LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF CALIFORNIA. 
 
 GIF^T Ot^ 
 
 Class 
 
SUMMARY 
 
 OF 
 
 RESEARCHES IN SOUND 
 
 CONDUCTED IN THE SERVICE OF 
 
 THE UNITED STATES LIGHT-HOUSE BOARD, 
 
 BY 
 
 JOSEPH HENRY, 
 
 DURING THE YEARS 1865 TO 1877 
 
 [FROM THE SMITHSONIAN REPORT FOR 1878. J 
 
 WASHINGTON: 
 
 GOVERNMENT PRINTING OFFICE 
 1879. 
 
(I 'V 
 
 
RESEARCHES IN SOUND: 
 
 WITH SPECIAL KEFERENCE TO FOG-SIGNALING. 
 
 BY JOSEPH HENRY. 
 [FROM THE ANNUAL REPORTS OF THE U. S. LIGHT-HOUSE BOARD.] 
 
 PREFATORY NOTE. 
 
 [The series of investigations undertaken by the late Professor Henry 
 in the interest of the light-house service, embracing not only observa- 
 tions more than usually laborious and extended, with regard to the 
 atmospheric conditions affecting the propagation of sound to a distance,, 
 but varied and elaborate experimental inquiries, as well, with reference- 
 to the most efficient character and form of sonorous instruments for 
 fog-signaling purposes, commenced as far back as the year 1865, and 
 continued to the last year of his life. 
 
 These important investigations, though in the language of an official 
 report, they "have resulted in giving us a fog-signal service conceded 
 to be the best in the world,"* have hitherto received so little publicity 
 and attention, owing to the purely official character and channel of their 
 presentation, that their collection and republication here (in advance of 
 of a possible edition of Henry's collected works), appears to be called 
 for in the interests of science, as well as of a just appreciation of the 
 value of his prolonged researches. 
 
 The first part, extracted from the Appendix to the Light House Ee- 
 port for 1874, comprises a preliminary statement and an account of ob- 
 servations and experiments made by the author, from 1865, at New 
 
 * Executive Document No. 94, Forty-fifth Congress, second session, Senate, p. 2. 
 This is a report to the Hon. Secretary, of the Treasury, made since Professor Henry's 
 death. To a similar effect, may be quoted an official statement made five years 
 earlier. In 1873, Major George H. Elliot, of the Light-House Board, commissioned to 
 make a tour of inspection of European light-house establishments, presented the re- 
 sults of his observations abroad in a very able and elaborate report, published by the 
 Senate in 272 octavo pages, with numerous illustrations. In his preliminary report to 
 the Board, dated September 17, 1873, he concludes, that while there are " many details 
 of construction and administration which we can adopt with advantage," (from the 
 British and French light-house systems,) "there are many in which we excel. Our 
 shore fog-signals, particularly, are vastly superior, both in number and power."' 
 (Executive Document No. 54, Forty-third Congress, first session; Senate, p. 12.) 
 
 455 
 
 130004 
 
456 RESEARCHES IN SOUND. 
 
 Haven, Conn., to 1872, at Portland, Me. The second part is a commu- 
 nication made to the "Philosophical Society of Washington n (of which 
 he was the president) December 11, 1872, embracing a discussion of 
 some abnormal phenomena of sound, and is extracted from the Bulletin 
 of the Society, vol. ii, appendix ix. The third part, forming the latter 
 portion of the Appendix to the Light-House Report for 1874, comprises 
 investigations extending from 1873, at Whitehead Station, off the coast 
 of Maine, to 1874, at Sandy Hook, New Jersey. The fourth part, extracted 
 from the Appendix to the Light-House Eeport for 1875, comprises his 
 investigations for that year. And the fifth part, from the Appendix to 
 the Light-House Eeport of 1877, embraces his latest observations dur- 
 ing September and October of the year 1877. 
 
 These papers necessarily are more fragmentary, and at the same time 
 involve more recapitulation than would have been the case, could they 
 have received the revision of their distinguished and lamented author. 
 It has however been considered more just to reproduce these contribu- 
 tions in this form, than to attempt either a compilation of extracts, or a 
 condensation of their substance in the form of an abstract.] 
 
 S. F. BAIRD. 
 
RESEARCHES IN SOUND, 
 
 INTRODUCTION.* 
 
 
 
 FOG. 
 
 Among the impediments to navigation none perhaps are more to be 
 dreaded than those which arise from fogs, and consequently the nature 
 of this impediment and the means which may be devised for obviating 
 it are objects of great interest to the mariner. Fogs are in all cases 
 produced when cold air is mingled with warm air saturated with moist- 
 ure. In this case the invisible vapor of the warmer air is condensed by 
 the cold into minute particles of liquid water, which, by their immense 
 number and multiplicity of reflecting surfaces, obstruct the rays of light 
 in the same way that a piece of transparent glass when pounded becomes 
 almost entirely opaque and is seen by reflection as a white mass. So 
 greatly does a dense fog obstruct light, that the most intense artifi- 
 cial illumination, such as that produced by the combustion of magnesium, 
 by the burning of oxygen and hydrogen in contact with lime, and that 
 produced between the charcoal points of a powerful electrical apparatus, 
 are entirely obscured at comparatively short distances. Even the light 
 of the sun, which is far more intense than that of any artificial illumina- 
 tion, is so diminished by a single mile of dense fog that the luminary 
 itself becomes invisible. Eecourse must therefore be had to some other 
 means than that of light to enable the mariner to recognize his position 
 on approaching the coast when the land is obscured by fog. 
 
 The only means at present known for obviating the difficulty is that 
 of employing powerful sounding instruments which may be heard at a 
 sufficient distance through the fog to give timely warning of impending- 
 danger. Investigations therefore as to the nature of sound and its 
 applications to fog-signals become an important object to those in charge 
 of aids to navigation. Such investigations are of special importance in 
 connection with the light-house service of the United States. The 
 northeastern coast of the United States on the Atlantic, and the entire 
 western coast on the Pacific, included in our territory, are subject, 
 especially during the summer months, to dense fogs, which greatly im- 
 pede navigation, as well as endanger life and property. 
 
 The origin of the fogs on our coast is readily explained by reference 
 to a few simple principles of physical geography. In the Atlantic Ocean 
 there exists a current of warm water proceeding from the Gulf of Mex- 
 ico, between Cuba and Florida, which flows along our coast to the lati- 
 
 * From the Report of the Light-House Board, for 1874. 
 
 457 
 
458 RESEARCHES IN SOUND. 
 
 tude of about 35, and then turning gradually to the eastward, crosses the 
 Atlantic and impinges against the coast of Northern Europe. Through- 
 out its entire course, on account of the immense capacity of water for 
 heat, the temperature of the stream is greater than that of the ocean on 
 either side. In addition to this stream, the Atlantic Ocean is traversed 
 by another current of an entirely opposite character, one of cold water, 
 which, coming from the arctic regions down Davis's Strait, is thrown, by 
 the rotation of the earth, against our coast, passing between it and the 
 Gulf-stream, and sinking under the latter as it approaches the southern 
 extremity of the United States. 
 
 These conditions are those most favorable to the production of fogs, 
 since whenever the warm air, surcharged with moisture, is blown from 
 the Gulf-stream over the arctic current and mingles with the cold air 
 of the latter, a precipitation of its vapor takes place in the form of fog. 
 Hence, especially in summer, when the wind in the eastern part of the 
 United States is in a southeasterly direction, fogs prevail. As we pro- 
 ceed southerly along the coast, the fog-producing winds take a more 
 easterly direction. 
 
 A somewhat similar circulation in the Pacific Ocean produces fogs on 
 the western coast of the United States. In this ocean a current of warm 
 water, starting from the equatorial regions, passes along the shores of 
 China and Japan, and, following the general trend of the coast, turns 
 eastward and continues along our shore. The northern part of this 
 current being warmer than the ocean through which it passes, tends to 
 produce dense fogs in the region of the Aleutian Islands and tlie coast 
 of Alaska. As this current descends along the American coast into 
 lower latitudes it gradually loses its warmth, and soon assumes the 
 character, in regard to the water through which it passes, of a compara- 
 tively colder stream; and to this cause we would attribute the preva- 
 lence of fogs on the coast of Oregon and California, which are most 
 prevalent during the spring and early summer, with wind from the 
 northwest and west. 
 
 From what has been said, it is evident that the fogs in the Aleutian 
 Islands occur chiefly in summer, when southwesterly winds prevail and 
 mingle the moist air from the warm current with the colder air of the 
 more northerly latitude. In winter, the wind being from the north 
 chiefly, the moist air is driven in an opposite direction, and dense fogs 
 therefore at this season do not prevail. 
 
 In regard to the fogs on the coast of Maine, the following interesting 
 facts were furnished me by the late Dr. Stiinpson, formerly of the Smith- 
 sonian Institution and of the Chicago Academy of Sciences, who had 
 much experience as to the weather during his dredging tor marine speci- 
 mens of natural history in the region of Grand Manan Island, at the 
 entrance of the Bay of Fundy. 
 
 a So sharply marked," says Dr. Stimpson, "is the difference of tem- 
 perature of the warm water from the Gulf-stream and that of the polar 
 
RESEARCHES IN SOUND. 459 
 
 current, that in sailing in some cases only a few lengths of a ship the 
 temperature of the water will change from 70 to 50. The fog fre- 
 quently comes rolling in with the speed of a race-horse 5 in some cases 
 while dredging, happening to turn my eyes to the south, a bank of fog 
 has been seen approaching with such rapidity that there was scarcely 
 time in which to take compass-bearing of some object on shore by which 
 to steer, before I would be entirely shut in, perhaps for days together." 
 He also mentions the fact that it frequently happened during a warm 
 day, while a dense fog existed some distance from the shore, close in to 
 the latter there would be a space entirely clear; this was probably due 
 to the reflection and radiation of the heat from the land, which converted 
 the watery particles into invisible vapor. 
 
 Dr. Stimpson has also noticed another phenomenon of some interest. 
 "When a dense fog, coming in regularly from the sea, reaches the land, 
 it gradually rises in the atmosphere and forms a heavy, dark cloud, which 
 is frequently precipitated in rain." This rising of fog is not due, accord- 
 ing to the doctor, to a surface- wind from the west pressing under it and 
 buoying it upward, since the wind at the time is from the ocean. It is 
 probably due to the greater heat of the land causing an upward cur- 
 rent, which, when once started, by its inertia carries the cloud up to a 
 region of lower temperature, and hence the precipitation. The height 
 of the fog along the coast is not usually very great, and can be frequently 
 overlooked from the mast-head. The deception as to size and distance 
 of objects as seen in a fog is also a remarkable phenomenon when ob- 
 served for the first time. A piece of floating wood at a little distance is 
 magnified into a large object, and after much experience the doctor was 
 not able to overcome-the delusion. It is said that the sailors in the Bay 
 of Fuudy prefer of two evils a fog that remains constant in density to 
 one that is variable, although the variation may be toward a greater 
 degree of lightness, on account of the varying intensity producing a 
 varied and erroneous impression of the size and distance of the object 
 seen through it. It is also his impression that sound can be heard as 
 well during fog as in clear weather, although there is a delusion even 
 in this, since the source of sound, when seen, appears at a greater dis- 
 tance than in a clear atmosphere, and hence the sound itself would 
 appear to be magnified. 
 
 Fogs also exist on the Mississippi, especially on the lower portion of 
 the river. They are of two classes, those which result from the cooling 
 of the earth, particularly during the summer in clear nights, with wind 
 probably from a northerly direction, followed by a gentle, warm wind 
 from the south surcharged with moisture, and the other induced by the 
 water of the river, which, coming from melting snow of northern regions, 
 is colder than the air in the vicinity. The air over the river being thus 
 cooled below the temperature of a gentle wind from the south, the 
 moisture of the latter is precipitated. This fog, which occurs in the last 
 of winter, during the spring, and beginning of summer, is very dense, 
 
460 RESEARCHES IN SOUND. 
 
 but is confined entirely to the atmosphere above the river, while the 
 other class of fog exists over the land as well. 
 
 FOG- SIGNALS. 
 
 The importance of fog-signals as aids to navigation, especially on the 
 northeastern portion of our coast, of which the shore is exceedingly bold 
 and to the approach of which the sounding-line gives no sure indication, 
 has been from the first an object of special attention. 
 
 At the beginning of the operations of the Light-House Board such 
 instruments were employed for producing sound as had been used in 
 other countries ; these consisted of gongs, bells, guns, horns, &c. The 
 bells were actuated by clock-machinery, which was wound up from time 
 to time and struck at intervals of regular sequence by which their posi- 
 tion might be identified. The machinery, however, by which these bells 
 were struck was of a rude character and exceedingly wasteful of power, 
 the weight continuing to descend during the whole period of operation, 
 including the successive intervals of silence. This defect was remedied 
 by the invention of Mr. Stevens, who introduced an escapement arrange- 
 ment, similar to that of a clock, which is kept in motion by a small 
 weight, a larger one being brought into operation only during the 
 instant of striking. 
 
 Bell-buoys were also introduced at various points. These consisted 
 of a bell supported on a water-tight vessel and rung by the oscillation 
 of the waves, but all contrivances of this kind have been found to be 
 untrustworthy; the sound which they emit is comparatively of feeble 
 character, can be heard at but a small distance, and is frequently ineffi- 
 cient during a fog which occurs in calm weather* Besides this, auto- 
 matic fog- signals are liable to be interfered with by ice in northern 
 positions, and in all sections to derangement at times when no substi- 
 tute can be put in their place, as can be in the cases of the bells rung 
 by machinery under the immediate control of keepers. A signal which 
 is liable to be interrupted in its warnings is worse than no signal, since 
 its absence may give confidence of safety in the midst of danger, and 
 thus prevent the necessary caution which would otherwise be employed. 
 
 Guns have been employed on the United States coast, first under the 
 direction of General Bates, engineer of the twelfth district, at Point 
 Bonita, San Francisco Bay, California. The gun at this station con- 
 sisted of a 24-pounder, furnished by the War Department. The neces- 
 sary arrangements being made, by the construction of a powder-house, 
 and laying of a platform, and employment of a gunner, notice to mari- 
 ners was given that after the 8th of August, 1856, a signal- gun would 
 be fired every hour and half-hour, night and day, during foggy or thick 
 weather. The first year, with the exception of eighty-eight foggy days, 
 omitted for want of powder, 1,390 rounds were fired. These consumed 
 5,560 pounds of powder, at a cost of $1,487, pay of gunner and incident- 
 als excluded. The following 3<ear the discharges were 1,&S2, or about 
 
RESEARCHES IN SOUND. 461 
 
 one-eleventh of the number of hours and half-hours of the whole time. 
 The fog-gun was found to answer a useful purpose ; vessels by the help 
 of it alone having come into the harbor during a fog at night, as well as 
 in the day, that otherwise could not possibly have entered. This signal 
 was continued until it was superseded by a bell boat. A gun was also 
 used at West Quoddy Head, near the extreme eastern part of Maine. 
 It consisted of a short piece, or carronade, 5 feet long, with a bore of 5J 
 inches, charged with four pounds of blasting-powder. The powder was 
 made up in cartridges and kept in chests in the work-house. The gun 
 was only fired on foggy days, when the steamboat running between Boston 
 and Saint John, New Brunswick, was approaching the light-house from 
 the former place. In going in the other direction the signal was not so 
 much required, because in the former case (of approach) the vessel had 
 been for some time out of sight of land, and consequently its position 
 could not be so well known. The firing was commenced with the hear- 
 ing of the steamer's whistle as she was approaching, and as the wind 
 during the fog at this place is generally from the south, the steamer 
 could be heard five or six miles. The firing was continued as frequently 
 as the gun could be loaded until the steamer answered by a signal of 
 three puff's of its whistle. The number of discharges was from one to 
 six, the latter exhausting a keg of powder valued at $8. The keeper of 
 the light-house acted as gunner, without compensation other than his 
 salary. The cost of powder was paid by the steamboat company. The 
 report of the gun was heard from two to six miles. 
 
 This signal has been abandoned, because of the danger attending its 
 use, the length of the intervals between the successive explosions, 
 and the brief duration of the sound, which renders it difficult to deter- 
 mine with accuracy its direction. 
 
 The lamented General Bache, of the Light-House Board, adopted a 
 very ingenious plan for an automatic fog-signal, which consisted in tak- 
 ing advantage of a conical opening in the coast, generally designated a 
 blow-hole. On the apex of this hole he erected a chimney which termi- 
 nated in a tube surmounted by a locomotive-whistle. By this arrange- 
 ment a loud sound was produced as often as a wave entered the mouth 
 of the indentation. The penetrating power of the sound from this 
 arrangement would not be great if it depended merely on the hydro- 
 static pressure of the wave, since this, under favorable circumstances, 
 would not be more than that of a column of water 20 feot high, giving 
 a pressure of about 10 pounds to the square inch. The effect however 
 of the percussion might add considerably to this, though the latter 
 would be confined in effect to a single instant. In regard to the prac- 
 tical result from this arrangement, which was continued in operation 
 for several years, it was found not to obviate the necessity of produc- 
 ing sounds of greater power. It is however founded on an ingenious 
 idea, and may be susceptible of application in other cases. 
 
462 RESEARCHES IN SOUND. 
 
 EXPERIMENTS BY PROFESSOR ALEXANDER, IN 1855. 
 
 The Light- House Board was not content with the employment alone 
 of the fog-signals in ordinary use, but directed a series of experiments 
 in order to improve this branch of its service. For this purpose the 
 board employed Prof. J. H. Alexander, of Baltimore, who made a 
 report on the subject, which was published among the documents. The 
 investigations of Professor Alexander related especially to the use of 
 the locomotive steam- whistle as a fog-signal, and in his report he details 
 the results of a series of experiments in regard to the nature and adjust- 
 ment of the whistle, the quantity of steam necessary to actuate it, with 
 suggestions as to its general economy and management. He found, 
 what has since been fully shown, that the power of the sound depends 
 upon the pressure of the steam in the boiler, and the pitch upon the 
 distance between the circular orifice through which the steam issues, 
 and the edge of the bell. He appears however to be under an erro- 
 neous impression that the sound is produced by the vibrations of the 
 metal of the goblet or bell, while in fact this latter portion of the appa- 
 ratus is a resounding cavity, which, as I have shown in subsequent 
 experiments, may be constructed of wood as well as of brass, in order 
 to produce the same effect. Mr. Alexander also mentions the efiect of 
 the wind in diminishing the penetrating power of sound when in an 
 adverse direction, either directly or approximately. He also recom- 
 mends the adoption of an automatic pump to supply the boilers with 
 water, and also to open and shut the valves at the proper intervals for 
 blowing the whistle. He states that the location of a sound can be 
 determined more precisely in the case of loud, high sounds than in that 
 of feebler or lower ones. On this point I am not prepared to concur 
 with him in experiments of my own. In all cases however loud sounds 
 are more desirable than feebler ones, in order that they may be heard 
 at a greater distance above the noise of the surf and that of the wind 
 as it passes through the spars and rigging of vessels. 
 
 The board however at this time were not prepared to adopt these 
 suggestions, and an unsuccessful attempt to use a steam-boiler, rendered 
 abortive by the incapacity of the keeper to give it proper attendance, 
 discouraged for a time efforts in this line. 
 
 Previous to the investigations of Mr. Alexander, at the expense of the 
 Light Hv>use Board, Mr. Daboll, of New London, had for several years 
 been experimenting on his own account with reference to a fog-signal. 
 His plan consisted in employing a reed trumpet, constructed after the 
 manner of a clarionet, and sounded by means of air condensed in a res- 
 ervoir, the condensation being produced by horse-power operating 
 through suitable machinery. Although the sound of this was more pene- 
 trating than that of bells, still the expense and inconvenience of the 
 maintenance of a horse, together with the cost of machinery, prevented 
 its adoption. Mr. Daboll however after this presented to the board a 
 
RESEARCHES IN SOUND. 463 
 
 modification of his invention, in which a hot-air engine of Ericsson's 
 patent was substituted as the motive-power, instead of the horse ; and 
 the writer ot this report, as chairman of the committee on experiments 
 in behalf of the board, examined this invention and reported in favor ot 
 its adoption. The other members of the committee made an unfavorable 
 report, on the ground that fog-signals were of little importance, since the 
 mariner should know his place by the character of his soundings in all 
 places where accurate surveys had been made, or should not venture 
 near the coast until the fog was dissipated. The board however estab- 
 lished Daboll trumpets at different stations, which have been in constant 
 use up to the present time. 
 
 PAET I. INVESTIGATIONS FKOM 1865 TO 1872.* 
 EXPERIMENTS NEAR NEW HAVEN, IN 1865. 
 
 The subject of sound, in connection with fog-signals, still continued to 
 occupy the attention of the board, and a series of investigations was 
 made in October, 1865, at the light-house near New Haven, under the 
 direction of the writer of this report, in connection with Commodore, now 
 Admiral, Powell, inspector, and Mr. Ledeile, acting engineer d>f the third 
 district. 
 
 The principal object was to compare the sound of bells, of steam- 
 whistles, and other instruments, and the effect of reflectors, and also the 
 operation of different hot-air engines. For this purpose the committee 
 was furnished with two small sailing-vessels. As these were very imper- 
 fectly applicable, since they could not be moved without wind, the writer 
 of the report devised an instrument denominated an il artificial ear," by 
 which the relative penetrating power of different sounding bodies could 
 be determined and expressed in numbers by the removal of the observer 
 to a comparatively short distance from the point of origin of the sound. 
 This instrument consisted of a conical horn, made of ordinary tinned 
 sheet iron, the axis of which was about 4 feet in length, the diameter of 
 the larger end 9 inches, and tapering gradually to If of an inch at the 
 smaller end. The axis of this horn was bent at the smaller end in a 
 gentle curve, until the plane of the section of the smaller end was at 
 light angles to the perpendicular section of the larger end, so that when 
 the axis of the trumpet was held horizontally and the larger section 
 vertically, then the section of the smaller end would be horizontal. 
 Across the smaller end a thin membrane of gold beater's skin was slightly 
 stretched and secured by a thread. On this membrane fine sand was 
 strewn. To protect the latter from disturbance by the wind, it was sur- 
 rounded by a cylinder of glass, cut from a lamp-chimney, the upper end 
 of which was covered with a plate of glass ; and, in the improved condi 
 
 * From the Report of the Light-House Board, for 1874, 
 
464 RESEARCHES IN SOUND. 
 
 tion of the instrument, with a magnifying lens, with which to observe 
 more minutely the motions of the sand. To use this instrument in com- 
 paring the relative penetrating power of souud from different sources, 
 as for example from two bells, the axis bring held horizontal, the mouth 
 was turned toward one of the bells, and the effect causing agitation of 
 the sand, was noted. The instrument was then removed to a station a 
 little further from the bell, and the effect again noted, the distance being 
 increased, step by step, until no motion in the sand could be observed 
 through the lens. This distance, being measured in feet or yards, gave 
 the number indicating the penetrating power of the instrument under 
 trial. The same experiment was immediately repeated, under the same 
 conditions of temperature, air, wind, &c., with the other sounding appa- 
 ratus, and the relative number of yards indicating the distance, taken 
 as the penetrating powers of the two instruments. It should be observed, 
 in the use of this instrument, that it is intended merely to concentrate 
 the rays of sound, and not to act as a resounding cavity ; since in that 
 case the sound, in unison with the resounding note, would produce effect 
 at a greater distance than one in discord. 
 
 The indications of this instrument were compared with the results ob- 
 tained by the ear in the use of the two vessels, and in all cases were in 
 exact accordance ; and it was accordingly used in the following investi- 
 gations, and has been found of great service in all subsequent experi- 
 ments on the penetration of sound. 
 
 The only precaution in using it is that the membrane shall not be of 
 such tension as to vibrate in unison with a single sound or its octaves ; 
 or, in other words, that the instrument must be so adjusted by varying 
 the length of the axis or the tension of ihe membrane that it shall be in 
 discordance with the sounds to be measured, and only act as a condenser 
 of the sonorous waves. 
 
 The first experiments made were with regard to the influence of re- 
 flectors. For this purpose a concave wooden reflector had been prepared, 
 consisting of the segment of a sphere of 16 feet radius, and covered with 
 plaster, exposing a surface of 64 square feet. In the focus of this, by 
 means of a temporary railway, a bell or whistle could be readily intro- 
 duced or withdrawn. The center of the mouth of the bell was placed in 
 the horizontal axis of the reflector. This arrangement being completed, 
 the sound of the bell, with and without the reflector behind it, was alter- 
 nately observed. Within the distance of about 500 yards the effect was 
 evidently increased, as indicated by the motion of the sand on the mem- 
 brane, but beyond this the difference was less and less perceptible, and 
 at the limit of audibility the addition of the reflector appeared to us en- 
 tirely imperceptible. This result was corroborated by subsequent ex- 
 periments in which a whistle was heard nearly as well in the rear of a 
 reflector as before it. It would appear from these results that while 
 feeble sounds, at small distances, are reflected as rays of light are, waves 
 
RESEARCHES IN SOUND. 465 
 
 of powerful sound spread laterally, and even when projected from the 
 mouth of a trumpet, tend at a great distance to embrace the whole circle 
 of the horizon. 
 
 Upon this and all the subsequent experiments, as it will appear, the 
 principle of reflection as a means of re-enforcing sound is but slightly ap- 
 plicable to fog-signals. It is evident however that the effect will be some- 
 what increased by augmenting the size of the reflector, and by more com- 
 pletely inclosing the source of sound in a conical or pyramidal reflector. 
 
 Another series of experiments was made to ascertain whether the pen- 
 etration of the sound was greater in the direction of the axis of the bells, 
 or at right angles to the axis ; or, in other words, whether the sound was 
 louder in front of the mouth of a bell or of its rim. The result of this 
 experiment was considered of importance, since, in one of the light- 
 houses, a bell has been placed with the plane of its mouth at right angles 
 to the horizon, instead of being placed, as usual, parallel to the same. 
 The effect on the sound in these two positions was similar to that pro- 
 duced by the bell with a reflector, the noise being greater at a short 
 distance with the mouth toward the observer than when the rim was in 
 the plane of the ear. At a distance however, the difference between 
 the two sounds was imperceptible. In practice therefore it is of very 
 little importance whether the axis of the bell is perpendicular or parallel 
 to the horizon. 
 
 The first fog-signal examined in this series of experiments was a double 
 whistle, improperly called a steam-gong, designed principally for a fire- 
 alarm and for signals for the comruencejnent of working-hours in large 
 manufacturing establishments. It consisted of two bells of the ordinary 
 steam-whistle on the same hollow axis, mouth to mouth, with a flat, hol- 
 low cylinder between them, through the upper and lower surface of which 
 the circular sheets of steam issue, the vibration of which produces the 
 sound. In the instrument under examination, the upper bell was 20 
 inches in length of axis, and 12 inches in diameter, and the lower- whistle 
 was of the same diameter, with a length of axis of 14 inches. The note 
 of the shorter bell was a fifth above that of the longer. This arrangement 
 gave a melodious sound, unlike that of the ordinary locomotive- whistle, 
 and on that account had a peculiar merit. The sound was also very 
 loud, and, according to testimony, had been heard under favorable 
 circumstances more than twenty miles. It however required a large 
 quantity of steam to give it its full effect, and the only means to obtain 
 an approximate idea as to this quantity was that afforded by observing 
 its action on a boiler of a woolen manufactory near Newport. It was 
 here blown with a pressure of at least 75 pounds. From theoretical con- 
 siderations however, it might be inferred that its maximum penetrating 
 power would be not greater than that of a single whistle using the same 
 amount of steam, and this theoretical inference was borne out by the 
 subsequent experiments of General Duane. But from th strikingly 
 S. Mis. 59 r-30 
 
466 RESEARCHES IN SOUND. 
 
 distinctive character of its tone it has, in our opinion, an advantage over 
 a single whistle expending an equal quantity of steam. 
 
 The fact that the vibration of the metal of the bell had no practical 
 effect on the penetrating power of the sound was proved quite conclu- 
 sively by winding tightly around each bell, over its whole length, a thick 
 cord, which would effectually stop all vibration. ^The penetration of the 
 sound produced under this condition was the same as that with the bells 
 free. It is true, the latter produces a difference in the quality of the 
 tone, such as that which is observed in a brass instrument and that of 
 one of wood or ivory. The inventor was not aware that the sound pro- 
 duced was from the resonance of the air within the bell, and not from 
 the metal of the bell itself, and had obtained a patent, not only for the 
 invention of the double whistle, but also for the special compound of 
 metal of which it was composed. 
 
 Another apparatus proposed to be used as a fog-signal was presented 
 for examination by the Marine Signal Company, of Wallingford, Conn. 
 It consisted of a curved tube of copper nearly in the form of the letter 
 C, and was supported on an axis passing through the center of the 
 figure. An ordinary bell- whistle was attached to each extremity of the 
 tube, the instrument being placed in a vertical position and partially 
 filled with water, then made to oscillate on its center of support. By 
 this means the air was drawn in at one end and forced out through the 
 whistle at the other. The motion being reversed the air was drawn in 
 at the end through which it had just made its exit and forced out through 
 the whistle at the other. By rocking the instrument, either by hand or 
 by the motion of the vessel, a continued sound could be produced. The 
 motive-power in the former case was muscular energy, and the experi- 
 ments which were made at this time, as well as all that have been made 
 subsequently, conclusively prove that the penetrating power of the 
 sound for practical use as a fog-signal depends upon the intensity of the 
 motive-energy employed. No instrument operated through levers and 
 pumps by hand-power is sufficient for the purpose. 
 
 One of these instruments with two 4-inch whistles gave a sound, as 
 indicated by the artificial ear, the power of which was about one- tenth 
 of that of a steam-trumpet. It was supposed however that this in- 
 strument would be applicable for light-ships ; and that if extended en- 
 tirely across the vessel, and armed with whistles of large size, it would 
 be operated by the rolling of the vessel, and thus serve to give warning 
 in time of thick weather. But as it frequently happens that fog exists 
 during a calm, this invention could not be relied upon to give warning 
 in all cases of danger. Besides this, the ordinary roll of a ship is not 
 sufficient to produce a hydrostatic pressure of more than five or six 
 pounds to the square inch, which is insufficient to give an effective 
 sound. It has however been proposed to increase the power by using 
 quicksilver instead of water; but, besides the first cost of this material, 
 and the constant loss by leakage and oxidation, the tendency to affect 
 
RESEARCHES IN SOUND. 467 
 
 the health of the crew is an objection to the introduction of this modi- 
 fication of the apparatus into light-ships. 
 
 The other instruments which were subjected to trial were an ordinary 
 steam-whistle and a Daboll trumpet. The bell of the whistle was 6 
 inches in diameter, 9 inches in height, and received the sheet of steam 
 through an opening of one-thirtieth of an inch in width ; was worked 
 by a pressure of condensed air of from 20 to 35 pounds per square inch, 
 and blown once in a minute for about five seconds. The air was con- 
 densed by a Eoper engine of one-horse power. The penetrating power 
 of the sound was increased by an increase in the pressure of the air, 
 and also the pitch. The tone however of the instrument was lowered 
 by increasing the distance between the orifice through which the circular 
 sheet of air issued at the lower rim of the bell or resounding cavity. 
 To prove conclusively that the bell performs the part of a mere resound- 
 ing cavity, a wooden one, on a subsequent occasion, was substituted for 
 that of metal without a change in the loudness or the pitch of the sound. 
 
 The penetrating power of the whistle was compared with a Daboll 
 trumpet, actuated by an Ericsson engine of about the same power ; the 
 reservoir for the condensed air of each machine was furnished with a 
 pressure-gauge, and by knowing the capacity of the condensing pumps 
 and the number of strokes required to produce the pressure, the relative 
 amount of power was determined. The result was that the penetrating 
 power of the trumpet was nearly double that of the whistle, and that an 
 equal effect was produced afc the same distance by about one-fourth of 
 the power expended in the case of the latter. It must be recollected 
 however, that the whistle sends sonorous waves of equal intensity in 
 every direction, while the greatest power of the trumpet is in the direc- 
 tion of its axis. This difference however is lessened on account of the 
 spreading of the sound to which we have before alluded.* The whistle 
 was blown, as we have said, with a pressure of from 20 to 35 pounds, 
 while the trumpet was sounded with a pressure of from 12 to 15 pounds. 
 In the case of the whistle, the pressure in the reservoir may be indefi- 
 nitely increased with an increase in the penetrating power of the sound 
 produced, while in the case of the trumpet a pressure greater than a 
 given amount entirely stops the blast by preventing the recoil of the 
 vibrating tongue ; this being made of steel, in the larger instruments 
 2 inches wide and 8 inches long, would receive a pressure of steam, at 
 only 10 pounds to the square inch, of 200 pounds, tending to press it 
 into the opening and to prevent its recoil, this circumstance limits, as it 
 were, the power of a trumpet of given dimensions. It is however well 
 fitted to operate with a hot-air engine, and is the least expensive in fuel 
 of any of the instruments now employed. The whistle is the simpler 
 and easier of management, although they both require arrangement of 
 machinery in order that they may be operated automatically. 
 
 * It is worthy of note however that in the case of a sound having primarily an axial 
 direction, the subsequent lateral diffusion must result in enfeebling the whole sphere 
 of expanding sound-waves in a more rapid ratio than the square of the distance. 
 
468 RESEARCHES IN SOUND. 
 
 It is a matter of much importance to obtain a hot-air engine of suffi- 
 cient power, and suitable for working fog-signals of all classes. This 
 will be evident when we consider the difficulty in many cases of obtain- 
 ing fresh water for producing steam, and the expense of the renewal of 
 the boilers in the use of salt- water, as well as that of the loss of power 
 in frequently blowing out the latter, in addition to the danger of the use 
 of steam by unskillful attendants. 
 
