vacuum which would otherwise be in the pipes, and which would tend to draw air from the house into the pipes instead of from the pipes into the house at any leakage. It would probably not be a minute before the people in the house would perceive the smell at various places. The manipulator of the peppermint must remain perched on the roof until those inside have had time to make their observations, otherwise he will infallibly bring the smell with him. The test described is an excellent one. It is searching, and is simple in application, but it has one drawback. It is impossible by means of it exactly to localise a leakage. This drawback does not apply to the smoke test. A smoke machine is nothing more nor less than a centrifugal pump attached to a vessel for generating smoke. The pump pumps smoke out by a pipe, which may be inserted in any pipe in direct communication with the drain or in an aperture made for the purpose. The test is, in all respects, similar to the peppermint one, except that the leakage is not smelt but seen. After the test has been performed the drain may be opened. This may be done by breaking into a pipe in front, by breaking off a collar, or by punching a round hole in the pipe. In any case it will be possible to judge much of the condition of the drain by the manner in which water runs through the pipes. If we have discovered that there is sufficient total fall, we can now see whether or not it is uniform. We shall, as remarked before, find in six cases out of every hundred examined that there is total stoppage, that no sewage whatever leaves the premises, and that consequently it must all be depositing under the basement. If the drain, after all tests so far applied, and from what can be seen of it, appear to be in good condition, it may be further tested by filling, or attempting to fill it with water. There is probably not an average of one drain in a thousand in London which would remain full of water for an hour. For the rest it is necessary to examine all appliances, to trace the pipes from them, and sometimes to test these pipes. The engineer has now completed his inspection, and has but to consider how he will make the best of a bad job, and put things to rights. At the beginning of his paper Burton expressed his intention of confining himself to a description of defects, and said he should not describe what he considered a perfect system; he, however, points out one or two of the chief features of the arrangements in the house which he calls well drained. [Illustration: 22. Disconnecting Chamber.]