214 Prof. Regnault on the Determination of 



neither observation nor numerical results to register. I think 

 the method I am about briefly to describe satisfies all these 

 conditions. 



Fig. 4, Plate I. represents an air thermometer, consisting of a 

 thin brass or silver tube a b, 20 centimetres long and 15 milli- 

 metre's in diameter. It is closed at a, and terminates at b in a 

 metallic capillary tube b c } provided with a stopcock r. Above 

 this stopcock is fitted, by means of a screw-press (fig. 5), such 

 as I often use in my apparatus, a short tube de } filled with 

 fragments of chloride of calcium. At the top of the tube de 

 there is a small aperture, so that the tube ab communicates 

 freely with the external air when the stopcock r is open; but 

 only perfectly dry air can penetrate. 



Let us suppose twenty such tubes fixed on a small board, so 

 that the reservoirs ab project beyond the board, and are in 

 contact with the surrounding air ; the capillary tubes being alone 

 fitted on the board, and the stopcocks at the command of the 

 aeronaut. These conditions are easily satisfied when the balloon 

 is raised above terrestrial obstacles. The reservoirs of the air 

 thermometers will not be influenced by the vicinity of the aero- 

 naut ; besides they can be removed from the boat as far as is 

 desired, by making the capillary tubes sufficiently long 



The w r hole of these twenty tubes, which are all numbered, 

 will form the thermometric system. 



To avoid accidents in the descent, or in the carriage, it is 

 better to arrange the tubes cylindrically as shown in fig. 6, the 

 capillary tubes b c being simply fixed round a central axis, and 

 the reservoirs ab remaining in free air. The system is then 

 placed, for the sake of carriage, in a metal case, and is only fixed 

 to the boat after the start. 



The barometric system consists of twenty tubes like those of 

 the thermometric system ; but their reservoirs are immersed in 

 melting ice during the whole balloon ascent. For this purpose 

 a brass or sheet-iron bath, AB C D, fig. 7, is used in which the 

 twenty reservoirs are arranged in parallel order in two rows, or 

 round an axis as in fig. 5. This bath is filled with melting ice 

 before the start. The water from the melted ice flows out by a 

 large tubulure, / v. But it is necessary to preserve the ice from 

 too rapidly melting at high temperatures, and especially from 

 sinking below 0° in the low temperatures which prevail in the 

 higher regions of the atmosphere. For that purpose the ice- 

 bath A B C D is placed in an annular bath, E F G H, filled 

 with warm water at starting. The interval between the two 

 baths is filled with a bad conductor of heat, such as cotton or 

 charcoal-ashes. Under these conditions the balloon may travel 

 several hours in regions where the temperature is —25° or —30° 



