394 Special Reports 



Wolff, and the necessary accessory apparatus, with the exception of a large capacity 

 storage battery, located in the main building, was set up in the Standardizing Observ- 

 atory by Mr. Fleming and myself. With the help of Mr. Fisk and Mr. Peters a long 

 series of simultaneous determinations of the horizontal intensity was then made with 

 the galvanometer and C. I. W. standard magnetometer No. 3. The later and better 

 part of these observations, together with a more complete series made in August by 

 Messrs. Fleming, Fisk, and Ives, is briefly treated by Mr. Fleming in a summary 

 published in this volume (Absolute Standard in Horizontal Intensity, p. 468). 



Throughout the work with the sine galvanometer the electrical measurements were 

 made by the method (2) of 21. The galvanometer was used according to the method 

 described in 19, H being calculated in absolute measure from equation (17). For each 

 setting of the coil a number of scale readings were taken at regular intervals in order 

 to follow the variations of H more closely and to secure better mean values. 



The half-period of the magnet for the values of d and d' used, viz, from 65 to 74, 

 corresponding to currents from 0.14 to 0.15 ampere, is about a second, and the motion is 

 heavily damped. At the same time the circle and scale are clear and well illuminated and 

 can be read with the greatest ease. Hence, with constant current, which is not difficult 

 to secure, a complete determination can be made very quickly. 1 With practice several 

 such determinations can be made in one minute, provided scale-readings are not multi- 

 plied. In the observations referred to a determination with multiplied scale-readings 

 ordinarily required two minutes or less. The advantages over the magnetometer in 

 saving time and in obtaining more nearly instantaneous measurements are quite appar- 

 ent, as a complete determination with a magnetometer requires at least from half an hour 

 to an hour. 



As shown by Mr. Fleming's table, the results obtained with the two instruments 

 agree with extreme closeness, the mean discrepancy being only 0.7?, or about 1 part in 

 25,000. Any agreement beyond 1 part in 10,000, or thereabout, would, however, have 

 to be considered accidental even if the electrical standardizations were given with the 

 greatest precision practicable; and such standardizations have not yet been obtained. 

 According to the Bureau of Standards, the certified e. m. f. of the Weston cell used is 

 correct to 1 part in 10,000; the certified resistance of the standard coil is correct within 

 1 part in 20,000; and the potentiometer can be relied on to within 1 part in 10,000. 

 Hence it is known only that the currents are not in error by as much as 1 part in 4,000. 

 It is unlikely, however, that the three contributing errors are all effective in the same 

 direction; and the standardizations are probably, as is usually the case with the Bureau 

 of Standards, more precise than the claims of the certificates. Hence it is probable 

 that the actual error in the currents is considerably less than 1 part in 4,000; but it will 

 be impossible to say how much less until more precise values of the standards have been 

 obtained. 



24. This work was undertaken at the request of Dr. Bauer, who has taken great 

 interest in its progress and who has seen that adequate facilities for it were provided. 

 My colleague, Mr. Fleming, has also taken great interest in the work and has made some 

 useful suggestions. The non-magnetic castings were made by Mr. C. Huff, who did most 

 of the work on the tripod and the magnetometer supports. The chief instrumental work, 

 including the most difficult parts, viz, the turning of the marble spool and the construc- 

 tion of the measuring devices, was done with great skill and patience by Mr. G. H. Jung, 

 whose ingenuity relieved me of many details. Dr. G. H. Wait and Dr. J. E. Ives have 

 done 1 lie greater part of the work involved in making and reducing the linear measure- 

 ments. 1 am indebted also to Mrs. Barnett for assistance in many parts of the work, 

 and to Mr. C. A. Kotterman for the illustrations. 



'When :i long series of observations is in progress, the work can be facilitated by using stops to keep 0+0' constant, 

 the variations in // for constant current being found from the scale-readings. 



