38 BULLETIN OF THE 



and generalizations reached by him, while they as strikingly discred- 

 ited as incongruous the rival hypothesis of hygroscopic flocculence in 

 the atmosphere as a notable occasion of acoustic disturbance. When 

 we consider the wide areas over which fog-signals are designed to 

 be conveyed — through which spaces the atmosphere can rarely be 

 uniform, either in its temperature or its movements — we can readily 

 understand that from these two prominent conditions of sound-re- 

 fraction, acoustic rays are commonly propagated in quite sensibly 

 curved or often serpentine directions ; and that while these inequal- 

 ities will sometimes favor audibility at given points, they will as 

 often impair or defeat it. Moreover, these deformations of sound 

 waves are not confined to vertical planes, since it has been shown 

 that lateral refractions may exist, giving false impressions of direc- 

 tion as well as of distance. 



As we have no means of either controlling or accurately deter- 

 mining these simultaneous differences of wind and temperature, we 

 are forced to admit that the practical difficulties attending these 

 anomalies of sound propagation are insoluble and incurable. But 

 we must not hence abandon sound-signalling as either hopeless or 

 inefficient, since it is the best — or rather the only — method at our 

 disposal of giving warning and guidance to the befogged mariner. 



Two partial alleviations of the recognized defects are suggested. 

 The first is to place the siren or the steam whistle at considerable 

 elevations, say on the top of skeleton towers, perhaps higher than 

 those ordinarily employed as light-towers ; at which points they 

 could readily be operated from the ground. This would, in many 

 cases, counteract the tendency to local acoustic shadows or bands 

 of silence, though in other cases it would be quite ineffectual. The 

 second expedient is, (if not too expensive,) to greatly multiply the 

 number of such signals at available points about dangerous coasts 

 or inlets, with proper distinctions to clearly specialize their indica- 

 tions, in order that the mariner failing to catch the sound from one 

 direction, might have the probability of picking up the sound from 

 a different azimuth. As these sound instruments may be operated at 

 considerable distances from the engine, and even at practically in- 

 accessible positions, on rocks or on buoys, danger points especially 

 should be guarded by fog-signals, not necessarily of great power, 

 but capable, at least, of covering the radius of actual insecurity. 



Remarks were made by Mr. William B. Taylor on the rela- 

 tion of fog and snow storms to audibility. 



