DEPARTMENT OF TERRESTRIAL MAGNETISM. 299 



The improvement shown by this record results jointly from the various 

 instrumental modifications referred to above and from the substitution 

 of the chloride-of-silver batteries for those previously employed. 



INSTRUMENT AND BUILDING WORK. 

 INSTRUMENT WORK. 



The work in the instrument shop during the year may be roughly 

 classified as follows: equipment, 41 per cent; improvements and repairs 

 of instruments, 41 per cent; experimental and survey, 9 per cent; 

 miscellaneous and stock, 9 per cent. 



Much of the instrumental work was concerned wdth the study and 

 design of the T. M. compass- variometer in an improved form for ship 

 work and of an inertia-gimbal system planned to eliminate dynamic 

 deviations and magnetic deviations in instruments on shipboard. 

 Much time was devoted also to consideration and development of 

 the use of this instrument as an intensity variometer suitable for the 

 rapid survey of locally-disturbed regions and in the application of the 

 principle involved to the use of intensity variometers for observatory 

 work. 



The improvements and repairs of instruments had to do largely with 

 the overhauling, remaking, and modifying of the magnetic, atmos- 

 pheric-electric, meteorologic, and miscellaneous equipment needed for 

 the Carnegie work. Opportunity was presented for the first time to 

 make careful magnetic tests of sea dip-circles of the type constructed 

 by Dover. It was found that the vertical circle of No. 189 was made 

 of German silver and was polarized. This doubtless explains to some 

 extent the erratic behavior of this dip circle, as indicated by the irregu- 

 lar and large corrections obtained for different needles by comparisons 

 wath standardized earth-inductors at stations in different magnetic 

 latitudes. AMien remaking the dip-circle, a new graduated vertical 

 circle of fine silver, carefully tested and found to be non-magnetic, was 

 supplied. 



The observations with the marine earth-inductor have heretofore 

 been made, using a balanced moving-coil marine galvanometer. The 

 unanimous opinion of our observers, borne out by study of the results, 

 has been that the accuracy of resulting values for inclination was limited 

 by the inaccuracies of the galvanometer readings. The difficulty arises 

 apparently from unavoidable change of balance with change in time 

 and temperature and jarring caused by the ship's motion which can not 

 be effectively controlled ; much time is lost in the tedious operation of 

 rebalancing. Accordingly, the string galvanometer (see p. 313), 

 designed and built by the Department, was arranged for use with the 

 marine earth-inductor on the Carnegie. Since the fiber of the gal- 

 vanometer is quite fine and its inertia thus practically negligible, it is 

 hoped that the resulting inclinations obtained with the earth inductor 



