the object in view to produce ice, and, in fact, there may be a very great advantage gained by not doing so. Thus, for instance, Mr. King has, by an ingenious contrivance, arranged the working of his apparatus in such a manner that the refrigeration is applied directly to the material to be cooled, without making ice or using any refrigerating medium; and by this means he has succeeded, with the apparatus just referred to as capable of producing five tons of ice in twenty-four hours, in obtaining, during the same time and with the same consumption of fuel, an effect equivalent to the production of no less than fourteen tons of ice. This reduces the cost of the work done very considerably below the estimate already given, and the fact serves well to show what great benefit may be derived from the judicious application of refrigerating apparatus.

Among other purposes to which it has been proposed to apply artificial refrigeration, is the cooling of the air in dwellings, and in passenger ships passing through tropical regions;* and there are, no doubt, many other cases in which it might be usefully employed.

V. ON SOME RECENT SPECTROSCOPIC RESEARCHES. By WILLIAM HUGGINS, F.R.S., Hon. Sec. R.A.S.

IT is the intention of the writer to give in this article an account of some of the more recent additions to our knowledge of the heavenly bodies which have followed from the application to them of that elegant and most searching method of analysis for which science is laid under lasting obligation to Kirchhoff.

The circumstance that in astronomy so large an amount of new and important knowledge has been gained, will not appear surprising when we consider how peculiarly spectrum analysis, which enables the observer to be independent of his distance from the source of light, and also of the length of time that the light has been on its way to him, is adapted for an examination of the light of the heavenly bodies, which are, for the most part, independent sources of light of a high temperature. Indeed the light of the sun, of the stars, of the nebula, and of comets, was written over with unread hieroglyphic characters, which, when the key was furnished by the German physicist, revealed to us information, such as it had appeared to man almost vain to hope ever to obtain.

Spectrum analysis, which consists in the skilled interpretation of the minute peculiarities by which spectra are distinguished, and

* Mr. Shand, 'Society of Arts Journal,' xvii., 97.

the mode of its application to terrestrial objects and to the heavenly bodies, are now too generally known to need description here. It may be well, however, to state the principal forms under which all spectra may be classed, and the interpretation which, in the present state of our knowledge, we are justified in giving to these different spectra, when the light has been emitted by bodies rendered luminous by a high degree of heat. It is necessary to make this distinction, as it is not purposed to describe the spectra of fluorescent and phosphorescent substances.*

The spectra of all highly heated bodies may be referred to three typical forms.

spec

First Form.-When the continuity of the coloured band into which the light is dispersed by the prism remains unbroken either by bright or by dark lines. Now, as a general rule, such a trum may be interpreted as telling us that the body which emitted the light is in the solid or liquid state. Further than this, a continuous spectrum gives to us no information of the nature of the source of the light, for all solid and liquid bodies are characterized by a continuous spectrum, whatever their chemical nature may be.

Such is the interpretation which, as a general rule, and unless there should exist circumstances rendering a solid or liquid condition of the source of light highly improbable, ought to be given to a continuous spectrum. In certain cases gases may give a spectrum which is continuous. Dr. Balfour Stewart has pointed out that as gases and vapours possess a power of general or indiscriminate absorption, in addition to the elective absorption peculiar to each gas, it would follow that a gas when luminous would also emit light of all refrangibilities, producing a continuous spectrum, in addition to the spectrum of bright lines which is peculiar to the gas, and further, that the intensity of this continuous spectrum would be in proportion to the opacity of the gas.

Besides this consideration, the researches of Plücker and Frankland have shown that, under certain conditions of temperature and density, the bright lines of some gases expand, and so a spectrum may be produced which would not be distinguishable from that of the light of a solid or liquid body.

Second Form.-Spectra belonging to this class consist of bright lines. These lines tell us that the source of the light is luminous. gas. Further, since, so far as observation extends, each gas and vapour is distinguished by a set of lines peculiar to itself, it becomes possible to discover if any of the substances known to us are present in the source of light. This method of analysis is not invalidated by the circumstance that the appearance of the lines may be greatly modified, or even altogether changed, under dif

For an account of the spectra of fluorescent and phosphorescent bodies, the reader is referred to 'La Lumière,' by E. Becquerel, vol. i. Paris, 1867.

ferent conditions of temperature and density, as is well known in the case of nitrogen, the vapour of sulphur, and some other substances, for throughout all these changes each gas behaves in a way peculiar to itself.

