;^.fo.25j> S. M-^O ^ ^ >■■'- ■■^•-^l^. 3CIENTIFIC MEMOIRS, SELECTED FROM THE TRANSACTIONS OF FOREIGN ACADEMIES OF SCIENCE, AND FROM FOREIGN JOURNALS, NATURAL HISTORY. EDITED BY ARTHUR HENFREY, F.R.S., F.L.S. &c., j ^ LECTURER ON BOTANY AT ST. GEORGe's HOSPITAL, AND THOMAS HENRY HUXLEY, F.R.S. / / LONDON: TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET, Printers and Publishers to the University of Ijondon. 1853. FLAMMAM. " Every translator ought to regard himself as a broker in the great intellectual traffic of the world, and to consider it his business to promote the barter of the pro- duce of mind. For, whatever people may say of the inadequacy of translation, it is, and must ever be, one of the most important and meritorious occupations in the great commerce of the human race." — Goethe, Kunst und Alterthum. CONTENTS. PART I, Page Art. I. — On the Circulation of Sap in Plants. By Dr. Hermann Hoffmann 1 Art. II. — Upon the Male of Argonauta Argo and the Hectocotyli. By Professor Heinrich Muller of Wiirzburg 52 Art. III. — A few Remarks upon Hectocotylus. By C. Th. von SiEBOLD, Professor at the University of Breslau 92 Art. IV. — Investigation of the Question: Does Cellulose form the basis of all Vegetable Membranes? By Hugo von MoHL 95 PART II. Art. IV. — Investigation of the Question : Does Cellulose form the basis of all Vegetable Membranes? By Hugo von MoHL {continued) 97 Art. V. — Memoir upon the Hectocotyli and the Males of certain Cephalopods. By MM. J. B. Verany and C. Vogt 119 Art. VI. — Organographical Observations on certain Epigynous Monocotyledons. By H. Cruger, of Trinidad 155 Art. VII. — Fragments relating to Philosophical Zoology. Se- lected from the Works of K. E. von Baer 1/6 IV CONTENTS. PART III. Page Art. VII. — Fragments relating to Philosophical Zoology. Se- lected from the Works of K. E. von Baer (continued) 193 Art. VIII. — On the Development of Zostera. By W. Hof- MEISTER 239 Art. IX. — On the Winding of Leaves. By M. Wichura .... 262 PART IV. Art. IX. — On the Winding of Leaves. By M. Wichura (con- tinued) 273 Art. X. — On the Development of the Asddians. By A. Krohn 312 Art. XI. — Observations on the Development of the Pectini- branchiata. By MM. Koren and Danielssen 330 TWELVE PLATES. SCIENTIFIC MEMOIRS. NATURAL HISTORY. Article I. On the Circulation of Sap in Plants. By Dr. Hermann Hoffmann. [From the Botanische Zeitung, vol. vi. p. 377, 1848; vol. viii. ^A7 et seq. ISfiO.] As, in animal physiology, many of the commonest vital phae- nomena were involved in obscurity previous to Harvey's great discovery of the circulation of the blood, so in the physiology of plants, not a few points are still inexplicable on account of our very meagre acquaintance with the nature of the circulation of the saps in vegetables. Every Manual of Botany gives a different view as to the anatomical system in which the saps are supposed to ascend or descend : many deny altogether the descent of the fluids, while others are so firmly convinced of the existence of this, that they have made it the basis of a peculiar kind of descending growth of roots, and stated that trees grow from above downwards, in- stead of upwards from below. We are likewise ignorant of the relations of the milk-sap to the other juices and to crude nutrient saps, so that it is still undecided whether it is to be SCIEN. M^U.—Nat. Hist. Vol. I. Part I. I Z HOFFMANN ON THE CIRCULATION regarded as a secretion, or as a circulating nutrient fluid analo- gous to blood. Even the function of the air-vessels is not clearly made out, or, otherwise, the recent statement, that the " so-called air-vessels " of the Ferns do not convey air, could not have been made. Under such circumstances, it can only be expected that the imperfect knowledge of the actual facts must leave the cause of the movement of the sap in utter uncertainty; and, indeed, on this very point the most wonderful notions are current, capillarity, contractiHty, and endosmose have to do duty in turn, as far as they will go. Where is the crude nutrient sap elaborated ? How^ does it arrive there ? By what path is it carried back to the other parts of the plant to furnish the material for development ? In what anatomical and physiological relation do the air-vessels stand to the system in which the saps circulate ? Injections cannot be applied in the endeavour to trace the paths by which fluids penetrate the vegetable tissues, on account of the minute size of the vessels, nor, indeed, without destruc- tion of the substance. But the spontaneous absorption of easily detected fluids accomplishes in plants what is done by injection in human anatomy. Coloured fluids, however, are rarely taken up by uninjured roots ; I have therefore made use of a very dilute solution of ferrocyanide of potassium, which may be readily detected in any spot to which it has penetrated, by the blue colour it assumes when chloride of iron is applied. And as the Prussian blue thus formed is insoluble in aqueous fluids, a little care in slicing and preparing the objects enables us to avoid the spreading pf the colour to unaffected parts, which would deceive the observer as to the boundaries within which the natural mo- tion of the sap takes place. In the first place, it is found that this fluid penetrates by a different path in uninjured roots, from that which it takes when the solution is caused to be absorbed directly by the cut surface of a cut plant ; in the next place, that the course which this fluid pursues is constantly the same, and peculiar to each plant; and, moreover, that by no means all cells and vessels take equal share therein, but that where, in general, vessels are met with, the sap enters first into these, and then, as in animals, passes far more slowly from these into all the remaining tissues of the plant. OF SAP IN PLANTS. 3 By this means it is possible to test, by direct experiment, the directions and course of the flowing saps from the root to the leaves, and backwards to all the other organs, which the follow- ing observations attempt to do in regard to the more important families of plants. I. ACOTYLEDONS. 1. Fungi. Clavaria rvgosa, Bull. — A number of these fungi (with unin- jured radical filaments imbedded in earth) were placed with the lowest part in an aqueous solution of ferrocyanide of potassium : in twenty-four hours they had become thoroughly imbued with it, so that a blue colour was produced when the tips were merely slightly touched with chloride of iron. When a cross section of the upper part of the fungus was made, and the surface of this tested, a blue colour was also pro- duced, but of very different intensity in different parts of the section. One part remained almost white, while the central portion, as w^ell as a zone within the layer of rind, was strongly coloured : the hymenial layer was tinged but very slightly, which appears to depend partly on the great density of the cellular tissue, offering more resistance to the passage of the juices, and partly on the similar eflfect of the horizontal position of its cells. The cells of the interior were of perfectly similar character, but they seemed to be more loosely packed under the cortical layer and in the central part, and the more active conduction of the fluid is connected with this. The walls of the cells, and also the fluid between them, were coloured blue. Longitudinal sections exhibited exactly the same results. Scaphophorum agaricoides, Ehrb. — The little branch to which the fungus was attached was dipped into the fluid, without wetting the fungus itself. In twenty-four hours the fluid had made its way to about the middle of the fungus ; when chlo- ride of iron was dropped on the surface, a blue colour w^as produced, and this arrested the further penetration of the ferro- cyanide of potassium toward the border, so that the latter did not exhibit any reaction even after two days. Although this fungus is composed of uniform elementary structures, the reaction of the different layers displayed very 1* 4 HOFFMANN ON THE CIRCULATION unequal intensity ; for while the sections of the lamellae were still white, an intense blue was sliown in the angles of their folds ; the fleshy substance in the vicinity was likewise white, while the same exhibited two strongly coloured layers farther up; the superficial layer was also coloured deep blue. Micro- scopical examination showed that the lower part of the cellular tissue which had remained white was more closely packed, with predominant transverse joints, whence the fluid could not pass so readily ; while the upper part, distinguished by its deep colour, facilitated the penetration by the laxity and the varied direction of its cells. Trametes suaveolens, Fries. — Several drops of the test fluid were poured upon the bark of the willow stem upon which the fungus grew, near the point of attachment, but without wetting the plant itself. In two days the fungus exhibited a tinge of blue, chiefly at the places where it was attached to the slice of bark, and consequently it must itself have contained a salt of the oxide of iron ; this slight blue colour was also visible on the out- side of uninjured fungi. On the application of chloride of iron to the surface of perpendicular sections, the colour became deeper near the base, and spread somewhat farther, especially in one point as far as the hymenial layer. Agaricus virgineus, Pers. — The patch of turf on which it grew was placed in the fluid, and within three hours the entire fungus was penetrated by the latter, which is not wonderful considering the great moisture of the plant. Stipe, pileus, flesh, and gills, were coloured deep blue in the reaction, the pileus darkest at the part where it is continuous with the hollow stipe, while the sections of the gills w-ere only weakly tinged, or did not become coloured at all ; but the passage of the blue-coloured into the uncoloured parts was quite gradual, and exhibited no distinctly marked boundaries. Examination by the microscope revealed a very lax interwoven cellular tissue ; in the lighter spots, cells were seen only half coloured blue, so that the fluid had not com- pletely penetrated them. From the foregoing it appears that the path of the circulation has no accurately fixed boundaries in the Fungi, and presents no anatomical peculiarities ; the fluid penetrates forwards and late- rally between and in the cells, proceeding most rapidly in those OF SAP IN PLANTS. 5 places ^^ here the laxity of the tissue and the direction of the cells oppose the smallest amount of resistance ; just as in blotting or other unsized paper. 2. Lichens. Cladonia subulata, Wallr. — The lumps of earth on which the plant grew were placed in the test fluid, and the whole was covered with a plate of g-ass, in order to maintain a moist atmo- sphere. Yet even after ten days the plant was not permeated throughout ; it had become coloured blue in some places, which indicated a natural existence of salts of oxide of iron in it. After the application of chloride of iron, moreover, to various sections, only a slightly deeper blue colour presented itself, which is readily explained by the extraordinary density of the tissue, and the consequent slow conduction of the juices in the lichen, which is itself of a rather dry nature. The blue colouring was uniform, and exhibited no marked limits. 3. Mosses, Syntrichia ruralis. — The moss and the soil supporting it were so placed in the fluid that the former remained unwetted ; the atmosphere was kept damp as above. The stem conveyed the fluid upwards ; the leav^es exhibited an uniform blue colouring at the base, which however spread very slowly over the surface, when the chloride of iron was dropped upon them ; the fluid advanced much more rapidly towards the point in a row of cells lying close beside the margin, while the mid-nerve acquired no perceptible colour. The chlorophyll granules did not seem to exert any especial influence on this conduction, since the lowest cells appeared coloured uniformly blue, in spite of their total absence. In the cells lying somewhat higher up, the blue colour did indeed correspond with the heaps of chlorophyll, which was perhaps only an optical phaenomenon : further up the cells were pure green. Barbula muralisy Timm. — Treated as above : half mature, like the foregoing. Here again a blue stripe of more delicate cells was seen along the margin up to the apex, after the application of the reagent, while no discoloration could be detected in the mid-nerve, and on the general surface of the parenchyma of the leaf only isolated spots occurred, appearing to indicate a very 6 HOPFMANX ON THE CIRCULATION imperfect conduction of the sap in these places. The fruit-stalk absorbed only a small portion of the fluid extremely slowly ; when cut across, and the cut surface dipped in the solution of ferrocyanide of potassium, it soon exhibited a very distinct re- action ; strongest below, where rind and pith were coloured ; weaker above, where only the pith vvas acted on (and this slightly). From this it appears that the passage of the sap goes on chiefly and most energetically in a peculiar layer of cells along the margins of the leaves, in these mosses, while the mid-nerve fulfils another function ; I could not detect air in the latter. The stem and the fruit- stalk conveyed fluid onward through all their tissues, but very slowly so long as they remained uninjured. When chloride of iron was applied to the uninjured rind, it pe- netrated extremely slowly, which is sufficiently explained by the solidity of the structure. Hypnum cupressiforme, L. — In this there are no peculiarly- formed marginal cells and no mid-nerve ; and no definite direction of the course of the sap could be detected here. This plant became gradually saturated with the fluid, so that even the peristome was coloured blue. 4. Ferns. Pteris serrulata, — The test fluid w^as dropped upon the mould in which the plant was rooted, without wetting or injuring the latter. In twenty-four hours it had already penetrated far up the petiole ; the reagent did not colour the brown internal layer or the vascular bundle in the central point ; the latter was white, surrounded by brown soft cells. The parenchyma was blue and rich in granules (chlorophyll and starch?) throughout. The scalariform vessels in the interior distinctly contained air, and when uninjured took up no test fluid, although the plant was in a most vigorous condition of growth. Poly podium crassifolium, — Treated like the preceding. In twenty-four hours the petiole was permeated by the fluid, and both the rather delicate cortical layer (of a green colour), containing no chlorophyll, and the entire parenchyma, which contained very numerous granules, were coloured blue by the reagent. The latter is traversed in the inner parts by isolated vascular OF SAP IN PLANTS. 7 bundles which are surrounded and supported by a hard brown layer of cells of a glass-like brittleness. These were not coloured blue ; neither were the vessels in their centre, which evidently contained air, and were to a slight extent unrollable. When this petiole was cut across and the open end dipped in the test fluid, the latter penetrated very rapidly upward into the white air-vessels, driving out the air. If the upper part of the leaf was dipped in the fluid, the phoenomena were the same as in the first case; under such circumstances nothing penetrated into the air-vessels, w^iich consequently are nowhere in communication with the surface (the stomates). Aspidium capense. — The phaenomena were exactly the same as in the preceding instance. Here again the stomates on the under side of the leaf conveyed no fluid into the air-vessels in the interior. Aspidium filix mas, Sw. — The results did not differ from the foregoing. The brown cells again were not coloured blue here. Hence the Ferns possess distinctly separate paths for the diffu- sion of air and sap. I could not observe any special path for the descending juices, and the existence of such may indeed be doubted, since these plants are supposed to grow only at the points. II, Monocotyledons. In the preceding section an attempt was made to prove that in the lower cellular plants, in accordance with their homogeneous structure, the fluids passing from the soil into the plants, took no fixed direction, but, soaking through from cell to cell, ad- vanced most rapidly wherever the laxity of the tissue opposed the minimum of resistance. In the Vascular Cryptogams, on the contrary, in the Ferns, it was found that special organs, the streaked vessels, already present themselves, exclusively des- tined to contain gaseous fluids, while the fluids absorbed from the earth first ascended within the looser cellular tissue in the vicinity of those vessels, and were from thence diffused through- out the remainder of the tissues of the vegetable ; not indeed without previously undergoing suitable elaboration and ame- lioration. In the Monocotyledonous plants, where the specialization of the anatomical systems becomes more distinctly marked, similar 8 HOFFMANN ON THE CIRCULATION results are met with, and it is especially seen here, that the function ordinarily attributed to the system of the spiral vessels and their allies, is devoid of all proof in fact, and has been de- duced from experiments in which sufficient regard was not paid to all the circumstances involved. At the conclusion of the third part of this essay, relating to the Dicotyledons, I shall examine these experiments more closely, and seek to demon- strate the causes which have so long restricted us to conjectures and opinions, in a matter which appears so simple. Since in the plants now to be considered, new^ organs, flowers, organs of fructification, and ovules, are added to those already examined ; since, moreover, a far greater variety becomes evident in the forms of the internal structure, and the physiological rela- tions of the roots, stem, leaves, &c., of the Monocotyledons, than exists in the majority of the Acotyledons ; the observation of the progress of the sap w ithin these different structures, pos- sessing peculiar forms in the diverse families, acquires greater importance, and admits of conclusions of a far more compre- hensive nature. I therefore take permission to enter more minutely into the details, in order to introduce the more extended remarks in fitting places. The experiments, as in the above-mentioned cases, were made, with the exception of Carina, on pot plants alone, which w ere treated in the ordinary manner. In order to discover the course of the sap, the earth was watered with a solution (always of pretty equal concentration) of ferrocyanide of potassium, and then, after this fluid had been absorbed by the uninjured roots, the parts where absorption had occurred were demonstrated in transverse and longitudinal sections of the plants, by means of sulphate of peroxide of iron. Anomatheca cruenta, Lindl. (Iridaceae). — Watered on the 16th of July ; in two hours the ferrocyanide of potassium could be detected in the bulbs, while it was sought in vain in the stems. In the petals also (even in those developed subsequently to the watering and investigated three weeks after), the sap could not be traced, although at first appearance it seemed to have entered them, since they contain a red colouring matter which is readily decomposed and changed into a blue colour; but a counter- experiment with pure water, with a careful apphcation of the OF SAP IX PLANTS. 