formed so readily that it is sometimes almost impossible to obtain solution of the mixed colours free from it.

an aqueous All these substances belong to a single natural group, distinguished by certain marked characters, and since the mixture of several of them has been named Phycoerythrin, it may be well to call it the Phycoerythrin group. It differs entirely from the Erythrophyll group, described in my paper "On the Colouring Matter of Leaves," in the fact that the position of the bands is very little, if at all, changed by the addition of weak acids or alkalies; but when they are stronger, actual decomposition occurs, some of them being more easily changed by acids and some by alkalies. The difference in the spectra shown in the woodcut cannot therefore be caused by any variation in the acid reaction of the juice of the plants, as is the case in the colouring matter of the petals of some flowers. The possibility of this must always be borne in mind in applying such a method of comparison, and the effect of various reagents must always be ascertained before the actual identity of the colouring matters can be inferred from the agreement in the bands in only one particular condition. Much may be learned by acting directly on the plants themselves.

The consideration of the various spectra described above seems to lead to the following conclusions :

1. When a spectrum shows two absorption-bands, like No. 5, they should not be considered due to one single substance until satisfactory evidence of the fact has been obtained. The solution should be allowed to undergo slow decomposition, and be repeatedly examined, in order to ascertain whether both bands disappear in the same proportion, and also the action of various reagents observed, in order to learn whether one band can be permanently removed without the other, making of course due allowance for any change that may depend merely on an acid or alkaline reaction.

2. When more than two bands are seen in the spectrum, as in No. 3, and they are not at nearly equal intervals, the compound nature of the substance may be considered so probable, that further examination should certainly not be neglected.

3. When there is broad shading about a narrow absorptionband, as in No. 2, it is important to ascertain whether or not it is due to the same substance. There are certainly many cases in which I have always concluded that both are due to the same, but examples like this evidently show that such an opinion ought not to be formed until after further examination.

The occurrence of so many associated colouring matters, as in Algæ, may be rare. It must not be supposed that I imagine that whenever there are two or more absorption-bands they are due to two or more independent substances. As an example of what I look upon as satisfactory proof of the contrary, I will describe some

facts connected with the well-known spectrum of blood. If after exposure in a dry state to the air for some weeks, until the hæmoglobin has been changed into methæmoglobin, a small quantity of the double tartrate of potash and soda be added to the aqueous solution, and afterwards a very minute portion of the double sulphate of protoxide of iron and ammonia, the methæmoglobin is deoxidized and reconverted into hæmoglobin, as described in my late paper "On Blood-stains." Here then we have a decomposition gradually effected by the atmosphere, and if two different substances had been present, it is extremely probable that they would have varied in the rate of change, so that there would finally have been an alteration in their relative proportion, and thus when deoxidized there would not have been the same relation between the absorption-bands as in fresh blood. I find, however, that the agreement is complete. Moreover, if the colouring matter had been a mixture of two substances, it is extremely probable that there would have been some such variation in their relative amount in the blood of very different animals, as occurs in the colouring matters of different Algæ. In order to ascertain whether this is the case, I carefully compared side by side the spectra of human blood and that of the small annelids so common in stagnant pools, and found that the position and relative intensity of the two bands was exactly the same.

Such then are the principal conclusions that have been forced on my attention in carrying out these investigations. For my own part I must say that they make me think that many of my previous observations require further examination, in order to ascertain whether I have not sometimes believed that I was examining a single substance when it was really a mixture. For the future I shall certainly be quite alive to the importance of the principles described in this paper, and trust that what I have said will serve to impress it on others, and assist them in carrying out similar inquiries.

*Monthly Micros. Journ.,' vi., 1871, pp. 9-17.

III.—On Spectra formed by the passage of Polarized Light through Double-refracting Crystals seen with the Microscope.

By FRANCIS DEAS, M.A., LL.B., F.R.S.E.

PLATE XCV.

Ir is familiarly known as one of the commonest experiments in optics that when a beam of polarized light is passed through a thin film of mica or selenite, and subsequently analyzed either by reflexion or by double refraction, two colours are seen complementary to one another, and alternating with one another at each 90° of a revolution of the analyzing plate or prism.

It might be expected that the coloured light thus obtained would, if thrown into the form of a spectrum by means of dispersion prisms, exhibit some peculiarities, and such is the case, as will be seen from the following experiments:

To make the experiments intelligible, it may be well in the first place to say a few words about the instrument employed, and the method of using it.

Any spectrum microscope ought to answer the purpose, provided that in addition to the spectroscopic arrangement a pair of Nicol's prisms can be attached, one below the stage and the other over the eye-piece. Both should be capable of being rotated, and it tends much to facility of working as well as to exactness of result that both the polarizing and the analyzing prism should carry graduated heads, so that their axes may readily be turned to any required degree of inclination to one another.

The instrument I employed was a large Smith and Beck. The spectroscopic arrangement consists of an adjustable slit attached to the under-part of the sub-stage below the achromatic condenser, and a set of direct-vision prisms which are inserted in the body of the instrument immediately above the object-glass.

