1893-94.] Prof. Knott and A. Shand on Magnetic Strains. 297 
1. In high fields and with high magnetisations there is decrease 
of volume in the interior of all three tubes ; hut, below a certain 
moderate field, the dilatation is positive. The field and magnetisa- 
tion at which this change of sign occurs are highest for the tube of 
thickest wall (III.) and lowest for the tube of thinnest wall (I.). 
2. With Tubes I. and II. the dilatation is, however, negative 
in still lower fields ; and the field and magnetisation at which this 
other change of sign occurs are higher for the tube of thicker 
wall (II.). Down to the lowest field in which a measurable change 
of volume for Tube III. was obtained, there was no evidence of 
a negative dilatation. But, as this lowest field was only 21 ’4 
(whereas it was possible to measure changes of volume in Tube I. 
in as low a field as 4*7), it is impossible to say anything definite 
regarding the behaviour of Tube III. in very low fields. 
3. The connection that is here indicated between the thickness 
of the wall and the field in which the dilatation changes sign, hints 
at a penetration of effect through the walls as the field increases in 
strength. 
4. It is curious to note that the dilatations in the tube of inter- 
mediate bore are, as a rule, numerically smaller than the correspond- 
ing dilatations in the tube of widest bore, and yet the tube of 
narrowest bore gives by far the greatest values for the same quantity. 
This is not what our experience with the iron tubes would have led 
us to expect. 
Prom experiments on different specimens of nickel strips and 
wires, Mr Bidwell * found for the elongations in field 600 the values 
-lOTxlO-’^ and - 240 x lO’h Assuming that our nickel tubes 
have similar elongations, we might calculate the elongations trans- 
verse to the lines of magnetisation. But such a calculation, though 
plausible enough in the case of the very thin- walled tube discussed 
in the earliest paper of 1891, is clearly quite out of the question here. 
In all probability the elongations of tolerably thick bars and tubes 
will differ materially from the elongations of wires and thin strips. 
* See his Paper “On the Changes Produced by Magnetisation, &c.,” Phil. 
Trans., vol. clxxix., 1888, A, pp. 205-230. 
