1913-14.] Resistance of Iron in Crossed Magnetic Fields. 263 
The various values are tabulated in the foregoing tables. The bracketed 
o o 
numbers in the first column on the left give the values of the longitudinal 
fields ; and the bracketed numbers in the horizontal rows give the values 
of the transverse fields. The remaining numbers are the changes of resist- 
ance produced by the field or combination of fields indicated by the symbol 
in the second column. 
In these experiments there is no evidence of what others have observed, 
namely, an increase of resistance in low and moderate transverse fields. 
For example, Grunmach (2), in three out of the four recorded experiments 
with iron, obtained increase of resistance up to fields of 7000 or 8000 Gauss, 
after which the change became a decrease rapidly increasing in value as 
the field was taken stronger. In like manner, he obtained with nickel an 
increase of resistance up to field 700, and thereafter decrease as the trans- 
verse field was made stronger. 
I have always been very doubtful of the reality of this initial increase 
of resistance ; and a recent paper by Messrs W. Morris Jones and 
J. E. Malam (3) seems to me to establish the fact that when nickel is ac- 
curately placed in the transverse field the change of resistance is always a 
decrease. In my own earlier experiments with nickel spirals in transverse 
fields I was never satisfied that I had the spiral absolutely perpendicular 
to the field until I had got rid of this apparent initial increase in low fields. 
When very thin wires are used, the difficulty of eliminating all chance of 
a resolved longitudinal effect becomes greatly increased. For the change 
of resistance depends undoubtedly upon the magnetization within the 
metal. In very thin wires the transverse magnetization cannot be very 
much greater than the transverse magnetizing force, whereas in the 
early stages the longitudinal magnetization is much greater than the 
longitudinal magnetizing force. A little consideration will show that 
a comparatively small resolved component of the magnetizing force along 
some part of the wire may easily be accompanied by a longitudinal 
magnetization large enough to produce a resistance change of positive 
sign able to overbalance the very small decrease due to the transverse 
magnetization. 
All this danger of having present an uneliminated longitudinal com- 
ponent is obviated in the experiments now described by the use of ribbons 
instead of wires of iron and nickel. For in the first place it is a compara- 
tively simple matter to set the coiled strip or ribbon with its width 
accurately along the lines of force ; and in the second place, even if the ad- 
justment were not quite accurate, the magnetization along the width of the 
metal would be considerable, so that any possible resolved longitudinal 
