have shown that this phenomenon is due to a change in the pressure or temperature of gases. Hydrogen, for instance, with medium rarefaction gives its three characteristic lines (H, a, 8, 7). With diminished pressure the least refracting line (Ha) disappears and the gas changes color. Under great pressure the lines increase in width and others make their appearance. Similar changes accompany a variation in temperature. Hence with a great pressure and a high temperature the hydrogen line spectrum is converted into a continuous one, resulting from the general widening of the lines. Now as absorption and radiation are proportional, it follows that a gas at a high temperature and under a considerable pressure will produce thick dark lines in its spectrum. The foregoing facts furnish an explanation of solar spots, which are generally thought to be due to the absorption of light from the photosphere as produced by masses ejected from the interior of the sun. These masses, when observed near the solar edge, take the form of flames, because they are then projected on a comparatively dark sky; but when observed over the luminous disk itself they produce absorption. Just as the sodium flame when interposed between the spectroscope and a dark background gives a bright yellow line, but a dark line is observed if the dark background be replaced by a luminous source. Before applying these principles to the spectra given by the stars, it will be advisable to examine the displacement of the lines, notably the C hydrogen line in the spectrum of the solar spots. This displacement is attributed to the backward and forward motion of hydrogen gas in the direction of the visual ray. But why so? The analogous deportment of optical and acoustic phenomena furnishes the key to the explanation of this fact. It is found in acoustics that the tone of a sounding body is higher if the sounding body approaches the ear, and lower if it recedes, than the true tone of the body. A rough verification of this generally admitted theory, termed Doppler's principle, is readily remarked in the change of the tone of the whistle of a locomotive on moving towards or receding from us. The explanation is briefly this: Under ordinary circumstances sound is propagated in air with a velocity of about three hundred and forty meters per second. Now suppose that a horseman, at the exaggerated speed of three hundred and forty meters per second, were to approach a sounding body, which is emitting four hundred vibrations per second, then the horseman's ear would receive twice four hundred vibrations, and the sound consequently would be raised to the octave. Hence the pitch of a sound will be raised or lowered proportionably to the speed with which one approaches the sounding body or recedes from it. An analogous phenomenon is very striking in the phonograph, which raises or depresses the tone of the original notes sung into the instrument accordingly as the cylinder is rotated more or less rapidly. As the different vibratory rates in sound are accompanied by a change in pitch, so a difference of vibration in light motion produces the various colors. Since, however, the motion producing a change in the luminiferous vibration must be proportioned to the number of vibrations of different colors-four hundred and eighty billions for red, eight hundred billions for violet— this motion must be infinitely more rapid than that which can effect a change in the sonorous vibrations. Doppler's principle, especially when applied to light, has been attacked by not a few, yet it seems to stand the test. It is true that Doppler's attempt to account for the variation of color in changeable stars on his principle, by supposing them to move towards or recede from us is inexact; for as there exist obscure rays beyond the red and violet, these colors on a change in the ether waves would assume those shades which would have disappeared for the red and violet respectively, so that the general color would remain unchanged. But although this circumstance precludes the application of Doppler's principle where there is question of the total color of the star, still this cannot be said with regard to the change of position in the absorbing lines. The rapid motion of a star changes, in fact, the position of all the dark lines observed in its spectrum. For this motion rendering all the ether waves of the rays emitted from the star longer or shorter, modifies in the same way the waves of the absorbed light. It happens, therefore, that the dark lines are shifted towards the violet or towards the red according as the motion is forward or backward. Their intensity, however, remains the same. As to the verification of the principle in its relation to light, we may mention that it has been applied to the motion of the planets which approach the earth and recede from it at wellknown rates. Besides, Fr. Secchi, Professor Langley, and others have applied it to the sun, one limb of which is approaching us and the other receding from us as it revolves on its axis. In both instances the displacement calculated, though small, subsequent observations verified. Before examining the application of this principle to solar phenomena as made in the new method, let us return for a moment to the widening of the lines above-mentioned, and see how this phenomenon when resulting from the temperature is explained by the supporters of the theory of thermodynamics. According to modern views the most minute particles of gas are subject to a rapid to and fro motion, which increases with the temperature. The tem perature being high, each particle of the violently agitated gas produces a spectrum line, either bright by radiation, or obscure by absorption. Now the line actually observed being a resultant of all these elementary lines, which are shifted some one way some another, is wider than when the temperature is lower, for the shifting is then less. And when the lines increase in width the edges appear not so well defined as the middle portion. This explanation coupled with the fact that a denser gas must produce wider lines, satisfactorily accounts for the phenomenon of the increase in width. We may remark that the darkening of the lines arises from increase in pressure and thickness, while their ill-defined appearance is due to temperature. Lockyer, accepting Doppler's principle as true when applied to the shifting of the lines in the spectra of rapidly moving gases, has measured the velocity of gaseous eruptions on the sun's surface. The gaseous masses thus raised have not only a rapid upward motion, but once raised they move like gigantic cyclones in the solar atmosphere. Both these motions, the upward and the lateral, similar to that of our winds, can be measured by Lockyer's method. The first by observing the spectrum of a spot whilst in the middle of the solar disk, the latter by observing that of a prominence near the sun's limb. If we observe a spot when in the middle of the disk, any upward motion is a motion towards us along the visual ray; we can, therefore, measure the