must be remembered that Mr. HAYFORD is in the position of having to make a special plea.

With his good instrument and this medley of observers, Mr. HAYFORD shows, among other conclusions :—

1. That the relative personal equation, that bugbear of astronomers in determining star positions and terrestrial longitudes practically vanishes,-that is, it is less than 0.050 in any case.

2. That he is justified in predicting that three nights of observing by the new method and without exchange of observers will serve to determine a longitude as accurately as ten nights of observing by the old method including an exchange of observers-a great economy of time and transportation.

3. That, for a practiced observer, the new method is truly equivalent to maintaining bisection upon a stationary point of light, and that the absolute or angular accidental error of observation is the same for all stars throughout the range of declination.

4. That good observations can be secured without previous practice.

5. That, owing to the rapidity with which the automatic signals can be recorded, a greater number of stars can be observed in a given time. The author states that his observinglist contained sixteen stars per hour to ten stars per hour in the customary lists of the Survey.

Mr. HAYFORD's fourth conclusion, as given above, when we consider the immense advantages of the method, is not so extravagant as it seems. However, the author would of course prefer an observer at the outset at least accustomed to the manipulation of delicate instruments. As the author notes, the remarks of observers at beginning plainly indicate their perturbation of mind as well as of hand in attempting to following a fast-moving star. But the persistent observer may be assured that some fine night there will come the satisfaction of seeing the star-images "go to sleep" on the apparently stationary thread. Mr. HAYFORD remarks that more practice simply reduces the accidental error by about twenty-five per cent. The present writer has had some experience with the new micrometer on a larger instrument. He found that his

probable error of a single star-signal started at ± 0.07 after observing a number of stars, and became±0.030, under favorable conditions, after considerable practice scattered over a number of months, or the same as for the old key method. As in the making of a telescopic objective, it is the last stages of improvement that take by far the most time and care.

In this connection, also, it may be remarked that Mr. HAYFORD calls attention, in the course of his discussion, to two or three "curious facts" which cropped out in the course of his investigations; but he does not mention the fact, as shown. in his Table V, that one observer, belonging to his very lowest class, takes the prize in the form of the smallest probable error of a single observation of a star, not a large amount of data involved, it is true, but more than for some of the other observers. Ladies always catch the largest fish in camping-parties, and here a young woman seems to incur the smallest error.

It should be admitted that all the advantages are not with the new micrometer; though its disadvantages, so far as seen, concern only the physical comfort of the observer, especially in cold weather. With the old method one could wear heavy gloves, but would hardly attempt to do so while turning a delicate micrometer. With the old method eve and hand could enjoy little intervals of rest between threads, while with the new both must be kept steadily at work; but this is modified if one is content with a short series of many signals close together, and with long practice the feeling of intense strain wears away. Still, it is the writer's experience that any unexpected noise or incident is more disturbing than with the old. method. But, of course, such incidents rarely occur in wellregulated observing.

. Mr. HAYFORD's conclusions are all borne out by the experience of the German observers. They have found the relative personal equations reduced to one tenth of their former value, and so small as to be masked, if existing at all, by the minute. outstanding accidental error. ALBRECHT found that the mean error of a single night's determination of longitude was reduced from values between 0.043 and 0.064 to values between 09.020 and ± 0.026. Also, they have found their results singularly free from systematic errors of all kinds, and

that a moderate amount of practice suffices for making tolerably good observations.

The last section of Mr. HAYFORD'S publication is a review of the literature of the transit micrometer. To a professional reader this is the most important part; for, as already noted, Mr. HAYFORD's work had a special purpose, while here he gives an excellent summary of a list of fifteen different publications-all but one by "those blessed Germans "-extending over the past sixteen years. The first published suggestion of this method came from Director CARL BRAUN, of Kalocsa, Hungary, in 1865. He attempted to construct a clockwork which should drive the movable thread, but was unsuccessful. This has been completely accomplished of recent years; but the observer's hand is required to maintain the finishing touch on the micrometer. Doubtless during the years following BRAUN's first publication, there were many suggestions among astronomers to the same purpose; and the author records one definitely made by Mr. F. D. GRANGER, of the Coast Survey, in 1878.

