the bottom of the pipe PP, and fills the space M where any 

 possible dirt collects. It then overflows the ring d and flows 

 in a thin film (shown by the arrows) and is thrown in the 

 gutter H, leaving through the spout with the thermometer L. 

 The exhaust steam from the flyer f goes up through eight 

 holes X X into the cylinder and heats the drum D. In run- 

 ning about 1,000 Ibs. of milk per hour I raised the milk from 

 -about 54 to 155 with the exhaust steam alone, but when I 

 ran 1,500 Ibs. an hour I had to use some live steam, which is 

 led through the pipe S under the bottom of the drum. If the 

 steam pressure and milk supply is uniform this apparatus 

 heats it steadily within a variation not to exceed 4 or 5 Fah. 

 A wooden jacket would be advisable to economise steam es- 

 pecially in winter. The condensed steam escapes by K and by 

 two small holes II in the bottom. 



Fig. 24. 



While the milk does not get any perceptible "boiled 

 taste," there is, after running 3,000 Ibs. through it, quite a 

 layer of coagulated albumen, which takes some labor to clean 

 off. Yet I would far rather clean that one drum instead of 

 five or six cans, even if the albumen deposit is less there. In 

 Fig. 24 the apparatus is shown as connected with three Baer 

 coolers. 



I also proposed to utilize the same idea by having the 

 bottom of the drum cast with shovels on, and rotate it by a 



