BY R. ghek; smith. 143 



when a yellow precipitate wliich melted at 176-178" separated 

 out. A vitreous yellow residue was obtained upon evaporating 

 the mother-liquor. When dissolved in water and evaporated, the 

 same vitreous residue remained. 



The third hydrolysis (5 % acid for 1 hour) yielded two fractions. 

 The first was a brown powder which when treated with hot dilute 

 acetic acid, to dissolve the brown constituent, furnished, upon 

 cooling, rosettes of microscopic needles which melted at 130^. 

 A dirty-brown vitreous residue was obtained upon evaporating 

 the mother-liquor. The second osazone was a yellow powder 

 which melted at 174-176°. Ether extracted a constituent which 

 dried as reddish-yellow crusts and when evaporated from aqueous 

 solution gave a slight yellow powder and a brown tar. Follow- 

 ing the ether treatment, the yellow powder was heated with 

 water upon the water-bath and filtered; the residue was again 

 heated with water and filtered. Both hot solutions deposited, 

 upon cooling, yellow precipitates of small microscopic sj^heres 

 which dried as brown skins. The first melted at 162-164"; the 

 second melted at 170\ The ftrst was heated with water and 

 filtered, when stellate tufts of pale yellov) needles tnelling at 162" 

 settled out. The residue, insoluVjle in the hot water, melted at 

 166°. The residue of the original yellow powder not dissolved 

 by the repeated treatment with hot water was dissolved in 

 alcohol, heated until much alcohol had evaporated, and then 

 cooled. The osazone that settled out melted at 182°, while the 

 residue oVjtained upon the evaporation of the mother-liquor 

 melted at 176*^. The osazone (182^) was treated with cold strong 

 alcohol which dissolved a constituent melting at 179°, leaving 

 undissolved a lemon-yellow osazone melting at 186-187^. An 

 osazone with a melting point higher than this could not be 

 obtained. 



The italics in the last paragraph indicate the osazones which 

 approached most closely to arabinosazone and galactosazone. 

 Although they simulated these compounds, there can be no doubt 

 that they were other substances, for it is a comparatively easy 

 matter to separate the definite osazones of arabinose and galac- 



Ilu LI BR AR Y 



