576 



NATURE 



[April 17, 1902 



case the ratio of 454 to 294, or, roughly, of 3 to 2. Thus, with 

 these assumptions, the density in the wave-front is half as great 

 again as the density of the undisturbed air. The wave resembles 

 that caused by an explosion rather than that due to an ordinary 

 sound. 



Several attempts were made to see the train of waves due to a 

 musical note, but on consideration of the facts just stated it is 

 not surprising that they were all unsuccessful. The notes due to 

 a shrill whistle and to a siren blown by a pair of foot bellows 

 were both ineffective. 



An attempt was also made to obtain a train of waves from the 

 oscillatory discharge of a condenser through a circuit possessing 



■self-induction. A large Leyden jar and a coil of wire were 

 inserted between the main terminals in Fig. 2. The frequency 

 of the oscillation lay between the limits 7920 and 1800 vibra- 

 tions per second, the wave-length of the corresponding sound- 

 wave lying between I 3 and 7 '3 inches. 



A number of photographs were obtained in which the first 

 wave-front had travelled 10 such a distance that the second wave 

 should have been clear of the terminals. But in none of them 

 is any trace to be seen of a second wave. Thus, as an 

 attempt to photograph a train of waves, this method too proved 

 a failure. 



Even as a failure the result is not without its interest, as it 

 brings out very clearly the difference in character between the 

 first discharge and the surgings that follow it when the spark is 

 an oscillatory one. 



In the words of Prof. Trowbridge, who has made a special 

 study of oscillatory discharges :—" Photographs of powerful 



electric sparks lead one to conclude that a discharge of lightning 

 makes way for its oscillations by first breaking down the resist- 

 ance of the air by a disruptive pilot spark — through the hole 

 thus made in the air the subsequent surgings or oscillations take 

 p lace." 



In the language of the modern ionic theory this would be 

 inter))rcted by saying that the first disruptive discharge results 

 in the production of a large number of free ions, and these ions 

 offer an easy passage to the subsequent oscillations. 



Since the refractive power of a gas is proportional to its 

 density, and that in turn depends on the temperature, the method 



NO. 1694, VOL. 65] 



will reveal the presence of any region in the air whose tempera- 

 ture differs materially from that of the surrounding atmosphere. 

 Such a case arises in every flame ; the products of combustion 

 rise from the flame as*a column of heated gas, and this is 

 revealed to the eye as a pillar of fire, or in the photograph as a 

 pillar of cloud. The photograph shows the effect due to the 

 flame of a spirit lamp (Fig. 3). 



In order to impress the fact that what is here seen is not the 

 flame, but the heated gas rising from the flame, a small fan was 

 arranged which could be set in rapid rotation and so drive the 

 hot gas all over the field of view (Fig. 4). The photograph 

 (Fig. 5) shows the effect when it has been set rotating. 



The flame of a Bunsen burner gives rise to a disturbance 

 similar to that due to the spirit lamp, but more voluminous in 

 character. The peculiar spiral form in which the column of gas 

 rises before breaking up into a cloud is very noticeable in some 

 of these photographs of flames. 



A number of photographs were taken of jets of gas issuing 

 from a narrow orifice. In one case the jet is formed by blowing 

 heated air through a brass tube; in another it consists of carbonic 

 .acid gas issuing from the generating flask. In consequence of 

 its great density, the gas begins to fall downwards soon after 

 leaving the nozzle. 



Several photographs were taken to show the mode of formation 

 of a vortex ring of heated air. These rings were produced in 

 the usual way by means of a box with an aperture in one side, 

 and the opposite side formed of some elastic material. On 

 giving this side a sharp tap, some of the enclosed air rushes out 

 with the formation of a ring. In this case the air in the box 

 was heated by placing a spirit lamp inside it so that the rings, 



being formed of hot air instead of smoke, were quite invisible 

 save by the use of the method of stri;e. Some of these photo- 

 graphs are reproduced in the paper and are very instructive as 

 showing the way in which the vortex motion is produced. The 

 air appears first of all to issue from the orifice in the form of a 

 column, but the tail is gradually left behind while the whirlpool 

 motion of the head is accentuated. Finally, little is to be seen 

 but the section of the ring itself, the spiral structure being 

 strongly marked (Fig. 6). In some cases a circular, in others an 

 elliptic orifice was used. 



The appearance of some of these photographs showing vortex 

 motion strongly recalls the published photographs of the nebuhe 

 of the heavens. 



THE TEMPERA TURE OF fN VERS ION OF THE 

 JOULE-KELVIN EFFECT FOR HYDROGEN.' 

 T X the year 1854 it was proved by Joule and Lord Kelvin that 

 hydrogen on tree expansion behaved diflerently to all other 

 gases. Air. when allowed to expand from a higher to a lower 

 pressure without performing external work, became cooled, the 

 fall of temperature being proportional to the dilVerence of pres- 

 sure ; hydrogen, on the other hand, became warmer. As is 

 well known, the Joule-Kelvin eflect has been applied by 

 Ilampson and Linde to the liquefaction of air in quantity, but 

 since for hydrogen the effect is of opposite sign, it was obvious 

 that the Hampson-Linde apparatus could not be directly applied 

 1 By K. OUzew,ki. Tr.insl.ltcd from the BulUtin dc fAc-idi-mie des 

 Sciences de Cracovic. (December, 1901.) 



