star.

well apart, and 1914, when the fainter | trometer wire across two certain scale star is seen on the photograph just divisions. They are convinced of the emerging from the proper-motion proportionality of the photo-current to light-intensity within the range of 22 magnitudes measured. By careful insulation and by keeping all surfaces dry they largely eliminate the "dark effect." The atmospheric extinction, depending greatly upon spectral type, is of course troublesome. The small variations of Cephei and other stars through four or five hundredths of a magnitude are well determined. Albrecht's conclusion that the light maxima of the Cepheids occur near passage of the descending node is not borne out. Provisionally, the probable error of one comparison is given as ±0.0060 magnitude.

An extensive programme has been undertaken at Greenwich of determining the parallax of stars down to the ninth magnitude with proper motion of more than 20" per century. Forty new parallaxes have just been published. The mean parallax of 22 of these, between proper motion 20" and 30′′, is 0′′.017. The probable error averages slightly less than 0".01 (M. N., lxxv, 592).

The first attempts to determine stellar parallaxes with the Mt. Wilson 6-ft. reflector have met with marked success. The Cassegrain combination, with an equivalent focal length of 80 ft., gives a one-fourth larger scale of plate than does the Yerkes 40-in. With this instrument it is intended to extend to stars of smaller proper motion but of the same spectral types the list, recently published by Adams and Kohlscheutter, of the absolute motions and absolute magnitudes of 100 stars with known parallaxes. Thus far five parallaxes have been completed. The mean probable error is not quite 0.006, a distinct gain. The mean parallax of the five stars is 0".020, while according to Kapteyn's table, based on statistical studies, it is, for this spectral type, magnitude and proper motion, 0".014 (Proc. Nat'l Acad., i, 187).

The selenium photometer has definitely decided the question of the variability of d Orionis. Stebbins finds it an eclipsing variable with a range of 0.15 magnitude, with a period in agreement with spectroscopic results. He finds evidence of increased brightness at periastron and of the brighter state of the advancing sides of the stars.

The

The details of the work of Coblentz with thermo-couples attached to the Crossley reflector are at hand (Lick Bull. No. 266). The thermo-element differs from the selenium and photoelectric cells in that it absorbs equally all the radiations of all frequencies falling upon it. Further, the response to the stimulus is proportional to the energy falling upon it. radiation sensitivity was such that, when used on the 3-ft. reflector, a deflection of one millimetre would have resulted, had the apparatus been exposed to a candle at a distance of 53 miles with no intervening atmosphere. In order, however, to do much successful work on stellar spectral energy curves, a sensitivity one hundred times as great is desirable, a sensitivity which is considered possible of attainment by using a 7-ft. reflector and by increasing the sensitivity of thermo-couple and galvanometer. Measurements could be made on stars as faint as the 6.7th magnitude. It was found that red stars emit from two to three times as much total radiation as blue stars of the same photometric brightness, but that a much larger portion of the radiation of the

Photometry and Radiometry.-A number of very important photometric studies have appeared during the year. The details of the work of Guthnick and Prager with the photo-electric cell are presented in the first publication of the Babelsberg Observatory. In accuracy and applicability the method is said to have come up to all hopes. One difficulty experienced in their study of variable stars is in finding comparison stars which are themselves not subject to variations not too small to be detected with the new apparatus. The velocity of the electrometer deflection has been found proportional to the brightness of the star, a fact which makes it possible to observe, as a measure of the light, the length of time between the transits of the elec

red stars lies outside the region to which the eye is sensitive. Assuming various estimates of the total light of the stars in terms of Polaris, and of the spectral distribution of the stars, he finds that if the total radiation from all the stars which falls upon one square centimeter of the earth's surface were absorbed and conserved, it would require from 100 to 200 years to raise the temperature of a gram of water one degree C.

A great many visual observations of variable stars have been made during the year. Shapley has published a voluminous and very valuable discussion of the light-curves of all eclipsing variables for which reliable data were available (Princeton Contribution No. 3). The number of systems is 90 and elements have been determined for them with much care on the hypothesis of uniform brightness of disk as well as of darkening toward the limb. Many interesting conclusions are drawn and suggestions made for future observation.

702). The observations of the shallow secondary minima are less numerous than those of the primary minimum but furnish considerable support to the author's conclusion that the line of apsides of each system revolves in the shorter of the two inequalities. This is about the rate to be expected from the measured ellipticity of the stars. From the discussion of observations of the eclipsing system RV Ophiuchi, not yet published in detail, Dugan finds an asymmetry in the light-curve, similar to one observed in the curve of RT Persei, which he considers should be attributed to the greater brilliance of the advancing sides of the stars (Proc. Am. Phil. Soc., liv, 52).