9 reagent to the petals (stripped of some of the epidermis by shaving with a razor), showed the error of the first conjecture. Into the sepals of the specimen, on the contrary, the sap ascended within a few days ; the same occurred in some half-ripe fruits which existed upon other specimens. The following is a detailed account of the anatomical conditions. The bulb does not consist of separate scales, but forms a kind of tuber which is chiefly composed of a perfectly homogeneous mass abounding in starch ; the outer coat is a reticulated shell composed of two membranes : a thick bundle of vascular cords runs up the centre. The starchy parenchyma does not react blue, neither does the vascular mass in the centre, in the separate vessels of which air may be detected by the microscope ; but a very dark blue colour was seen in innumerable closely-crowded little points in the cortical layer or shell. And the presence of the ferrocyanide in the cortical substance did not depend upon a penetration from without during the watering of the specimen, and thus upon a penetration into the interior by an unusual path ; for the outer coating of all showed no reaction whatever with the salt of iron, while the blue colour was scarcely per- ceptible inside the finest of the root-fibrils. The shell of the bulb contains no air-vessels or spiral structures, but is traversed in all directions by numerous anastomosing liber bundles, which are sharply defined against the looser cellular substance surrounding them, and give the shell the open reticulated aspect already mentioned. These liber bundles exhibited no reaction ; in the transverse section they look like large yellowish circles, which are clearly contrasted against the surrounding blue cel- lular spots. The cells lying between these bundles are of two kinds ; those lying next them are narrower, more slender and more elongated than the rest, which exhibit a rounded four- or five-angled form ; it was these elongated cells which were coloured blue by the application of the salt of iron, while the remaining cells remained for the most part altogether uncoloured. The blue-coloured cells frequently lie several one behind the other, so that it is easy to trace the straight line in which the sap ad- vances, without deviating to the side ; in other cases, especially in the rest of the parenchyma, they frequently lie isolated, whence it appears to follow, that the motion of the sap is not 10 HOFFMANN ON THE CIRCULATION only less perfect in general here, but also takes place in an in'c- giilar manner, by no means in one and the same plane (parallel to the surface of the section). This tuberous bulb is, evidently, remarkably well fitted to afford a clear notion of the organs con- veying the sap in plants, since here all the important parts, the spiroids, the parenchyma, the elongated vessels, and the liber, are distinctly, and in fact — as in the Dicotyledons — widely sepa- rated from each other, to the great convenience of the observer. Underneath the bulb is the true root, which is of a pointed conical form. Here also the vascular bundles lie in the middle, while the blue colouring occurred in the parts more toward the circumference. In the stem no discoloration could be perceived, even in the cases where distinct reaction occurred, not only in the bulb below, but also in the capsule at the upper end. From this it seems as if the absorbed sap rested or became accu- mulated in particular parts of the plant, while it only hurried through the others as passages. In the cases where reaction was observed in the calyx, microscopical examination showed that the air-carrying, delicate unreliable spiral vessels or annular vessels were never the organs conveying the sap, but the elon- gated cells often lying close beside them. In specimens which were taken up nine days after watering, the absorbed ferrocyanide could be detected inside the half-ripe, otherwise fully-developed, capsules. It occurred within the elongated cells, which always accompanied the numerous delicate air-vessels in the axis, and seen through a lens after the reaction, formed a great abundance of little blue points and streaks. The soft, unripe seeds also exhi- bited distinct reaction in particular places ; here the fluid had also been conveyed from cell to cell through the seeds; no vessels of any kind exist in the seeds. Tigridia pavonia, Pers. (Iridaceae). — The plant was watered with the ferrocyanide while in full vegetation, on the 25th of July. After six days the reaction was observable not only in the bulb but also throughout the whole stem up to the floral organs, but not in these or in the interior of the germen. In nine days, however, the germen also had absorbed the salt, close beneath the surface, but neither the interior nor the ovules exhibited any reaction. The current had advanced more externally to the upper i)arts, to supply the floral organs with sap in the first OF SAP IX PLANTS. 11 instance, since a strong influx of sap and an active vegetation do not exist in the ovary and the ovules until they are fully developed. In the flower- bud of a blossom unfolded on the same day, the reaction could be detected in the petals, the stamens and in the pistil up to the stigma. — The root : this is placed underneath the bulb, and is of a swollen conical form. Internally it is homogeneous, and in the centre is found the bundle of striped vessels ; the blue colour was seen to a slight extent surrounding this, but much more marked at the peri- pherical layer. — Bulb : it consists of the bases of some six stems, enveloped by the sheaths at the bases of leaves ; the vascular bundles are scattered irregularly in the former; no reaction could be detected in the bases of the stems. Inside the bulb- scales the vascular bundles lie at some distance from each other in the median layer, while the reaction brought out a great number of blue spots, both in the mesophyllum and, more espe- cially, beneath the inner and outer epidermis of the bulb-scales. In the outermost coats of the bulb the blue colour was not limited to isolated points, but diffused uniformly throughout the entire mesophyllum, with the exception of the vessels. Micro- scopical examination of longitudinal sections showed that the vessels contained air and were uncoloured, while the cellular tissue and especially a portion of the elongated laxer cells were coloured deep blue. It is remarkable that only a portion and not the whole of the elongated vessels were concerned in the first conveyance of the sap. In longitudinal sections it was seen that these cells were much more easily separated from each other laterally than at their ends, by pressure, which seems not unim- portant in reference to the conveyance of the sap. — Stem : in a cross section half-way between two nodes, the reaction made visible a large number of blue points, both inside the cortical layer and in the layers of the central parenchyma. The cortical layer is composed of few, narrow, striped vessels, liber bundles with very long cells colourless as glass, and of elongated cells which were in part coloured blue ; the layer allied to liber found at the circumference of the central parenchymatous substance of the stem, which is very thick in the internodes, while it is far weaker at the nodes, did not become coloured. The vasciUar bundles are scattered in the internal substance and were not 12 HOFFMANN ON THE CIRCULATION coloured ; they are connected with the compact parenchyma- tous cellular tissue by a lax tissue of delicate prosenchymatous cells, forming an investment around them. When the prepa- ration was dried these were partially torn away, and the vascular bundles hung loose in the tubular cavities thus produced. In longitudinal sections the vascular bundles were seen most di- stinctly of a yellowish colour ; between and beside these the blue streaks and points which indicated the sap-bearing cellular tissue. — Flower : when the epidermis had been removed it was easy to detect the reaction here also, and it was seen uniformly through- out the parenchyma (containing colouring matter) of the petals, while the vessels and the delicate parenchymatous cells accom- panying them remained free. The stamens, conjoined into a tube, in the angles of which two vascular bundles are found, exhibited blue points both at their inner and outer boundaries, while the style lying free in the tube of the filaments was only spotted w ith blue in its outer periphery. The stigma w^as found very uniformly permeated with the salt. Dioscorea bulbifera, L. (Dioscoreaceae) . — Within seven days after the watering the salt had already ascended five feet up in the stem. — Root : upon reacting on a longitudinal section the w^hole tuber was found spotted with blue points and streaks ; no colouring occurred in the yellowish- white pieces of vascular bundles, which, composed of striped vessels containing air, were visible here and there, nor in a considerable number of reddish spots which microscopic examination proved to depend upon large drops of yellowish red oil occurring in many of the cells. The blue was seen inside the parenchyma of which the tuber is almost exclusively formed, in the same cells wiiich contained starch. In a cross section of the Stem are seen a number of small and also very large vessels not combined into bundles. The larger are in many places situated at pretty regular di- stances, forming something like a hexagonal figure in the paren- chyma of the stem, while the small are scattered quite irregularly, and often enveloped in a pretty thick coat of elongated cells. Inside the central parenchyma or pith occurred isolated blue points ; the broad woody layer on the contrary was wholly devoid both of these and also of vessels, while by far the greatest mass of the salt had ascended in the elongated cells of the liber and OP SAP IN PLANTS. 13 cortical layers, and had become more or less diffused from thence into the surrounding parenchyma of the rind. The large vessels of the stem just mentioned are of so great diameter that a hair may be easily passed more than an inch down them. When the stem was cut up, some of the fluid almost always escaped from the wounded sap-bearing cells of the circumference and pene- trated into the interior, as also in the garden balsam, where the size and transparency of the tracheae invite such experi- ments, giving an appearance as if these canals were not really air-vessels but sap-conduits. Disregarding the fact that, in my experiments, their walls and their rather fluid contents never exhibited a blue colour, the original presence of air in them may be easily directly demonstrated. For if a stem held the wrong end upwards is cut across and the w^ound immediately dipped in a drop of thick solution of gum, there is no difficulty in seeing bubbles of air emerge slowly from these tubes and inflate the dense fluid, when the stem is squeezed from the root upwards toward the w^ound. The ascent of bubbles of air begins directly, even when the pressure is commenced six inches below the wound, and it continues without interruption if the pressure is advanced slowly upwards to the cut surface. To conclude from the wholly accidental presence of water in these tubes, that they are nor- mally conductors of liquid, is like deducing from the presence of blood in the air-passages of a decapitated animal, that the lungs are naturally intended for that fluid. While the smaller vessels of the stem are usually spiral vessels, the structure of the larger is very well worthy of notice, for their walls look as if formed of small, nearly quadrangular frameworks, over which a delicate, closely spotted membrane appears stretched. The petiole, like the stem, is very rich in starch-bearing parenchyma ; it is coated by a rind, the partly elongated cells of which (but not the liber cells) become deep blue ; the internal tissue exhibited very few blue points, and the isolated spiral vessel bundles, in which air was clearly seen under the microscope, bore no share in the colouring. Rhapis flabeUiformis, Ait. (Palmaceae). — The plant, a specimen three feet high, was penetrated throughout all parts by the fluid in ten days after watering. — Stem : a transverse section, at least in the upper part of the plant, displayed this enclosed in the 14 HOFFMANN ON THE CIRCULATION sheath of a leaf (like the scale of a bulb), which was especially rich in sap and reacted strongly, while at the lower part of the stem this part was already dead and in a great measure decayed away. In this sheathing base of a leaf the elongated cells had chiefly absorbed the salt, while the liber bundles were not dis- coloured. The principal mass of the vessels (their walls present short streaks) lie in some degree concentrically, at equal distances from the centre and the circumference ; the vascular bundles, with their woody envelopes, contained air and were not disco- loured 5 but in the rounded, four-sided, starchy parenchyma cells between them, blue points were observed, which occurred in proportionately largest quantity in the vicinity of the central point and of the periphery. In a longitudinal section it was seen that the reacting cells exhibited scattered points or little streaks, and never lines continued in one plane. The vascular bundles run pretty nearly parallel in the stem ; even the nodes exert no influence upon the arrangement of the internal parts lying opposite to them. When the stem was cut through higher up, where it begins to lose itself in the leaves, a different appearance was seen. Yet microscopical examination showed that here again the salt had not ascended inside the vessels, but only in their immediate neighbourhood, inside the elongated cells by which they are accompanied. The upper leafy, as well as the peripherical, parts of this plant reacted far more strongly than the interior of the stem. Commelina ccelestis, Willd. (Commelineae). — This plant was watered on the I7th of July; the examination was made on the 24th and 26th of July, and the 9th and 15th of August, but in all cases the salt was sought in vain in the parts of the stem situated between the nodes, while in the nodes themselves, from the lowest to the uppermost, the reaction could be easily de- tected. From this it appears, taking the other experiments into account, that we must conclude the sap to rest for a time in particular parts of plants, as in tubers, nodes, buds and ovaries, while it hurries rapidly through others, especially the internodes of the stem. The cross section of the nodes of the stem exhibits a distinct separation into pith and rind. The bundles of spiral vessels, which displayed bubbles of air in their interior when ex- amined with the microscope, exist both in the pith and the OP SAP IN PLANTS. 15 cortical layer ; in the internodial parts of the stem the cortical layer is contracted further inwards, so that there is not so sharp a line of demarcation there. The points displaying the blue reaction were among isolated roundish cells which occur inter- nall}'^ and externally in the neighbourhood of the vascular bundles. In longitudinal sections the blue points appeared in an irregular band passing transversely through the node. In buds examined four weeks after the watering (they had been developed during this period), blue colouring could be detected inside the elongated cells of the calyx and the stigma, and inside the minute parenchymatous cells of the (still white) petals and of the connective, while the dehcate spiral vessels, even in the youngest, still imperfect organs, never displayed any colouring. It is worthy of notice that the epidermis of the petals opposes such a resistance to the penetration of the salt of iron, that no reaction occurs even when the entire petal is dipped in sulphate of iron ; hence it is necessary to tear the epidermis carefully away to admit the sulphate of iron to the parenchyma when we wish to bring out the reaction. Commelina pubescens, — The presence of the ferrocyanide could be shown in the cells of the stem and leaves five days after the watering. Commelina clandestina, Mart. — Here also the absorption could be certainly demonstrated after a short period. Commelina angicstifolia, Michx. — Detected certainly after four days. Commelina tuberosa, — The absorption demonstrable in two days, both inside longish cells and in isolated large cells of the central parenchyma. Gladiolus psiitacinns, Hook. (Iridaceae). — Here again it was easy to trace the salt many inches up the stem, and above all in the elongated cells, in which it was very uniformly diffused, and which are in part situated close to the air-bearing spiral and annular vessels, and in part at a distance from these and near the inner and outer epidermis of the separate leafy sheaths, of which the stem is almost entirely composed. The inner and softer leaf-sheaths of the stem reacted most strongly and conse- quently appeared to convey the sap most actively. These plants must contain a salt of peroxide of iron as a natural constituent, 16 HOFFMANN ON THE CIRCULATION for in four weeks after the watering, the leaves spontaneously exhibited a deep blue colour at their tips. The plants all died in little more than a month, apparently in consequence of the watering with the ferrocyanide, since they were placed in con- ditions otherwise favourable. Tritonia fenestrata, Ker. (Iridaceae). — Within two days after the watering, the salt could be detected in the bulb and in the lower part of the tuft of leaves. The lower part of the bulbous mass (which is the bulb formed the year before, or real tuber) was coloured very deep blue throughout; in the upper bulb (of the current year) it was observed that the parts giving a blue reaction were little scattered points and streaks, which were universally distributed. The microscope showed that those lying in the central part were the elongated cells which sur- round the air-bearing vascular bundles ; in the parenchymatous part of the tuber the blue colour was deepest in the starch- filled, irregular parenchymatous cells of which this structure is composed. The circumstance that in all cases, and especially distinctly here in the longitudinal section, the blue-coloured cells were mostly scattered and only exceptionally formed unin- terruptedly continuous lines, proves that the passage of the sap does not take place uniformly through all parts indifferently, but peculiarly through certain cells, which do not lie all in the same place (parallel to the surface of the sections), but more or less scattered, causing a ramification of the currents of sap often of a very irregular character. Allium neapolitanum, Cyr. (Liliaceae). — Not even four weeks after the watering could the presence of the ferrocyanide be de- tected either in the interior of the bulb or in the stem. On the other hand, the outermost dead scale of the bulb was coloured of an uniform blue tint, evidently by accidental imbi- bition of the fluid in contact with it, not only in the cells, but both in the reticularly striped and the true unrollable spiral ves- sels, which, moreover, still here and there displayed air in their interior. Canna indica (Cannese). — This plant was growing in the open air and was watered on the l7th of July. In seven days the blue reaction could be readily detected in the stem. This exhibits somewhat the same structure as the Palms in the cross section OF SAP IN PLANTS. 1? (vide Rhapis flabelliformis above), except that the vascular bundles of the true axis are very irregularly scattered in the interior, while they are arranged in a tolerably even radial order towards the circumference. In the main axis itself the salt could not be demonstrated with sufficient certainty at this time, but with the greatest ease in the sheath (the base of a leaf) which tirmly enveloped the whole main axis. Here the blue points were found in greatest abundance close beneath the outer epi- dermis, but also in large numbers immediately surrounding the liber bundles (enclosing unroUable spiral vessels). The micro- scope proved that the blue colour existed exclusively in uncom- monly continuously stained cellules (containing large nuclei) arranged in very straight longitudinal rows close beside the vas- cular bundles. The arrangement of the spiral vessels, liber- bundles and sap-conduits in the leaves was similar to that in the stem. Panicum plicatum, Lam. (Graminaceae).