76 PRINCIPAL FORBES ON AN EXPERIMENTAL INQUIRY INTO 



the same temperature.* The gas flame and the violently heated currents of air 

 thence arising were prevented from playing against any part of the bar by a piece 

 of metal plate fastened by wire to the crucible against the face a b (but to prevent 

 confusion not shown in the figure), while the whole conduction bar was farther 

 protected from heated currents, and from radiation from the crucible and gas 

 chimney, by three polished metal screens d, e,f, placed parallel to one another, two 

 to the left and one to the right of the wooden support E. The square apertures in 

 the screens were 0-25 inch wider than the dimensions of the bar, and they were 

 supported in such a manner as not usually to touch it in any part. These screens 

 very effectually defend the thermometers, as well as the bar, from extraneous heat. 

 The first thermometer only — that at 3 inches distance from the zero line b c — is 

 seen at g in the section, fig. 1. 



47. The conduction-bars have already been described in Arts. 17, 18. The 

 more perfect one was 1-25 inch square, and fully 8 feet long, reckoning from the 

 zero line above mentioned. It was used in two states, first, with a naked polished 

 surface, and, secondly, when coated with thin paper. The other bar, also of 

 wrought-iron, was 1 inch in the side and 7 feet long. In the present paper I shall 

 discuss separately the results obtained with these two bars, presenting three quite 

 independent cases, but which, as they ought to lead to an identical value of the 

 conductivity of iron (assuming the quality of the bars to be alike), put the method 

 here proved to a severe trial. 



48. Throughout the remainder of this paper, when I speak of Case I., I mean 

 the 1^-inch bar with moderately polished surface ; Case II. is the same bar with 

 paper surface ; Case III. is the 1-inch bar with naked, but less brightly polished 

 surface. 



49. The thermometers were inserted in holes in the bar 028 inch diameter 

 and about £ inch deep. They were surrounded by mercury or (in the hotter holes) 

 by fusible metal. (See Art. 20, and also Plate II. fig. 1.) Nine or ten thermo- 

 meters were usually employed, and in the case of the principal bar (Cases I. and II.) 

 they were usually (though not always) spaced as follows, reckoning from the 

 zero line a b on the face of the crucible : — 



0-25, 0-5, 0-75, 1, 1-5, 2 or 2-5, 4, 6, 8 feet. 



50. The method of using a single standard thermometer for obtaining final re- 

 sults by the method of stepping, with its advantages, have been fully explained 

 at Art. 22. 



51. " The free temperature, or that to be deducted from the readings of the 

 thermometers, in order to get the true excess of statical temperature along the bar, 

 was obtained by inserting a well-compared thermometer into a hole containing 

 mercury, drilled in a similar but short bar of iron, supported in the free air of the 



* See, however, note to Art. 70, below. 



