Distribution of Stations 23 



could not be carried out, the data for some one element being in consequence 

 lacking. Practically all of the secondary stations lack two of the magnetic ele- 

 ments, either because they were generally established merely for investigating the 

 possible existence of local disturbance, or because the time available was insuf- 

 ficient for complete observations. 



Of the 204 "C.I.W. repeat localities" listed in Table 6, which furnish secular- 

 variation data for localities previously visited by observers of the Department of 

 Terrestrial Magnetism, the reoccupations for each locality listed involve from 1 

 to 4 stations and may be classified as follows: exact, 94; close, i. e., within less 

 than 30 meters, 74; practical, i. e., within less than 300 meters, 18; and proximate, 

 i. e., within less than 5 kilometers, 18. For many of these localities the repeat 

 observations have been obtained not only at several stations, but also at different 

 times during 1914 to 1920. Furthermore, fully 150 of the stations were points at 

 which the magnetic elements had been determined previously by other organiza- 

 tions or observers; about one-half of these were reoccupations within 300 meters 

 and the remainder within 5 kilometers. Thus secular-variation and correlation 

 material result from over 20 per cent of the data given in the present volume. 

 The stations include those occupied for intercomparison observations at 17 mag- 

 netic observatories; several of these have been occupied more than once, thus 

 affording further data regarding the question as to the degree of accuracy within 

 which the instrumental constants can be maintained under strenuous field condi- 

 tions. The results of the comparisons of magnetic standards made at these obser- 

 vatories will be given in a later special report. 



The Department of Terrestrial Magnetism has continued to furnish instru- 

 mental and other assistance in cooperating with various organizations. The 

 "Maud Expedition" of Captain Roald Amundsen, begun in 1918, was supplied 

 with a complete magnetic outfit, including specially adapted magnetometer 

 and dip circle (see page 8 and Plate 2), accessories, instrumental constants, 

 and detailed instructions for proposed work. There has thus already been ob- 

 tained the valuable series of observations along the north coast of Siberia to be 

 found in the Table of Results. The Department likewise has loaned a magnet- 

 ometer to Government- Astronomer G. F. Dodwell of South Australia; Mr. Dod- 

 well and his assistants have observed at numerous stations in South Australia, 

 the data obtained being given in the Table of Results. Extensive cooperative 

 work in connection with special observations, as, for example, during solar eclipses, 

 has also been accomplished. 



CONCERNING GEOGRAPHIC POSITIONS. 



Full use in theoretical discussions of accurate magnetic observations requires that 

 the geographic coordinates of stations be known with a fair degree of accuracy (see 

 Volume I, pp. 22 et seq.). The determination of latitude is comparatively simple, and in 

 general, as already stated for the methods followed (see p. 5), the error in this coordinate 

 is usually less than 0'.5, and usually within about 0'.2. The determination of longitude, 

 on the other hand, is subject to a greater uncertainty. 