— Had not absorbed the least in four weeks ; the same happened in Ruscus aculeatus and Ananassa saliva. Arum divaricatum, L. (Aroidea^). — These plants had, for the most part, absorbed nothing at all even after four months ; in the few cases in which a weak reaction could be detected in the tuber, the blue colour was found in the large parenchymatous cells of the general substance, which contained the unusually thick starch-granules, of beautiful pyramidal and polyhedral forms with rounded faces ; while the scattered yellowish bundles of partly round, partly angular annular vessels were never dis- coloured. Tradescantia discolor, L'Herit. (Commelinaceae). — The reac- tion was visible in the stem in four days after the watering. The investigation is most difficult in the upper leafy parts, although they are coloured deepest, because immediately the cortical layer is cut across a great quantity of thick gummy liquid exudes, which covers the whole surface of the section and con- founds all the previously distinct fluids together. But when the stem is cut lower down, a few inches from the ground, the in- terior is found far less rich in sap, only a little mucilage exudes, and it is more clearly seen when the absorption of the salt has SCIEN. MEM.— iVrt^ Wst. Vol. I. Part I. 2 18 HOFFMANN ON THE CIRCULATION occuri'ed. A cross section at this point presents the following appearance. At the outside a cortical layer, without vessels, the starchy parenchyma of which is separated from the large central parenchyma of the stem by a kind of liber-layer : the central parenchyma, forming the great body of the stem, con- tains scattered vessels divided into two sets, a central and a peripherical, between which lies a ring of parenchyma free from vessels. The cells which became blue lie in the vicinity of the vascular bundles ; the gummy cortical layer was devoid of them. It is worthy of notice that both the boundary between the liber and parenchyma, properly the outermost layer of the central cellular mass, and the immediate environs of the vessels scat- tered in the central mass, emitted a small quantity of a tenacious white sap, in which were suspended a few starch-granules and an extraordinary quantity of minute raphides. No ferrocyanide could be detected in the cortical layer even after the space of a month. Chlorophytum StembergianmHy Steudl. (Liliaceae). — In this plant, which appears to derive the greatest part of the requisite moisture from the atmosphere, it is difficult to diffuse a quantity of the salt sufficient for the reaction in the ordinary way. The slender filiform stem to which the tuft of leaves with their aerial rhizomes and roots is attached, died in a few weeks, in conse- quence of the watering, without much of the liquid having been absorbed. In this case the cells of the bases of the leaves, as also the aerial roots and the rhizomatous tubers, exhibited a weak blue colour when cut surfaces were wetted with sulphate of peroxide of iron. A clearer view may be obtained in a differ- ent way. The aerial roots readily absorb fluids, for instance water, when they are dipped into it ; moreover, sulphate of iron was decomposed, the oxide fell as a yellow powder to the bot- tom of the glass, while the w^ater penetrated into the plant; on the other hand, a large aerial root which was already buried in the earth and had taken root there, readily absorbed so much ferrocyanide that this could be easily detected in longitudinal and transverse sections several inches above the ground. The central layer containing the dotted vascular bundles was found free from colour ; but all the rest of the tissue, composed of OF SAP IN' PLANTS, 19 elongated cells, even close in the vicinity of the former, had ab- sorbed a considerable quantity of the salt, and exhibited innu- merable blue dots and streaks in the longitudinal section. Caladium viviparum, Roxb. (Aroideae). — The fluid could be detected in all parts of the tuber and petiole twelve days after the watering. Minute examination showed the root to be a true tuber, which in three weeks after the watering was found soft- ened, collapsed, and, in short, dead (probably in consequence of the watering) ; on reaction it turned uniformly blue all through, just as a sponge would do under similar circumstances. On this tuber is seated the bulb-like base of the stem, which en- closes within it the true terminal bud of the axis. Both the base of the stem and the central bud exhibited distinct reaction in the same way ; in longitudinal and cross sections they became covered with countless blue points, which were recognized under the microscope as elongated cells. A leaf-bud which had just been developed from the tuber had likewise absorbed the salt, especially in its central region. The investigation is difficult in the fully developed leaf-stalk, since this so abounds in sap, that the fluid exuding when it is cut across readily spoils the expe- riment with the reagent. The blue sap is found in the elon- gated cells which surround the numerous round liber- bundles of the thin peripherical or cortical layer, the latter enclose isolated delicate unroUable spiral vessels*. The whole of the interior of the leaf-stalk is composed of parenchymatous cells, inside which run not only ordinary liber-like cells with vessels, but also several canals, so large that a hair may be readily introduced into them. These canals, still filled with fluid in the bud above mentioned, contain air in the fully developed petiole ; when the latter is squeezed even six inches from a cross-cut surface, air- bubbles are forced out of them ; just as was described above of the Dioscorea. The walls of these air-cavities are composed of large cells ; these tubes therefore form a transition to those of Dioscorea^ which exhibit a great affinity to the ordinary pitted * The numerous thick- walled raphides-cells occurring here slowly expelled their contents when accidentally injured, and these spread out like a tuft of feathers, displaying a peculiar backward movement not unlike that of the Oscillatorieae. 2* 20 HOFFMANN ON THE CIRCULATION wood-vessels. In the interior of the petiole the absorption of the sap was indicated only by a few spots becoming blue. Aloe picta, D.C. (Liliaceae). — Not even after five weeks, any more than at several shorter periods, could I detect the slightest trace of the salt in the roots, buds, stem or leaves of these plants, which were perfectly healthy, and which scarcely betrayed any injury even at the point of the main root in consequence of the watering. It is true the application of the salt of iron to a cross section discoloured the yellow fluid (the aloe-bitter) which ex- uded, transparent, from the large, very much elongated cells near the spiral vessels at the whole periphery of the leaf, but the stain was brownish black, not blue ; a reaction which occurred also with Aloes of this species never watered with the ferro- cyanide, and which therefore must be attributed to tannic acid. The absence of absorption after the stated period is the less re- markable, since these plants remain for seven months of every year in a dry part of the conservatory, without ever being watered, therefore without requiring any other fluid than that which they receive through evaporation when other plants are watered. Zephyranthes grandiflora, Lindley (Amaryllidaceae). — Three days after the watering no reaction could be detected in the bulb, petals, or ovules ; four days later, however, the bulbs ex- hibited distinct blue spots both in their outer and also in their inner scales, but the yellowish vascular bundles were not stained. The different lamellae of the bulb consist here of two epidermal surfaces, of parenchyma-cells full of starch, of elongated cells which are very soft and loose, and especially contained the blue fluid, and next to these, in the median line, of imroUable spiral and annular vessels, surrounded by a few delicate prosenchy- matous cells, as is usual in herbaceous plants. At this period no salt had penetrated into the stem, not even into the lowest part next the bulb. The plants in the same pot all died within eleven days. Cyperus monandms, Roth (Cyperaceae). — Nine days after the watering the ferrocyanide was found in the stem and branches. The rhizome, which was dug up after thirteen days, was, like the whole plant, perfectly stiff and healthy, as was the case with OF SAP IN PLANTS. 21 almost all of the plants mentioned above, so far as the researches here described go to show. The transverse section showed that the salt had perfectly penetrated this, with the exception of the yellowish vascular bundles which run in the thick parenchy- matous cylinder ; otherwise a great quantity of dark blue cells were found in all parts among the uniformly light blue mass of cells. In the lower parts of the stem the reaction was only weak ; it was more distinct in the peripherical part (no rind ex- isted here) further up ; it was strongest at the apex, where the separate flowering branches go off, for here not only the periphery but also the interior of the stem was closely sown with blue spots (streaks in the longitudinal section), which under the mi- croscope were seen to be elongated cells : the angular, dotted, and small streaked vessels which were situated partly near the periphery and partly deep in the parenchyma of the stem, con- tained air and had no share of the colouring. III. Dicotyledons. In these plants we meet with a phagnomenon which is not observed in those divisions of the vegetable kingdom above in- vestigated, at all events not in Germany, namely, the ^* bleed- ing,^^ or "tears,^^ which occur on wounds of many of the shrubs and trees of this division in early spring. This phaenomenon is so peculiar that it will be advantageous to examine it separately. In the following experiments I inves- tigated, 1st, the course of the spring sap ; and 2ndly, that of the summer sap, independently. 1. The Spring Sap, 1. On the 27th of February 1850, a root A, of ^ an inch dia- meter, coming from the west side of a white birch 6 inches in diameter, was laid bare, cut across, and the upper cut end placed in a cylindrical glass, of about 1 cubic inch capacity, filled with a solution of ferrocyanide of potassium. This was done at four o'clock in the afternoon. By the next morning the fluid was absorbed as far as the root reached. On the 27th also, several holes of 1| line diameter were bored i an inch deep, and into these introduced and cemented quills 22 HOFFMANN ON THE CIRCULATION of the same section, with the inner ends cut otF obliquely and the orifice looking upwards. These borings were made, — B, 3 feet above the ground, on the west side. C, 3^ feet above the ground, on the south side. D, 5 feet high, on the south side. On the 7th of March a clear fluid was first found in a vessel attached below the quill (B), perhaps caused by rain which had fallen on the 6th, and had run down from the stem. On the 8th of March the water (exuded from the stem) trickled from B and C ; D remained dry. The fluid gave neutral reaction with litmus paper. D did not become moist until the 11th, M hile B and C poured out abundance. The two last were closed and cemented with resin on the 12th ; D was bored out 1 inch deep, and a new hole was made of the same depth, — E, 2| feet from the ground, on the south side. The sap flowed immediately from both ; E delivered /O drops in five minutes, D only 15. On the 13th new traces flowed from D; E gave 19 drops in five minutes. Both holes were stopped and two new ones bored, — F, 6 feet high, north side. G, 1 foot high, same side. F gave only traces, while G gave 32 drops in five minutes. On the 14th F gave nothing, G 10 drops in five minutes. On the 15th F was dry ; G gave 45 drops in five minutes. On the 14th a new hole was bored, — I, 8 feet high, on the north side, which gave off no fluid on the 14th or 15th. All these experiments were made at noon. From the foregoing, the exudation and decurrence of the sap occurred observably earlier in the lower than in the upper part of the stem ; it was caused by a warmer temperature. All these fluids were tested for the ferrocyanide with sulphate or acetate of peroxide of iron, without results. It is evident that none of the holes entered the current of saj) passing up from the root A {vide infra). On the 10th of March the ascending root A was cut off and examined. The section 5 inches above the point where it dipped in the fluid, reacted strongly with salt of iron and excess of hydrochloric acid, and the same occurred with the other parts j the ferrocyanide had even descended into the lateral branches of the root. The longitu- OF SAP liN PLANTS. 23 dinal section showed that the liber had absorbed strongly ; and light blue, not sharply defined streaks, corresponding to the course of the tracheae, were seen in all parts of the wood ; the thin pith remained unaffected. On the whole the reaction was weak, because the great mass of the fluid had already ascended into the higher parts of the plant. 2. A young sycamore stem {Acer plat anoides), 3 inches in dia- meter, was watered with the ferrocyanide in the ground ; but this could not be detected in a small quantity of fluid which ex- uded from boring. It appeared to have been wholly decomposed by the (ferruginous) soil, since this had acquired a blue colour. 3. The soil at the base of a young sycamore of 1|^ inch dia- meter was watered, on the 26th of February, with a weak solu- tion of the ferrocyanide. On the 28th holes were bored into the stem, but the drops which exuded did not react with salt of iron ; neither did the ^ of a drachm of fluid that was found on the 1st of March. No more exuded after this. The soil around was coloured blue, and when the stem was cut off on the 10th, no ferrocyanide could ^be detected in it. The small efflux of sap in these two young trees, 2 and 3, was remarkable. 4. The earth round the trunk of a sycamore, 4 inches in dia- meter and 35 feet high, was watered with solution of the ferro- cyanide on the 2nd of March ; a hole. A, was then bored in the stem 1 foot above the ground ; the vessel attached to this was rapidly filled with fluid. Tolerable quantities of fluid flowed out, and in part down, continuously during the following days, but on the 7th of March no ferrocyanide could be discovered in it. On this day the soil round the tree was watered a second time with one ounce of a more concentrated solution ; the orifice con- tinued to pour out freely. On the 9th the soil was watered a third time, and a new hole was made (on the north side) at the same height as the old one (on the south side). On the 10th of March the fluid which had trickled from A gave a dark blue precipitate with sulphate of iron, which was not the case with B. On the 11th no reaction on either side. On the 12th no more had exuded ; on the 14th therefore the tree was cut ofl" at the base, and three pieces of it separated and examined, from the base, the middle, and the summit. The two lowest pieces 24 HOFFMANN ON THE CIRCULATION only gave a reaction, which was moreover very weak ; the trans- verse section displayed a few light blue bleared spots in the wood ; the liber and bark were unaffected. Longitudinal sections, on the contrary, displayed at particular spots, very sparingly, sharply defined blue points or even streaks (some of them in the direction of the medullary rays) ; microscopic examination showed that the few places in which some of the absorbed sap still remained, were tracheae (spiroids) and the neighbouring prosenchyma of the wood. In the middle fragment, 10 feet above the ground, the cross section gave the same result as before ; longitudinal sections led to none. 5. On a young sycamore, with a trunk 4^ inches in diameter, a branch A, arising 7 feet above the ground on the north side, was bent down to a certain degree and fixed in this position, the outer part cut off, and the point of the portion remaining attached to the stem, dipped 4 inches deep into a cylindrical glass full of solution of the ferrocyanide. A hole B was bored, 1 inch deep, 1 foot above the ground on the south side of the trunk. This was done on the 2nd of March. On the. 3rd the branch A had absorbed all the fluid, as far as it dipped iir it ; the glass attached under B was filled with watery fluid, which did not react with salt of iron. On the 5th the glass at A was refilled. On the 4th the vessel at B was only half-filled; on the 5th again quite filled ; rain which fell on the 4th seemed to have increased the flow of the sap. By the 7th (warmer weather than the 6th) the glass A had been again emptied, and was filled anew. On the 9th again empty, but up to this day, when the efflux at B ceased, no ferrocyanide could ever be detected in the fluid poured out. On this day therefore a new hole C was bored at the same height as B, on the north side of the trunk, corresponding to the absorbing branch A ; and in spite of B having stopped, fluid was poured out freely at C, which on the 10th reacted deep blue with the iron salt; the glass A was emptied down to a quarter of its contents (it held altogether about 1 cubic inch). On the 11th the freshly effused fluid ex- liibited a very strong reaction. The here evidently strong suction of a cut branchy or of a cut root, as in Experiment 1, thus both in the ascending and de- scending directions, results from the pressure of the atmosphere. OF SAP IN PLANTS. 25 The larse surface of the tree with its hundreds of shoots causes an evaporation even before the buds svi^ell ; the vacuum resulting from this, causes any fluid which may be brought into free com- munication with the interior of the tree, at any spot, to be driven in by the vis a tergo. The power of suction thus produced increases in equal proportion as the surface of the shoots is in- creased by the presence of leaves, and the capacity of the air to retain moisture by the heat of summer, as Hales (Vegetable Statics) and, more recently, Dassen (Wiegmann's Archiv, xiii. 2. p. 311) have proved; under the most favourable circum- stances it sustains a column of mercury 12 inches high. It will not be thought strange that the height of the column of mercury did not attain 18 inches, as has been observed in curved glass tubes under similar circumstances. The bark of the shoots of a tree does not form a solid envelope like the glass, for even with a smaller weight of the column of mercury, the air itself pene- trates through the bark into the interior, and thus puts an end to all further ascent. Hence the loose-barked vine absorbs but weakly; the denser Prunus domestica raised a column to ^th of a Flemish ell (6f inches) ; Betula nana, 0*240 (13J inches) (Das- sen). Hales observed the column rise to 4 inches in the vine, and 12 inches on an apple branch (equal to 13 feet 8 inches of water) . It is remarkable how accurately the fluids ascending or de- scending in this way are retained in that side of the stem which corresponds to the absorbing branch (or root) ; this is explicable by the straight, uninterrupted, little-branched course of the tracheae ; the anatomical examination of the preceding cases also proved this directly. The absorbing branch was cut off on the 11th of March and anatomized, with the following results. In a cross section 1 line above the lower end, which had been dipped in the fluid, all parts, bark, liber, wood and pith, reacted deep blue ; but even at a distance of 2 lines the pith was found uncoloured ! Four inches further up, that is, close above the upper boundary of that portion of the branch w^hich had been immersed in the fluid, the heart-wood in the vicinity of the pith was also found uncoloured, as was the case everywhere above this. From this point the pitted and striped trachea of the softer external por- tion of the wood, with the liber, had taken on the conduction of 26 HOFFMANN ON THE CIRCULATION the fluid, and on reaction formed sharply defined dark blue lines, which could be easily traced even in the stem, both up- wards and downwards, more than 1 foot from the absorbing. branch. Microscopic investigation showed that the blue colour had only spread half or the whole width of a row of cells, around the tracheae. A longitudinal section of the (upper) axil of the branch indicated the course of the trachese by curvilinearly ar- ranged blue spots and streaks, and not by continuous blue lines. Consequently these tracheae do not run so perfectly on one and the same plane here, that their entire course can be at once dis- played by a single section. It is worth notice, that only that side of the stem on which the branch was seated was coloured blue, the opposite not at all. That the pith does not conduct, is comprehensible, since the air it contains could not readily be displaced by w^ater pene- trating, and each of these cells containing air is a completely closed sac. But it is difficult to understand why the medul- lary sheath and the heart-wood do not conduct sap, since here communicating unrollable spiral and striped tracheae are present in abundance. This non-conducting layer amounted, at the point of attachment of the branch, to a quarter of the woody layer (the entire diameter of the stem at this point amounted to 1:J: inch). It is certain that the air visibly contained in these tracheae is no hindrance, for it is none in those of the outer, younger wood. The elongated cells of the medullary sheath are far narrower and closer than the prosenchymatous cells of the w^ood; they are crowded with starch- granules, w^hich are absent from the rest of the prosenchyma of the wood up to the medullary rays. Can it be some starchy or gummy mucilage which stops the progress of the fluid ? I believe not ; there is no dissolved starch present ; iodine distinctly reveals the unaltered starch-granules. And as to gum, it is not evident how this could close up the tracheae from the cells, since the microscope shows the former to be filled wdth air-bubbles. 