We photographed no northern stars there, except the Pleiades and the Præsepe. Of the Pleiades I brought home sixteen plates, with two impressions of the whole group upon each, made in five different years from 1872 to 1882 inclusive. Although the centre of the cluster never attains a greater attitude at Cordoba than 34° 50', some of the plates contain seventy stars. All but one of Bessel's stars are there, which belong within the limits of the field, the missing one being of the magnitude 94; and there are yet other stars of the magnitudes 10, 10 and 11. Of the Præsepe, there are five plates, and with a correspondingly increased number of stars.

About seventy southern clusters have been repeatedly photographed at Cordoba, comprising all those of the southern hemisphere which seemed important; also somewhat more than a hundred double stars, being a sufficient number to serve as a good test of the method. The total number of photographs now at hand is somewhat less than 1300, only few having been preserved in which the images were not circular.

Especial attention, however, was given for many years, to taking frequent impressions, at the proper seasons, of four stars selected, on account of their large proper-motion, as likely to manifest appreciable annual parallaxes. The refined and elaborate observations of Drs. Gill and Elkin at Cape Town have been made, computed and published, while the Cordoba photographs have lain untouched in their boxes. There is but one of my four stars, 3 Hydri, which is not included in their list. Still it will be a matter of much interest to apply the photographic investigation to the same problem, even if for no other purpose than a comparison of the results of the two methods.

I am convinced that the Cordoba plates contain a large number of stars as faint as the eleventh magnitude of Argelander's scale; and believe that these are much the earliest photographs of stars fainter than Mr. Rutherfurd's of 1865 and 1866. There are several plates, covering about a degree square, which cannot contain less than 550 stars, and I believe that some of them contain a greater number. Such are those of the cluster Lac. 4375, and that near 2 Carinae. The region in the vicinity of η Carinae and that magnificent tract in Sagittarius, which is too densely sown with stars to be considered merely a portion of the Milky Way, and yet too large and undefined to be regarded simply as a cluster, were both of them taken several times, during the years 1875 to 1882, in a series of overlapping photographs, each containing about a square degree, and recorded upon a glass surface of 9 X 12 cm. In their present form they are of course of small value for scientific use, inasmuch as the stars are too crowded for their configurations to be easily perceived; and--although these two series form in fact maps of considerable regions in the sky,still the record is of a very perishable nature, and of small avail for use by astronomers until it shall have been translated into an enduring and numerical form by micrometric measurement.

In this connection I may say that one of the greatest of my present anxieties, regarding the Cordoba photographs, arises from a discovery of the

ease with which the collodion or gelatine film may become detached from the glass. The Argentine government has assigned a moderate sum for the prosecution of the measurements, and with this some progress has already been made. It is but right to add that the full amount was given for which I asked; still it is now quite inadequate, in consequence of the unfortunate depreciation of the national currency; and, in the present financial crisis there, I cannot reasonably expect more. Yet this matter of prompt measurement appears to me at present much more important than it did while I was unaware of the facility with which the film can blister and peel.

In 1883, after Mr. Common's brilliant success in photographing nebulas with his great three-foot reflector, he proposed to me a joint arrangement for photographing the whole heavens. My work at Cordoba was so near its close that it was out of the question to undertake anything new; but the immense labor requisite for the measurement of the plates would under any circumstances have tended to determine. It is an undertaking demanding the joint energy, application and material resources of a large number of persons, if the results are to be made available for astronomical use; indeed, I see no other astronomical value in the unmeasured photographs than the possibility of confirming at some future epoch, the existence of relative motion previously detected or made probable by some other investigations.

Since then this process of photographic charting is said to have been systematically undertaken by the Brothers Henry at Paris. I have seen none of the plates; but their sharpness is highly spoken of, and the work appears to be prosecuted with much skill and very sensitive plates. There can of course be no question as to the value of any permanent record whatsoever, corresponding to a known date; yet I cannot feel that any essential advance is likely to be made in this way until the photographic record shall have been brought within the range of numerical expression.

The measurements of the Cordoba photographs, thus far completed, are those of the double stars, of the four stars with large proper-motion, of the Pleiades, of the Præsepe and of the clusters Lac. 4375 and Crucis. The corresponding computations have been made, as yet, only for a portion of the Pleiades impressions, but I am hopeful of completing these at a comparatively early date. We shall then be able not only to compare the results with Bessel's of forty-five years ago, but to test the deduced values of the proper-motions by means of the photographic determinations of 1865 and 1866. Meanwhile the valuable memoir of Wolf has been published, giving closely approximate positions for 571 stars of the group, and Dr. Elkin has recently been executing at New Haven a heliometric triangulation of the principal stars. Our photographic results will have to be confronted with his delicate heliometric ones; and should they bear this test with tolerable success, it will be all that can reasonably be desired.

SOME ACCOUNT OF A NEW CATALOGUE OF THE MAGNITUDE OF SOUTHERN STARS. By EDWIN F. SAWYER, Cambridge, Mass.

