396 Professor Dewar [Jan. 19, 



liquid air sponge to another portion of the surface would cause no 

 visible deposit. This is exactly what takes place. If, however, two 

 spheres, one much larger than the other, are joined together by means 

 of a tube about 2 mill, in diameter and 50 mill, long, the whole 

 space being a Torricellian vacuum (with some excess of mercury) 

 then on decanting, the mercury may be transferred to the smaller 

 sphere, as is represented in Fig. 3. Now if an air sponge is applied 

 to a portion of the surface of the larger sphere, a mercury mirror 

 instantly deposits, but on applying a new air sponge to another 

 portion of the surface, no further mercury mirror is formed. The 

 narrow glass tube prevents the excess of liquid mercury in the 

 small bulb supplying vapour rapidly to the larger one, so that the 

 local cooling to — 180° C. of a portion of the surface has practically 

 condensed all the mercury in the larger space, although the small 

 one is still filled with saturated vapour and a free communication 

 exists between them. If while in this condition the small bulb is 

 inclined so as to allow a drop of liquid mercury to fall into the lower 

 side of the large bulb, which has not been cooled, instant deposition 

 of mercury takes place on the liquid air cooled portion of the upper 

 surface. Under very small pressure of vapour, therefore, equalisation 

 of pressure of two bulbs communicating by a narrow tube is a very 

 slow process. There are cases, however, in which the application of 

 a sponge of liquid air to the surface of a vessel causes no visible 

 deposit, and yet the inference is that something has been condensed. 

 The best arrangement to show this effect is to select highly ex- 

 hausted vacuum tubes containing phosphorescent materials like 

 alumina and other minerals, and to arrange the induction coil 

 spark gap of a little greater resistance than the vacuum tube. On 

 starting the coil the current passes solely by the vacuum tube, but 

 immediately the liquid air sponge cools a portion of the surface of 

 the bulb, the discharge shifts to the air gap. During the cooling the 

 phosphorescence of the glass tube is greatly increased, but finally the 

 resistance may become so great that all discharge in the vacuum tube 

 ceases. Some old tubes belonging to the late Dr. de la Eue have 

 given visible deposits near the electrodes, and in many the diameter 

 and distribution of the strias are materially changed during the local 

 cooling to — 190° C. When large vessels containing nothing but 

 mercury or iodine vapour as a residuum of the vacuum space are 

 rubbed with a cotton wool sponge of liquid air in a dark room, 

 luminous glows filling the vessel take place occasionally, or bright 

 flashes of light which enable the shape of the vessel to be seen. The 

 ordinary mercury vacuum vessels show the same phenomena, which is 

 doubtless due to electric discharges caused by friction and cooling. 



The optical properties of bodies cooled to the temperature of 

 boiling liquid air will require long and patient investigation. An 

 interesting fact easily observed is the marked change in colour of 

 various bodies. Thus, for instance, oxide, sulphide, iodide of mercury, 

 bichromate of potash, all become yellow or orange ; while nitrate of 

 uranium and the double chloride of platinum and ammonium become 



