PLATE VI.

This Plate gives various very beautiful transformations of the circular solar spectra as seen by viewing the solar disk of the mercurial globule portrayed obliquely on the stage, which vary in their forms according as the glasses are under- or overcorrected, as depicted in the figures, which it is needless here to describe in detail.

Fig. 34 is another form of mercurial globule, beautifully defined.

The figures displayed by the magnified artificial star for oblique reflection render it probable that the obliquely illuminated mercury globule, viewed directly in close proximity to the front glass of the microscope upon the stage, is a very imperfect test; and the methods here described are submitted as possessing very superior delicacy and convenience.

XXII. "On Comparative Vegetable Chromatology." By H. C. SORBY, F.R.S. &c. Received June 9, 1873.

The study of colouring-matters as described in the following paper requires the use of a somewhat special kind of chemistry and of spectro

483

scopic instruments designed for the purpose, but yet it cannot be looked upon as being either chemistry or spectrum-analysis. It is also requisite to make use of light as a reagent, and yet the subject could not be called photography or even photo-chemistry; and though most intimately connected with biology, it could scarcely be regarded as a branch of that science. It appears to me most desirable to adopt some name for the whole subject, including whatever modifications of other branches of science are requisite for its efficient study, and none would meet the requirements of the case better than Chromatology. I trust that the facts. I am about to describe will be sufficient to show that we have such a wide branch of science before us that it deserves some special name. Then, again, since I now purpose to describe the points of agreement and difference in the colouring-matters of various classes of plants, it appears to me that no better title could be adopted than the one I have chosen, viz. Comparative Vegetable Chromatology.

My attention has been for a long time directed to the colouringmatters of plants, but it is only within the last year that I have discovered the means of carrying out the inquiry in a satisfactory manner, and convinced myself that the most valuable conclusions are to be learned, not by looking for the rarer or more attractive substances which give remarkable spectra, but from the qualitative and quantitative determination of the different coloured constituents of complicated mixtures, some of which substances have no striking properties, and might easily be overlooked, though perhaps of great importance in connexion with the life of particular plants. The successful carrying out of this inquiry will necessarily be the work of years; for not only is the whole field comparatively unexplored, but it will be requisite to examine all classes of plants and many of animals, at various seasons of the year and when growing in different natural and artificial circumstances, and also to study the relation of the different colouring-matters to one another, and the action on them of light and chemical reagents. In every department of the subject very interesting questions remain to be answered; and the following paper must be looked upon merely as a general outline of what I have hitherto been able to learn, and as an indication of what may be expected to result from further research.

The Absorption-band-raising Power of Solvents.

The identification of the individual colouring-matters has been effected by means of their spectra, sometimes as seen by examining the plants themselves, but generally when dissolved in various liquids, with or without the addition of appropriate reagents; but any other peculiarities were taken advantage of, according to the circumstances of the case. For further particulars connected with the general method I refer to my former papers*.

* Proc. Roy. Soc. 1867, vol. xv. p. 433. Quart. Journ. of Microscopical Science,

Here, however, I would remark that the exact position of the absorption-bands given by particular substances varies considerably, according to the nature of the solvent, even though it is quite neutral and exerts no chemical action on the colouring-matter. Thus, for example, the bands may lie much nearer to the red end when dissolved in bisulphide of carbon than when dissolved in benzole or absolute alcohol. This kind of fact has been often noticed by others, but its bearing on some important questions has generally been overlooked. Apparently, as a rule, the bands are more raised towards the blue when the colouring-matter is dissolved than when in a solid state, and this power of raising them varies with the solvent; but the extent also depends very much on each particular substance. I have carefully determined the order in which various liquids thus raise bands, and compared it with the order of their specific gravity, and of their refractive and dispersive power, but have hitherto failed to recognize any simple connexion between what may be called the band-raising power and any other physical property. In some cases the position of the bands given by the solid substance appears to vary according as it is in a colloid or a crystalline state. Though at first this difference in position may seem to be an objection in the use of the spectrum method, yet it only shows the need of comparing together our specimens in exactly the same condition, and it may be taken advantage of as a means for distinguishing closely allied substances which may be influenced in a very different manner and for enabling us to ascertain the condition of a colouring-matter as it exists in an animal or plant. It is thus possible to determine whether it is in a free state or dissolved in an oil or wax, and in the latter case whether it is liquid or solid. In this manner I have been able to detect a difference in the relative quantity of oil in different parts of the same leaf. Various forms of spectrum-microscopes were employed, according to the requirements of the case; and the amount of material at command would often have been too small for examination with any other kind of instrument; so that I think I may fairly say that comparative vegetable chromatology is the first great subject which has resulted from the use of the spectrummicroscope.

Separation of Colouring-matters by chemical means.

