106 [May 15, 



ence in the succession of positions of the bars, difference of time 

 allowed for rest, difference in the violence of the blows, &c. 

 The principal results appear to be the following : 



1 . The greatest amount of magnetism which a bar can receive, 

 appears to be such as will produce (on the average of bars) a com- 

 pass-deviation of about 11, the bar being 5 inches below the com- 

 pass. It was indifferent whether the bars rested on stone or on 

 wood, or whether they were struck with iron or with wood, the bars 

 lying on the dip plane while struck. 



2. When the bars, thus charged, lay on the plane transverse to 

 the dip, they lost about one -fifth of their magnetism in one or two 

 days, and lost very little afterwards. 



3. When the charge of magnetism is smaller than the maximum, 

 the diminution in a day or two is nearly in the same proportion as 

 for the maximum. 



4. The effect of violence on the bars, when lying on the plane 

 transverse to the dip, is not in all cases to destroy the magnetism 

 completely, sometimes it increases the magnetism. 



5. The Cold-Rolled Iron receives (under similar violence) or parts 

 with (under similar violence) a greater amount of magnetism than 

 the Hot- Rolled Iron, in the proportion of 6 to 5. 



6. There is some reason to think that the Hot-Rolled Iron has a 

 greater tendency to retain its primitive magnetism than the Cold- 

 Rolled Iron has. 



7. There is some reason to think that, when lying tranquil, the 

 Hot-Rolled Iron loses a larger portion of its magnetism than the 

 Cold-Rolled Iron loses in the same time. 



IV. "On the Analytical Theory of the Conic." By ARTHUR 

 CAYLEY, Esq., F.R.S. Received May 8, 1862. 

 (Abstract.) 



The decomposition into its linear factors of a decomposable quadric 

 function cannot be effected in a symmetrical manner otherwise than 

 by formulae containing supernumerary arbitrary quantities ; thus, for 

 a binary quadric (which of course is always decomposable) we have 



'' ^ ^0c- 