6. Repetition of the preceding experiment. — On a sycamore trunk 3 inches in diameter, a branch A, at a height of 5 feet, was cut off, bent down, and fastened : the wound dipped in a glass of 1 cubic inch capacity, filled with concentrated solution of ferrocyanide of potassium. At the side (the south) of the OF SAP IN PLANTS. 2? trunk a hole B was bored at 1 foot above the ground, and a glass attached to catch the exuding fluid. On the following day (March the 5th) the vessel at A was almost empty and was refilled. In the glass at B there was no fluid ; but on March the 6th ^th of a cubic inch, which reacted deep blue with sulphate of iron. On the 6th the glass A was only half empty, but was refilled. On March the 7th, at 2 o'clock p.m. (as usual), A was half empty; nothing had flowed from B, but the orifice reacted deep blue with salt of iron. A new hole C was bored ^rd of a foot above the ground, on the east side, but no fluid exuded. On the 8th of March A was half empty ; nothing had flowed from C. On March 9th A was quite empty ; C had given off nothing. The tree w^as cut down on the 9th of March in order to trace the course by which the fluid had descended from the absorbing branch to the orifice. The liber and albumen reacted blue, but not the bark, heart-wood or pith. In the albumen the colouring w^as streaky, and the microscope showed that it was in the tracheae : these contained blue fluid with a few air-bubbles. The liber was not coloured nearly so far as the woody portion ; at 3 feet down the stem from the absorbing branch (on the same side), the wood reacted very distinctly, but not the liber. From this it follows that capacity of conducting is much inferior in the liber-layer, to what it is in the tracheae of the wood. Distinct reaction in blue streaks could also be observed in the trunk more than 2^ feet above the branch, on the same side (but not on the other) ; here the liber w^as affected, although not so strongly. A transverse section showed that about |^rd of the circumference of the trunk below the branch, |th above it, had taken part in the diffusion of the sap : the base of the branch occupied about J-th of the circumference. Dissection of the point of insertion of the branch on the stem gave exactly the same results as in No. 5. Examination of the wood in the vicinity of the orifice, and 2 inches lower down by radial longitudinal sec- tions, show'cd that the trachece had conveyed down the solution : the cellular tissue and medullary rays did not react. It is evident, therefore, that both in normal absorption by uninjured roots, and in abnormal by wounds, the conduction of the spring sap occurs chiefly through the trachea and liber ; a 28 HOFFMANN ON THE CIRCULATION result which stands in contradiction with the experiments already related on Monocotyledons and Acotyledons at the time of the full activity of the leaves {vide infra). Comparison of the Experiments 5 and 6 with 4, shows how- much more quickly the fluid runs through the tissues when it is placed in contact w ith open wounds, than when the absorption is allowed to take place in the normal way by the root. In 5 and 6 the ferrocyanide traversed some 8 feet, down from the cut branch to the orifice, in one day; in 4 it occupied seven days for the salt with which the soil (root) was watered to ascend to an orifice 1 foot above the ground. In the former there were open communicating tubes, in the latter closed membranes only permeable through endosmose. 7. In a young sycamore of about 3 inches diameter, a root on the west side, A, ^ an inch in diameter, was laid bare, cirt through at the distance of a foot from the trunk, and the upper end placed in a cylindrical glass of 1 cubic inch capacity, containing solution of ferrocyanide. An orifice B w-as made at a height of 1 foot above the soil, on the south side of the trunk, and a glass vessel fixed under the quill introduced into it. On the following day (March 5th) the vessel at A was found empty, and was refilled ; B was half full of fluid (^ a cubic inch). On March 6th, A, again empty, was again refilled ; B contained ^rd of a cubic inch of fluid. On the 7th, A as the day before ; nothing emitted from B. On the 9th, A empty. On the 10th, A was refilled ; B contained fluid again {^ a cubic inch), evidently in consequence of warm weather the day before. None of the exuded fluid gave reaction with salt of iron. On the 11th, nothing emitted from B. A new hole C w^as bored on the west side, corresponding to the root A, 6 inches above the ground, but nothing flowed from it up to the LSth. Although, there- fore, the fluid had been absorbed, it was impossible this time to draw it from the stem here by tapping ; evidently because the orifice B, which alone emitted fluid, was not made on the cor- responding side of the stem. The observations were made at 2 P.M. each day. On the 13th of March the absorbing root was cut off" and dissected. Reaction with sulphate of iron and hydrochloric acid showed that the exceedingly small pith had not conducted. At OF SAP IN PLANTS. 29 a distance of about 10 inches from the immersed lower end (the root dipped in about 3 inches), the longitudinal section gave a blue reaction throughout the woody substance ; the microscope showed liber, tracheae and wood-cells stained ; the tracheae most strongly, both in the inner and outer parts of the wood. The solution had not only ascended, but also passed down as much as 2 inches in small lateral branches of this root. At these points the liber no longer took part in the conveyance of the fluid, while many (especially the central) tracheae reacted deep blue. It is worth notice that the (ferruginous) soil in the vicinity of the entire root acquired a blue colour from the 6th of March onward ; perhaps by secretion through the peripherical parts of the root, perhaps through the injured points of the little radical fibres. 8. A sycamore trunk of } of a foot diameter was bored with similar holes, 1 inch deep, in two places, and all the exuding fluid was caught. A, orifice 1 foot from the ground, south side. B, orifice 3 feet from the ground, south side, 1 inch further westward. Experiment commenced at 3 o'clock on the 5th of March. Five o'clock : A had emitted 1 Paris cubic inch of clear watery fluid ; B i a cubic inch. March 6th, 8 a.m.: A and B had secreted nothing (cold night, below freezing point). 2 30 P.M. : A 9} cubic inches; B 12| cubic inches. 5 30 P.M. : A and B had each given off^th of a cubic inch. March 7th, 8 a.m. : the flask at A and B empty. 2 P.M. : A |ths of a cubic inch; B J^th of a cubic inch. 5 30 P.M. : the flasks at A and B empty. ' March 8th, the same. According to the foregoing, night limits the force of the cur- rent (by a lower temperature). When we reflect on the unequal duration of the bleeding in the trees 2 — 8, w^hich all stood near together, it becomes a question what was the cause of this phaenomenon ? The varying thickness of the trunk does not seem to explain the matter. No. 4 was 4 inches in diameter and bled ten days ; No. 2 was 3 inches in diameter and bled five days ; No. 8 was 30 HOFFMANN ON THE CIRCULATION 9 inches in diameter and bled only three days. Neither does the height of the orifice above the soil explain the phaenomenon, for the period of bleeding differed much at equal heights, as No. 8 and No. 4 show. Just as little influence is exerted by the aspect of exposure. On the other hand, the higher or lower position of the trees appeared to have an essential influence hydrostatically, They all stood on the southern slope of a hill 30 feet high, in the Botanic Garden at Giessen, in the following order, from above downwards : — «, No. 8, which bled 3 days. b, — 2, 5 . . 5+3 days. c. 6, d, - 3, — 7, e, — 4, 4 2 7 10 No. 4 stood 5 feet above the bottom of the hill, and was 35 feet high ; the rest were only about 27 feet, in spite of their unequal diameter. This difference of height also appeared to have some influence. The epochs of tapping the different trunks were not far enough apart for me to ascribe any considerable importance to that point. 9. A trunk of Acer campestre, \\ foot in diameter (at the bottom), was tapped at equal depths of 1 inch, in two places, at 3 o'clock on the 5th of March, and the exuding fluid all collected (as in all cases, with care that no rain-water should penetrate) : — A, orifice 1 foot from the ground, south side. B, the same, north side. By about 5 o'clock A had delivered l|^ths of a cubic inch, B also IJ^ths. From that time till 8 a.m. on the 6th of March not a drop flowed out. About 2 o'clock there were 13fths cubic inches in A, and 19J cubic inches in B. At 5^ 30^ there was i a cubic inch from each. On the 7th at 8 o'clock there was no fluid. At 2 o'clock there was ^th of a cubic inch at A, ^ a cubic inch at B. At 4 o'clock no more had flowed out ; neither was there any at 2 o'clock on the 8th. From the little that the short duration of the bleeding allowed to be observed here, the north side would OP SAP IN PLANTS. 31 appear to pour out more sap than the south. The checking influence of nocturnal cold was again distinctly visible here. 10. On the 10th of March a young sycamore stem {Acer platanoides)^ 3 feet high and | an inch in diameter, was plenti- fully watered with solution of ferrocyanide of potassium. On the 15th it was taken up, root and all. When dissected, the reagents showed that absorption had not commenced in the stem. At the base alone, weak, bleared blue spots were pro- duced on a cross section. In one place the colour was deep and sharply- defined enough to admit of microscopic examination : this proved that the walls of a few streaked and spotted tracheae were coloured blue. 11. A white birch, with a trunk 1 foot in diameter, was tapped 1 foot above the ground, on the west side, on the 4th of March ; up to the 11th nothing flowed out. On this day a new orifice was made in the west side of the stem, at a height of 7 feet. Up to the 14th neither gave off any liquid. Only a few sucking flies appeared to indicate that the sap was beginning to rise. It is worthy of notice, how much later the bleeding occurs in the birch than in the sycamore. 12. On the ^\h of March, at three o'clock, a sugar maple [Acer saccharinum), with a trunk li foot in diameter, was tapped : — A, orifice 1 foot from the ground, south side. B, orifice 5 feet from the ground, south side, but 1 inch further to the east. About eight days previously several small branches had been cut off further up the stem ; some sap ran down from the wounds. The fluid exuding from the borings was collected, and gave the following results : — March 7. 3^ o'clock, at A, 3 cub. in. ; at B, l^ cub. in. .. 5i ... 3i .... .... 7i .... 8. 8 ... 16J .... .... a few drops .. 2 ... 8| .... .... 4 cub. in. .. 5i ... 11 .... .... 4 .... 9. 7i ... 6i .... .... none. .. 2i ... 17 .... .... 7j cub. in. .. 5i ... 7i .... .... 2} .... 32 HOFFMANN ON THE CIRCULATION March 10. 7i o'clock, at A, 3 cub. in. ; at B, — 11. — 12. — 13. 1^ 5 7 2 7 2 5 7i 2 5i 7i .. 2 15. 7i .. 11 14. .. 38(!) .. 9 .. 13i .. 9 .. 5i .. 4 17 .. 4i .. .. 6i .. .. 15i .. .. 4i .. .. 13i .. 23 1 none. . . 5 1 ^ cub. 11 4 3 3i a few drops. 1 cub. in. a few drops. ditto. 2^ cub. in. i cub. From this it follows that the efflux, or fulness of sap, is greater in the lower part of the stem than further upwards. This phaenomenon is not hydrostatical (as a barrel emits a more pow- erful stream from a hole nearer the bottom, than from one at the top, on account of the higher pressure of water), but depends on the force of the water making its way upwards, as is seen by a comparison with Experiment No. 1. It is also again seen what an obstacle the nightly cold is. Lastly, a comparison of the temperatures of the air in the sun, which I observed, shows how much the ascent of the sap is favoured by the heat of the air. From the following table, especially from the two last columns, it is seen that on the whole the outflow runs parallel with the temperature. The 10th of March alone forms an exception, evi- dently on account of the unusually favourable weather on the 9th, the after-influence of which is seen here. OF SAP IN PLANTS. 3J I i : : : : : : = ^ = = = "St* H|« eclo -J« «|i» «l* mK» "p --Iw iNk> -\fi) -+) |-S9 5 (u o »f3 IC W5 «>» (M ©< ep -^ ^» ifS •^ n ^ OS »>.«>» CO op op "^ OS «b '•^ CO iti ^ ^ CO ^ S "» 2S '^ S l>» O O ift OS »;» f-H- "!« -"Iifi _^ <-l«>.0OOS»f5l>» »f5-^£0»-<»ft 1-1 ^ CO ^ ^ (M 3, : : : : : i ; : ^ M|« M5 » -lei (N 0, ^ 1 i : :. : i •• i : : : : : «^ CO : OS d ,-• : (N CO : ■^ : "^ I I I I I SCIEN. MEM.—Nat. Hist. Vol. I. Part I. 34 HOFFMANN ON THE CIRCULATION The specific gravities of the effused fluid exhibited the fol- lowing scale. They were determined in a narrow-neck globular bottle, of about 1 oz. capacity. Rain water, 1st filling . . . 25*660 grammes. 2nd 25-661 3rd 25-665 Sap of the Sugar Maple, Orifice A. Orifice B. March 7. 3^ o'clock 25*877 gramin. — 8. 4 .... 25-875 .... 25-918 gramm •— 9. 2i . . . . 25-871 — 10. 2 .... 25-861 — 11. 2 .... 25-889 — 12. 2 .... 25-883 25-918 .... — 13. 2 .... 25-890 — 14. 2 .... 25-896 According to this, the specific gravity increases pretty rapidly, and the sweetness of the sap also is readily detected by the tongue. The upper part of the stem contains a less aqueous sap than that near the soil. The reaction of the exuded fluid was neutral to blue litmus and to turmeric paper in a fresh condition on the 8th of March ; on the 9th slightly acid, as also on the 15th. The amount of sugar contained was tested by adding solution of potash, a few drops of solution of sulphate of copper, and boihng; after a long- continued boiling only was a very small quantity of copper reduced. Some ammonia was set free in this operation. When the mixture was merely allowed to stand at ordinary temperatures, not the least copper was reduced in two hours. It was therefore cane-sugar. 13. A birch (Betula pubescens, Ehrh.) was tapped in two places at half-past two o'clock on the 8th of March. The base of the trunk was 1 J foot in diameter. A, orifice 1 foot from the ground. B, ... 7 feet ... Both orifices 1 inch deep. Quills were cemented in and the effused fluids caught. On the upper part of this birch a few small branches had been cut off, from which some sap exuded, which, however, did not run down to the bottom. The following quantities flowed out : — OF SAP IN PLANTS. 35 March 8. Up to 4 — 9 — 10 — 11 12. H 2i 5i 7i 2 5 7 21 5i 7i 2 , 9i cub. in. ; at B, 7 7 ... 4 39^.. ... 8i 231 .. ? 9 ... 6 5J .. . ... If 18i .. ... 15i 7 ... 12i 13i . . ... 4J H ■■ ... ^? 3i .. ... 6i 5} .. ... 7J 19| . . . .. 38 cub — 13. — 14. 4f 31i 33 12 As there was no fixed result here, the holes A and B were bored out again to remove any accidental obstruction. March 12. Up to 5 o'clock at A, 5 cub. in. ; at B, 9 cub. in. .. 7\ .... 2 .51 .... 2 .. 7i .... ^ - .. .2 .... 3 J A was closed up, and a new hole C bored near it. March 14. Up to b\ o'clock, from C, 5 cub. in. ; fromB, l;^cub. in. — 15. . . 7i .... I . . none. — .. ..11 .... 3.. 4 drops. Thus, from the 8th to the 11th of March more flowed from below; from that time the proportion was reversed, perhaps in consequence of the more powerful swelling of the lower {wetter) wood at A, and a contraction of the orifice resulting from this. After a new hole was bored, the proportion was as in ordinary cases (compare Experiments 20 and 21). The specific gravities of the fluids from A, B, and C, exhi- bited the following scale : — March 8. Spec. grav. of A, 25*705 gramm. ; of B, 25*715 — 9 25*690 10. 11. 12. 13. 14. 25*712 25*728 25*722 25*718 25*714 — 15. Spec. gfav. of C, 25*732 25*717 25*719 3* 36 HOFFMANN ON THE CIRCULATION According to this, there was increase at A from the 9th to the 11th, then from the 12th (through a fall of snow on the 11th) decrease; on the 15th increase (through the preceding warmth, and resulting evaporation of moisture from the earth ?) ; at B increase. The sap from the lower orifice was not so dense as that from the upper. The reaction of the fresh sap was neutral to blue litmus and turmeric papers on the 8th and 9th of March; on the 15th slightly acid. On the 8th and on the 15th the taste was indistinctly sweetish and earthy ; but chemical testing demonstrated the presence of ^r«j9e- sugar, when the fluid was warmed with solution of pot- ash and sulphate of copper, since a red powder was rapidly thrown down. 14. A birch [Betula puhescens, Ehrh.) 1^ foot in diameter was tapped in various places, in order to discover whether more fluid was effused from the upper or lower part of the stem. This took place on the 14th of March, at three o'clock. A, East side, 1^ foot high, gave in 5 minutes 335 drops. B, 8 feet .... 120 . . A and B were then closed. C, North side, 2 feet high, gave in 5 minutes 118 drops. D, 8 . . .... 80 . . The lower orifice therefore gave more than the upper. The fluids of A and B were neutral with test-papers. 15. Repetition of the preceding experiment in another birch of the same species and of the same size. Experiment made afler three o'clock on the 14th of March. A, N.E. side, 1 foot from the ground, gave in 5 minutes 93 drops. B, . . 9 feet .... .... 103 . . A and B were closed, and two new holes bored. C, North side, 1 foot from the ground, gave in 5 minutes 63 drops. D, .. 9 feet .... .... 51 .. Here also the lower orifice usually emitted more than the upper. The striking and frequent anomalies which appeared here, as in Experiment 13, will appear abundantly explicable when we reflect what an important influence a very small difference in the condition of the orifices (in regard to depth, diameter, quan- OF SAP IN PLANTS. 