REVIEWS. Untersuchungen über die Spectra von 528 Sternen. H. C. VOGEL und J. WILSING. Publicationen des Astrophysikalischen Observatorium zu Potsdam. No. 39. Bd. XII, Stuck 1. Pp. 73. Engelmann, Leipzig, 1899. WITH the systematic mode of presentation and fullness of detail that has made the Potsdam Publications the models of their kind, this work gives the results of the investigations of the photographic spectra of the brighter northern stars belonging to spectral Class I. The photographs were made by Professor Wilsing, with a small one-prism spectrograph attached to the 13-inch astrographic refractor, chiefly during the years 1894 to 1897; the examination of the negatives with a magnifying glass and the classification of the spectra was the work of the two authors together; while the measurement of the spectra and the comparison with the results of other observers was made by Professor Vogel. In the introduction a rather unfavorable opinion is expressed of the sufficiency of the photographic method, as contrasted with visual observations, for distinguishing between spectra of Types II and III, and for differentiating between Types IIIa and IIIb. While hardly agreeing with this opinion, in view of the recent great improvements in sensitizing photographic plates, and of the actual results obtained at the Yerkes Observatory in photographing spectra of the two latter types (this JOURNAL, August 1899), we may heartily assent to the statement of the advantage of the photographic plates for studying spectra of Type I. Upon the discovery of helium and the accurate determination of its spectrum by Runge and others, during the progress of this work, Professor Vogel applied the new results in extending, and making more precise, his well-known classification of stellar spectra. This is given in some detail, and although his paper on the subject has already appeared in this JOURNAL (2, 333, December 1895), it may be well to recall the new features of the system. Class Ia now receives three subdivisions, essentially as follows: Ia. Spectra in which only the broad and strongly developed hydrogen lines appear. Ia2. Those containing other metallic lines, especially of calcium, magnesium, and sodium, but no lines of helium. K is sharply bounded, but does not reach the breadth of the hydrogen lines. Lines of other metals are faint. Ia3. K is of nearly equal intensity to the hydrogen lines; is rarely sharply bounded; or may be broader and stronger than the hydrogen lines. Helium lines cannot be seen, but numerous and strong lines of different metals, particularly of iron, appear. This subclass constitutes the transition to Class II, in which the hydrogen lines are no longer dominant, and only the calcium lines H and K are striking on account of their extraordinary breadth and diffuseness. Ib. Spectra characterized by the presence of the helium lines, in addition to conspicuous hydrogen lines, and more or less lines of calcium, magnesium, sodium, and iron. Class Ic is now also subdivided, into Ici. Spectra with bright hydrogen lines. Ic2. Those having bright lines of helium, calcium, magnesium, and other metals in addition to those of hydrogen. As good examples of the different subclasses we may cite the following stars: Iai. A Bootis; Draconis; ß Cygni, seq.; λ Cygni. Ia2. a Cygni; ẞ Aurigae; ẞ Equulei. Ia3. Leonis; r Cygni; B Cassiopeiae; & Geminorum; a Canis Ib. It is distinctly stated that the discrimination between Iar and Ib, and Iar and Ia2 offers the most difficulty, and that with strong dispersion, and good conditions generally, these spectra would presumably show the additional lines characteristic of subclasses Ib and Ia2, thus leaving Iar without any representatives, so that it could be dropped from the classification. It is also pointed out that the new classification is adapted as closely as possible to the original form, so that the new first type stars with bright lines, which modern views assign to the earliest stage of development, are retained under Ic rather than made to precede Ia. This is certainly a logical procedure, and reduces the confusion which might otherwise ensue. The reviewer adheres to the opinion that all schemes of stellar classification are at present provisional, and that it will be many years before a radically new and improved order of succession can be established. Such a classification ought to be based on the results obtained with the highest instrumental power attainable, including all the measurable radiations received from the bodies--electrical (if any), thermal, actinic, visual. Meanwhile the reviewer regards Vogel's classification with its present extension as the best system for practical use. The spectra themselves, on which this research was based, extended from about 3800 to 4500 with almost uniform sharpness, due to the fact that the collimator and camera lenses were achromatized in precisely the same manner as the photographic refractor to which the spectrograph was attached. The range from λ3700 to λ4480 was 10 mm; the most convenient breadth for measuring was 0.2 mm; the average exposure was from 12 to 15 minutes. The slit was accurately set in the focus of the objective for Hd (4102). An arrangement was provided for furnishing a comparison spectrum by a vacuum tube, but was seldom used. A comparison spectrum could also be obtained by bringing a star of known spectrum immediately above and below the center of the slit before and after the exposure on the star to be investigated. Guiding was