We next proceeded to put on one side ten of our forty flasks, to serve for subsequent corroboration; in ten others, by means of the tube which is represented on the right hand side of the flask (Fig. 8), we put a few drops of the water in which the bunch of grapes had been washed; in a third series of ten flasks we put a few drops of the same liquid, after having previously boiled it. Lastly, we introduced into the ten remaining flasks a drop of grape-juice, taken from the inside of an uninjured grape. To do this we had to bend the right-hand tube of each of our last ten flasks, drawing it out to a fine point and closing it in the flame, as represented in Fig. 9 A. This fine closed point was filed round near its extremity and then thrust, as represented in Fig. 9 B, into a grape placed on a hard substance; when the point _b_ was felt to touch the substance supporting the grape it was broken off by a slight pressure sideways at the point _a_, where the file marks had been made. We had taken care to secure a slight vacuum in the flask beforehand; this now caused a drop of the juice to be drawn into the flask. We then drew out the fine point, and closed it immediately in the flame of a spirit lamp. The vacuum was produced by heating the flasks in our hands, or over the flame of a lamp, thus causing a little air to be forced out through the end of the bent tube, which we then closed up with the lamp. When the flask was cool, the slight difference of pressure sufficed to force into it some of the juice contained in the grape, as we have just described. The drop of juice that is sucked into the flask generally remains in the curved part of the fine tube; to mix it with the must we must incline the flask so as to bring the must in contact with the drop; after that we may replace the flask in its natural position.