Joly — On a Hydrostatic Balance. 351 



the bottom of the containing sphere, and adhesion at these points is 

 avoided. The arrangement also obviously secures the advantage 

 of guarding the fine wire against the effects of a weight erroneously 

 added in excess of the power of the balance : indeed, if care is 

 taken in adding and removing weights, the wire remains uniformly 

 in the one state of strain. It is thus kept straight and true. Again, 

 should the wire break, the descent of the pan is arrested by the eye 

 and link. The arrangement, too, renders a slow motion of dis- 

 placement very readily observable. I may observe, that the eye 

 encircling the link is gapped at one point, to enable the link to be 

 removed if desired. The construction of this link is simple, but ne- 

 cessitates the exercise of a little care in the process of equilibration, in 

 order that the effect of adhesion at the link may be guarded against. 

 The process of equilibrating is as follows : — The larger weights 

 being added successively in the usual way, the equilibrium, we 

 will suppose, determined to 1 centigramme, and the milligrammes 

 reached, the 5 is added : if there is no immediate effect we 

 grasp the link with the ivory forceps (used with the weights), and 

 bring it down to the centre of its run, then release it, and observe 

 its motion. It will most probably ascend; but there may have 

 been adhesion between the link and eye, and we may find that it 

 descends slowly. Its velocity of motion is in either case instruc- 

 tive after a little practice, and, as with the chemical balance, will 

 enable us to save trials. If the link in the present case very slowly 

 descends, we replace the 5 with 4 mgrs., and repeat the process of 

 drawing down the link. We take care also to close the glass 

 door while observing the motion. The result will be perfectly 

 definite. It will ascend with the 4, suppose ; descend with the 5, 

 the starting point in each case being the centre of its run. And I 

 may observe that it will, in this way, indicate a less quantity even 

 than 1 milligramme. Thus recently using this balance with a new 

 set of weights, I detected discrepancies in the 1 centigramme 

 weights, as compared with the added milligrammes, which, on sub- 

 sequently evaluating on a delicate chemical balance, were severally 

 found to be '7 and *3 of a milligramme. It would, doubtless, be 

 easy to arrange, so that the adhesion necessitating these precau- 

 tions when weighing with the milligrammes would be eliminated. 

 In fig. 3 two kinds of bearing are suggested: one — (a) where there 

 is contact at two points ; the other, (b) at one point only, when the 