 The merits of the two engines however under consideration could not 
 be fully tested by the short trial to which they were subjected during 
 these experiments. The principal objection to the Ericsson engine was 
 the size of the fly-wheel and the weight of the several parts of the ma- 
 chine ; the Eoper engine was much more compact, and appeared to work 
 with more facility, but from the greater heat imparted to the air the 
 packing was liable to burn out and required to be frequently renewed. 
 Although at first the impression of the committee was in favor of the 
 Eoper engine, yet in subsequent trials of actual practice it was found 
 too difficult to be kept in order to be employed for light-house purposes, 
 and its use has consequently been abandoned ; another hot-air engine 
 has been employed by the board, the invention of a Mr. Wilcox, which 
 has also been discontinued for a similar reason. I was assured by the 
 person last named, a very ingenious mechanician, that when the several 
 patents for hot-air engines expired, a much more efficient instrument 
 could be devised by combining the best features of each of those now in 
 use. 
 
 For determining the relative penetrating power of these instruments, 
 the use of two vessels had been obtained, with the idea of observing the 
 sound simultaneously in opposite directions. 
 
 Unfortunately however the location which had been chosen for these 
 experiments was of a very unfavorable character in regard to the em- 
 ployment of sailing-vessels and the use of the artificial ear. It was fully 
 open to the ocean only in a southerly direction, navigation up the bay 
 to the north being limited to three and a half miles, while on shore a 
 sufficient unobstructed space could not be obtained for the proper use 
 of the artificial ear. With these obstructions and the necessity of beat- 
 ing against the wind, thereby constantly altering the direction of the 
 vessel, exact comparisons were not possible, yet the observations made 
 were sufficiently definite to warrant certain conclusions from them as to 
 the relative power of the various instruments submitted to examination. 
 
 The following is a synopsis of the observations on four different days. 
 Before giving these however, it is necessary to observe that at each 
 stroke of the piston of the hot-air engine a loud sound was produced by 
 the blowing off of the hot air from the cylinder, after it has done its work. 
 In the following statement of results the noise thus produced is called 
 the exhaust. On the first day but one set of observations was made, the 
 vessel's course being nearly in the line of the axis of the trumpet. The 
 
RESEARCHES IN SOUND. 469 
 
 order of penetrating power was as follows: 1, trumpet ; 2, exhaust; 3, 
 bell; these instruments being heard respectively at 5-J, 3J, and 2 miles. 
 The whistle was not sounded. 
 
 The second day simultaneous observations were made from two ves- 
 sels sailing nearly in opposite directions. The results of the observations 
 made on the vessel sailing in a southerly direction were very irregular. 
 The trumpet was heard at 3f miles and lost at 4f miles with the wind 
 slightly in favor of the sound, and heard at 6 J miles with the wind some- 
 what against the sound ; it was heard even at 7-jj miles from the mast- 
 head, though inaudible from the deck. In all these cases the position 
 of the vessel was nearly in line with the axis of the trumpet. 
 
 The whistle and exhaust were heard at 7J miles with a feeble oppos- 
 ing wind, and lost at 6J miles when the force of the wind became greater. 
 
 The order of penetration in this series of observations was : 1, trumpet 
 and gong; 2, whistle; 3, exhaust. 
 
 In the case of the vessel sailing northward, its course being almost 
 directly against the wind and in the rear of the trumpet, all the sounds 
 were lost at less distances than in the case of the other vessel. The 
 observations showed very clearly the effect of the wind, the bell at a 
 certain distance being heard indistinctly with a strong opposing wind 
 and more and more plainly as the wind died away. The trumpet was 
 heard only as far as the whistle, the vessel being in the rear of It. 
 
 The third day observations were made from the two vessels, both 
 however sailing to the south. From the vessel sailing at right angles 
 to the direction of the wind the order of penetration was : 1, trumpet ; 
 2, whistle; 3, exhaust; 4, bell. 
 
 In the case of the other vessel the opposing effect of the wind was 
 greater, and the sounds were heard to a less distance; the order was: 
 1. trumpet; 2, whistle; 3, exhaust; 4, bell; 5, rocker. 
 
 On the fourth day two trips were made by the same vessel in the 
 course of the day, one being northward and the other southward. In 
 the first case the trumpet was lost at 3J miles, the vessel being nearly 
 in its rear; in the second case, the wind being almost directly opposed 
 to the sound, the large bell was heard at 1J miles, and lost at J- of a 
 mile, probably due to increase of the force of the wind; the trumpet was 
 was lost at 3^ miles. 
 
 In all these observations, owing to the unfavorable conditions of the 
 locality, and the direction of the wind, we were unable to obtain any 
 satisfactory observations on sound moving with the wind. In all cases 
 the results were obtained from sounds moving nearly against the wind, 
 or at right angles to it. From the results of the whole it appears that 
 the sound was heard farther with a light opposing wind than with a 
 stronger one, and that it was heard farthest of all at right angles to the 
 wind. From this latter fact, however, it should not be inferred that in 
 this case sound could be heard farther at right angles to the wind than 
 with the wind, but that in this direction the effect of the wind was neu- 
 
470 EESEARCHES IN SOUND. 
 
 tralized. The results also exhibited, in a striking manner, the diverg- 
 ency of sound from the axis of the trumpet, the trumpet being heard in 
 the line of its axis in front at 6 miles and behind at 3, the wind being 
 nearly the same in both cases. 
 
 All the observations were repeated on land with the artificial ear as 
 far as the unfavorable condition of the surface would permit. Although 
 the limit, as to distance, at which the sand might be moved was not in 
 most cases observed, yet the relative degree of agitation at a given dis- 
 tance established clearly which was the most powerful instrument, the 
 result giving precisely the same order of penetration of the different in- 
 struments as determined by direct andition. 
 
 During this series of investigations an interesting fact was discovered, 
 namely, a sound moving against the wind, inaudible to the ear on the 
 deck of the schooner, was heard by ascending to the mast-head. This 
 remarkable fact at first suggested the idea that sound was more readily 
 conveyed by the upper current of air than the lower, and this appeared 
 to be in accordance with the following statement of Captain Keeney, 
 who is commander of one of the light-house vessels, and has been for a 
 long time on the banks of Newfoundland in the occupation of fishing : 
 "When the fishermen in the morning hear the sound of the surf to the 
 leeward, or from a point toward which the wind is blowing, they take 
 this as an infallible indication that in the course of from one to five hours 
 the wind will change to the opposite direction from which it is blowing at 
 the time. 77 The same statement was made to me by the intelligent keeper 
 of the fog-signal at Block Island. In these cases it would appear that 
 the wind had already changed direction above, and was thus transmit- 
 ting the sound in an opposite direction to that of the wind at the surface 
 of the earth. 
 
 Another remarkable fact bearing on this same point is established by 
 the observations of General Duane. At Cape Elizabeth, 9 miles south, 
 easterly from the general's house, at Portland, is a fog-signal consisting 
 of a whistle 10 inches in diameter ; at Portland Head, about 4 miles from 
 the same city, in nearly the same direction, is a Daboll trumpet. There 
 can be no doubt, says the general, that those signals can be heard much 
 better during a heavy northeast snow-storm than at any other time. 
 "As the wind increases in force, the sound of the nearer instrument, the 
 trumpet, diminishes, but the whistle becomes more distinct ; but I have 
 never known the wind to blow hard enough to prevent the sound of the 
 latter from reaching this city." In this case, the sound comes to the city 
 in nearly direct opposition to the course of the wind, and the explanation 
 which suggested itself to me was that during the continuance of the 
 storm, while the wind was blowing from the northeast at the surface, 
 there was a current of equal or greater intensity blowing in an opposite 
 direction above, by which the sound was carried in direct opposition to 
 the direction of the surface current. The existence of such an upper 
 current is in accordance with the hypothesis of the character of a north- 
 
RESEARCHES IN SOUND. 471 
 
 east storm, which sometimes rages for several days at a given point on 
 the coast without being felt more than a few miles in the interior, the air 
 continuously flowing in below and going out above. Indeed, in such 
 cases a break in the lower clouds reveals the fact of the existence above 
 of a rapid current in the opposite direction. 
 
 The full significance, however, of this idea did not reveal itself to me 
 until in searching the bibliography of sound I found an account of the 
 hypothesis of Professor Stokes in the Proceedings of the British Associa- 
 tion for 185G, * in which the effect of an upper current in deflecting the 
 wave of sound so as to throw it down upon the ear of the auditor, or 
 directing it upward far above his head, is fully explained. This subject 
 will be referred to in the subsequent parts of the report, in the attempt 
 to explain various abnormal phenomena of sound which have been ob- 
 served during the series of investigations connected with the Light- 
 House Board. 
 
 During these investigations an attempt was made to ascertain the 
 velocity of the wind in an upper stratum as compared with that in the 
 lower. The only important result however, was the fact that the veloc- 
 ity of the shadow of a cloud passing over the ground was much greater 
 than that of the air at the surface, the velocity of the latter being de- 
 termined approximately by running a given distance with such speed 
 that a small flag was at rest along the side of its pole. While this 
 velocity was not perhaps greater than six miles per hour, that of the 
 shadow of the cloud was apparently equal to that of a horse at full 
 speed. 
 
 During this and subsequent investigations, inquiries were made in 
 regard to the effect of fog upon sound, it being a subject of considera- 
 ble importance to ascertain whether waves of sound, like the rays of 
 light, are absorbed or stifled by fog. On this point, however, observers 
 disagree. At first sight, from the very striking analogy which exists in 
 many respects between sound and light, the opinion largely prevails 
 that sound is impeded by fog; although observers who have not been 
 influenced by this analogy have, in many instances, adopted the opposite 
 opinion, that sound is better heard during a fog than in clear weather. 
 For instance, the Kev. Peter Ferguson, of Massachusetts, informs me 
 that from his own observations, sound is conveyed farther in a fog than 
 in a clear air. He founds this opinion on observations which he has 
 made on the sound of locomotives of several railways in passing over 
 bridges at a distance. Unfortunately, the question is a difficult one to 
 settle, since the effect of the wind, in order to arrive at a true result, 
 must be carefully eliminated. Captain Keeney, who has previously 
 been mentioned, related the following occurrence, in the first part of 
 which he was led to suppose that fog had a very marked influence in 
 deadening sound, though in a subsequent part he came to an opposite 
 conclusion : He was sailing during a dense fog, with a slight wind bear- 
 
 * Report of British Association, 1856 ; Abstracts, p. 22. 
 
472 RESEARCHES IN SOUND. 
 
 ing him toward a light-vessel, the locality of which he expected to find 
 by means of the fog-signal. He kept on his course until he thought 
 himself very near the ship, without hearing the stroke of the bell. He 
 then anchored for the night, and found himself next morning within a 
 short distance of the light- vessel, but still heard no sound, although he 
 was assured when he got to it that the bell had been ringing all night. 
 He then passed on in the same direction in which he had previously 
 sailed, leaving the light- vessel behind, and constantly heard the bell for 
 a distance of several miles, the density of the fog not perceptibly dimin- 
 ishing. In this case it is evident that the deadening of the sound was 
 not due to the fog, but, as we shall hereafter see, in all probability to the 
 combined action of the upper and the lower currents of air. 
 
 On returning to Washington the writer took advantage of the occur- 
 rence of a fog to make an experiment as to the penetration of the sound 
 of a small bell rung by clock-work, the apparatus being the part of a 
 moderator-lamp intended to give warning to the keepers when the sup- 
 ply of oil ceased. The result of the experiment was contrary to the sup- 
 position of absorption of the sound tiy the fog, but the change in the 
 condition of the atmosphere as to temperature and the motion of the 
 air, before the experiment could be repeated in clear weather, rendered 
 the result not entirely satisfactory. 
 
 EXPERIMENTS AT SANDY HOOK IN 18G7. 
 
 The next series of experiments was made from October 10 to October 
 18, 1867, under the direction of the writer of this report, in connection 
 with General Poe, engineer-secretary of the Light-House Board, Com- 
 modore (now Admiral) Case, then inspector of the third light-house dis- 
 district, and Mr. Lederle, acting engineer of the same district. 
 
 The principal object of these investigations was to compare different 
 instruments, and to ascertain the improvements which had been made 
 in them since the date of the last investigations, especially the exami- 
 nation of a new fog-signal called the siren, and the comparison of it with 
 the Daboll trumpet, although other investigations were made relative to 
 the general subject of sound in relation to fog-signals. The locality 
 chosen was Sandy Hook, a narrow peninsula projecting northward, 
 about five miles into the middle of the Lower Bay of New York, and al- 
 most at right angles to its coast, having a width of about half a mile. 
 Near the northern point on the east shore a temporary building was 
 erected for the shelter of the engines and other instruments. 
 
 The comparisons in regard to penetrating power were made by the 
 use of the artificial ear, heretofore described, by carrying this off a 
 measured distance until the sand ceased to move. This operation was 
 much facilitated by previous surveys of members of the Engineer 
 Corps, who had staked off a straight line parallel with the shore, and 
 accurately divided it into equal distances of 100 feet. 
 
 On account of the character of the deep and loose sand, walking 
 
RESEARCHES IN SOUND. 473 
 
 along this distance was exceedingly difficult, and, to obviate this, a car- 
 riage with broad wheels, drawn by two horses, was employed. An 
 awning over this vehicle protected the observer from the sun, and ena- 
 bled him, without fatigue and at his ease, to note the agitations of 
 sand on the drum of the artificial ear, the mouth of which was directed 
 from the rear of the carriage toward the sounding instrument. 
 
 For these and other facilities we were indebted to General Humphreys, 
 Chief of the Engineer Bureau, who gave orders to the officer in charge 
 of the military works at Sandy Hook to afford us every aid in his power 
 in carrying on the investigation. 
 
 The instruments employed were 
 
 1st. A first-class Daboll trumpet (the patent for which since the 
 death of Mr. Daboll, is owned by Mr. James A. Eobinson,) operated by 
 an Ericsson hot-air engine. It carried a steel reed 10 inches long, 2| inches 
 wide, and J inch in thickness at the vibrating end, but increasing gradu- 
 ally to an inch at the larger extremity. This was attached to a large 
 vertical trumpet curved at the upper end into a horizontal direction and 
 furnished with an automatic arrangement for producing an oscillation 
 of the instrument of about GO in the arc of the horizon. Its entire 
 length, including the curvature, was 17 feet. It was 3J inches at the 
 smaller end and had a flaring mouth 38 inches in diameter. The engine 
 had a cylinder 32 inches in diameter, with an air-chamber of 4J feet in 
 diameter and 6 feet long, and was able to furnish continually a five-second 
 blast every minute at a pressure of from 15 to 30 pounds. 
 
 2d. A siren, originally invented by Cagniard de Latour, and well 
 known to the physicist as a means of comparing sounds and measuring 
 the number of vibrations in different musical notes. Under the direc- 
 tion of the Light-House Board, Mr. Brown, of New York, had made a 
 series of experiments on this instrument ill reference to its adoption as 
 a fog-signal, and these experiments have been eminently successful. 
 The instrument as it now exists differs in two essential particulars 
 from the original invention of Latour: 1st, it is connected with a 
 trumpet in which it supplies the place of the reed in producing the agi- 
 tation of the air necessary to the generation of the sound ; and 2d, the 
 revolving disk, which opens and shuts the orifices producing the blasts, 
 is driven not by the blast itself impinging on oblique openings, as in 
 the original instrument, but by a small engine connected with the feed- 
 pump of the boiler. 
 
 The general character of the instrument may be understood from the 
 following description: Suppose a drum of short axis, into one head of 
 which is inserted a steam-pipe connected with a locomotive-boiler, while 
 the other end has in it a triangular orifice, through which the steam is 
 at brief intervals allowed to project itself. Ira mediately before this head, 
 and in close contact with it, is a revolving disk, in which are eight 
 orifices. By this arrangement, at every complete revolution of the disk, 
 the orifice in the head of the drum is opened and shut eight times in 
 
474 EESEARCHES IN SOUND. 
 
 succession, thus producing a rapid series of impulses of steam against the 
 air into the smaller orifice of the trumpet placed immediately in front of 
 the revolving disk. These impulses are of such intensity and rapidity 
 as to produce a sound unrivalled in magnitude and penetrating power 
 by that of any other instrument yet devised. 
 
 The siren was operated by an upright cylindrical tubular boiler, with 
 a pressure of from 50 to 100 pounds on the square inch. For this form 
 of, boiler has been subsequently substituted an ordinary horizontal 
 locomotive-boiler with a small engine attached for feeding it and for 
 rotating the disk, the latter being effected by means of a band passing 
 over pulleys of suitable relative dimensions. 
 
 3d. A steam-whistle 8 inches in diameter. Through some misunder- 
 standing a series of whistles of different diameters was not furnished 
 as was intended. 
 
 The first experiments to be noted were those in regard to the com- 
 parison of penetrating power of the siren and the whistle, the fitting up 
 of the Daboll trumpet not having been completed. The principal object 
 of this however was to test again the truthfulness of the indications of 
 the artificial ear in comparison with those of the natural ear. 
 
 An experiment was made both by means of the artificial ear on land 
 and by actually going off on the ocean in a steamer until the sounds be- 
 came inaudible to the natural ear. By the latter method the two sounds 
 ceased to be heard at the distances of six, and twelve and a half miles, 
 respectively. The indications of the artifical ear gave a similar result, 
 the distance at which the sand ceased to move in one case being double 
 that of the other. In both cases the conditions of wind and weather 
 were apparently the same. In the case of the steamer the distance was 
 estimated by noting the interval of time between the flash of steam and 
 the perception of the sound. 
 
 Comparison of the Daboll trumpet and the siren. The pressure of the 
 hot air in the reservoir of the hot-air engine of the trumpet was about 
 20 pounds, and that of the steam in the boiler of the siren about 75 
 pounds. These pressures are however not considered of importance in 
 these experiments, since the object was not so much to determine the 
 relative amount of motive power employed as the amount of penetrat- 
 ing power produced by these two instruments, each being one of the 
 first of its class. 
 
 1. At distance 50 the trumpet produced a decided motion of the sand, 
 while the siren gave a similar result at distance 58. The two observa- 
 tions being made within ten minutes of each other, it may be assumed 
 that the condition of the wind was the same in the two cases, and hence 
 the numbers above given may be taken as the relative penetrating power 
 of the two instruments. 
 
 2. Another series of experiments was instituted to determine whether a 
 high or a low note gave the greatest penetration. For this purpose the 
 
RESEARCHES IN SOUND. 475 
 
 siren was sounded with different velocities of rotation of the perforated 
 disk, the pressure of steam remaining at 90 pounds per square inch. 
 The effect upon the artificial ear in causing greater or less agitation of 
 sand was taken as the indication of the penetrating power of the differ- 
 ent tones. The number of revolutions of the disk in a given time was 
 determined by a counting apparatus, consisting of a train of wheels and 
 a series of dials showing tens, hundreds, and thousands of revolutions ; 
 this was temporarily attached to the projecting end of the spindle of 
 the revolving disk by pushing the projecting axis of the instrument into 
 a hole in the end of the spindle. 
 
 From the whole of this series of experiments it appeared that a revolu- 
 tion which gave 400 impulses in a second was the best with the siren 
 when furnished with a trumpet. On reflection however it was con- 
 cluded that this result might not be entirely due to the pitch, but in 
 part to the perfect unison of that number of impulses of the siren with 
 the natural tone of the trumpet. To obviate this complication, a series 
 of experiments was next day made on the penetration of different pitches 
 with the siren alone, the trumpet being removed. The result was as 
 follows : 
 
 The siren was sounded at five different pitches, the artificial ear be- 
 ing at such a distance as to be near the limit of disturbance by the 
 sound. In this condition the lowest pitch gave no motion of sand. A 
 little higher, slight motion of sand. Still higher, considerable motion 
 of sand ; and with a higher pitch again, no motion of sand. The best 
 result obtained was with a revolution which gave 3GO impulses in a 
 second. 
 
 3. An attempt was made to determine the most effective pitch or tone 
 of the steam-whistle. It was started with what appeared to be the funda- 
 mental note of the bell, which gave slight motion of sand ; a higher 
 tone a better motion ; still higher, sand briskly agitated; next, several 
 tones lower, no motion; higher, no motion; still higher, no motion. The 
 variation in the tone was made by altering the distance between the bell 
 and the orifice through which the steam was ejected. 
 
 The result of this experiment indicated nothing of a definite charac- 
 ter, other than that with a given pressure there is a maximum effect 
 produced when the vibrations of the sheet of air issuing from the circu- 
 lar orifice are in unison with the natural vibrations from the cavity of 
 the bell, a condition which can only be determined in any case by actual 
 experiment. In practice, Mr. Brown was enabled to produce the best 
 effect by regulating the velocity until the trumpet gave the greatest 
 penetrating power, as indicated by an artificial ear of little sensibility, 
 in order that it might be employed for determining the relative power 
 while the observer was but a few yards from the machine. 
 
 These experiments have been made in an apartment of less than 80 
 feet in length, in which the sounding apparatus was placed at one end 
 
476 EESEARCHES IN SOUND. 
 
 and the artificial ear at the other, substituting fine shot instead of sand. 
 
 The experiments with the siren however indicate the fact that neither 
 the highest nor the lowest pitch of an instrument gives the greatest pen- 
 etrating power, but one of a medium character. 
 
 Another element of importance in the construction of these instru- 
 ments is the volume of sound. To illustrate this, it may be mentioned 
 that a harpsichord- wire stretched between two strings of India rubber, 
 when made to vibrate by means of a fiddle-bow, gives scarcely any 
 appreciable sound. We attribute this to the want of quantity in the 
 aerial wave; for if the same wire be stretched over a sounding-board 
 having a wide area, the effect will be a comparatively loud sound, but 
 of less duration, with a given impulse. It was therefore suggested that 
 the width of the reed in the Daboll trumpet, the form and size of the 
 holes in the disk of the siren, and the circumference of the vibrating 
 sheet of air issuing from the circular orifice of the whistle, would affect 
 the power of the sound. The only means of testing this suggestion is 
 by using reeds of different widths, sirens with disks of different-shaped 
 openings, and whistles of different diameters. In conformity with this 
 view, Mr. Brown has made a series of empirical experiments with open- 
 ings of different forms, which have greatly improved the operation of 
 the siren, while Mr. Wilcox has experimented on several forms of reeds, 
 of which the following is the result : 
 
 The best reed obtained was 2J inches wide, 8 inches long in the vi- 
 brating part, | inch thick at the butt, and J inch thick at the free end. 
 This sounded at a pressure of from 20 to 30 pounds. The thinner reeds 
 gave a sound at a less pressure, from 5 to f pounds, the thicker at from 
 20 to 30 pounds. A reed 8.J- inches long in the vibrating part, 1 inch thick 
 at the butt, f inch thick at the end, and 3 inches wide, did not begin to 
 sound until a pressure of 80 pounds was reached, then gave a sound of 
 a dull character. Another reed of the same width, - inch thick at the 
 butt, and -fa inch at the end, and same length, gave a sound at 75 pounds 
 pressure, but still dull and of little penetrating power. These reeds 
 were evidently too heavy in proportion to their elasticity. These were 
 made without the addition of a trumpet, and therefore to produce the 
 best result when used with a trumpet, the latter must be increased or 
 diminished in length until its natural vibrations are in harmou3 r with 
 those of the former, as will be seen hereafter. General Duane has also 
 made experiments on whistles of different diameters, of which the result 
 will be given. 
 
 Another consideration in regard to the same matter is that of the am- 
 plitude of the oscillations of the tongue or steel reed in its excursion in 
 producing the sound ; the time of oscillation remaining the same, that 
 is, the pitch, the amplitude will depend upon the elasticity of the reed, 
 the power to surmount which will again depend upon the pressure of 
 steam in the boiler, and hence we might infer that an increase of pres- 
 
RESEARCHES IN SOUND. 477 
 
 sure in the boiler with an increase of the elasticity of the reed, every- 
 thing else being the same, would produce an increase in penetrating 
 power. From the general analogy of mechanical effects produced by 
 motive power, we may denote the effect upon the ear by the expression 
 mv*, in which m expresses the mass or quantity of air in motion, and v 
 the velocity of the particles in vibration. 
 
 If this be the expression for the effect upon the ear, it is evident that 
 in case of a very high note the amplitude of the vibration must be so 
 small that the effect would approximate that of a continued pressure 
 rather than that of distinct alternations of pressure, giving a vibrating 
 motion to the drum of the ear. 
 
 4. Next, experiments were made to determine the penetrating power in 
 the case of the siren under different pressures of steam in the boiler. 
 The experiments commenced with a pressure of 100 pounds. The pres- 
 sure at each blast was noted by two observers, and to compare these 
 pressures with the indications of the sand, the time of the blasts was 
 also noted. 
 
 The following are the results : 
 
 Pressure. Relative distances n L , which 
 
 the sand ceased to move. 
 
 100 61 
 
 90 59 
 
 80 58 
 
 70 57 
 
 60 57 
 
 50 56 
 
 40 55 
 
 30 53 
 
 20 51 
 
 From this series of experiments, it appears that a diminution of pres- 
 sure is attended with a comparatively small diminution in the pene- 
 trating power of the siren. 
 
 In regard to this unexpected result of great practical importance, the 
 following appears to be the explanation. It is a well-known principle 
 in aerial mechanics that the velocity of the efflux of air from an orifice 
 in a reservoir does not increase with an increase of condensation, when 
 the spouting is into a vacuum. This is evident when we reflect that 
 the weight of density of the air moving out is increased in proportion 
 to the elasticity or pressure; that is, the increase in the propelling force 
 is proportional to the increase in the weight to be moved, hence the 
 velocity must remain the same. 
 
 In the foregoing experiments with high pressures large in proportion 
 to the resistance of the air, the velocity of efflux should therefore be 
 but little increased with the increase of pressure, and inasmuch as the 
 velocity is the roost important factor in the expression mv' 2 , which indi- 
 cates the effect on the tympanum, the penetrating power of the sound 
 should be in accordance with the above experimental results. 
 
478 RESEARCHES IN SOUND. 
 
 A similar result cannot be expected with the use of the whistle or 
 the trumptt, since in the former the stiffness of the aerial reed depends 
 upon its density, which will be in proportion to the pressure in the 
 boiler, and in the case of the latter no sound can be produced on the 
 one hand unless the pressure be sufficient to overcome the resistance of 
 the reed, and on the other the sound must cease when the pressure is 
 so great as to prevent the recoil of the reed. 
 
 5. An experiment was made to determine the effect of a small whistle 
 inserted kico the side of a trumpet near the small end. The whistle 
 being sounded before and after it was placed in the trumpet, the result 
 was as follows: The penetrating powers were in the ratio of 40: 51, 
 while the tone was considerably modified. From this experiment it 
 appears that a whistle may be used to actuate a trumpet or to exercise 
 the functions of a reed. In order however to get the best results, it 
 would be necessary that the trumpet and whistle should be in unison, 
 but it may be doubted however whether an increase of effect, with a 
 given amount of power, would result from using such an arrangement; 
 it might nevertheless be of advantage in certain cases to direct the 
 sound of a locomotive in a definite direction, and to use a smaller 
 whistle, especially in cities, in which the locomotive passes through long- 
 streets; perhaps in this case the sound might be less disagreeable than 
 that of the naked whistle, which sends its sound-waves laterally with 
 as much force as in the direction of the motion of the engine. 
 
 6. General Poe called attention to the sound produced by the paddle- 
 wheels of a steamer in the ofh'ng at a distance estimated at four and a 
 half miles. The sound was quite distinct when the ears were brought 
 near the surface of the beach. 
 
 In this connection he stated that he had heard the approach of a 
 small steamer on the northern lakes when its hull was still below the 
 horizon, and was even'enabled to designate the particular vessel from 
 among others by the peculiarity of the sound. 
 
 The sound in the case of the steamer is made at the surface of the 
 water, and it might be worth the trouble to try experiments as to the 
 transmission of sound under this condition, and the collection of it by 
 means of ear-trumpets, the mouths of which are near the water, the 
 sound being conveyed through tubes to the ears of the pilot. In order 
 however to determine in this case the direction of the source of sound, 
 two trumpets would be necessary, one connected with each ear, since 
 we judge of the direction of a sound by its simultaneous effects on the 
 two auditory nerves. This suggestion, as well as many others which 
 have occurred in the course of these researches, is worthy of special in- 
 vestigation. 
 
 7. A series of experiments was made to compare trumpets of different 
 materials and forms having the same length and transverse areas, all 
 blown at a pressure of 9 J pounds. 
 
EESEARCHES IN SOUND. 
 The following table gives the results : 
 
 479 
 
 No. 
 
 Material of trumpet. 
 
 Cross-section. 
 
 Relative distances at which 
 the sand ceased to move. 
 
 1 
 
 2 
 3 
 4 
 
 Wood. 
 Brass. 
 Cast iron. 
 Wood. 
 
 Square. 
 Circular. 
 Circular. 
 Circular. 
 
 13 
 23 
 24 
 30 
 
 From these experiments it would appear that the material or elasti- 
 city of the trumpet had little or no effect on the penetrating power of 
 the sound, although the shape appeared to have some effect, the pyri- 
 midal trumpet, or one with square cross-section (No. 1), giving a less re- 
 sult than the conical ones of the same sectional area. A comparison 
 was made between a long straight trumpet and one of the same length 
 curved at its upper end, which gave the same penetrating power with 
 the same pressure. It is probable that a thin metallic trumpet would 
 give greater lateral divergency to the sound, and also a slightly differ- 
 ent tone. 
 
 8. The effect of a hopper-formed reflector was next tried with the 
 whistle, the axis of which was about 5 feet in length, the mouth 6 feet 
 square, and the small end about 18 inches. When the whistle was 
 sounded at the small end of this reflector, the distance at which the 
 sand ceased to move was 51 ; the sound of the same whistle without the 
 reflector ceased to move the sand at 40. The ratio of these distances 
 would have been less with a more sensitive instrument at a greater dis- 
 tance on account^ of the divergency of the rays. 
 
 9. In order to determine the diminution of sound by departing from 
 the axis of the trumpet, a series of experiments was made with a ro- 
 tating trumpet, the axis of which was at first directed along the gradu- 
 ated line of observation, and subsequently deflected from that line a 
 given number of degrees. The following were the results : 
 
 Direction of the trumpet. 
 
 Relative distance at 
 which sand moved. 
 
 Along the line 
 
 26 
 
 Deflected 30 
 
 23 
 
 Deflected 60 fc 
 
 21 
 
 Deflected90o 
 
 18 
 
 Deflected 120 
 
 13 
 
 
 
 These results illustrate very strikingly the tendency of sound to spread 
 on either side of the axis of the trumpet ; had the experiments been made 
 with a more sensitive instrument, and at a greater distance, the effect 
 
480 RESEARCHES IN SOUND. 
 
 would have shown a much greater divergency. It should be observed 
 however that the mouth of the trumpet in this case was 36 inches, which 
 is unusually large. 
 
 From the experiments made near New Haven, and also from those at 
 this station, it appears that the actual amount of power to produce sound 
 of a given penetration is absolutely less with a reed trumpet than with a 
 locomotive whistle. This fact probably finds its explanation in the cir- 
 cumstance that in each of these instruments the loudness of the sound 
 is due to the vibration of the air in the interior of the trumpet and in 
 the bell of the whistle, each of these being a resounding cavity ; and 
 furthermore that in these cavities the air is put in a state of sustained 
 vibration by the undulations of a tongue, in the one case of metal, in 
 the other of air ; and furthermore it requires much more steam to set 
 the air in motion by the tongue of air than by the solid tongue of steel, 
 the former requiring a considerable portion of the motive power to give 
 the current of which it consists the proper degree of stiffness, if I may 
 use the word, to produce the necessary rapidity of oscillation. But what- 
 ever may be said in regard to this supposition, it is evident, in case reli- 
 able hot-air engines cannot be obtained, that the Daboll trumpet may be 
 operated by a steam-engine, although at an increased cost of mainte- 
 nance, but this increase we think will still not be in proportion to the 
 sound obtained in comparison with the whistle. 
 
 Another question which naturally arises, but which has not yet been 
 definitely settled by experment, is whether both the siren and the whis- 
 tle would not, equally with the trumpet, give more efficient results when 
 worked by condensed air than by steam. 
 
 From hypothetical considerations this would appear to be the case, 
 since the intensity of sound depends upon the density of the medium in 
 which it is produced; and as the steam is considerably lighter than air, 
 and as the cavities of all of these instruments are largely filled with steam, 
 the intensity of sound would, on this account, seem to be less than if 
 filled with air. 
 
 At the conclusion of the experiments at Sandy Hook, the siren was 
 adopted as a fog-signal, in addition to the reed-trumpet and the locomo- 
 tive-whistle, to be applied to the more important stations, while large 
 bells were retained for points at which fog-signals were required to be 
 heard at but comparatively small distances. These instruments of the 
 first class being adopted, it became of importance to determine, in actual 
 practice, the cost of maintenance, the best method of working them, and 
 any other facts which might have a bearing on their use. 
 
 But as investigations of this kind would require much time and pecu- 
 liar advantages as to location and mechanical appliances, this matter 
 was referred to General Duane, the engineer in charge of the 1st and 
 2d light-house districts, who had peculiar facilities near his residence, 
 at Portland, Me., in the way of workshops and other conveniences, and 
 who, from his established reputation for ingenuity and practical skill 
 
RESEARCHES IX SOUND. 481 
 
 ill mechanism, was well qualified for the work. The assignment of this 
 duty to General Duane by the Light-House Board was made during my 
 absence in Europe, in 1870, and as my vacation in 1871 was devoted to 
 light-house duty in California, I had no opportunity of conferring with 
 him on the subject until after his experiments were completed. His re- 
 sults are therefore entirely independent of those obtained under my di- 
 rection, and I give them herewith in his own words, with such com- 
 ments as they may suggest and as are necessary to a proper elucidation 
 of the subject. 
 
