( 224 ) 



experimental tube, tbe latter being immersed in liquid air; al last 

 the experimental tube, still immersed in liquid air, is connected with 

 the mercury airpump, and fused off. By weighing it is ascertained 

 that the desired quantity has been transferred to the experimental tube. 



5. A mode of heating which does not give rise to convection 

 currents inside the tube, is considered of great importance also by 

 Tkichnkr and Traube, to prevent mixing of what is at the upper 

 end and what is at the lower end of the tube. Tkichner does not 

 accept Young's refutation of his experiments (see Teichner, loc. cit.) 

 because in Young's apparatus convection currents are not so well 

 prevented as in that of himself. The thermostat used by us, however, 

 satisfies much higher demands than that of Teichner. The tube is 

 immersed in a liquid bath in a double walled non-silvered vacuum- 

 glass which is hermetically closed with a badly conducting lid. The 

 glass is provided with a valve-stirrer and an arrangement to heat 

 the bath electrically from above, and is itself again immersed in a 

 bath with double glass walls, which is likewise provided with a 

 valve-stirrer. With the exception of two windows, the space between 

 these two walls is tilled with cotton wool, by which the outer wall 

 too is surrounded. Like the bath in which the improved apparatus of 

 de Heen (§ 2j was placed, the outer bath was kept at constant 

 temperature (up to 0°,02) by flowing water with the aid of a xylene 

 thermoregulator (§ 2). In this way the temperature of the bath in 

 which the experimentaltube is, can be kept constant up to : ',002. The 

 heating takes place according to the indication of thermometers divided 

 into 50 th of a degree, which are placed in the inner and in the outer 

 bath, and is regulated in such a way that everything (hat might give 

 rise to convection currents is as much as possible avoided. 



6. When manipulated the tube was always efficiently shielded 

 for the protection of the observer. 



§ 7. Observations. The bulbs correspond with the densities 0.365, 

 0.390, 0.405, 0.421, 0.443, 0.450, 0.466, 0.483, and 0.510. 



The position of the bulbs and that of the meniscus (indicated in 



what follows by the read number of the mark of division between 



]) was read on a millimeter scale etched on the tube; in the middle 



of the tube the mark 30 is found, the zero point is 20 mm. above 



the bottom. 



In the first experiment the heating took place very slowly; at 

 first also the inner bath was (electrically) . heated, t refers to the 

 outer bath, t, to the inner one. After three hours small gas bubbles 

 were seen to rise from the downmosl iherino-elenienl, probably 



