For taking the density of a liquid, we plunge the end, B, into it, and then suck, and afterward close the rubber tube with the clamp. It is essential that this latter shall hold well, so that the levels may remain constant. We now do the reading. Suppose, for example, we read 250.3 mm. on the pipette, and 147.7 mm. and 152 mm. on the branches of the gauge. Having these data, we loosen the clamp, and allow the liquid to flow. On account of capillarity, there remains a drop in B; and of this we read the height, say 6 mm. A height 250.5 mm - 6 = 244.5 mm. of liquid raised is, then, balanced by a column of water of 147.5 + 152 = 299 mm. Now the heights of these two liquids is in the inverse ratio of their densities: _d_ 299.5 --- = -----, whence _d_ = 1.22. 1 244.5 We obtain _d_ by a simple division. When the instrument has been carefully graduated, and has been constructed by an expert, the accuracy of the first two decimals may be relied upon. With a little practice in estimating the last drop, we may, in trying to estimate the density of water, even reach a closer approximation. In order to measure the height of the drop accurately, one should read the maximum height to which the liquid rises between the fall of two drops at the moment when the last ones are falling, since at that moment, and only at that, can it be ascertained that the lower level of the bubble is plane. The error in such reading does not reach half a millimeter, and, as a suitable height of the apparatus permits of having columns that vary between 13 and 30 centimeters, an error of this kind is but 1-300. This is the limit of precision of the method. The clamp might be advantageously replaced by a glass cock, or, better still, A might terminate in a rubber bulb; and a lateral tubulure might be fixed to the pipette, and be closed with a rubber stopper. This little apparatus is more easily maneuvered than any of those that have hitherto been devised upon the same principle. It is capable also of replacing areometers in ordinary determinations, since it permits of correcting the error in capillarity that is neglected in instruments; and, moreover, one can, when he desires to, easily verify for himself the accuracy of the graduation.--_La Nature._ * * * * * USEFUL BAGS, AND HOW TO MAKE THEM. By JOHN T. HUMPHREY.