462 



Mr. T. Andrews. 



[June 16, 



accuracy, the results recorded being the average of many observa- 

 tions. In some instances somewhat higher results were noticed. 



It is almost' impossible to obtain two pieces of iron (even when 

 forming adjacent parts of one polished rod) which when in solution 

 are devoid of some slight galvanic action between themselves ; but 

 the greatest care was exercised in the special preparation of the iron 

 used, so that this variation might be reduced to a minimum. 



To ensure success in the experiments it was found essential that the 

 iron bars should possess an excellent polished surface, free from mag- 

 netic or other oxide or impurities ; the solutions were also concen- 

 trated, and both discrimination and manipulative skill were requisite 

 in obtaining the practical galvanic equilibrium of the bars at com- 

 mencement. The time needed to ensure this seemed to vary consider- 

 ably with different solutions according to circumstances. A sensitive 

 galvanometer was also a requisite of success in these observations, 

 and telescopic readings were necessary, as in some cases the effects 

 were small. 



It seems desirable here to add a few remarks on the possible 

 influence of temperature on the reactions, and to state the means 

 used in the endeavour to minimise errors from this source. In con- 

 ducting the experiments, I should have preferred using greater 

 battery power, but employed only one bichromate cell ; the wire of 

 the coil was also of considerable thickness to prevent undue heating* 

 from resistance. The centre of the wooden coil bobbin was also about 



inch in thickness, so as to act as a central non-conductor. More- 

 over, an air space was allowed of J inch between the wooden centre 

 of the coil and the enclosed limb of the U-tube. The other limb for 

 the unmagnetised bar was enclosed by another coil, which, when not 

 in use, acted as an external protective jacket. Notwithstanding these 

 precautions, there was a slight increase of temperature in the interior 

 of the coil C. Thermometers inserted in test solutions, one in each 

 limb, gave an average difference of about 1° Fahrenheit at the end of 

 an hour, this increase of temperature in the solution in the coil tube 

 being, however, very gradual. It would be untenable to state that 

 this difference of temperature, arising from the action of the coil, did 

 not to some slight extent influence the results of the experiments 

 with apparatus fig. 1 ; but the results obtained therewith could 

 certainly not be regarded as due only to differences of temperature 

 conditions between the two tubes. Most of the experiments with 

 that apparatus afford within themselves evident proof to the con- 

 trary ; thus, it will be seen, that the magnetised bar assumed a pro- 

 minent positive position almost immediately after magnetisation, in 

 the case of nitric acid, before any perceptible difference of tempe- 

 rature could obtain between the respective tubes (see Table A, 

 columns 6 and 7). 



