imperfect evidence now existing in regard to the one under consideration; and it must be remembered that no matter how inadequate this evidence may at first sight appear to be, it is yet as full, minute and complete as that relating to any single nebula except that of Orion; and if the drawings here considered are not sufficient to prove a change or the absence of a change, we are reluctantly driven to the conclusion that the work done by astronomers in this direction has been largely wasted. I hope that the evidence here adduced may be deemed of sufficient importance to warrant the great expense of time and labor necessary to a detailed monographic study of this nebula, which may serve for future reference. There is proba bly no nebula visible in the northern hemisphere more worthy of such examination. U. S. Naval Observatory, April, 1876. ART. XLIII.—Brief Contributions from the Physical Laboratory of Harvard College. No. XVIII.-On the effect of thin plates of Iron used as armatures for Electro Magnets, and a new form of Induction Coil; by JOHN TROWBRIDGE, assistant Professor of Physics. IN a paper presented to the American Academy of Arts and Sciences, April 13, 1875, I showed that the application of armatures to two straight electro-magnets, which formed the primary circuit of a Ruhmkorff coil more than doubled the strength of the induction current produced by breaking the primary cir cuit. When, however, the circuit of the secondary coil was not closed, and a spark was allowed to jump across the interval be tween its poles, the striking distance of the spark and its power to charge a condenser did not seem to be notably increased by the application of armatures to the electro-magnets of the pri mary circuit. My experiments at that time were made with solid iron cores; and I now resume these experiments with bundles of fine iron wires in place of the solid iron cores. The mechanical difficulty of making the ends of the bundle of fine iron wires constituting the cores plain surfaces was overcome by dipping them in melted solder and then filing the ends smooth. In this way I had no trouble in applying the armatures so that they should lie upon a plain surface. The resistance of each of the two induction coils covering the two straight electro-magnets was 6000 ohms; and that of each of the straight electro-magnets, 34 of an ohm. The diameter of the bundles of fine iron wires constituting the cores was 5 cm. and the length of the electro-magnets 28 cm. Condensers of 2nt various sizes were placed in the primary circuit. The results given in this paper were obtained by the use of a condenser of about one farad. The method of experimenting was to charge a condenser of one-third of a farad; and then to discharge this condenser through a galvanometer. If we express the quantity of electricity received by the condenser by Q, the electromotive force by E, and the capacity of the condenser by C, we have Q= sin, where n is the reduction factor of the galvanometer, t the time of vibration of the magnet, and the angle through which it swings under the effect of the change. Knowing the reduction factor of my galvanometer, I had thus the means of reducing my results to absolute measure. But I speedily found that the relative results obtained by the proportions π Q: Q' sin; sin0'-E: E' would present the points of this investigation in as clear a manner as if the results had been reduced to absolute magnetic measure. My first experiments were made with solid arma In table I the numbers are the deflections of the reflecting galvanometer expressed in millimeters, and the distance of the scale from the magnet was one meter. In this case the gain by the use of the armatures was trifling, being only about fourteen per cent. These results were obtained by charging the condenser of one-third of a farad, by sparks one millimeter in length. On a closed secondary circuit, however, a gain of one hundred per cent was clearly seen in the strength of the induced currents produced by breaking the primary circuit. The question, how to make this increase in the strength of the induced current by the employment of armatures apparent on a broken secondary circuit, became an interesting one. It seemed at first as if the application of armatures, by maintaining the temporary magnetization of the iron cores would be detrimental rather than otherwise. I next tried the effect of bundles of thin iron plates, which were placed, as armatures, upon both poles of the electro-magnet, thus making a magnet of a horse-shoe form. On charging the condenser, I found a very great increase in quantity, which was manifested by the swing of the galvanometer needle; the indicator being entirely off the scale. Table II shows the results obtained by the use of iron plates one and one-quarter of an inch in thickness, twenty in number, constituting each armature. Here a gain of four hundred per cent was manifested by the use of thin plates. The next step was to ascertain how many plates were necessary to obtain the maximum effect. The difficulty of obtaining plates of the same homogeneity, made it impossible to obtain smooth curves. To this difficulty was added that of breaking the primary circuit in a regular manner. If the results of Table III are plotted, it will be seen that the increase within small limits, is very nearly proportional to the number of thin plates, which were of an inch in thickness. On increasing the number of plates a point was reached where there was no additional effect. The best result was obtained when the mass of the armatures was approximately equal to that of the cores of the electro-magnets. Plates of of an inch were also used, but no advantage resulted in their employment over those of of an inch. It would seem that the thin plates