and Retractions of Rods of Iron and Steel. 191 



sending the current with its maximum strength, the rod elon- 

 gated 1-3 division of the scale; but if we now keep the circuit 

 closed but gradually diminish the strength of the current, we 

 observe that the scale-readings do not correspond to those given 

 when the corresponding strengths of current were reached by 

 going from their minimum to their maximum, as the following 

 Tables, giving the differences of scale-readings in the two cases, 

 clearly show. 



First Series of Experiments. 



Fraction of length of fine \ i 3 11 1 a 



interpolar wire. . . J ** 5 * *' w ' ' 



iSif: h -}s^ »" 55.8 55-9 55-95 56-1 



Differences ... -65 -7 -8 '7 -35 



Second Series of Experiments. 



Fraction of length of fine "I i s i i in 



interpolar wire . . J *" 2 * *' ¥ ' 



On , gradually diminishO 55 . 4 55-9 56-1 



ing the current . J 

 Ongraduallyincreasingl 



the current . . J 



Differences . . . -45 -6 -8 -8 "^5 ~0~ 



We thus see that the rod tends to persist in the elongation 

 it acquired on first passing the maximum current ; for it does 

 not retract in proportion to the diminished strength of this cur- 

 rent; and the experiments show that even when the current is 

 so far diminished in strength that it would, if suddenly thrown 

 through the helix, be unable to elongate the rod sufficiently to 

 be measurable, yet this feeble current holds the rod elongated 

 •45 of a division in the second series of experiments; but on 

 breaking the circuit the rod instantly retracts *45 of a division 

 in the second series of experiments and *65 of a division in the 

 first series, and regains the length it had before the current was 

 passed round it. 



On passing the current with the whole of the fine wire in the 

 circuit, we have in the first series of experiments an elongation 

 of '05 of a division ; but on making the circuit without the in- 

 terposed fine wire we have an elongation of 1*3 division; and if 

 we now do not break the circuit, but gradually diminish its 

 strength by increasing the interpolar resistance, we find that, 

 when the whole of the fine wire is again in the circuit, the 



