IRON, STEEL, AND NICKEL TUBES IN THE MAGNETIC FIELD. 535 



Consider, first, the experiments with the iron cap. 



It will be seen at a glance that Nos. I. and VII. — the narrowest and widest bores — 

 have quite opposite behaviours in a series of ascending magnetic fields. The internal 

 space in No. I. begins by increasing in low fields, recovers its original volume in field 

 100, and thereafter continues decreasing as the field is raised. In No. VII. , on the con- 

 trary, there is decrease in the lower fields and increase in the higher. Nevertheless, a 

 careful study of the forms of all the graphs reveals that, as we pass from I. to VII. through 

 the intermediate conditions of bore, there is a gradual transition in form from the first 

 to the last. Any contiguous two, in fact, behave in similar fashion, the corresponding 

 graphs presenting similar features, and evidently being of the same type. 



Thus No. II. is very like No. I. in the form of its graph, the main difference being a 

 smaller diminution of volume in the high fields. A minimum point, followed later by a 

 maximum, occurs in No. II. soon after the change of sign. To this in No. I. there corre- 

 sponds a hump, which just falls short of being a true minimum point. 



Passing to No. III., we find still smaller negative change and very much more dis- 

 tinct minimum and maximum points than in the case of No. II. After the change of 

 sign the curve still lies wholly below the axis. 



But with No. IV. a new feature presents itself — namely, a second change of sign 

 about Field 160. Thus there is increase to Field 75, decrease to Field 160, and increase 

 ever after, up at all events to the highest fields used. But after the second change of 

 sign a distinct maximum exists in Field 300, so that it is possible that a third change of 

 sign may occur in still higher fields. The curve preserves, however, the same general 

 form as that of No. III., the second change of sign being, so to speak, the result of the 

 upward displacement of the curve as a whole. 



No. V. presents the same general features as No. IV, the second change of sign 

 occurring in a much lower field. 



With No. VI. the initial positive change in low fields has vanished quite away, 

 and with it, of course, the first change of sign. The point at which the change of sign 

 from decrease to increase occurs shifts to still lower fields. This backward drift of the 

 field of zero change continues as we pass to No. VII. Neglecting the initial stages in 

 low fields, we see at a glance that Nos. V, VI., VII. are all of the same type of curve. 



Broadly stated, the negative change of volume which is characteristic of the 

 narrowest-bored condition gradually diminishes as the bore is increased, until it is 

 largely replaced by positive change of volume. As this process goes on, the field at 

 which the positive change succeeds the negative change diminishes in value. 



When the brass-cap is substituted for the iron-cap, the general character of the 

 phenomena is very much the same. The tendency is for the " brass-cap " curves to be 

 higher than the corresponding " iron-cap " curves. For example, the changes of volume in 

 field 500 for the brass-capped tubes are (with a single exception) all greater positive 

 numbers, or smaller negative numbers, than the corresponding changes for the iron-capped 

 tubes. One particularly interesting case is No. III., the curve of which just comes 



VOL. XXXVIII. PART III. (NO. 13). 4 C 



