Page:Proceedings of the Royal Society of London Vol 69.djvu/500

480 broad bands by absorption. I have never expressed the opinion, that the disappearance of the light between the D-lines in the absorp- tion spectrum of dense sodium vapour is only a result of the strong dispersion ; but I warned against always ascribing the observed dark bands to absorption only.

Wood's recent researches are very important as a contribution to our knowledge of dispersion in general. For the present their bearing on the spectral phenomena exhibited by the light from the chromo- sphere and from sun-spots, seems not to be so direct, because, most probably, the density of the vapours is much less in the solar atmo- sphere than in the dispersion tubes used by "Wood in his brilliant experiments.

In the following note I wish to draw attention to a connection which appears to exist between the electric conductivity of certain alloys of iron and the specific heats, and hence atomic masses, of the particular elements with which the iron is alloyed. In a paper published in the ' Transactions ' of the Royal Dublin Society the electric conductivity and magnetic permeability of a very large and, I believe, unique collection of alloys of iron is given. These alloys, it may be mentioned, have been prepared with great care by my friend, Mr. R. A. Hadfield, Managing Director of the Hecla Steel Works, Sheffield. Of the alloys made, 110 different specimens were found homogeneous, and could be forged and rolled these were analysed at the Hecla works, and submitted to similar heat treatment all being carefully annealed under the direction of Mr. Hadfield.

The specimens were in the form of rods, of nearly circular cross-section, about one-half a centimetre in diameter and 104 cms. long. The conductivities were found by the potential method, a standard of pure copper being employed. Although the determination of the mean sectional area of the specimens was made with great care, by numerous measurements of each rod with a micrometer screw, and also by means of water displacement, yet, owing to slight irregularities in the diameters of the rods, and the numerical importance of this value in