Photographic photometry now rests on fairly secure foundations. Many reference stars over a wide range of brightness are now available. In the The observations of RT Persei and Astrophysical Journal (xli, 206, 259) Z Draconis by Dugan (Princeton Conappear summaries of the elaborate in- tributions Nos. 1 and 2), supplementvestigations carried on by Seares at ed by recent observations by him and Mt. Wilson since 1910. He has de- photographic observations at Hartermined the photographic magni-vard running back many years, show tudes of 617 stars and the photo- distinctly two inequalities in the pevisual magnitudes of 339 stars near riod of each (M. N., lxxv, 692, the North Pole. The former range from the 2.5th to the 20th, the latter from the 2d to the 17th magnitude. The agreement of the Mt. Wilson and the Harvard scales was already close from the 10th to the 15th magnitude. This range of agreement has been extended in both directions by correcting the Harvard results for color and distance from the center, and by removing the corrections applied for difference varying steadily with the order of exposure. Seares believes that the plate comes to its final state of equilibrium soon after the beginning of the first exposure. The Cepheid variable RR Lyrae has been investigated photographically by Martin and Plummer (M. N., lxxv, 7). They believe that two minor sin- Spectroscopy.-Observers at Lick uosities in the light curve are estab- have found that the nuclei of planlished by the observations. These etary nebulae show the characteristic secondary maxima are found to oc- lines of Wolf-Rayet stars and that, on cur at times when Kiess (from a the other hand, the gas envelopes surcomparatively small amount of data) rounding certain Wolf-Rayet stars found that the star's light becomes show the nebular lines. In the enbluer, and an attempt is made to ex-velope of one star there are no helium plain the curve by radial pulsations in the atmosphere of the star. A photographic determination of the difference in brightness of the components of many double stars has been carried out by Hertzsprung (A. N., 4783). When combined with the knowledge of the difference in visual magnitude, these data give the difference in color of the components.

lines, while this element is well represented in the star itself. Wright concludes that the envelope is condensing into the star, helium being the first to settle.

When the observed velocities of a spectroscopic binary are of about the size of accidental errors, Schlesinger finds that the shape of the frequency curve of observed velocities often tells

him whether orbital motion really ex-tude 3 which stops the further ists and the character of the orbit. increase in effective wave-length with The method has, of course, a wide ap- decreasing absolute brightness (Ap. plication to all periodic phenomena J., xlii, 97, 111). (Ap. J., xli, 162).

Frost has verified with the spectrograph the interferometer observations of Fabry and his associates (4. Y. B., 1914, p. 587) which showed marked differences in the radial velocities of neighboring portions of the Orion Nebula. Within 2' of the Trapezium Frost finds velocities ranging from 6 to 17 km. per sec.

General Studies.-Turner has made counts in various zones of the Astrographic Catalogue and finds that the ratio of faint to bright stars varies with right ascension. He attributes a defect of faint stars to the presence of obscuring matter and from his counts is able to draw an approximate chart of the obscured patches in the sky (M. N., lxxv, 57, 143, 148, 465, 601).

Lindemann adopts the theory that novae are produced by collisions with dark stars, and finds that there are about 4,000 times as many dark stars as there are bright ones, and that the average life of a star, i. e., the period between collisions, is of the order of 10 years (M. N., lxxv, 178). From an examination of the Franklin-Adams charts, Melotte finds that the great majority of the star clusters are distributed over a zone within ±30° of the galactic plane, and that only a few, mostly globular, are to be found outside these limits. Further, it is to be noted that there is a distinct tendency for the globular clusters to crowd together about longitude 325°, and that opposite this point the clusters appear spread out over a somewhat wider range in latitude and include some of the most extended clusters (Mem. R. A. S., lx, Pt. v).

Some years ago Campbell announced his discovery that the radial velocities of stars are a function of the spectral type, the solar-type stars moving more rapidly than the helium stars. Naturally this refers only to stars bright enough to be observed with the spectrograph. In making this selection a relatively large num ber of near stars of types F to M have been observed. Adams and Kapteyn find that for all spectral types the average radial velocities show a regular increase with the proper motions, and that the very distant stars of types F to M show a low average velocity. The observed relation between spectral type and velocity may, then, be largely one between distance and velocity, the stars near the sun moving more rapidly than the more distant ones. There are also some indications, found independently by several writers, of a change in radial velocity with absolute luminosity; the brighter stars moving more slowly than the fainter ones. The average velocity of stars of very low luminosity is very great. If these very faint stars are stars of small mass, we are led to the conclusion that the motion of stars is a function of mass (Ap. J., M. N.).

