A— MATHEMATICAL AND PHYSICAL SCIENCES 41 



The loudness of a noise depends of course on its remoteness and to a 

 less degree on its environment. The inverse-square law appears to be 

 followed fairly exactly in the open for average sounds such as those of 

 motor vehicles. In a room or a tunnel, the law does not hold and a noise 

 sounds louder than it would outside, owing to the building up of the 

 sound level by multiple reflection. Similar conditions prevail in a narrow 

 busy street and contribute to the noise discomfort of the occupants in 

 the upper stories, though to a pedestrian, the noise of, say, a passing car 

 is but little louder than in an open space. Incidentally, drivers of cars 

 are familiar with the sudden access of high-pitched components from the 

 . engines, exhausts or tyres of their cars as they pass reflecting walls or 

 fencing or even minor way-side objects such as tree trunks or telegraph 

 poles. 



One has also to remember that the path of a sound may appreciably 

 modify its composition. For example, the high-frequency components 

 may be abnormally reflected or absorbed as compared with lower notes 

 which tend to pass through or round obstacles. Even in the open, a 

 hedge row or a barrage of trees may, to a limited extent, so serve as a 

 muffler of traffic noise. The high-pitched components of a sound may 

 further be selectively enfeebled in passing over different types of ground, 

 for which the sound absorption may be three or four times as much for 

 high notes as for low. To judge by experience, the absorption figures for 

 newly fallen snow must be rather high, though I am not aware that they 

 have been measured. High-frequency components may also be selectively 

 absorbed by the air itself if it is humid. Knudsen has shown that the 

 eff'ect is due to interaction between the oxygen and water molecules, the 

 nitrogen playing no part. Incidentally, he estimates that if we lived in 

 an atmosphere of oxygen at a humidity of about 20 per cent., the high 

 notes of the violin and piccolo would be completely inaudible 50 yards away. 



Noise on the Railway. 



The background of noise (70-90 phons) which prevails in the com- 

 partments of most express trains is normally so near the borderline that, 

 even with the windows closed, conversation between ' diagonal-wise ' 

 passengers, though possible, is not ' comfortable.' The noise patently 

 originates from the rapid intermittent pounding of the steel tyres against 

 irregularities in the steel rail, largely supplemented by recurrent rail- 

 joint impacts. Much of the noise is of low pitch, and, as simple observa- 

 tion shows, it enters the coach partly through the floor, but mainly through 

 the glass of the windows. The remainder of the noise, which is of some- 

 what higher pitch and more directional, forces its attention on us whenever 

 the train passes a reflecting surface such as a platform, wall, or another 

 train, on which occurrence, the normal background of noise in a com- 

 partment is supplemented by a burst of higher-pitched noise deflected 

 through the windows. The conditions are aggravated (by as much as 

 10 phons) when a train is passing at high speed under a bridge or through 

 a cutting or tunnel, so that conversation then becomes difficult. 



An obvious palliative for rail-joint tap takes the form of longer rails : 



c 2 



