On Newton's Rule for Imaginary Roots. 363 



hydrochloric acid, and yields the same products as with baryta- 

 water. 



Husemann and Marme* have described the preparation of a 

 new base from the leaves and stems of the Lycium barbarum, L. 

 (Chinese Box-thorn). The aqueous decoction of the plant pre- 

 ferably of the leaves) was treated with acetate of lead, and the 

 filtrate from the precipitate treated with sulphuric acid to re- 

 move the lead, then neutralized with soda and evaporated. The 

 salts of the new base are so soluble that its precipitation was 

 impossible by the ordinary methods, and a new one was made 

 use of. This consisted in the addition of phosphomolybdate 

 of soda, with which the base forms an insoluble compound. The 

 precipitate produced in this way was decomposed with moist 

 levigated chalk, and the evaporated mass exhausted with boiling 

 alcohol. 



This alcoholic extract left, on evaporation, a deliquescent crys- 

 talline mass. To obtain the base in the free state, its sulphate 

 is decomposed with carbonate of baryta. The free base crystal- 

 lizes from water and from alcohol in small plates and prisms ; in 

 ether it is almost insoluble. Analyses of its salts give for it the 

 formula C^H^NO 2 ; it is isomeric with butalanine, the base 

 found by Gorup-Besanez in the salivary glands of the ox ; but 

 it differs from it materially ; and it most resembles sarcosine, 

 with which it is also homologous. Its poisonous action is in- 

 considerable. It forms a series of beautifully crystallizing salts 

 and double salts. 



LI. On Newton's Rule for Imaginary Roots, 



To the Editors of the Philosophical Magazine and Journal. 



Priory Cottage, Peckham, S.E., 

 Gentlemen, October 14, 1865. 



AT the top of page 291 of my paper in the last Number of 

 this Journal there is printed in italics a statement which the 

 remarks that follow were intended to show to be an admissible 

 truth. I now perceive, however, that the inference drawn from 

 those remarks at the bottom of that page does really assume 

 the principle contended for, and that that principle has not the 

 axiomatic character I had considered it to have. All therefore 

 that I have satisfactorily done, in reference to Newton's theorem, 

 is to show that the condition (1) at page 290 always implies a 

 pair of imaginary roots, and that, however numerous be the con- 



* Liebig's Annalen (Supplement), vol. ii. p. 383 ; and vol. iii. p. 245. 



