88 DRS. J. A. HARKER AND P. CHAPPUIS ON A 



mixture of nitrates of potassium and sodium, heated externally by a double 

 circulation of hot gases, and stirred continuously by a system of rotating screws. 



The cast-iron vessel which forms the bath has a depth of 50 centims. and an 

 exterior quadrangular section of 20 centims. by 12 centims. ; the angles are slightly 

 rounded at the corners and on the bottom. The interior cross-section of the bath 

 approaches an ellipse, a form found by one of us especially favourable to thorough 

 stirring, when the shaft carrying the rotating screw-blades is placed at one of the foci. 



The casting is supported by four substantial feet, 6 centims. in length, on a massive 

 iron plate. This plate is pierced in the centre by a large circular hole, 8 centims. 

 diameter, connected to a suitable sheet-iron chimney, to take away the products of 

 combustion. Around the bath is fixed the first envelope of stout sheet-iron resting 

 on the base-plate, whose height is a little less than that of the bath. Over this is 

 placed a second envelope, open below, slightly pyramidal, and protected on the 

 exterior by several lay era of asbestos-card and wool. This rests on the upper edge of 

 the l>ath, and may easily be detached from the rest of the apparatus. 



A special rectangular burner, fitted with several gas taps, is placed round the inner 

 envelope, near the lower opening. The hot gases rise first in the space between the 

 two covers, then descend between the bath and the inner one, finally escaping by the 

 chimney. The top of the bath and the whole of the hot portions of the apparatus 

 which are exposed are prevented, as far as possible, from disturbing the temperature 

 of the room by covering them with thick layers of asbestos-wool. 



It would be dangerous to expose the thermometric reservoirs to the direct action of 

 the melted salts. We therefore fixed in the bath thin weldless steel tubes closed at 

 their lower ends, and projecting a few centimetres above the surface of the liquid. The 

 thermometers were introduced into these tubes, which they fitted almost exactly. 

 The tube containing the reservoir of the nitrogen thermometer was provided with a 

 brass lid closely surrounding the capillary tube, a few washers of asljestos completing 

 the joint. 



The bath is stirred by two sets of screw-blades fixed to a vertical steel shaft, which 

 extends to a height of about 40 centims. above the top. The upper end of this shaft 

 is suspended directly by a piece of rubber tube from the axis of a small electric motor 

 worked by four accumulators, 



The system of heating which we have just described allows a very satisfactory 

 constancy of temperature to be attained, but several hours are required in order to 

 obtain another equilibrium at a different temperature. To facilitate this, the bath 

 was heated continuously during the whole course of the experiments at high 

 temperatures. For the first set of comparisons, which were interrupted by an 

 accident, and which consisted of a small number of measurements, we employed the 

 reservoir of " verre dur " described previously. 



In the second, and more complete series, the porcelain reservoir was used 

 throughout. 



