SCIENTIFIC NEWS. 17^, 



in rendering It as perfectly white as that of my former experi- Sugar of 

 pents. I have remarked, that the sugar which forms granular^ 

 crystals is easy to be divided, and yields readily to theoperation 

 of claying: but it is not so with what reiuains thin, fat, like 

 honey, and consequently imperfectly freed from its syrup ; the 

 humidity of the argil penetrates it too slowly, and it dissolves 

 and carries away too much sugar. All the efforts of the 

 sugar-baker should therefore, in my opinion, be d. reeled to- 

 wards obtaining the crystallization in the most granulated 

 state possible ; and I am encouraged to hope that the difti- 

 ciilties which may appear in the way of our object will be 

 surmounted, as grape sugar crystallizes considerably quicket 

 than that made from cane : The candy which we see so fre- 

 quently on sv,/eetmeats of every kind affords a daily proof of 

 this. Ail that i have examined for several years past in the 

 confectionary of my own house, was a sugar perfectly analo- 

 gous to that from grapes, without admixture of that froitt 

 cane. The latter is therefore much less crystallizable. 



Grape sugar is not so white, I must repeat it, as that from 

 cane ; but its Havo\ir is full, pure, and \Vithout the least 

 remaiiis of a ven;etable taste or smell. 



XII. 

 SCIENTIFIC NEWS» 



Magnetical Telescope, 



R. Edward Trough tori has constructed a new tclcscbpe for Mr. Trough- 

 determining the magnetical meridian. It consits of a tube of *7a magneu" 

 steel, containing a set of lenses with cross wires or spiders the foi-m of a 

 weebs, in the usual manner. It will easily be understood that a'J''a^eSope, 

 an instrument of this kind, after receiving the magnetic power, 

 may traverse upon pivots or by any other similar mode of sus- 

 pension, and will dispose itself in the magnetic meridian. One 

 of the difficulties attending the magnetic bar of the usual form 

 is, that its line of direction may not be parallel to its side; 

 .?.nd it is not easy io determine the quantity of error, by re- 

 A a 2 versing 



