EXPERIMENTAL METHODS IN WATER- AND SEWAGE-WORKS 27 



spent more than $50,000. More recently their devices, when tested, 

 have been purchased at the beginning. 



While this paper is devoted essentially to methods of purifying 

 water and sewage by works partly or wholly of artificial construc- 

 tion, mention should be made of the important advances in the 

 allied field of water supply from storage reservoirs, and the disposal 

 of sewage by dilution. Among the more prominent investigations 

 of this kind should be stated those field surveys and laboratory 

 studies made in connection with the Chicago Drainage Canal, the 

 additional water supply of New York City, the improvement of 

 the Charles, Mystic, and Neponset Rivers in Massachusetts, and 

 foreshore pollution along the Massachusetts coast. Several hundred 

 thousands of dollars have been expended on these investigations. 



OBJECT AND ADVANTAGES OF EXPERIMENTAL METHODS. 



The purposes of applying experimental methods to problems 

 of water and sewage purification are chiefly threefold, as follows: 



1. To provide data for the official and technical authorities, 

 to enable them to adapt new works most advantageously to the 

 local conditions, and to indicate dimensions and other physical 

 conditions permitting contract plans to be prepared and the cost 

 of construction to be approximately estimated. 



2. To educate the non-technical public, who as citizens and 

 taxpayers are interested in public works. 



3. To provide data so that the officials can operate effectively 

 the works after they are completed, and forecast the approximate 

 cost of operation. 



Technical data. In regard to the first object accomplished, 

 that of enabling city officials and their technical advisers to design 

 economically works of a suitable character, it goes without saying 

 that this has been of the greatest importance, and is a strong factor 

 in explaining the rapid strides in successful sanitary works accom- 

 plished during the past few years. It has frequently been the advice 

 of technical men, in dealing with problems which differ from those 

 successfully solved elsewhere, to make tests for a year or so at a 

 cost approximating the interest for one year for the works contem- 

 plated. In this way the cost of errors and unbalanced designs 



