masses. It is simple but effective, and so delicate in its indications that the utmost care was necessary to avoid interference for external causes, often difficult to control. Full details are given, as they are of great interest, especially to those who contemplate the use of a quartz torsion fibre. It is interesting to note that the author was never able, throughout a long series of experiments, to control absolutely the zero point of his balance. Although quartz is enormously superior to any other suspension thus får proposed, it is still defective in this respect. For some cause which Dr. Mackenzie is unable to give, the zero was constantly shifting. He does not clearly say whether this partakes of the nature of a 'drift' in one direction or not. In a long series of experiments, made by direction of the writer of this notice, for the purpose of trying to improve the existing form of the vertical force magnetometer, quartz fibres were used. Although apparently well protected from convection currents and changes in temperature, the mirror attached to them was never actually at rest. When this shifting and drifting is small, as it usually is, and observations are of the nature of those described by Dr. Mackenzie, that is, not in themselves extending over long periods, the error arising from it may be readily and correctly eliminated. The apparatus used for observing the attraction of isotropic masses was of the same character, and similar to that used by Professer Boys. The conclusion reached, the experimental results being in agreement within one or two-tenths of one per cent., is that neither in the case of crystalline nor isotropic masses was any deviation from the law of Newton detected. The author fails to note the very ingenious and interesting method of attacking the problem of the attraction of crystalline masses proposed by Poynting in his Adams Prize Essay on the Density of the Earth. Poynting proposes to test the question of there being different properties as to attraction along different axes of crystals by the directive action which must exist when one sphere of a crystal is in the field of another. He made some experiments along that line, and his work probably preceded by a year or two that of Dr. Mackenzie. At the present moment, with library out of reach, I am unable to say whether he has published any further results. The Influence of Temperature on the Transparency of Solutions, by E. S. Nichols and Mary C. Spencer, is another prominent article of the Review. Transparency to various wave-lengths was tested and a number of color solutions were examined. There are also papers on the Electric Conductivity of Certain Salt Solutions, by A. C. MacGregory, a continuation of the paper on Forces between Fine Solid Particles totally Immersed in Liquids and among the minor contributions is one interesting and useful on the Variation of Internal Resistance of a Voltaic Cell with Current, by Professor Carhart. T. C. M. NEW BOOKS. Die Chemie des Chlorophylls. L. MARCHLEW- Magnetismus und Hypnotismus. G. W. GESSMAN. Vienna, A. Hartleben. 2d edition. Pp. xiv + 205. Bulletin of the Geological Institution of the University of Upsala. Edited by HJ. SJÖGREN. Upsala, Almqvist & Wiksells. 1893-1894. Pp. 95, 293. SCIENCE. EDITORIAL COMMITTEE: S. NEWCOMB, Mathematics; R. S. WOODWARD, Mechanics; E. C. PICKERING, AS- J. S. BILLINGS, Hygiene; J. McKEEN CATTELL, Psychology; NATIONAL ACADEMY OF SCIENCES. REPORT OF THE WATSON TRUSTEES ON THE AWARD OF THE WATSON MEDAL TO SETH C. CHANDLER. On the recommendation of the Board of Trustees of the Watson Fund, the Academy last year unanimously awarded the Watson medal to Seth C. Chandler, of Cambridge, Mass., for his investigations relative to variable stars, his discovery of the period of variation of terrestrial latitudes, and his researches on the laws of that variation. It is the pleasant duty of the Trustees to set forth the grounds on which this award was recommended. It is a result of the well-known laws of dynamics relating to the rotation of a rigid body, as the earth is assumed to be, upon its axis, that the poles of the earth may be determined in two ways. Our globe, being a spheroid flattened at the poles and protuberant at the equator, has a certain axis passing between the points of greatest flattening. This axis has no direct connection with the rotation of the earth; it would exist if the latter, retaining its present form, did not rotate at all. It is called the axis of figure, being determined altogether by the shape of the earth. But the earth has also an axis around which it rotates. Now, assuming the earth to be a rigid solid, there is no necessity that the axis of rotation should correspond to that of the axis of figure just described. We could take a solid body, pass an axis through it in any direction, and make it rotate on that axis. It was shown by Euler, more than a century ago, that if a solid body rotated on an axis different from that of figure, the position of the axis of rotation in the body would be subject to a slow change, consisting in a constant revolution around the axis of figure. Were this body the earth, the latitude of a place, as determined by astronomical observation, would change in the same way. The time of one revolution of the pole would