472 



SCIENCE 



[N. S. Vol. XXIX. No. 742 



0.82 to 1.31, in a field whose intensity was 400 

 units, which is the intensity usually found in 

 ordinary galvanometer fields. 



By plotting the relation between M/T and the 

 field strength when the values of the latter were 

 both increasing and decreasing, a marked hystere- 

 sis was found, which explain the hysteresis ob- 

 served in galvanometer defleetions whose magni- 

 tude depends somewhat on the direction from 

 which the coil comes to its deflected position. 



The relation between the strength of the mag- 

 netic field and the " set in the fiber " obtained 

 after a reversed deflection was found to be pro- 

 portional to the strength of the field, except that 

 in weak fields there was no observable set. This 

 shows, as previously explained by the writer, that 

 the set is practically all due to a change in the 

 strength and the direction of magnetization of the 

 impurities in the coil. This magnetization grad- 

 ually returning to its normal strength and direc- 

 tion explains also, in part at least, the shifting 

 of the zero point with time. 

 The Three Temperature Coefficients of the Moving 



Coil Galvanometer and their Relation to the 



Temperature Coefficients of its Various Parts: 



Anthony Zelent and O. Hovda, University of 



Minnesota. 



Tlie values of the temperature coefficients for 

 galvanometers having cliilled cast-iron magnets 

 are given in the following table, where B is the 

 temperature coeflTicient of a particular circuit. 



Measurement 

 Current, ■! 



Suspension 



Phos. bronze, 

 Steel, 



-f 0.00018 

 -1-0.00005 



Potential ; Phos. bronze, 4-0.00018-J? 

 rotenuai, \ g^eel, -|- 0. 00005 — J? 



Ballistic, < 



Phos. bronze, —0.00017 



Steel, 



-0.00017 



The temperature coefficient for current meas- 

 urements is shown to be 



dk' = Fk + t„ + L,c-D^, 



(1) 



where die', Ftc, tje, Lk, Die, are the temperature co- 

 efficients respectively for deflections, field strength, 

 period of vibration of the coil, and the linear ex- 

 pansion of cast iron and of copper. 



Tlie temperature coefficient for potential meas- 

 urements can be calculated from 



dk" =^ du' — B, 



(2) 



where B, as given above, is the temperature co- 

 efficient for the resistance of a particular circuit. 

 The temperature coefficient for ballistic throws 



die ^ die — ik* 



(3) 



These equations enable any one of the three 

 temperature coefficients to be calculated from the 

 known temperature coefficients of the various 

 parts of the galvanometer. 



The temperature coefficients of a galvanometer 

 with a magnet other than chilled cast iron can be 

 calculated from 



iC' = ^-f (i^/c' -I- 0.00040), (4) 



where K represents the value of any particular 

 coefficient given in the above table, corresponding 

 to the one desired, and Fh' is the temperature 

 coefficient of the field strength for the magnet of 

 the galvanometer whose temperature coefficient is 

 to be determined. 



A New Method for the Aisolute Measurement of 



Resistance: E. B. Eosa, Bureau of Standards, 



Washington. 

 A Plea for Terrestrial and Cosmical Physics: L. 



A. Baueb, Carnegie Institution, Washington. 



This paper will be published in full in Science. 



The Ellipticity of the Earth is Not a Proof of a 



Former Liquid State: John F. Hatford, Coast 



and Geodetic Survey, Washington. 



The idea is often expressed, even by physicists 

 of high rank, that the observed ellipticity of the 

 earth is a proof of a former liquid state. This 

 idea is based upon a gross misconception of the 

 magnitude of the stresses which would be pro- 

 duced within the earth by any departure of the 

 actual ellipticity from the value corresponding to 

 the rate of rotation. Sir George Darwin has com- 

 puted that a departure of only one seventh part, 

 of the actual ellipticity from that corresponding 

 to the rotation, would produce stress-differences 

 in the interior of the earth as great as five tons 

 per square inch. Even the best granite will ordi- 

 narily fail under a stress-difference less than five 

 tons per square inch. Therefore, unless the 

 earth in its inner parts is stronger than the best 

 granite it will yield to the stresses and take a new 

 shape before the actual ellipticity has departed 

 from that due to the rotation by as much as one 

 seventh part. 



Any one who will start from this as a basis 

 and consider the improbability of the earth being 

 as strong as the best granite throughout, even if it 

 is solid, consider the improbability of the ma- 

 terial in the earth being incompressible under 

 stresses applied continuously for ages, and con- 

 sider the uncertainty introduced into the evalua- 

 tion of the theoretical ellipticity due to rota- 

 tion on account of this evaluation being af- 

 fected by the assumed relation of depth and 