As far as our knowledge goes, there appears to be but one exception to the statement that a spectrum of bright lines indicates luminous gas. Bunsen found that when solid erbia is heated to incandescence, the continuous spectrum contains bright bands.

Third Form. This type embraces all spectra in which the continuity of the colours of the spectrum are interrupted by dark lines or bands. These gaps in the spectrum, which indicate that light of certain refrangibilities is wanting, do not teach us anything of the source of light itself, but show the existence, without the source of light, of vapours at a lower temperature, which, by a selective power of absorption peculiar to them, have quenched the light of certain periods of vibration only, and have not been able to make up for the light they have taken, by light of their own. the kinds of light absorbed by each vapour correspond precisely with the set of bright lines which that vapour emits when in a luminous state, a comparison of the bright lines of substances that are known with the dark lines seen in a spectrum will show whether the vapours through which the light has passed are those of any of the bodies with which we are acquainted. The bright spectrum, in which the dark lines occur, must be questioned for information of the source of the light itself, according to the principles stated under the first form of spectra.

The foregoing statements form the canons of interpretation by which we are to be guided in our explanation of the spectra of the heavenly bodies.

The most important recent information obtained respecting the fixed stars results from the application of spectrum analysis in a new direction.

Under certain conditions, which will be stated, the spectrum of a luminous body is adapted to tell us whether that body is moving towards or from the earth.

It may be well, however, to point out in how remarkable a manner this new application of prismatic analysis supplies a want which astronomers had come to regard as one that could not be met by any method of observation within our reach. The stars, though apparently so immovable that they can serve as the figures on the dial-plate of the heavens to which all sensibly moving objects may be referred the fixed stars, as it is still convenient to call them, are not absolutely motionless, like fiery studs riveted in the canopy of heaven. These brilliant points are found to shift their places to a minute extent relatively to each other. Small displacements are found which must be interpreted to represent a proper motion.

peculiar to each. Now, by ordinary methods of observation, that part only of a star's motion which is transverse to the line of sight can be detected, for the motion a star might have in the visual direction, either directly towards the earth or from it, would not cause any apparent change of position of the star, and would therefore remain undetected.

The method of photometry was clearly too coarse and too uncertain. Too coarse, for the eye, even when aided by suitable instruments, could not hope to distinguish the minute increase or decrease of brightness which would correspond to any velocity we could with probability assign to the stars, even if the observations were repeated at intervals of a thousand years. Too uncertain, for if the variations in the transparency of our atmosphere could be certainly eliminated, the large number of stars, of which the light is known to be variable, would forbid us to regard any alterations of brightness that might be observed, as trustworthy indications of the approach or recession of the star.

Now this radial motion of a star, which eluded our methods of observation, does record itself most fortunately in small alterations, which can be distinguished in the spectrum of its light. If the star be approaching the earth, a line, either light or dark, in its spectrum will be found to have moved from its proper place towards the blue end of the spectrum; if the star be receding from the earth, the line will have moved in the opposite direction, towards the red. The amount of shift of position produced by a star's motion will express exactly the proportion which its motion bears to the velocity of light. As the rate of propagation of light is known, the velocity of the star may be found.

The refrangibility, or the colour of a ray of light, or what is the same thing, the place in the spectrum which the light would take after it had passed through a prism, is determined by the number of pulsations or waves which meet the eye, or fall upon the prism, in a second of time. The special character which distinguishes red light from violet light consists in that the waves of red light are nearly as long again as those of extreme violet light. Now the velocity of propagation through the ether is precisely the same for all the colours of the spectrum. Red, yellow, green, blue light, emitted by a distant star, would reach the earth at the same instant, and it is for this reason that a new star, at the first moment that its light falls upon a human eye, appears of its true colour. Light travels with a velocity of about 185,000 miles in a second of time, a series of waves therefore extending through 185,000 miles enters the eyes each second. Now as it is upon the length of the waves, or upon the number contained in the series that enters the eye in a second, that a judgment is formed of the colour of the light, or the place of the light in the spectrum after