

 salt analysed, and that gallic acid therefore occurs to a considerable 

 extent ready formed in Divi-divi. 



As the tannin of nut galls yields pyrogallic acid when distilled, I 

 was induced to try if the tannin of Divi-divi had the same property. 

 I therefore precipitatedja quantity of its tannin by slowly adding sul- 

 phuric acid to a saturated solution of Divi-divi. A very scanty dark 

 brown precipitate fell, it was collected on a cloth filter, strongly com- 

 pressed and washed with a little cold water to free it as much as 

 possible from adhering sulphuric acid. When dried and subjected to 

 distillation, it did not yield any trace of pyrogallic acid. It gave off 

 scarcely any empyreumatic products, and left a very bulky charcoal. 



