45G JOURNAL Olf THE WASHINGTON ACADEMY OF SCIENCES VOL. 11, NO. 19 



dom orientation with respect to each other. Between the crystalline 

 grains there are localities in which the atoms are not regularly ar- 

 ranged, but the metal is more or less amorphous in character, and 

 in these localities there is a residual resistance that persists to ab- 

 solute zero. The curve of resistance against temperature of such 

 a metal becomes asymptotic at absolute zero to a low but finite re- 

 sistance. Suppose, however, the metal is one having a polymorphic 

 transition in the region of low temperatures. Normally the growth of 

 the polymorphic form will start from a single nucleus, so that after 

 the transition the wire will consist of a single crystalline grain with- 

 out amorphous regions. On raising the temperature again beyond 

 the transition point the wire recovers its original crystalline structure 

 and its original resistance. The recovery of the original structure 

 is of course evidence of some sort of residual nuclear structure in 

 the low-temperature modification. This appears at first sight not 

 what one would expect, but it is exactly what is found in the transition 

 of ordinary polymorphs at higher temperatures. 



Apart from its intrinsic probability and the fact that this view 

 does not need to invoke a new mechanism to explain the discon- 

 tinuity, there is considerable evidence in its favor. In the first place 

 the original belief that supraconductivity was in some way connected 

 with the purity of the metal, and that only those metals would show 

 that it could be obtained in a state of extreme purity, has not turned 

 out to be true. It has been found that amalgams show the effect, 

 and that gold, which has been obtained in a higher state of purity 

 than lead or tin, does not show the effect. 



The appearance of supraconductivity is to a certain extent ca- 

 pricious, and occasionally samples are prepared which for some un- 

 known reason do not become supraconducting. This is to be ex- 

 plained by some accident of structure or handling which caused 

 the growth of the polymorphic form to start from more than one 

 nuclear center, so that regions of separation between the different 

 grains persist beyond the transition point. 



The temperature of the discontinuity of resistance is known to 

 be depressed when the current in the conductor is increased, and 

 also when the conductor is placed in a magnetic field. Silsbee'* has 

 shown in this Journal that the probability is high that the phenomena 

 are not unrelated, but that the threshold value of the current is that 



at which the magnetic field within the wire due to the current itself 

 *F. B. SiLSBEE. This Journal 6: 597-G02. 1916. 