velocity of gases in their motion by examining the shifting of the lines in the spectrum of a spot thus located. So in the same manner by observing the shifting of the lines produced by a prominence near the sun's limb, we can ascertain the motion of the gases producing it in the direction of the visual ray, which being tangent to the sun represents the lateral motion of the gases, like that of a cyclone in our atmosphere. By this method it has been found that the hydrogen in the sun has sometimes the wondrous velocity of several hundred miles a second. Though these deductions met with some contradiction on their first announcement, eventually they have been fully indorsed by the most eminent astronomers, among whom may be mentioned Secchi, Huggins, Young, Respighi, Rayet, Ferrari, etc. From the preceding, and from other features in the new method of spectroscopic observation-features which are here omitted either because too technical or not bearing directly on our subject-it has been deduced that the nucleus of the sun (regarding which nothing certain is known) is surrounded first, by a very luminous layer, termed the photosphere, which, from all indications, is quite thick and exceedingly mobile. A second, called the reversing or absorbing layer, subtending about two seconds of an arc, equal to about one thousand miles in thickness, incloses the first. This second layer, which appears bright during eclipses, gives the Fraunhofer lines, which at times are observed to be enlarged towards the sun's limb. Outside the second layer, and apparently five times as thick, is the chromosphere layer, which is mainly composed of hydrogen and an unknown substance, by many termed helium, giving a line near that of sodium. Above the chromosphere is the coronal atmosphere, the principal constituents of which are hydrogen, and a second unknown substance giving the green line 1474 on Kirchoff's map. During late eclipses, particularly that of 1878, phenomena were observed which lead to the conviction that this gaseous atmosphere is surrounded by minute meteoric bodies, sometimes called meteoric dust, whose presence explains the coronal light surrounding the sun when obscured by the lunar disk. Whatever may be the nature of this last appendix of the sun, which some connect with the zodiacal light, it is certain that the layers surrounding the photosphere form a kind of grating which lessens in a measure the intensity of its light; and when somewhat thinned, either by the uprisings of the photosphere or by other causes, they give rise to the faculæ, or luminous points seen on the sun's surface, just as interior eruptions of gas by thickening this layer produce the spots by their absorption, as abovementioned. The consideration of solar spectroscopic observations having taught us to interpret the language of the stars, we can return to the solution of the questions proposed in the beginning, and frequently alluded to in the preceding pages concerning stellar spectra. We shall begin our explanation from the second stellar type, because the appearance of this type being altogether similar to that of our sun, nothing further need be added regarding these stars, as the temperature and pressure of their gases must be about the same as for the sun. In the first type considerable variety is noted. We remarked that the stars of this group, when examined with the spectroscope, present a spectrum bearing large ill-defined lines of hydrogen. absorption, while but few metallic lines make their appearance; therefore their gaseous atmosphere must be subject to a higher temperature and a greater pressure than in the second group. Their absorbing metallic lines are also fewer, and often not easily detected, a phenomenon which Fr. Secchi attributed to the dense atmosphere through which the lines are seen. If we should admit, for the moment, Lockyer's late theory regarding the so-called chemical elements as founded on fact, the increase in the width of the lines should be explained by the high temperature of the gases surrounding the stars of the first group,-a temperature capable of decomposing the metals, which, according to him, are reduced to simpler substances, especially to hydrogen. But, whatever be said on this point, it is evident that there exists no marked difference between the stars of the first and second groups. They gradually pass from one into the other. A similar gradation or connecting link is noted between the other two groups. The stars of these third and fourth types give indications of a comparatively lower temperature than that to which those of the first and second types are subject, the temperature decreasing from the first to the fourth. The spectra of the third and fourth groups of stars are comparable, as stated above, to that of nitrogen; that is, they are fluted. This appearance, according to the recent researches of Frankland, Lockyer, and others, seems to be owing to the presence of compound bodies not raised to a high temperature. If, in fact, the electric spark be passed through a compound body without raising the temperature sufficiently to decompose it, the spectrum produced will be composed of fluted spaces or band lines. In this respect nitrogen behaves like a compound body. If the temperature increase while the spark is passing, then, at the moment of decomposition, the spectrum changes, and, as a rule, the spectrum of the metal present becomes predominant. This being the case, it may be asked what compounds give spectra resembling the spectra produced by the stars of the third and fourth types? Fr. Secchi found-and so far as we know at the present writing the result of his experiments has not been denied that the spectrum of benzine, or, in general, of the carbon compounds, is ŝimilar to those given by these stars. It is next to impossible, however, to define which of the carbon spectra in particular corresponds to each star, as the appearances of the carbon spectra vary considerably under different circumstances. The main difference, however, between the spectra of the third and fourth types seems to be that those of the fourth are bright band spectra, while those of the third contain absorption lines. Admitting this distinction, we can regard the third type as made up of stars whose brilliant portion-the photosphere-is surrounded by compound bodies which, absorbing a portion of the stellar light, produce fluted spectra. These compound bodies could not be present were the temperature as high as it is in the stars of the first and second groups. The spectra of the fourth group show that the light of these stars arises principally from the radiation of compound gases, whose temperature, therefore, must be lower still than that of the preceding type. The following is a brief explanation of these stellar types as given by Fr. Secchi: "The stellar spectra of the first and second types produce lines of metallic absorption like those of our sun. The spectra of the third and fourth types give, besides the metallic lines, others corresponding to gases, very probably to carbon |