37 tity of fragments of wood remaining in, &c.) and the unequal expansion of the wood must have ; a difficulty which I could not master. Here, therefore, useful results could only be ob- tained by a number of observations. II. The Summer Sap. The circulation of the sap during the summer, at the period of the greatest activity of the leaves, displays at once much agree- ment with and many striking differences from that of the early spring, among the latter of which stands above all the circum- stance that the trees hitherto mentioned no longer bleed from wounds inflicted on them, although, as a little reflection must reveal, the quantity of fluid actually passing in the stem is far greater ; a fact also demonstrated by direct observation. In summer, as in spring, there exists a rapid ascent of the crude sap ; in addition to this, a descent of unelaborated fluids after every shower of rain ; and, lastly, a descent of the elabo- rated fluids from the leaves into all parts of the plant. Since there apparently exists no means of tracing accurately the mode and course of the last phaenomenon directly, I have restricted myself to the first, namely to the roads which the un- elaborated fluids traverse in their ascent and descent in plants ; but the results obtained could not but give ground for the de- duction of many conclusions as to the behaviour of the elabo- rated saps. The following pages therefore will be devoted to the investigation of the paths by which the crude summer sap ascends or runs down under conditions as natural as possible, and afterwards also under various abnormally arranged conditions, especially when wounds have been made in the plant. A. The Ascending Sap. 1. With normal absorption of the Sap by the Root. For the purpose of tracing the course of the sap, the earth I ^Rround the plants to be experimented on was watered with dilute ^Hfiolution of ferrocyanide of potassium ; after which cross slices of ^Hthe plant were tested for that solution with a mixture of acetate ^Bof iron and hydrochloric acid. It is not advisable to make these 38 HOFFMANN ON THE CIRCULATION fluid is here spread about too much and too unequally ; hence the absorption becomes very uncertain, or even fails to take place at all^ as I have experienced several times to my dis- comfort in vines^ plums, and sycamore trees. I therefore pre- ferred such plants as had been kept a longish time in pots, taking only such as exhibited a full activity of vegetation. Euphorbia terracina, L. — Watered on the 5th of June ; taken up by the root on the 8th. The saline solution was detected in the inner layer of the bark (the liber), and in a few tracheae or spiroids of the outer layer of wood. Watered on the 15th ; withering on the 24th ; taken up on the 25th ; the saline solu- tion could be detected last, as far as 2^ inches above the collar of the root, in the striped spiroids of the outer layer of wood and in the liber, in which it ascended farthest. These vessels were only partially filled with the saline solution ; the majority still contained air and did not react. Since in these and several similar cases, not only the cellular tissue, but also — in direct opposition to the preceding observa- tions on the Monocotyledons — the air-vessels took part more or less in the conduction of the sap, the first object was to clear up the contradiction. The Monocotyledons used in my investigations, whatever their other differences, w^ere almost without exception furnished w^ith tuberous or bulbous rhizomes. The conjecture was not far-fetched, that the predominant subterraneous stem-structure, the whole character of which is, moreover, accumulative and retentive, retarded the conveyance of the fluid into the upper portions of the stem, and thereby exerted essential influence over it. Hence arose the question, whether, in the Dicoty- ledons also, varied conditions in the conduction of the sap would occur according to the rapidity of the absorption, accord- ing to superabundant or scanty watering, &c. a. Accelerated absorption of the Fluid. Balsamina hortemis. — The root was carefully freed from earth and immersed in a large vessel full of solution of the ferrocy- anide ; then the stem was cut across obliquely at a height of 8 inches, and sucked with the mouth. After the sucking had been continued for half an hour, reaction occurred at this point ; OF SAP IN PLANTS. 39 the solution had penetrated into several of the large and small spiral vessels of the stem, and still more had ascended in the delicate parenchymatous tissue surrounding the vascular bun- dles ; the pith and remaining portions of the cellular tissue had taken no part. When the lower end of a fragment of a balsam stem 3 inches long was dipped in ink and the upper end sucked, the ink rose instantaneously. From this is evident how consi- derable an obstacle the uninjured epithelium of the root opposed to the forced penetration of the fluid in the preceding case. Balsamina hortensis. — The plant was allowed to stand dry from the 14th to the 19th of June, until the withering stem had collapsed considerably. The soil was then well watered with 14 cubic inches of dilute solution of the ferrocyanide, which was wholly retained by the mould, as in a sponge (the plant stood in a pot 5 inches high and 4 inches in diameter). On the 21st it began to wither, the leaves exhibited spots and died, while the stem was still partly elastic. On the 23rd the plant was analysed. All parts had absorbed. In the cellular tissue of the pith and rind, the i7itercellular spaces or passages, especially, were found deep blue, so that the rind-cells which contained a red sap, presented red spots enclosed in blue frames; in the pith the whole of the cell-contents were coloured blue in many places. The vessels, striped as well as unroUable spirals, to- gether with the immediately adjacent prosenchyma, w^ere almost without exception dyed deep blue ; air-bubbles were met with in very few, but sometimes even in those vessels which contained blue fluid. b. Retarded absorption of the Fluid. Oxalis tetraphylla, — Watered very slightly with the saline solution from June 25th to July 27th ; the plant was exposed to the atmospheric moisture in the open air. About this time the leaves began to lose their colour and wither. — Analysis. The bulb is composed of two distinctly separated circles of scales, from the interior of which springs the leaf-stalk. The solution was principally met with in the periphery of the inner portion of the bulb ; and the elongated cells at the surface of the separate fleshy scales, but not the spiroids, had also absorbed it in very small quantity. — Leaf-stalk. This contained a loose circle of 40 HOFFMANN ON THE CIRCULATION five vascular bundles ; the salt had ascended in the prosenchy- matous cellular tissue surrounding these, but not in the tracheae themselves ; the latter were filled with air. The cortical layer had also conducted, and indeed in the intercellular passages. Euphorbia terracina, L. — Treatment as in the preceding case ; taken up at the end of four weeks. The solution had ascended in small quantity, especially in the inner cortical layer, the liber. No saline solution in the vessels of the wood. Hibiscus Trionum. — Watered with 1 cubic inch of the saline solution during heavy rain ; taken up after three days. Only the root had absorbed up to this time, and chiefly in the central layer of wood, where the prosenchymatous cells in the vicinity of the vessels were coloured blue in spots : the tracheae took no part. These experiments showed that when small quantities of liquid are absorbed by the root, the sap of herbaceous Dicotyledons ascends, just as in the Monocotyledons above described, in the cellular tissue, and with especial ease in the delicate prosen- chyma surrounding the vessels; w^hile when the absorption is hastened and superabundance of fluid present, the vessels also take part in the conduction of the sap, at the same time propor- tionately parting with the air they contain. 2. Behaviour of the Ascending Sap in abnormal absorption. Salix alba. — Absorption through the exposed wood. — A young leafy shoot 10 inches long was stripped of its bark for 2 inches at the bottom, and dipped 1 inch in the solution ; 2 inches of bark were also removed at the upper end, and this part rolled up in blotting paper. Then, to prevent drying, a glass tube closed at the upper end was passed over the upper half of the shoot. After one day the paper was already moist and reacted strongly blue ; after six days the shoot was analysed ; it was filled in all parts wdth the saline solution, especially, however, in the wood-vessels which were gorged with sap up to the very top; much salt had crystaUized out on both surfaces of the leaves, especially at the bases. Here, with the mouths of the vessels of the wood standing open, a rapid ascent was observed not only in the longitudinal direction, but also horizontally, into the blotting paper in contact only with the alburnum. OF SAP IN PLANTS. 41 Salix alba, — A young leafy shoot 12 inches long was stripped of its bark for 2 inches at the bottom, and dipped 1 inch in the jfluid ; 1 inch below the top (the upper oblique cross section of the shoot) an annular piece of bark 3 lines wide was removed. Protected from drying as before. After six days the salt was found to have ascended into the very apex, and indeed into all parts, but most strongly in the medullary sheath. This con- tained dark blue dotted vessels and unrollable spirals. Only the epidermis of the upper piece of bark gave no reaction. S. alba. — A piece 12 inches long was cut out of a young leafy shoot, 2 inches of bark removed at the bottom, and dipped 1 inch into the solution. In the upper part a little ring of bark was cut out, and the whole allowed to stand without protection from evaporation. After six days, all parts, up to the top of the shoot, even the externally dry, exposed part of the wood which had been laid bare, gave a reaction. At the extreme point the vessels of the medullary sheath no longer took part, but the inner prosenchymatous cells of the wood reacted deep blue. From these experiments it is seen how little share the bark takes in the conduction of the sap, and how readily a horizontal movement of the sap takes place from the gorged young wood into the bark. S, alba. — Absorption through the barh — A piece 12 inches long, of a young leafy shoot, was taken and the bark slit up 2 inches from the bottom drawn back, and the exposed cylinder of wood, 2 inches long, removed ; then the lower end (merely bark) was dipped 1 inch into the solution. After six days the shoot was found remarkably dry, from insufficient supply of fluid. — Analysis. The whole of the stripped piece of bark, even the epidermis, reacted strongly. The wood had likewise ab- sorbed fluid from the bark into all parts at the lower end, but not uniformly ; particular vessels and cells did not react at all. — Cross section 2 inches higher up. The bark and medullary sheath had absorbed most, the pith least. At 3 inches distance from the lower end of the wood, the salt was found only in liber and wood ; the epidermis and pith no longer reacted. Consequently, under favourable circumstances, a movement of the fluids in the bark, and horizontally from the bark into the wood, undoubtedly occurs, although to a very slight extent. 42 HOFFMANN ON THE CIRCULATION The isolating power of the epidermis against moisture is worthy of notice. B. The Descending Sap. It seemed advisable in this case also to examine the different conditions separately, since it must be influential whether the fluids ascend from the uninjured roots before descending from the peripherical parts of the stem, or make their way down di- rectly from the leaves, or from the points of cut shoots, &c. 1. The Descending Sap when absorbed through the leaves. To warrant this experiment on physiological grounds, it suf- fices to refer to the fact of such a condition occurring in nature in every fall of dew or rain, wherein it in fact constitutes a condition essential to the well-being of plants. Salisc fragilis, L. — On June 7th a large uninjured leaf was immersed in the solution, and on the 10th the shoot which bore it was cut off". — Analysis. In the outer part of the shoot -all the systems reacted ; nearer the base at length only isolated trachecB of the wood, and to the greatest distance on the side of the twig on ivhich the leaf arose. S, fragilis, L. — Experiment as before, with the modification that the larger shoot which bore the absorbing twig was notched deeply on the corresponding side. In three days the solution could be traced in the absorbing twig, farthest in the dotted vessels of the woody layer, and above all in the unrollable spirals of the medullary sheath. In the main shoot the fluid had de- scended over the boundary between the inner and outer layers of wood, but not beyond the notch. S. fragilis, L. — Fresh leaves of a young twig were immersed one after another for several days in the solution, until a large quantity of it had been absorbed. The main shoot which bore the absorbing twig was ringed 3 lines broad down to the wood, and the ring of bark removed. — Analysis after twelve days. In the peripherical part of the absorbing twig all systems had again absorbed, but only the medullary sheath conducted far down ; from this the fluid had passed out to the cicatrices of the partly fallen leaves, while the buds in the axils of these, at that time without any vascular connexion with the medullary sheath, were wholly passed over. It had descended a good way in the OF SAP IN PLANTS. 43 interior of the wood of the main shoot, and here again at the boundary between the outer and inner layers of wood (but espe- cially in the latter), while in the bark and liber it had not reached the ringed portion^ much less passed beyond it. The fluid had, moreover, not merely descended, but also ascended, in the main shoot, and in the same region of the woody system. In the absorbing twig itself it likewise passed into the leaves situated below the absorbing leaves, and could readily be detected in the unrollable spiral vessels of their petioles. It must be observed that the two layers of the wood differed very much in their whole character ; the outer was gorged with sap and evidently still in active process of development (June 27th), while the inner was white and dry, and therefore much better fitted to convey the crude juices. It therefore only remains remarkable, that the fluid which penetrated most easily in the medullary sheath of the twig, left the neighbourhood of the pith in the main (older) shoot, and passed to vessels situated further out. I observed the same in Salioc acuminata. Smith, in which the solution was traced through three communicating generations or systems of branches. The still green absorbing twig behaved as above ; the shoot from which this arose possessed three layers of wood ; in this also the fluid had descended chiefly near the pith. The second shoot passed into (or arose from) a third thicker branch in the wood, in which four layers could be distinguished, and here the fluid had passed down at the boundary between the inmost and the next succeeding layer of wood, and not next the pith. It perhaps would not be erroneous to attribute this circumstance to the difference of the annual course of growth, assuming that the fluid always kept to one and the same tract, to vessels of the same age, in passing from the youngest shoot into the older. At all events this is not contradicted by the observed occurrence of several layers of wood, for I have seen distinctly (in Salioo alba) that at least three succeeding systems (or generations) of shoots may be developed in one and the same year, the lowest and thickest containing two clearly distinguishable layers of wood, which however were indicated even in the last and thinnest (in the beginning of July). In order therefore to obtain a surer basis for the decision of the differences of age in the young systems of branches, I examined the condition of that small 44 HOFFMANN ON THE CIRCULATION vascular bundle which diverges at certain points from the me- dullary sheath, and runs into the petiole of the leaf which subtends the bud. Subsequently (in the succeeding year), these vessels, which are torn off externally at the fall of the leaf, are covered up and buried by degrees by the new wood formed in the young shoot (produced by that axillary bud) ; but they may still be discovered, even in old branches, if carefully sought. In the above case, in Salix acuminata^ it was found that the saline solution had descended in the old main shoot, as mentioned, between the innermost and next succeeding layer of the wood, and thus externally over that little vascular bundle (originally going to a leaf) belonging to the inmost layer of wood. Balsamina hortensis. — The solution absorbed by the leaves could in three days be traced upwards in all parts and farthest, and downwards in thevessels and the (wood) prosenchyma accom- panying them ; the latter, however, had conducted very much more fluid, since in cross sections the colourless large vessels w^ere ordinarily perceived surrounded by a delicate ring of very small blue cellular points. In another case also, when the plant had absorbed very little of the solution, it was observed that this had descended principally in the internal cortical layer, and in the prosenchymatous cells in the vicinity of the vessels, not how- ever in the latter themselves. "While, therefore, the tracheae conducted most readily in woody plants, in the succulentbalsam the neighbouring prosenchymatous cells were decidedly overcharged. Perhaps the cause lay in the prosenchymatous cells of the woody plants being in many cases filled with air (which is not the case in the balsam), whereby, of course, the passage of the solution from cell to cell through the, moreover dry, membranes might be rendered more difficult. Lactuca saliva. — The solution absorbed by the leaf during eleven days was contained in especial abundance in the spiral vessels and their surrounding prosenchyma, on the corresponding side ; on the other side of the stem, however, only in the pros- enchyma surrounding the vessels, and in the lower parts of the stem the same. Here, moreover, the rind and the pith also had conducted, the pith principally in the cells at the boundary of the medullary cavity. Enphm'bia t€7racina, L. — After four days' absorption through OF SAP IN PLANTS. 