TO THE MOVEMENT OF MACHINES. 



515 



too narrow ; at all events it ought to exceed half an inch. I have also 

 made experiments with voltaic pairs arranged like the calorimotor 

 of Hare, but there were reasons for rejecting these also. It is a very 

 different thing to make an isolated observation, and to put in requisition 

 the galvanic action for whole hours and days. It is in the latter case 

 that for practical purposes measures are required, the necessity of which 

 had not been before anticipated. It will be also necessary to reject the 

 use of copper as a negative metal ; the expense of employing silver, 

 platina, or at least copper well plated with silver, gold, or platina, must 

 be no obstacle. The solution of the copper in the sulphuric acid, how- 

 ever weak it may be, and its reduction into a metallic state, by the se- 

 condary effects of the nascent hydrogen, give rise to partial galvanic 

 effects, by which the principal action is much affected, and to avoid 

 which the greatest pains must be taken. In fact the motion of the 

 magnetic apparatus was sometimes suddenly slackened or entirely in- 

 terrupted, and on examining more closely I found that metallic parti- 

 cles of cementing copper or of iron had been deposited all along the 

 pieces of wood interposed, or upon the bottom of the troughs, and thus 

 formed a partial circuit. I shall speak of zinc hereafter. The following 

 is the table of observations which I have made on the magnetic power 

 of the horseshoe bar of iron above dsecribed. 



The values given in the last column have been calculated according 



to the formula A = ^^ i—, the constants of which were found by 



20 +a; ^ 



the method of the least squares. It is true that there are considerable 

 differences between the obsei-vations of the same series, but there was 

 no reason to choose those which agreed the best with each other, and 

 to attribute the differences to an error of observation. 



10. 

 I have read in an extract from the memoir of the Abbe dal Negro 

 inserted in the "Annali delle Sc, 1813 [1833?], Marzo e Aprile, 



• The first series of experiments, which were made with the pair of plates of 

 144 square inches, presented such diflerent values that no use could be made 

 of them. I have sought in vain for the cause of these anomalies. After a fort- 

 night the experiments were repeated, and gave values but little diflerent. This 

 IB proved by llie Table. 



Vol. I Pakt IV. 2 n 



