Q. I think the plumber is right and you are wrong about your advice to Mr.---------, when you say you can carry the return below the floor and then rise to the boiler and have circulation. If he does as you say he will get a trap where he rises, and the cold water will stay there and he can't get any circulation. I don't think it makes much difference how high the circulation is, either. I would say to Mr.--------- it is a waste of time and money to run the circulation-pipe as he wants to do it.

Figure 103.

A. We must still maintain that a cir-culation will take place in the pipe below the floor. It depends, as stated in the original article, on the cooling of the upper or flow-pipe. Then the density and the weight of the colder air in the descending pipe, at the left of the original diagram, indicated by the descending arrow, are greater than in the rising pipe, and it is this greater weight which causes the circulation.

We, of course, are not to be understood as maintaining that a circulation will take place at every distance below the water-back, and under all circumstances. How far below the water-back the return can be carried will depend on the height of the rising column. In the case submitted to us it is assumed that the upper pipe is at least six feet above the water-back, and the lower pipe not more three feet below it.

Figure 104.

Figure 103 represents the plan we first proposed. Figure 104 is the method suggested by a firm in New York City which has had extensive experience in the fitting up of hot-water apparatus. It will be seen tha the lower return is here connected with the water-back return-pipe.