followed the same law as that of the bundle of fine iron wires which constitute the cores of induction coils of the present day, and that only a moderate degree of discontinuity in the mass of iron submitted to magnetic influence is necessary to prevent the formation of currents of induction, which prolong the magnetism of the cores, and prevent the quick demagnetization necessary to produce intense currents of induction. The effect of insulating the thin plates with the dielectrics was also tried with no gain in effect. There appeared to be a slight gain by placing the plates edgewise on the poles of the magnets instead of allowing them to repose on their flat faces. This was doubtless due to better contact of the metallic surfaces. Since the above results proved conclusively a very great gain in quality and electrometric force by the application of thin plates as armatures, I next measured the striking distance of the spark. Table IV gives the results which are the mean of many trials. TABLE IV. A curious fact came to light in this connection; the lengthening of the spark was not shown when the spark leaped directly between the poles of the induction coil; the increase in quantity and electromotive force, was only made manifest to the eye by the employment of condensers in the secondary circuit. The results in Table IV were obtained by the employment of a leyden jar of large capacity. The increase in the quantity and electromotive force was not only shown by the increased length of the spark, but also by its increase in volume, and its louder snap. The spark consisted of a thick central bolt surrounded by curious thin, detached sparks. An attempt was made to measure the increase of light in the Geissler tubes by Vierodt's photometric apparatus, but it was found too inexact for this purpose; if, indeed there was any increase of light, which remains to be proved. I know of no results which bear upon the relation of the increase of light to the increase of electromotive force of the induction spark. Without condensers in the secondary circuit, however, the increased electromotive force of the spark was shown by its greater constancy in leaping over a given resistance of air. Unless an instrument is desired for popular scientific lectures, length is not so much to be desired as quantity of electricity of a spark, and in this form of induction coil the gain is principally in quantity, although it is true that with the aid of leyden jars, the striking distance is increased one hundred per cent. The principal points of this paper can be thus summed up: 1. The application of thin plates of soft iron upon the poles of two straight electro-magnets, with bundles of fine iron wires for cores, increases the strength of the spark produced at the poles of the secondary coils surrounding the electro-magnets, four hundred per cent. 2. The length of the spark is increased one hundred per cent. This gain in length is only manifested by the employment of leyden jars of large capacity, which are connected with the secondary circuit. 3. Instead of distributing the fine wire of a Ruhmkorff coil upon a straight electro-magnet, as is done at present, this wire should be distributed equally upon two straight electro-magnets, whose poles should be provided with armatures of bundles of thin plates of soft iron. ART. XLVI.-Communications from the Laboratory of Williams College. No. VL-Concerning Phosphorus Oxychloride; by IRA REMSEN. THE fact was recently established* that carbon monoxide, though it must be considered as an unsaturated compound, does not readily combine with the oxygen from ozone to form the saturated dioxide. Indeed it was impossible to discover any conditions under which such a combination takes place. Although it is known that ozone does readily oxydize many substances, it seemed to me desirable to further test its action upon bodies which are generally recognized as unsaturated. For this purpose I have first employed phosphorus trichloride in the hope of obtaining the oxychloride, POCI,. The method of formation of the oxychloride thus indicated would be interesting from more than one stand-point, as will be pointed out below. It has already been shown by Brodiet that, when oxygen is passed into phosphorus trichloride at the boiling temperature of the latter, a partial transformation into the oxychloride takes place; and Michaelis subsequently showed that this transformation or oxydation is exceedingly incomplete, even though the process be continued for two or three days. An analogous experiment has also been performed by Henry, who proved that, when sulphur and phosphorus trichloride are heated together in a sealed tube at 130°, the sulphochloride PSCI, is formed. It is plain that, in both of these experiments, one of the forces which opposes the combination is that which binds together the atoms of oxygen in the molecule of oxygen, and the atoms of sulphur in the molecule of sulphur; and hence, if we could employ free atoms of oxygen or sulphur instead of their molecules, we would expect the action to take place much more readily. In the case of sulphur, it is not possible, as far as we know at present, to obtain free atoms or unstable molecules which by their breaking up yield free atoms. ozone, however, we have such an unstable molecule of oxygen. As we have seen in the experiment with carbon monoxide, above referred to, ozone does not always appear to furnish free atoms of oxygen when we might expect it to, and hence the formation of phosphorus oxychloride by the action of ozone could not be predicted with any certainty. Experiment proved, however, that the formation actually does take place with *This Journal, vol. xi, p. 136. Odling's Handbook, i, 297. Gmelin-Kraut's Handbuch der Chemie, I, i, 391. In |