January 13, 1905.] 



SCIENCE. 



47 



at least have great historic interest are 

 utterly lost to us, because we can not make 

 even a resijectable guess at the units used. 

 The adoption of an international luiit of 

 length and its necessary auxiliary, a coiu- 

 iiion thermometric scale, and the provision 

 which the various governments made for 

 the reference of their measuring apparatus 

 to a common unit was a step of funda- 

 mental importance. 



The association as such has no control 

 over the geodetic operations conducted by 

 the different governments. Its function is 

 to be the intermediary where cooperative 

 action is needed, and to discover and point 

 out along what lines the greatest need for 

 information exists. 



In pursuance of these duties it has helped 

 to perfect the European systems of trian- 

 gulation by showing where missing links 

 should be supplied, not only by measure- 

 ment of angles and bases but also by addi- 

 tional astronomical observations. It has 

 made absolute gravity determinations with 

 all the accuracy demanded by modern sci- 

 ence and has caused suitable connection to 

 be made by relative measures between wide- 

 ly scattered pendulum base stations, and it 

 has instituted unicjue relative gravity meas- 

 ures, to which further reference will be 

 made. It organized and maintains the sta- 

 tions for observing the variation of lati- 

 tude in regard to which it should be re- 

 marked that it is the desire of the associa- 

 tion to continue the observations beyond 

 the year 1906, which marks the end of the 

 ten-year period for which that service was 

 tentatively organized. The association 

 strongly desires not only to continue but 

 to extend the service to the southern hemi- 

 sphere and other latitudes than those now^ 

 occupied by the permanent stations, and to 

 obtain the cooperation of suitably situated 

 observatories in their endeavor to discover 

 the cause of the phenomena. 



That the problem of determining the 



earth's dimensions could not be solved by 

 simply measuring two arcs in suitable lo- 

 calities was brought home to geometers by 

 the anomalous results obtained in the eight- 

 eenth century. For instance, according to 

 Lindenau the combination of the two Amer- 

 ican arcs, Mason and Dixon's, measured in 

 1764, and that of Peru, measured a quarter 

 of a century earlier, gave a value of one 

 five-hundredths for the earth's compression. 

 The value derived from those measured in 

 Great Britain alone was about nine times 

 as great, or one fifty-fifth, while those 

 made in France, considered by themselves, 

 gave one one-hundred-and-fiftieth. It is 

 not now important to inquire whether these 

 differences are not in part due to the crudi- 

 ties of the methods of measurement em- 

 ployed. They were sufficiently real to 

 throw doubt on the belief that the earth 

 could be represented by a regular mathe- 

 matical figure. Finally, the existence of 

 local deflections of the vertical as affecting 

 the amplitude of arcs was recognized, but 

 not taken into account save, perhaps, by 

 arbitrary exclusion of stations showing ex- 

 ceptionally large deflections. 



The method of finding an osculating 

 spheroid from arc measures remained in 

 its essence that of taking averages of meas- 

 urements reduced to the geoidal surface. 

 The differences between the observed direc- 

 tions of the vertical and those computed on 

 an assumed spheroid of reference were 

 treated as if they were accidental errors of 

 observation. At the present time it is the 

 aim of geodesists to assign to the deflections 

 their proper place in the computations and 

 to interpret them by discovering through 

 them and through gravity measurements 

 the manner of the distribution of masses in 

 the interior of the earth. Thus geodesy is 

 trenching on the domain of geophysics and 

 geology. 



In India, in Europe and in the United 

 States the study of these deflections is re- 