depend upon the figure of the earth. The flattening of the earth is such that, were it a perfectly rigid body, the time of revolution would be about 305 days; that is to say, the north pole would make its circuit in a period of 305 days. There being no necessity that the two poles should coincide, the question was naturally raised whether, perhaps, there might actually be such a difference of the two poles, and, in consequence, a change of latitude of every place on the earth's surface having a period of 305 days. The first to investigate this question with all the refinements of modern astronomy was C. A. F. Peters, who, half a century ago, was an assistant at the Pulkowa Observatory. In his classic paper on the parallax of the fixed stars, one section is devoted to the question of the variability of the latitude in a period of 304 days, which, according to the then accepted value of the flattening of the earth, would be the time of one revolution of the poles. He found a coefficient of 0".079, with a probable error of 0".017. This result was so extremely minute that it might have arisen from unavoidable sources of error; and the conclusion therefore reached was that if there was any such separation of the two poles, it was too small to be certainly detected by the most refined observations. In 1862 our late fellow member, Professor Hubbard, of the Naval Observatory, commenced a series of observations with the prime-vertical transit of that institution, which would be available for the same research. They were interrupted after a little more than a year, by his untimely death, but were continued four years longer by his successors. The result was the same as that reached by Peters; no change having a period of 305 days could be detected. In 1873 the question was investigated by Nyrén in connection with a longer series of observations on the latitude of the Pulkowa Observatory. His results were somewhat discordant, and the only conclusion that could be drawn from them was that the variation could not be certainly detected by these most refined observations. Ten years later Nyrén repeated the determination, in connection with his observations for the determination of the constant of abberation. These observations, made with the prime-vertical transit, were carried through with the minutest attention, and the utmost care to avoid every conceivable source of error. Curious discordances were nevertheless found in the results for the constant of abberation. In 1885 Küstner showed that they could be accounted for by supposing a change going on in the latitude. But nothing could be inferred respecting the law or the cause of the change. As a result of these investigations, the coincidence of the earth's axes of rotation and of figure has, until within a very few years, been assumed by astronomers as a practically established fact; and all their methods of observation have rested upon the idea of absolute coincidence. This confidence has not been disturbed until within a few years, when the question has been reopened. But it has now apparently been settled upon a new and firmly established basis. Dr. Chandler's work upon this subject began with observations made by him in 1884-85, using a novel form of astronomical instrument of his own invention. These observations, continued uninterruptedly for thirteen months, revealed a progressive change of a pronounced periodical character in the instrumental values of the latitude. In publishing these results in 1885 he announced his intention to continue the research throughout the remainder of that year. Yet circumstances prevented him from carrying out his intention at that time, and he did not resume his examination of the subject until six years later. Meanwhile Dr. Küstner, at the Observatory of Berlin, in 1888, published a memoir on the Constant of Aberration, as deduced by him from a series of observations also made in 1884-85, simultaneously with Chandler's series, which brought to light anomalies of an entirely analogous character. Küstner's series was not continuous enough to show the periodic nature of the phenomenon; but, by an exhaustive examination of the possible subjective sources of error, he clearly demonstrated that it was no longer permissible to retain the hypothesis of an invariable position of the pole, and he recommended that properly organized observations at various places be instituted to settle the question definitely. It was doubtless this work of Küstner's which compelled the attention of astronomers to the subject. As a result, by the coöperation of three German observatories, under the auspices of the International Geodetic Association, and the independent action of that at Pulkowa, the fact of the variability of terrestrial latitude was placed beyond question, and, by a corresponding series made at the Sandwich Islands, the further fact was established that the variable element is the position of the axis of rotation with respect to the earth's body, and not its position in space. It was just before this point that a renewal of Chandler's connection with the problem began. The results are published in a series of eighteen papers in the Astronomical Journal (1891-94), exclusive of a series of five papers upon a topic closely related thereto, and involving it; namely, the abberation-constant, which will be separately spoken of later. The keynote of these investigations, and the undoubted cause of the success which has attended them, lies in the fact that at the outset he first recognized the necessity of deliberately disregarding all teachings of the adopted theory, which had misled previous investigators, and of examining the facts by a purely inductive process, taking nothing for granted, and basing all conclusions strictly upon the observations themselves. It is impossible to give here more than a bare statement of the principal results thus established, which we arrange in their natural order, and not in the historical order of their derivation. 