538 PROFESSOR KNOTT ON THE STRAINS PRODUCED IN 



higher fields, is very like No. V. But here again the transition from No. V. to No. VI. 

 differs essentially from the transition from No. IV. to No. V. It is curious to note that, 

 in spite of all this irregularity, the successive tubes from II. to VI. come in order in the 

 very highest fields, No. II. being lowest and No. VI. highest. 



But the most extraordinary change is the transition from No. VI. to No. VII, a 

 transition which it is hopeless to bring into line with the other transitions. No. VII. 

 is, indeed, much more like No. II. than any of the others. If it were drawn to one- 

 eighth or one-ninth of its true scale, it would differ very slightly from No. II. But 

 if (§ 5) we compare the volumes of the bores in the two cases, we shall see that the 

 dilatations or changes of unit volume are very much the same ; for the volume of 

 No. VII. is nearly eight times that of No. II. This result, however, can hardly be 

 regarded as other than accidental (see Appendix). Very remarkable in the case of 

 No. VII. are the comparatively great increase of volume in lower fields and the very 

 great diminution of volume in the higher fields. 



The broad characteristics of Plate III. are reproduced in Plate IV., which shows the 

 results when the brass cap is substituted for the steel cap. As in the case of the iron 

 tubes, so here. There is a tendency, especially in the higher fields, for the brass-cap 

 curves to lie higher than the steel-cap curves. In other words, there is a slightly 

 greater tendency to diminution of volume when the tubes are steel-capped than when 

 they are brass- capped. The explanations already suggested in § 8 will apply equally 

 well here. 



The greatest deviations in form between the two sets of curves occur in the lower 

 fields. In Plate IV. the curves are smoother in their march than in Plate III. Com- 

 pare, in particular, the two III.'s and the two V.'s. Had it not been for the curious 

 undulation in iron-capped No. V., the much less marked undulation in the correspond- 

 ing brass-capped tube's curve might have been ascribed to some flaw in the experiment. 

 It will be noticed that Nos. III. and IV., brass-capped, have diminished volumes in the 

 lowest fields, whereas the iron-capped III. and IV. begin with increase of volume. In 

 our ignorance of the character of the strain produced in the tubes, it is quite impossible 

 to suggest any explanation of these differences in detail. 



The complex and inexplicable relations which hold among the successive steel tubes 

 are the more extraordinary because of the comparative simplicity in the case of the iron 

 tubes. In their purely magnetic qualities the iron and steel tubes are very similar, the 

 steel having slightly higher permeability, but neither having any appreciable residual 

 magnetism. Yet they are markedly different in their behaviour as regards strain in the 

 magnetic field. Possibly the iron may be more homogeneous than the mild Bessemer 

 steel ; but, however this may be, all the facts serve to show that we are dealing with a 

 very complex set of phenomena. 



§ 10. The Nickel Tubes Nos. I. to VII. — The results for the nickel tubes with brass 

 cap are shown graphically in Plate V. To prevent folding of the plate, and yet to keep 

 the scale the same as in the other plates, the nickel curves are represented only to field 



