ELECTRICAL CONDUCTIVITY OF CERTAIN SALINE SOLUTIONS. 61 
bring the method to its full efficiency. During that time the results of our 
experiments were untrustworthy and inconsistent. But after every difficulty 
had been overcome, they were obtained in a perfectly satisfactory manner. 
We think we need have no hesitation in saying, that the above method of 
measuring the resistance of liquids meets all the difficulties of the problem, 
and, though somewhat laborious, is capable, if proper care be taken, of giving 
results on which the utmost reliance may be placed. 
We found that the effect of temperature was so marked as to make it abso- 
lutely necessary to ascertain the resistance of all the solutions at the same 
temperature. We adopted 10° centigrade as our standard, and all our observa- 
tions, both of density and resistance, were made at this temperature. 
The salt we first examined was normal sulphate of zinc—ZnSO, + 7H,0. 
Before use the salt was freed from all impurities, of which at first it was 
very full. Nineteen solutions in all were preparéd, most of them by dis- 
solving a known weight of salt in a known weight of distilled water, from 1 in 
40 parts, and so on, down to a saturated solution. The resistance of the weak 
solutions was found to be very great. The rate of diminution was at first rapid, 
but gradually fell off as the density increased, till it became very slow as the 
point of minimum resistance was approached. After that the resistance slowly 
rose again up to the point of saturation. The following table gives our numerical 
results with reference to this salt. Column III. contains the resistance in B. A. 
units of the liquid as measured in the tube. Column IV. contains the specific 
resistance in B. A. units. By “specific” resistance is meant the resistance to 
conduction between a pair of opposite faces of a cubic centimeter of the sub- 
stance. This quantity varies with the size of the cube adopted (being directly 
proportional to the length of the edge, and inversely proportional to its square), 
and has therefore no claim to be called specific ; but the term is now in general 
use, and is convenient for purposes of calculation. In order to reduce the 
results of column III. to column IV., we had to make an accurate deter- 
mination of the length of the tube and also of its cross-section, the latter being 
done by finding the weight of mercury it could hold. The ends were assumed 
to be parts of cones, and their resistance found by integration. The coefficient 
by which the figures in column III. are multiplied to give those in column I[V., 
is for this tube ‘022301. 
