140 PROCEEDINGS OF THE AMERICAN ACADEMY. 



affected by this so-culled von Waltenhofen phenomenon, we could 

 not discover the least difference in our results whether the rod to be 

 used was or was not surrounded by a thick cop])er tu))e, though the 

 tube makes the throws a trifle more regular. 



It is well known, also, that under low excitations, the magnetic 

 moment acquired by a rod in a solenoid under a given final excitation 

 may be much increased if the rod be constantly tapped while the mag- 

 netic changes are taking place. So far as we can make out this effect 

 is entirely lacking at very high excitations. We used a large electric 

 tapping apparatus made by Mr. Coulson to give many sharp blows 

 per second to a brass rod butted upon the specimen in the solenoid, 

 and Table VII shows characteristic results. 



Some years ago I encountered three specimens of very pure Norway 

 Iron, each of which showed a very high specific magnetism, when 

 tested by a modification of the Isthmus Method. Each piece was 

 about 8 cms. long and 1.26 cms. in diameter. They were presumably 

 from different sources. 



TABLE VIIL 



Each of these numbers comes from a series of closely consistent values 

 for different field strengths, and I have no reason to think that the 

 determinations were not good, but I consider the probable error some- 

 what greater in all work I have done with isthmuses than with such 

 experiments as I have made with larger specimens, with the help of a 

 solenoid. 



A single slender isthmus cut from the bar from which No. 2 in 

 this tal)le was taken, gave the very large value 1796 for I in fields 

 above 6000, but other larger pieces from the same bar showed lower 

 values for /. According to my experience very small bits taken from 