On the present occasion I propose to consider almost exclusively those colouring-matters which are soluble in bisulphide of carbon and in fixed oils, but insoluble in water. These appear to be by far the most important in connexion with plant life, or at all events are more uniformly distributed. Those soluble in water and insoluble in bisulphide of

1869, vol. ix. pp. 43 & 358; 1871, vol. xi. p. 215. Monthly Microscopical Journal, 1870, vol. iii. p. 229; 1871, vol. vi. p. 124. Quarterly Journal of Science, 1871, vol. i.

p. 64.

carbon or fixed oils are much more numerous, and often very partially and, as it were, accidentally present: those of the former class may often be separated from one another by means of bisulphide of carbon, alcohol, and water, mixed in proper proportions. As an illustration I take the very simple case of olive Algæ, like Fucus or Laminaria. They should be well crushed, slightly dried, and heated in spirits of wine of the usual strength. Absolute alcohol must not be used, since it dissolves out the chlorophyll in such a manner that it is not carried down as usual, when agitated with bisulphide of carbon. When cold the solution should be agitated in a test-tube with so much excess of the bisulphide that a considerable quantity subsides to the bottom, carrying with it the whole of the orange xanthophyll* and the greater part of the blue chlorophyll, leaving some of the latter in the alcohol along with nearly all the fucoxanthine and chlorofucine. This solution should then be removed by a pipette, and agitated with a fresh quantity of bisulphide, the process being repeated until it subsides with only a very slight tinge of green. The alcoholic solution, on being evaporated and treated with bisulphide, yields approximately pure fucoxanthine. Uniting the different lots of bisulphide containing chlorophyll, and agitating them over and over again with fresh spirit, taking care that a small excess of bisulphide is always present, the whole of the chlorophyll may be removed, and the orange xanthophyll left in the final residue. On adding a little water to the washings containing the chlorophyll, the disolved bisulphide is precipitated, carrying down all the chlorophyll. By proceeding in a similar manner, with or without the addition of weak acids or alkalies, various other colouring-matters found in plants may be more or less perfectly separated. All these processes should be carried on without exposure to strong daylight, or else some of the constituents might be in great measure destroyed.

Photochemical Analysis.

There are many cases in which certain colouring-matters cannot be separated in a satisfactory manner by merely chemical methods, and then fortunately we are able to resort to what I propose to call photochemical analysis. The action of light on the coloured constituents of plants was made the subject of an excellent memoir by Herschel, published in the Philosophical Transactions for 1842, p. 181; but the object he had in view and the methods he employed were altogether different to mine. He adopted no means for separating the different constituents of the plants, and some of his solutions probably contained as many as ten different coloured substances; and, moreover, at that time no attention had been paid to their spectra. He exposed paper coloured with such mixtures to the solar spectrum, so as to learn the effect of each portion; and his results are of great interest, as showing that the decomposition

*The exact use of all these terms will be explained in the sequel.

is due to the luminous rays, which act with variable intensity in different parts of the spectrum, according to the nature of the substance. On the contrary, my method of photochemical analysis consists in using either the whole or particular rays of the sunlight passing through coloured glasses as reagents to decompose some of the constituents of a mixed solution, and leave others in a state of approximate purity. In employing light for such a practical purpose, it would be almost impossible to use a spectrum. Without this method the study of the colouringmatters of plants could not be carried out in a successful manner. I have sometimes studied a substance for several weeks before I was able to prove that it was a mixture by ordinary chemical methods, whereas I afterwards found that this could be easily proved by simply exposing the solution to the sun. I, however, now give only such examples as have an important bearing on the subject before me. For instance, in the case of some Algae it is impossible to separate the xanthophyll from the orange xanthophyll, and the spectrum of the mixed solution in bisulphide of carbon gives absorption-bands in an intermediate position, not agreeing with those characteristic of either one or the other. It might thus be imagined that these Alga differed entirely from allied species, in containing a single and peculiar kind of xanthophyll instead of the usual two, thus making an apparent complete break in what is really perfect continuity; but by exposing the solution to the open sun the orange xanthophyll is so much more rapidly destroyed than the xanthophyll, that in a while the absorption-bands of the latter are seen in their normal position; and knowing this fact, and comparing the spectrum with that of the original, it is easy to see that at first there must have been a mixture of the two kinds. Of course in all such experiments care must be used to prevent overexposure, and the specimens should be examined from time to time. In most cases it is also extremely difficult to separate yellow xanthophyll from lichnoxanthine; but by exposure to the sun this xanthophyll is soon decomposed, its absorption-bands disappear, and the uniform absorption of the lichnoxanthine remains. Solutions which originally give very different spectra can thus be proved to be merely variable mixtures of two or more substances, each well known in a pure state. There are, however, cases in which both constituents would be destroyed by exposure to white light; but one is changed chiefly by one kind of rays, and the other by another. For example, when the mixed solution in bisulphide of carbon of phycoxanthine and orange lichnoxanthine, obtained from certain Alge and lichens, is exposed to the open sun, the phycoxanthine is soon destroyed, but at the same time some of the orange lichnoxanthine, and it is almost impossible to cease exposing at the proper point; when kept under deep green glass, however, it is easy to decompose nearly all the phycoxanthine without materially diminishing the amount of the orange lichnoxanthine. Other