45 a leaf, the solution was traced in the entire stem, chiefly in the wood cells and isolated tracheae (unroUable spirals and dotted vessels), which latter also contained a few air-bubbles at the same time, Tropaolum majus, — After five days' absorption the solution was found inside the intercellular passages of the rind and pith, still more in the prosenchyma of the wood, and most of all in some tracheae. Cucumis Melo. — After one day's absorption all the intercellular passages, for several inches upwards and downwards in the stem, were found filled with the saline solution, while the fluid cell- contents themselves did not react; particular portions of the (wood-) prosenchyma, and above all some of the large tracheae, had also absorbed. The size of the vessels here permitted a decision, for this and all other cases, of the question whether the blue reaction so often observed in the interior of the exposed tracheae might not result from the process of preparation, in short from the cutting of the sections, since the knife might indeed readily spread reacting fluids from the neighbouring prosenchymatous cells into the open mouths of the vessels. If such were the case, if therefore the reaction in the interior of the vessels were exclusively caused by an unavoidable smearing at the time of the analysis, the blue vas- cular points arising from the reaction would not, at all events, always occupy the same place in a succession of transverse slices from a stem where the course of the vessels was exceeding straight. But this actually took place, and it follows beyond doubt that the tracheae can in some cases take up and carry forward fluids. Vitis vinifera. — Here again it was found that the fluid pro- ceeded both downwards and upwards from the absorbing leaf into the shoot bearing it, and in both cases in the same situation, namely, chiefly in the medullary sheath and in the portions of cellular tissue enveloping the liber-bundles on the inner side ; apparently somewhat more had descended externally, and some- what more ascended internally. Cucurbita Pepo, — This time a tendril, instead of a leaf, was immersed in the solution, but it absorbed very little, probably in consequence of continued wet weather. Analysis showed 46 HOFFMANN ON THE CIRCULATION that the fluid had ascended in the prosenchyma accompanying the vessels, but not in these themselves. From these and similar experiments it follows that in the ab- sorption of fluids through the leaves, they are conveyed most readily by the tracheae or the prosenchyma closely surrounding these; in plants gorged with sap more readily in the latter, and the reverse in dry woody plants. But even in the most succu- lent vegetables, only a somewhat longer continuance of the in- troduction of the fluid, or a greater quantity of it, is requisite to cause it to pass very readily into the air-tubes, and at length into all parts. It would therefore be erroneous to assume that any particular anatomical system is exclusively charged with the conveyance of unelaborated fluids in the ascending or descending direction. It was above all seen, that the tracheae do usually convey air in summer, but very readily become temporarily more or less, or even wholly filled with fluids which displace the air. In fact, chemical reasons led me to consider the existence of the gas in the spiral vessels and spiroids as nothing more than a result of the absorption of crude fluids from the soil, which, ascending in the higher and warmer layers of the plant, at once give off almost unaltered the gases dissolved in them, these being diffused through those communicating passages, and so gra- dually evaporated outwards and upwards without doing any mischief. In this point of view the vessels would be regarded as ' tubes of safety.^ It merits some attention, that, as the last experiments prove, no parts take so little share in the conduction of the solution downward as the layers of the bark. I therefore took occasion to investigate the capability of the bark to convey fluids by a direct experiment. This was done by removing every other passage but the bark from the descending fluid. Salia; vitellina. — On the 1 9th of June a fresh pendent twig 1^ line in diameter, had the bark slit up for the length of 1 inch at the side, 5 inches from the end ; the bark was turned back and the wood within completely removed for a length of 2 lines ; then the bark was returned into its place, rolled up in a living leaf to prevent drying, and the shoot strengthened by a splint. Lastly, a leaf situated below the excised wood was immersed in the solution. After twenty-four hours the fluid had advanced OF SAP IN PLANTS. 4? very little in the pith, but elsewhere in all parts (bark and liber included) as far as the cross-section of the wood, but not beyond this even in the bark. The result was similar after absorption for four days. Even after seven days' absorption, the solution had not made its way over the bridge of bark ; indeed this did not itself react, although it was in part perfectly fresh and living. In this case, moreover, the whole of the lower outer part of the twig was densely filled with the saline solution ; even the badly conducting pith of the lower portion of wood reacted distinctly ; the medullary sheath and the peripherical portion of the wood reacted most strongly, especially on that side of the wood corre- sponding to the absorbing leaf. From this it is evident that a far stronger penetration of the fluids than that which occurred here, is requisite to overcome the resistance which liber and bark oppose to the descending sap (see Sect. 4 below). This observation rendered it necessary to apply the saline solution immediately to those parts to which the business of conveying the sap down is most frequently attributed, in order to bring the question nearer to a decision. 2. The Descending Sap with direct absorption by the Cambium layer, Salix acuminata, Sm. — In a branch \^ inch thick, the bark was slit up, separated to a certain extent, and a piece of filtering paper, many folds thick, soaked in the solution, inserted under it, the wound being then loosely bandaged. After one day (June 15) it was found that the salt had neither ascended nor descended beyond the exposed portion of the wood ; it had only penetrated extremely superficially even in the liber which lay directly upon the paper, and not at all into the rest of the bark or the wood. Salix arbuscula, Whlbg. — Bark slit up for 2 inches ; branch 1 J inch thick : otherwise as above. After four days, only those parts directly in contact with the paper reacted ; the salt had not passed beyond in any direction ; the paper was still moist. Salix hippophaefolia, — Branch 1 inch thick ; bark slit up as in the preceding, but the fluid was actually dropped in on the 18th and 19th of June. On the 20th it was found that the solution had not gone beyond ; even the layer of sap-wood was scarcely penetrated ^th of a line deep. 48 HOFFMANN ON THE CIRCULATION Salix arhuacula. — Branch 1^ inch in diameter. Repeatedly wetted, as in the preceding case, during eight days: result almost the same. The solution had only advanced 1 line beyond the exposed spot, and quite uniformly upwards, downwards, and to the side ; the sap-wood reacted ^ of a line deep. When the ex- periment was continued for eighteen days the result was the same. Consequently there is no layer in the whole tree less favour- able than the cambium for the conduction of the sap. In oppo- sition to the views of many inquirers, this part, being in the most active condition of development, most energetically arrests the fluids. It is clear how unfitted the bark is for the transport of fluids ; they occupy a longer time there than in most other parts, in changing their place. And in this experiment, we must not be led away by the results of what are called the " magic rings*" on trees. For if a thickened border is formed on them at the upper cut edge, this only proves that the sap in general has a descending motion ; not, however, that this does not take place far better and more easily in the totally uninjured woody layer. When we reflect that even in the oldest trees a continual in- crustation of the cells, a continual increase of that transforma- tion of the saps, goes on deep in the interior of the wood, the result of which is the concentric growth of the heart-wood, at the expense of what is at first sap-wood, it is seen at once that it would be a great mistake to regard the wood, on account of its solidity, as lifeless and unengaged in the conduction of the sap. 3. The Descending Sap in absorption by the root. Salia? alba, L. — A piece l^ foot long, of a shoot ~ an inch thick, was placed in the ground on the 24th of February, and kept at a moderate temperature, so that roots were formed, and by the 20th of April leaves had already burst out. On the 6th of June the rooted portion was carefully split up the middle, from below upwards, and one of the halves immersed in solution of the ferrocyanide, the other in a vessel of pure water standing close beside. On the 14th all the leaves were dead, the roots still fresh and liealthy. On the 4th of July the height of the fluids in the two vessels was not perceptibly altered, whether the levels were previously alike or different, as counter-experiments * Made by removing a ring of bark running all round the tree. OP SAP IX PLANTS. 49 proved ; therefore no siphon action had been exerted. On this day the twig was analysed. The solution had not only ascended to the upper end in the one part, but also descended in the other part (at the water side), and indeed just to the surface of the water ; it had penetrated farthest of all in the vascular part of the outermost wood, which, at the upper part, was in contact with the absorbing half of the shoot ; in the liber and bark it fell about 1 inch short of this, while the inner wood, the medul- lary rays, and the pith, did not react. The surrounding water exhibited no reaction, which, it may be remarked in passing, does not speak much in favour of the hypothetical '^ root-secre- tion.^' The half dipping in the saline solution, when examined upwards, reacted most in the medullary sheath, and in the (two) outer layers of wood ; also, however, in the liber and bark ; while the inner (third) layer of wood and the pith had not conducted so far. At the upper free and undivided extremity, the branch reacted only at one side, that corresponding to the vessel con- taining the saline solution ; therefore the solution had not passed round by the top to descend into the other half (to the water), but had gone over (in extremely small quantity) hori- zontally from wood to wood further down. When the experiment was stopped sooner, in other cases, it was found (June 12th) that the solution had merely ascended, and not descended ; in another piece of a shoot, the solution had descended half way in the water-half by the 15th of June. In one case, when the experiment was kept in action longer, the solution had descended | an inch down below^ the level of the water in the half dipping in the latter, — not, however, to the highest of the little roots ; here also the water exhibited no re- action ; in fact, the portion of the inner layer of wood here laid bare by splitting the shoot had not conducted. Whether, in these cases, the absorption of the saline solution took place through the roots, or also through the lowest exposed portion of the inner layer of wood, it is certain that here again the liber and bark were decidedly less concerned than the tra- cheae of the wood, in the descent of the sap. It is seen that the descending sap, when it ascended from the roots and penetrated horizontally from wood to wood, avoided the medullary sheath, while it was shown in previous experiments that it very readily SCIEN. MEM.— A^a/. Ilkt. Vol.1. Part I. 4 50 HOFFMANN ON THE CIRCULATION passes into the latter when it is brought into the plant by the leaves ; a circumstance which is doubtless to be explained by the intimate anatomical connexion between the vessels of the me- dullary sheath and those of the petioles in young shoots. 4. The Descending Sap after direct absorption by cut surfaces OF THE WOOD. Salix mtellina. — The end of a pendent young leafy shoot was cut off, and the lower peripherical portion stripped of bark for 2 inches ; the part thus laid bare was immersed I inch in the solution. At 2 inches further up in the same shoot, the bark was sUt up at the side, and the cylinder of wood cut out for a length of 2 lines ; the bark being returned to its place, and the wound wrapped in fresh leaves, the whole shoot was supported by a splint, to keep it in a fixed position. After four days the shoot had absorbed the fluid as far as it dipped in it. In this case the saline solution \i2i^ passed the bridge of bark, had advanced 4i inches beyond the vacancy in the wood, and into all parts ; furthest, however, in the bark and wood, principally in the me- dullary sheath and the peripherical part of the wood. Repeated experiments gave the same result, but sometimes the liber, some- times the wood, had conducted a little farther. Consequently, here, where a forced entrance of the fluid had accomplished the passage through the bridge of bark, the solution had again penetrated in the horizontal direction through the wood above this bridge, and sometimes even advanced further in it than in the bark itself. S. alba. — A portion 1^ foot long of a leafy young shoot was stripped of its bark for 2 inches at the upper end, and dipped, with its wood wrong end upward, 1 inch deep in the solution. Then 1 inch of bark was peeled from the other free end, and a closed glass tube turned down over it to prevent de- siccation. After seven days the fluid had ascended through all parts; whence, comparing this with the cases mentioned in section A 2, it results, that in absorption by exposed layers of wood, it makes no difference in the conduction of the sap whether the shoot is immersed in the fluid upright or in a reversed posi- tion. In this case also, some salt had crystallized out upon the leaves, When the free end of the wood was enveloped in blot- OF SAP IN PLANTS. 51 ting paper, the latter absorbed a great deal of the solution (in the horizontal direction from the wood) even when the bark much lower down was unaffected. Or if only a ring of bark was cut out on the upper part of the shoot, this interruption was no hindrance to the advance of the solution ; it was found at the end, both in the wood and in the bark. The epidermis did not give a blue reaction even after remaining one hour in contact with the salt of iron. Salix acuminata, Sm. — The question investigated in this case was, how far a horizontal conduction of the fluids can take place under favourable circumstances in the wood itself, through layers of different ages. For this purpose the point of a small twig was cut off, at the end of June, and the open end immersed in the solution. The main shoot (6 lines thick) which bore the foregoing was so notched circularly in four different places, that there was no immediate communication with the vessels of the stem in any place : the wounds were enveloped in fresh leaves. After eight days it was found that the solution had advanced exclusively in that side of the main shoot which corresponded to the absorbing twig, and indeed only as far as the notch which interrupted the vascular communication 4 inches further up. Here also the outermost layer of wood and the liber had conducted principally, and not the cellular intermediate layer of the bark. From this we see what difficulties are opposed to the assumption of a horizontal movement of the sap through the medullary rays; although, at the same time, a horizontal movementof the sapsin the young wood, generally, under very favourable circumstances, as in the experiments of Hales (/. c), cannot be disputed. Postscript. — With regard to the behaviour of the milk-sap, which I had an opportunity of observing in several EuphorhvB, in Sonchus oleraceus, &c., in reference to the conduction of sap, I am led to assume, from all that I could notice, that it takes no part whatever in this, whether the fluid penetrate into the plant through the leaves or through the roots, setting aside all the anatomical reasons against a circulation of the milk-sap, the most decisive of which is, that in the majority of plants the milk-sap passages have no continuity or general distribution. [A.H.] 4* 52 MULLER ON THE MALE OP Article II. Upon the Male of Argonauta Argo and the Hectocotyli. By Professor Heinrich Muller of Wurzburg. [From Siebold and KoUiker's Zeitschrift fiir Zoologie, June 1852.] Among the many perplexities presented by the sexual rela- tions of the Cephalopoda, we have had to reckon, even up to the present time, the statement of the majority of observers, that they had found none but female Argonauts. I believe that in the present essay I for the first time describe the perfect male Argonaut, as one of the arms of which the so-called Hectocotylus Argonauta is developed. The Hectocotyli, which, from the first, Cuvier called "truly extraordinary'^ creatures, will none the less deserve that title. It is well known that Kolliker* has endeavoured to show that the Hectocotylus of the Argonaut, described by Delle Chiajef and afterwards by Costa J, is the male of this Cephalopod ; that the newly discovered Hectocotylus Tremoctopodis also is the male of Tremoctopus violaceus, D. Ch. ; and Von Siebold § has assented to his views. More lately Veranyll, in his work upon the Cephalopoda, communicated some very important discoveries with regard to the Hectocotylus of an Octopod. He found, that among five specimens of a peculiar species which he had previously named Octopus Carena, in three the third arm upon the right side was longer and stronger than the others, and was provided with a vesicle at its extremity. The fourth specimen had in the same position a short pedunculated vesicle ; and the fifth possessed simply the peduncle without either arm or vesicle. Filippi noticed that the longer arm, which in one instance was observed ♦ Annals of Natural History, 1845. Linnaean Transactions, vol. xx. Bericht von d. Zootomischen Anstalt zu Wurzburg, 1849. t Descrizione, iii. p. 137, tab. 152. X Annales d. Sciences Nat. 1841, p. 184 and pi. 13. § Vergleichende Anatomie, p. 363. II Mollusques Mediterraneans, l^^e partie, Genoa, 1847-51. ARQONAUTA ARGO AND THE HBCTOCOTYLI. 53 to drop off on being touched, resembled the Hectocotyhis Octo- podis of Cuvier*, and Verany concludes from thence that this Hectocotylus Octopodis is a deciduous arm bearing male organs which are probably periodically developed. With regard to the Hectocotyli of the Argonaut and of Tremoctopus on the other hand, Verany beUeves that they cannot be arms of the corre- sponding Cephalopoda. These statements rendered the subject of the Hectocotyli far more difficult than ever. It could hardly be believed that the Hectocotylus of the Octopus could be really distinct in its nature from the two examined by KoUiker ; and yet upon the other hand, there were many reasons for hesitating to apply the con- clusions drawn from the former to the latter. The Hectocotylus octopodis differs in many respects from the others ; its sexual relations are less certain, while those of the Octopods to which it was attached, either in the mantle or as an arm, are wholly unknown ; and finally, the positive assertions of Madame Power and Maravigno (see KoUiker, /. c.) seemed to prove that the Hectocotylus Argonautce was developed as such in the ova of the Argonaut. While at Messina, in the past autumn, I was very desirous of repeating the observations of Madame Power; but notwith- standing the examination of many thousand ova of all the Ar- gonauts which I could procure, I merely found embryos of the ordinary form more or less developed ; never those vermiform young, whose description had led to the belief that the Hecto- cotyli were developed in especial bunches of ova. At last, at the end of September and in the beginning of Oc- tober, there were brought to me, among many very small Argo- nauts which had not yet acquired a shell, a few of a quite pecu- liar form. Their cephalic extremity presented a little sac, which projected between the arms, as the animals swam about with their peculiar retrograde movement. On closer inspection f one could perceive seven arms, which all terminated in points like the six lower arms of other Argonauts of the same size. The * Annales d. Sc. Nat. 1829, p. 147, pi. 11. t The relations of the parts were clearest when the animals fixed themselves during life, within a glass, so that one could look from without straight down upon the oval surface of the head ; or after death, by placing them in a waxen, pit so as to obtain a similar view. 54 MULLER ON THE MALE OP two upper and the two lower arms were longer than the lateral ones ; of the latter on the right side three only were present ; while on the left side there existed, in all the specimens which I received, in the place of the lower lateral arm, the sac in question, supported by a short and delicate pedicle, as if it were constricted. The pedicle arose from a small depression between the second and fourth arms and the mouth, from which the sac could be easily drawn out a little. The membrane which unites the base of the arms of the Argonaut passed upon the left side from the second to the fourth arm, without immediately invest- ing the sac, whose position was somewhat internal to it. The sac itself was not so long as the arms in the smallest specimens, whilst in the larger it equalled or exceeded them in length. In shape it was not exactly round, but somewhat elongated and compressed in such a manner that the diameter in a radial direction from the mouth was greater than in the line of the two neighbouring arms. The colour, like that of the rest of the body, was intensely reddish brown when the chromatophora were dilated, more greyish when they were contracted. Only on the inner, oral side was there a white streak without chro- matophora, which however did not extend over the whole length of the sac. In all cases a Hectocotylus Argonautae lay coiled up withm Ihe sac. It was curved towards the side which bears the suckers, so that the back of the thick part corresponded longitudinally with the internal convexity of the sac. The part described by Kol- liker as a silvery sac, forms at this place, immediately under the skin of the sac in large specimens, a ridge-like elevation visible externally, through which a whitish tint often glistens. The thinner part of the sucker-bearing body is bent back along the inner convexity of the sac towards the base, and the filiform ap- pendage lies betw^een them in multitudinous convolutions. This position of the Hectocotylus is frequently obvious from without, especially during the lively movements which it often makes ; and still more clearly on the opening of the sac, when it uncoils itself from its narrow cell under the eye of the spectator. The relation of the Hectocotylus to the capsule in which it lies, and the change which the latter undergoes after its eversion, are very remarkable. ARGONAUTA ARGO AND THE IIECTOCOTYLI. 