 EXPERIMENTS AT PORTLAND, ME. 1871, BY GENERAL DUANE. 
 
 The apparatus employed consisted of the first-class siren, first-class Daboll trumpet 
 and steam-whistles of various sizes. 
 
 The points to he decided were : 
 
 1st. The relative power of these machines ; i. e., the distances at which they could be 
 heard under various conditions of the atmosphere. 
 
 2d. The amount of fuel and water consumed by each. 
 
 3d. The attention and skill required in operating them. 
 
 4th. Their endurance. 
 
 5th. Whether they are sufficiently simple in construction to permit of their being 
 managed and kept in running order by the class of men usually appointed light- 
 house keepers. 
 
 In conducting these experiments the following method was pursued : 
 
 The signals were sounded at alternate minutes, and their sound compared at dis- 
 tances of two, three, and four miles, and from different directions. On every occasion 
 the quantity of fuel and water consumed per hour by each was carefully noted, and 
 the condition of each machine examined, both before and after the trial, to ascertain 
 whether any of its parts had sustained injury. 
 
 Before giving the results of these experiments some facts should be stated, which 
 will explain the difficulty of determining the power of a fog-signal. 
 
 There are six steam fog- whistles on the coast of Maine ; these have been frequently 
 heard at a distance of twenty miles, and as frequently cannot be heard at the distance 
 of two miles, and this with no perceptible difference in the state of the atmosphere. 
 
 The signal is often heard at a great distance in one direction, while in another it 
 will be scarcely audible at the distance of a mile. This is not the effect of wind, as 
 the signal is frequently heard much farther against the wind than with it. For ex- 
 ample, the whistle on Cape Elizabeth can always be distinctly heard in Portland, a 
 distance of nine miles, during a heavy northeast snow-storm, the wind blowing a gale 
 dieectly from Portland toward the whistle, 
 
 [In this sentence, General Duane certainly does not intend to convey 
 the idea that a signal is frequently heard " at a much greater distance 
 against the wind than with it," since this assertion would be at variance 
 with the general experience of mankind; but the word " frequently " 
 applies to the whistle on Cape Elizabeth, which has been already men- 
 tioned as a remarkably exceptional case, in which the sound is heard 
 best against the wind during a northeast snow-storm.] 
 
 The most perplexing difficulty, however, arises from the fact that the signal often 
 appears to be surrounded by a belt, varying in radius from one to one and a half miles, 
 from which the sound appears to be entirely absent. Thus, in moving directly from 
 a station, the sound is audible for the distance of a mile, is then lost for about the 
 same distance, after which it is again distinctly heard for a long time. This action is 
 common to all ear-signals, and has been at times observed at all the stations, at one 
 of which the signal is situated on a bare rock twenty miles from the main -land, with 
 no surrounding objects to affect the sound. 
 
 All attempts to re-enforce the sound by means of reflectors have hitherto been un- 
 successful. Upon a large scale, sound does not appear, on striking a surface, to be 
 reflected after the manner of light and heat, but to roll along it like a clpud of smoke. 
 S. Ms. 52 31 
 
482 RESEARCHES IN SOUND. 
 
 [This statement is in a measure in accordance with results which I 
 have previously found in connection with investigations at the light- 
 house near New Haven, in which the conclusion was arrived at, that 
 although rays of feeble sounds, and for a short distance, observe the 
 law that the angle of reflection is equal to the angle of incidence after 
 the manner of light, yet powerful sounds tend to diverge laterally to 
 such a degree as to render reflectors of comparatively little use.] 
 
 In view of these circumstances, it will be obvious that it was extremely difficult to 
 determine the extent of the power of the various signals under examination. 
 
 It should be remembered that while the sound from the whistle is equally distrib- 
 uted in all directions,* that from the two other signals, both of which tire provided 
 with trumpets, is not so distributed. 
 
 [The difference is apparent near by, but, as we have seen before, on 
 account of the tendency of sound to spread it is imperceptible at a dis- 
 tance.] 
 
 In the siren the sound is most distinct in tbjs axis of the trumpet. 
 
 In the Daboll trumpet it is usually strongest in a plane perpendicular to this axis. 
 
 [This is at variance directly with any observation I have myself made.] 
 
 Relative power. From the average of a great number of experiments the following 
 result was obtained : 
 
 The power of the first-class siren, 12" whistle, and first-class Daboll trumpet, may 
 be expressed by the numbers 9, 7, 4. 
 
 The extreme limit of sound of the siren was not ascertained. That of the 12" whistle 
 is about twenty miles, and of the trumpet twelve. 
 
 Consumption of fuel and water. The siren, when working with a pressure of 72 pounds 
 of steam, consumes about 180 pounds of coal and 126 gallons of water per hour. 
 
 The 12" whistle, with 55 pounds pressure of steam, consumes CO pounds of coal and 
 40 gallons of water per hour. 
 
 The Daboll trumpet, with 10 pounds pressure of air in the tank, consumes about 20 
 pounds of coal per hour. 
 
 The relative expenditure of fuel would be : siren, 9 ; whistle, 3 ; trumpet, 1. 
 
 The siren. Of the three machines this is the most complicated. It uses eteam at a 
 high pressure, and some of its parts move with very great velocity, the siren spindle 
 making from 1,800 to 2,400 revolutions per minute. The boiler must be driven to its 
 full capacity in order to furnish sufficient steam. A large quantity of strain is, at 
 intervals, suddenly drawn from the boiler, causing a tendency to foain, and to eject a 
 considerable amount of water through the trumpet. 
 
 The constant attention of the keeper is required to regulate the fire, the supply of 
 water to the boiler, of oil to the journals, &c. 
 
 In general terms, it may be stated that the siren requires more skill and attention 
 in its management than either of the other signals. 
 
 The Daboll trumpet. As the caloric engine, which has been hitherto employed to 
 operate this signal, requires little fuel, no water, and is perfectly safe as regards 
 danger from explosion, it would, at the first glance, appear to be the most suitable 
 power that could be applied to fog-signals, and was accordingly at first exclusively 
 adopted for this purpose. It was however found to be so liable to accident and so 
 difficult to repair that of late years it has been almost entirely rejected. In the steam- 
 boiler the furnace is surrounded by water, and it is impossible, under ordinary circum- 
 stances, to heat the metal much above the temperature of the water. The furnace of 
 the caloric engine is surrounded by air, and is therefore liable to be burned out if the 
 fire is not properly regulated. 
 
 The working-piston is packed with leather, and as it moves horizontally, with ita 
 whole weight resting on the lower side of the cylinder, the packing at its lower edge 
 is soon worn out. 
 
 If the engine is allowed to stop with the piston at the furnace-end of the cylinder, 
 
 *The sound of the whistle is equally distributed horizontally. It is, however, 
 much stronger in the plane containing the lower edge of the bell than on either side 
 of this plane. Thus, if the whistle is standing upright, in the ordinary position, its 
 sound is more distinct in a horizontal plane passing through the whistle than above 
 or below it. 
 
EESE ARCHES IN SOUND. 483 
 
 the leather is destroyed by the heat. The repacking of a piston is a difficult and 
 expensive operation, requiring more skill than can be expected among the class of 
 men from whom light-house keepers are appointed. 
 
 Another accident to which these engines are subject arises from a sudden check in 
 the velocity of the piston, caused either by the jamming of the leather packing or the 
 introduction of dirt into the open end of the cylinder, in which case the momentum of 
 the heavy, eccentrically-loaded fly-wheel is almost sure to break the main rocker-shait. 
 
 The expense of repairs is considerably increased by the fact that these engines are 
 not now in general use, and when important repairs are required it is usually neces- 
 sary to send to the manufacturer. 
 
 This signal requires much attention. The fires must be carefully regulated to avoid 
 burning out the furnace, the journals thoroughly oiled, and the cylinders well sup- 
 plied with tallow. 
 
 The steam-whistle. This machine requiring much less steam than the siren in propor- 
 tion to the size of its boiler, there is not the same necessity for forcing the fire ; the 
 pressure of steam required is less, and the point from which it is drawn much higher 
 above the water-level in the boiler, and there is consequently no tendency to foam. 
 
 The machinery Ls simple ; the piston pressure very light, producing but little strain 
 on the different parts of the engine, which is therefore not liable to get out of order 
 and requires no more attention than a common stationary engine. 
 
 One marked advantage possessed by this signal is that should the engine become 
 disabled, the whistle may still be sounded by working the valve by hand. This is not 
 the case with the two others, where an accident to any part of the machinery renders 
 the signal for the time useless. 
 
 It will thus be seen that the siren is the most expensive of the fog-signals as regards 
 maintenance, and that it is adapted only to such stations as are abundantly supplied 
 with water and situated in the vicinity of machine-shops where the necessary repairs 
 can be promptly made. 
 
 On the other hand, as it is the most powerful signal, there are certain stations where 
 it should have the preference ; as, for example, Sandy Hook, which from its import- 
 ance demands the best signal that can be procured, regardless of cost. Such stations 
 should be provided with duplicate apparatus, well supplied with spare parts, to guard 
 against any possibility of accident. 
 
 There should be a keeper whose sole business must be to attend the signal, and who 
 should have sufficient mechanical skill to make the ordinary repairs. He should 
 moreover be a licensed engineer. 
 
 There will also be required an assistant, who may be one of the light-keepers, to 
 relieve him during the continuance of foggy weather. 
 
 The steam- whistle is the simplest in construction, most easily managed and kept in 
 repair, and requires the least attention of all the fog-signals. It is sufficiently power- 
 ful for most localities, while its consumption of fuel and water is moderate. 
 
 It has been found on this coast that a sufficient quantity of rain-water can be col- 
 lected to supply the 12" whistle at nearly every station. This has been the case for 
 the last two years at Martinicus. 
 
 The Daboll trumpet, operated by a caloric engine, should only be employed in ex- 
 ceptional cases, such as at stations where no water can be procured, and where, from, 
 the proximity of other signals, it may be necessary to vary the nature of the sound. 
 
 The trumpet however may undoubtedly be very much improved by employing 
 steam power for condensing the air. The amount of work required, which is that of 
 compressing 70 cubic feet of air to an average pressure of 8 pounds per inch, would be 
 less than two-horse power. For this purpose the expenditure of fuel and water would 
 be moderate ; indeed, the exhaust steam could be condensed and returned to the cis- 
 tern, should the supply of water be limited. 
 
 The siren also is susceptible of improvement, especially as regards simplification. 
 
 [In the foregoing remarks we think the general has expressed a some- 
 what undue partiality for the whistle, and somewhat overestimated the 
 defects of the other instruments. The trumpets, with Ericsson engine, 
 have not been abandoned, except partially in the two districts under 
 the direction of General Duane, to which he probably intended to con- 
 fine his statement. They are still in use in the third district, where 
 they are preferred by General Woodruff, who finds no diflicultyin keep- 
 ing them in repair, having employed a skilled machinist who has made 
 these instruments his special study, and who, visiting them from time 
 to time, makes repairs and supplies new parts.] 
 
484 KESEARCHES IN SOUND. 
 
 Th^ intermittent action of fog-signals makes it necessary to employ a peculiar form 
 of boiler. The steam used is at a high pressure, and drawn off at intervals; conse- 
 quently there is a tendency to foam .and throw out water with the steam. To obviate 
 this difficulty the form of boiler found by experience to be best adapted to this service 
 is a horizontal tubular boiler (locomotive), with rather more than one-half of the inte- 
 rior space allowed for steam-room. The steam-dome is very large, and is surmounted 
 by a steam pipe 12" in diameter. Both the dome and pipe were formerly made much 
 smaller, but were gradually enlarged as long as any difficulty with regard to foaming 
 was noticed. The steam is drawn off at a point 10" above the water-level in the 
 boiler. The main points to be observed are to have plenty of steam-room, and to draw 
 the steam from a point high above the water-level. It will be readily perceived that 
 a vertical tubular boiler is entirely unsuited to this work. 
 
 It is essential, both as regards economy of fuel and the efficient working of the sig- 
 nal, that the boiler, including the dome and stand-pipe, should be well covered with 
 some good non-conductor of heat. A material, called salamander felting, manufac- 
 tured in Troy, N. Y., was used on the fog- whistle boiler at House Island during the 
 winter of 1870. There resulted a saving of more than 20 per cent, of fuel over that 
 consumed in the same boiler when uncovered. Where this material cannot be pro- 
 cured, a thick layer of hair felting, covered with canvas, will be found to answer a 
 good purpose. 
 
 Various expedients have been proposed with the view of keeping the water in the 
 boilers hot when the signals are not in operation, that the signal may always be ready 
 to sound at a very short notice, and that the water in the boiler and pipes may be 
 prevented from freezing in extremely cold weather. One of these contrivances is 
 " Button's circulating water-heater." It consists essentially of a small, vertical, tubu- 
 lar boiler, entirely filled with water, and connected with the boiler or tank which 
 contains the water to be heated, by two pipes on different levels. As soon as the 
 water in the heater is warmed, a circulation commences, the hot water flowing through 
 the upper pipe into the boiler, and the cold through the lower pipe from the boiler to 
 the heater. As the furnace in the heater is very small but little fuel is consumed, and 
 nearly the entire heat produced by the combustion is utilized. 
 
 The apparatus has been extensively employed in heating the water in tanks designed 
 frr filling the steam fire-engine boilers, when the alarm of fire is first given, and appears 
 admirably adapted to this purpose. If used in connection with a steam-boiler, it should 
 be disconnected before steam is raised in the latter, as, from its construction, it is not 
 calculated to withstand any considerable pressure. 
 
 An arrangement, similar in principle, has been used in the first light-house district, 
 consisting of a small cylinder coal-stove, of the ordinary pattern, around the interior 
 of which, and above the grate, is introduced a single coil of f" pipe. This coil is 
 connected with the boiler by two pipes, one entering near the bottom, the other about 
 2 feet higher. It has been found that in consequence of the rapid circulation of the 
 water through this coil, and the great capacity of water for heat, that nearly all the 
 heat from the fire in the stove is transferred to the water in the boiler. This arrange- 
 ment possesses the advantage of the " pipe, being strong enough to stand any press- 
 ure that can be used in the boiler, rendering it unnecessary to disconnect it at any time. 
 
 Experience has, however, proved that none of these contrivances are essential. It 
 is seldom that an attentive keeper cannot foresee the approach of fog or snow in time 
 to have the apparatus in operation as soon as required, even when obliged to start his 
 fire with cold water in the boiler. 
 
 Keepers should be directed to watch the state of the weather carefully, and to light 
 their tires at the first indication of fog or snow-storm. As soon as the water in the 
 boiler is near the boiling point, should the necessity for sounding the signal have not 
 yet arisen, the fire may be banked, and in this state the water may be kept hot for any 
 length of time at a moderate expenditure of fuel. With proper care, no more fuel is 
 required to keep the water at the requisite temperature by means of a banked fire than 
 by any other method, and it is a matter of great importance to avoid complicating 
 fog-signal apparatus by unnecessary appendages. 
 
 The same plan should be adopted in extremely cold weather to prevent the water 
 iu the boiler from freezing. There should be a small air-cock in the draught-pipe near 
 its junction with the feed-pump, and in cold weather this should be opened when the 
 pump is not in use, in order to allow the pipe to empty itself. 
 
 When the draught-pipe cannot be protected from the cold, and the well is at a con- 
 siderable distance from the engine, the following expedient has been employed with 
 success : The pipe is inclosed in an India-rubber hose of about double its diameter, 
 and from time to time steam is forced through the space between the hose and draught- 
 pipe by means of a small pipe from the boiler. 
 
 Although the laws governing the reflection of light and heat are undoubtedly, in u a 
 great measure, applicable to sound, there are yet so many disturbing influences, such as 
 inflection, refraction, caused by the varying density of the atmosphere, &c., interfering 
 with the reflection of the latter, that but little use can be made of this property in 
 
RESEARCHES IN SOUND. 485 
 
 directing and condensing the waves of sound issuing from a fog-signal. This fact may 
 be illustrated by an account of some experiments made during the last year. 
 
 A whistle being sounded in the focus of a large parabolic reflector, it was very per- 
 ceptible to an observer in the immediate vicinity that the sound was louder in the 
 front than in the rear of the reflector. As the distance of the observer from the whistle 
 was increased this disparity rapidly diminished, and at the distance of a few hundred 
 yards entirely disappeared. The lyeam of sound had been dissipated and the shadow 
 had vanished. The effect of a horizontal sounding-board 10 feet square, suspended over 
 the whistle to prevent the escape of sound in a vertical direction, was inappreciable at 
 the distance of a quarter of a mile. 
 
 The employment of a trumpet with the whistle was rather more successful. The 
 trumpet was constructed of wood, in the form of a square pyramid ; the lower base 
 being 10' by 10', the upper base 2' by 2', and the height 20'. The axis was horizontal 
 and the whistle placed at the smaller end. By this arrangement the increased power 
 of the sound could be perceived at the distance of a mile, the action being similar to 
 that of a speaking-trumpet. 
 
 It is probable that some modification of this form of whistle may be advantageously 
 employed in certain localities, but there is however a disadvantage attending the use 
 of a Irumpet with fog-signals. 
 
 The sound from a trumpet not being uniformly distributed, it is difficult to estimate 
 the distance of the signal, or, as the pilots term it, " to locate the sound." This has 
 been observed in the siren and Daboll trumpet. The sound from these signals being 
 stronger on one course than any other, may be distinctly heard from a vessel when 
 crossing the axis of the beam of sound, but as its distance from this lino increases, the 
 sound appears fainter and more remote, although the vessel may be approaching the 
 signal. 
 
 From an attentive observation, during three years, of the fog-signals on this coast, 
 and from the reports received from captains and pilots of coasting vessels, I am con- 
 vinced that in some conditions of the atmosphere the most powerful signals will be 
 at times unreliable. 
 
 Now it frequently occurs that a signal, which under ordinary circumstances would 
 be audible at the distance of fifteen miles, cannot be heard from a vessel at the dis- 
 tance of a single mile. This is probably due to the reflection mentioned by Hum- 
 boldt. 
 
 The temperature of the air over the land where the fog-signal is located, being very 
 different from that over the sea, the sound, in passing from the former to the latter, 
 undergoes reflection at their surface of contact. The correctness of this view is ren- 
 dered more probable by the fact that when the sound is thus impeded in the direction 
 of the sea, it has been observed to bo much stronger inland. 
 
 When a vessel approaches a signal in a fog, a difficulty is sometimes experienced in 
 determining the position of the signal by the direction from which the sound appears 
 to proceed, the apparent and true direction being entirely different. This is undoubt- 
 edly due to the Detraction of sound passing through media of different density. 
 
 Experiments and observation lead to the conclusion that these anomalies in the pen- 
 etration and direction of sound from fog-signals are to be attributed mainly to the want 
 of uniformity in the surrounding atmosphere, and that snow, rain, fog, and the force and 
 direction of the wind, have much less influence than has generally been supposed. 
 
 [In the foregoing I differ entirely in opinion from General Duane as 
 to the cause of extinction of powerful sounds being due to the unequal 
 density of the atmosphere. The velocity of sound is not at all affected 
 by barometric pressure, but if the difference in pressure is caused by a 
 difference in heat, or by the expansive power of vapor mingled with the 
 air, a slight degree of obstruction of sounds may be observed. But this 
 effect we think is entirely too minute to produce the results noted by 
 General Duane, while we shall find in the action of the currents of wind 
 above and below, a true and sufficient cause.] 
 
 The experimental whistles were of the following dimensions, viz: 2|", 3", 4", 5", 6", 
 10", 12", and 18" in diameter. Those of 2|", 3", 5", and 10" were fitted, instead 
 of the ordinary bell, with long cylinders, provided with movable pistons, so that 
 the effective length of the bell could be altered at pleasure. The pitch of the 
 blast was found to vary with the length of the bell, and the power of the whistle 
 with its diameter. The ratio of the power to the diameter was not accurately ob- 
 tained, but it is probable that the extreme range of sound of a whistle is proportional 
 to the square root of its diameter. 
 
486 RESEARCHES IN SOUND. 
 
 [This result, that the pitch varies with the length of the bell, is in 
 conformity with well-established principles of resounding cavities; and 
 that the power should increase with the extent of the aerial reed, the 
 vibrations of which give motion to the resounding air within the cavity, 
 is also, as we have seen, in accordance with hypothetical considerations; 
 but as the density of this stream of steam, and consequently the rapid- 
 ity of its vibrations, depends upon the pressure of the steam in the 
 boiler, a perfect whistle should have the capability of changing its di- 
 mensions, not only in relation to the width of its throat, but also in 
 regard to the pressure of the steam in the reservoir.] 
 
 The pitch giving the greatest range appears to he at the middle of the scale of sound. 
 It is certain that a good result cannot be obtained from either a very shrill or a bass 
 note. This remark is applicable to all varieties of signal. 
 
 The 10" and 12" whistles are recommended for ordinary use. The 18" whistle is 
 more powerful, but the increase of power bears too small a proportion to that of the 
 expenditure of fuel to render its employment generally advisable. The best results 
 were obtained by giving the whistle the following proportions : The diameter of the 
 bell equaling two- thirds of its length, and the set of the bell, i. e., the vertical dis- 
 tance of the lower edge above the cup, the one-third to one-fourth of the diameter for 
 a pressure of 50 to 60 pounds of steam 
 
 A bell, whether operated by hand or by machinery, cannot be considered an efficient 
 fog-signal on the sea-coast. In calm weather it cannot be heard half the time at a 
 greater distance than one mile, while in rough weather the noise of the surf will 
 drown its sound to seaward altogether. 
 
 Ou approaching a station I have frequently seen the bell rung violently by the 
 keeper, without being able to hear the sound until I had landed. 
 
 Nevertheless, all important stations should be provided with bells, as there are occa- 
 sions when they may serve a useful purpose, but it should be well understood by mari- 
 ners that they must not expect always to hear the bells as a matter of course. 
 
 Bells should not be omitted at stations furnished with steam fog-signals, especially 
 when the latter are not in duplicate, and mariners should be warned that the bell will 
 be sounded when the regular signal is disabled. 
 
 It has been observed that a bell rung by hand can be heard further than when 
 sounded by machinery, and many of the steamboat companies on this coast pay the 
 keepers of bells rung by clock-work to ring them by hand when the boats of their 
 line are expected to pas's. 
 
 [We think the difference in the effect of ringing of bells by hand or 
 by machinery is so slight as to be inappreciable except at a short dis- 
 tance. It is true, as I have before observed, that the sound is louder 
 when the mouth of the bell is directed toward the hearer than when the 
 edge is so directed, but on account of the spreading of this sound the 
 effect is lost in a small distance, and indeed in one light-house the bell 
 is permanently placed with the axis of its mouth directed horizontally, 
 and in this position, if the bell Were struck interiorly with a hammer, 
 which would give it a larger vibration than when struck exteriorly, I 
 doubt whether any difference would be observed between the two meth- 
 ods of ringing ; and if any existed it would probably be in favor of the 
 fixed bell rung by machinery.] 
 
 On rivers, narrow channels, and lakes, where the difficulty from the noise of the surf 
 does not exist, this species of signal may be used to advantage, as its maintenance 
 requires but a small expenditure of either money or labor, and by a proper arrange- 
 ment of the machinery the intervals between the strokes of the bell may be so regu- 
 lated as to avoid the danger of confounding the signals, however near together. 
 
 Although a bell may be heard better when sounded by hand than by clock-work, yet 
 in thoroughfares where the signal must be kept'in constant operation during the en- 
 tire continuance of a fog, it would be impracticable to make use of the former method, 
 and recourse must be had to machinery. 
 
RESEARCHES IN SOUND. 487 
 
 In arranging the signal the bell and machinery rnnst be placed as low as possible, 
 as the sound is heard much more plainly on the water when the bell is near its sur- 
 face, and also as the machinery, when thus situated, is steadier and more readily ac- 
 cessible. 
 
 Particulars as to the siren. The boiler of a second-class apparatus is 12 feet long, 42 
 inches in diameter, and has 300 feet heating-surface. The dome is 2 feet in diameter 
 and 3 feet high. 
 
 The cylinder of the engine is 4 inches in diameter and 6 inches stroke. The prolon- 
 gation of the piston-rod forms the plunger of the feed-pump. The main shaft carries 
 three pulleys, the larger driving the siren-spindle; the second, the worm and screw 
 gear ; and the third, the governor. 
 
 In the worm-gear the wheel makes two revolutions per minute, and is provided with 
 a cam, which, acting on a lever, opens the valve, admitting steam through the siren - 
 disks. The cam has such a length as to hold the valve open for about seven seconds. 
 A counter-weight closes the valve as soon as the lever is released by the cam. 
 
 The siren itself consists of a cylindrical steam-chest, closed at one end by a perfor- 
 ated brass plate. The perforations are twelve in number, equidistant from each other, 
 and arranged on the circumference of a circle, whose center is in the axis of the cyl- 
 inder. The other end is closed by a cast-iron head. The heads are connected by a 
 brass pipe, through which the spindle passes. 
 
 The perforated head is covered on the exterior by a brass disk, attached to the 
 spindle, having twelve rectangular notches corresponding to the apertures on the for- 
 mer, and so arranged that by its revolution these apertures are simultaneously opened 
 and closed. The spindle is driven by a belt from the large pulley on the main shaft. 
 This shaft makes 180 revolutions per minute; the epindle, 1,620; and as there are 12 
 apertures in the disks, from each there will issue jets of steam at the rate of 19,440 
 per minute. The sound produced by these impulses may be rendered more or less 
 acute by increasing or diminishing the velocity of revolution. 
 
 The valve and valve-seat are disks similar to those already described, having how- 
 ever four openings instead of twelve. The valve revolves on the brass tube inclosing 
 the siren-spindle, "and is worked by a bevel gear. The trumpet is of cast-iron. 
 
 The Daboll trumpet. The apparatus used in the foregoing experiments is a second- 
 class trumpet, operated by an Ericsson caloric-engine. The air-pump is single-acMng. 
 Its cylinder is 12" in diameter by 12" stroke. The engine makes forty strokes per min- 
 ute. There is a screw-thread raised on the main shaft, which, acting on a wheel, 
 drives a bevel gear, giving motion to a cam- wheel. The latter makes one revolution 
 in two minutes, and is furnished with three equidistant cams. These cams, pressing 
 on the valve-lever, throw the valve open once in forty seconds, admitting the com- 
 pressed air through the reed-chest into the trumpet. 
 
 The quantity of air forced into the tank should be in excess of that needed for the 
 trumpet, the surplus being allowed to escape through a delicate safety-valve. This is 
 necessary to provide against a deficiency in case of leakage, and also to allow the 
 pressure of air to be regulated to accommodate the reed. Each reed requiring a differ- 
 ent pressure, it is necessary to alter the pressure of the valve-spring whenever a reed 
 is changed. 
 
 The first-class trumpet differs only in size from that described. 
 
 The caloric-engine for the first class has a 30" cylinder. The air-pump is 16^" by 
 15" stroke. 
 
 The steam-whistle. The boiler of this machine is that of the siren. On the forward 
 part of the boiler the bed-plate of a small engine is secured by two cast-iron brackets. 
 The cylinder of this engine is 4" by 9". The fly-wheel shaft carries an eccentric, 
 which, acting through a rod and pawl on a ratchet-wheel, gives the required motion 
 to the cam-wheel shaft. 
 
 The cam-wheel, which makes one revolution per minute, is provided with one or 
 more cams, depending on the number of blasts to be given in a minute ; the length of 
 the blast being regulated by that of the cams. 
 
 The valve for admitting the steam into the whistle is a balance-valve, the diame- 
 ters of the two disks being respectively 3" and 2f ", which difference is sufficient to 
 cause the pressure of steam to close the valve tight without requiring too great a force 
 to open it. The valve is worked by a stem attached to the rocker-shaft at the lower 
 part of the steam -pipe. This shaft passes through a stuffing-box in the steam-pipe, 
 and is provided with a collar which the pressure of the steam forces against the inte- 
 rior boss on the pipe, thus making the joint steam-tight. The exterior arm on this 
 rocker-shaft, as well as that on the engine, is perforated in such a manner as to allow 
 the throw of the valve to be adjusted. 
 
488 RESEARCHES IN SOUND. 
 
 In the comments we have made on the report of General Duane, the 
 intention was not in the least to disparage the value of his results, which 
 can scarcely be too highly appreciated ; but inasmuch as the true explana- 
 tion of the phenomena he has observed has an important bearing on the 
 location of fog-signals and on their general application as aids to naviga- 
 tion, and are as well of great interest to the physicist, who values every 
 addition to theoretical as well as practical knowledge, we have not only 
 thought the remarks we have offered necessary, but also that special 
 investigations should be made to ascertain more definitely the condi- 
 tions under which the abnormal phenomena the general has described 
 occur, and to assign, if possible, a more definite and efficient cause than 
 those to which he has attributed them. 
 
 We have, therefore, given much thought to the subject, and since the 
 date of General Duane's report, have embraced every opportunity which 
 occurred for making observations in regard to them. The first step we 
 made toward obtaining a clew to the explanation of the phenomena in 
 question resulted from observatiops at New Haven, namely : 1st, the 
 tendency of sound to spread laterally into its shadow 5 2d, the fact that 
 a sound is frequently borne in an opposite direction to the wind at the 
 surface by an upper current ; and 3d, that a sound moving against a wind 
 is heard better at a higher elevation. The first point to consider is in 
 what manner the wind affects sound. That it is in some way connected 
 with the distance to which sound can be heard is incontestably settled by 
 general observation. At first sight, the explanation of this might seem 
 to be very simple, namely, that the sound is borne on in the one direc- 
 tion and retarded in the other by the motion of the wind. But this ex- 
 planation, satisfactory as it might appear, cannot be true. Sound moves 
 at the rate of about 780 miles an hour, and therefore, on the above sup- 
 position, a wind of 7.8 miles per hour could neither retard nor accele- 
 rate its velocity more than one per cent., an amount inappreciable to 
 ordinary observation ; whereas we know that a wind of the velocity we 
 have mentioned is frequently accompanied with a reduction of the pene- 
 trating power of sound of more than 50 per cent. 
 
 The explanation of this phenomenon, as suggested by the hypothesis ot 
 Professor Stokes, is founded on the fact that in the case of a deep current 
 of air the lower stratum, or that next the earth, is more retarded by fric- 
 tion than the one immediately above, and this again than the one above it, 
 and so on. The effect of this diminution of velocity as we descend toward 
 the earth is, in the case of sound moving with the current, to carry the 
 upper part of the sound-waves more rapidly forward than the lower parts, 
 thus causing them to incline toward the earth, or, in other words, to be 
 thrown down upon the ear of the observer. When the sound is in a 
 contrary direction to the current, an opposite effect is produced, the 
 upper portion of the sound-waves is more retarded than the lower, which 
 advancing more rapidly, in consequence inclines the waves upward and 
 directs tjiem above the head of the observer. To render this more clear, 
 
RESEARCHES IN SOUND. 489 
 
 let us recall the nature of a beam of sound, in still air, projected in 
 a horizontal direction. It consists of a series of concentric waves per- 
 pendicular to the direction of the beam, like the palings of a fence. 
 Now, if the upper part of the waves has a slightly greater velocity 
 than the lower, the beam will be bent downward in a manner some- 
 what analogous to that of a ray of light in proceeding from a rarer to a 
 denser medium. The effect of this deformation of the wave will be 
 cumulative from the sound-center onward, and hence, although the ve- 
 locity of the wind may have no perceptible effect on the velocity of sound, 
 yet this bending of the wave being continuous throughout its entire 
 course, a marked effect must be produced. 
 
 A precisely similar effect will be the result, but perhaps in a consid- 
 erably greater degree, in case an upper current is moving in an opposite 
 direction to the lower, when the latter is adverse to the sound, and in 
 this we have a logical explanation of the phenomenon observed by 
 General Duane, in which a fog-signal is only heard during the occurrence 
 of a northeast snow-storm. Certainly this phenomenon cannot be ex- 
 plained by any peculiarity of the atmosphere as to variability of density, 
 or of the amount of vapor which it may contain. 
 
 The first phenomenon of the class mentioned by General Duane, which 
 I had the good fortune to witness was in company with Sir Frederick 
 Arrow and Captain Webb, of the Trinity House, London, in their visit 
 to this country in 1872. At the distance of two or three miles from an 
 island in the harbor of Portland, Maine, on which a fog-signal was 
 placed, the sound, which had been distinctly heard, was lost on ap- 
 proaching the island for nearly a mile, and slightly regained at a less 
 distance. On examining the position of the fog-signal, which was sit- 
 uated on the farther side of the island from the steamer, we found it 
 placed immediately in front of a large house with rising ground in the 
 rear, which caused a sound-shadow, into which, on account of the late- 
 ral divergence of the rays, the sound was projected at a distance, but 
 not in the immediate vicinity of $he island. In the same year I made 
 an excursion in one of the light-house steamers, with Captain Selfridge, 
 to an island on the coast of Maine, at which abnormal phenomena were 
 said to have been observed, but on this occasion no variation of the 
 sound was noted, except that which was directly attributable to the 
 wind, the signal being heard much farther in one direction than in the 
 opposite. 
 
490 RESEARCHES IN SOUND. 
 
 PAET II. EEMAEKS ON SOME ABNOEMAL PHENOMENA OF 
 
 SOUND.* 
 
 The communication which I propose to make this evening is brought 
 forward at this time especially on account of the presence of Dr. Tyn- 
 dall, he being connected with the light-house system of Great Britain, 
 while the facts I have to state are connected with the light-house service 
 of the United States, and must therefore be of interest to our distin- 
 guished visitor. The facts I have to present form part of a general re- 
 port to be published by the United States Light-House Board. 
 
