208 REPORT OF THE COMMISSIONER OF FISHERIES. 
6. Gelatin not liquefied; stab cultures and plate cultures give char- 
acteristic growths. 
7. Nitrates reduced to nitrites. 
Bacterium lactis aerogenes is a closely allied form, but differs from 
Z. coli in that it is nonmotile; it produces larger amounts of gas in 
dextrose broth (75 per cent), and it does not produce indol. It is 
nonpathogenic. 
B. cloacz also produces large quantities of gas in dextrose bouillon’ 
(from 65 to 75 percent). It hquefies gelatin, casein, and blood serum, 
and produces indol and nitrates. 
Samples of water to be tested were collected in sterile 25 c. ec. tubes 
by means of an apparatus similar to that suggested by Professor 
Bolley for use in deep wells. The tubes were made from large 8-inch 
test tubes by drawing out slightly in a Bunsen flame the open end of 
the tube, bending the lengthened portion to a right angle with the 
rest, and finally drawing it out into a fine capillary tube. These 
tubes were sterilized, and after a partial vacuum had been secured by 
heating, the fine tube was sealed in a flame. A rack holding 20 of 
these tubes was easily carried in a small grip. The collecting appa- 
ratus consisted of a solid block of brass 9 inches long by 14 inches 
wide by three-fourths inch thick, against the flat side of which the 
tube was firmly held by two sets of clamps, the sealed capillary tube 
passing through a hole bored in the upper end of the block. In col- 
lecting the water samples the apparatus was lowered by a stout cord 
to the desired depth and the sealed tube broken by a metal slide, 
which was operated by allowing a weight to run down the line on 
which the apparatus was lowered. The partial vacuum in the tubes 
usually fitled them one-half to three-fourths full of water. These 
tubes were again placed in the rack and carried to the laboratory 
unsealed, for a length of the bent tube sufficient to protect the sample 
from outside contamination usually remained after the sample had 
been collected. When the tubes reached the laboratory, at no more 
than four or five hours after collection of the water samples, the tops 
were passed through a flame and enough of the glass broken away 
with sterile forceps to allow the entrance into the tube of a sterile 
1c. c. pipette. Samples were immediately transferred from these 
tubes to the different culture media, as already described. 
When samples were taken in deep water, two collections were 
usually made at each locality visited, one a foot below the surface of 
the water and a second a foot off the bottom of the river. In the 
shallow water near the shores samples were collected by plunging 
sterile bottles below the surface of the water. In examining clam 
flats and mussel beds left uncovered by the tide, samples of sand and 
mud were collected at low tide and samples of the water covering 
these grounds on the flood tide. 