exceedingly accurate and easy with the 9-inch visual telescope of the photographic refractor. The work falls in three parts: I. The catalogue of the observed spectra; II. Measurements on plates of 130 stars; III. Comparison with the results of other observers. The catalogue gives the name, magnitude, position for 1900, and date of plate and type of spectrum, together with abbreviated but readily intelligible remarks. One good plate only was taken of a star, unless the object appeared for some reason of especial interest. Among the 528 stars, 66 of Class II and 3 of Class III fell outside of the original scope of the investigation, which was the study of spectra of Class I. Omitting these and the stars whose spectra were too faint for inclusion in the classification, and those which could not be assigned with certainty to any of the subclasses, the remaining 437 were distributed as follows: Iaz to Iar. Iar 44. Ia2 68. Ib. 100. Ia3 to IIa. 19. Ic.. I. Iaz to Ib.. 5. Double spectra: Iai and Icı, 1; Ib and Ic 2, 2; Ib and? ICI, 1. The remarks include data as to strength, character, and number of the lines, separately for the hydrogen and for the metallic lines, with estimates of the comparative appearance of H (including He) and K. Exclamation points attract instant attention to conspicuous examples of types. From this column, and the extended notes accompanying the measures in part II, we extract interesting details of some of the notable spectra. K is lacking in the spectrum of Persei, of which towards 90 plates were taken. No lines at all were visible on many plates. The spectrum is considered variable. No mention is made of the detection of bright Hẞ and Hy, originally discovered by Campbell, which are double (as is always the case with y Cassiopeiae) on a good plate obtained with the 40-inch refractor by the reviewer. Hence the assignment of this star to Class Ia 2?- of course the only legitimate assignment on the basis of the plates and apparatus employed-is incorrect; more powerful dispersion shows it to belong to Class Icı. A similar remark would apply to Persei, and some other stars of this variety. The spectrum of y Cassiopeiae appeared purely continuous on the five plates obtained. It is not surprising that the dark lines in this spectrum were not seen, for it is vexatiously difficult, even with powerful apparatus, to obtain the necessary contrast for revealing them. The objective prism has a singular advantage in this respect, as shown by the results at Harvard and Stonyhurst. However, the presence of the dark lines has been abundantly confirmed by the spectrographs of the Allegheny, Lick, and Yerkes Observatories. In view of the complexity of the spectra of stars of the y Cassiopeiae type, it will no doubt be necessary in future to discriminate more sharply in regard to the varieties of Class Ic. We further note that in the star in Cassiopeiae, 37 Hev., K was double on three plates; this is also found to be the case for ẞ Aurigae. It was suspected in the stars o Ursae Majoris, 68 Ophiuchi and 109 Virginis. Thirteen plates of the last star were obtained, and on several plates K was double, or, at any rate, two distinct lines were seen at the position of K. On some plates of έ Ursae Majoris the hydrogen lines were double or brightened up in the center. The spectrum of Algol showed no changes in character having relation to the light period, but in some cases the hydrogen lines were brightened up in the center. Numerous plates were obtained of a Cygni, as slight changes were suspected in its spectrum. The micrometric measures of wave-lengths, made on the spectra of 130 of the stars, were reduced graphically from numerous solar and metallic spectral plates. They are given to the tenth-meter, and are based on Rowland's scale. In most cases a single measurement of a spectrum was sufficient to certify to the presence of helium lines, and an agreement within ±1.5 tenth-meters was regarded as adequate for identification - an accuracy which all will agree was abundant for the purpose. A valuable accompaniment to the measures is the estimate, on a decimal scale, of the intensities of the lines. The average number of lines measured was, of course, not large in this class of spectra — perhaps twenty- but in the case of a Cygni as many as fifty-five lines were determined. ẞ Lyrae and P Cygni are both classified as "Ib and Ic 2" and are described at some length. Comment on these extraordinary spectra would lead us at once to transcend the bounds of a review. A tabular exhibit is made of the helium lines found in stars of Class Ib, with their intensity in the spectrum of each star thus measured; and another table summarizes the lines not due to helium or hydrogen, with the stars in which they are found. The results obtained are compared with the "Draper Catalogue" and with Miss Maury's "Spectra of the Brighter Stars." After pointing out cases where the "Draper Catalogue" gives discordant type assignments of the same bright star from different plates, and remarking that this is not to be wondered at, in view of the difficulty of securing a proper exposure for plates containing so many stars of different magnitude Professor Vogel comments that "on the average, however, the comparison with our results gives a right good agreement, and full recognition surely cannot be witheld from the great undertaking of the production of a catalogue of the spectra of 10351 stars." gracefully acknowledging the value of Miss Maury's work he considers. |