EUpaiion of Water hy its lateral ABion. 495 



its width, and greatly retard the expenditure, it will be beft to introduce them all through 

 holes made in the circumference of the tube, and to take ofF their ends even with its infidc. 

 When water is thus raifed in an additional number of columns, the water tubes ftiould 

 be made ftill fliorter than was recommended before, and to prevent the water which is 

 raifed by the fecond feries from being wafted, the rarefying tubes fhould be carried up to 

 fuch height above each veflel, that the column may counterbalance the a£live preflurc of 

 the atmofphere before it arrives at the top. Recourfe may be had to this method when 

 feveral conical tubes are employed, as is reprefented in Fig. 3, PI. XIII. In this laft cafe, 

 however, it will be unneceflary to make any addition to the loweft veflel, as the ex- 

 penditure can only take place from the loweft ciftern. 



In all the arrangements that have been hitherto defcribed, the ftream through the conical 

 tube is fufFered to run one half of its time without producing any ufeful effeft. It will, 

 however, be of very great advantage to keep the lateral action conftantly in work in rarefy- 

 ing the air in different parts of the apparatus. The raifed water may then be let out of 

 one fet of veflels while it is rifing into another fet, and the water that flows out of one de- 

 fcending veflel, and has opened the valves of the firft fet, may, by being carried into a 

 {mall ciftern a fecond time, and delivered from thence by afyphon into another defcending 

 vefl"el, be made to open the valves of the fecond fet. 



As it may fometimes happen that all the valves may be half open, and much air be thus 

 admitted into the ftream, it is likely that cocks fixed in the rarefying tubes, and placed in 

 water, would on this occafion be preferable to valves, particularly when the rarefying 

 tubes can be filled with water, for then no air can defcend into the jet. I think, how- 

 ever, that this can only be neceflary when feveral air pipes are connected with the conical 

 tube, and the air would enter the ftream on all fides. 



In this manner by the aflTiftance of the lateral action of a ftream of water through one 

 conical tube, and the preflure of the atmofphere, an ufeful quantity of water may be 

 raifed to a confiderable height. 



I am, SIR, 



Your humble fervant, 



WILLIAM CLOSE. 



V. On 



