20 SCIENCE IX SHORT CHAPTERS. 



smartly and simultaneously by the opposite sides of the choir, 

 the effect was very disagreeable. 



This defect, however, was not observable from the Press 

 Gallery, which is placed as nearly as may be to the focus of the 

 orchestral curve, so that radial lines drawn from the auditor to 

 different parts of the orchestra do not differ so much in length 

 as to effect perceptible differences in the moment at which the 

 different sounds reach the ear. 



My conclusion, therefore, is that if any amendment is to be 

 made in the numbers of the Handel Festival choir, it should 

 rather be done by a reduction than an increase ; that the four 

 thousand voices should rather be reduced to three thousand 

 than increased to five thousand. With greater severity of 

 selection as regards quality, power, and training of each indi- 

 vidual voice, and with better packing, the three thousand 

 would be more effective than the four thousand. 



