CKEAP AND SIMPLE AIR PUMP* fiJS 



(which W€r€ secured with oiled leathers, in the usual man« lisatllon 

 uer,) from leaking: but, even with this imperfection, the 

 mercury has frequently descended, in the siphon-gage, 

 to within ^V» ^"<i once or twice to ^^ of an inch oj^ 

 the level. I have thought of a plan of obviating this 

 imperfection, (to which, by the by, every pump is more 

 or less liable) and shall be happy in communicating it to 

 you, if it answers my expectations. There is one objection, 

 which may be made to this pump; and that is, that the 

 valve d will be exposed to censiderable friction. This may 

 be obviated, in some degree, however, by making it of iron, 

 or steel, which is known to wear very little in brass; and 

 still farther, by fixing it loosely to the end of the rod, in the 

 same manner as the valves are fixed to the keys in Potter's 

 patent flutes ; or it may be entirely removed, at the expense 

 of a trifling inaccuracy in the pump : as follows. C, (PI. 

 Vin, fig. 3,) is the piston-rod, D the piston, andm a bra-js 

 plate, fixed to the end of the rod, instead of the valve d: lo , 



this plate a piece of leather is screwed, which is shaded 

 darker in the figure; this leather, when the rod is pub^hed 

 down, covers the hole g g, and prevents the return ol the 

 air into the upper part of the barrel. The hole rnay be 

 made so small, as not to contain, in a barrel of 10 inches ca- 

 pacity, more than '00001 of the air, independent of the size 

 of the receiver. I have had a pump constructed with a 

 vidve of this description ; and find it to answer perfectly well. 

 It may be farther improved, by fixing a small wire, or 

 valve, to the brass piece, (as is shown by the dotted lines) 

 which will fill the hole g g, when the piston is pressed 

 downwards : only care must be taken, that it do not project 

 so far, as to open the valve c in the bottom of the barrel. 



The price of a single-barrelled pump, of this description, 

 (as represented, PI. ViU, fig. 1,) made by Messrs. Dollond, 

 will not exceed 10 guineas. It may also be made with two 

 barrels. It is evident, that this pump will also act as a 

 condenser, with a very little additional apparatus; which, 

 as it will not diff'er from that usually employed, it is unne- 

 cessary to describe. I must beg leave to add, that the 

 juncture, I formerly recommended, for bent tubes*, will be 



» SecTol. X>wV,p. 584. 



very 



