166 On the Refraction of Sound by the Atmosphere. [Jan. 6, 



distances to which sounds were heard on this day is supposed to be due. 

 The diminution in the temperature of the air being downwards, the sound, 

 instead of being lifted as it usually is, was brought down, and thus 

 intensified at the surface of the water, which, being perfectly smooth, 

 was thus converted into a sort of whispering-gallery. 



The report of the pistol and the sounds of the voice were attended with 

 echoes, but not so the reports of the rockets ; and it is suggested that 

 these so-called echoes may be found only to attend sounds having a 

 greater intensity in one direction than in another. 



The second part of the paper refers to a phenomenon noticed by Arago 

 in his report of the celebrated experiments on the velocity of sound made 

 on the nights of the 21st and 22nd of June, 1822. 



It was then found that, although the guns fired at Montlhery could be 

 distinctly heard at Villejuif (11 .miles distant), those fired at Villejuif 

 could not be heard at Montlhery without great attention, and at times 

 (particularly on the second night) they were not heard at all — although 

 on both nights the wind was blowing from Villejuif to Montlhery, the 

 speed of the wind, which was very light, being about 1 foot per second. 

 No explanation of this phenomenon was offered by the observers, although 

 it was much commented on. And on the second night the gun at Villejuif, 

 which on the previous night had been pointed upward, was brought down 

 in the hope that this might improve its audibility (this step, however, was 

 found to render matters worse than before). 



Erom this lowering of the gun at Villejuif it seemed as though there 

 was probably some difference in the conditions under which the guns at 

 the two stations were placed, as if that at Villejuif was fired from a 

 level, while that at Montlhery might be fired over a parapet. An inspec- 

 tion of the district confirmed this view ; for Villejuif is on a low flat hill, 

 while Montlhery is on the top of a steep cone ; and not only is it 80 feet 

 above Villejuif, but it is surmounted by the mound of an old castle, which 

 is supported by a vertical wall towards Villejuif and surrounded by a low 

 rampart. Hence it is suggested that in all probability the advantage of 

 the gun at Montlhery was due to its being fired over this parapet, while 

 that at Villejuif was fired from the level ground. 



The fact that the wind blowing from Villejuif did not reverse this 

 advantage, suggested the possibility that at night, when the diminution of 

 temperature is downward, a light wind may not produce the same effect 

 upon sound as when the diminution of temperature is upward, as it 

 generally is during the day. 



To ascertain if this is the case, some observations were made on some 

 calm nights in May and June of the present year, from which it was 

 found : — 



(1) That when the sky was cloudy and there was no dew, the sound of an 

 electric bell 1 foot above the grass could always be heard further with the 

 wind than against it ; but 





