COOLING HOT-BOTTLED PASTEURIZED MILK. 3 
Thirty crates of bottles, stacked in six piles each five crates high 
inside the box, were cooled at one time. Before the crates were 
stacked on the platform the movable sections, were removed, so that 
the space under the crates was open and the rest of the platform 
closed. By this arrangement it was possible to force all the air 
directly through the crates of bottles. 
When air was forced up through the crates the operation was as 
follows : Air was drawn by blower A through pipe B from the outside 
of the building and forced through pipe C into the bottom of the 
cooling box. From that point it was forced up through the piles of 
crates and passed out from the top through pipe D. In this case 
damper No. 1 in pipe C was open and damper No. 3 in pipe F was 
closed, while damper No. 2 in pipe E was closed and damper No. 4 
in pipe D was open. 
When air was to be forced down through the crates, and conse- 
quently had to enter at the top of the box, it was necessary simply to 
change the dampers in the pipes. To force air in at the top through 
pipe E, damper No. 1 was closed and No. 3 opened; in the outlet 
pipes damper No. 2 was opened and No. 4 closed. This system of 
pipes and dampers would not be necessary for cooling milk on a 
commercial scale, but was necessary for experimental purposes. 
DESCRIPTION OF INSTRUMENTS AND METHOD OF RECORDING DATA. 
Temperatures * were determined with copper-constantin thermo 
elements made from No. 30 wire, tested to the required degree of 
accuracy for thermoelectric homogeneity. Junctions were soldered, 
wound with silk thread, and inclosed in thin glass tubes. Where 
leads were exposed to steam or air blast the double-silk insulation was 
protected by being incased in black-rubber tubing. Each element 
was composed of two pairs of couples in series. A calibration curve 
for each element was constructed from the potential given when the 
cold junction was immersed in a thermos bottle containing crushed 
ice, and the hot junction was immersed (1) in melting ice, (2) in a 
fusion of recrystallized sodium sulphate (Richard's point), and (3) in 
steam at known barometric pressure. In each case the temperature 
was also taken with a thermometer calibrated by the Bureau of 
Standards. From the calibrations so obtained deviation curves were 
constructed and used with Adams's 2 table. 
Potentials were measured with a Leeds and Northrup potentiometer 
and galvanometer. The potentiometer was calibrated by the Bureau 
i We are greatly indebted to Dr. W. M. Clark, of the Dairy Division, for preparing, describing, and oper- 
ating a system for temperature measurements by means of thermocouples. 
2 Adams, L. H. Journal of American Chemical Society, 36, p. 65, 1914. 