55 1 may here observe, that in one case the sac burst before my eyes, along the inner, oral side, in consequence of the violent movements of the Hectocoiylus, when the process about to be described took place in exactly the same manner as in other spe- cimens which were artificially opened. It is observed, in the first place, that the thick end of the Hec- tocoiylus is fixed to the pedicle of the sac or forms it ; next, that the membrane of the sac is perfectly distinct from the filiform appendage and from the neighbouring parts of the sucker-bear- ing body ; but that, upon the thick portion of the body, while it leaves the sucker- side free, it is attached along the back behind the suckers, and forms the covering of the silvery sac above men- tioned. The so-called pigmented testis capsule of KoUiker, however (as it is observed in the Hectocotyli which are found free upon female Argonauts), does not yet exist, and is subsequently formed from the membrane of the sac. As soon indeed as the appendage and the thinner part of the body, which usually become twisted upon their axis at the same time, are evolved, the thick part bends forcibly back in the op- posite direction to the previous curvature, that is towards the back. By this means the longitudinally cleft membrane of the sac is inverted, so that its inner surface comes to be exterior, and the edges of the torn part are turned back towards the back of the Hectocoiylus, which is now concave. The previously external pigmented layer of the sac now lies in the pit between these edges, and when the latter have united, there is left only a small cleft, a process which can naturally not be directly traced. We have just such a pigmented capsule formed as has been already found in the dorsal crest of Hectocoiylus Argonautce. In this way we readily account for the singular fact, that a colourless layer is constantly found upon the exterior of the dorsal crest, while the layer of chromatophora lies internally upon the so-called capsule of the testis. The membrane of the sac then belongs to the future Hectoco- tylus. This was seen most clearly in that specimen in which, as has been already noticed, the sac opened spontaneously ; for upon touching the Hectocoiylus frequently it detached itself from its delicate pedicle so as to carry away the inyerted sac with it. 56 MUL.LER ON THE MALE OF However, as the cleft in the sac had not extended quite so far as the insertion of the pedicle, the first suckers still remained hidden by the pigmented sac, whose borders began to be reverted only opposite to the fourth sucker. The case here cited hardly allows us to doubt that the Hecto- cotylus once formed is intended to become detached from the rest of the animal ; as might indeed already be concluded from the fact that all the Hectocotyli seen by Delle Chiaje, Costa and KoUiker, to which I can add thirteen others, were found sepa- rated and associated with female Argonauts. Hence also it would seem to be probable that the detachment of the Hectocotylus is preceded by the bursting of the sac; though I have found no specimen in which when captured the Hectocotylus had already made its exit from the sac. When and in what manner the separation of the Hectocotylus, and its transport to the female, go on ; whether any act of copulation, for instance, takes place, were points upon which I had no opportunity of making any observations. I will now first consider a few points with regard to the ex- ternal and internal structure of Hectocotylus Arffonautce, and I will then compare the Hectocotyli of Octopus and Tremoctopus with it. For the most part I have only to confirm Kolliker's observations, though of course my interpretation of them must be somewhat different. The name " Hectocotylus '' may very well be retained, without any implication of independent ani- mality. Hectocotylus Argonautee. As to external form, there w^ere two portions to be distinguished in all the Hectocotyli, whether free or enclosed in sacs, which 1 examined : the one thick, and carrying suckers ; the other called by KoUiker the filiform appendage, thin and suckerless, but di- rectly continuous with the former. In the free Hectocotyli the body and its appendage sometimes attained the length of an inch or more each ; in other cases, each was some lines shorter. A few of the Hectocotyli which were just set free from the sac had this latter size. In three speci- mens, in which the body and head of the whole animal, as far as the base of the arms, were about 4 lines long, the sucker- bearing portion of the Hectocotylus-divm measured 8-10 lines, and ARGONAUTA ARGO AND THE HECTOCOTYLI. 57 the appendage about as much more. In an animal of 3 lines long, each part of the Hectocotylus-a.rm was a few lines shorter. The smallest specimen which I met with measured 2 lines to the base of the arms ; body and appendage of the Hectocotylus-2irm. each 3-4 lines. The length of the uninjured sac was about 1 line in an animal of 2^ lines ; on the other hand, it was 3 lines in a specimen whose body, as far as the arms, was 4 lines long*. At the thick end of the detached Hectocotyli is the point, where the constricted axis must finally have divided ; it is drawn a little towards the dorsal side, while the first suckers project somewhat forward. No trace of any rent is to be seen, but the surface is quite smooth as if cicatrized, and the fringe which unites the suckers upon their dorsal side is also present between the oblique anterior pair, so that the one series of suckers passes in a continuous curve into the other. The point of transition of the thicker body into the filiform appendage is sharply marked in all the free Hectocotyli and in the larger enclosed ones. The suckers with the fringe which unites them cease suddenly, the axis of the body becoming thinner and passing into the appendage. In the smallest speci- men before mentioned, on the other hand, the transition was far more gradual. The suckers in the posterior broad part of the body, which did not measure more than '15 of a line across, be- came gradually smaller and more rudimentary, and finally ap- peared as mere transverse elevations ; when they ceased the dia- meter of the body was still 0*1 of a line. The membranous lobes described by Kolliker at the origin of the suckerless part of the body were present in all free Hectoco- tyli ; but it could generally be clearly observed that there is pro- perly speaking only a single lobe, which in its highest part crosses the body of the appendage transversely, and then passes gradually upon each side into a slight fold. These two folds run along the appendage for a considerable distance : in one case, the most elevated portion of the lobe was prolonged into two elongated processes. The height of the transverse portion * An eighth specimen, in which both the body and the unopened sac surpass the above dimensions, is in the possession of M. Verany, who immediately re- collected it on seeing my specimens. It had previously been brought by Krohn from Messina. 58 MULLER ON THE MALE OF varied from an inch to more than half an inch (1 bis uber ^''), The fibrous tissue of which the lobe consists is contractile, and frequently moves very vivaciously by itself. In the yet attached Hectocotyli the lobe was in the same way more or less developed, and was wanting only in the smallest specimen. This is un- favourable to the view that the lobe is a residue of the torn sac ; which might otherwise suggest itself, especially since the strong resistant epithelium which externally coats the appendage is absent upon the lobe. The edges of the lobe were generally smooth and did not appear torn. Of the tentacular cirrhi which Costa (/. c. fig. 2" e and/) de- picts at the anterior end of the Hectocotylus, I could never find any trace, and I am inclined to believe that they were some ac- cidentally adhering foreign bodies, since nothing could be lost or torn away, in the Hectocotyli otherwise perfect, which were taken out of the sac. Such specimens as the latter are also important for the deter- mination of the interpretation which is to be put upon the dorsal pigmented capsule * and the position of the appendix in it. Kolliker has called this capsule the capsule of the testis ; — inasmuch as in one specimen he saw the filiform appendage enter it through a cleft in the back, and become connected therein with a coil of seminal canals, to which he gave the name of ^ testis.^ I believe that the presence of semen there is acci- dental, and that another interpretation must be given to the position of the appendage. It has been already stated, that there exists no capsule in the Hectocotyli while still included in the sac ; the appendage is always free, and nothing is to be seen of any seminal canals. In the free Hectocotyli the capsule was indeed always formed, but in many instances it was quite empty, the appendage also lying outside it. In other cases the appendage passed, in the manner depicted by Kolliker (pi. 1. fig. 9 and pi. 2. fig. 17)^ through the cleft of the dorsal ridge into the pigmented capsule, but lay free therein, no seminal canals being present. * The chroma tophora here exhibit in free Hectocotyli the same movements as elsewhere. The radial muscular fibres are clearly recognizable, contracted or relaxed according as the chromatophora appear large or small. Muscular fibres are also found in the deeper layers of the cutis in Hectocotyli as else- where among the Cephalopoda, e, g, in the dorsal ridge. ARGONAUTA ARGO AND THE HECTOCOTYLI. 59 This could be seen partly upon opening the capsule, partly from without during the movements of the Hectocotylus. The appendage not only twisted about in the capsule, but crept al- ternately out and in, so that even the thin part of the sucker- bearing body, as far as it would go, became hidden in the cap- sule. The Hectocotylus then crawled about in a very peculiar manner w ith this inserted part, which formed about the middle of the body, directed forwards. On the other hand, if the Hecto- cotylus were disturbed by touching, it not unfrequently drew its appendage quite out of the capsule, and could then be no longer distinguished from the first form which has been mentioned. It appears then, that the appendage makes the pigmented capsule its residence either from being accustomed to its pre- vious imprisonment in the sac, or as a sort of presentiment of its proper position *. The exceptional occurrence of seminal canals in the capsule of Kolliker's Hectocotylus is explained, if we consider the route which the semen must take to be poured out. I Kolliker has exactly described the course of the vas deferens between an aperture in the neighbourhood of the point of the appendage and a thick silvery sac which lies under the pigment- capsule. He called that sac a penis, or finally, vesicula semi- nalis ; assuming that the semen passes out of the capsule (testis) along the appendage, then into the ductus deferens along the back, and finally out of the silvery sac at the thick end of the Hectocotylus. However, in the most free Hectocotyli, and even in the largest of the included ones, we find this sac completely filled with semen. Sometimes this distension extends into the ductus deferens to a greater or less extent, and evidences itself to the naked eye even, by an intense white streak along the back and appendage of the Hectocotylus. Lastly, on one occasion a Hec- tocotylus passed a whole coil of a thread, about '06 of a line thick and consisting of spermatozoa, from the aperture of the append- age, and the thread remained attached to it, so that the appear- * Considering the occurrence of gills in Hectocotylus Tremoctopodis, we might ask whether that form developes a digestive organ, for which Cuvier took the capsule in //. Oclopodis. At present however we have no evidence upon the point. 60 MUL.LER ON THE MALE OF ance figured by Kolliker (pi. 2. fig. 19) arose: only the ap- pendage was free, while in KoUiker's specimen it was inserted in the pigmented capsule. We may hence assume that the semen during ejaculation passes from the thicker sac towards the point of the appendage*. It agrees very well with this conclusion, that a copulation very probably takes place during which the appendage represents the penis (vide infra). KoUiker's Hectocotylus therefore only committed an error loci when it deposited its semen in the pigmented cap- sule. The presence of an investment to the seminal coil which Kolliker found in the pigment-capsule, is not, as I believed at first, any argument against its secondary deposition there, for a structureless layer was also very visible in the free seminal cy- linder, at least in some portions of it. It is perhaps only analogous to the structureless mass which is to be found else- where in the sexual canals of the Cephalopoda, and is deposited around the semen when excreted. In a second instance I could find no such investment. Since the pigmented capsule upon the back of the Hectoco- tylus could not be the testis, the latter was to be sought for elsewhere. At first, I was tempted to consider the silvery sac as its representative ; since not only was this full of perfect spermatozoa in all free Hectocotyli, but also in that Hectocotylus- arm already referred to which had burst its sac after its spon- taneous detachment. Nevertheless it was surprising that in other Hectocotylus-arms just taken out of their sac, the silvery capsule had not its white colour, and neither perfect semen nor any stages of its development were to be perceived therein. Subsequently I convinced myself that there is unquestionably a testis in the abdomen of the animal which carries the Hectoco- tylus as an arm. Behind the gills and venous appendages a great part of the mantle- cavity is taken up by a capsule, whose free lower wall is very remarkable on account of its isolated chromatophora scattered over as hining golden ground. Behind, * In most cases I could not exactly make out the place of the aperture, though in the two upper thirds of the appendage the ductus deferens is usually easily recognizable, and even far forward has a diameter of 0*05 of a line when it is not collapsed. On one occasion I succeeded in pressing the semen from the silvery sac to within two lines of the point where the ductus deferens only measured 003 of a line. ARGONAUTA ARGO AND THE HECTOCOTYLI. 61 it adheres to the mantle. In the capsule lies a white mass which consists of little cylinders or caeca which are united at one ex- tremity: their length is about 1 Hne, their thickness 0-06--1 of a line. A clearly-defined tunica propria could not be distinctly recognized in these spirit specimens for each cylinder, but ex- panded membranous coverings could be frequently observed between them. In the cylinders themselves large pale cells lay at the periphery ; the interior was in one case occupied by masses which consisted of numerous granules of about 0-002 of a line in diameter, and which had frequently a delicate process in an oblique direction with regard to the axis of the cylinder. In a second specimen there could be no doubt that these lumps were forms of the development of the spermatozoa. There lay in the same position more or less developed bundles of spermatozoa, whose somewhat wavy threads had the same oblique direction with regard to the axis of the little cylinder. This appeared, consequently, to be quite a fibrous streak. The length of the single bundles was about 008 of a line. In these two animals provided with full testes, the generally white and distended capsule of the Hectocotylus was colourless and collapsed. In a third animal again, which had carried the detached Hectocotylus-arm filled with spermatozoa, the shining golden capsule was indeed present, but it was empty. If we connect all these facts together, it becomes very probable that the semen is produced in the testis, and that it is then transferred into the Hectocotylus, although I could not recognise with cer- tainty this portion of the ductus deferens, which must lie under the skin of the head. The silvery capsule, then, would be neither penis nor testis, but vesicula seminalis ; and so long as the Hectocotylus remains connected with the rest of the animal, the essential distinction from other cephalopod males must consist in this, that the aper- ture of the ductus deferens, instead of lying in the mantle- cavity, is placed at the end of the peculiarly-developed arm. The structure of the silvery capsule harmonizes very well with Ihis interpretation. It is, as KoUiker has shown, very muscular *, and in its interior there lay in all the free, and in the largest of the * The musculav fibres are distinguished from those of the rest of the body by a peculiar development, a point to which I shall recur elsewhere. 62 MULLER OX THE MALE OF included Hectocotyli, the coils of a thread of 006-0-08 of a line in diameter, consisting of perfect spermatozoa. I have not seen the aperture of this organ, stated to exist by KoUiker at the ex- tremity of the thick end of the Hectocotylus, If the semen be actually passed out of the testis into the capsule, such an open- ing must exist at one period or other ; but it probably becomes closed behind the deposited semen before the detachment of the Hectocotylus takes place. The spermatozoa of the Argonaut consist of a very delicate thread, at one of whose ends is a somewhat thicker fusiform body. They are therefore analogous to those of Tremoctopus, but smaller, since they measure, as we see especially in the bundles, only 0*08-0-09 of a line in length, of which we may consider the body to form 0*01 of a line. In general the bodies lie grouped together, and from them the threads pass nearly parallel, like the cilia of a colossal ciliated epithelium. On one occasion almost every bundle of bodies was spirally twisted. This occurred in the appendage of a Hectocotylus which I found in the ovarian capsule of a female Argonaut. In this one case I saw a lively movement in the spermatozoa, the groups of which formed regular progressive waves like the sea after a strong breeze. What is there in the fnuscular tube which forms the axis of the Hectocotylus ? is a most important question. Since we know that the Hectocotylus is developed as an arm, it may be surmised, a priori, that the structure of the whole axis will nearly resemble that of other arms, as indeed Kolliker has already shown it does, so far as the muscular tube is con- cerned. In fact, there lies in its interior a chain of ganglia, which answer to the suckers. We see them best in longitudinal sections which have been placed in solutions of chromic acid or corrosive sublimate, and the single ganglia may be separated and demonstrated as far as the root of the filiform appendage. From this point the muscular tube may be very readily traced to the extreme end, but it is difficult to make out what it contains : certainly not the vas deferens ; for this, as Kolliker has shown, is only attached superticially. In fresh specimens I saw a few times a clear tube-like streak, which gave off lateral branches as well in the axis of the sucker-bearing portion as in that of the AttGONAUTA ARGO AND THE HECTOCOTYLI. 