IRON, STEEL, AND NICKEL TUBES IN THE MAGNETIC FIELD. 529 



fixed into the hole of the cap by means of marine glue. But this method proved in 

 many ways inconvenient. The plan finally adopted was to provide the cap with a 

 secondary cap or plug, through which the capillary tube also passed. This plug screwed 

 down upon a rubber washer which encircled the capillary tube and which, being 

 squeezed hard against the inner wall of the cap and the outer surface of the capillary, 

 made a water-tight junction between the capillary and the cap. 



The cap itself screwed down upon a rubber washer which rested on the ledge or 

 shoulder where the true bore of the tube began. In the earlier experiments the cap 

 was screwed in as tightly as possible so as to compress the washer to the uttermost. 

 But it finally appeared that, however much the washer was mechanically compressed, it 

 could still be compressed a little more when the cap was iron or steel fitted to an iron or 

 steel tube in a high magnetic field. For it was found that with the iron or steel cap 

 much greater diminution of volume was obtained than when the brass cap was used. 

 This was clearly due to the iron or steel cap being pulled down hard upon the washer, 

 because of the strong magnetic attraction between it and the tube. A similar effect, 

 though very much smaller, was produced with the nickel cap and tube. The result was, 

 of course, a diminution of volume superposed upon the true effect, and in certain cases 

 quite obliterating it. Unfortunately, this source of error was not detected until Steel 

 and Nickel Tubes No. II. had been bored out. The results for No. I. Steel with the 

 steel cap were utterly untrustworthy and are not here considered. The results for 

 No. I. Nickel have, however, been tabulated, because the false effect is certainly 

 comparatively small. 



The simplest way to prevent this mechanical displacement of the cap as a whole was 

 to introduce between the projecting flange of the cap and the top of the tube a brass 

 ring of suitable thickness. It was thin enough to allow the rubber washer to be 

 compressed so as to be water-tight, and thick enough to be in hard contact with both 

 cap and tube. 



A similar false effect was also produced when the plug which fixed the capillary to 

 the cap was made of iron, steel, or nickel. Hence, although originally each cap was 

 provided with a plug of the same material, only the brass plug was used in any of the 

 experiments to be here described. 



In every case the capillary, cap, washers, ring, and tube were put together under 

 water so as to prevent the retention of any air bubbles. When through haste or 

 carelessness an air bubble happened to be entrapped, its presence was at once detected 

 by the manner in which the change of volume occurred. Instead of a sharp, abrupt 

 motion of the meniscus in the capillary when the magnetic field was applied, there 

 was a slow, gradual motion with a corresponding gradual return on removal of the 

 field. When this happened, the whole apparatus was taken to pieces and carefully 

 put up again. 



Before each experiment the capillary tube was carefully cleaned, strong nitric acid and 

 caustic potash being in succession drawn through it. In the preliminary experiments, 



