BACTEE1AL COUNT OF MILK AND DIET TEST. 5 
Tbis table, like the others, shows considerable variations; No. 1, 
which had a bacterial count of 768,000, tested "fair" by the sediment 
test, and No. 8, which has a count of 7,200, tested "bad." These 
disks are shown in Plate II (lower). 
COMPARISONS WITH FILTERED MILK. 
After comparing the bacterial count with the various sediment tests 
of unfiltered market milk, it was decided to make a comparison after 
the milk was filtered through such substances as are frequently used 
as strainers by farmers to remove dirt. Twenty samples were filtered 
through 4-ply cheesecloth and the Lorenz disks compared with the 
bacterial count. 
The table below shows the results obtained from 10 average samples 
out of 20y filtering through cheesecloth. 
Table 4. — Comparison of bacterial count with Lorenz sediment test (milk filtered through 
cheesecloth). 
Sample No. 
Bacteria 
per cubic 
centimeter. 
Character 
of sediment. 
Sample No. 
Bacteria 
per cubic 
centimeter. 
Character 
of sediment. 
1 t . 
109,000 
67,000 
46,000 
24, 000 
639,000 
Good. 
Do. 
Do. 
Do. 
Do. 
6 
33,000 
84,000 
93,000 
54,000 
316,000 
Good. 
2 
7 
Do. 
3 
8 
Do. 
4 
9 
Do. 
5 
10 
Do. 
Twenty samples were filtered through one ply of Canton flannel 
and the bacterial count compared with the Lorenz disks. Table 5 
shows the results obtained from 10 average samples out of 20. 
Table 5. — Comparison of bacterial count with Lorenz sediment test (milk filtered through 
1-ply Canton flannel) . 
Sample No. 
Bacteria 
per cubic 
centimeter. 
Character 
of sediment. 
Sample No. 
Bacteria 
per cubic 
centimeter. 
Character 
of sediment. 
1 
78,000 
31,000 
41,000 
108,000 
18,000 
Good. 
Do. 
Do. 
Do. 
Do. 
6 
19,400 
316,000 
129, 000 
149,000 
119,000 
Good. 
2 
7 
Do. 
3 
8 
Do. 
4 
9 
Do. 
5 
10 
Do. 
Twenty samples were filtered through 1-ply ordinary absorbent 
cotton, covered above and below with 1-ply cheesecloth. The Lorenz 
disks were compared with the bacterial count, as in the preceding 
table. Table 6 shows the results obtained from 10 average samples 
out of 20. 