1. The phenomenon is not a local or a regional, but a terrestrial one; also it is a displacement of the earth's axial rotation. with reference to the principal axis of inertia, and not of the direction of the former in space. 2. The axis of rotation, although fixed as regards its direction in space, performs a relative revolution about that of inertia in a period of 428 days. This motion is circular, with an average radius of about fourteen feet, and its direction is from west to east. 3. Simultaneously with the above motion, the actual position of the principal axis of inertia on the earth's surface is in motion about a mean position, in a period of a year. Its direction is also from west to east, but is in an ellipse, three or four times as long as broad, the major and minor axes being about twenty-five feet and eight feet respectively. The major axis is inclined at present, by about 45° to the Greenwich meridian. The motion is central, obeying the law of proportionality of times to areas described by the radius vector about the center of the ellipse. 4. Both the radius and period in the circular 428 days' revolution are systematically variable; the former between about eight feet and eighteen feet, the latter between about 423 and 434 days; in a long period of apparently about sixty-six years. In this inequality of motion the average angular velocity is attained when the size of the circle is least or greatest when the circle has its mean dimensions. 5. Similarly there are simultaneous changes in the apparent dimensions and velocity in the annual period, which are complementary in their character to those in the 428 days' revolution; but whether they are the result of real changes in the form and dimensions of the ellipse, or the effect of an apsidal motion of long period, cannot at present be determined from the observations available. All that can be said is that observations during five years show that the line of apsides is either fixed, or, if variable, revolving only at a very slow rate. 6. Besides these two motions of relatively short period, there is distinct evidence of a third motion of rotation in a much larger term, probably not far from twelve years, with a radius of ten or fifteen feet, which reconciles similar indications of slow changes which had been pointed out by other investigators. (A. J., XII., 178; XIII., 35, 36.) The results thus established are the outcome of the examination of an immense number of observations, covering the whole interval since the era of refined practical astronomy began, and in fact practically exhaust the materials which may be drawn for this purpose from existing astronomical annals. The endeavor to make the discussion exhaustive in this respect made it neces sary to completely reduce, from the original instrumental readings, extensive older series of observations. It has, incidentally, for example, rescued from almost complete oblivion the series of Pond, 1825-36, and shown that work to be of a character which will compare favorably with the most refined observations made with the meridian instruments of the present day. Intimately connected with the work on the variation of latitude are five additional papers, containing a redetermination of the value of the aberration-constant from eight different series of observations at the Pulkowa Observatory, with the prime vertical transit and the vertical circle. The correct value of this fundamental element is one of the most important questions occupying the astronomy of the day. VARIABLE STARS. THE subject of variable stars was erected into a distinct branch of astronomical science by Argelander, beginning in 1843, and occupied a large share of his activity and interest during a score of years. His classical labors were succeeded or overlapped by those of Schönfeld, who assumed the principal charge of the subject for another score of years, when his devotion to the great work of the Southern Durchmusterung, and later his failing health, left opportunity for other hands to take up and continue the work where they had left it. Since the issue of Schönfeld's Second Catalogue the number of known variables has more than doubled, while the fund of observations pertaining to them has vastly increased. Chandler's work in this direction, besides the incidental work of observation and discovery which he has contributed to it, has involved the collection of all the data in astronomical history, their discussion, and the formulation of the elements of their light-variations into numerical laws. The catalogues of 1888 and 1893, while filling a |