302 PROCEEDINGS OF SECTION A, 



12.— ON MEASUREMENT OF DOUBLE STARS. 

 By H. C. RUSSELL, C.3I.G., B.A., F.R.S. 



13.— THERMO-ELECTRIC DIAGRAMS FOR SOME 

 PURE METALS. 



By W. HUEY STEELE, M.A. 

 Plate IX. 



The paper by Professor P. G. Tait in the Tx-ansactions of the 

 Society of Edinburgh, 1878, " A First .Approximation to a Thermo- 

 electric Diagram," has hitherto been admitted to contain the best 

 work done on the thermo-electric diagram. This paper, however, 

 was stated by the author to be merely a preliminary to more 

 accurate results that were to follow, but which have not yet 

 appeared. The paper is mostly taken up with the discvission of 

 the peculiarities in the iron line, and it is not stated whether the 

 metals used were pure or otherwise, what the limits of accuracy 

 of observation were, nor what methods of observation were used. 

 Being in possession of a piece of thallium, whose line was not 

 determined by Professor Tait, I determined its position relatively 

 to silver, in order to determine its position on the diagram, and 

 then, finding its line cutting that of copper at about 70° C. 

 according to Professor Tait's results, I measured its position 

 relatively to copper, and found that the two results were utterly 

 inconsistent, according to Professor Tait's results. On measuring 

 the relative positions of copper and silver, using fairly pure 

 specimens, copper was found to be \erj close to and above silver, 

 while Professor Tait puts it a considerable distance below. I 

 therefore proceeded to construct the diagram afresh for as many 

 pure metals as I could obtain. 



The first essential in accurate thermo-electric measurement is a 

 .sensitive galvanometer, the one indicating the least current not 

 necessarily being the best for the purpose, but the one that 

 indicates the least current in proportion to its resistance, or, in 

 other words, the one that will indicate the lowest e.m.f. applied 

 to its terminals. Out of half a dozen types of sensitive 

 galvanometers I found the best for my purpose an astatic 

 instrument Avith one coil and a resistance of about half an ohm. 

 To magnetise the needles as strongly as possible I made a coil of 

 a very large number of turns of fine wire, and, putting the needles 

 into it, flashed as great a current through it as the wire would carry ; 

 the astatic pair was then suspended and the stronger needle 