 The Light-House Board of the United States has from its first estab- 
 lishment aimed not only to furnish our sea-coast with all the aids to 
 navigation that have been suggested by the experience of other countries 
 and to adopt the latest improvements, but also to enrich the light-house 
 service with the results of new investigations and new devices for the 
 improvement of its efficiency, or, in other words, to add its share to the 
 advance of a system which pertains to the wants of the highest civiliza- 
 tion. 
 
 Among the obstructions to navigation none are more serious, especially 
 on the American coast, than those caused by fogs. 
 
 Fog, as it is well known, is due to the mingling of warmer air sur- 
 charged with moisture with colder air, and nowhere on the surface of 
 the earth do more favorable conditions exist for producing fogs 
 than on both our Atlantic and Pacific coasts. On the Atlantic the 
 cold stream of water from the polar regions in its passage southward, 
 on account of the rotation of the earth, passes close along our eastern 
 coast from one extremity to the other, and parallel to this but opposite 
 in direction, for a considerable distance is the great current of warm 
 water known as the Gulf Stream. Above the latter the air is constantly 
 surcharged with moisture, and consequently whenever light winds blow 
 from the latter across the former, the vapor is condensed into fog, and 
 since in summer along our eastern coast the southerly wind prevails, 
 we have during July, August, and September, especially on the coast 
 of Maine, an almost continuous prevalence of fogs so dense that distant 
 vision is entirely obstructed. 
 
 On the western coast the great current of the Pacific, after having 
 been cooled in the northern regions, in its passage southward gives rise 
 to cold and warm water in juxtaposition, or, in other words, a current of 
 the former through the latter, and hence whenever a wind blows across 
 the current of cold water, a fog is produced. 
 
 From the foregokig statement it is evident that among the aids to 
 navigation fog-signals are almost as important as light-houses. The 
 application however of the science of acoustics to the former is far less 
 advanced than is that of optics to the latter. Indeed, attempts have 
 been made to apply lights of superior penetrating power, as the electric 
 
 * Made before the " Philosophical Society of Washington," December 11, 1872. 
 
RESEARCHES IN SOUND. 491 
 
 and calcium lights, to supersede the imperfect fog-signals in use. When 
 however we consider the fact that the absorptive power of a stratum of 
 cloud, which is but a lighter fog, of not more than two or three miles in 
 thickness, is sufficient to obscure the image of the sun, the intensity of 
 the light of which is greater than that of any artificial light, it must be 
 evident that optical means are insufficient for obviating the difficulty in 
 question. 
 
 The great extent of the portions of the coast of the United States 
 which are subject to fogs renders the investigation of the subject of fog- 
 signals one of the most important duties of the Light-House Board. 
 
 In studying this subject it becomes a question of importance to ascer- 
 tain whether waves of sound, like those of light, are absorbed or stifled 
 by fog 5 on this point however, observers disagree. At first sight, from 
 the very striking analogy which exists in many respects between light 
 and sound, the opinion has largely prevailed that sound is impeded by fog. 
 But those who have not been influenced by this analogy have in some 
 instances adopted the opposite opinion that sound is better heard during 
 a fog than in clear weather. To settle this question definitely the Light- 
 House Board have directed that at two light-houses on the route from 
 Boston to Saint John the fog-signals shall be sounded every day on 
 which the steamboats from these ports pass the station, both in clear 
 and foggy weather, the pilots on board these vessels having, for a small 
 gratuity, engaged to note the actual distance of the boat when the sound 
 is first heard on approaching the signal and is last heard on receding 
 from it. The boats above mentioned estimate their distance with con- 
 siderable precision by the number of revolutions of the paddle-wheel as 
 recorded by the indicator of the engine, and it is hoped by this means 
 to definitely decide the point in question. We think it probable that 
 fog does somewhat diminish the penetrating power of sound, or in 
 other words, produce an effect analogous to that on the propagation 
 of light. But when we consider the extreme minuteness of the parti- 
 cles of water constituting the fog as compared with the magnitude of 
 the waves of sound, the analogy does not hold except in so small a de- 
 gree as to be of no practical importance, or, in other words, the exist- 
 ence of a fog is a true, but, we think, a wholly insufficient cause of di- 
 minution of sound, which view is borne out by the great distance at 
 which our signals are heard during a dense fog. 
 
 Another cause, which without doubt is a true one, of the diminution 
 of the penetrating power of sound is the varying density of the atmos- 
 phere, from heat and moisture, in long distances. The effect of this, 
 however, Avould apparently be to slightly distort the wave of sound 
 rather than to obliterate it. However this may be, we think, from all 
 the observations we have made, the effect is small in comparison with 
 another cause, viz, that of the influence of wind. During a residence of 
 several weeks at the sea-shore, the variation in intensity of the sound 
 
492 RESEARCHES IN SOUND. 
 
 of the breakers at a distance of about a mile in no case appeared to be 
 coincident with the variations of an aneroid barometer or a thermome- 
 ter, but in every instance it was affected by the direction of the wind. 
 
 The variation in the distinctness of the sound of a distant instrument 
 as depending on the direction of the wind is so marked that we are war- 
 ranted in considering it the principal cause of the inefficiency in certain 
 cases of the most powerful fog-signals. The effect of the wind is usu- 
 ally attributed, without due consideration, to the motion of the body of 
 air between the hearer and the sounding instrument ; in the case of its 
 coming towards him it is supposed that the velocity of the sound is re- 
 inforced by the motion of the air, and when in the opposite direction 
 that it is retarded in an equal degree. A little reflection, however, will 
 show that this cannot be the cause of the phenomenon in question, since 
 the velocity of sound is so vastly greater than that of any ordinary 
 wind that the latter can only impede the progress of the former by a 
 very small percentage of the whole. Professor Stokes, of Cambridge 
 University, England, has offered a very ingenious hypothetical explana- 
 tion of wind on sound, which we think has an important practical bear- 
 ing, especially in directing the line of research and subsequent applica- 
 tion of principles. 
 
 His explanation rests upon the fact that during the passage of a wind 
 between the observer and the sounding instrument the velocity of this 
 will be more retarded at the surface of the earth on account of friction 
 and other obstacles, and that the velocity of the stratum immediately 
 above will be retarded by that below, and so on, the obstruction being 
 lessened as we ascend through the strata. From this it follows that the 
 sound wave will be deformed and the direction of its normal changed. 
 Suppose, for example, that the wind is blowing directly from the olS- 
 server. In this case the retardation of the sound wave will be greater 
 above than below, and the upper part of the wave-front will be thrown 
 backwards so that the axis of the phonic ray will be deflected upwards, 
 and over the head of the observer. If, on the other hand, a deep river 
 of wind (so to speak) is blowing directly towards the observer, the upper 
 part of the front of the wave will be inclined down and towards him, 
 concentrating the sound along the surface of the earth. 
 
 The science of acoustics in regard to the phenomena of sound as ex- 
 hibited in limited spaces has been developed with signal success. The 
 laws of its production, propagation, reflection, and refraction have been 
 determined with much precision, so that we are enabled in most cases 
 to explain, predict, and control the phenomena exhibited under given 
 conditions. But in case of loud sounds and those which are propagated 
 to a great distance, such as are to be employed as fog-signals, consider- 
 able obscurity still exists. As an illustration of this I may mention the 
 frequent occurrence of apparently abnormal phenomena. General War- 
 ren informs me that at the battle of Seven Pines, in June, 1862, near 
 Bichmond, General Johnston, of the Confederate army, was within three 
 
RESEARCHES IN SOUND. 493 
 
 miles of the scene of action with a force intended to attack the flank of 
 the Northern forces, and although listening attentively for the sound of 
 the commencement of the engagement, the battle, which was a severe 
 one, and lasting about three hours, ended without his having heard a 
 single gun. (See Johnston's report.) Another case of a similar kind 
 occurred to General McClellan at the battle of Games' Mills, June 27 7 
 1862, also near Richmond. Although a sharp engagement was progress- 
 ing within three or four miles for four or five hours, the general and his 
 staff were unaware of its occurrence, and when their attention was 
 called to some feeble sound they had no idea that it was from anything 
 more than a skirmish of little importance. (See Eeport of the Commit- 
 tee on the Conduct of the War.) A third and perhaps still more re- 
 markable instance is given in a skirmish between a part of the Second 
 Corps under General Warren and a force of the enemy. In this case 
 the sound of the tiring was heard more distinctly at General Meade's 
 headquarters than it was at the headquarters of the Second Corps itself, 
 although the latter was about midway between the former and the point 
 of conflict. Indeed the sound appeared so near General Meade's camp 
 that the impression was made that the enemy had gotten between it and 
 General Warren's command. In fact so many instances occurred of wrong 
 impressions as to direction and distance derived from the sound of guns 
 that little reliance came to be placed on these indications. 
 
 In the report of a series of experiments made under the direction of 
 the Light-House Board by General Duane of the Engineer Corps is the 
 following remark : " The most perplexing difficulty arises from the fact 
 that the fog-signal often appears to be surrounded by a belt varying in 
 radius from one to one and a half miles. Thus in moving directly from 
 a station the sound is audible for the distance of a mile, is then lost for 
 about the same distance, after which it is again distinctly heard for a 
 long time." 
 
 Again, in a series of experiments at which Sir Frederick Arrow and Cap- 
 tain Webb, of the Trinity Board, assisted, it was found that in passing in 
 the rear of the opposite side of an island in front of which a fog-signal 
 was placed, the sound entirely disappeared, but by going further off to 
 the distance of two or three miles it reappeared in full force, even with 
 a large island intervening. Again, from the experiments made under the 
 immediate direction of the present chairman of the Light-House Board, 
 with the assistance of Admiral Powell and Mr. Lederle, the light- 
 house engineer, and also from separate experiments made by General 
 Duane, it appears that while a reflector, in the focus of which a steam 
 whistle or ordinary bell is placed, reinforces the sound for a short dis- 
 tance, it produces little or no effect at the distance of two or three miles, 
 and, indeed, the instrument can be -as well heard in still air at the dis- 
 tance of four or five miles in the line of the axis of the reflector, whether 
 the ear be placed before or behind it. From these results we would 
 infer that the lateral divergency of sound, or its tendency to spread lat- 
 
494 RESEARCHES IN SOUND. 
 
 erally as it passes from its source, is much greater than has been sup* 
 posed from experiments on a small scale. The idea we wish to convey 
 by this is that a beam of sound issuing through an orifice, although at 
 first proceeding like a beam of light in paralled rays, soon begins to 
 diverge and spread out into a cone, and at a sufficient distance may in- 
 clude even the entire horizon. 
 
 We may mention also in this connection that from the general fact 
 expressed by the divergence of the rays of sound, the application of 
 reflection as a means of reinforcing sound must in a considerable degree 
 of necessity be a failure. 
 
 By the application of the principle we have stated and the effect of 
 the wind in connection with the peculiarities of the topography of a 
 region and the position of the sounding body, we think that not only 
 may most of the phenomena we have just mentioned be accounted for, 
 but also that other abnormal effects may be anticipated. 
 
 In critically examining the position of the sounding body in the experi- 
 ment we have mentioned, in which Sir Frederick Arrow and Captain Webb 
 assisted, it was found that the signal was placed on the side, of a bank 
 with a large house directly in the rear, thereof of which tended to deflect 
 the sound upwards so as to produce in the rear a shadow, but on account 
 of the divergency of the beam this shadow vanished at the distance of 
 a mile and a half or two miles, and at the distance of, say, three miles 
 the sound of the instrument was distinctly heard. I doubt not that, on 
 examination, all the cases mentioned by General Duane, with one ex- 
 ception, might be referred to the same principle, the exception being 
 expressed in the following remarkable statement in his report to the 
 Light-House Board: " The fog-signals have frequently been heard at a 
 distance of twenty miles and as frequently cannot be heard at the distance 
 of two miles, and with no perceptible difference in the state of the atmos- 
 phere. The signal is often heard at a greater distance in one direction, 
 while in another it will be scarcely audible at the distance of a mile. 
 For example, the whistle at Cape Elizabeth can always be distinctly heard 
 in Portland a distance of nine miles during a heavy northeast snow- 
 storm, the wind blowing a gale nearly from Portland towards the whistle." 
 
 This is so abnormal a case, and so contrary to generally received opinion, 
 that I hesitated to have it published under the authority of the board 
 until it could be verified and more thoroughly examined. In all the 
 observations that have been made under my immediate supervision, the 
 sound has always been heard farther with the wind than against it. It 
 would appear, therefore, from all the observations that the normal effect 
 of the wind is to diminish the sound in blowing directly against it. 
 
 There is however a meteorological condition of the atmosphere dur- 
 ing a northeast storm on our coast which appears to me to have a 
 direct bearing on the phenomenon in question. It is this : that while a 
 violent wind is blowing from the northeast into the interior of the country, 
 a wind of equal intensity is blowing in an opposite direction at an ele- 
 
RESEARCHES IN SOUND. 495 
 
 vation of a mile or two. This is shown by the rapid east wardly motion 
 of the upper clouds as occasionally seen through breaks in the lower. 
 
 As a further illustration of this principle I may mention that on one 
 occasion (in 1855) I started, on my way to Boston from Albany, in the morn- 
 ing of a clear day, with a westerly wind. The weather continued clear 
 and pleasant until after passing the Connecticut Eiver, and until within 
 fifty miles of Bosfon. We then encountered a storm of wind and rain 
 which continued until we reached the city. On inquiry I learned that 
 the storm had commenced in Boston the evening before, and, although 
 the wind had been blowing violently towards Albany for twenty hours, 
 it had not reached inwardly more than fifty miles. At this point it met 
 the west wind and was turned back above in almost a parallel current. 
 This is the general character of northeast storms along our coast, as 
 shown by Mr. Espy, and is directly applicable to the phenomenon men- 
 tioned by General Duane, and which, from the frequency with whicli he 
 has witnessed the occurrence, we must accept as a fact, though by no 
 means a general one applicable to all stations. While a violent wind was 
 blowing towards his place of observation from Cape Elizabeth, at the 
 surface of the earth, a parallel current of air was flowing above with 
 equal or greater velocity in the opposite direction. The effect of the 
 latter would be to increase the velocity of the upper part of the wave 
 of sound, and of the former to diminish it ; the result of the two being 
 to incline the front of the wave of sound towards the observer, or to 
 throw it down towards the earth, thus rendering the distant signal audi- 
 ble under these conditions when otherwise it could not be heard. I think 
 it is probable that the same principle applies in other cases to the abnor- 
 mal propagation of sound. 
 
 For the production of a sound of sufficient power to serve as a fog- 
 signal, bells, gongs, &c., are too feeble except in special cases where the 
 warning required is to be heard only at a small distance. After much 
 experience, the Light-House Board has adopted, for first-class signals, 
 instruments actuated by steam or hot-air engines, and such only as de- 
 pend upon the principle of resonance, or the enforcement of sound by a 
 series of recurring echoes in resounding cavities. 
 
 Of these there are three varieties. First, the steam-whistle, of which 
 the part called the bell is a resounding cavity, the sound it emits having 
 no relation to the material of which it is composed ; one of the same 
 form and of equal size of wood producing an effect identical with that 
 from one of metal. Another variety is the fog-trumpet, which consists 
 of a trumpet of wood or metal actuated by a reed like that of a clario- 
 net. Tlte third variety is called the siren trumpet, which consists of a 
 hollow drum, into one head of which is inserted a pipe from a steam- 
 boiler, while in the other head a number of holes are pierced, which are 
 alternately opened and shut by a revolving plate having an equal num- 
 ber of holes through it. This drum is placed at the mouth of a large 
 trumpet. The sound is produced by the series of impulses given to the 
 
496 EESEARCHES IN SOUND. 
 
 air by the opening and shutting of the orifices and consequent rushing out 
 at intervals with explosive violence of the steam or condensed air. The 
 instrument, as originally invented by Cagniard de Latour, of France, 
 was used simply in experiments in physics to determine the pitch of 
 sound $ but Mr. Brown, of New York, after adding a trumpet to it, and 
 modifying the openings in the head of the drum and the revolving plate, 
 offered it to the Light-House Board as a fog-signal, and as such it has 
 been found the most powerful ever employed. 
 
 In ascertaining the penetrating power of different fog-signals, I have 
 used with entire success an instrument of which the following is a 
 description: A trumpet of ordinary tinned iron of about 3 feet in 
 length, and 9 inches in diameter at the larger end, and about 1 inch at 
 the smaller, is gradually bent so that the axis of the smaller part is at 
 right angles to the axis of the larger end 5 on the smaller end is soldered 
 a cgne, of which the larger end is about 2 inches in diameter. Across 
 the mouth of this cone is stretched a piece of gold-beater's skin. When 
 the instrument is used, the opening on the larger end is held before 
 the iustrment to be tested, the membrane being horizontal, and the mouth 
 of the trumpet vertical ; over the membrane is strewed a small quantity 
 of fine sand, which is defended from the agitation of the air by a cylinder 
 of glass, the upper end of which is closed by a lens. When the instru- 
 ment under examination is sounded, being sufficiently near the sand, is 
 agitated ; it is then moved further off, step by step, until the agitation 
 just ceases ; this distance, being measured, is taken as the relative pene- 
 trating power of the sounding instrument. The same process is repeated 
 with another sounding instrument, and the distance at which the sound 
 ceases to produce an effect on the sand is taken as the measure of the 
 penetrating power of this instrument, and so on. On comparing the re- 
 sults given by this instrument with those obtained by the ear on going out 
 a sufficient distance, the two are found to agree precisely in their indica- 
 tions. The great advantage in using this contrivance is that the rela- 
 tive penetrating power of two instruments may be obtained within a 
 distance of a few hundred yards, while to compare the relative power 
 of two fog-signals by the ear requires the aid of a steamer and a de- 
 parture from the origin of sound in some cases of 15 or 20 miles. 
 
 PART III. INVESTIGATIONS DURING 1873 AND 1874.* 
 
 OBSERVATIONS ON SOUND AND FOG-SIGNALS, IN AUGUST, 1873. 
 
 
 Professor Henry, chairman, and Commander Walker, naval secretary 
 
 of the Light-House Board, left Portland August 12th, 1873, at 3 o'clock 
 p. M. in the steam-tender Myrtle, Captain Foster, for Whjtehead light- 
 station, at which place abnormal phenomena of sound had been observed. 
 
 * From the Report of the Light- House Boajxl,, foj; 1874. 
 
RESEARCHES IN SOUND. 497 
 
 WJiitehead light-station is on a small island about a mile and a half 
 from the coast of Maine, on the western side of the entrance to Penob- 
 scot Bay, and in the direct line of the coasting-steamers and other ves- 
 sels from the westward bound into the Penobscot Bay and Biver. The 
 light-house and fog-signal are situated on the southeast slope of the 
 island, the surface of which consists almost entirely of rock, the middle 
 being at an elevation of 75 feet above the mean tide-level. 
 
 The phenomena which had been observed at this and other stations 
 along the coast consisted of great variation of intensity of sound while 
 approaching and receding from the station. As an example of this we 
 may state the experience of the observers on board the steamer City of 
 Bichmond on one occasion, during a thick fog in the night in 1872. The 
 vessel was approaching Whitehead from the southwestward, when, at a 
 distance of about six miles from the station, the fog-signal, which is a 
 10-inch steam-whistle, was distinctly perceived and continued to be 
 heard with increasing intensity of sound until within about three miles, 
 when the sound suddenly ceased to be heard, and was not perceived 
 again until the vessel approached within a quarter of a mile of the sta- 
 tion, although from conclusive evidence furnished by the keeper it was 
 shown that the signal had been sounding during the whole time. The 
 wind during this time was from the south, or approximately in an opp.o- 
 site direction to the sound. Another fact connected with* this occur- 
 rence was that the keeper on the island distinctly heard the sound of 
 the whistle of the steamer, which was commenced to be blown as soon 
 as the whistle at the station ceased to be heard, in order to call the at- 
 tention of the keeper to what was supposed to be a neglect of his duty 
 in intermitting the operations of his signal. It should be observed in 
 this case that the sound from the steamer was produced by a 6-inch 
 whistle, while that of the station was from an instrument of the same 
 kind of 10 inches in diameter ; or, in other words, a lesser sound was 
 heard from the steamer, while a sound of greater volume was unheard 
 in an opposite direction from the station. It is evident that this result 
 could not be due to any mottled condition or want of acoustic transpar- 
 ency of the atmosphere, since this would absorb the sound equally in 
 both directions. The only plausible explanation of this phenomenon 
 is that which refers it to the action of the wind. In the case of the 
 sound from the steamer, the wind was favorable for its transmission, 
 and hence it is not strange that its sound should be heard on the island 
 when the sound from the other instrument could not be be heard on the 
 steamer. To explain on the same principle the fact of the hearing of 
 the sound at the distance of six miles, and afterward of losing it at the 
 distance of three miles, we have only to suppose that in the first in- 
 stance the retarding effect of the wind was small, and tj^at in the sec- 
 ond it became much greater on account of a sudden increase in the rel- 
 ative velocity of the current in the upper and lower portions. 
 
 After making a critical examination of the island and the position of 
 S. Mis. 59 32 
 
498 RESEARCHES IN SOUND. 
 
 the machinery, and also in regard to any obstacle which might interfere 
 with the propagation of the sound, the keeper was directed to put the 
 instrument in operation and to continue to sound it for at least two 
 hours, or until the steamer was lost sight of, which direction was com- 
 plied with. In passing from the island, almost directly against a light 
 wind, the intensity of the sound gradually diminished as a whole, with 
 the increase of distance, but varied in loudness from blast to blast, 
 now louder, then again more feeble, until it finally ceased at a distance 
 of about fifteen miles, as estimated by the intervals between the blasts 
 and the sight of the steam as seen through a spy-glass, and also from 
 points on the Coast-Survey charts. 
 
 The result of this investigation clearly showed the power of the appa- 
 ratus in propagating sound under conditions not entirely favorable, 
 since the wind, though light, was in opposition to the sound. 
 
 Cape Elizabeth Light- Station, Maine, August 29, 1873. The fog-signal 
 at this place is on a prominent headland to which the course of all ves- 
 sels is directed when bound from the southward into Portland Harbor. 
 It is furnished with two light houses 919 feet apart and 143 feet above 
 sea-level. The easterly tower is connected with the keeper's dwelling 
 by a wooden-covered way 200 feet long and about 12 feet high ; the 
 station is furnished with a 10-inch steam fog-whistle, placed to the 
 southward of the easterly tower, at a distance of about 625 feet and 
 about at right angles with the covered way ; it therefore has a back- 
 ground, including the covered way, of about 65 feet above the height of 
 the whistle, which was found to reflect a perceptible echo. The whistle 
 was actuated by steam at 55 pounds pressure, consuming from 60 to 65 
 pounds of anthracite coal per hour. The whistle itself differs from the 
 ordinary locomotive-whistle by having a projecting ledge or rim around 
 the lower part through which the sheet of steam issues to strike against 
 the lower edge of the bell. What effect this projecting ledge or rim may 
 have is not known to the observers. This whistle is provided, (for the 
 purpose of concentrating the sound in a given direction,) with a hollow 
 truncated pyramid 20 feet long, 10 feet square at the large end, and 2J 
 feet square at the small end, the axis of the pyramid being placed par- 
 allel to the horizon, with the whistle at the smaller end. In order to 
 ascertain the effect of this appendage to the whistle the simplest plan 
 would have been to have noted the intensity of sound at various points 
 on a circle of which the whistle would have been the center. This being 
 impracticable on account of the intervention of the land, the observations 
 were confined to points on the three arcs of a circle of about 120, of 
 which the axis divided the space into 80 and 40 and a radius of one, 
 two, and three miles. The result of these observations was that, starting 
 from the axis of the trumpet on the east side, the sound grew slightly 
 less loud until the prolongation of the side of the trumpet was reached, 
 when it became comparatively faint and continued so until the line be- 
 
RESEARCHES IN SOUND. 499 
 
 tween the whistle and observer was entirely unobstructed by the side 
 of the trumpet, when the sound was apparently as loud as in the pro- 
 longation of the axis itself. On the west side, of the axis of the trumpet 
 the sound in a like manner diminished from the axis until the prolonga- 
 tion of the side of the trumpet was reached, when it became feeble 
 again, slightly increased, and then gradually diminished until the line 
 of direc ion made an angle of about 80 with the axis of the trumpet, 
 when it ceased to be heard at a distance of about one and a half miles. 
 It should be observed, however, that at this point the line of sight of 
 the observers was obstructed by the side of the trumpet and the smoke- 
 stack of the boiler. The wind was light, at south-southwest, approxi- 
 mately in direct opposition to the direction of the sound when it ceased 
 to be heard. We are informed that complaints had previously been made 
 by officers of steamers passing near this point that the sound was here 
 inaudible previous to the introduction of this trumpet ; it would there- 
 fore follow that it is of no use in increasing the effect on the western 
 side of the axis and is of injury to the sound on the lines of prolonga- 
 tion of its sides. If the sound ceased to be heard at the point mentioned, 
 when the trumpet is removed the only apparent cause of the phenome- 
 non will be the prevailing direction of the wind, which, coming from the 
 southwest, will be in opposition to the sound of the whistle ; but in the 
 case of the present investigation the force of the wind was so small that 
 it scarcely appeared adequate to produce the effect, and this question, 
 therefore, must be left for further investigation. It may be important 
 to state that in the case where the sound ceased to be heard it was re- 
 gained by sailing directly toward the station about one mile, or at half 
 a mile from the station. After making the foregoing observations as to 
 the intensity of sound in different directions from the station, the obser- 
 vations were closed by sailing directly along the axis of the trumpet 
 until the sound, which gradually grew fainter as the distance increased, 
 finally ceased to be heard at a distance of about nine miles. In com- 
 paring this last result with an instrument of about the same power at 
 Whitehead, which gave a perceptible sound at a distance of fifteen 
 miles, the only apparently variable circumstance was the velocity of the 
 wind, in both cases adverse to the direction of the sound ; but in that 
 of Cape Elizabeth it was of considerable more intensity. 
 
 During the foregoing experiments, when the vessel was about a mile 
 from the station, steaming directly outward, in the prolongation of the 
 axis of the instrument, there was heard after each sound of the whistle 
 a distinct echo from the broad, unobstructed ocean, which was attributed 
 at the time, as in other cases, to reflections from the crests and hollows 
 of the waves, a similar phenomenon having since been referred to a re- 
 flection from air of a different density. This observation becomes im- 
 portant in regard to the solution of the question as to the abnormal 
 phenomena of sound. 
 
500 KESEARCHES IN SOUND. 
 
 Cape Ann Light- Station, Massachusetts, August 31, 1873. This is one of 
 the most important stations on the New England coast. It is furnished 
 with two first-order lights, and a 12-inch steam- whistle, actuated by 60 
 pounds pressure of steam. The present is the fourth engine which has 
 been erected at this station, in consequence of the complaints either as 
 to the inefficiency of the sound or its failure to be heard in certain direc- 
 tions. It was at first proposed to sail entirely around the island in order 
 to test the intensity of the sound in different directions, but this was 
 i found impracticable on account of want of depth of water on the inland 
 Wde ; the observations were therefore confined to the direction in which 
 complaints had been made as to the deficiency of the signal, namely, in 
 a southerly direction. The result of these observations, the points of 
 which included an arc of 120, was that the sound was heard with equal 
 intensity except when the direction of the station was to the northward 
 and eastward of tke observers ; then, in one instance, the sound became 
 very indistinct, and in another was entirely lost, both at a distance of 
 about two miles. In these cases the line of sight between the observers 
 and the signal was interrupted, in the first by a small building, the gable- 
 end of which was within 10 feet of the whistle, and in the second by the 
 south light-tower, which is within 30 feet of the whistle. In this series 
 of experiments, as with the last, the wind was against the sound ; the 
 effect was noted by passing over the arc several times at different dis- 
 tances. The wind was from the southward and westward and very 
 light, and the sound was finally lost at about six miles, and in the direc- 
 tion of the obstructions. 
 
 Boston Light- Station, August 31, 1873. The light-house is situated on 
 a low, rocky island, on the north side of the main outer entrance to Bos- 
 ton Harbor, nine miles from the city. It is furnished with three caloric 
 engines^ two of the second class and one of the first. The two second- 
 class engines are so arranged as to act separately or together, and in the 
 latter arrangement serve to duplicate the larger engine. At the time 
 the observations were made, the larger engine was about being repaired, 
 and one of the smaller engines with the double air-reservoir was used. 
 The larger engine is used with 12 pounds pressure of air, which falls to 
 8 pounds in producing the sound. The smaller engine, with the double 
 reservoir, is started with i) pounds pressure, which falls to 8 pounds. 
 This difference in the pressure of air in the two engines is caused by the 
 larger ratio of the reservoir to the size of the reed. With a greater 
 pressure than 12 pounds to the square inch in the larger engine and 9 
 pounds in the smaller no sound is produced ; the reed is unable to act 
 against the pressure, and, consequently, the orifice remains closed. The 
 trumpet of the larger of the engines is reported to have been heard 
 eighteen miles at sea, which, in consideration of the results obtained at 
 Whitehead, we thought very probable. The time required, from start- 
 ing fires, to get a good working-pressure, is about half an hour. The 
 amount of coal consumed per hour is 17 pounds. 
 
RESEARCHES IN SOUND. 501 
 
 There is moreover at this station a bell, operated by a Stevens clock, 
 not at present used. It is placed on a high, wooden frame-structure, on 
 which one of the ancient bell-striking machines was originally erected. 
 The most proper position for the fog- signal is on the ground occupied by 
 this bell-tower, but as this was not removed at the time of the erection 
 of the trumpets, they were placed in such positions as to have the line 
 of sound interrupted to the northeastward by the bell and light towers. 
 It was therefore thought probable that this was the cause of the de- 
 ficiency of sound in this direction. To test this the vessel was caused 
 to traverse the arcs of several concentric circles, in the portion of the 
 horizon where the sound was most required as a signal. The first arc 
 traversed was about one and one-half miles from the signal. The vessel 
 on this crossed the axis where the sound was quite loud, and proceeded 
 northward until the sight of the trumpet was obscured by the before- 
 mentioned towers, when the sound became almost inaudible. The ves- 
 sel next returned across the axis, on a circle of about three miles radius, 
 with similar results ; but after crossing the axis the sound on the south- 
 ern side continued to be but little diminished in intensity along an arc of 
 two and a half miles, or as far as the land would allow the vessel to go. 
 The vessel was next put upon an arc, of which the radius was one and 
 a half miles, and on the south side of the axis, and sailed to the north- 
 ward until the axis was reached, it was then turned and ran for the en- 
 trance of the harbor, hugging the southern shore, keeping as far from the 
 signal as possible. Throughout this passage the sound was clear and 
 loud, showing very little, if any, diminution of power as the several posi- 
 tions deviated more and more from the direction of the axis, until the 
 vessel was at right angles with the axis, the land not permitting any 
 greater distance. The vessel approached to within three-quarters of a 
 mile of the signal and then continued still farther around, until nearly in 
 the rear of it, the sound still continuing clear and loud. The vessel next 
 proceeded up the harbor, nearly in the line of the axis of the trumpet 
 prolonged in the rear, still continuing to hear the signal distinctly until 
 the keeper, losing sight of the vessel, stopped sounding the instrument. 
 These observations were made under very favorable circumstances, it 
 being nearly calm. What wind did exist was about equally favorable 
 to points on either side of the axis. The inference from these observa- 
 tions is, first, that small objects placed near the source of sound tend to 
 diminish its intensity in the direction of its interruption, and should, 
 therefore, if possible, be removed, or the instrument so placed as to ob- 
 viate such obstructions ; and, second, that, even with the trumpet, the 
 sounjl so diverges from the axis as to be efficient even in the rear of the 
 instrument. 
 
 OBSERVATIONS ON FOG-SIGNALS, AUGUST 25, 1874. 
 
 The first of these was on board the steamer Putnam, at Little Gull 
 Island, with Admiral Trenchard, inspector of lights of the third dis- 
 
502 RESEARCHES IN SOUND. 
 
 trict, accompanied by Governor Ingersoll, of Connecticut, and Captain 
 Upshur, U. S. K 
 
 At this place are two sirens, the one to replace the other in case of 
 an accident. One of the sirens was sounded with the pressure of 50 
 pounds per square inch. The wind was across the axis of the trumpet, 
 and almost precisely at right angles to it. 
 
 The steamer was headed against the wind, on a line at right angles 
 to the axis of the trumpet. The sound in this case also travelled against 
 the wind, which was at an estimated velocity of from 4 to 5 miles per 
 hour. The distance travelled before the sound became inaudible was 
 estimated, by the speed of the steamer, at 3J miles. 
 
 The steamer was next headed in an opposite direction and returned 
 along its previous path, across the mouth of the trumpet of the siren, 
 the sound gradually increasing in strength without any marked irregu- 
 larity, until the siren was reached, and on leaving this, the course re- 
 maining the same, the sound gradually diminished in intensity, but with 
 less rapidity than before, until it was finally lost at a distance of 7 
 miles. In the latter instance the movement of the sound was with the 
 wind. The result of these observations was conformable to that gene- 
 rally obtained from previous observations, namely, that the sound is sel- 
 dom or never heard at the same distance in different directions, and, more- 
 over, that it is generally heard farther with the wind than against it. 
 
