High Temperatures, 173 



nicated may undergo in the further investigation of the prob- 

 lem cannot sensibly affect the matter in hand. 



The method may be described briefly as follows : The bar is 

 laid for the measurement of its expansion symmetrically in an 

 electrically heated fire-clay tube, 80 cm long, in which two lateral 

 holes are bored 10 mm in diameter and 485 mm apart. 



This tube is protected against loss of heat by radiation by 

 two concentric fire-clay jackets with corresponding lateral open- 

 ings, and the whole placed in a horizontal position upon two 

 brick pillars in such a way that the openings are directed 

 downward. Underneath these, two microscopes are mounted 

 with the aid of which readings can be made of the position of 

 certain marks cut upon a flat polished surface near the ends of 

 the expanding bar. 



The heating coil which is wound upon the innermost tube 

 extends for a good distance beyond the ends of the bar so that 

 the temperature opposite the opening differs in maximum only 

 10 per cent from that at the middle. On account of the 

 increase in the coefficient of expansion with the temperature 

 an average temperature for the bar is not sufficient if consider- 

 able local differences exist, hence it is eminently desirable that 

 the differences of temperature along the bar be not too great. 

 We are accordingly making an effort to improve upon our 

 present coil in this direction in the hope that an even greater 

 uniformity in the temperature distribution may be secured. 



The microscopes were mounted upon sliding bases upon a 

 long horizontal support and could be fixed in the proper posi- 

 tions with clamping screws. To protect them from the radia- 

 tion of the oven a diaphragm through which water circulated 

 constantly was introduced between. 



With this apparatus the expansion was then measured with 

 the eyepiece micrometers of the microscopes, the reading 

 before heating being verified after the oven had cooled down 

 again. No measurable displacement of the microscopes was 

 observed throughout the series but displacements of the bar 

 occurred now and then at the higher temperatures; symmetri- 

 cal observations upon both ends, however, eliminate any error 

 from this cause in the results. 



The observations were usually made at 250°, 500°, 750° and 

 1000°, and for the same temperature upon different days agree 

 to within ± - 01 mm , whereby it should be remarked that one 

 whole turn of the micrometer screw corresponds to about 



1 mm 



The divisions, which were situated in the axis of the bar to 

 diminish the effect of a possible slight bending or warping, 

 were cut with the dividing engine, rather deep but with sharp 

 edges, and rubbed with Parisian red (iron oxide). They were 



