1877] WRITINGS OF JOSEPH HENRY. 503 



Another hypothesis has been suggested, that it is due to a 

 flocculent condition of the atmosphere, or to an invisible 

 acoustic cloud of a density differing from that of the general 

 atmosphere at the time. To test this hypothesis experiment- 

 ally, the large trumpet of the siren was gradually elevated 

 from its usual horizontal position to a vertical one. In con- 

 ception this experiment appears very simple, but on account 

 of the great weight of the trumpet it required the labor of 

 several men for two days to complete the arrangements nec- 

 essary to the desired end. The trumpet in its vertical posi- 

 tion was sounded at intervals for two days, but in no instance 

 was an echo heard from the zenith, but one was in every case 

 produced from the entire horizon. The echo appeared to be 

 somewhat louder from the land portion of the circle of the 

 horizon than from that of the water. On slowly restoring 

 the trumpet to its former direction, the echo gradually in- 

 creased on the side of the water until the horizontal position 

 was reached, when the echo as usual appeared to proceed from 

 an azimuth of about twenty degrees of the horizon, the middle 

 of which was in the prolongation of the axis of the trumpet. 

 A similar experiment was made with one of the trumpets of 

 the two sirens at Little Gull Island. In this case the trumpet 

 was sounded in a vertical position every day for a week with 

 the same result. 



From these experiments it is evident that the phenomenon 

 is in some way connected with the plane of the horizon, and 

 that during the continuance of the experiment of sounding 

 the trumpets while directed towards the zenith, no acoustic 

 cloud capable of producing reflection of sound existed in the 

 atmosphere above them. 



Another method of investigating this phenomenon oc- 

 curred to me, which consisted in observing the effects pro- 

 duced on the ear of the observer by approaching the origin 

 of the echo. For this purpose, during the sounding of the 

 usual interval of twenty seconds of the large trumpet at Block 

 Island, observations were made from a steamer which pro- 

 ceeded from the station into the region of the echo and in 

 the line of the prolongation of the axis of the trumpet, with 

 the following results : 



