ttieafuriiig the mutual Gravitation of Bodies. 447 



Mr. Michell had prepared two wooden ftands, on which the leaden weights were to be 

 ■fupported and puflied forwards till they came almoft in contaft with the cafe ; but he ftems 

 to have intended to move them by hand. 



As the force with which the balls are attradled by thefe weights is cxcefilvely minute, not 

 more than one fifty-millionth part of their weight, It is plain that a very minute difturbing force 

 ■will be fufficient to deftroy the fuccefs of rhe experiment : and from the following experi- 

 ments it will appear, that the difturbing force moft difficult to guard againfl is that arifing 

 from the variations of heat and cold ; for, if one fide of the cafe is warmer than the other, the 

 air in conta£l with it will be rarefied, and in confequence will afcend, while that on the 

 other fide will defcend, and produce a current which will draw the arm fenfibly afide *. 



As Mr. Cavendifh v/as convinced of the neceffity of guar^ling againfl; this fource of er- 

 ror, he refolved to place the apparatus in a room which (hould remain conftantly fliut, and 

 to obferve the motion of the arm from without by means of a telefcope, and to fufpend the 

 leaden weights in fuch a manner that he could move them witliout entering into the room. 

 This difference in the manner of obferving rendered it neceflary to make fome alteration in 

 Mr. Michell's apparatus : and as there were fome parts of it which Mr. Cavendifli thought 

 not fo convenient as could be wlftied, he chofe to make the greateft part of it afreflr. 



Plate XIX. is a longitudinal vertical fe£lion through the inftrument, and the building in 



•which it is placed. ABCDDCBAEFFE is the cafe, x and x are two balls wliich 



are fufpended by the wires h x from the arm gh mh, which is itfelf fufpended by the flen- 



der wire gl. This arm confifts of a flender deal rod h m h, ftrengthened by a filver wire 



.hgh\ by which means it is made ftrong enough to fupport the balls, though very light f. 



The cafe is fupported and fet horizontal by four fcrews, refting on pofts fixed firmly into 

 the ground : two of them are reprefented in the figure by S and S ; the two others are not 

 reprefented, to avoid confufion. G G and G G are the end walls of the building. W 

 and W arc the leaden weights which are fufpended by the copper rods Rr PrR and the 

 wooden bar r r from the centre pin ?/>. This pin pafTes through a hole in the beam H H, 

 perpendicularly over the centre of the inftrument, and turns round in it, being prevented 

 from falling by the plate />. M M is a pulley faftened to this pin, and M/« a cord wound 

 round the pulley and paffing through the end wall, by which the obferver may turn it 

 round, and thereby move the weights from one fituation to the other. When the weights 

 are in the poGtion reprefented in the engraving, both confpire, by their adlion on the fmall 

 weights X Xy to draw the arm in the fame direftion ; but when, by means of the pulley 



* Mr. Caffini, in obferving the variation-compafs placed by him in the Obfervatory (which was conftrufled 

 fo as to make very minute changes of pofitiou vifible, and in which the needle was fufpended by a filk thread), 

 found that ftanding near the box, in order to obferve, drew the needle fenfibly afide ; which I have no doubt 

 was caufed by this current of air. It muft be obfer\'ed that his compafs-box was of metal, which tranfmits heat 

 fafter than wood, and alfo was many inches deep ; both which caufes ferved to increafe the current of air. To 

 tJiminifli the tifefl of this current, it is by all treans advifable to make the box in which the needle plays no» 

 much deeper than is neceflary to prevent the needle from ftriking againft the top and bottom. C. 



f Mr. Michell's rod was entirely of wood, and was much ftronger and ftitfer than this, though not much 

 tieavier ; but as it had warped when it came to Mr. Cavendifli, he chofe to make another, and preferred this 

 form, partly as being eafier to conflruft and meeting with lefs refiftance from the air, and partly becaufe, from 

 its being of a left complicated form, he could more eafily compute how much it was attrafted by the weights. 



3 M 2 MM, 