The average radial velocity of seven irregular, extended nebulae observed at Lick is 10 km. per sec., about the same as the velocity of the B-type stars. Of 73 planetary and regular formed nebulae, the 39 large ones are found to have an average velocity three times, the 34 of diameter less than 5" five times, as great. The old hypothesis that helium stars have in general evolved from planetary nebulae appears, therefore, hardly tenSeveral investigations of the color able. Campbell suggests that the of stars have been published during planetary nebulae may result from the year. They all bear evidence to stellar collisions, the transformation the greater redness of faint stars. being more far-reaching than in the Hertzsprung finds that with each pho- case of novae. These 73 nebulae are tographic magnitude the effective apparently a part of our stellar syswave-length increases about 30 A. tem, as they show a marked preferProceeding, however, from the abso-ence of motion in directions making lutely brightest stars to the fainter but small angles with Kapteyn's two ones, a new element evidently comes stream motions (Proc. Natl Acad., into action at about absolute magni-i, 496).

XXIII. GEOLOGY, METEOROLOGY, AND GEOGRAPHY

GEOLOGY

DYNAMICAL AND STRUCTURAL | part of the closing Cretaceous period

Geophysics. In an article "On the Earth Considered as a Heat Engine" (Proc. Nat. Acad. Sci., i, 81-86, 257258), G. F. Becker discusses the results of recent physical experiments having a bearing on the state and temperature of the earth's interior in geological time. Of the results published by Hayford, Helmert and their

and that the sedimentary formations ment, especially where the latter is were originally continuous over the calcareous." existing mountains. As a corollary to these conclusions, it is pointed out that no marked mountain-building affected the region prior to the Upper Cretaceous. The well known but neglected doctrine of shifting zones of sedimentation under the control of change of level of land and sea, proposed by Huxley and brought out by Robert T. Hill in the case of the Cre-associates, Becker remarks that, taceous of the Black Prairies of Texas, are applied by Lee to the explanation of the Dakota sandstone problem. (See also Physical Geography, infra.)

Faults.-E. M. Kindle and L. D. Burling (Canadian Museum Bull. 18, Geol. Ser., 28, 1915, pp. 1-23) describe the important group of faults which traverse the Palæozoic sediments of the St. Lawrence valley and form the boundary between the sediments in the border of the Pre-Cambrian shield of Canada in that region. The faulting is of the normal kind and is shown to be of post-Ordovician date. The authors point out that the recognition of this fault greatly alters the interpretation of the distribution of land and sea about the Canadian shield as shown on palæogeographic maps. From 2,000 to 4,000 ft. in thickness of Palæozoic strata must once have extended northward over the Pre-Cambrian area of the Lauren

tian highland, extending far towards the north the area presumably invaded by the early Palæozoic seas.

"since they have compelled us to concede that the earth is even now in a condition of approximate isostatic equilibrium, it seems impossible to believe that it has not been so in the past." Arguments are given for the thesis that the mechanism of the earth regarded as a heat engine is competent "to bring about all of the dynamical effects with which geology has to deal."

Coral Reefs.-The geological aspects of the coral-reef problem continue to call forth contributions in which the nature and cause of the change of level of coral islands in relation to the sea are the chief points of interest. To the classic work of Darwin and Dana, W. M. Davis, as the result of a Shaler Memorial expedition to the Pacific Ocean in 1914, brings confirmatory observations and conclusions. To quote a paragraph from his summary:

dence, supported by Dana's principle Darwin's original theory of subsiof shore-line development, gives by far the most satisfactory explanation of ited in the Pacific or studied on largeall the barrier reefs that I have visscale charts, and as atolls often occur in association with barrier reefs, Darwin's theory of subsidence appears to give the best explanation of such atolls also. Atolls that are not associated with barrier reefs may be of some other origin, but this seems very improbable.

The apparent cause of the submergence of coral islands is regarded as subsidence of the ocean floor (Am. Jour. Sci., xi, 223-271).

Details of geological structure are described in various papers recently issued. W. J. Miller ascribes the intraformational contorted limestone bed at Trenton Falls to differential movements within the mass of the limestone coincident with overthrust faulting, in opposition to the hypothesis advocated by Hahn that such contortions are due to the slipping of strata during the period of deposition (N. Y. State Museum Bull. 177, 1915). In the same bulletin J. M. R. A. Daly (Proc. Am. Acad. Arts Clarke returns to the geology of Per- and Sci., v, 157-251) restates in amcé, province of Quebec, whence he plified form the glacial-control theory draws illustrations of the theme that of coral reefs, in which the author "mountains of the Catskill type and holds that a submarine plateau is the age" probably owe their isolation as dominant feature of coral seas. The mesas "to the rifting of their sides theory implies a lowered wave-base by solution of the underlying pave-and a lowered base-level for rivers