On the Theorems relating to " Canonic Roots." 453 



"Howardite" class, and the specific gravity as recorded by 

 other observers (3*07), are^quite inconsistent with our very highly 

 chondritic stone called by this name being a true specimen of the 

 Loutolax fall. Its specific gravity is 3 , 671-3 , 674. Mr. Greg 

 imagined it might be a specimen of Timochin erroneously labelled. 

 But it differs from that aerolite very widely in the character of 

 its spherules, which are singularly round, numerous, and distinct, 

 — and also in the rust-stains — a very good characteristic where 

 aerolites exhibit them, as the mode of distribution of the stain 

 depends on the nature of the porosity of the stone and on the 

 mode of dissemination of the iron, which are both important aud 

 constant features. 



The iron in both is very similarly distributed ; but the stains 

 are more patchy in the Timochin stone, and the spherules are 

 not so sharply relieved by their colourless character from the 

 rust around them. Their specific gravities are also very nearly 

 the same, that of Timochin being 3*636. 



I cannot as yet find any meteorite quite like the stone in the 

 British Museum, though, until I had sections made, I believed 

 Barbotan to be so. It is, however, a very different stone. I 

 should make one remark which I do not think is without im- 

 portance. This stone was obtained from that sound mineralo- 

 gist and generally accurate observer Heuland. His original 

 description of it, preserved in the archives of the Museum, de- 

 scribed it as a stone that fell at TViborff in March 1814; whereas 

 the stone known as Loutolax fell on December 13, 1813. Heu- 

 land does not mention the name Loutolax. It is not impossible 

 that aerolitic falls took place in the neighbourhood ofWiborg on 

 these two several dates ; but until some further facts are ascer- 

 tained on the subject, either confirming or condemnatory of this 

 hypothesis, I shall place the " Wiborg " stone of the Museum 

 among the many specimens of problematical aerolites. 



LXI. Sequel to the Theorems relating to "Canonic Roots" given 

 in the last March Number of this Magazine. By J. J. Syl- 

 vester, F.R.S.* 



THE theorems kindly communicated from me by Mr. Caylcy 

 in the March Number of this Magazine were originally 

 designed to appear as a note or excursus to a memoir in prepa- 

 ration on the extension of Gauss's method of approximation from 

 single to multiple integrals by a method which invariably leads 

 to the construction of a canonizant whose roots are all real. To 

 establish this reality, recourse may advantageously be had to a 

 * Communicated by the Author. 



