24 Prof. Magnus on the Diathermancy 



of its position. For if the telescope, with the scale fixed below 

 it, is placed at 2 metres distance from the mirror, every altera- 

 tion of one millimetre in the position of the image corresponds 

 to a change of 51 seconds of arc in the position of the mirror; 

 consequently the daily variations in the intensity of the earth's 

 magnetism, amounting as they do to several minutes, cause a very 

 perceptible variation in the direction of the mirror. No harm, 

 however, can arise from this, so long as no observations are taken 

 as belonging to the same series except such as are made in close 

 sequence, and if it is also borne in mind that the observed 

 deflections have not an absolute, but only a relative value. 

 Nevertheless they are always proportional to the intensities of 

 the current, even when they amount to one or two hundred 

 millimetres. This proportionality, together with the certainty 

 ensured by reading off through a telescope, make this instru- 

 ment specially fitted for thermo-electrical experiments. 



A galvanometer of this construction was used in repeating 

 Prof. Tyndall's experiments. As sources of heat I used two 

 blackened vessels filled with boiling water, which, in order to 

 avoid the disturbing effect of flames, was kept boiling by means 

 of steam. In order to hinder the accidental cooling of the 

 vessels, each of them was surrounded by a pasteboard box, in 

 which the only opening, besides that through which the rays of 

 heat escaped, was a small one in the cover. The screen of the 

 compensating source was placed inside its box. The boiler for 

 producing the steam was outside the boxes. The rays from both 

 sources passed through tubes, open at both ends, to the thermo- 

 pile furnished on both sides with its conical reflectors. 



The experimental tube was 0*66 metre long, and had an opening 

 in the side near each end. One of these openings was connected 

 with an air-pump ; and through the other, air was pressed into the 

 tube by means of a pair of bellows. As in Prof. TyndalPs expe- 

 riments, this air could be made to pass through several tubes 

 containing chloride of calcium and broken glass moistened with 

 sulphuric acid, or through tubes containing broken glass moist- 

 ened with water, before it entered the experimental tube, and 

 could in this way be used either dry or saturated with moisture. 

 With this arrangement I got, on allowing dry or moist air to 

 flow through the tube, deflections of the galvanometer which 

 corresponded to those described by Prof. Tyndall. But I did not 

 always get them ; and what particularly surprised me was, that 

 the deflection of the needle did not correspond to an absorption 

 of heat by its passage through moist air, but that, on the con- 

 trary, when moist air was passed through the tube, the face of 

 the pile which was turned towards the tube was found to be 

 most heated. In order to clear up the already mentioned uncer- 