 The observations of this day also illustrate the spread of the sound- 
 wave on either side of the axis of the trumpet, a fact which has fre- 
 quently been observed in other investigations. It may be well to men- 
 tion that the siren trumpet at this locality is directed horizontally with 
 its prolonged axis passing over, immediately in front of the mouth of 
 the trumpet, a space of very. rough ground, the surface of which is 
 principally composed of bowlders, one of which, of very large size, is 
 directly in front of the trumpet, and the idea occurred to me that this 
 rough surface might produce some effect on the transmission of sound 
 to a distance. I observed by strewing sand upon a paper that the 
 former was violently agitated when held near the surface of the large 
 bowlder just mentioned, during the blast of the siren trumpet. 
 
 At this station, during the visit of Sir Frederick Arrow, the sound 
 was lost in the direction of the axis of the trumpet at a distance of two 
 miles, and then again regained with distinctness at the light- vessel, a 
 distance of four and one-half miles ; this was what we have denominated 
 as an abnormal phenomenon, which we think was due to a slight varia- 
 tion in the velocity of the lower or upper part of the current of air, but, 
 unfortunately, the demand for the use of the vessel as a light-house 
 tender prevented the attempt to ascertain whether the same phenomenon 
 would be observed a second time and to further investigate its cause. 
 
 The second investigations this season were September 1, 1874, with 
 General Barnard, of the Light- Ho use Board, and General Woodruff, 
 
RESEARCHES IN SOUND. 503 
 
 engineer of the third district. We proceeded on this occasion in the 
 steamer Mistletoe to Block Island, one of the outer stations of the 
 Light-House Board, fully exposed, without intervention of land, to the 
 waves and storms of the ocean. 
 
 On the southerly side of this island a light-house is about being erected, 
 and a siren station at this locality had been established and was in full 
 operation. 
 
 There are here two sirens attached to one boiler, one to be used in case 
 of an accident to the other. For the sake of experiment they are of 
 slightly different qualities, one with a larger trumpet with a revolving 
 disk of the old pattern, giving a lower tone ; the other a smaller trumpet, 
 having a revolving disk with openings allowing a much more sudden 
 full blast of steam, and revolving with greater velocity so as to give a 
 higher pitch. The latter is far the superior instrument, as was evident 
 to us by the sound which it produced, and as had been established by 
 the use of the artificial ear in the manufactory of Mr. Brown. The 
 effect on the unguarded ear was scarcely endurable, and the very earth 
 around appeared to tremble during the blast. The keeper (an intelli- 
 gent man who has been promoted from the station of assistant keeper 
 at Beaver Tail light to this station) informed us that a fleet of fishing- 
 vessels coming in distinctly heard it at a distance estimated by their 
 rate of sailing at scarcely less than thirty miles ; this was on two sepa- 
 rate occasions. The keeper had been directed to note and record the 
 date at which he heard the sound from other signals ; he reported that 
 he had frequently heard the fog-signal at Point Judith, a distance of 
 seventeen miles, and that the observer at the latter place frequently 
 heard his signal ; but on comparing records the two sounds had not 
 been heard simultaneously by the two keepers ; when it was heard from 
 one station it was not heard from the other, illustrating again the gen- 
 eral rule that sound is not transmitted simultaneously with equal inten- 
 sity in opposite directions. 
 
 This occasion also furnished very favorable conditions for observing 
 the remarkable phenomenon of the ocean-echo. At the cessation of 
 each blast of the trumpet, after a slight interval, a distinct and pro- 
 longed echo was returned from the unobstructed ocean. It is important 
 to observe, in regard to this phenomenon, that the siren is placed near 
 the edge of a perpendicular cliff, at an elevation of from 75 to 100 feet 
 above the ocean, and, furthermore, that the direction of the wind formed 
 an angle of about 35 with the axis of the trumpet. Now, the loudness 
 of this echo was not the greatest at the siren-house, but increased in 
 intensity until a point was reached several hundred yards from the 
 trumpet, approximately more in accordance with a reflection from the 
 waves. The wind was blowing from the shore with the direction of the 
 sound as it went off from the trumpet, and nearly against it on the re- 
 turn of the echo. I have attributed this phenomenon, which was first 
 observed in 18GG at East Quoddy Head on the coast of Maine, and since 
 
504. RESEARCHES IN SOUND. 
 
 at various stations, at which the trumpet or siren has been used, to the 
 reflection of the sound from the crests and slopes of the waves, and the 
 observation we have mentioned would appear to favor this hypothesis. 
 In connection with this explanation, I may mention that my attention 
 has been called by General Meigs, of the United States Army, to an 
 echo from the palings of a fence, and also from a series of indentations 
 across the under side of the arch of one of the aqueduct bridges of the 
 Washington water- works. The fact that the sound was much louder at 
 a point considerably distant from the trumpet was noted by one of the 
 party entirely unacquainted with the hypothesis. 
 
 The keeper at this station confirmed without a leading question the 
 statement of Captain Keeney, that it frequently happens that a feeble 
 sound of a distant object, as the roar of the surf, can be heard against 
 the direction of the wind, and that in this case it always betokens a 
 change in the weather, and is, in fact, used generally by the fishermen 
 as a prognostic of a change in the direction of the wind, which will, in 
 the course of a few hours, invariably spring up from an opposite quar- 
 ter. In such case, it is highly probable, as has been stated, that a change 
 has already taken place in the direction of the upper strata of the air, 
 although, from theoretical considerations, we might infer that the same 
 result would be produced if the wind were stationary above and moving 
 with a considerable velocity in a direction opposite to the sound at the 
 surface of the earth, the velocity gradually diminishing as we ascend, 
 for in this case, also, the inclination of the sound waves would be down- 
 ward. 
 
 The third series of investigations, September 23, 24, 1874, was made 
 in company with Captain John Davis and Major Hains, both of the 
 Light-House Board, and General Woodruif, engineer of the third dis- 
 trict, and Mr. Brown, patentee of the siren. For the purpose three 
 light-house tenders were employed, viz: Mistletoe, Captain Keeney j 
 Putnam, Captain Field ; Cactus, Captain Latham. 
 
 The place of operation chosen for the first day's series was about 1J 
 miles from the northern point of Sandy Hook. 
 
 From the experience gained by the accumulated observations which 
 had been made, it was concluded that the phenomena of sound in re- 
 gard to perturbing influences could not be properly studied without 
 simultaneously observing the transmission of sound in opposite direc- 
 tions. It was therefore concluded to employ at least two steamers in 
 making the investigations. 
 
 In regard to this point the commission was fortunate in being able to 
 command the use, for a limited period, of the three tenders mentioned 
 above, which happened to be at the time assembled at the light-house 
 depot, Staten Island, and could be spared from their ordinary opera- 
 tions for a few days without detriment to the service. It was also for- 
 tunate in selecting for the scene of the investigations an unobstructed 
 
RESEARCHES IN SOUND. 505 
 
 position in the lower bay of ]Sew York, and perhaps still more fortunate 
 in the season of the year when, on account of the heat of the snu, a land 
 and sea breeze, which changed its direction at a particular hour of the 
 day, enabled results to be obtained bearing especially on the phenomena 
 to be investigated. 
 
 Attention was first given to the character of the several steam-whistles 
 which were intended to be used as the sources of the sound during the 
 series of investigations. 
 
 These whistles, which were sounded during the whole of the observa- 
 tions with 20 pounds of steam on each boiler, gave at first discordant 
 sounds, and were found by their effect upon an artificial ear to be con- 
 siderably different in penetrating power; they were then adjusted by 
 increasing or diminishing the space between the bell and the lower cylin- 
 der by turning a screw on the axis of the bell intended for that purpose, 
 until they produced the same effect upon the sand in the membrane of 
 the artificial ear; but in order to further be insured of the equality of 
 the penetrating power of the several whistles, the three steamers abreast, 
 forming as it were a platoon, were directed to proceed against the wind, 
 sounding all the time in regular succession the Cactus first, then, after 
 an interval of a few seconds, the Mistletoe, and then the Putnam until 
 the stationary observers lost the sound of each. They became inaudible 
 all very nearly at the same moment. The sound of the Putnam was 
 thought to be slightly less distinct ; it was therefore chosen as a station- 
 ary vessel, from which the observations of the sound of the other two 
 were to be made. 
 
 The Putnam being anchored at the point before mentioned, arrange- 
 ments were made for sending off the other two vessels in opposite direc- 
 tions, one with and the other against the wind, with instructions to return 
 when the sound became inaudible to those on the stationary vessel, this 
 to be indicated by a flag-signal. It should be mentioned that the velocity 
 of the wind was measured from time to time during the subsequent ex- 
 periments with one of Eobinson's hemispherical cup anemometers, made 
 by Casella, of London. The velocity of the wind first observed by this 
 instrument, just before the starting of the vessels, was 6 miles per hour, 
 the instrument being freely exposed on the paddle-boxes of the steamer. 
 A sensitive aneroid barometer marked 30.395 in. and continued to rise 
 gradually during the day to 30.43 in. the temperature was 71 F. 
 
 The vessels left at 11:18 A. M. the wind being from the west, Captain 
 Davis taking charge of the sounding of the whistle on the Cactus, which 
 proceeded east with the wind, the sound coming to the ear of the observer 
 Against the wind; while the sounding on the Mistletoe was in charge of 
 General Woodruff, and, as the vessel steamed against the wind, the sound 
 came to the observers on the stationary vessel with the wind ; the other 
 members of the party remained on the Putnam, at anchor at the point 
 before mentioned, off the Hook, Major Hams having charge of the sig- 
 nals. The sound of the first of the vessels was heard faintly at 14 min- 
 
506 EESEAECHES IX SOUND. 
 
 utes after leaving, but not heard at 16 minutes; we may therefore assume 
 that it became inaudible at 15 minutes. And within a minute of the 
 same time, by a mistake of the signal, the other ceased to advance, and 
 commenced to come back ; the sound from it, however, was very distinct, 
 while at the same moment the sound from the other was inaudible. On 
 account of the mistake mentioned, the relative distance at which the 
 sounds from the two vessels might have become inaudible cannot be 
 accurately given; but the fact observed, that the sound which came 
 with the wind was much more audible than the other, is in conformity 
 with the generally observed fact that sound is heard farther with the 
 wind than against it. In the mean time the velocity of the wind iiad 
 sunk to 1 J miles per hour. 
 
 Next, the vessels, leaving at 11:55 A. M. changed positions ; the Cac- 
 tus, under Captain Davis, steamed west, directly in the direction from 
 which the wind caine, while the Mistletoe, under General Woodruff, 
 steamed east, directly before the wind. The result of this trial was well 
 marked in all respects ; the sound of the Mistletoe was lost in 9 minutes, 
 which, from the speed of the steamer, was estimated at about 1 miles, 
 while the sound of the Cactus was heard distinctly for 30 minutes, or 
 at an estimated distance of 5 miles. The wind at the middle of this 
 trial had sunk to 0.42 mile per hour, or nearly to a calm. The result of 
 this trial was somewhat abnormal, for though the wind had sunk nearly 
 to a calm, the sound was still heard three times as far in the direction of 
 the slight wind as against it. 
 
 After a lapse of an hour and a half a third trial was made , in the 
 mean time the wind had changed within two points of an exactly 
 opposite direction, blowing, from the indications of the anemometer, at 
 the rate of ten and one-half miles per hour. 
 
 The Cactus again steamed in the eye of the wind, which was now 
 however from nearly an opposite point of the compass, while the other 
 vessel steamed in an opposite direction. The sound of the Cactus was 
 lost at the end of twenty-seven minutes, with the wind, or at a distance 
 of four and a half miles. 
 
 The sound of the Mistletoe was lost at the end of thirty minutes, or 
 at a distance of five miles, moving against a brisk wind then blowing. 
 
 This result was entirely unexpected and much surprised every member 
 of the party, since it was confidently expected that an increase in the 
 intensity of the wind of more than ten miles per hour, and a change to 
 the opposite direction, would materially affect the audibility of the 
 sound, and give a large result in favor of the sound, which moved in 
 the same direction with the wind, but this was not the case. In the 
 course of all the observations in several years in which investigations 
 have been carried on under the direction of the chairman of the board, 
 this is the only instance in which he had heard a sound at a greater dis- 
 tance against the wind than with it, although, as before stated, a mm- 
 
RESEARCHES IN SOUND. 507 
 
 ber of eases have been reported by other observers in which, under 
 peculiar conditions of the weather, this phenomenon has been observed. 
 
 To briefly recapitulate the results, we have in this case three instances, 
 in succession, in which a sound was heard farther from the west than 
 from the east, although in the mean time the wind had changed to nearly 
 an opposite direction. Had these results been deduced from the first 
 observations made on the influence of wind on sound, or, in other 
 words, without previous experience, the conclusion would have been 
 definitely reached that something else than wind affected the convey- 
 ance of sound, and this conclusion would have been correct, if the sug- 
 gestion had been confined to the wind at the surface ; but from previous 
 observations and theoretical conclusions, the observed phenomena aro 
 readily accounted for by supposing that during the whole time of ob- 
 servation the wind was blowing from the west in the higher part of the 
 aerial current, and that the calm and opposing wind observed were con- 
 fined to the region near the surface. To test this hypothesis, Major 
 Hains constructed a balloon of tissue-paper, which, after being com- 
 pleted, was unfortunately burned in the attempt to inflate it with heated 
 air. 
 
 The remainder of this day was devoted to observations on the sound 
 of the siren at the light-house at Sandy Hook. For this purpose the 
 Cactus, under Captain Davis, was directed to steam in the eye of the 
 wind, while the Mistletoe, under General Woodruff, steamed before the 
 wind, and the Putnam steamed at right angles to the wind. Unfor- 
 tunately, on account of the diminution of light at the closing in of the 
 day, nothing could be observed. The only result obtained was that one 
 of the duplicate sirens was heard more distinctly than the other, namely, 
 the one with the higher note. 
 
 Experiments September 24, 1874. The place chosen for the observa- 
 tions of this day was still farther out in the ocean, at the Sandy Hook 
 light- vessel, 6 miles from the nearest point of land. The pressure of the 
 atmosphere was a little greater than the day before, being 30.52 ; the 
 temperature about the same, 72 Fahr. wind light, from a westerly 
 direction, as on the previous day, with a force, as indicated by the 
 anemometer, of 1.2 miles per hour. Having been provided with a num- 
 ber of India-rubber toy balloons, the two vessels were sent off in opposite 
 directions the Mistletoe toward the west, against the wind, the Cactus 
 toward the east, with the wind, leaving at 10:40 A. M. A change was 
 also made in observing the sound. In these observations the sound was 
 noted at each vessel from the other, the speed of the steamers being the 
 same ; the distance between them when the Mistletoe lost the sound of 
 the Cactus was two miles, while the Cactus continued to hear the Mistle- 
 toe's sound coming with the wind until they were four miles apart. Simul- 
 taneously with this observation a balloon was let off from the Putnam 
 at the light-vessel, which, in its ascent, moved continuously obliquely 
 
508 RESEARCHES IN SOUND 
 
 upward in a line slightly curving toward the horizon, in the direction of 
 the wind at the surface, as far as it could be followed with the eye, in- 
 dicating a wind in the same direction in the several strata through which 
 it passed, but of a greater velocity in the upper strata. 
 
 The vessels now changed places, the Cactus steaming west, the Mis- 
 tletoe east, the wind having entirely ceased at the surface of the earth. 
 In this case the Cactus lost the sound of the Mistletoe when the vessels 
 were two miles apart, while the Mistletoe continued to hear the sound 
 of the Cactus until they were three miles apart. A balloon let off as- 
 cended vertically until it attained an elevation of about one thousand 
 feet, when, turning east, it followed the direction of the previous one. 
 The sound in this case from the east was heard three miles, while that 
 from the west was heard two miles, while in the preceding observations 
 the distances were as 2 to 1 ; the only changing element, as far as could 
 be observed, was that of the wind at the surface, which became less. 
 
 Third trial, 12:45 P. M. The wind previous to this trial had changed 
 its direction 10 points or about 112 round through the south, and as 
 indicated by the anemometer at a velocity of 4.8 miles per hour. In 
 this case the Cactus, going against the wind, lost the Mistletoe's sound 
 coming to her against the wind when the vessels were 1 mile apart, 
 while the Mistletoe heard the Cactus, the sound coming to her with the 
 wind when the vessels were 1J- miles apart. The several balloons set 
 off at this time were carried by the surface wind westwardly until nearly 
 lost to sight, when they were observed to turn east, following the direc- 
 tion of the wind observed in the earlier observations. The results of 
 the whole series of observations are extremely interesting. In all the 
 experiments the difference in the audibility of the sound in different di- 
 rections was very marked, and indeed it rarely happens that the sound 
 is equal in two directions, although from the hypothesis adopted this 
 may be possible, since, according to this hypothesis, both the upper and 
 lower currents have an influence upon the audibility of sound in certain 
 directions. From the first trial, the motion of the air being in the same 
 direction both below and above, but probably more rapid above than 
 below on account of resistance, the upper part of the sound-wave would 
 move more rapidly than the lower, and the wave would be deflected 
 downward, and therefore the sound, as usual, heard farther with the wind 
 than against it. In the third experiment of the same day, in which the 
 wind changed to an almost opposite direction, if the wind remained the 
 same above, as we have reason to suppose it did from the observations 
 on the balloons on the second day, the sound should be heard still 
 farther in the same direction or against the wind at the surface, since, 
 in this case, the sound-wave being more retarded near the surface would 
 be tipped over more above and the sound thus be thrown down. 
 
 The observations of the second day are also in conformity with the 
 same hypothesis, the change in the wind being probably due to the' 
 heating of the land, as the day advanced, beyond the temperature of 
 
RESEARCHES IN SOUND. 509 
 
 9 
 
 the water, and thus producing 1 a current from the latter to the former, 
 while the wind observed in the morning from the west was the land- 
 wind due to the cooling of the latter. 
 
 In the morning the wind was blowing from the wesk both in the 
 higher strata and at the surface of the earth, and in this condition the 
 sound was heard farther with the wind than against it. 
 
 The wind at the surface about midday gradually ceased, and shortly 
 afterward sprang up from an east direction 5 in this condition the sound, 
 with the wind at the surface, was heard at a greater distance. This is 
 also in strict conformity with the theory of a change in the form of the 
 sound- wave, as in the latter case the lower portion would be retarded, 
 while the upper portion of the wave would be carried forward with the 
 same velocity, and hence the sound would be thrown .down on the ear 
 of the observer. To explain the result of the third trial of the second 
 day, we have only to suppose that the influence of the upper current 
 was less than that of the lower. The conditions for these observations 
 were unusually favorable, the weather continuing the same during the 
 two days, and the change of the wind also taking place at nearly the 
 same hour. 
 
 The fact thus established is entirely incompatible with the supposition 
 that the diminution in the sound is principally caused by a want of 
 homogeneity in the constitution of the atmosphere, since this would 
 operate to absorb sound equally in both directions. 
 
 In May, 1873, Professor Tyndall commenced a series of investigations 
 on the subject of the transmission of sound, under the auspices of the 
 Trinity House, of England, in which whistles, trumpets, guns, and a 
 siren were used, the last-named instrument having been lent by the 
 Light-House Board of the United States to the Trinity House for the 
 purpose of the experiments in question. The results of these investi- 
 gations were, in most respects, similar to those which we had previously 
 obtained. In regard to the efficiency of the instruments, the same order 
 was determined which has been given in this report, namely, the siren, 
 the trumpet, and the whistle. Professor Tyndall's opinion as to the 
 efficiency of the siren may be gathered from the following remarks. 
 Speaking of the obstruction of sound in its application as a fog-signal, 
 he says, "There is but one solution of this difficulty, which is to make 
 the source of sound so powerful as to be able to endure loss and still 
 retain sufficient residue for transmission. Of all the instruments hitherto 
 examined by us the siren comes nearest to the fulfillment of this condi- 
 tion, and its establishment upon our coasts will, in my opinion, prove an 
 incalculable boon to the mariner/' Professor Tyndall arrived at the 
 conclusions which the information we had collected tended to establish, 
 that the existence of fog, however dense, does not materially interfere 
 with the propagation of sound ; and also that sound is generally heard 
 farther with the wind than against it, although the variation of the in- 
 
510 RESEARCHES IN SOUND. 
 
 tensity of the sound is not in all cases in proportion to the velocity of 
 the wind. The result of his investigations in regard to the pitch of sound 
 was also similar to those we have given ; and, indeed, all the facts which 
 he has stated are, with a single exception as to the direction of the echo, 
 in strict accordance with what we have repeatedly observed. We regret 
 to say, however, that we cannot subscribe to the conclusions which he 
 draws from his experiments as to the cause of the retardation of sound 
 that it is due to a flocculent condition of the atmosphere, caused by the 
 intermingling with it of invisible aqueous vapor. 
 
 That a flocculent condition of the atmosphere, due to the varying den- 
 sity produced by the mingling of aqueous vapor, is a true cause of ob- 
 struction in the transmission of sotind is a fact borne out by deduction 
 from the principles of wave-motion, as well as by the experiments of the 
 distinguished physicist of the Eoyal Institution of Great Britain ; but 
 from all the observations we have made on this subject we are far from 
 thinking that this is the efficient cause of the phenomena under consider- 
 ation. A fatal objection, we think, to the truth of the hypothesis Pro- 
 fessor Tyndall has advanced is that the obstruction to the sound, what- 
 ever may be its nature, is not the same in different directions. We think 
 we are warranted in asserting that in the cases of acoustic opacity which 
 he has described, if he had simultaneously made observations in an op- 
 posite direction, he would have come to a different conclusion. That a 
 flocculent condition of the atmosphere should slightly obstruct the sound 
 is not difficult to conceive ; but that it should obstruct the ray in one 
 direction and not in an opposite, or in a greater degree in one direction 
 than in another, the stratum of air being the same in both cases, is at 
 variance with any fact in nature with which we are acquainted. We 
 would hesitate to speak so decidedly against the conclusions of Professor 
 Tyndall, for whose clearness of conception of physical principles, skill 
 in manipulation, and power of logical deduction we entertain the highest 
 appreciation, were the facts which were obtained in our investigations 
 of a less explicit character. 
 
 While the phenomena in question are incompatible with the assump- 
 tion of a flocculent atmosphere as a cause, they are in strict accordance 
 with the hypothesis of the refraction of the waves of sound due to a 
 difference in velocity in the upper and lower portions of the currents of 
 air. We do not say, however, that the transmission of sound in the 
 atmosphere is fully investigated, or that the abnormal phenomena which 
 are said to have been observed in connection with fog-signal stations 
 have been fully explained. So far from this, we freely admit we are as 
 yet in ignorance as to how the hypothesis we have adopted is applicable 
 to the critical explanation of the obstruction to sound in the abnormal 
 cases mentioned by General Duane. We feel, however, considerable 
 confidence in its power to afford a rational explanation of these pheno- 
 mena when the conditions under which they exist shall have been accu- 
 rately determined. 
 
RESEARCHES IN SOUND. 511 
 
 We are farther confirmed in our conclusion by the publication of an 
 interesting 1 paper in the proceedings of the Royal Society by Professor 
 Osborne Eeynolds, of Owens College, Manchester, intended to show that 
 sound is not absorbed by the condition of the atmosphere, but refracted 
 in a manner analogous to the hypothesis which has been adopted in the 
 preceding report. 
 
 Much further investigation is required to enable us to fully under- 
 stand the effects of winds on the obstruction of sound, and to determine 
 the measure of the effect of variations of density in the air due to in- 
 equality of heat and moisture. But such investigations can only be 
 made under peculiar conditions of weather and favorable localities, with 
 the aid of a number of steamers, and a series of observers, by whom the 
 transmissibility of the air may be simultaneously observed in different 
 directions. The position which we were so fortunate to obtain in our 
 experiments in the lower bay of New York at the season of the prev- 
 alence of land and sea breezes was exceptionally favorable for the study 
 of the action of wind upon sound. It is the intention of the Light- 
 House Board to continue observations in regard to this matter, and to 
 embrace every favorable opportunity for their prosecution under new 
 and varied conditions. 
 
 LIGHT-HOUSE BOARD, October, 1874. 
 
 PAET IV. INVESTIGATIONS IN 1875.* 
 PRELIMINARY REMARKS. 
 
 In the Appendix to the Light-House Report of 1874 1 gave an account 
 of a series of investigations relative to fog-signals, which had been made 
 at different times under the direction of the chairman of the committee 
 on experiments. 
 
 These investigations were not confined to the instruments for produc- 
 ing sound, but included a series of observations on sound itself, in its 
 application to the uses of the mariner. In the course of these investiga- 
 tions the following conclusions were early arrived at : 
 
 1st. That the rays of a beam of loud sound do not. like those of light, 
 move parallel to each other from the surface of a concave reflector, but 
 constantly diverge laterally on all sides ; and, although at first they are 
 more intense in the axis of the reflector, they finally spread out so as to 
 encompass the whole horizon, thus rendering the use of reflectors to 
 enforce sound for fog-signals of little value. 
 
 2d. That the effect of wind in increasing or diminishing sound is not 
 confined to currents of air at the surface of the earth, but that those of 
 higher strata are also active in varying its transmission. 
 
 3d. That although sound is generally heard farther with the wind than 
 against it, yet in some instances the reverse is remarkably the case, espec- 
 
 * From tlie Report of the Light-House Board, for 1875. 
 
512 RESEARCHES IN SOUND. 
 
 ially in one locality, in which the sound is heard against a northeast 
 snow-storm more distinctly than when the wind is in an opposite direc- 
 tion. This anomaly was referred to the action of an upper current in 
 an opposite direction to that at the earth, such a current being known 
 to exist in the ca.se of northeast storms on our coast. But in what man- 
 ner the action of the wind increased or diminished the audibility of 
 sound was a problem not solved. It could not be due, as might be 
 thought at first sight, to the acceleration of the sonorous impulse by the 
 addition of the velocity of the wind to that of sound, on the one hand, 
 nor to the retardation of the latter by the motion of the wind, on the 
 other. The inadequacy of this explanation must be evident when we 
 reflect that sound moves at the rate of 750 miles an hour, and therefore 
 a wind of 7 J miles an hour would only increase its velocity one per cent, j 
 whereas the actual increase in audibility produced by a wind of this 
 intensity is in some instances several hundred per cent. 
 
 In this state of our knowledge, a suggestion of Professor Stokes, of 
 Cambridge, England, which offered a plausible explanation of the action 
 of the wind, became known to us, and was immediately adopted as a 
 working hypothesis to direct investigations. 
 
 This suggestion, the importance of which appears to have escaped 
 general recognition, is founded on the fact that the several strata into 
 which a current of air may be divided do not move with the same veloc- 
 ity. The lower stratum is retarded by friction against the earth and by 
 the various obstacles it meets with, the one immediately above by friction 
 against the lower, and so on; hence the velocity increases from the ground 
 upward a conclusion established by abundant observation. Now, in 
 perfectly still air, a sounding instrument, such as a bell, produces a series 
 of concentric waves perfectly spherical ; but in air in motion the differ- 
 ence of velocity above and below disturbs the spherical form of the 
 sound-wave, giving it somewhat the character of an oblique ellipsoid, 
 by tending to flatten it above to the windward, and to increase its con- 
 vexity above to the leward ; and since the direction of the sound is per- 
 pendicular to the sound-wave, against the wind it will be thrown upward 
 above the head of the observer, and in the opposite direction downward 
 toward the earth. A similar effect will be produced, but with some varia- 
 tions and perhaps greater intensity, by a wind above, opposite to that at 
 the surface of the earth. 
 
 These propositions will be rendered plain by the following illustrations 
 (Figures 1, 2, and 3), for which I am indebted to an article in the Ameri- 
 can Journal of Science, by William B. Taylor. 
 
 Fig. l. 
 
RESEARCHES IN SOUND. 
 
 513 
 
 In these, Figure 1 represents the effect of a favorable wind in depress- 
 ing the waves of sound, S being the signal- station and O the point of 
 observation. The wind blowing from W to E, as the spheroidal faces 
 of the sonorous waves become more pressed forward by the greater ve- 
 locity of the wind above, assuming it to be retarded at the surface by 
 friction, and the direction of the acoustic beam being constantly normal 
 to the wave-surfaces, the lines of direction of the sound will gradually 
 be bent downward and reach the ear of the observer with an accumu- 
 lated effect at the point O. 
 
 Fig. 2. 
 
 Figure 2 represents the ordinary effect of an opposing wind blowing 
 from E to W against the sound ; the wave-faces being more resisted 
 above than below, assuming as before a retardation at the surface, the 
 sound-beams are curved upward, and the lowest ray that would reach, 
 in still air, the distant observer at O, is gradually so tilted up that it 
 passes above the ear of the listener, leaving him in an acoustic shadow. 
 
 Fig. 3. 
 
 o . 
 
 Figure 3 represents the disturbing effect of two winds, the lower in 
 opposition to the sound at the surface, and the upper with it. In this 
 case the principal effect will be a depression of the sound-beam, similar 
 to that shown in Figure 1, but more strongly marked, as the difference 
 of motion will be greater as we ascend. Attending this action, says 
 Mr. Taylor, there will probably be some lagging of the lower stratum 
 by reason of the surface-friction, the tendency of which will be to dis- 
 tort the lower part of the sound-waves, giving them a reverse or serpen- 
 tine curvature. In this case the upper ray of sound would only have a 
 single curvature, similar to that shown in Figure 1, while the lower rays 
 would be represented by the lower line S O, rendering the sound less 
 audible at an intermediate point, , than at the more distant station O. 
 This hypothetical case of compound refraction offers a plausible ex- 
 planation of the paradox of a nearer sound being diminished in power 
 by the wind which increases the effect of a more distant one. 
 S. Mis. 59 33 
 
514 EESEARCHES IN SOUND. 
 
 In these figures and all the succeeding ones the direction of the wind 
 is indicated by arrows. 
 
 The hypothesis we have adopted in connection with the fact of the 
 lateral spread of sound gives a simple explanation of various abnormal 
 phenomena of sound such as has been observed in the previous investi- 
 gations, and of which the following are examples : First, the audibility 
 of a sound at a distance, and its inaudibility nearer the source of sound ; 
 second, the inaudibility of a sound at a given distance in one direction, 
 while a lesser sound is heard at the same distance in an opposite direc- 
 tion ; third, the audibility of the sound of an instrument at one time 
 at the distance of several miles, while at another time the sound of the 
 same instrument cannot be heard at more than a fifth of the same dis- 
 tance ; fourth, while the sound is heard generally farther with the wind 
 than against it, in some instances the reverse is the case $ fifth, the sud- 
 den loss of sound in passing from one locality to another in the same 
 vicinity, the distance from the source of the sound being the same. 
 
 The first four of these phenomena find a ready explanation in the 
 hypothesis adopted by supposing an increase or diminution in the rela- 
 tive velocity of the currents of wind in the upper or lower strata of air. 
 The fifth is explained by the interposition of an obstacle which casts, 
 as it were, a sound-shadow, disappearing at a given distance by the 
 divergence of the rays on each side of the obstacle into what would be 
 an optical shadow. 
 
 Accounts of these investigations were presented from time to time to 
 the Light-House Board, and to the Philosophical Society of Washington 
 in 1872. Subsequently a series of investigations on the same subject 
 was instituted in England by the Elder Brethren of the Trinity House, 
 under the direction of their scientific adviser, the celebrated physicist, 
 Dr. Tyndall. While in the latter investigations various abnormal phe- 
 nomena, similar in most instances to those we have mentioned, were 
 observed, they were referred by Dr. Tyndall to an entirely different 
 cause, viz, to the existence of acoustic clouds, consisting of portions of 
 the atmosphere in a flocculent or mottled condition, due to the unequal 
 distribution of heat and moisture, which, absorbing and reflecting the 
 sound, produce an atmosphere of acoustic opacity. While we do not 
 deny the possible existence of such a condition of the atmosphere, we 
 think it insufficient to account for all the phenomena in question, and 
 believe that a more general and efficient cause is that of the wind, in 
 accordance with the hypothesis of Professor Stokes. 
 
 We regret to differ in opinion from Dr. Tyndall, and have published 
 our dissent from his views in no spirit of captious criticism or desire to 
 undervalue the results he has obtained, some of which are highly im- 
 portant. Our only object in our remarks and in our investigations is 
 the establishment of truth. 
 
 The determination of the question as to the cause of the abnormal 
 phenomena of sound we have mentioned, and the discovery of new phe 
 
KESEAECHES IN SOUND. 515 
 
 noinena, are not mere matters of abstract scientific interest, but are of 
 great practical importance, involving the security of life and property, 
 since they include the knowledge necessary to -the proper placing of 
 fog-signals, and the instruction of mariners in the manner of using 
 them. 
 
 The hypothesis we have adopted, that of the change of direction of 
 sound by the unequal action of the wind upon the sound-waves, is 
 founded on well-established mechanical principles, and offers a ready 
 explanation of facts otherwise inexplicable. It is also a fruitful source 
 from which to deduce new consequences to be verified or disproved by 
 direct experiment. It would however ill become the spirit of true 
 science to assert that this hypothesis is sufficient to explain all the facts 
 which may be discovered in regard to sound in its application to fog- 
 signals, or to rest satisfied with the idea that no other expression of a 
 general principle is necessary. An investigation however to be fruit- 
 ful in results, as a general rule, must be guided by a priori conceptions. 
 Hap-hazard experiments and observations may lead to the discovery of 
 isolated facts, but rarely to the establishment of scientific principles. 
 There is danger however in the use of hypotheses, particularly by those 
 inexperienced in scientific investigations, that the value of certain 
 results may be overestimated, while to others is assigned less weight 
 than really belongs to them. This tendency must be guarded against. 
 The condition of the experiment must be faithfully narrated, and a scru- 
 pulously truthful account of the results given. While we have used 
 the hypothesis above mentioned in the following investigations as some- 
 thing more than an antecedent probability, we have not excluded ob- 
 servations which may militate against it, and we hold ourselves ready 
 to admit the application of other principles, or to modify our concep- 
 tion of those we have adopted, when new facts are discovered which 
 warrant such changes. But we require positive evidence, and cannot 
 adopt any conclusions which we think are not based upon a logical 
 correlation of facts. 
 
