552 BELL SYSTEM TECHNICAL JOURXAL 



itself, since specialization is required — specialization of a type which 

 eventually disqualifies most men from undertaking other lines of 

 work most cfl'ecti\el\'. 



MaTIIKMATU S IX Kl.KCTKICAl. CoMMlXH ATIOX 



In order to make the foregoing observations somewhat more spe- 

 cific, I will refer to a few applications of mathematics in the industrial 

 research of the Bell Telejihone System. This field is selected because 

 I am more familiar with it than with other industrial activities. 



Certainty of prediction is the basic requirement in the develop- 

 ment and operation of the telephone system; no vital need of the 

 system can be left to chance or to fortuitous development. For 

 this reason, the Bell System is highly organized under research con- 

 trol. The telephone situation is studied as a whole; all depart- 

 ments cooperate; each problem is considered from every point of 

 view. Every attempt is made to master a situation in advance 

 of the necessity of action, so that the most effective and economical 

 means for electrical communication may be adopted W'ith each ex- 

 pansion of the system. Much more than the immediate require- 

 ments of the hour must be known; preparation for all eventualities 

 must be made. Fortunately, the executi\es have carried out this 

 program with a pro|)lK'tic api)rcciati()ii of the \alue and necessity of 

 mathematics. 



The importance of the theory and practice of probabilities was 

 recognized as soon as the telephone reached a thoroughly commercial 

 basis. It has prov'cd invaluable during the great expansion which 

 has already carried the number of telephones in the city of New 

 York to over a million. Meeting the pea'k load demand of the million- 

 odd telephones in New York City, on a practically no-delay basis, 

 with the minimum amount of e(iuipmcnt, is a highly complex and 

 important problem. Without probaliility studies of the situation, 

 the equipment installed at one point would be inadequate, while 

 at other points it would be superabundant. The superfluous equip- 

 ment would involve a waste of capital, while the inadequate ctjuip- 

 ment would mean inconvenience to the public and a loss of possible 

 revenue. K(|iiipment engineering involves a large number of prob- 

 ability [jroblems which arc novel, difficult, and financially most 

 important. The aggregate cost of all such studies is large, but the 

 resulting saving to the leleiihonc-using public is much greater. Satis- 

 factory telephone service in metropolitan areas is as deiicndrni ujjon 

 applied probability as is the success of life insurance. 