THE present short paper is merely intended to announce the completion, in the immediate future, of a work undertaken in 1882, namely, the determination of the relative magnitude of the stars included between the equator and 30° S. Dec., and not fainter than the 7th magnitude. In fact the work is a revision of so much of Dr. Gould's valuable catalogue the Uranometria Argentina, as is included in the above limits, and his catalogue has been used as the basis of the undertaking. The observations have been made by Argelander's well known method of step estimations; the stars being gathered into convenient sequences, and these sequences formed from the stars as closely adjacent as possible; thus reducing to a minimum errors arising from atmospheric causes and the unfavorable situation of the stars, except in a few sequences formed from bright stars necessarily more widely scattered. In making the observations, an operaglass magnifying two and one-half diameters has been employed and the glass placed slightly out of focus, expanding the stars into disks of light; this method after repeated trials appearing to give the most reliable results, especially where the stars are colored. The observations have generally been made during evenings free from the moonlight and clouds and haze. As at first planned, each star was to have been observed but once (owing to the press of other astronomical work on our leisure time): but after the observations had been practically completed in 1885 (comprising some 3500 stars) and while their reduction was under way, it was decided to reobserve each star once in ordinary cases, repeatedly where large discordances appeared. This duplication of the observations, while it increased the amount of time and labor which I had originally intended to devote to the undertaking, would I felt, be justified by the enhanced value of the results; observations were therefore resumed in 1886, and it is hoped that another year will find the work completed and ready for publication.

The number of stars comprised will approximate 3500, and the average number of observations for each star will be about three and one-half. During the progress of the work one variable star has been discovered, and large discordances in the observations of several stars may possibly lead to the detection of others.

As a test of the character of the work, the following results have been deduced from a partial discussion of the observations of about 900 stars observed from two to five times each.

To determine the accidental errors of observation, I have found from 593 stars, observed twice, that the average difference between two independent determinations of a magnitude of a star is .112 of a magnitude, which corresponds to a probable error of a single observation of ± 065 of a magnitude. There appears to be no sensible difference in the values of this element dependent on the zenith distance.

The process of reduction adopted presumably excludes a liability to any systematic deviation from the system of magnitude of the Uranometria

Argentina, which, as is known, is based on that of the Uranometria Nova. But a direct test has been made to verify this point, by employing stars observed three or more times. Taking the difference between my mean magnitude and that of Dr. Gould's, and classifying according to magnitude we have the following table :

The values in the third column are merely nominal and afford satisfactory evidence of the coincidence between my magnitude scale and that of Dr. Gould's. From the fourth column, it appears as would naturally be expected that the estimates for the fainter magnitudes are ratably the more accurate. The probable error of a difference G.-S. is on the average .091 of a magnitude. Assuming for want of a criterion that both catalogues have equal weight, this corresponds to a probable error in each of .064 of a magnitude.

I have not limited the observations to the stars only contained in Dr. Gould's Catalogue, but have inserted all objects which have appeared to be in the neighborhood of his fainter class.

The number of such objects is surprisingly small, probably not over seventy-five; and this, in connection with the observed small probable error appears to be a most satisfactory evidence of the high degree of precision of the magnitudes of the Uranometria Argentina.

A

COMPARATIVE ESTIMATE OF METHODS AND RESULTS IN STELLAR PHOTOM

ETRY. BY S. C. CHANDLER, jr., Cambridge, Mass.

[ABSTRACT.]

THE main object of this paper is to give the results of an examination of various catalogues of stellar magnitude which have appeared during

the last twelve years, with special reference to the question of the comparative value of instrumental and non-instrumental methods in stellar photometry.

Among the topics discussed are the light ratio for the unit of magnitude; the wide discordance in the measurement of light differences by various photometers; their inability to furnish a trustworthy iso-photometrical scale; the existence and probable causes of the various systematic mutual deviations which they exhibit; and the conclusions which may be drawn as to the present condition of stellar photometry.

The ground is taken that instrumental photometry is so far practically a failure, in that it furnishes contradictory solutions of the problems for which it was especially needed, and has besides failed to attain the degree of precision which pertains to the non-instrumental method.

COMPARISON OF THE BOSS AND AUWERS DECLINATION-STANDARDS.
HENRY FARQUHAR, U. S. C. and G. Survey, Washington, D. C.
[ABSTRACT.]

By

THE standard of Dr. Auwers' "Fundamental Catalogue" (A. G. Pub. 14, 17) is, for declinations, that of the recent Pulkowa observations; to which all other authorities used in the catalogue are adjusted by systematic corrections, from a comparison in order of declination and another in order of right-ascension. Prof. Boss's N. B. Survey Catalogue having been similarly compared with the Pulkowa declinations, the mean differences (shown in order of declination in an accompanying table; those in order of right ascension having no significance) are remarkably uniform in sign, all those north of 50° being positive, increasing to 0.17 about 80°, and all south of 50° negative, larger for lower stars, and surpassing 0.3 south of -10°. Applying to these differences the correction — 0.06 +0.30 sin(being the zenith-distance at Pulkowa, 60°-) they are much reduced, and the uniformity in sign disappears; there is a suggestion, however, of an undulatory inequality, of a 30° period. The latter correction, if it really exist, probably denotes a periodic graduationerror in a meridian circle; the former pretty clearly denotes an imperfectly corrected tube-flexure. The error may be in the Pulkowa declinations, or in some of the sources drawn upon by Boss. The residuals could be further diminished by a small term in cos, indicating flexure from a zenith south of Pulkowa; or by a small term in tan ?, showing an inaccurate allowance for refraction in one or other catalogue; but the gain is too uncertain to justify a positive conclusion.