502 Proceedings of Royal Society of Edinburgh. [sess. 
show that when the nickel is in the form of wire wound on a bobbin 
whose axis lies along the direction of the field, the sign of the 
change in transverse fields at 13° C. depends on the strength of the 
field. The results are given in the tables subjoined. In the first 
table four distinct experiments are given, with three different 
values of the current used in measuring the resistance. The values 
of this current (i) are entered at the top of each column. They are 
the currents in the nickel wire itself. It seemed necessary to 
make a direct inquiry as to a possible disturbance arising from 
the circular magnetization due to this current. The results are 
practically the same for all three values of current. 
The mean of these four experiments was then considered. A 
curve in terms of the current used to magnetize the electromagnet 
was drawn with great care ; and having already calibrated the field 
of the air-gap of the electromagnet by means of a flat bismuth coil 
placed successively at convenient positions along the air-gap, we 
were able to find the change of resistance for any assigned field 
within the range of the experiments. The results as finally reduced 
are given in the second table, and the representative graph is shown 
on the plate. It is curious to note that the curve when plotted in 
terms of the magnetizing current deviates distinctly in the higher 
fields from a straight line ; but, as will be apparent at a glance, it is 
very approximately rectilinear above field 1500 when plotted in 
terms of the field. Thus we find that in fields below 1300 the re- 
sistance of nickel wire magnetized transversely is increased, reaching 
a maximum in field 900 ; above 1300 the sign of the change alters, 
and thereafter up to the highest fields attainable the resistance is 
decreased and goes on decreasing at a fairly constant rate as the 
field is increased. 
It is possible that the initial increase may be due to the existence 
of the longitudinal effect along with the transverse effect proper. 
In coiling a wire in the manner described, there must he a certain 
pitch in the helix, so that there will be a component of field along 
the wire. The longitudinal effect being so much more easily 
produced than the transverse effect, might under certain conditions 
and in low fields mask the latter entirely. To test this point two 
coils were wound with the same kind of wire and with every- 
thing as nearly as possible the same except that the pitch in the 
