February 2, 1900.] 



SCIENCE. 



183 



the parallels at right angles, nor do they meet 

 in one point, the apex of the developed cone. 

 By sacrificing the equidistance of the parallels, 

 conical projections may be constructed which are 

 either equivalent or conformal (without angu- 

 lar distortions) and yet retain either one or two 

 true parallels. Lambert and Gauss have de- 

 voted considerable study to these projections ; 

 but there is one, devised by Albers in 1805, 

 ■which has equivalence and two true parallels, 

 qualities which should entitle it to special con- 



Association ' should persist in making use of 

 Bonne's projection in their reports. 



Regarding the polyconic projection, devised 

 by the Coast Survey and very extensively used 

 in this country, Blundau has not much to say 

 that might be considered as very flattering. He 

 says in substance that the distortion which in 

 the equidistant conic projections is most per- 

 ceptible near the upper and lower borders is 

 here shifted to the more distant parallels, and 

 that it is of real advantage only when applied 



sideration, but hitherto it has not received any 

 special trial. Amongst the pseudo or conven- 

 tional conical projections, Bonne's occupies the 

 first rank ; although as long ago as 1880, Tissot 

 exposed its glaring defects, it has nearly up to 

 the present date retained almost undisputed 

 possession of the principal atlases. This has 

 sometimes been ascribed to undue French influ- 

 ence, but it may just as well be a consequence 

 of custom and convenience. It is somewhat 

 surprising that after all that has been said by 

 Tissot and Zoppritz the ' Geodetic International 



to the, representation of regions of predomina- 

 ting meridional dimensions. It is a fact that In 

 maps of the United States on this projection, like 

 those issued by the General Land Office, the 

 Geological Survey and the Census, the exagger- 

 ation of the meridians and areas near the At- 

 lantic and Pacific borders reaches fully 6^ per 

 cent., which is nearly three times as much as 

 might be considered a fair allowance, and it 

 seriously interferes with the use of these maps for 

 the measurement of either distances or areas 

 near these borders. But there are several 



