24 ANTHONY McGILL ON 



ÏI- — The weight of coffee used with a fixed volume of water must be calculated upon 

 the dry substance. The hygroscopic moisture in coffee is so very variable, depending, 

 among other conditions, upon the length of time since grinding, the conditions under 

 which it has been kept, and the quantity of chicory or other adu.lterant present, that the 

 comparison of one extract with another in regard to density, would otherwise be entirely 

 misleading and worthless. 



Ill- — The boiling must begin within a fixed time after the water is poured upon the 

 coffee, and be continued for a definite time after ebullition begins. Thus, in sample No. 

 1407, an extract made in duplicate, and boiled sixty minutes, gave density at 62° Fah. 

 (a.) 1-01461, {b.) 101452 — mean density, 1-01456 ; while an extract boiled seventy minutes 

 gave density 101496. In No. 887, the extracts boiled sixty minutes gave (a.) 101408 and 

 (b.) 101398— mean density being 101403 ; while an extract boiled seventy minutes gave 

 101460, and one boiled ninety minutes gave 101571. 



I'V. — The time elapsing between the removal of the flame and the subsequent filtra- 

 tion, must be constant, and the conditions as to temperature of the liquid in the meantime 

 must be constant. 



V- — The conditions of filtration must be constant. 



These conditions I have sought to secure in the following way : — 



(1.) In my early experiments I pulverised the substance until the whole passed 

 through a sieve of 400 meshes per square inch. This I found a very difficult task. I 

 could not get a mill that would grind even pure coffee to such a degree of fineness, and 

 had to separate the coarse part by a sieve, and treat in a mortar again and again. I have 

 now, for sometime, adopted the following standard, which can easily be attained by the 

 use of a drug mill, such as is made by the Enterprise Manufacturing Company of Phila- 

 delphia. I have used their No. 3 mill. At least three-fourths of the powdered coffee, etc., 

 must pass through a 400-mesh sieve, and the remainder, with the exception of a few frag- 

 ments of the papyraceous membrane of the coffee berry, through a sieve of 100 meshes to 

 the inch. 



(2.) About one to two grammes of the coffee so prepared is dried on a watch-glass in 

 the water-oven at 100° C, and repeatedly weighed, between watch-glasses, until two suc- 

 cessive weighings, made at an interval of half-an-hour, differ by less than one milli- 

 gramme. This operation is always done in duplicate, and from the mean result is 

 calculated the percentage of moisture. For purposes of extraction, I take such a weight 

 of the uiidried substance as corresponds to 10 grammes of the dry substance, calculated 

 upon the data obtained as above. Thus, if M = percentage moisture, then, 10 grammes 

 dry = jj™i grammes undried. 



(3.) The substance is introduced into a flask of about 400 c.c. capacity, and 100 c.c. 

 distilled water added. The flask is immediately placed on a small sandbath, and connected 

 with a back-flow condenser already in position. Heat is applied by means of a large spirit- 

 lamp, or a moderate-sized gas flame, so that boiling begins in from ten to fifteen minutes 

 from the time the water is poured on the coffee. The moment of ebullition is accurately 

 noted, and the boiling is continued for exactly one hour. During this time the flask is 

 lifted from the sand-bath three or four times, without detaching the condenser, and, by a 

 circular movement, any particles which have been projected above the level of the water 

 are washed down from the sides of the flask. 



