WORK OF P. B. DAVIS. 183 



of flow in the glycerol viscosimeter. The viscosimeters were filled by means of care- 

 fully calibrated pipettes with such a volume of liquid as was required to fill them 

 from the middle of the upper to the middle of the lower bulb. With such an arrange- 

 ment, according to Appleby 1 the alteration in hydrostatic pressure due to small 

 variations in the volume of the liquid is a minimum. 



On account of the steaming of the viscosity thermostat at high temperatures, the 

 viscosimeters after filling were provided with a simple apparatus designed to exclude 

 dust and moisture, and to eliminate the danger of obstructing the free passage of air 

 from one arm of the viscosimeter to the other during the flow of the liquid through 

 the capillary. This consisted of two T-shaped guard tubes each provided with a 

 small bulb in the long arm. These bulbs were filled with cotton- wool to filter out 

 dust particles. One end of the cross-arm of a guard tube was connected to each arm 

 of the viscosimeter, and the opposite ends joined by a short piece of rubber tubing. 

 The liquid was drawn up in the capillary arm of the viscosimeter by compressing the 

 rubber tube and attaching the bulb arm of the guard tube on that side to an aspirator. 

 The air entering through the guard tube on the opposite (reservoir) arm is thor- 

 oughly dried by passing over calcium chloride and freed from dust by the cotton wool 

 mentioned above. When the liquid has risen to a point slightly above the upper 

 mark on the viscosimeter, the compression on the rubber tube is released and the air 

 pressure immediately equalizes itself on both sides of the viscosimeter. By this 

 arrangement no air from the outside enters during the actual flow of the liquid 

 through the capillary. 



Viscosimeters, while being read, were supported in the bath by means of special 

 clamps with cork-lined jaws, attached to a heavy stand which was carefully leveled. 



Specific-gravity determinations were made by means of the Ostwald pycnometer 

 as modified for liquids with large expansion by Jones and Veazey. 



All measuring flasks and pipettes were carefully calibrated either by direct weigh- 

 ing or by the method of Morse and Blalock 2 to hold aliquot parts of the true liter at 

 20, and solutions were brought to within 0.1 of this temperature before being 

 diluted to the mark. 



SOLUTIONS. 



Because of a limited supply of salts at hand all solutions were made up at 20 and a 

 temperature correction made for temperatures above 25. The expansion coefficient 

 for glycerol was taken as 0.00049 and was determined from specific-gravity measure- 

 ments in the viscosity work. 



Solutions of the N/0.75, N/2, N/4, and N/10 were made by direct weighing of 

 the anhydrous salt, which had been previously dried for some time in an air-bath at 

 130 to 135. From the N/10 solution the N/50 and N/100 solutions were made 

 by dilution. These in turn served as mother-solutions for the N/200 and N/40.0, 

 from which the N/800 and N/1600 solutions were made in a similar manner. The 

 latter was made by diluting the N/400 solution four times. 



Because of the hydroscopic nature of the solvent and the difficulty in obtaining 

 proper drainage of burettes containing it, measuring flasks were used exclusively in 

 the dilution of solutions in glycerol. In every case the proper measuring flask was 

 filled with the mother-solution, set in a thermostat at 20, warmed in an air-bath to 



Uourn. Chem. Soc, 972, 2000 (1910). ^Amer. Chem. Journ., 16, 479 (1904). 