By proper focussing, an image of the slit is thus formed by the achromatic condenser in the focus of the object-glass, and a fine spectrum obtained filling the whole field.

EXPLANATION OF PLATE XCV.

FIG. 1.-Illustrating the dividing and re-union of the bands, as described. 2.-Spectra formed by ordinary and extraordinary ray, partly overlapping. The blue of the upper spectrum seen through the black band belonging to the lower spectrum. The yellow of the under-spectrum seen through the black band of the upper.

3.-Rings and brushes of crystal of sugar-candy.

4. Showing displacement of the rings and absence of brushes of same crystal when light circularly polarized before and after its passage through the crystal.

5.-Lemniscates and brushes of crystal of nitrate of potash.

VOL. VI.

L

This arrangement, it will be seen, differs considerably from the spectrum microscope in common use in which the dispersion prisms are placed close to the observer's eye, the slit being in the focus of the eye lens. The former arrangement has this manifest advantage, that owing to the distance of the prisms from the eye, the spectrum fills the whole field; also, that the apparent breadth of the spectrum can be varied at pleasure by a change of the magnifying power employed. Each form of arrangement has, however, its advantages as well as disadvantages, which it would be out of place to discuss here.

The polarizing part of my apparatus consists of two Nicol's prisms, for one of which, when desired, a double-image prism can be substituted.

The polarizing prism is carried on the sub-stage. It is inserted just above the slit in a short tube in which it can be freely turned by a graduated head. The analyzing prism is placed in the usual way-in a cap over the eye-piece.

The film of selenite to be examined, having first been mounted in balsam between two thin glasses, is placed on the stage of the microscope like an ordinary object.

It is a great convenience in this class of experiments to have the stage of the microscope not only capable of rotation in the optical axis of the instrument, but graduated.

By this means we can at any time, without displacing the film under examination, adjust its neutral axes at any required angle to the plane of polarization.

With regard to the mounting of the selenite films for examination the following method will be found convenient :-Make in the turning lathe several wooden disks about two inches diameter and one-eighth of an inch thick. Through the centre of each a hole must then be bored of about half an inch diameter. A small portion round the hole is then scooped out so as to form a cup, and in this the selenite is placed and secured with sealing-wax.

The axes of the selenites are then determined and marked on the rims of the disks.* In this way any two or more selenites can be used in combination with their axes set at any required angle to one another.

It remains only to trace the course of a beam of light in passing through the foregoing combination. First the ray, having been reflected from the mirror, passes through the slit. It is then polarized by the first Nicol's prism, after which it passes through the lenses of the achromatic condenser, and appears as an image of the slit in the focus of the object-glass. Having passed through the selenite and the object-glass, the ray enters the dispersion prisms and is drawn out into a spectrum. This is magnified by the eye

*The graduated rotatory stage above mentioned, and which is supplied by Smith and Beck, affords a ready means of doing this.

piece through which the ray, having passed, is lastly analyzed by the second Nicol's prism.

The loss of light from the number of the above media is not so great as might be supposed, still an intense source of light is desirable for satisfactory results. A good artificial light placed close to the mirror will be found the best. In diffused daylight rays are apt to enter the object-glass by reflexion from the brasswork without first passing through the polarizer, by which the beauty of the spectrum is impaired.

To understand the bearing of the experiments, it is necessary to keep in view the different effects of a doubly-refracting film upon polarized light, according to the position of its axes, with respect to the planes of polarization.

Suppose we take a film of selenite, such as those commonly sold as an adjunct to the polarizing microscope, giving, as its two colours, a pinkish red and its complementary green. Such a film will, if examined between two Nicol's prisms, act on the light according to the following laws:

1st. When a neutral axis of the film is in the plane of primitive polarization, the film will exercise no influence on the light; if therefore the prisms are set with their axes perpendicular the field will remain dark, if the prisms have their axes parallel the field will contain only white light.

2nd. If the prisms are placed with their axes perpendicular, and the film is made to rotate, there will be four points of darkness at each quarter of a revolution, viz. when an axis of the film is in the plane of polarization, and between these four points, the same colour (say green) will occur.

3rd. If the prisms are set with their axes parallel, and the selenite is rotated, the field will be white at the four points where it was previously dark, and of the complementary colour (red) between each of these four points.

4th. If the selenite is fixed with its neutral axis inclined 45° to the plane of primitive polarization, and the analyzer made to rotate, the field will be alternately red and green in the four quadrants.

5th. The colours are always of maximum brightness when the axes of the prisms are perpendicular or parallel, and the axes of the selenite inclined 45° to the plane of polarization.

Suppose, now, we repeat the above experiments, using the polarizing spectrum microscope above described, and let us call the point in the revolution of the selenite at which either of its axes is in the plane of primitive polarization, the zero point, from which the number of degrees through which it is turned are measured.

Let the prisms be set with their axes perpendicular to one another, and the selenite rotated on the stage. The spectrum will