576 



NA TURE 



[October 28, 1922 



tions has recently been conducted by Mr. F. R. 

 Watson, also of Illinois University, which are described 

 in Bulletin No. 127 of the University. The results 



have led to conclusions somewhat at variance with 

 generally accepted ideas. 



Sound, on striking an object, is reflected, absorbed, 

 or transmitted, and usually all 

 three results occur. In any par- 

 ticular case a definite amount of 

 energy has to be got rid of in these 

 ways, and for sound - proofing one 

 may aim chiefly at reflection or ab- 

 sorption. When sound waves in one 

 medium encounter another medium 

 having a different density, the 

 progression of the waves is dis- 

 turbed, a certain amount of reflec- 

 tion takes place, some of the energy 

 is absorbed, that is, converted into 

 heat, while the amount transmitted 

 through the medium will depend on 

 its thickness and properties, such as 

 porosity and rigidity. 



In practice the materials used to 

 separate rooms or buildings are 

 usually of a complex character, and 

 their rigidity will depend not only on their nature and 

 thickness, but on the area of the separating wall. 



NO. 2765, VOL. I IO] 



Hence the problem of assessing sound transmission is 

 a very complex one. The author of the bulletin cited 

 directs attention to the very detrimental effect as 

 regards sound insulation of even 

 small apertures caused by ill-fitting 

 doors or by ventilators ; he also 

 makes a distinction between sounds 

 due to air waves striking a separat- 

 ing medium and vibrations such as 

 those caused by machinery, the 

 former best resisted by heavy and 

 rigid walling, the latter by arranging 

 for absorption of the vibrations by 

 beds of sand or like loose material. 



From a useful resume of previous 

 experiments on sound transmission, 

 the conclusion is drawn that rigidity 

 is a deciding factor in sound pre- 

 vention, and some experiments 

 recently conducted by Prof. P. E. 

 Sabine are cited which showed that 

 a plate of glass three-sixteenths of 

 an inch thick transmitted less sound 

 than two glass plates with a sheet 

 of celluloid sealed between them of 

 the same total thickness. A series 

 of tests made at the Music Building, 

 Chicago, in 1895, i s quoted, which 

 tends to show that an air space 

 between materials forming the two 

 sides of a partition is of much less 

 value for sound prevention than is 

 commonly supposed, and that 

 benefits which accrue from such 

 space are almost wholly negatived 

 by the inevitable connexion at 

 intervals for structural reasons 

 between the two sides. 



In Mr. Watson's experiments 

 use was made of the Rayleigh 

 disc resonator, which admits of much more accurate 

 and comparable results than are possible by 

 aural comparisons adopted by many earlier experi- 



JLCTION 



Fig. 3. — Section of Acoustic Research Building. 



menters. A very large number of materials were 

 tested, and these were in all cases of satisfactory area— 



