for preventing Water-Pipes Bursting during Frost. 239 



ing. I may add to this, that the practice of building up water-pipes 

 in the centre of a wall, with the view of preserving them from burst- 

 ing, is dangerous in the extreme. The ill effects of the frost on the 

 pipes is often imperceptible at the time. The expansion sustained 

 by the pipe may not be sufficiently great to burst it ; but in that 

 case it produces a regular enlargement of the pipe, which part ne- 

 cessarily becomes the weakest ; and it accordingly yields sooner or 

 later to the mere pressure of the water. Cases of burst pipes in 

 summer have thus not unfrequently to be attributed to the frost of 

 the previous winter. 



The only really practicable and unfailing means of preserving 

 water-pipes from bursting with frost, is to adopt the simple precau- 

 tion oi keeping them empty. This has been well ascertained to be the 

 only security in every country where precautions are adopted. The 

 best of those in use at present, consists in attaching a stop-cock to 

 the lowest part of the pipes, and another immediately above it. By 

 the one, the water is shut off when a low temperature approaches, 

 and by the other, the pipes are emptied of the water they contain. 

 To render these precautions of any avail however, the utmost watch- 

 fulness is necessary, and even where the cocks exist, they are rarely 

 used in time. Besides, a severe frost will frequently occur in course 

 of a single night, and in that case the opportunity of shutting off 

 the water is lost. The same objection applies to the mode of heat- 

 ing by fires, of circulating the water, and to all other precautions of 

 a similar kind. 



Reasoning on this, I have conceived the possibility of employing 

 some self-acting apparatus, which, on the approach of a low degree of 

 temperature, would of itself shut off the water and empty the pipes ; 

 or, in other words, of having a machine so constructed and regu- 

 lated, that it would shut a cock before the freezing-point of water, 

 and open it again, when the temperature assumed its normal state. 



I first thought of applying a machine on the principle of the ther- 

 mometer. If, in place of the small bulbous glass of an ordinary 

 thermometer, I took a vessel of much larger dimensions, and in place 

 of being sealed at the top, it was closed by means of a small sliding 

 piston on the surface of the mercury ; — if it were possible to adjust 

 this, so as to elevate the piston at a temperature of 60°, it is evident 

 that a decrease of the temperature to the freezing-point would pro- 

 duce a diminution of bulk in the mercury, and a corresponding de- 

 scent of the piston. A stop-cock on the supply-pipe, attached to the 

 piston-rod, might thus be made to stand open at all ordinary tem- 

 peratures, and to shut precisely at the freezing point. 



Such an apparatus, however, would be subject to a few serious and 

 insurmountable objections. In the first place, it would require some 

 intricate mechanical arrangement, to prevent it diminishing the sup- 

 ply of water at low temperatures, although still above the freezing- 

 point. In the next place, its whole available power in shutting the 

 cock would necessarily be small, since it could only be derived from 



