92 Mr. T. G. Bedford on tJw Expansion of 



copper leads. The head also carried a second pair of terminals 

 to which was attached a thick platinum wire running to the 

 upper mark and back to act as a " compensator.''' 



Connexion was made between the pyrometer-head and the 

 resistance-box by means of a carefully equalized and insulated 

 fourfold cable. The lower end of the porcelain tube was 

 closed by a porcelain ping. 



The resistance measurements were made by the usual 

 Wheatstone-bridge method, the arms of the bridge being 

 equal. The resistance-box used contains a set of 7 coils of 

 nominal values 10 to 640 mean box-units, and in series with 

 the coils a bridge- wire 40 cm. long, the effective resistance of 

 1 cm. of which is equal approximately to 1 box-unit. By 

 means of a vernier '01 cm. can be read. The mean box-unit 

 is roughly *01 ohm *. A careful standardization of both coils 

 and bridge-wire was made in January j 898, by Mr. E. H. 

 Griffiths and Mr. C. F. Green ; and again in May 1899, 

 since the completion of these experiments, the coils have been 

 restandardized by Mr. C. F. Green, and show no perceptible 

 change since the previous standardization. The resistance of 

 the pyrometer-wire was about 260 box-units at 0° C. The 

 current from a dry cell through 20 ohms was employed, and 

 was found to produce no serious heating-effect. Thermo- 

 electric effects were eliminated by the use of Griffiths' 

 Thermo-electric key (Phil. Trans. 1893 A, p. 398). A dead- 

 beat mirror galvanometer was used. 



The reading-microscopes, for the use of which the writer 

 is indebted to Prof. Ewing, were made by the Cambridge 

 Scientific Instrument Company. They were supported on 

 solid stone blocks standing on a firm stone bench. The 

 microscope-carriage moves along its slides by means of a screw 

 of *5 mm. pitch, the head being divided into 100 parts and 

 thus reading to *005 mm. Tenths of these divisions could be 

 estimated, but the diamond marks on the porcelain were not 

 sufficiently good to enable settings of the cross- wire to be 

 made to such accuracy. The gas furnace in which the tube 

 was heated w 7 as supported on a stand provided with levelling 

 screws, by adjusting which the marks were kept in focus as 

 the tube was heated or cooled. Both the microscopes and 

 bench were screened by bright tinned plate from the radia- 

 tion from the furnace. As a standard length the distance 

 between two diamond marks on a glass tube kept in a trough 

 of melling ice was used. This distance was equal to 91*394 cm. 

 During each experiment the readings for the glass tube were 

 taken at intervals. 



* There is a full description of a box of this type, known as the 

 " Student's Box," in the ' Electrical Review/ Aug. 18, 1899. 





