246 Mr. C. V. Boys on Water-wipes 



and this the more easily as its section departs more from the 

 circle; so the very fact of its becoming more circular makes it 

 less ready to change its form. In a very little while, there- 

 fore, though originally weaker, it will become as strong as 

 neighbouring portions : therefore an elliptical pipe under the 

 action of frost is in a state of stable equilibrium, and, instead 

 of giving way in the bulbous manner of a circular pipe, it uni- 

 formly becomes more circular. Now the expansion of water 

 in becoming ice is known: and therefore it is easy to calculate 

 by compound interest how many complete freezings (that is, 

 freezings from one end of the pipe to the other) any given 

 section of piping will stand before it becomes round. Of 

 course, in practice, the whole length of a pipe does not get 

 frozen; yet if it were originally all of it elliptical, the unfrozen 

 portions would be effective in preventing the more exposed 

 parte from bursting ; because as soon as the exposed portions 

 have become rounder than the rest, the latter, and not the 

 former, will yield. If two places in the pipe become com- 

 pletely frozen through and then the intermediate portion 

 freezes, it is true that parts beyond the frozen plug will have 

 no effect. 



Mr. Powell and I tried a series of experiments on the sub- 

 ject to see if, in practice, the pipes behaved as we expected. 

 We obtained a quantity of f-inch lead pipe, about ^j inch 

 thick, and some thin composition pipe of the same size. The 

 piping was cut into lengths of about three feet ; half of them 

 were squeezed into an approximately elliptical form, and the 

 rest left circular in section. The degree of ellipticity was 

 such that the major axis was a little more than twice the minor 

 axis. One end of each pipe was squeezed together and soldered. 

 Into the other ends brass plugs cut with a sharp thread were 

 screwed while hot, having been previously smeared with a 

 cement of rosin, beeswax, and red-ochre. After these plugs were 

 inserted and while still hot, the lead was, as an additional pre- 

 caution, squeezed over a narrower portion of the plug above 

 the screw-thread. In each plug a hole had been drilled and 

 tapped. Through these holes the pipes were filled with water; 

 and then iron screws, with washers of leather boiled in bees- 

 wax and tallow, were used to make a tight joint. The pipes 

 were then all laid together in a long box, and surrounded with 

 a freezing-mixture. When a short test-pipe of the same dia- 

 meter showed that the water was completely frozen, the pipes 

 were removed and thawed. The round composition pipe was 

 burst. The round lead pipe was swollen in an irregular 

 manner. The elliptical piping had become slightly rounder, 

 but was perfectly uniform in shape from end to end, which 



