per wire, No. 14, wound upon cylinders of wood (a).* Wooden rods (c), clamped so as to move with gentle friction, bear a wire (w), from which an astatic system of needles is suspended by means of a silk fiber. The upper needle is midway between the centers of the two coils. The lower needle plays over a scale (s) photographed on glass, beneath which is the horizontal condensing lens of the vertical lantern. The needles are ordinary sewing needles, and are each 15 inches in length. Each coil is composed of 34 7 meters of wire, the resistance of which is 0.444 ohms. Each coil should have the same number of windings, and the same resistance. This is easily effected by care in winding. By sliding the rods (b) in or out, the distance between the coils may be varied from 2 cm. to 10 cm., the image of the lower needle being in all cases perfectly distinct. In this way the instrument is adjusted to currents of any strength. Scales cut in the rods (b) serve to regulate the distances. D ૧. 2. E 어 C B 어 h A On the outside of the box are six plates of brass, whose form and arrangement are shown in fig. 2. The extremities of the coils are connected with the four plates A, B, C, D. This connection may be made by means of binding-screws on the inside of the box, in which case the coils may be replaced with ease by others of greater or less resistance. The plates are put in metallic contact by means of *For Duboscq's lantern, the coils must be placed lower than here represented. The lower needle may be replaced by a bristle from a painter's brush, or some other light pointer, the upper one being damped by magnets as recommended by Mayer. brass plugs, inserted at a, b, c, d, e, g, h, k. Putting plugs at h and e, and connecting the poles of a galvanic cup at the bindingscrews A and C, and the current runs successively through the two coils R, each causing deflection in the same direction. Let R represent the resistance of one coil of the galvanometer, then the resistance of the galvanometer will be 2R. This arrangement is used in working with ordinary galvanic currents. If instead of the former connections, plugs be put at a, d, g, and h, the wires from the source of electricity being connected at E and F, then the galvanometer resistance becomes R. This arrangement is to be used with circuits of small resistance, such as thermo-currents. For this kind of work the instrument is thoroughly adapted. This instrument can also be used as a differential galvanometer. To do this, put the positive pole of the battery at E. Plug a and c. Divide the negative wire into two equal branches which are to be connected at B and D. The circuit being thus closed, the needle evidently remains at zero. Introducing any wire the resistance of which is to be determined, into one branch, bring the needle to zero again by introducing known resistances into the other, and the unknown resistance is readily determined. In measuring fractions of an ohm, a rheochord is, all things considered, the best. The contacts are good, and an audience obtains a better idea of what is meant by electrical resistance than when a resistance box alone is used. Using platinum wire weighing 737 grams per meter, the resistance of which is one ohm to 192.9 cm. of wire (which is 96:45 cm. on the instrument scale), and thousandths of an ohm can be measured direct. If ground connections are made the negative pole of the battery is sent to ground direct, and the branches of the current from B and D are sent to ground through the unknown resis tance and the resistance box respectively. Shunts may be introduced into either of the half circuits. This may be done by introducing coils of resistance R or R, between the binding screws A, B or C, D. These wires may also be wound upon metallic plugs, which have been split lengthwise, the parts being insulated and each being connected with one extremity of the wire. Permanent shunts may be introduced by connecting one extremity with plates A or D, the other extremity being attached to an insulated plate, to be put in contact with B or C by means of a solid metallic plug. These shunts are used in Latimer Clark's differential galvanometer, and their use in measuring resistance is too well known to need further explanation. The advantages possessed by this galvanometer are: 1. It is easily adjusted to any vertical lantern, from which it can be removed in a moment if desired. 2. The distance between the deflecting coils being readily varied, it can be adjusted to currents of various intensity. 3 The resistance of the galvanometer is quickly varied from one-half, to twice the resistance of one of the galvanometer coils. 4. The coils may be replaced by others when desired. 5. It can instantly be converted into a differential galvanometer and used in measuring resistance. 6. It can be constructed in any work-shop at a very small expense. St. Louis, Oct. 25, 1875. ART. XI.-On a new occurrence of Tartronic Acid, with some remarks on the Molecular Structure of Glyceric Acid; by SAMUEL P. SADtler. (Read before the American Philosophical Society, September 17, 1875.) IN the Propyl series, nine normally formed acids are possible, besides several isomeric unsymmetrically formed ones. They are: and the following are the acids considered as having the molecular structure just given : I, propionic acid; II, lactic acid (of fermentation); III, pyruvic or pyro-racemic acid; IV, ethylene lactic acid; V, glyceric acid; VI, carbacetoxylic acid; VII, malonic acid; VIII, tartronic acid; IX, mesoxalic acid. In one or two of these cases, however, there is still a difference of opinion as to whether the acid named is the one possessing the normal molecular structure given above, or is only an isomer of it, having its carbon atoms differently united. Notably with glyceric acid is this yet an open question. Some results lately obtained in the course of a study of this acid appear to me to be of value for the solution of this question. The other view of the molecular structure of glyceric acid makes it unsymmetrical, two of the carbon atoms being doubly united. The formula given is CH,.OH d.OH 01 As will be seen, this formula does not contain the carboxyl group, hitherto supposed to be the inevitable characteristic of an organic acid. The author of this theory is Prof. Wislicenus, of Würzburg, and the following are the reasons given in support of it. If lactic acid be acted upon with hydrogen iodide, a iodo-propionic acid is formed, according to the following reaction: This when heated to 150° with strong HI is changed into propionic acid. If, on the other hand, glyceric acid be acted upon with hydrogen iodide, ẞ iodo-propionic acid is formed. If this had the formula CH,I CH, co.OH, on treatment with moist silver oxide, it would pass into ethy lene lactic acid. It does not, however, do this, but a new acid isomeric with ethylene lactic acid is formed-hydracrylic CH2.OH 01 That the molecular structure of this acid is essentially different from that of ethylene lactic acid is proved by the oxydation products of the two. Ethylene lactic acid yields malonic acid, while hydracrylic does not yield a trace of this, breaking up into glycolic and oxalic acids and carbonic dioxide. Moreover, hydracrylic acid on heating yields acrylic acid, a derivative of allyl alcohol, instead of the lactid yielded by the lactic acids. Prof. Wislicenus, however, frankly gives one experiment made by himself, the result of which tends the other way. He reduced the iodo-propionic acid by sodium amalgam and obtained what appeared to be the normal propionic acid, showing the regular molecular structure. In favor, moreover, of the normal structure for the molecule of glyceric acid is the formation of pyruvic or pyroracemic acid. CH3 ćo from glyceric acid upon heating this to 140°, explained by the following reaction: The structure of this pyruvic acid is known from the fact that acted upon by nascent hydrogen it gives normal lactic acid. A strong additional argument would be had, if we could show a connection between glyceric acid, Hitherto tartronic acid had not been formed from glyceric acid, but only in an indirect way, by the spontaneous decomposition of nitro-tartaric acid, according to the following reaction: However this mode of formation was interesting as tending to show its symmetry of structure. For that matter a dibasic, triatomic acid could hardly exist, except by the assumption of two carboxyl groups. I have been fortunate enough to find tartronic acid associated with glyceric acid in the oxydation products of glycerine. The preparation of the two acids was as follows: One part by weight of glycerine is mixed with one part of water, and to the mixture is added, by means of a long funnel tube reaching to the bottom of the cylinder, about one and a quarter parts of |