 The investigations described in the following account, though simple 
 in their conception, have been difficult and laborious in their execution. 
 To be of the greatest practical value they were required to be made on 
 the ocean, under the conditions in which the results are to be applied to 
 the use of the mariner, and therefore they could only be conducted by 
 means of steam-vessels of sufficient power to withstand the force of 
 i ough seas, and at times when these vessels could be spared from other 
 duty. They also required a number of intelligent assistants skilled in 
 observation and faithful in recording results. 
 
 OBSERVATIONS IN AUGUST, 1875, AT BLOCK ISLAND. 
 
 The party engaged in these investigations consisted of the chairman 
 of the Light-House Board,- General Woodruff, U. S. A., engineer third 
 light-house district; Dr. James C. Welling, president of Columbian Uni 
 
516 RESEARCHES IX SOUND. 
 
 versify, Washington, D. C.; Mr. T. Brown, of New York, patentee of 
 the siren; Mr. Edw. Woodruff, assistant superintendent of construction ; 
 and Captain Keeney, Commander of the light-house steamer Mistletoe. 
 They arrived at Block Island on the afternoon of the 4th of August, 
 1875. This place was chosen as the site of the experiments, first, on 
 account of its insular position, being as it were in the prolongation of 
 the axis of Long Island, distant fifteen miles from the most easterly 
 part of the latter, and entirely exposed to the winds and waves of the 
 Atlantic Ocean; and, secondly, because there are on Block Island two 
 light-houses, one of which is of the first order, and connected with it are 
 two fog-signals, one of them with the latest improvements. (See Fig. 4.) 
 
 OBSERVATIONS IN REGARD TO THE AERIAL ECHO. 
 
 This phenomenon has been frequently observed in the researches of 
 the Light-House Board, in case of powerful sounds from the siren and 
 from the fog-trumpet.* It consists of a distinct reflection of sound as if 
 from a point near the horizon in the prolongation of the axis of the trum- 
 pet. The question of the origin of this echo has an important bearing, 
 according to Dr. Tyndall, on the explanation of the abnormal phenom- 
 ena of sound we have mentioned. He refers it to the non-homogeneous 
 condition of portions of the air, which reflect back the waves of sound 
 in accordance with the analogy of the reflection of light at the common 
 surface of two media of different densities. We have adopted, as a pro- 
 visional hypothesis, that it is due to the reflection from the waves and 
 the larger undulations of the surface of the ocean, in connection with 
 the divergency of beams of powerful sounds. To bring these hypotheses 
 to the test of a crucial experiment, arrangements were made, under the 
 direction of Mr. Brown, to change the direction of the axis of one of the 
 sirens from the horizontal to the vertical position. 
 
 The first observations were made August 5, with the siren in its usual 
 horizontal position, while the air was so charged with fog as to render 
 the sound of the instrument necessary for the guidance of the mariner, 
 the image of the sun being obscured and the land invisible, from the 
 sea. Under these conditions an echo was heard when the pressure of the 
 steam reached 50 pounds per square inch. The reflection in this case, 
 as usual, was from a point in the sea-horizon in the prolongation of the 
 axis of the trumpet. It was not, however, heard more distinctly when 
 standing near the origin of the sound than at several hundred feet on 
 either side of it. The interval between the cessation of the original 
 sound and the commencement of the echo was not as marked as in some 
 previous observations, not being more than four or five seconds. The 
 
 * The same phenomenon is mentioned by Froissart in his account of the embarkation 
 of the expedition of the French and English to the coast of Africa to assist the Genoese 
 against the pirates in 1390. " It was a beautiful sight/' says the chronicler, " to view 
 this fleet, with the emblazoned banners of the different lords fluttering in the wind, 
 and to hear the minstrels and .other musicians sounding their pipes, clarions, and trum- 
 pets, whose sounds were re-echoed back by the sea." (See Illustrations of Froissart 
 by H. N. Humphrey, Plate IV.) 
 
RESEARCHES IN SOUND. 517 
 
 duration of the echo was on the average about eight seconds, beginning 
 with the time of its first perception, and not with the cessation of the 
 sound of the trumpet. General Woodruff and Doctor Welling both noted 
 the peculiar character of the echo, which was that of a series of reflec- 
 tions varying in intensity from a maximum, near the beginning, and 
 gradually dying away. The wind was nearly at right angles to the axis 
 oi the trumpet and also to that of the crests of the swell of the ocean, 
 which was rolling in from the effects of a commotion without. The 
 barometer at 12 M. indicated 30.2 inches; the dry-bulb thermometer 
 73 F. the wet-bulb 70 F. indicating a remarkable degree of aqueous 
 saturation. During the whole day the air in all the region around 
 Block Island was undoubtedly in a homogeneous condition. 
 
 August 6. On this day the weather was nearly the same. The fog- 
 signal on the 5th instant was kept in operation for the use of the mariner 
 nineteen hours, and on this day it was blown twenty hours continuously. 
 The barometer marked 30.20 inches ; the thermometer 70 F. ; the fog- 
 not as equally distributed as on the preceding day ; the north end of the 
 island, distant four miles, being distinctly visible. The wind was S. W. 
 to S., making an angle of about 60 with the axis of the fog-trumpet. 
 The echo continued to be heard distinctly with a sound varying in 
 intensity, but was not as loud as we have heard it on certain occasions 
 in previous years. 
 
 During this and the preceding day, workmen were employed .under 
 Mr. Brown in inserting a flexible India-rubber tube, two inches in diam- 
 eter, between the revolving plate of the siren and the smaller end of the 
 trumpet, so that it might be brought into a vertical position. This work, 
 though apparently simple, was difficult in execution, since it involved 
 the necessity of strong supports for the cast-iron trumpet, which in itself 
 weighed eight hundred pounds, and also of a union of the parts of suf- 
 ficient strength to resist the pressure of the steam at fifty pounds to the 
 square inch. 
 
 August 7. Wind from the S. S. W. Fog continued; the workmen had 
 not as yet completed the attachment. 
 
 August 9. Barometer 30.30 inches at 12 M. Dry -bulb thermometer 
 740 F.; wet bulb 71.5. Wind S. S. W. Fog dense along the south 
 coast, but light over all the northern portion of the island. The echo 
 was heard all day, not very loudly, but distinctly. Siren still horizontal, 
 the arrangement for elevating it not having been, at 10 a. m., completed. 
 Experiments were made on the reciprocal sounds of the whistles from 
 two steamers, the results to be given hereafter. At 5 p. M. the adjust- 
 ment of the flexible tube to the smaller end of the trumpet was finished, 
 which, giving an additional length to the instrument of about 5 feet, 
 threw it out of unison with the siren proper. To restore this unison the 
 speed of revolution of the perforated plate was diminished, and after this 
 the trumpet, still being horizontal, was sounded. An echo similar in 
 
518 RESEARCHES IN SOUND. 
 
 character to those which had been observed 011 the preceding day, and 
 the earlier part of the same day, was produced. 
 
 August 10. Barometer 30.10 inches. Dry bulb 74 ; wet bulb 69 F. 
 Wind W. S. W. 5 atmosphere hazy. Observations first made with the 
 trumpet horizontal. Echo as that of preceding days, distinct but not 
 very loud, and coming principally from the portion of the horizon in the 
 direction of the axis of the trumpet. The position of the trumpet was 
 then changed, its axis being turned to the zenith in order to make what 
 was thought might be a crucial experiment. When the trumpet was 
 now sounded a much louder echo was produced than that which was 
 heard with the axis of the trumpet horizontal, and it appeared to encircle 
 the whole horizon $ but though special attention was directed to the point 
 by all the party present, no reverberation was heard from the zenith. 
 The echo appeared however to be more regular and prolonged from the 
 ocean portion of the horizon than from that of the land. 
 
 In this experiment, while there was no reflection from the zenith in 
 which the sonorous impulse was strongest, there must have been 
 reverberations from the surface of the land and the ocean. This will 
 be evident when we consider the great divergency of sound by which 
 sonorous waves from a vertical trumpet are thrown down to the 
 plane of the horizon on every side, some of which, meeting oblique sur- 
 faces, must be reflected back to the ear of the observer near the source 
 of the. sound. This inference will be more evident when it is recol- 
 lected that the reflected rays of sound diverge as well as those of the 
 original impulse. Hence reflection from the surface of the sea is a true 
 cause of the echo, but whether it be a sufficient one may require further 
 investigation. For this explanation it is not necessary that the sea 
 should be covered with crested waves ; a similar effect would take place 
 were the surface perfectly smooth but in the form of wide swells, s which 
 in places exposed to an open sea are scarcely ever absent. Moreover, 
 the increased loudness of the echo is a fact in accordance with the same 
 view. 
 
 The observations were repeated with the same effect on succeeding 
 days, until this class of experiments was ended by the bursting of the 
 India-rubber tube. Had a distinct echo been heard from the zenith, the 
 result would -have been decidedly in favor of the hypothesis of a reflec- 
 tion from the air ; but as this was not the case the question still remained 
 undetermined, especially since the atmosphere during these experiments 
 was evidently in a homogeneous condition. We do not agree however 
 in the position taken in the report of the Trinity Board, that on the 
 origin of this echo depends the whole solution of the problem as to the 
 efficient cause of the abnormal phenomena of sound. The ingenious ex- 
 perimental illustrations of the reflection of sound from a flame or heated 
 air, establish clearly the possibility of such reflection; but it must be 
 remembered they were made under exaggerated conditions, the atmos- 
 phere being in a state of extreme rarefaction in a limited space, and the 
 
RESEARCHES IN SOUND. 519 
 
 sound of a feeble character, while the phenomena in nature are produced 
 with a comparatively small difference of temperature and'with powerful 
 sounds. 
 
 EXPERIMENTS AT BLOCK ISLAND 
 
 RELATIVE TO THE EFFECT OF ELEVATION ON AUDIBILITY. 
 
 For this investigation the first-order light-house at Block Island offered 
 peculiar facilities. It is situated near the edge of a perpendicular bluff t 
 152 feet above the sea. The tower being 52 feet above the base, gives a 
 total height to the focal plane of the lens of 204 feet, on the level of which 
 the ear of the observer could be placed. 
 
 The first and second experiments of this class were made on the 10th 
 of August, with two light-house steamers, the Putnam and the Mistletoe, 
 moving simultaneously in opposite directions. The barometer indicated 
 30.10 inches of atmospheric pressure ; the dry-bulb thermometer indi- 
 cating 74 F., and the wet-bulb 69. The wind at the time of the ex- 
 periments was from the west, and of a velocity of seven miles per hour. 
 The vessels started from the point C, Fig. 4, opposite the light-house, 
 A, about one mile distant, a position as near the shore as it was con- 
 sidered safe to venture. The Putnam steamed with the wind, the Mis- 
 tletoe steamed against the wind, each blowing its whistle every half 
 minute. The duration of the sound was noted at the top of the tower and 
 at the level of the sea, Mr. Brown being the observer at the latter sta- 
 tion, while the chairman of the board, with an assistant, observed at the 
 former. On comparing notes, the watches having been previously set 
 to the same time, it was found 
 
 First. That the duration of the sound on the tower, when coming 
 against the wind, was nine minutes, while at the base of the cliff it was 
 heard only one minute. It was afterward found from the records on 
 board of the Putnam, the sound of which came against the wind, that 
 this vessel was moving, during the experiment, at half-speed, and hence 
 the duration of the sound on the tower should be considered as 4 minutes, 
 and the difference in favor of audition on the tower 4 minutes instead 
 of 8, as given by the first record. 
 
 Second. That the sound of the Mistletoe, coming to the observer with 
 the wind, was heard on the tower during 15 minutes, while it was heard 
 at the base of the cliff during 34 minutes, the difference being 19 minutes 
 in favor of hearing at the level of the sea. This result, which differs from 
 that of all the other experiments of the same class, deserves special at- 
 tention. 
 
 After making the foregoing experiments of this class, and others, on 
 the effect of wind on sound, to be described in the next section, the ves- 
 sels were called off for other duty, and the investigations were not re- 
 sumed until August 17, when the following experiments were made : 
 
 The third experiment. The wind was from the E. ~S. E., at the rate of 
 
520 
 
 RESEARCHES IN SOUND. 
 
 about five miles per hour at the surface, and a greater velocity at the 
 height of the tower. Barometer, 30.25 ins. ; thermometer, 72. 
 
 Fig. 4. 
 
 BLOCK ISLAND 
 
 L. 9 Buoy 
 
 :LH. 
 
 O 1 
 
 Scale of miles 
 2 3 
 
RESEARCHES IN SOUND. 521 
 
 In tliis and the subsequent experiments of the same day but one 
 steamer the Mistletoe was employed. It started at 10:30 A. M. from 
 the point C, Fig. 4, at the foot of the cliff, and steamed W. S. W. along 
 C B for about 12 minutes, or a distance of two miles, blowing the whistle 
 every half-minute. To note the duration of the sound, Dr. Welling was 
 stationed at the foot of the cliff, at the level of the sea, while the chair- 
 man of the Light-House Board, with an assistant who acted as clerk, 
 was on the upper gallery of the tower, the ears of the latter being almost 
 precisely 200 feet above those of the observer at the foot of the cliff. 
 
 The watches having been previously set to the same time, on compar- 
 ing results it was found that the whistle was heard at the top of the 
 tower for twelve minutes and at the bottom of the cliff for five and one- 
 half minutes, making the difference in favor of audition on the tower six 
 and one-half minutes. In this experiment the sound came to the ob- 
 servers nearly against the wind. 
 
 The fourth experiment consisted in directing the Mistletoe to proceed 
 in the opposite direction from the same point, along the line C D. It 
 started at 11:5 A. M. the breeze being light at the time, and proceeded 
 about two and one-half miles before the sound was lost to the observers. 
 On comparing notes it was found that the sound was heard at the top 
 of the tower during fifteen minutes, and at the level of the sea for eleven 
 minutes, giving a difference in favor of the hearing on the top of the 
 tower of four minutes. 
 
 In the fifth experiment, the Mistletoe steamed again in the Direction 
 with the wind, the sound from its whistle coming to the ears of the ob- 
 servers against the wind. Starting about 11:45 A. M. and steaming about 
 two miles, the sound was heard on the tower during twelve minutes 
 and at the foot of the cliff' during five and one-half minutes, making a 
 difference of six and a half minutes in favor of audition on the tower. 
 Previous to this experiment the wind had veered one point to the west, 
 bringing the direction of the sound to the observers in less direct oppo- 
 sition to the wind than in the last experiment. 
 
 The siyth experiment. In this case the steamer was directed to proceed > 
 in the opposite direction, or against the wind, so that the sound of the 
 whistle would reach the ear of the observers in the same direction as 
 that of the wind. It started at 12:19 p. M. and proceeded two and one- 
 sixth miles ; the whistle was heard during thirteen minutes on the top of 
 the tower, and at the bottom of the cliff during precisely the same time, 
 the difference between the top of the tower and the bottom of the cliff 
 in this case being nothing. 
 
 The vessel having again been called off on other duty the seventh ex- 
 periment was made the 1st of September. On this day the wind was 
 northeast ; the velocity at the top of the tower was thirteen and a half 
 miles per hour, and at the bottom of the tower eleven miles per hour. 
 The barometer indicated 30.20 inches pressure, the dry bulb 72, and the 
 wet bulb G7o. 
 
522 RESEARCHES IN SOUXD. 
 
 The theoretical conditions for exhibiting the effect of height on audi- 
 tion in this experiment were much more favorable thah any of the pre- 
 ceding. First, the velocity of the wind was greater ; second, the differ- 
 ence between the velocities at top and bottom of the tower was well 
 marked, and the direction of the wind was more favorable for direct 
 opposition to the sound as it came to the ear of the observer. In this 
 case General Woodruff was the observer at the bottom of the cliff, while 
 the chairman of the Light-House Board and his assistant, with several 
 visitors, were at the top of the tower. 
 
 The steamer started at 10:58 A. M. and proceeded during eight minutes, 
 or a mile and one- third, when the sound was lost at the top of the tower. 
 In this case, though the sound was heard for eight minutes from the top 
 of the tower, and the first five blasts marked on the notes as quite loud, 
 it was not heard at all at the bottom of the cliff at least a hundred yards 
 nearer the source of the sound. 
 
 This result, which interested and surprised a number of intelligent 
 visitors, who were in the tower at the time, strikingly illustrates the 
 effect of elevation on the audibility of sound moving against the wind. 
 The result was so important that it was thought advisable to imme- 
 diately repeat the experiment under the same conditions. 
 
 In the eighth experiment, the Mistletoe was again directed to proceed^ 
 in the direction of the wind, along the line it had previously traversed. 
 It started at 11:25 A. M., and proceeded during six minutes, or one mile, 
 when the sound was lost at the top of the tower. In this case, the first 
 blast of the whistle was feebly heard at the base of the cliff, but no 
 other, while thirteen blasts were heard at the top of the tower, of which 
 the first six were marked as loud. 
 
 That this remarkable effect was not produced by an acoustic cloud or 
 a flocculent atmosphere is evident from the experiment which imme- 
 diately succeeded. 
 
 The ninth experiment. In this trial, the Mistletoe was directed to pro- 
 ceed against the wind, so that the sound of its whistle should come to 
 the ears of the observers with the wind. It started at 11:48 A. M., and 
 proceeded during sixteen minutes, or two and two-thirds miles, when 
 the sound of its whistle was lost to the observers on the top of the tower. 
 In this case the sound of the whistle became audible at the bottom of 
 the cliff as soon as the position of the vessel became such as to bring the 
 sound to the observers approximately with the wind, and continued to 
 be audible during fifteen minutes, or within one minute as long as the 
 sound was heard at the top of the tower. 
 
 It may be mentioned as an interesting fact, that an assistant who was 
 observing the sound with General Woodruff at the foot of the cliff, when 
 the sound could not be heard at the level of the sea, in the sixth experi- 
 ment perceived it distinctly by ascending the side of the cliff to a height 
 of twenty-five or thirty feet. 
 
 All the conditions and results of these experiments are strikingly in 
 
RESEARCHES IN SOUND. 
 
 523 
 
 conformity with the theory of the refraction of sound which we have pre- 
 viously explained. 
 
 The following recapitulation of the results of the foregoing experi- 
 ments will exhibit their correspondence with the general theory : 
 
 Sound heard coming against the wind. 
 
 Experiments. 
 
 Duration at the 
 top of the tower. 
 
 Duration at the base of 
 the cliff. 
 
 Difference in favor of 
 audition on the tower. 
 
 First 
 
 44- minutes 
 
 i minute 
 
 4 minutes. 
 
 Third 
 
 12 minutes 
 
 54- minutes 
 
 &J minutes. 
 
 Fifth. ... 
 
 12 minutes 
 
 5^ minutes .- 
 
 6^ minutes. 
 
 Seventh. 
 
 8 minutes 
 
 Not heard 
 
 8 minutes. 
 
 Eighth 
 
 6 minutes 
 
 First blast heard, but 
 
 5 minutes. 
 
 Average 
 
 42 minutes 
 &J minutes . 
 
 no other, 
 minute after starting. 
 
 12 minutes. 
 
 30 minutes. 
 6. 1 minutes. 
 
 
 
 
 
 Sound heard coming with the icind. 
 
 Experiments. 
 
 Duration at the 
 top of the tower. 
 
 Duration at the base of 
 the cliff. 
 
 Difference in favor of 
 audition at base of cliff. 
 
 Second 
 
 15 minutes 
 
 34 minutes 
 
 19 minutes 
 
 Fourth . . 
 
 ] 5 minutes . . 
 
 11 minutes . 
 
 4 minutes. 
 
 Sixth 
 
 13 minutes ..... 
 
 13 minutes. 
 
 minutes. 
 
 Ninth 
 
 16 minutes 
 
 15 minutes 
 
 1 minute. 
 
 
 59 minutes . . 
 
 73 minutes 
 
 14 minutes. 
 
 
 
 
 
 Average 
 
 14 minutes 
 
 18^ minutes 
 
 34- minutes 
 
 
 
 
 
 From the first of the foregoing tables it appears that the elevation of 
 the observer has a marked effect on the audition of sound moving against 
 the wind while, from the second, with one important exception, it has 
 very little, if any, effect on sound moving with the wind. Another ex- 
 periment relative to the same class of phenomena was made on the 19th 
 of August (see Fig. 4), the wind being S. S. W. Two observers, General 
 Woodruff and Dr. Welling, starting from the bottom of the cliff imme- 
 diately below 7 the light-house, went along the beach, the one in the direc- 
 tion A /, and the other in direction A e. General Woodruff found that 
 the sound of the siren was distinctly heard all the way to the break- 
 water, and was so loud that it probably could have been heard for sev- 
 eral miles in that direction. Dr. Welling, on the contrary, entirely lost 
 the sound within a quarter of a mile of the light-house. This result is 
 readily explained as a case of lateral refraction 5 the wind was in the 
 direction traversed by General Woodruff, and contrary to that pursued 
 by Dr. Welling. In the one case the wind, retarded by the surface of 
 
524 RESEARCHES IN SOUND. 
 
 the cliff, moved with less velocity than it did farther out, and conse- 
 quently the sound was 'thrown against the face of the cliff, and on the 
 ear of the observer, and in the other thrown from it, thus leaving, as it 
 were, a vacuum, of sound. The effect in the case was very striking, since 
 the siren was pointed toward the zenith, and the sound in still air could 
 have been heard for miles in every direction. 
 
 INVESTIGATIONS AT BLOCK ISLAND 
 
 IN REGARD TO THE EFFECT OF WIND ON AUDIBILITY. 
 
 These were made by the aid of two steamers. Captain Walker, naval 
 secretary of the board, having completed a series of inspections in the 
 Third District, sent the steamer Putnam, under Captain Fields, to aid 
 the Mistletoe in the investigations. They were commenced on the 9th 
 of August, at 12 o'clock. The wind was S. S. W. with a velocity of 7J 
 miles per hour. Barometer, 30.3 inches 5 thermometer, dry bulb, 74 F. 
 wet bulb, 71J F. 
 
 The two steamers started from a buoy near the north end of the island, 
 the one steaming against the wind, and the other with it, each blowing its 
 whistle every minute. The distance travelled by each steamer was esti- 
 mated by the running time, which, from previous observations, was found 
 to be ten miles per hour. Each vessel was furnished with a whistle of 
 the same size, of 6 inches diameter, actuated by the same pressure of 20 
 pounds of steam, and which, by previous comparison, had been found to 
 give sound at this pressure of the same penetrating power. The obser- 
 vations on the Mistletoe were made by General Woodruff, and on the 
 Putnam by Dr. Welling, each assisted by the officers of the respective 
 vessels. The two steamers proceeded to buoy off the north end of the 
 island, in which position the wind was unobstructed by the land a'low 
 beach. Indeed, the island being entirely destitute of trees, and consist- 
 ing of a rolling surface, the wind had full sweep over it in every direction. 
 
 First experiment. The Putnam went against the wind and the Mistle- 
 toe in the opposite direction. The Putnam lost the sound of the whistle 
 of the Mistletoe in two minutes and stopped, but continued to blow the 
 whistle. The Mistletoe continued on her course and heard the Putnam's 
 whistle for twenty minutes in all. During the first two minutes both 
 vessels were in motion, and therefore the space through which the sound 
 was heard moving against the wind would be represented by 4, while the 
 space through which the sound was heard moving with the wind would 
 be represented by 20 + 2 = (22), the ratio being 1 : 5J. 
 
 Second experiment. In this the Putnam went with the wind and the 
 Mistletoe in the opposite direction. The Mistletoe lost the sound of the 
 Putnam's whistle in two minutes. The Putnam then stopped and re- 
 mained at rest, while the Mistletoe continued on her course until the 
 Putnam lost sound of her whistle, twenty-six minutes later. As both 
 steamers were separating during the first two -minutes with equal speed, 
 
RESEARCHES IN SOUND. 525 
 
 the distance travelled by the sound heard moving against the wind is 
 represented by 4, while the distance of the sound heard with the wind 
 is represented by 26 -f 2=28, the ratio being 1:7. It should be men- 
 tioned, however, that the notes in this experiment are defective and 
 somewhat discrepant. 
 
 Third experiment. The Putnam went against the wind, the Mistletoe 
 in the opposite direction. The Putnam lost the sound of the whistle of 
 the Mistletoe in two minutes, while the Mistletoe continued to hear the 
 whistle of the Putnam ten minutes longer. Owing to a misunderstand- 
 ing, one of the, steamers stopped for two minutes and then resumed its 
 course. As both steamers were separating during the first two minutes 
 with equal speed, the distance of the sound heard moving against the 
 wind is represented by 4, while the sound heard with the wind through 
 a space denoted by 2 x 10 -f 42=22, the ratio being 1 : 5J. 
 
 Fourth experiment. The vessels again changed directions, the Putnam 
 going with the wind, and the Mistletoe in the opposite direction. The 
 Mistletoe lost the sound in two minutes, and the Putnam nine minutes 
 later. As each steamer was moving from the other at the same rate, 
 the distance of the sound heard moving agaiast the wind would be rep- 
 resented by 4, while the distance of the sound moving with the wind 
 would be represented b}^ 9 X 2 + 4=22, the ratio being again 1 : 5J. 
 
 Fifth experiment. This experiment was made August 10, by the same 
 vessels and same observers, wind W. S. W., of about the same intensity 
 as on previous days ; barometer, 30.1 ins. ; dry bulb, 74 F. wet bulb, 69. 
 The Putnam steamed against the wind, and the Mistletoe in the opposite 
 direction. The Putnam lost the sound in two minutes, and the Mistle- 
 toe nine minutes later. The two vessels moving apart with equal veloc- 
 ity, the space traversed by the sound moving against the wind was rep- 
 resented by 4, while that in the opposite direction was represented by 
 22, viz, 9x2 + 4=22. 
 
 Sixth experiment. The vessels were next separated in a direction at 
 right angles to the wind, when each reciprocally lost the sound of the 
 other on an average of six minutes, giving a distance travelled by the 
 sound, while audible, of twelve spaces. 
 
 Seventh experiment. The vessels were next directed along an interme- 
 diate course between the direction of the wind and a line at right angles 
 to it, with the following results : The Mistletoe, against the wind, lost 
 the sound in about two minutes, while the Putnam heard the sound 
 seven minutes longer. As in the previous case, the two vessels moving 
 apart with equal velocity would in two minutes be separated by a space 
 represented by 4, which would indicate the audibility of the sound mov- 
 ing against the wind, and for the same reason the other vessel, hearing 
 the sound seven minutes longer, would have the additional space repre- 
 sented by 14, and adding to this four spaces, we have 18 to represent 
 the audibility of the sound in the direction approximating that of the 
 wind. 
 
RESEARCHES IN SOUND. 
 
 The following table exhibits at one view the results of the foregoing 
 experiments, wJiieh relate to sound moving against tlxe wind, and with 
 the wind, reduced to miles. 
 
 Experiment. 
 
 Sound with the wind. 
 
 Sound against wind. 
 
 1 
 
 Miles. 
 3.66 
 
 Miles. 
 066 
 
 2 ^ 
 
 4.66 
 
 0.66 
 
 3 ^ 
 
 3.66 
 
 0.66 
 
 4 . 
 
 3 66 
 
 066 
 
 5. ^ 
 
 3.66 
 
 0.66 
 
 
 
 
 These results are in accordance with those of all the direct observa- 
 tions which had previously been made on sound in regard to wind by 
 the Light-House Board, with the exception of those at Sandy Hook in 
 September, 1874, as given in the last report, in which the sound was 
 heard from a steamer farther against the wind than in the direction of 
 the wind. This anomaly was explained by the existence of an upper 
 currejit of air, moving in an opposite direction to that at the surface, jn 
 accordance with the hypothesis of the refraction of sound. 
 
 It will be observed that four of the experiments give exactly the same 
 distances to represent the audibility of sound with and against the 
 wind. This coincidence was not observed until after the notes were 
 collated for discussion, and, if not accidental, was due to the equal ve- 
 locity of the wind and the general conditions of the atmosphere on the 
 two days. 
 
 To give a definite idea of these relations we have plotted the results 
 obtained on August 10, in Fig. 5, converting the distances into miles and 
 referring them to a common center, and tracing through the several ex- 
 tremities of the lines representing the distances a continuous line, which 
 may be designated as the curve of audibility. C being the center to 
 which the" sounds are referred, O A. represents the distance at which the 
 sound was heard against the wind, and CB in the direction of the wind, 
 while E and D represent the distance at right angles to the wind ? 
 andj? and G- the distances respectively with and against the wind 
 on an intermediate course. 
 
RESEARCHES IN SOUND. 
 FIG. 5. 
 
 527 
 
 The curve which is presented in the foregoing figure may be considered 
 as that which represents the normal limit of audibility during the two 
 days in which the experiments were made. The line D E divides the 
 plane of the <?urve into two unequal portions, D A F E. and D G B E, 
 the former representing the audibility of sound moving against the 
 wind, and the other the audibility of sound moving with the wind. 
 
 We can scarcely think that any other condition of the air than that 
 of its motion could produce a result of this kind. It exhibits clearly the 
 fact that sound is not heard as a general rule at right angles to the wind 
 farther than with the wind, as has been asserted. In this case the ratio 
 of the latter to the former is as 11 to 6, or nearly double. 
 
 The investigation of the relation of wind to the penetration of sound 
 was renewed in a series of subsequent experiments, the results of which 
 are to be given in a succeeding part of this report, 
 
 It should be observed, in comparing Fig. 5 with the subsequent fig- 
 ures representing the curve of audibility, that the arrow representing 
 the direction of the wind points in the longest direction to the figure, 
 whereas in other figures the pointing is in the opposite direction. The 
 difference arises from the fact that in Fig. 5 the sound is supposed to 
 radiate from the center, C, while in the others the sound converges to .the 
 center as a point of observation. The foregoing diagram and all that 
 follow in this report were* plotted by Mr. Edward Woodruff, assistant 
 superintendent of construction of the third Jight-house district. 
 
528 EESEARCHES IN SOUND. 
 
 EXPERIMENTS AT LITTLE GULL ISLAND, SEPTEMBER, 1875. 
 
 The next series of experiments made during this season was at Little 
 Gull Island, at the east end of Lond Island Sound. This location was 
 chosen on account of its convenience of approach from the harbor of 
 Kew London, seven miles distant, at which the light-house steamers of 
 the third district usually remain when not engaged in active service, 
 and also because there is a light-house on the island furnished with two 
 sirens of the second order, and an extent of water on every side which 
 would allow the vessels used in the experiments to proceed from the 
 island as a center to a considerable distance in every direction. The 
 island itself is a small protuberance above the water, merely sufficient 
 in area to support a raised circular platform of about 100 feet in diam- 
 eter, on which the light-house and other buildings are erected. The 
 following sketch (Fig 6) will give an idea of the position of Gull Island 
 relative to the main-land and the islands in the vicinity. 
 
 From this it will be seen that the position was not the most favorable 
 for a stable condition of the atmosphere. As the heat of the sun in- 
 creases during the first part of the day, the temperature of the land 
 rises above that of the sea ; and this excess of temperature produces 
 upward currents of air, disturbing the general flow of wind both at the 
 surface of the sea and at an elevation above. But although the locality 
 was unfavorable for obtaining results tending to exhibit the effects of 
 broad currents of wind flowing in one direction, it had the advantage 
 of offering more varied phenomena than could otherwise have been ex- 
 hibited. Before commencing the experiments, directions were given to 
 attach a rotating iron neck to the trumpet of one of the sirens, in order 
 that it might be directed to the zenith , while the other siren remained 
 with its axis in a horizontal direction. The observers in these investi- 
 gations consisted of the chairman of the board 5 General Woodruff, 
 engineer third district; Porter Barnard, assistant superintendent of 
 construction; Captain Keeney, and other officers of the Mistletoe; with 
 an assistant who acted as one of the observers and recording clerk. The 
 Mistletoe was daily employed, though on two occasions the Cactus, 
 another of the light-house steamers, rendered assistance. 
 
RESEARCHES IN SOUND. 
 
 529 
 
 Fig. 6. 
 
 J\/eir LondorL j 
 
 BartletCs Reef Lt.V. 
 
 Plum. Id. 
 
 Fisher s Id. 
 Rctce Rock 
 
 jMtle Gull Id. LH. 
 
 Little- Gull 
 8? Vicinity 
 
 Scale of rrvil.es 
 
 S. Mis. 59 34 
 
530 RESEARCHES IN SOUND. 
 
 OBSERVATIONS ON THE ECHO AT LITTLE GULL ISLAND. 
 
 The first observations to be mentioned are those relating to the echo ; 
 the results, however, in regard to this are not very satisfactory. The 
 sirens were of the second order, and therefore the echoes produced were 
 not as distinct as those from the larger instrument at Block Island. 
 The echo from the horizontal trumpet was distinct, and in the prolonga- 
 tion of its axis ; the interval however between the blast of the siren 
 trumpet and the commencement of the echo was very brief; so short, 
 indeed, that the ending of the one and the beginning of the other were 
 generally difficult to distinguish. A slight leak in the apparatus of the 
 siren produced a continuous hum, which interfered somewhat with the 
 distinct appreciation of the sound of the echo. The keeper thought the 
 weather was not favorable for the protiuction of echoes. He thinks 
 they are heard most distinctly during a perfect calm, which did not occur 
 during the course of these investigations. 
 
