THE UPPER ATMOSPHERE DOBSON 185 



ever, very rough, chiefly because it was not possible to get accurate 

 measurements of the brightness and speed of the meteors. 



The idea that the air at a height of some 50 kilometers might be 

 warm was found to fit in with the observations of sound heard at 

 a great distance from its source, and such observations have now been 

 used to give us a much better knowledge of the temperatures at 

 heights of betAveen 40 and 70 kilometers above the surface. When 

 a large explosion occurs, the sound is heard for many miles around, 

 but not to such great distances as might be expected. That the 

 sound is not heard farther away is largely due to the fact that 

 sound travels faster in warm air than in cold air. Since in the 

 lower atmosphere, the temperature falls with increasing height, 

 sound travels faster through surface air than through the air a 

 little higher up. Thus the sound waves near the surface run ahead 

 of those above, with the result that the sound ray is bent upward, 

 away from the ground. 



The bending of sound waves upward is well seen in the fact that 

 sounds in a valley are often heard much more clearly on the hills on 

 either side than at places the same distance away in the bottom of 

 the valley. Now, in the case of very large explosions it is fre- 

 quently found that while the sound cannot be heard for more than 

 20 or 30 miles round the explosion, it is heard again at places vevj 

 much farther away, perhaps a hundred miles from the explosion, 

 and often heard quite loudly. Such cases as the firing of salvos by 

 the fleet often show this phenomenon, and in a recent case when the 

 fleet was off Portland Bill the firing was heard as far away as 

 Surrey and Oxfordshire but not at many intervening places. 



The sound waves which are heard at these great distances are 

 found to have taken some 2 minutes longer over their journey than 

 they should have done had they traveled direct through the lower 

 atmosphere. From this and the fact that they are not heard at 

 intermediate distances, it seems clear that the sound has traveled 

 up into the upper atmosphere and then been deflected back to the 

 ground. This curious behavior had been known for a long time and 

 several explanations had been suggested to account for the fact that 

 the sound is deflected back to earth by the upper air. One sug- 

 gestion was that the upper air was largely composed of hydrogen, 

 and since it is known that sound travels faster in hydrogen than in 

 oxygen and nitrogen, this would account for the downward bending 

 of the sound rays. None of these explanations, however, carried 

 conviction until Dr. Whipple pointed out that if the temperature 

 rose again at very great heights, as was suggested from the meteor 

 results, then the sound, rays would be bent down on entering this 

 upper warmer region in just the same way as they are bent upward 

 by the fall of temperature in the air near the ground. 



