528 The Rev. T. R. Robinson's Experimental Researches on the 



In examining the different series of L given by the various magnets with 



which I have experimented, it will be observed that — — -, the rate of increase 



of L for a given change of ^ decreases as this variable increases, and so as to 

 indicate that in each of them L has a limit which it cannot pass by any amount 

 of excitation. When I first noticed this, I explained it by supposing that the 

 cause of magnetism, whatever it be, is present in each molecule only in finite 

 quantity : if it be a hypothetic fluid decomposed into two of opposite powers, 

 when all of it is so resolved, the process must stop. If we assume Ampere's 

 currents, when all of them are forced into the same plane and direction, there 

 can be no further development of magnetism. This still seems to me to be 

 the case, with a slight modification, which makes the limit depend also on the 

 length of the circuit ; for as we must consider the force at each pole to be the 

 resultant of those which I have described as M and M' ; and as each of these 

 is the resultant of two others transmitted by induction, it follows that even an 

 infinite -^ would not produce the total magnetization of the polar surfaces, as 

 long as the portions of the circuit through which this induction takes place 

 have a. finite magnitude. The hollow magnet which I first used, and which 

 with T^ = 4639'27 gave L — 76975, could not have passed 800 with any ^, to 

 judge from the curve of those variables. Had this magnet been solid, the limit 

 might be expected to be one-third higher, or — 1067 ; but in the solid 2-inch 

 magnet, with a circuit reduced to 16''2 instead of 36', it is seen from Table xn. 

 that this is exceeded by a i^ not a fourth of the above, and with a ■^ = 2548-59 

 I got on one occasion i = 1379'll. With the small iron magnet the maxi- 

 mum is about 300, its section being one-third, and its circuit about three-fourths 

 of those of the hollow one. These facts show that the limit is determined not 

 merely by the section of a magnet, but by the other causes which I have pointed 

 out as influencing the efficiency of the helices. More precise information can- 

 not be expected till we possess an exact theory of the electro-magnet, but I 

 may be permitted to give here an empirical formula which represents a consi- 

 derable portion of these experiments with sufficient precision. 



It was suggested by examining the march of the sets of — -. If L were 

 as T^, -y- would be as — r- ; but it is not so, and the second must be multiplied 



