274 Exptrimtnts on the laitral Cammun't cation of Motion h Fluids. 



diture through additional tubes, whatever may be their pofition. We fhall in the next place 

 examine the mode of aftion hy which the atmofphere produces this augmentation, and de- 

 termine the refult from its caufe. I {hall begin with the cafe beft adapted to favour the 

 action of the atmofphere, which is, that of conical diverging tubes of a certain form, which, 

 we have not yet confidered. 



Let the extremity AB, fig. lo, Plate VHI, of the tube ABEF be applied to an orifice 

 formed in a thin plate. The part A B C D is nearly of tha figur? of the contra6led vein, 

 which form has been fliewn to make no perceptible alteration in the expenditure (Experi- 

 ment IV.} The fluid which iffues through C D is difpofed to continue its courfe in the 

 cylindrical form C D H G. But if the lateral parts of the diverging conical tube C E G,, 

 D F H, contain a mafs of the fluid at reft, the cylindrical flream C D H G will commu- 

 nicate its motion to the lateral parts (by Prop. I.) fuccefQvely from part to part. And pro- 

 vided the divergence of the fides CE, DF, be fuchas is beft adapted to the fpeedy and com- 

 plete lateral communication of motion, all the fluid contained in the truncated cone CDEF 

 vAW at length acqnire the fame velocity as that of the ftream which continues to iffue 

 through C D. On this fuppofition, while the fluid ftratum C D Q_R, preferving its velocity 

 and thicknefs, would pafs into R Q_TS, a vacuum would be formed in the folid zone- 

 Rm r S Q_n o T. Or othervvife, if it be fuppcfed that the ftratum C D Q_I\i preferving its 

 progreflive velocity, fliould enlarge in RQ_TS, this cannot happen without its becoming 

 thinner, and detaching itfelf from the ftratum which follows, and by that means leaving a 

 ■" vacuum equal in magnitude to the zone laft mentioned. A fimilar effeOi would take place- 

 through the whole of the tube C E; and if the quantity C m be fuppofed to be invariable, 

 the fum of all thefe void fpaces will be equal to the folid zone VExGzYFH. 



From this confideration, we fee that the lateral communication of motion caufes the fame 

 effeiSl in a conical tube, whether horizontal or vertical, as gravity produces in the defcendi 

 ing tube of Propofition IV. The atmofphere in this cafe alfo renders part of its preffure 

 adlive on the refervoir, and at EF. If the aftion of the atmofpliere upon the refervoir 

 increafes the velocity of the fedion CD, this velocity will communicate itfelf likewife t^ 

 the whole fluid C D F E, and the tendency to a vacuum will take place as before ; but 

 fmce the a£tion of the atmofphere is exerted equally at E F, it will take away at E F all the 

 velocity which it added at C D. ; fo that, being dedufled from the fame mafs, and in tbfe. 

 fame time, at E F, the fluid will not ceafe to be continuous in the pipe. It is found by 

 computation, that this will happen when the velocity of CD is increafed in th& ratio of 

 C Dmo E ¥\ 



By applying the general laws of motion to the lateral fluid filaments of the ftream which-, 

 ;fl"ues through AB, it is found that they tend to describe a curve which commences withia. 

 the refervoir, for exam.ple, at A, and continues towards C S E. To determine the nature of 

 this curve, it is requifite taknow, and to combine together by calculation, the mutual con- 

 yergency of the fluid filaments in A B, the law of the lateral communication of motibti 

 between the filaments themfelves and their divergent progreffion from C to £. Thefe cam>. 

 binations and calculations are perhaps beyond the utmoft efforts of analyfis.. While tUt 

 tube ABFE poflefles a diflMsrent figure from this natural curve, the refults of experiment 

 will always differ more or lefs from the theory. 



JExpmmetit XIII. The compound tube A fi E jE ©f the fame fig. io> having the following 



