310 Mary Davies Swartz, 



still warm into about three times its volume of acetone, which expe" 

 rience showed to be a more satisfactory precipitant than alcohol* 

 Most of the carbohydrate came down very soon, in large, fiocculent, 

 yellowish-white masses, but a portion remained in suspension as a 

 fine white powder, which made filtration difficult. The bulk of the 

 precipitate was therefore removed by filtering through three or four 

 thicknesses of fine gauze, and the rest obtained by distilHng off the ace- 

 tone, concentrating the residue, and reprecipitating the carbohydrate 

 in solution with acetone. This precipitate was very hydroscopic, and 

 was therefore transferred immediately to 95 per cent alcohol. This 

 was replaced by fresh alcohol after a few hours, and the whole boiled 

 on a reflex condenser for half an hour. A yellowish, granular powder 

 was thus obtained, which was filtered, washed with ether, and the ad- 

 herent ether allowed to evaporate. It was then redissolved in a small 

 volume of water, filtered hot through paper, on a jacketed funnel, 

 reprecipitated with acetone, again put into 95 per cent alcohol, and 

 finally into absolute alcohol, in which it was allowed to stand for 

 several weeks. It was then filtered off, washed with ether, and dried in 

 vacuo over sulphuric acid. The product was a cream-white powder, 

 and apparently not at all hydroscopic. From about two kilograms of 

 crude commercial dulse, approximately 75 grams of this material were 

 obtained, and used subsequently for feeding experiments. 



An attempt made to remove the dark red coloring matter by extrac- 

 tion with 1 per cent sodium carbonate, led to the discovery that this 

 carbohydrate is readily extracted by dilute alkaline solutions. For 

 preparations on a large scale, it was therefore found more satisfactory 

 to use the following method, based on Salkowski's method (139, 140) 

 of obtaining xylan and araban by precipitation with Fehling's solu- 

 tion. This method could be applied exactly as described, but there 

 was an evident tendency for the carbohydrate to dissolve in the Feh- 

 ling's solution. 



The dulse was accordingly extracted with 1 per cent potassium 

 hydroxide solution for 48 hours, with occasional stirring, the extract 

 removed by a hand press, and the extraction with fresh alkali repeated 

 for 24 hours.i These extracts were filtered through several thicknesses 

 of gauze, and to this filtrate a solution of copper sulphate was added 

 till the reaction was just neutral. A flocculent, bluish-green precipi- 

 tate formed. Into this solution was stirred carefully the alkaline 

 Rochelle salt-potassium hydroxide solution used for Fehling's solu- 

 tion, until the precipitate clumped together in heavy granular masses. 



^A third extraction contained so little of the material that it was discarded. 



