12

     In the following experiments as many of the large ventilating holes as could
be filled up were closed with tightly packed paper and the room was made as
air-tight as most native huts can be made.

     In the floor of the room four types of rat burrows were present, each
consisted of a small wooden box to contain rats and fleas and one or two passages
leading from it to the floor of the room. The boxes measured 21 by 10¾ by
12 inches. They are fitted with hinged lids and are sunk in the ground. Inside
each box is a rack to hold test-tubes containing fleas. The cage for the rat
can be placed inside the box. The channels consist of a series of short sections
of earthenware pipes sunk in the ground. Each section is made up of concave
country roofing tiles placed together to form a somewhat tapering pipe, the
narrow end of one section fitting into the broad end of the other. The narrow
end of the lowest of these sections projects slightly into the box. These
"burrows" measured from 5½ to 12 feet.

     There are four types of burrows :—

        Type A.—A blind burrow with a single opening.

        Type B.—A blind burrow with a single opening inside the room. The
box for the cage was outside the room for facility of obser-
vation.

        Type C.—Two burrows with S-shaped curves on the horizontal, leading to
the box containing the cage.

        Type D.—Same as type C, but with the curves vertical.

     Over the tiles 4 to 6 inches of earth were placed and, after the fleas and rats
had been placed inside, the lids of the boxes were carefully sealed down with moist
clay.

     These burrows and observation chambers were originally devised by Captain
Gloster, I.M.S., for his experiments with the Clayton gas apparatus.

     The results of experiments in this room are detailed in Experiments VII to
IX.

     Experiment VII shows the results when the gas was generated outside the
room and introduced by tubes without pumping. Some tubes were also led a
little distance into the burrows and the mouth of the burrows closed up. The gas
is lighter than air and better results were obtained at the upper layer of the room
than on the floor. Neither rats nor fleas in the burrows were affected by the
gas. At the level of the floor many fleas in cages with sand at the bottom
managed to escape by burrowing into the sand, while those in bran were killed.

     In Experiment VIII the gas was introduced by larger calibre tubes and
pumped in by a fan; four of the tubes were led into the burrows and the mouth
of the opening closed up with mud round the pipe. This was a most successful
experiment. Fleas in all situations,—in the room, in cages with sand and bran at