Following closely upon the announcement by REPSOLD of the successful performance of his first transit micrometer, verifications of its superiority were presented in 1890 and 1891 by Dr. TH. ALBRECHT before the (European) International Geodetic Association and by Professor E. BECKER, who had applied the new micrometer to transits of the broken type. In a publication of the Prussian Geodetic Institute in 1901, ALBRECHT declared that the superiority of the new method was so complete that it should be employed in all primary longitude work. The experience of OERTEL and COHN in Germany and of the Washburn Observatory in this country with the new micrometer on meridian-circles-as transit instruments of the straight telescope type-demonstrated its superiority for the larger observatory instruments. And there would seem to be left no question of the desirability of the new micrometer for the present instruments of the U. S. Coast and Geodetic Survey. A. S. FLINT.

The Fifth Satellite of Jupiter.-Number 580 of the Astronomical Journal contains an article by Professor E. E. BARNARD, giving a long series of observations of the fifth satellite

of Jupiter, made with the large refractor of the Yerkes Observatory during the years 1903 and 1904. These measures, as well as previous ones by the same observer, were made by measuring with a micrometer the distance of the satellite from the limb of Jupiter, and then applying a reduction factor for the semi-diameter of the planet. The values of the reduction factor were taken from a table previously constructed by Professor BARNARD from an elaborate series of observations of the apparent diameter of Jupiter. A few measures were made by determining distance and position-angle of the fifth satellite with respect to one of the brighter satellites. The two methods are quite different, and each has its advantages and each its disadvantages.

There is a third method of determining the position of a satellite which seems not to be used by observers of these objects. Why should not the position of a satellite be determined in the same way that the position of a comet or an asteroid is determined,-namely, by measuring its position on the sky with respect to a star or stars of known coordinates? Such observations can be made as accurately, and in most cases more accurately, than by either of the two methods usually employed; and it would seem, theoretically at least, to be a decided advantage to the computer who handles the observations to have the position of the satellite referred to fixed points rather than to constantly moving points. The object of all observations of position of satellites is to obtain material from which to compute the orbit of the body, and this can be done as easily, or more easily, from accurate right ascensions and declinations of the body than from the observations ordinarily made.

This question is worth looking into, and I hope at no very distant date to examine it more critically in order to determine just what observations should be made upon a satellite in order to compute the elements of its orbit with the greatest possible accuracy and facility. It may be that a combination of the observations mentioned should be made. S. D. T.

Figure of the Sun.-In number 104 of these Publications attention was called to an article by Dr. C. L. POOR on the variable figure of the Sun. In the Astrophysical Journal for

December Dr. Poor has contributed a second article on the same subject. This second article deals with the results obtained from an elaborate series of observations upon the diameter of the Sun made by Messrs. SCHUR and AMBRONN, with a six-inch Repsold heliometer of the Göttingen Observatory. These observations extend over a whole sun-spot period, from 1890 to 1902.

Dr. Poor finds in these observations further confirmation of the deductions obtained from his previous investigations,— namely, that the equatorial diameter of the Sun increases, with respect to the polar diameter, at the same time that the number of spots increases, and vice versa. The amount of the variation in the ratio between the equatorial and the polar diameters, however, is not so great in the observations of SCHUR and AMBRONN as was obtained from the Rutherford photographs.

The Astronomische Jahresbericht, prepared hitherto by the late Professor WISLICENUS, will now be undertaken by Professor BERBERICH, of the Recheninstitut in Berlin.

The following notes have been taken from recent numbers of Science:

The Paris Academy of Sciences has awarded the Lalande prize to Professor WILLIAM HENRY PICKERING, of Harvard University, for his discovery of the ninth and tenth satellites of Saturn.

At the New York meeting of the Astronomical and Astrophysical Society of America, on December 28-30, 1905, the following officers were elected for the ensuing year: President, E. C. PICKERING; First Vice-President, G. E. HALE; Second Vice-President, W. W. CAMPBELL; Secretary, G. C. COMSTOCK; Treasurer, C. L. DOOLITTLE; Councilors, E. B. FROST and HAROLD JACOBY. Councilors ORMOND STONE and W. S. EICHELBERGER hold over from the preceding year. The time and place of the next meeting will be determined by the Council.

The will of the late CHARLES T. YERKES, who owed his large fortune to the direct application of recent advances in science, makes provision for three important institutions, which are to bear his name. The Yerkes Observatory, to