420 



SCIENCE. 



[N. S. Vol. XXIII. No. 585. 



than others, the change of resistance is 

 due to the formation of crystals in unstable 

 equilibrium. 



Another theory is that in the selenium 

 cell there is a form of selenium called met- 

 allic, which conducts electricity well and 

 which is a sort of solution with the non- 

 conducting selenium. Light causes the 

 metallic selenium to make better contact 

 and thereby reduces the resistance. 



As selenium has a coefficient of expansion 

 about five times that of ordinary metals, 

 the author was led to the study of pres- 

 sure effect, thinking that the change of 

 resistance might be due to contact differ- 

 ences. But this is, at present, only a 

 theory. 



Elastic After-effects in Crystals: J. R. 



Benton, Geophysical Laboratory of the 



Carnegie Institution. 



The elastic properties of solid bodies 

 vary with different specimens of the same 

 substance, and in the same specimen when 

 it is subjected to varying preliminary treat- 

 ment. There is reason to believe that the dis- 

 crepancies are due to irregularity of struc- 

 ture, such as is Imown to exist in metals 

 and many other solids. If this explanation 

 is correct, there should be no irregularity 

 in the elastic behavior of single crystals. 

 To test this, experiments were planned for 

 observing the elastic after-effect, elastic 

 hysteresis and permanent set, in crystals. 

 The present paper describes the first part 

 of these experiments, which deals with 

 elastic after-effect. Observations were 

 made on the torsion of mica, and on the 

 flexure of selenite, kunzite and rutile; they 

 show that the elastic after-effect is not en- 

 tirely absent, as was hoped would be the 

 case, but is very small as compared with 

 that in most solids. The reason why it ap- 

 pears at all pi'obably lies in the fact that 

 absolutely perfect crystals can not be se- 

 cured for the experiments. 



The Percentage Bridge: A. C. Longden, 



Knox College. 



Five or six years ago, a paper entitled 

 'A Percentage Bridge' was presented to 

 the American Association for the Advance- 

 ment of Science, and also to the American 

 Institute of Electrical Engineers, by Mr. 

 H. C. Parker, of Columbia University. 



The instrument as described by Mr. 

 Parker is essentially a four-gap slide wire 

 bridge in which the two inner gaps are 

 used for the comparison of resistances, and 

 the auxiliary resistances in the outer or 

 end gaps have such a relation to the resist- 

 ance of the bridge wire that a change of a 

 hundredth of one per cent, in the ratio of 

 A'j to ^2 shall produce a change of one 

 millimeter in the position of the balancing 

 point on the bridge wire. The test coil, 

 however, is not balanced directly against 

 the standard, but one of the resistance gaps, 

 2?i, for example, is used as a substitution 

 gap, and a standard resistance in this gap 

 is balanced against a resistance approxi- 

 mately equal to it, and then the test coil is 

 substituted for the standard and the bridge 

 is again balanced. The distance in milli- 

 meters on the bridge wire between the two 

 balancing points indicates the difference be- 

 tween the two coils in hundredths of one 

 per cent. 



This method seems to have a number of 

 real advantages over the Carey Foster 

 method for comparing standard resistances. 



The simplicity of the percentage method 

 is greatly in its favor and ought at least 

 to entitle it to serious consideration. It 

 does not eliminate the resistance of the end 

 connections, nor does it necessarily make 

 them so small as to be negligible, but it 

 makes the total value of the end resistances 

 so large that even if they differ by a hun- 

 dredth of an ohm, the error in the result 

 will only be one part in twenty million! 



A mercury commutator is suggested for 