 The axis of the siren with the movable trumpet being directed to the 
 zenith, strict attention was given by all the observers to any echo which 
 might be produced from it; but in this case, as in that at Block Island, 
 the slight echo which was heard came from all points of the horizon. 
 On one occasion General Woodruff called attention to a small cloud pass- 
 ing directly over the zenith, from which a few drops of rain fell upon the 
 platform on which the light-house is erected. Advantage was taken of 
 this occurrence to direct strong blasts of the siren toward the cloud, 
 but no perceptible echo was returned. We have failed, therefore, in 
 this series of investigations, to obtain any positive facts in addition to 
 those already known as to the character of the echo. In regard to the 
 hypothesis offered for its explanation, if we found little in its support, 
 we have met with nothing to invalidate it. But whatever may be the 
 cause of the phenomenon, we do not consider it an important factor in 
 explanation of the results we have obtained, since it was too feeble to 
 produce any effect in the way of absorbing any notable part of the 
 original sound. Its importance from Dr. Tyndall's point of view is its 
 apparent support of the hypothesis of a flocculent condition of the 
 atmosphere. 
 
 OBSERVATIONS ON EFFECT OF ELEVATION ON AUDIBILITY. 
 
 The next class of experiments at Little Gull Island had relation to 
 the effect of elevation on sound. The conditions here, however, for ar- 
 riving at definite results on this point were by no means as favorable as 
 those at Block Island. The height which could be commanded was only 
 that of the tower of the light house, the gallery of which is 74 feet 
 above the platform upon which the buildings are erected, and 92 feet 
 above the level of the sea, much less than that at Block Island. Be- 
 sides this, the variableness of the wind at the surface of the ocean and 
 at heights above was not favorable for the illustration of the point in 
 question. 
 
RESEARCHES IN SOUND. 
 
 531 
 
 The theoretical conditions in order that the sound may be heard with 
 greater distinctness at an elevation than below are, as we have said 
 feefore, that the wind be moving with a greater velocity in a given di- 
 rection at an elevation than at the surface of the earth, and that the 
 difference in the velocities may be against the sound-wave, so that its 
 upper part may be more retarded than the lower. In this case the di- 
 rection of a beam of sound will be curved upward, leaving as it were 
 a va-cuum of sound beneath. The distance of the origin of sound, how- 
 ever, must not be too great relatively to the elevation of the observer; 
 otherwise it will pass over his head, as well as over that of the observer 
 at the surface of the earth. In most instances the sound was not con- 
 tinuous, but was interrupted heard for a time, then lost; again be- 
 coming audible, it was heard until it finally became imperceptible. 
 Besides this, it was difficult to determine when the sound ceased to be 
 heard, since this depended on the sensibility of the ear and the greater 
 or less attention of the observer at the time of the observation. To 
 obviate these difficulties as well as the unfavorable condition of too 
 great a distance of the origin of sound from the observer, it was con- 
 cluded to adopt as the duration of the sound the elapsed time between 
 its beginning and the period when it was first lost. 
 
 The observer on the tower was Mr. P. Barnard, while the one below 
 was General Woodruff. From the records of the observations of these 
 gentlemen the following tables are compiled, the first of which indicates 
 the relative duration of sound on the top of the tower and at the bottom, 
 the sound moving against the wind; the second, the same duration, the 
 sound moving with the wind; and the third, the same with the sound 
 at right angles to it. 
 
 TABLE 1. Sound against tlie wind. 
 
 Date. 
 
 Heard at 
 top of 
 tower. 
 
 Heard at 
 bottom. 
 
 1875. 
 September 2 
 
 min. sec. 
 5 30 
 
 min. sec. 
 4 00 
 
 4 . 
 
 4 30 
 
 3 30 
 
 4 
 
 5 30 
 
 3 00 
 
 4 .... 
 
 5 00 
 
 4 00 
 
 6 
 
 7 00 
 
 2 15 
 
 6 
 
 4 00 
 
 3 00 
 
 7 
 
 5 00 
 
 2 15 
 
 8 
 
 6 00 
 
 4 00 
 
 8 
 
 5 30 
 
 3 45 
 
 8 
 
 3 30 
 
 2 15 
 
 8 , 
 
 3 00 
 
 1 15 
 
 
 
 
 Mean , 
 
 4 57 
 
 3 01 
 
 
 
 
532 
 
 RESEARCHES IN SOUND. 
 
 It appears from Table 1 that without a single exception the duration of 
 the sound was greater at the top of the tower than at the bottom, al- 
 though the difference in favor of the top of the tower in the several experi- 
 ments is very variable. These results are in accordance with what was 
 
 anticipated. 
 
 TABLE 2. Sound with the wind. 
 
 Date. 
 
 Heard at 
 top of 
 tower. 
 
 Heard at 
 foot of 
 tower. 
 
 September 2 . . . 
 
 min. sec. 
 30 00 
 
 min. sec. 
 30 00 
 
 3 
 
 16 30 
 
 18 00 
 
 4 
 
 2i 00 
 
 20 30 
 
 4 
 
 18 00 
 
 23 30 
 
 6 
 
 12 30 
 
 12 30 
 
 7 
 
 6 30 
 
 5 30 ' 
 
 
 
 
 Mean . . . 
 
 17 25 
 
 18 20 
 
 
 
 
 In these observations the duration of the sound at the bottom and top 
 are nearly the same, from which we might infer that the elevation of 
 the observer has little effect on the hearing of sound moving with the 
 wind. Were it not for the result of the first experiment of this class at 
 Block Island, we should not hesitate to adopt this as a general conclu- 
 sion. 
 
 TABLE 3 Sound heard nearly at right angles to the wind. 
 
 Date. 
 
 Heard at 
 top of 
 tower. 
 
 Heard at 
 foot of 
 tower. 
 
 September 2... < .. 
 
 min. sec. 
 6 00 
 
 min. sec. 
 4 00 
 
 2 . . .. 
 
 6 45 
 
 10 00 
 
 2 , 
 
 25 00 
 
 23 00 
 
 2 ....... 
 
 16 30 
 
 4 00 
 
 3 
 
 21 00 
 
 19 15 
 
 3 
 
 16 00 
 
 14 30 
 
 3 
 
 23 30 
 
 16 45 
 
 4 . . 
 
 19 30 
 
 17 30 
 
 6 
 
 6 30 
 
 5 30 
 
 7 
 
 5 00 
 
 6 45 
 
 7 . 
 
 12 00 
 
 12 30 
 
 8 
 
 4 15 
 
 3 15 
 
 8 . 
 
 9 30 
 
 5 00 
 
 
 
 
 Mean 
 
 13 12 
 
 10 55 
 
 
 
 

 
 RESEARCHES IN SOUND. 
 
 533 
 
 From the result of this table it would appear that the sound can be 
 heard moving at right angles to the wind better at an elevation that at 
 the surface a result not anticipated. 
 
 OBSERVATIONS ON THE EFFECT OF WIND ON SOUND. 
 
 This series was commenced on the 2d of September. Barometer, 30.3 
 inches ; thermometer, dry-bulb, 70.5 F. wet-bulb, 67.5. Wind at the 
 surface of sea was six miles per hour, and variable ; at 3 p. m. the ve- 
 locity was eight miles at the surface. (See Fig. 7.) 
 
 Fig. 7. 
 
 1 sl course 
 
 The experiments were made by means of the steamer Mistletoe, which 
 proceeded from the light-house, as a center, in different directions, 
 blowing the whistle every half-minute, and returning when, from a sig- 
 nal, the sound was lost ; the time being noted by different observers, 
 and the distance estimated by the position of the steamer in reference 
 to known objects on the Coast- Survey chart, as well as by angles of 
 azimuth and time of sailing. The steamer was directed to proceed, as 
 indicated in Fig. 7, 1st, against the wind, so that the sound would come 
 to the observers with the wind ; 2d, at right angles to the wind ; 3d, in 
 an intermediate direction between the last course and the direction of 
 the wind ; 4th, approximately with the wind, so that the sound would 
 come to the ears of the observers against the wind ; 5th, in an interme- 
 diate direction j and, 6th, again at right angles to the wind. It was 
 
534 RESEARCHES IN SOUND. 
 
 supposed that by this arrangement a symmetrical curve of sound would 
 be obtained ; and we think this would have been the case had the wind 
 remained constant in direction. It did remain nearly the same during 
 the time of describing the first, second, and third courses, and only 
 slightly varied during the fourth ; but previous to running the fifth and 
 sixth courses the wind had changed to a direction nearly at right angles 
 to its first course. 
 
 As is shown in Fig. 7, the first, second, third, and fourth courses form 
 a normal curve of audition ; the fifth and sixth courses, however, give 
 discordant results, being much longer than a symmetrical curve would 
 indicate, showing a change in the condition of the medium from that 
 which existed during the running of the other courses ; this change was 
 evidently that of the wind, which, veering, as we have seen, through 
 an arc of a little more than 90, brought it nearly at right angles to the 
 fifth course, and approximately in the direction of the sixth course ; the 
 wind also increased its velocity. These changes are sufficient, without 
 other considerations, to give a rational account of the phenomena ob- 
 served. They both tend to increase the distance at which the sound 
 would be heard. 
 
 In these experiments, as in subsequent ones, it is to be regretted that 
 for want of balloons the motion of the air above could not be ascer. 
 tained, as was done at Sandy Hook in September, 1874. Previous to 
 sailing from the depot at Staten Island attempts had been made to se- 
 cure a supply of toy balloons, but none could be found at that time in 
 the city of New York. Arrangements were therefore made for procur- 
 ing a reservoir of condensed hydrogen, by which India-rubber balloons 
 could be inflated at the time they were wanted. Unfortunately this ap- 
 paratus did not arrive in time to be of much avail in this series of experi- 
 ments. Besides this, on account of the smallness of the balloons, the 
 ascent was too slow compared to the horizontal motion to indicate the 
 direction of the wind at a considerable elevation above the points of 
 observation. They were however of use in pointing out definitely the 
 direction of the wind and the changes it was undergoing. Moreover, at 
 the time of leaving New York we were only able to procure one ane- 
 mometer, whereas we ought to have had a number, one for the top of 
 the tower, one for the bottom, and one for each vessel. 
 
 Experiments of September 3. Barometer, 30.02 inches ; thermometer, 
 dry bulb, 72.5 F. wet bulb, 70 ; wind from the east, but too slight to 
 move the cups of the anemometer ; it soon however spang up from the 
 opposite direction, in which it continued during the remainder of the 
 day, attaining a velocity of five and a quarter miles per hour. 
 
 In these experiments two light-house steamers were employed, the 
 Mistletoe and Cactus, which enabled us to obtain the results in half the 
 time, and thus to obviate the effect, in some degree, of any change in 
 the direction of the wind. On this occasion the sound was noted at the 
 light-house as it converged to a center from the whistle of each vessel, 
 
RESEARCHES IN SOUND. 
 Fig. *. 
 
 535 
 
536 RESEARCHES IN SOUND. 
 
 and also simultaneously by each vessel as it diverged from the vertical 
 siren. 
 
 We were enabled, in this way, to produce two curves by a reverse 
 process. These are plotted in Fig. 8, and exhibit a remarkable degree 
 of similarity. The corresponding parts of the two curves, being in each 
 case reversed, exhibit the fact that, through the same space in opposite 
 directions, the audibility of the sound was similarly increased with the 
 wind and diminished against it. The effect however of the wind in the 
 experiments of this day was less marked than on any in the whole 
 series, and consequently the two curves of audition more nearly ap- 
 proximate circles. 
 
 We can see in this result no other effect than that which would be 
 produced from a wind flowing with a uniform but slow velocity at the 
 surface, but having a slightly increased velocity above. Had there been 
 no wind, according to this view the two curves would have exhibited 
 two concentric circles. 
 
 Experiments of September 4. Barometer, 29.85 inches, falling ; ther- 
 mometer, dry bulb, 77 F. wet bulb, 73.25. Wind south by west, 
 twelve and one-fourth miles per hour at the top of the tower and nine 
 and one-fourth miles below ; variable. 
 
 These experiments were also made with two vessels. The distances 
 and directions are given in Fig. 9. With the exception of the fourth 
 course of the Cactus, the other courses would form nearly a symmetri- 
 cal curve, but in this case the sound of the whistle of the Cactus was 
 lost at the point a at a distance of one mile, and was afterward regained 
 at the point &, and continued audible until the steamer reached the 
 point c. 
 
 This presents one of the abnormal phenomena of sound which might 
 in part be accounted for by the existence of a flocculent cloud between 
 a and &, but why the sound could be heard so much farther in this direc- 
 tion than in the others is not easy to explain on that hypothesis. 
 
 The line I c was described after all the lines of Fig. 9 had been com- 
 pleted, and therefore the curve given in the figure correctly represents 
 the boundary of the area of audition while these courses were being run, 
 the point a being the termination under that condition of the fourth 
 course of the Cactus. To explain the abnormal line b c, we have only to 
 suppose that a change in the velocity of the wind afterward took place, 
 by which its opposition to the sound-wave was diminished ; this will ac- 
 count for the greater length of the line ; the change however did not 
 reach the light-house until after the vessel had passed the point &. 
 
RESEARCHES IN -SOUND. 
 
 Fig. 9. 
 
 Cactus 
 ^t- 7 ^ course 
 
 537 
 
 As affording evidence in support of this hypothesis, it may be men- 
 tioned that on examining the records of the Signal-Service, of which there 
 is a station at New London, seven miles north of the position at which 
 these observations were made, it was found that the wind in the morn- 
 ing of that day was south, in the afternoon southwest, and in the even- 
 ing northwest, and that it was probable, as in other cases, that the wind 
 had changed above while the part of the course & c was run. 
 
 Experiments of September 6. Barometer, 29.93 inches ; thermometer, 
 dry bulb, 74.5 F. wet bulb, 67 ; wind from northwest to southwest, 
 seventeen miles per hour. The wind, though of higher velocity than on 
 any other occasion, was variable. On this day the experiments were 
 principally made with the Mistletoe. The Cactus, being obliged to leave 
 on other duty, ran one course a distance of two-thirds of a mile before 
 the sound of her whistle was lost at the light-house. She afterwards 
 steamed off in the direction C & (Fig. 10), noting the sound of the siren, 
 which was lost at the point &, afterward regained, and heard distinctly 
 ten and one-half miles distant. 
 
 During the passage of the first course of the Mistletoe, the wind at 
 the surface and above was from southwest, the latter being indicated 
 
538 
 
 RESEARCHES IN SOUND. 
 
 by a cloud passing the zenith. During the second course the wind was 
 variable, changing its direction about 90, principally from the north- 
 west ; while during the third course the wind was again from the south- 
 west. The long course of the Cactus marked on the figure indicates the 
 sound of the siren, from the center outward, as it was heard seven and 
 one-fourth miles, then lost for an interval, and afterward heard again at 
 a distance of three and one-fourth miles farther, making, in all, ten and 
 
 one-half miles. 
 
 Fig. 10. 
 
 Experiments of September 7. Barometer, 30.1 inches ; thermometer, 
 dry-bulb, 73 F. wet-bulb, 62 ; wind, eight miles per hour above, and 
 five miles per hour below, tower. The wind was variable, as indicated 
 by the letting-off of balloons, which however did not rise to any great 
 height. The direction of the wind is shown in Fig. 11 by arrows. There 
 is nothing remarkable in the curve of audition of this day. It indicates, 
 as usual, a greater distance toward the side on which the sound was 
 moving with the wind. 
 
 Experiments of September 8. Barometer, 30.3 inches $ thermometer, 
 dry -bulb, 71 F. wet-bulb, 64.5; wind, west-southwest, fifteen miles 
 per hour at' top of tower, nine miles per hour below. Fig. 12 indicates 
 the curve of audition of the vertical siren as compared with that of the 
 
EESEAECHES IN SOUND. 
 Fig. 11. 
 
 539 
 
 horizontal siren. The steamer first proceeded along the line C a nearly 
 in the direction of the axis of the horizontal trumpet. For the distance 
 of the first three miles the horizontal trumpet was the louder. At the 
 
 Fig. 12. 
 
 point a, four miles distant, the two were distinct and very nearly equal. 
 At b they were distinct, also very nearly equal, the vertical perhaps a 
 little more distinct. At c very nearly equally distinct. At d the verti- 
 cal siren was decidedly more distinct just before entering the optical 
 
540 RESEARCHES IN SOUND. 
 
 shadow of the light-house tower and the keeper's dwelling. This shadow 
 continued to the point e, which was nearly the extent of the acoustic as 
 well as of the optical shadow, since from d to e the sound was heard 
 from neither instrument, and the origin of sound was too near to cause 
 much diiference between these two shadows. From/ to a, through the 
 point </, the two instruments continued to be fully heard the vertical 
 the more distinct. The effect of the wind in this figure is also very 
 distinctly marked, the longer lines indicating the distance the sound 
 was heard with the wind, and the shorter against it. The curve of this 
 figure is not traced through points at which the sound was absolutely 
 lost, but at which it was heard feebly and with nearly equal distinctness. 
 
 Thus far all the facts we have observed, if not in strict conformity 
 with our conception of the hypothesis of Professor Stokes, are at least 
 not incompatible with it. We are now however to direct attention to 
 a fact of much interest, which may not have escaped the attention of 
 the reader ; namely, the remarkable difference in the area of audition 
 as exhibited in the several figures, all drawn to the same scale. Com- 
 pare, for example, the curve of Fig. 10 with the inner curve of Fig. 8. 
 It might at first sight be inferred that the srnallness of the curve in 
 the former case was due to a mottled condition of the atmosphere, 
 which, by absorbing the sound, diminishes the sphere of audition ; but, 
 unfortunately for this explanation, it would appear from the observa- 
 tions made by the Cactus within the hour of obtaining the data for 
 describing the curve, that the air was then in a remarkably favorable 
 condition for the transmission of sound, since it was heard ten and a 
 half miles, the ordinary limit of the maximum penetrating power of the 
 instrument a siren of the second order; while on the 3d of September, 
 the day on which the large curve, Fig. 8, was described, the greatest 
 distance at which the sound of the same instrument could be heard was 
 eight and a quarter miles. 
 
 The only difference in the condition of the air observed during the 
 time of describing the curve of audition given in the figure, and the 
 hearing of the sound by the Cactus for ten and a half miles, was a 
 change in the direction, and perhaps in the intensity, of the wind, in the 
 latter case the direction being the same as that of the course of the 
 Cactus. 
 
 Before therefore admitting any other solution of the question as to 
 the cause of the difference in the area of audition, we must inquire 
 whether it is not possible to refer it to the action of the wind itself. 
 
 The most marked difference in the conditions which apparently af- 
 fected the phenomenon on the days in question was that of the greater 
 velocity of the wind, both at the surface of the sea and at the top of the 
 tower, and by comparing the several figures in regard to the wind it 
 will be seen that where the condition of the air was nearest that of a 
 calm the larger was the curve of audition, and the nearer the figures 
 approach to a circle, of which the point of origin of sound or the point 
 
EESEARCHES IN SOUND. 541 
 
 of perception is the center. From these facts we are inclined to think 
 that sound is not heard as far during a time of high wind in any direc- 
 tion as it is during a perfect calm, and that it is heard farthest with a 
 gentle wind. This conclusion, which was not anticipated at the begin- 
 ning of these investigations, is we think in strict conformity with the 
 hypothesis adopted. In the case of sound moving against a strong 
 wind, the sonorous waves being thrown up above the ears of the ob- 
 server, the sphere of audition in that direction is without question 
 greatly diminished ; and that it should be also diminished when sound 
 is moving with a strong wind having a greater velocity above than be- 
 low is not difficult to conceive. In this case the sound-wave will be so 
 thrown down against the earth, and so much of it absorbed, as to weaken 
 the intensity of that part which reaches the ear, while in the case of a 
 feeble wind, moving faster above than below, the portion of the wave 
 thrown down from above will only be sufficient to compensate for the 
 smaller loss by friction, and thus the sound may be heard at a greater 
 distance than in still air. But on this point, as well as others, further 
 experiments are required. 
 
 While we consider the wind as the principal agent in producing the 
 abnormal phenomena of sound, we do not by any means regard it as 
 the sole agent. Prof. Osborn Reynolds, of Owens College, Manchester, 
 without any knowledge of what was doing in America on this subject, 
 instituted a series of experiments on the effect of wind upon sound, and 
 finally adopted precisely the same hypothesis which we have used for 
 generalizing the observed phenomena. He has however in a very in- 
 genious and important paper, presented to the Royal Society in 1874, 
 extended the same principle to the effect of heat in changing the form 
 of the sound-wave, and has shown, both by reasoning and experiment, 
 that the normal direction of the sound-wave in still air, instead of pro- 
 ceeding horizontally should be turned upward, on account of the greater 
 velocity of sound near the earth, due to the greater heat of the strata in 
 that position than of those above. This principle, which indicates the 
 existence of a true refraction of sound independent of the motion of 
 the medium, is undoubtedly applicable as a modifying influence to the 
 phenomena we have recorded. It produces however only a slight ef- 
 fect in the case we have last mentioned, since the observation on board 
 the Cactus shows the condition of the air was that of little acoustic ab- 
 sorption. It would nevertheless favor the hypothesis that sound in 
 perfectly still air of homogeneous density could be heard farther than 
 sound in a moving medium, or in one of unequal temperature. This is 
 also in accordance with the fact repeatedly observed in arctic regions, 
 in which the sound of the human voice is heard at great distances dur- 
 ing times of extreme cold. In this case, the air is of a uniform temper- 
 ature above and below, but of diminished elasticity, and should, on this 
 account, transmit sound with less intensity ; and yet the audibility is 
 increased, which is explained by the assumption that its stillness and 
 
542 RESEARCHES IN SOUND. 
 
 uniformity of temperature more than compensate for the diminished 
 elasticity. The same may be said with regard to the audibility of sound 
 during a fog, which usually exists during extreme stillness of the air. 
 
 Whatever be the cause of the variation in the limit of audition as 
 exhibited in the diagrams, it is less efficient than the ordinary action of 
 the wind in producing the same phenomena. This is evident from the 
 fact that while the ratio of the extreme variation in the limits of audi- 
 tion in the first case is not more_ than 1:3, in the second it is that of 
 1:5. 
 
 Moreover, when the effect of the wind on the audition of sound in 
 relation to elevation is considered, we think we are fully warranted in 
 asserting, as we did in our last report, that the wind is a more efficient 
 cause of the variability of the penetration of sound than the invisible 
 acoustic clouds adopted by Professor Tyndall for the explanation of the 
 phenomena. 
 
 The object of these investigations, as stated at the beginning of this 
 report, was to obtain facts which might serve to establish the true theory 
 of the abnormal phenomena of sound, an object, independent of its 
 scientific interest, of much practical importance in its application to fog- 
 signals. Although the observations were not as perfect as we could 
 wish in many respects, from want of certain appliances, they are yet 
 sufficient we think to establish principles of much practical value. For 
 example, if the mariner in approaching a fog-signal while the wind is 
 blowing against the sound fails to perceive it on deck, he will probably 
 hear it by ascending to the mast-head ; or, in case a sound from a given 
 station is constantly obscured in a certain direction, while it is audible 
 in adjacent directions, we may attribute it to a sound-shadow produced 
 by some interposing object. If again, the obscuration of sound in a 
 given direction is only observed during a wind moving against the sound, 
 the cause will probably be found in a lateral refraction, due to the re- 
 tardation of the current of wind against a perpendicular wall or cliff, as 
 in the 3ase observed at Block Island, August 19. The subject however 
 is one of great complexity, and requires further investigation, but the 
 results thus far obtained may be considered as furnishing the preliminary 
 data on which to found more precise observations. These should be 
 made with the aid of a number of steamers simultaneously employed, 
 each furnished with anemometers and balloons for determining with more 
 accuracy the direction and velocity of the wind. 
 
 We hope to renew the investigations during next summer, and in view 
 of this have directed that in the mean time the light-keepers at Block 
 Island and at Point Judith shall continue to sound their sirens a certain 
 length of time every Monday, noting the direction and velocity of the 
 wind, the temperature and pressure of the air, and the audibility of the 
 sound as it comes reciprocally from each instrument. 
 
 It is shown, from the results thus far obtained from these reciprocal 
 observations, that sound is occasionally heard more distinctly against 
 
RESEARCHES IN SOUND. 543 
 
 the wind than in a contrary direction. We think however that these 
 instances are generally followed by a change in the direction of the wind 
 at the surface of the earth. 
 LIGHT-HOUSE BOARD, October, 1875. 
 
 PAET V. INVESTIGATIONS IN 1877.* 
 
 On account of the occurrence of the Centennial Exhibition, which 
 absorbed most of the time of the officers of the Light-House Board not 
 devoted to ordinary light-house service, but few observations were made 
 relative to sound in 1876, and an account of what were made is incor- 
 porated in the following report. 
 
 Agreeably to previous engagement, I visited Portland, Me., to make 
 SDine investigation in regard to an abnormal phenomenon of sound, 
 which was noticed in a former report. We left Portland on the after- 
 noon of September 3, 1877, in the steamer Iris, which had been fitted up 
 during the year under the direction of the inspector, Commander H. F. 
 Picking, and was in excellent condition, and well adapted to the duty 
 of a light-house tender. The party consisted of General J. C. Duane, 
 engineer of the first district ; Commander H. F. Picking, inspector of 
 the first district ; Mr. Edward L. Woodruff, assistant engineer of the 
 third district ; Mr. Charles Edwards, assistant engineer of the first dis- 
 trict, and myself. 
 
 We first examined one of the automatic whistling-buoys invented by 
 Mr. Courtenay, of New York. This was in place and emitting sounds 
 at a station called Old Anthony, off Cape Elizabeth, about nine miles 
 from Portland. On approaching it at right angles to the direction of 
 the wind, we heard it at the distance of a mile. But the sound did not 
 appear loud even within a few rods. It was however of considerable 
 quantity, being from a locomotive whistle of ten inches in diameter. 
 The instrument is operated by the oscillation of the waves, which at 
 this time were not of sufficient height to move it vertically through a 
 space of more than one foot. It emitted a sound at each oscillation. 
 This invention consists of a large pear-shaped buoy about twelve feet in 
 diameter at the water-surface, and floats about twelve feet above the 
 same plane. In the interior of this buoy is a large tube or hollow cyl- 
 inder, three feet in diameter, extending from the top through the bottom 
 to a depth of about thirty feet below the latter. This tube is open at 
 the bottom, but projects air-tight through the upper part of the buoy, 
 and is closed with a plate having three orifices in it, two for letting in 
 the air into the tube, and one between the others for letting it out to 
 operate the whistle. These orifices are connected with three tubes 
 which extend downward to near the level of the water, where they pass 
 through a diaphragm which divides the cylinder into two parts. 
 
 * From the Report of tlie Light-House Board for 1877. 
 
544 RESEARCHES IN SOUND. 
 
 When the buoy rises, the water in the cylinder by its inertia retains 
 its position, and a partial vacuum is formed between the head of the 
 column and the diaphragm, into which the air is drawn through two of 
 the tubes, and when the buoy descends, the escape through the injec- 
 tion-tube being prevented by valves, it is forced out of the inner tube 
 and actuates the whistle. 
 
 The mooring-chain, which is sixty fathoms in length, is attached to 
 the cylinder at a point just below the buoy, and is secured to a large 
 stone weighing about six tons. The apparatus rides perpendicularly. 
 
 The sound in this instrument is not produced merely by the difference 
 in hydrostatic pressure of the water in the two positions of the buoy, 
 but by the accumulated effect of impulse generated by the motion of 
 the apparatus. 
 
 Plans have been devised, but have not yet been perfected, to condense 
 the air in the buoy by the effect of repeated oscillations, until a valve 
 loaded to a definite pressure would open automatically and allow the air 
 to escape. In this way the sound from the accumulated pressure would 
 be produced at intervals to a greater or less extent, and would serve to 
 diversify the character of the sound so as to enable the mariner to dis- 
 tinguish different locations. The invention, as it is, is considered a 
 valuable addition to the aids to navigation, has received the unqualified 
 approbation of all navigators on this coast who are acquainted with its 
 operation, and will probably be introduced in all countries where its 
 merits are known. Experience has shown that it can be permanently 
 moored in deep water, and that vessels can safely approach it within the 
 nearest distance, and take perfect departure from it. 
 
 The Light-House Board has adopted this buoy as one of its permanent 
 aids to navigation, and will in time introduce it at all points where its 
 presence will be of importance to the navigator. In order to obtain 
 reliable data as to the operations of the automatic buoy, Commander 
 Picking has established a series of observations at all the stations in the 
 neighborhood of the buoys, giving the time of hearing it, the direction 
 of the wind, and the state of the sea, from which it appears that in the 
 month of January, 1877, one of these buoys was heard every day at a 
 station one and one-eighth miles distant ; every day but two at one two 
 and one-quarter miles distant ; fourteen times at one seven and one-half 
 miles distant, and four times at one eight and one-half miles distant. 
 It is heard by the pilots of the New York and Boston steamers at dis- 
 tances of from one-fifth to five miles, and has been frequently heard by 
 the inspector of the first light-house district at a distance of nine miles, 
 and even, under the most favorable circumstances, fifteen miles. 
 
 We sailed around the buoy and observed the difference in the intensity 
 of sound in regard to the direction of the wind, which was at the time 
 a fresh breeze of from twelve to fifteen miles per hour, from the west- 
 ward, the greatest intensity being apparently at points forty-five degrees 
 on either s:de of the axis of the wind. The effect however was not 
 very definitely marked, though the sound on the whole appeared to be 
 
RESEARCHES IN SOUND. 545 
 
 greater on the semi-circumference of the circle to the leeward ; but the 
 velocity of the wind was so great that the noise produced by it on the 
 rigging of the vessel prevented the effects from being definitely observed. 
 
 Experiments have been made with this buoy carrying whistles of dif- 
 ferent sizes, the result being that a whistle of less than ten inches diam- 
 eter does not give a sound which can be heard as far as one of the latter 
 size, although it appears to the ear near by equally loud. 
 
 There is a difference between the quantity of sound and the loudness. 
 Two sounds may be equally loud when heard near by, yet differ very 
 much in regard to their being heard at a distance, the loudness depend- 
 ing upon the intensity of sound or on the amplitude of vibration of the 
 sounding body, while the quantity of sound depends on the extent of 
 the vibrating surface. 
 
 The size of the whistle must be limited by the quantity of air ejected 
 at each oscillation of the buoy. The fact that the ten-inch whistle gives 
 a sound which can be heard farther than one of eight inches appears to 
 have a bearing on the question (the actuating force being the same) of 
 the united effect of two sounds of the same quantity and pitch, since 
 the sound from several parts of the circumference of the larger whistle 
 may be considered as a union of several sounds of less quantity. 
 
 After these observations on the automatic buoy we proceeded along the 
 coast to White Head, at the entrance of Penobscot Bay, a distance of 
 sixty miles, which we reached at about twelve o'clock at night, and cast 
 anchor in Seal Harbor, near the White Head light-house. 
 
 Our first operation next morning was the examination of an automatic 
 fog-bell, invented by Mr. Close, and which has been erected by a special 
 appropriation of Congress. It is very simple in conception, and would 
 do good service in southern latitudes, where it would not be affected by 
 the ice. It consists of an upright shaft thirty-two feet long, fastened to 
 the rock beneath the water, and kept in a vertical position by a series 
 of iron rods serving as braces. Around this shaft is a hollow metallic 
 float, having sufficient buoyancy by the motion of the waves to elevate 
 a vertical rod having at the upper end a rack gearing into a ratchet- 
 wheel. By means of projecting pins on the surface of the wheel the 
 hammer of the bell is elevated and the bell sounded at each descent of 
 the float. This arrangement is the most simple and efficient of any of 
 the kind of which we have any knowledge. 
 
 The objection to it is its liability to be deranged by the action of ice 
 and the rusting of the parts from exposure to the weather. 
 
 Our next operation at this place was the examination of the remark- 
 able abnormal phenomenon, which was the principal object of our 
 excursion. It has been frequently observed by the captains of the 
 steamers plying between Boston and New Brunswick, and has also been 
 noticed on two different occasions by officers of the light-house estab- 
 lishment. The phenomenon, as reported by these authorities, consists 
 S. Mis. 59 35 
 
546 RESEARCHES IN SOUND. 
 
 in hearing the sound distinctly on approaching the station at the distance 
 of from six to four miles, then losing it through a space of about three 
 miles, and not hearing it again until within about a quarter of a mile of 
 the instrument, when it becomes suddenly audible in almost full power. 
 This phenomenon is always noticed when the vessel is approaching the 
 signal from the southwest, and the wind is in the same or in a southerly 
 direction, and therefore opposed to the direction of the sound from the 
 station, which is the case during a fog. Commander Picking, having 
 frequently received complaints from masters of vessels as to losing the 
 sound at this place, concluded to verify the facts by his own observation. 
 For this purpose he embraced the opportunity of an inspection-tour in 
 July, 1877, to approach the station from the southwest during a fog. 
 In his own words, he heard the sound distinctly through a space of from 
 six to four miles, then lost it and could hear nothing until within a 
 quarter of a mile of the station, when the blast of the whistle burst 
 forth in full sound. The wind at this time was from the southward, or 
 against the sound. This cessation in the hearing of the sound could 
 not have been due to the failure of the instrument to emit sound, since 
 its operation is automatic when once started, and in this case the fog so 
 lifted on nearing the station as to admit the observation of the puffs of 
 steam emitted at each blast of the whistle. 
 
 On a previous occasion General Duane and Mr. Edwards on approach- 
 ing the same signal from the southwest heard the sound at about six 
 miles distance, then lost it, and did not again hear it until within about 
 a quarter of a mile. The wind in this instance was also the same as that 
 in the observation of Commander Picking, namely, from the southwest. 
 
