sulphuric acid, or failed to give a brownish precipitate with stannous chloride.  As the sample contained a considerable quantity of potassium carbonate, in which the resin is soluble, it was thought that by neutralizing this it might render the resin more easy of extraction.  This was found to be so, but it was accompanied by such a mass of extractive as made it in the long run more troublesome, and hence it was abandoned.  Thinking the spirit employed might be too weak, an experiment with commercial absolute alcohol was carried out as follows:  One hundred grains of a red sample, No. 4, were thoroughly dried, powdered finely, and boiled in 2 ounces of the alcohol, filtered, and the residue treated with half an ounce more.  This required to be repeated with fresh half ounces of the alcohol until in all 71/2 were used; the time occupied from first to last being almost three hours.  This was considered unsatisfactory, besides being very expensive, and so it, also, was set aside, and a series of experiments with methylated spirit alone was set in hand.  The results showed that the easiest and most satisfactory way was to take 100 grains (this amount being preferred, as it reduces error to the minimum), dry thoroughly, powder finely, and macerate with frequent agitation for twenty-four hours in a few ounces of spirit, then to boil in this spirit for a short time, filter, and repeat the boiling with a fresh ounce or so; this, as a rule, sufficing to completely exhaust it of its resin.  Wynter Blyth says that the red resin, or bixin, is soluble in 25 parts of hot alcohol.  It appears from these experiments that much more is required to dissolve it out of commercial annatto.

The full process followed consisted in determining the moisture by drying 100 grains at 212 deg.  F. till constant, and taking this dried portion for estimation of the resin in the way just stated.  The alcoholic extract was evaporated to dryness over a water-bath, the residue dissolved in solution of sodium carbonate, and the resin precipitated by dilute sulphuric acid (these reagents being chosen as the best after numerous trials with others), added in the slightest possible excess.  The resin was collected on a tared double filter paper, washed with distilled water until the washings were entirely colorless, dried and weighed.

The ash was found in the usual way, and the extractive by the difference.  In the ash the amount soluble was determined, and qualitatively examined, as was the insoluble portion in most of them.

The results are as follows: