this much easier than they might imagine, and that no one accustomed to the delicate manipulation of microscopic work would find the slightest inconvenience from this cause.

For micrometric purposes, however, I have adoped the method which has been for some time in use, i. e. the image of a photograph scale reflected from the upper surface of the top prism. This image passes through two collimating lenses to render the rays parallel, and can be focussed by a sliding movement of the lenses in an inner tube by means of two pins passing through slots: I have had the head of one of the pins made to screw on, so that when the right position is obtained, a slight turn of the head renders it immovable by clamping the inner tube. The photograph image may be moved in three directions; it may be placed across the top, bottom, or middle of the spectrum by rotating the cap, which carries the arm; and it may be rendered parallel to the spectrum by rotating the arm on its own axis-a clamp is furnished for fixing it in the latter position; thirdly, the image may be made to traverse the spectrum so as to get any division to correspond accurately with any band by a micrometer screw divided into 100ths of an inch, the head of the screw being also divided into 100ths gives a reading with a fixed index of 10,000ths. The scale reads in the same direction as the wave-lengths, that is, the highest numbers on the scale are found at that end of the spectrum which has the greatest number of wave-lengths.

For the comparison prism I have adopted one which has been in use a short time for astronomical purposes, that is, a rightangled prism with a slot ground in it; through this slot the light passes from the object on the stage of the microscope, whilst the portions of the prism on either side the slot transmit the light from the object on the comparison stage of the spectroscope. The effect of this is that the spectrum is divided into three parts, the upper and lower of which are the same, if the aperture in the comparison stage is covered by one object; but by a proper arrangement of two objects on the comparison stage, it is quite easy to display three different spectra at the same time; I do not, however, think that there is much practical advantage to be gained by so using three spectra. But I think in the case of two spectra but slightly different from each other, in the position of their lines, two comparisons are better than one. It is as though you made two obser

vations at the same time.

The face of the prism from which the reflection is obtained, is covered externally with a plate of very thin sheet copper, the surface of which is blackened. The comparison prism is moved into or withdrawn from position by the fitting to which it is attached, sliding in brass dovetail grooves; this, I think, is more steady than the usual method of mounting it on a cylindrical rod;

VOL. I.

2 A

there are four screws in the brass angle piece, connecting the prism with the slide, by which adjustment in two directions may be obtained; one renders the edge of the prism parallel with the slit, and the other renders the base of the prism parallel with the internal platform of the instrument.

Quite recently another method has been used in comparing spectra together, and I have within the last week made this the last addition to my instrument. Instead of having the spectra one above the other, in this arrangement they are actually superposed so that the whole of the field is occupied by each of the two, or by withdrawing the reflector a little, one half the field displays the spectrum from the stage of the microscope, and the other half this spectrum combined with that from the comparison stage.

This result is obtained in the most simple manner. To the end of the prism is cemented a small bit of glass with parallel surfaces in such a way that it forms, as it were, a continuation of the hypothenuse of the triangle of the comparison prism. It will be seen at once that it acts precisely in the same way as the glass in Dr. Beale's neutral tint reflector; some of the rays from the stage of the microscope pass through, and some of the rays from the comparison stage are reflected at right angles, and so join the path of the former.

The comparison stage is, I think, a much better form than the one in ordinary use; it consists of a flat plate with two hard white metal springs, beneath which the usual microscopic slide may be readily inserted, or in the same manner the tube carrier. This consists of a strip of brass, to which two curved springs are attached for holding in position the tubes in which liquids for spectroscopic examination are generally sealed.

To save the expense of a second spectroscope, I had an additional tube made with an adjustable slit to take the same direct vision prisms, the whole closing in a case for the pocket, which is all that is required for the preliminary examination of many objects, or to ascertain the presence of certain absorption lines in the solar spectrum caused by vapour in the atmosphere, which, according to Professor Piazzi Smyth,* is a far more delicate means of foretelling rain than the variations of the barometer.

I cannot conclude without expressing my indebtedness in these matters to Mr. Hilger for the readiness with which he carried out my suggestions, and for the practical hints he has given me; but, above all, for his kindness in placing his valuable collection of prisms at my disposal, and his permission to make use of his workshop and tools for any experiments I might wish to undertake.

*Astronomical Register,' Sept. 1877; 'Journal of the Scottish Meteorological Society,' vol. v. p. 84.

NOTES AND MEMORANDA.

The Gemmiparous and Fissiparous Reproduction of the Noctiluce (Noctiluca miliaris, Suriray).-I sum up in these few lines the following facts, not hitherto pointed out, or imperfectly known, which I shall elsewhere describe in detail with others already observed, in a memoir about to be published.

The disappearance of the tentacle, of the basilar tooth, of the flagellum, and of the infundibular furrow-like depression of the Noctiluca before their reproduction, has been noticed by Brightwell (1857), as well as by Cienkowski (1871). I have proved that this disappearance is constant, and not accidental before fission, and that it takes place by atrophy properly so called, and not by retraction of the tentacle into the interior of the body. I have besides been able to follow the phases of the obliteration of the buccal slit, as the precursory phenomenon of gemmation. Before fission, this obliteration does not take place. The flagellum and the tentacle only fall off.

The buccal obliteration brings the Noctiluca to the condition of a cell properly so called, closed on all sides, spherical, provided with a proper wall, represented by the envelope of the animal, and with wellknown sarcodic contents, with a nucleus without nucleolus, also spherical. But there is here nothing to be compared with the encystment preceding the reproduction of different infusoria (Euglena, &c.).

Far from disappearing before the formation of the gemmæ, as has been said and figured by Cienkowski, the nucleus of these unicellular adult animals of 3 mm. to 6 mm. diameter plays a direct and important part in the constitution of the contents of each gemma, as also does the yellowish substance of the cellular body which surrounds it; the cellular wall of the animal rises in a conoidal projection to form directly that of each gemma.

One individual with another, there are produced from 256 to 512 gemmæ, by gradual bisegmentation of the nucleus and of the cellular body, with a corresponding production of as many projections or gemmæ of the cell-wall, as fills one of the nucleo-cellular segments resulting from this bisegmentation. The total duration of these phenomena is from ten to twelve hours in a temperature of 12° to 18° in April and May.

The following are the phases of each division of the nucleus in this segmentation: it is lengthened into a cylinder, blunt at each end, and becomes very finely granulated, instead of remaining homogeneous. Immediately after, it becomes very finely striated longitudinally; the striæ are distinct, and evidently result from the juxtaposition of very thin colourless filaments, which compression shows to be soft and flexible. This fibrillar production, following the greater axis of the nucleus, is a constant fact in the fission of the nucleus of plants and animals, as Auerbach, Strasbürger, Bütschli, and E. Van Beneden have successively proved. About ten minutes later, the two extremities of the nucleus which had remained granulated, become

spherical, whilst continuing united to one another by the bundle or little band of fibrillæ just formed. These two extremities thus constitute two spherical nuclei, finely granulated, between which exists the little band of fibrille whose extremities remain in continuity with these new nuclei. Gradually the little band of fibrillæ becomes attenuated towards the middle of its length, as if it were drawn out, and folds up more or less on itself, in such a way as to bring the two nuclei nearer to each other. This attenuation soon leads to the rupture of the continuity of the fibrilla, of which each half then gradually withdraws into that one of the two new nuclei to which it had remained attached by one end. The bisegmentation is thus complete from an hour, to an hour and a half at most, after its commencement, presenting, in one individual with another, varieties of secondary importance as well as a few other peculiarities not pointed out here.