 So well established was this phenomenon that General Duane at- 
 tempted to remedy the evil by elevating the duplicate whistle (with 
 which every station is provided) to a height twenty-two feet above the 
 level of the other whistle, by placing it on the upper end of a tube. 
 But this arrangement produced no beneficial effect. 
 
 In the morning of September 4, 1877, on which we commenced our ex- 
 periments, the weather was clear, the wind west-southwest, the velocity 
 from ten to twelve miles, remaining nearly constant during the day. 
 Our first object was to verify by direct observation the several features of 
 the phenomenon, and for this purpose we steamed to the southward, or 
 directly to the windward, from the station through the region in which 
 the abnormal phenomena had been noticed. The pressure of the atmos- 
 phere, as indicated by an aneroid barometer, was 28.9 inches. The tem- 
 perature of the air was 67 Fahrenheit; that of the water at various 
 points along our course was 58, except at two points where the ther- 
 mometer indicated 57. This difference was too small to have any per- 
 ceptible effect on the density of the rapidly moving air which was 
 passing over the surface of the water. As we increased our distance 
 from the signal the sound slightly diminished in loudness until the dis- 
 tance was between a quarter and half a mile, when it suddenly ceased 
 
RESEARCHES IN SOUND. 547 
 
 to be heard, and continued inaudible through a distance of about a mile, 
 when it was faintly heard and continued to increase in loudness until we 
 reached the distance of four miles ; at this point it was heard with such 
 clearness that the position of the station could be located with facility ; 
 but on proceeding farther in the same direction it appeared to diminish 
 gradually except at one point, when a blast, as indicated by the steam 
 issuing from the whistle, was inaudible; but on turning the vessel around 
 the next blast was distinctly heard. 
 
 As a second experiment we retraced the same line back to the station 
 and observed the same phenomena in a reverse order. The sound was 
 heard the loudest at a point four miles from the station ; afterward it 
 diminished and then became inaudible through a space of two miles, 
 and then suddenly burst forth nearly in full intensity at the distance of 
 a quarter of a mile, and continued loud until the station was reached. 
 
 As a third experiment the same line was traversed again, the only 
 difference in the condition of the experiment being that the whistle on 
 the steamer was sounded every minute between the blasts of the signal 
 at the station ; and while the observers on the vessel noted the sounds 
 from the latter, those at the station observed the sound from the former. 
 The same phenomena as described in the previous experiments were 
 witnessed by those on board the vessel, but on receiving the report of 
 the observers at the station, it was found that no cessation of the sound 
 from the steamer was observed through the whole distance traversed by 
 the vessel. It should be noted that the whistle at the station is ten 
 inches in diameter, actuated by a pressure of sixty pounds of steam, 
 and that on board the vessel six inches in diameter with twenty-five 
 pounds of steam. It appears from this remarkable result that a feeble 
 sound passes freely through what has been called the region of silence 
 when sent in the direction of the motion of the wind, when a louder 
 sound does not pass in the opposite direction. 
 
 As a fourth experiment the vessel proceeded northward on the oppo- 
 site side of the station to that before traversed, but in the prolongation 
 of its previous course. The sound in this case from the signal to the 
 observers on the vessel was with the wind, while that from the vessel 
 to the observers at the station was against the wind. In this experi- 
 ment no cessation was observed on the vessel in the hearing of the 
 sound from the station ; it was heard with varying intensity to the dis- 
 tance of four and a half miles, and could probably have been heard 
 much farther had our progress not been interrupted by land. On return- 
 ing to the station the observers there reported that after the vessel had 
 left the station and was scarcely more than a hundred yards distant not 
 a single blast of its whistle was heard. In this case the phenomena 
 which had been observed on the southerly side of the station were 
 exhibited in a reverse order on the northerly side. 
 
 In what may be considered the fifth experiment, the vessel being at a 
 distance of four miles from the station on the line traversed in the first 
 
548 RESEARCHES IN SOUND. 
 
 two experiments, the sound was heard slightly. The vessel then altered 
 its course so as to steam, as it were, around the signal, keeping at the 
 same distance until the direction of the station from the vessel was 
 nearly at right angles to the direction of the wind ; at this point no 
 sound was heard from the station, although it had been slightly heard 
 at points along the curved line traversed in reaching the point men- 
 tioned. The vessel then proceeded toward the station in a straight line, 
 but no sound was heard until it approached the latter within one-fourth 
 of a mile. The observers at the station however heard the sound from 
 the vessel through the whole distance. 
 
 This experiment was made to ascertain the truth of the general impres- 
 sion that at this place the sound was always heard better coming at right 
 angles or across the wind than in the direction in which it was blowing. 
 The experiment however was found in conformity with the general rule 
 previously established, that the sound was usually heard farthest with 
 the wind, less against the wind, and at an intermediate distance across 
 the wind. 
 
 The primary object of these investigations is to determine the mechan- 
 ical causes to which the phenomena may be referred and from which 
 new conclusions may be deduced, to be further tested by experiment, 
 and such definite views obtained as may be of value in the employment 
 of fog-signals for the uses of the mariner. 
 
 For this purpose a number of different hypotheses may be provision- 
 ally adopted and each compared with the actual facts observed. 
 
 The first hypothesis which has been suggested for the explanation of 
 the phenomena in question is that they are due to some configuration of 
 the land ; but on inspecting the Coast-Survey chart of this region it will 
 be seen that the nearest land consists of a series of broken surfaces not 
 rising above the ocean enough to reflect sound or in any way to produce 
 sound-shadows in the region through which the phenomena are observed. 
 This hypothesis therefore is inadmissible. 
 
 Another hypothesis is that of what have been called invisible acoustic 
 clouds or portions of atmosphere existing over the water at the region 
 of silence, which might absorb or variously refract the sound. That 
 such a condition of a portion of the atmosphere really exists in some 
 cases is a fact which may be inferred from well-established principles of 
 acoustics, as well as from experimental data. They would occur espe- 
 cially in the case of dissolving clouds, which would be accompanied by 
 local diminutions of temperature, and also from portions of air which 
 have been abnormally heated by contact with warm earth. But, if the 
 phenomena in question were produced by a cloud of this kind, its pres- 
 ence ought to be indicated by transmitting through it the usual set of 
 meteorological instruments. This was done in the foregoing experi- 
 ments, but no change was observed in the indications either of the 
 thermometer or barometer. Unfortunately we had not a hygrometer in 
 our possession, but this observation was less necessary, since from 
 
RESEARCHES IN SOUND. 549 
 
 abundant testimony it is established that the same phenomena are ex- 
 hibited during a dense fog, in which all parts of the atmosphere for 
 miles in extent must be in a homogeneous condition. Furthermore, a 
 local cloud could not -continue to exist in a given space for more than an 
 instant while a wind Avas blowing with a velocity of from ten to twelve 
 miles an hour. Again, this hypothesis fails entirely to explain the fact 
 that this phenomenon is always observed at nearly the same place, espe- 
 cially during a fog, when the wind is in a southerly direction. Finally, 
 it is impossible to conceive of a cloud so arranged as a screen produc- 
 ing a sound-shadow of greater intensity on one side than on the other. 
 
 Another hypothesis is that of the refraction of sound due to the action 
 of the wind. It is an inference from well-established theory, as well as 
 from direct observation, that the sound is refracted by the wind, that it 
 tends to be thrown upward when moving against the wind, and down- 
 ward with the wind. This result is attributed very properly to the dif- 
 ferent velocities of the strata, that next the surface being retarded, 
 those above being less retarded. 
 
 The upper part of the front of the wave is thus thrown backward, and 
 the direction of the wave turned upward. In the case of the experi- 
 ment south of the station, the wind passing over a long line of rough 
 sea was moving less rapidly in its lower stratum than in the higher, and 
 consequently the sound-wave was thrown backward above, and, as it 
 issued from the instrument, tended to rise above the head of the observer, 
 and at a certain distance from the origin of the sound, depending upon 
 the difference of velocity above and below, was lost entirely to the ob- 
 server, and a sound-shadow was thus produced by refraction which is 
 closed in again by the lateral spread of the sound at a given distance. 
 
 In the experiment on the other side of the signal, the vessel proceed- 
 ing to the north, the wind coming to the observer on the vessel had to 
 pass over a rougher surface than that of water, and consequently the 
 difference of velocities above and below, and therefore the refraction 
 would be greater, and consequently the sound from the vessel was almost 
 entirely lost to the observer at the station, while the sound from the 
 station was heard uninterruptedly on the vessel, since it was moving 
 with the wind. 
 
 On examining the records of experiments of previous years, I find a 
 number of cases recorded where sounds were heard at a greater dis- 
 tance, while inaudible at a less distance, especially one in connection 
 with the fog-signal at Gull Island, in 1874. In this case the sound, in 
 passing from the signal, was heard distinctly at the distance of about 
 two miles against the wind, then lost for a space of about four and a 
 half, and heard again distinctly for a distance of perhaps one mile. At 
 the same station, during the experiments of 1875, the sound of the 
 whistles of the steamers was heard for a certain distance, then ceased 
 to be heard for a considerable interval, and was then heard again. Fur- 
 thermore, the pilots of the steamboats from New York to Boston report 
 
550 RESEARCHES IN SOUND. 
 
 that the automatic buoy is found to intermit its sound, being heard at 
 a distance, then becoming inaudible, and heard again as the steamer 
 approaches the source of sound. 
 
 From all the facts which we have gathered on this subject, I think it 
 highly probable that in all cases in which sound moving against the 
 wind is thrown up above the head of the observer it tends to descend 
 by the lateral spread of the sound-wave and to reach the earth at a dis- 
 tance ; the conditions however for the actual production of this effect 
 are somewhat special, and will depend upon the amount of the initial 
 refraction and the quantity of the sound-waves. Besides the lateral 
 spread of the sound-wave there are two other causes sufficient, in certain 
 cases, to biing a portion of the sound-waves which have been elevated 
 in the air back again to the earth : the first is when an upper current 
 of wind is moving in an opposite or approximately opposite direction 
 to that at the surface of the earth, in which case an opposite or down- 
 ward refraction would take place ; and the second is the case in which 
 the surface- wind is terminated above by strata of still air ; in this case, 
 also, a reverse refraction, but of less amount, would take place, which 
 would tend to bring the sound-wave downward. 
 
 We can readily imagine that an isolated island, cooled by the radia- 
 tion of the heat by night, would send every morning, in all directions, 
 a current of cold air from its center. In this case, the sound from a 
 whistle placed in the center of the island would be inaudible in a space 
 entirely surrounding it, and thus give rise to a condition mentioned by 
 General Duane, in which a fog-signal appeared to be surrounded by a 
 belt of silence. 
 
 The next experiment was made on the morning of the 5th, on leaving 
 the station. In this case we proceeded along the direction of the same 
 line in which the first, second, and third experiments were made the day 
 before. The wind had changed about four points to the southward. 
 As in the preceding experiments, the sound was lost again at the dis- 
 tance of about one-fourth of a mile, but was not distinctly regained, 
 though some of the observers thought they heard it at a distance of two 
 and one-half miles. 
 
 The only perceptible difference in the wind on the 5th was that it was 
 a little less rapid, and four points more to the southward. 
 
 From a subsequent report of the keepers, the whistle of the vessel 
 was heard continuously as far as the puffs of steam could be observed, 
 a distance six or seven miles. In this case the sound was moving with 
 the wind. These results therefore are in accordance with those pre- 
 viously obtained. 
 
 The next experiments were made at Monhegan, an island sixteen miles 
 southwest of White Head. On this island there is a Daboll trumpet 
 actuated by a hot-air engine.. 
 
 We departed from this station in. a westerly direction at an angle of 
 
RESEARCHES IN SOUND. 551 
 
 45 to the right of the direction of the wind, and after proceeding about 
 one mile, as estimated by time, we lost the sound of the signal. We 
 then turned at right angles to our former course and proceeded toward 
 the leeward, keeping about the same distance from the signal, when the 
 sound was regained at a point which probably depended upon the direc- 
 tion of the wind and the axis of the trumpet combined. From this point 
 it was heard to a point to the leeward, and thence we retraced our 
 course at about the same distance and proceeded across the axis of the 
 trumpet toward the windward, where the sound was again lost. The 
 only definite result from this experiment was another case of the sound 
 being heard farther to the leeward than to the windward. 
 
 After this experiment we returned to Portland. 
 
 An interesting fact may be mentioned in connection with this station, 
 having a bearing upon the protection of light-houses from lightning. 
 The fog- signal is placed on a small island separated from the large island 
 by a water-space of about one-eighth of a mile. General Duane, desir- 
 ing to connect the light-house and fog-signal by an electrical communi- 
 cation, suspended a wire between the two points and attempted to form 
 a ground connection by depositing a plate of metal in the ground on 
 each island, but to his surprise, though the arrangements were made by 
 a skilled telegrapher, no signal would pass. The two islands being 
 composed of rock and the soil limited in thickness, the conduction was 
 imperfect, and it was only by plunging the plate of metal into the water 
 on each side of the space between the two islands that a signal could 
 be transmitted. 
 
 No further experiments on sound were made during this excursion, 
 because the vessel could no longer be spared from more pressing light- 
 house duty in the way of inspection and the stated supply of materials 
 to the stations. 
 
 On my return to New York, accompanied by Mr. Woodruff, J took the 
 route by the Western Railway to the Hudson Eiver at Troy. This line 
 was chosen in order to make some investigations relative to any pecu- 
 liarities of sound which might be observed in the Hoosac tunnel, through 
 which the railroad passes. For this purpose we spent a day at East 
 Windsor, a village situated near the west end of thje tunnel, and were 
 very cordially received by the engineers in charge. 
 
 The tunnel is four and three-quarters miles in length, twenty -four feet 
 wide, and twenty feet high to the crown of the arch. It ascends slightly 
 from either end to a point near the center, where there is a ventilating- 
 shaft 1,028 feet high extending to the outer air above. In winter, when 
 the external temperature is less than that within the tunnel, there is a 
 constant current from each end toward the center, and in the summer, 
 when the temperatures are reversed, there is a current out of the tunnel 
 at either end, except when the external wind is sufficiently strong, es- 
 pecially from the west, to reverse the direction of the current from one 
 
552 RESEARCHES IN SOUND. 
 
 half and direct tbe stream entirely out of the other entrance. At the 
 time of our visit, there was a gentle current flowing out of both ends. 
 
 The only peculiarity of sound which had been observed, as stated by 
 the engineers, was that it was greatly stifled at the time by the smoke 
 with which the air was filled immediately after the passage of a loco- 
 motive. So great was this in some cases that accidents were imminent 
 to the workmen, who are constantly occupied in the tunnel in lining the 
 crown of the arch with brick, by the sudden appearance of a locomo- 
 tive, the approach of which had not been heard. 
 
 That the audibility of sound should be diminished by smoke was so 
 contrary to previous conceptions on the subject, since sound is not prac- 
 tically interrupted by fog, snow, rain, or hail, that I was induced to 
 attribute the effects which had been observed to another cause, and to 
 regard the phenomenon as due to an exaggerated flocculent condition 
 of the air in the tunnel ; adopting in this instance the hypothesis ad- 
 vanced by Dr. Tyndall, and so well illustrated by his ingenious ex- 
 periments. The effect which would be produced in the condition of 
 the air in the tunnel by the passage of a. locomotive is indicated by 
 the appearance of the emitted steam extending behind the smoke- 
 stack of a locomotive in rapid progress before the observer at a dis- 
 tance. This consists of a long stream composed of a series of globu- 
 lar masses produced by the successive puffs of steam which are emitted 
 at equal intervals. Allowing the diameter of the driving-wheels to be 
 five feet, then since four puffs are made at each revolution of the wheels, 
 a puff of hot steam would be given out at every four feet travelled by 
 the engine, and these puffs mingling with the air at the ordinary tem- 
 perature would produce an exaggerated iiocculeut condition. On our 
 expressing a desire to witness the effect upon sound of the passage of a 
 locomotive through the tunnel, Mr. A. W. Locke, one of the engineers 
 who had charge of the western section, politely offered us the means of 
 experimenting on this point, and also of passing leisurely through the 
 tunnel on a hand-car. 
 
 To observe the effect of a locomotive on the sound we took advantage 
 of the entrance of a freight train, impelled by two engines, the extra 
 one being necessary to drive the load up the inclined plane to the mid- 
 dle of the tunnel, where it was detached and returned along the same 
 line, while the train was drawn the remaining distance along the eastern 
 decline by a single engine. In order to make the experiment with re- 
 gard to sound the time was accurately noted during which the noise of 
 the entering engines could be distinctly heard, which would give ap- 
 proximately the distance the sound travelled through the flocculeut at- 
 mosphere produced by the locomotive before becoming inaudible, and 
 again the time was noted from the first hearing of the returning engine 
 until it reached the end of ^ the tunnel. In the mean time the current of 
 air blowing through the tunnel had removed a considerable portion, at 
 least, of the flocculent atmosphere, so that the seund in this case came 
 
RESEARCHES IN SOUND. 553 
 
 through an atmosphere of comparative uniformity of temperature, or 
 one much less flocculent than the other ; the result was that the dura- 
 tion of sound in the first case was about a minute, while in the second 
 it was upward of two minutes. The darkness in the tunnel, on account 
 of the smoke, was so profound immediately after the passage of a loco- 
 motive, that with two large torches, charged with mineral oil, the sides 
 of the tunnel at a distance of six feet could scarcely be observed ; while 
 in the other half of the tunnel, where no smoke existed, the eastern 
 opening could be observed like a star in the distance of upward of two 
 miles. It was therefore not surprising that the stifling of the sound 
 which was observed should be referred to the smoke as a palpable cause, 
 and that the more efficient one of the varying density or flocculent con- 
 dition should be disregarded. 
 
 The method of determining by experiment the question as to which of 
 these causes was the efficient one did not occur to me until we had left 
 the tunnel, and then the simple expedient suggested itself to me, for the 
 purpose of repeating the experiment, that instead of locomotives charged 
 with wood, two locomotives charged with charcoal or coke which emit 
 no smoke, but only transparent gases principally carbonic acid should 
 be used in an experiment similar to the one just described. This ex- 
 periment Mr. Locke has kindly promised to perform as soon as a con- 
 venient opportunity shall occur. 
 
 The opportunity was embraced while at the mouth of the tunnel to 
 make some observations which might have a bearing uj>on the phenom- 
 ena of the aerial echo. For this purpose, a d vantage was taken of a 
 large tool-chest, which happened to be place d about twenty or thirty 
 feet within the western mouth of the tunnel. By slamming down vio- 
 lently the cover of this chest, a loud sound of an explosive character 
 was produced, from which a prolonged echo was returned from the in- 
 terior of the tunnel. This echo was slightly intermittent, suddenly in- 
 creasing in loudness at intervals for a moment, and again resuming its 
 uniform intensity. This effect was attributed to projecting pieces of 
 rock in that part of the tunnel which had not been lined with brick. 
 An echo was however evidently returned from that portion of which 
 the sides were not projecting, which I would consider an effect of the 
 same cause which produces the aerial echo. 
 
 AERIAL ECHOES. 
 
 During the year 1877, (as also in 187G,) series of experiments were 
 made on the aerial echo, in which I was assisted in the first series by 
 General Woodruff, engineer of the third light-house district, and in the 
 second series by Edward Woodruff, assistant engineer of the same dis- 
 trict. These experiments were made principally at Block Island, but 
 also at Little Gull Island. Especial attention has been given to this 
 phenomenon, which consists in a distinct echo from the verge of the 
 horizon in the direction of the prolongation of the axis of the trumpet 
 
554 RESEARCHES IN SOUND. 
 
 of the siren , "because the study of it has been considered to offer the 
 easiest access to the solution of the question as to the cause of all the 
 abnormal phenomena of sound, and also because it is in itself an object 
 of much scientific interest. 
 
 In my previous notice of this phenomenon, in the report of the light- 
 house board for 1874, 1 suggested that it might be due to the reflection 
 from the crests of the waves of the ocean ; but as the phenomenon has 
 been observed during all conditions of the surface of the water, this ex- 
 planation is not tenable. 
 
 Another hypothesis has been suggested, that it is due to a flocculent 
 condition of the atmosphere, or to an acoustic invisible cloud, of a 
 density in different parts differing from that of the general atmosphere 
 at the time. To test this hypothesis experimentally, the large trumpet 
 of the siren was gradually elevated from its usual horizontal position to 
 a vertical one. In conception, 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 necessary to the 
 desired end. The trumpet, in its vertical position, was sounded at in- 
 tervals 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 restoring the 
 trumpet to its horizontal position, the echo gradually increased on the 
 side of the water, until the horizontal position was reached, when the 
 echo, as usual, appeared to proceed from an angle of about twenty de- 
 grees 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. On one occasion it happened that a small cloud passed 
 directly over the island on which the light-house is erected, and threw 
 down on it a few drops of rain. At the moment of the passage of this 
 cloud the trumpet was sounded, but no echo was produced. 
 
 From these experiments it is evident that the phenomenon is in some 
 way connected with the plane of the horizon, and that during the contin- 
 uance of the experiment of sounding the trumpets while directed toward 
 the zenith no acoustic cloud capable of producing reflection of sound ex- 
 isted in the atmosphere above them. 
 
 Another method of investigating this phenomenon occurred to me, 
 which consisted in observing the effects produced on the ears of the ob- 
 server 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 proceeded 
 from the station into the region of the echo and in the line of the pro- 
 longation of the axis of the trumpet, with the following results : 
 
 1. As the steamer advanced, and the distance from the trumpet was 
 
RESEARCHES IN SOUND. 555 
 
 increased, the londness of the echo diminished , contrary to the effect of 
 an echo from a plane surface, since in the latter case the echo would 
 have increased in loudness as the reflecting surface was approached, 
 because the whole distance travelled by the sound-wave to and from the 
 reflector would have been lessened. The effect however is in accord- 
 ance with the supposition that the echo is a multiple sound, the several 
 parts of which proceed from different points at different distances of the 
 space in front of the trumpet, and that as the steamer advances toward 
 the verge of the horizon, it leaves behind it a number of the points from 
 which the louder ones proceed, and thus the effect upon the ear is dimin- 
 ished as the distance from the trumpet is increased. 
 
 2. The duration of the echo was manifestly increased, in one instance, 
 from five seconds, as heard at the mouth of the trumpet, to twenty 
 seconds. 
 
 This would also indicate that the echo is a multiple reaction of vary- 
 ing intensities from different points, and that at the place of the steamer 
 the fainter ones from a greater distance would be heard, which would 
 be inaudible near the trumpet. 
 
 3. The arc of the horizon from which the echo appeared to come was 
 also increased in some cases to more than three times that subtended by 
 the echo at the place of the trumpet. This fact again indicates that the 
 echo consists of multiple sounds from various points at or near the sur- 
 face of the sea, the angle which the aggregate of these points subtend 
 necessarily becoming greater as the steamer advances. 
 
 But perhaps the most important facts in regard to the echo are those 
 derived from the series of observations on the subject, made by Mr. 
 Henry W. Clark, the intelligent keeper of the principal light-house sta- 
 tion on Block Island, and by Joseph Whaley, keeper of the Point Judith 
 light-house. Mr. Clark was furnished with a time-marker to observe the 
 duration of the echo, and both were directed to sound the trumpets every 
 Monday morning for half an hour, noting the temperature, the height of 
 the barometer, the state of the weather as to clearness or fog, the direc- 
 tion and intensity of the wind, and the surface of the ocean. 
 
 From the observations made at these two points, for more than two 
 years at one station and over a year at the other, the echo may be con- 
 sidered as produced constantly under all conditions of weather, even 
 during dense fogs, since at Block Island it was heard 106 times out of 
 113, and at Point Judith 50 times out of 57, and on the occasions when 
 it was not heard the wind was blowing a gale, making a noise sufficiently 
 intense to drown the sound of the echo. These results appear to be 
 sufficient to disprove the hypothesis that the phenomenon is produced 
 by an acoustic cloud accidentally situated in the prolongation of the axis 
 of the trumpet. It must be due to something more permanent in its 
 effects than that from a portion of air differing from that of the general 
 atmosphere in temperature or density, since such a condition cannot 
 exist in a dense fog embracing all the region of the locality of the phe- 
 
556 RESEARCHES IN SOUND. 
 
 iioinenon. Indeed, it is difficult to conceive how the results can be pro- 
 duced, even in a single instance, from a flocculent portion of atmosphere 
 in the prolongation of the axis of the trumpet, since a series of patches 
 of clouds of different temperature and densities would tend to absorb or 
 stifle by repeated reflections a sound coming from their interior rather 
 than to transmit it to the ear of the observer. 
 
 The question, therefore, remains to be answered : what is the cause 
 of the aerial echo ? As I have stated, it must in some way be connected 
 with the plane of the horizon. The only explanation which suggests 
 itself to me at present is that the spread of the sound which fills the 
 whole atmosphere from the zenith to the horizon with sound-waves 
 may continue their curvilinear direction until they strike the surface 
 of the water at such an angle and direction as to be reflected back to 
 the ear of the observer. In this case the echo would be heard from a 
 perfectly flat surface of water, and as different sound-rays would reach 
 the water at different distances and from different azimuths, they would 
 produce the prolonged character of the echo and its angular extent along 
 the horizon. 
 
 While we do not advance this hypothesis as a final solution of the 
 question, we shall provisionally adopt it as a means of suggesting further 
 experiments in regard to this perplexing question at another season. 
 
 GENERAL CONCLUSIONS. 
 
 From all the experiments which have been made by the Light-House 
 Board in regard to the transmission of sound in free air and those de- 
 rived from other observations which can be fully relied upon, the follow- 
 ing conclusions may be considered established, subject however to such 
 further modification and extension as subsequent investigation may 
 seem to indicate: 
 
 1. The audibility of sound at a distance (the state of the atmosphere 
 being constant) depends upon the character of the sound. The distance 
 through which a sound may be heard is governed by the pitch, the loud- 
 ness, and the quantity of sound. The pitch or frequency of the impulses 
 in a given time must not be too high, otherwise the amplitude of vibra- 
 tion will be too small to allow a sufficient quantity of air to be put into 
 motion ; neither must the pitch be too low, for in this case the motion of 
 the atoms of air in the sound-wave will not be sufficiently rapid to con- 
 vey the impulse to a great distance. Again, the greater the loudness 
 of the sound, which depends upon the amplitude of the vibrations of the 
 sounding-body, the greater will be the distance at which it will be heard. 
 And finally, the greater the quantity of sound, which depends upon the 
 magnitude of the vibrating surface, the greater will be the distance to 
 which it is audibly transmitted. These results are derived from observa- 
 tions on the siren, the reed-trumpet, and the automatic buoy. The effect 
 of quantity of sound is shown in the fact that in sounding different in- 
 
RESEARCHES IN SOUND. 557 
 
 struments at the same time, it was found that two sounds apparently of 
 the same loudness were heard at very different distances. 
 
 2. The audibility of sound depends upon the state of the atmosphere. 
 A condition most favorable to the transmission of sound is that of per- 
 fect stillness and uniform density and temperature throughout. This 
 is shown by the observations of Parry and other Arctic explorers ; al- 
 though in this case an efficient and co-operating cause is doubtless the 
 downward refraction of sound due to the greater coldness of the lower 
 strata of air, as first pointed out by Professor Eeynolds. Air however is 
 seldom in a state of uniform density, but is pervaded by local currents, 
 due to contact with portions of the earth unequally heated, and from 
 the refractions and reflections to which the sound-wave is subjected in 
 its passage through such a medium it is broken up and lost to the ear 
 at a less distance. 
 
 3. But the most efficient cause of the loss of audibility is the direct 
 effect produced by the wind. As a general rule, a sound is heard far- 
 ther when moving with the wind than when moving against it. This 
 effect, which is in conformity with ordinary observation, is not due to an 
 increase of velocity of the sound-wave in one direction and a diminution in 
 the other by the motion of the wind except in an imperceptible degree; for 
 since sound moves at the rate of about seven hundred and fifty miles an 
 hour, a wind of seven miles and a half an hour could increase or dimin- 
 ish the velocity of the sound-wave only one per cent, while the effect ob- 
 served is in some cases several hundred per cent. It is however due to 
 a change in its direction. Sound moving with the wind is refracted or 
 thrown down toward the earth ; while moving against the wind it is 
 refracted upward and passes over the head of the observer, so as to be 
 heard at a distance at an elevation of several hundred feet when in- 
 audible at the surface of the earth. 
 
 4. Although, as a general rule, the sound is heard farther when mov- 
 ing with the wind than when moving against it, yet in some instances 
 the sound is heard farthest against the wind ; but this phenomenon is 
 shown to be due to a dominant upper wind, blowing at the time in an 
 opposite direction to that at the surface of the earth. These winds are 
 not imaginary productions invented to explain the phenomena, but 
 actual existences, established by observation, as in the case of the 
 experiments made at Sandy Hook, in 1874, by means of balloons, and 
 from the actual motion of the air in the case of northeast storms, as 
 observed at stations on the coast of Maine. 
 
 5. Although sound issuing from the mouth of a trumpet is at first con- 
 centrated in a given direction, yet it tends to spread so rapidly that at 
 the distance of three or four miles it fills the whole space of air inclosed 
 within the circuit of the horizon, and is heard behind the trumpet nearly 
 as well as at an equal distance in front of its mouth. This fact precludes 
 the use of concave reflectors as a means of increasing the intensity of 
 sound in a given direction j for although at first they do give an increase 
 
558 RESEARCHES IN SOUND. 
 
 of sound in the direction of the axis, it is only for a comparatively short 
 distance. 
 
 6. It lias been established, contrary to what has formerly been thought 
 to be the case, that neither fog, snow, hail, nor rain, materially interferes 
 with the transmission of loud sounds. Tlje siren has been heard at a 
 greater distance during the prevalence of a dense and widely-extended 
 fog than during any other condition of the atmosphere. This may 
 however be attributed to the uniform density and stillness of the air at 
 the time. 
 
 7. In some cases sound-shadows are produced by projecting portions 
 of land or by buildings situated near the origin of the sound, but 
 these shadows are closed in by the spread of the sound-waves, and thus 
 exhibit the phenomenon of sound being heard at a distance and after- 
 wards lost on a nearer approach to the station. 
 
 8. It frequently happens on a vessel leaving a station, that the sound 
 is suddenly lost at a point in its course, and after remaining inaudible some 
 time, is heard again at a greater distance, and is then gradually lost as 
 the distance is farther increased. This phenomenon is only observed 
 when the sound is moving against the wind, and is therefore attributed 
 to the upward refraction of the sound-wave, which passes over the head 
 of the observer and continues an upward course until it nearly reaches 
 the upper surface of the current of wind, when the refraction will be re- 
 versed and the sound sent downward to the earth j or the effect may be 
 considered as due to a sound-shadow produced by refraction, which is 
 gradually closed in at a distance by the lateral spread of the sound- 
 wave near the earth, on either side, in a direction which is not affected 
 by the upward refraction. Another explanation may be found in the 
 probable circumstance of the lower sheet of sound-beams being actually 
 refracted into a serpentine or undulating course, as suggested in the 
 Appendix to the Report of the Light-House Board for 1875. (See page 
 513.) Such a serpentine course would result from successive layers of 
 unequal velocity in an opposing wind ; as being retarded at and near 
 the surface of the earth, attaining its maximum velocity at a height of 
 a few hundred feet, and then being again retarded at greater elevations, 
 by the friction of upper counter currents or stationary air. In some 
 cases the phenomenon is due to one or the other of these causes, and in 
 other cases to all combined. That it is not due to the obstructing or 
 screening effects of an abnormal condition of the atmosphere is shown 
 by the fact that a sound transmitted in an opposite direction, through 
 what is called the region of silence, passes without obstruction. It is 
 probable from all the observations, that in all cases of refraction of a 
 sound moving against the wind it tends again to descend to the earth 
 by the natural spread of the sound. 
 
 9. The existence of a remarkable phenomenon has been established, 
 which is exhibited in all states of the atmosphere during rain, snow, and 
 dense fog, to which has been given the name of aerial echo. It consists 
 
RESEARCHES IN SOUND. 559 
 
 of a distinct echo, apparently from a space near the horizon of fifteen or 
 twenty degrees in azimuth, directly in the prolongation of the axis of 
 the trumpet. The loudness of this echo depends upon the loudness and 
 quantity of the original sound, and therefore it is produced with the 
 greatest distinctness by the siren. It cannot be due to the accidental 
 position of a flocculent portion of atmosphere, nor to the direct reflection 
 from the crests of the waves, as was at first supposed, since it is always 
 heard except when the wind is blowing a hurricane. 
 
 As a provisional explanation, the hypothesis has been adopted that 
 in the natural spread of the waves of sound, some of the rays must take 
 such a curvilinear course as to strike the surface of the water in an 
 opposite direction and thus be reflected back to the station or location 
 of the origin of the sound. 
 
 LIGHT-HOUSE BOARD, October, 1877. 
 
 ^ 
 OF THE 
 
 UNIVERSITY 
 
 Of 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 AN INITIAL FINE OF 25 CENTS 
 
 WILL BE ASSESSED FOR FAILURE TO RETURN 
 THIS BOOK ON THE DATE DUE. THE PENALTY 
 WILL INCREASE TO SO CENTS ON THE FOURTH 
 DAY AND TO $1.OO ON THE SEVENTH DAY 
 OVERDUE. 
 
 
 
 MAY 7 fSta 
 
 
 ^*TV 
 
 
 MAY a6W 
 
 
 
 
 
 
 ~jg 
 
 
 S? 
 
 
 
 
 
 
 REC'D LD 
 
 
 DEC 191333 
 
 
 
 
 MfW go 1993 
 
 
 
 
 
 
 IAN 3 1^94 
 
 
 
 
 
 
 
 LD 21-20m-5,'39 (9269s) 
 
U.C.BERKELEY LIBRARIES 
 
 ___"'