With solutions twice as dilute as those above, the relative conditions are still more sensible, and the different sugars arrange themselves in the same order, with the exception of levulose, which shows a slight advantage over sorbine and acquires the first rank. From the above determinations, it is shown that levulose and sorbine give vastly greater quantities of osazones, under given conditions, than the other reducing sugars. It would be easy, therefore, to distinguish them by this reaction and to recognize their presence also even in very complex mixtures, where the polarimetric examination alone would furnish only uncertain indications. It is remarkable that these two sugars are the only ones among the isomers or the homologues of dextrose, actually known, which possess the functions of an acetone. They are not, however, easily confounded, since the glucosazone forms beautiful needles which are ordinarily visible to the naked eye, while the sorbinosazone is still oily and when heated never gives perfectly distinct crystals. This method also enables us to distinguish between dextrose and galactose, of which the osazone is well crystallized and melts at almost the same temperature as the phenylglucosazone. Finally, it is observed that the reducing sugars give less of osazones than the sugars which are not capable of hydrolysis, and consequently differ in their inversion products. It is specially noticed in this study of the polyglucoses (bioses, trioses), that this new method of employing the phenylhydrazin appears very advantageous. It is sufficient to compare the weights of the osazones to that which is given under the same conditions by a known glucose, in order to have a very certain verification of the probabilities of the result of the chemical or optical examination of the mixture which is under study. All the polyglucoses which have been examined from this point of view give very decided results. The numbers which follow have reference to one gram of sugar completely inverted by dilute sulfuric acid, dissolved in 100 cubic centimeters of water, and treated with two grams of phenylhydrazin, the same quantity of acetic acid, and five grams of crystallized sodium acetate. All these solutions have been compared with the artificial mixtures and corresponding glucoses, with the same quantities of the same reagents. The following are the results of the examination: Weight of the osazone. Character of the sugar. gram. 1 {Saccharose, ordinary 0.71 {Glucose and levulose (.526 g each) 0.73 2 {Maltose 0.55 {Glucose (1.052 g) 0.58 3 {Raffinose, crystallized 0.48 {Levulose, glucose and galactose (.333 g each) 0.53