At the same time the layer of sarcodic substance (protoplasm), which is in immediate contact with the nucleus, is segmented without showing anything peculiar. But the whole of the sarcodic filaments, anastomosed into a network which is spread around the abovementioned layer, present some curious phases of segmentation. The extreme peripheric portion of this network is condensed into a thin yellowish layer or bordering, circumscribing that part of this reticulum which remains interposed between this homogeneous bordering and the equally homogeneous layer in contact with the nucleus. The whole contracts itself "en bissac" with a wrinkling of the surface, simulating torsion of the sarcodic substance, at the level of the constriction of the nucleus which precedes its division. The contraction of this "bissac" increases and culminates in a complete separation or segmentation of the sarcodic substance, which ends a few minutes after the completion of the nuclear fission. Afterwards, the substance forming a peripheral bordering to the reticulum, gradually approaches the perinuclear homologous layer (in consequence of the contraction of the intermediate network itself) until the latter disappears and the whole is blended into a yellowish, homogeneous, cellular body, with an undulating surface, lodged in a projection, corresponding to a gemma, of the wall of the body of the parent animal.

That is the case until the end of this double formation by simultaneous segmentation and gemmation, associated together in the gemmiparous reproduction of the Noctilucæ. The wall of each gemma and its cellular contents contract at their point of continuity with their homologues of the unicellular parent, and separate from them when the length of each is reduced to an average of 018 mm., that is to say, when their number is either 256 or 512. They are so many new unicellular individuals like the one which generates them (to die afterwards), and which from the time of their ulterior evolutionary growth, always remain unicellular. At least no evolutionary phase higher than the tentaculated form has up to the present time been observed.

Before the gemmæ are completely separated from the parent and swim freely, a flagellum six or seven times their length is developed on their plane surface (the other being rounded), nearer to their

blunt extremity (which is still adherent to the point of gemmation), than to the other end. This extremity is always foremost when the freed gemma swims upon its plane surface propelled by the undulations of the flagellum which trails behind it. One or two pulsatile vacuoles or vesicles of 004 mm. diameter are seen in the cellular body of each gemma between the plane surface and the nucleus. This vesicle is absent in the adult.

No observer has been able up to the present to follow the evolutions of Noctilucæ from the gemma state up to that of the adult individual. It is not yet known whether the flagellum of the gemma from 10 mm. to 12 mm. in length remains like the flagellum of the adult, which is only 06 mm. to 07 mm. long. The smallest Noctiluca which I have seen were of 15 mm. diameter, spherical, without mouth or infundibular depression, and with neither flagellum nor tentacle. I have followed the formation of the mouth at the level of the nucleated cellular body adherent to the internal face of their wall. It begins by a linear wrinkle of the latter, which is thickened a little on each side of this fold by the production of two or three mammillated projections. This thickening becomes gradually yellowish, and takes the form of the lips of the buccal slit in the adult. These phenomena last about three-quarters of an hour, after which the lips open a little from time to time. Then begins the formation of the infundibular depression and of the rectilinear dorsal fold, as well as that of the tentacle. That of the flagellum only takes place after the complete development of this latter organ.

Let us note that some Noctilucæ are found scarcely larger than the greater number of the others which are anatomically double, that is to say, provided with two cellular bodies, with two alternate buccal slits, each accompanied with corresponding tentacle and flagellum. It is probable that they proceed from some gemma in which the fission raising their number from 128 to 256, or from 256 to 512, has failed, whilst it took place in the others; a gemma which nevertheless has continued to develop like the others.

The fission of the Noctiluca has been pointed out by M. de Quatrefages (1850), and by Krohn (1852), who has seen that it begins. by the fission of the nucleus, since better studied by Brightwell. The fission of the nucleus presents the same phases as at the time of its first division in the cases of gemmiparity. The total division of the Noctiluca takes place in such a manner that, once complete, each of the two new individuals has a buccal slit, in which one of the lips has grown from one of those of the buccal slit of its generator. A tentacle is developed on the side of the mouth of each new Noctiluca from the end of the total fission, or immediately after the separation of the two new individuals. It is about an hour before the freed organ begins its movements.

In all cases, the production of the tentacle begins by the formation of a short prolongation of the yellowish substance of the cellular body which raises the tegument near one of the lips of the buccal slit, showing itself a little. Below this prolongation and in continuity with it, a second is raised, conoidal at first, and gradually taking the