Glycerol 1,2-and 1,3-dinitrateJ prepared from the corresponding dibromohydrins of glycerol, are characterised by reaction with aromatic acid chlorides. The 1,Z-dinitrate is also obtained by the action of iodinesilver nitrate on allyl alcohol; derivatives of the dinitrate are prepared similarly from allyl esters and by nitration of l-glycerides. Partial nitration of glycerol gives glycerol 1,3-dinitrate, also obtained by mild nitration of glycidol (2,3-epoxypropan-l-o1) nitrate. Denitration of glycerol trinitrate by sodium nitrite or potassium hydroxide in aqueous ethanol and by sulphuric acid confirms that the secondary nitric ester group is displaced more readily.

FOR the past sixteen years I have adopted the aluminium method for the estimation of nitrates as the most convenient for the purposes of water analysis, and the following remarks are to be regarded simply as a record of my experience in the working of the process during that period, without laying any claim whatever to originality. I have had so many proofs of the reliability of the procew, its convenience and economy of time, that I have thought the matter worth the attention of the Society of Public Analysts. My procedure is as follows :-I take 70 C.C. of a sample of water, 30 C.C. of a fenper cent. solution of caustic NaHO-of which more presently-and a slip of thin sheet aluminium, weighing about 0.75 grammes ; these are placed in a narrow-mouth vial (402.) the mouth of which is closed with an indiarubber stopper pierced with one hole, in which is inserted a tube 4 inches long, 6 inch diameter, with a drawn out narrower portion about 3 inches long (this latter is inserted in the hole of the stopper), and the wider part is nearly filled with glass-beads. The whole arrangement is then placed under a bell-jar, which may be capacious enough to take several such if desired. As soon as the action is over, and the metal dissolved, the contents are transferred to a 200 C.C. flask and made up to mark with the washings of the beads bottle, &c., and the whole mixed by shaking, Where time is an object, the mixture may be filtered at once (rejecting the first '' runnings,'' and a moderate quantity of the clear liquid diluted to standard bulk and Nesslerized) otherwise the whole is put into an 802. stoppered bottle and allowed to settle for 3 or 4 hours, and a portion of the clear liquid Nessleriwd after proper dilution.