63 appendage, in which it measured only '012 of a line; but of its nature I can say nothing. The thickness of the whole axis measured in one case at the end of the seminal capsule 0*39 of a line ; towards the end of the sucker-bearing portion 0-3 of a line; in the beginning of the appendage 0*15 of a Une ; at the end of it 0*03 of a line. The inner tube measured at the same points 0*24, 0*18, 0*08, and 0*025 of a line. The development of the Hectocotylus as an arm, while it ex- plains the presence of a ganglionic chain, equally accounts for the absence of any special organs of sense. But on this ground the sensibility of the skin is very considerable. Since the muscular tube is filled by the chain of ganglia, the supposition that an intestinal canal exists there, which Kolliker himself considered doubtful, must be given up ; at least I have perceived nothing of the kind. With respect to the circulation, I can unfortunately give no information as to the connexion existing between the Hectoco- tylus and the rest of the animal. In the separate Hectocotyli the investigation is beset with difficulties, since they are for the most part very restless, and wind and twist about in the most determined manner. Yet a progressive wavy motion can readily be observed in the trunks which lie upon each side of the back and are immediately continued into the appendage. In one instance I could distinctly perceive that this somewhat slow movement passed upon the right side (the appendage being supposed to be posterior and the suckers below) as far as the extreme point of the appendage, and then returned in the opposite direction. After each wave towards the point, however, there succeeded not merely a centripetal one upon the other side, but centri- petal movements frequently arose, which commenced from the point of the appendage. In other cases I met with a different rhythm in the longitudinal trunks of the one and of the other side, and a few times it appeared to me as if in the same vessel the movement went on sometimes in one sometimes in the other direction, as in the Tunicata. However, two vessels lying close together might readily cause a deception in this case. Whether any distinct central organ of the circulation or heart exist, I cannot as yet decide. There occur indeed considerable dilatations in the vessels; 64 MULLER ON THE MALE OP especially in one specimen, in which I found after the circulation had ceased, a sacculated space '15 of a line long by 'OlS-'OS broad in one of the longitudinal vessels of the back, somewhat behind the extremity of the seminal capsule. At both ends of the dilated portion more delicate lateral branches were given off. Somewhat further behind upon the same vessel, and in a corre- sponding position upon that of the other side, there was a rounded expansion of -OS-'OG of a line in diameter. If these dilatations are to be called hearts — in which case they might be considered to undergo a further development — we must suppose that such exist in many places, which would agree with the general arrangement in the Cephalopoda. But it may well be that these spots had merely remained ac- cidentally dilated after death, for the narrower parts of the ves- sels were evidently contracted, and one of them, further on in the thick part of the body, was evenly dilated to 005 of a line. The differences in breadth, which may be successively observed in the same place during life, are again very considerable. A vessel between the axis and the seminal capsule measured in its condition of expansion 0*05 of a line, and contracted to '02 of a line. This rhythmical expansion and contraction of the larger ves- sels goes on in somewhat different modes. Frequently one por- tion of a vessel is suddenly distended by the wave propelled by the contraction of that portion which lies immediately behind it, and then collapses again. At other times one part is slowly dis- tended by the blood which streams in gradually, especially out of the smaller vessels, and at last contracts with a jerk, whereby the vessel in consequence of its elongation becomes much bent. The independent share of single portions of the vascular system in the centripetal movements of the blood is very clear here as in other parts of the Cephalopoda, e. g. the gills. More or less rhythmical and swift contractions are seen to drive the blood from the smallest venous twigs into the larger trunks ; these help it on further either by an immediate contrac- tion, which is a continuation of that of the smaller branches, or only after they are more dilated by the repeated contractions of the latter. That this venous movement is nowise propagated from the arteries, is clear from the circumstance that single ARGONAUTA ARGO AND THE HECTOCOTYLI. 65 ramifications often pulsate rapidly, while the neighbouring ones are either still or move at a different pace. These relations are especially recognizable in those ramifications which come off from a transverse branch of the longitudinal trunks, between every pair of suckers, and spread in the membranous fringe connecting these. Here, as in many other places in the Cephalopods, one may readily convince oneself of the existence of capillary vessels. Concerning the development of the male Argonaut I can state nothing, since I unfortunately obtained no more ova advanced in development, after I had discovered the Hectocotylus to be an arm ; and inasmuch as I had not previously paid sufficient attention to the form of the arms, in the expectation of finding the totally different vermiform embryos described by Madame Power. Indubitably, however, the male embryos are not to be found in especial bunches of ova, but have been frequently seen among the female ones by Kolliker and myself, although from the similarity of their form no further notice has been taken of them. One is the more justified in this supposition, since the sac with the Hectocotylus appears to be relatively smaller the younger the animal, and from its shape might be easily confounded with the yelk-sac. In some preserved ova of the Argonaut far advanced in deve- lopment, I believe that I can recognize the arm of the Hectoco- tylus. The statements of Madame Power and of Maravigno quoted by Kolliker (/. c. p. 84) contain truth and fiction ; but they may be thus interpreted. What are denominated Hectocotyli three days old are without doubt Hectocotyli ; the description of them which is given, no less than the statement that only two or three are developed in the maternal shell, accords very well with this view. For if we procure Argonauts with advanced ova, in a short time we see the hatched young swim about in innume- rable multitudes. The statement that the seven other arms spring out as buds from the vermiform animal while it is assu- ming the form of the common Argonaut, leads one almost to believe that Madame Power saw entire male Argonauts with Hectocotyltts- arms everted from their sacs in the shell of the female Argonaut. SCIEHt MEN.— iVoY. Hist. Vol. I. Part I. 5 66 MULLER ON THE MALE OF If this were demonstrable from the original manuscript, it would afford an important evidence in favour of the immediate transportation of the Hectocotylus from the whole male upon the female Argonaut. In the following pages I will bring forward for comparison all that is to be said concerning the two other known kinds of HectocotyluSy for it is precisely the very striking differences and similarities of the three Hectocotyli with one another, that pro- mise in course of time to afford an insight into the meaning of the separate organs and the nature of this singular creature as a whole. For this purpose I refer to the description by Kolliker of the Hectocotylus Tremoctopodis, and to that of the Hectocotylus Octo- pedis by Cuvier and Verany. Although the identity of the spe- cies of Octopus, in which Cuvier and Verany found their Hecto- cotyli, is not demonstrated, yet the Hectocotyli are so nearly re- lated that they may well be considered together. Hectocotylus Octopodis, As Kolliker has shown, the Hectocotylus of the Octopus is di- stinguished from that of the Argonaut not only by its greater size, but because at one extremity instead of the filiform appendage there is a vesicle containing a filament ; in other respects the two forms essentially agree. Upon comparing the figures of Cuvier and Verany, we find that the " solid cylindrical body '' which Cuvier indicates as the origin of the silky thread is the muscular axis ; the supposed nervous threads of Cuvier are the vascular trunks, which in Hectocotylus Argonautm take on a similar appearance in spirit ; the sac [e) filled with the coils of a white thread is the thicker seminal capsule ; the canal {h) is the ductus deferens upon the back ; the " ston^iach '^ {d ) corresponds with the pigmented dorsal capsule of Hectocotylus Argonautce. The position of the aperture of this capsule alone differs, since in Hectocotylus Ar~ gonaut(B it remains at the posterior end, while Cuvier has depicted it at the anterior (/). According to this figure and Cuvier's statement that this opening (which he calls a mouth) in the fresh state is slit-like and leads into the pigmented capsule, as well as from the analogy of Hectocotylus Argonauta, I cannot agree ARGONAUTA ARGO AND THE HECTOCOTYLI. 67 with Kolliker (pp. 79 & 80) that this aperture belongs to the seminal capsule. Whether a second aperture for this exists at the anterior end is not certainly determined by Cuvier himself, and he does not state that Laurillard had seen the semen poured out anteriorly. The connexion which, according to Cuvier, exists between the thread (i) everted from the terminal vesicle, on the one hand with the axis of the body, and on the other with the canal (h), which, coming from the seminal capsule, is evidently analogous to the ductus deferens in the Argonaut, is singular. If with KoUiker we may be permitted here to suppose an error on the part of Cuvier, who only examined spirit speci- mens, I should imagine, keeping in view the fact that Verany found a filament with a free and pointed end in the terminal vesicle of the Octopus, that the latter was overlooked by Cuvier. If this be the case, the analogy between the filament exserted from its vesicle and the filamentous appendage (penis) of Hecto- cotylus Argonautce becomes striking. It would then require to be made out whether and how this appendage makes its exit from its vesicle, and we might con- ceive similar relations to those which I shall subsequently show obtain in Hectocotylus Tremoctopodis, In the latter case the presumed error of Cuvier would be easily explicable. The above- mentioned differences in the position of the pigmented sac might also be connected with the difference in the position of the appendage, since in Hectocotylus Octopodis the supposed appendage is perhaps never meant to pass into that pigmented capsule as in Hectocotylus Argonautce, From the structure of the Hectocotylus Octopodis, especially the presence of the silky thread in the capsule, e, and the asser- tion of Dujardin {Helminthes, p. 131) that the white thread consists of spermatozoa, the male nature of this Hectocotylus also may be concluded. Its development is probably quite similar to that of the Hec- tocotylus of the Argonaut. Verany and Filippi (see above) have already proved that the Hectocotylus of the Octopod arises as one of its arms. Verany has observed in the same place a sac, which, according to the figure in which chromatophora are indicated, appears rather to be analogous to the sac of the male Argonaut than to the 5* fiS miJller on the male op vesicle which otherwise terminates the Hectocotylus Octopodis, The pigmented capsule upon the back of the latter again is evi- dence of a process of eversion similar to that of the Argonaut ; and in the specimen of Octopus, which M. Verany had the good- ness to show me, the pigmented spot upon the back of the Hectocotylus-orm appeared very similar, if I do not err, to that upon the back of the Hectocotylus-arm of the Argonaut when just everted from its sac. Here also then the eversion of the Hectocotylus from its sac precedes its separation. Whether the pigmented capsule is already developed in one of Verany's spe- cimens I know not. It is interesting that Cuvier describes one of his Hectocotyli as the arm of the Octopus. Four out of five individuals were found in the mantle of the Octopod ; " the fifth had attached itself to one arm of the Poulpe, and had changed it into a kind of sac in which it had imbedded its head, whilst the rest of the body remained free externally (p. 150) ; " and " it has almost destroyed the arm, and appears to replace it so com- pletely that at first it might be taken for the arm itself ^^ (p. 149). It is hardly to be doubted that the animal "qu'un parasite devore" was a male Octopus, which carried a Hectocotylus newly freed from its sac, but not yet separated. The last link of the series finally is formed by that Octopus, in which Verany found merely a short stump in the place of the Hectocotylu9-ann, which in all probability had already fallen off*. Considering the general resemblance of the Hectocotyli of the Argonaut and of the Octopus, it is very remarkable, that, accord- ing to Verany, in all cases the third arm of the right side is the abnormally developed one in Octopus, whilst in the Argonauts it always is the third arm of the left side. Cuvier does not state in the place of what arm his Hectocotylus was fixed. It is much to be regretted that Cuvier has given no informa- tion as to the sexual organs of the Octopods which carried the Hectocotyli in their mantle or as an arm, and it is the more de- sirable that such Octopods should be closely examined, since their more considerable size will without doubt allow a more easy and better determination of many points than the small Argonauts permit. ARGONAUTA ARGO AND THE HECTOCOTYLl. 69 Hectocotylus Tremoctopodis. The third kind, the Hectocotylus Tremoctopodis, is interme- diate in size between the two others; in form, however, it differs far more from the Hectocotylus of the Argonaut than from that pf the Octopus, Nothing is known of its development. However, besides the structure of the muscular tube of the body already pointed out by Kolliker, that of the suckers, and the presence of genuine chromatophora, we have one very im- portant point of resemblance with the Hectocotylus Argonauts in the presence of a longitudinal series of ganglia. This ganglionic chain, which has already been recognized by Von Siebold, passes from the anterior end of the Hectocotylus to the commencement of the capsule in the abdomen. The single ganglia are so disposed that one lies upon each of the alternating suckers ; they are thence closely appressed. If we make a longitudinal section, not perpendicularly between the suckers, but horizontally, we obtain precisely the view given by Kolliker, pi. 2. fig. 14. It is therefore obvious that the co- nical masses of granular substance described by him, at p. 74, were these ganglia. The doubts as to the presence of any intestine expressed by Kolliker are quite just in this case. The opening which he also gives as doubtful, at the anterior extremity of the body somewhat towards the back, was perhaps only the end of the axis whose already attenuated tube here terminates — the inner layers contributing to form a blind end round the last ganglion, while the outer layers united with the skin form a more or less distinct knob*. If in many cases an opening is actually present, this would indicate even more than the mode of termination which has been described, and which agrees with that of the thick end of the Hectocotylus Argonautcs, that supposing the Hectocotylus of the Tremoctopus to be developed as an arm like the others, this is its attached end. The structure of the opposite end renders this conclusion • Von Siebold also has concluded from the want of such an aperture that there is no digestive organ in Hectocotylus Tremoctopodis, and Cuvier affirms that in Hectocotylus Octopodit the axis has no opening anteriorly. 70 MiJLLER ON THE MALE OF probable ; the ovate or pyriform capsule in the abdomen being similar to that which in Octopus is certainly placed at the free end of the Hectocoiylus-^Lrm. Besides, both in form and position that abdominal capsule is analogous to the membranous lobe on the appendage of Hecto- cotylus Argonautce, Among eighteen Hectocotyli of Tremoctopus, twelve had the abdomen constructed as KoUiker has described it. In six, how- ever, the capsule of the abdomen had a cleft upon its dorsal side ; this commenced close behind the last sucker of the left side (the suckers being supposed to be below and the capsule behind), and extended somewhat obliquely as far as the commencement of the delicate process into which the capsule is prolonged ; the latter lay consequently wide open, and it could readily be seen that it was empty ; that is, that it contained neither the convo- lutions of the seminal capsule nor the ductus deferens, which in other cases are found in the closed capsule*. Upon careful examination, however, the cleft is, even in all ordinary Hecto- cotyli^ to be seen as a streak occupying the direction indicated. There is a ridge upon the right side ; pale, more or less closely adherent, and capable of being raised. It can be observed then that the subjacent layer, upon the left side, passes for a cer- tain distance under the other, and appears to be as it were rolled up at its anterior end. In Hectocotylus Tremoctopodis this cleft then always exists, and it would be interesting to know whether it also occurs in the capsule of Hectocotylus Octopodis, When the cleft is open and the capsule is empty and relaxed, it has a considerable resemblance in external character to the membranous lobe of Hectocotylus Argonautae, which also at times forms a deep pit. Both organs exhibit a lively undulatory movement. In order to comprehend more exactly the relative position of the capsule, it is previously necessary to show that the penis of * I presume, that at least in most cases the capsule has been burst after the Hectocotyli were taken, by contact with fresh water or the like. In three Hec- tocotyli of the ordinary form which I had thrown into a dilute solution of chromic acid, I also found after a few days the capsule open and the contents fallen out. ARGONAUTA ARGO AND THE HECTOCOTYL.I. Jl the Hectocotylus Tremoctopodis is analogous to the filiform ap- pendage of Hectocotylus ArgonautcR, Of the six specimens in which the cleft in the capsule was open, three were particularly distinguished by having no penis visible externally. It was not torn off, as might have been con- cluded from the absence of the opening out of which it generally passes * ; but it lay spirally coiled up under the skin inferiorly behind the last suckers, and, indeed, more towards the right side. It was very short, but relatively thick. Here then it was clearly evident that the penis is the immediate continuation of the muscular axis, which could not be so well demonstrated in specimens with a long free penis. Close behind the last suckers, the end of the thick part of the axis, which contains the last ganglion, lies in such a manner in the lower wall of the capsule, that it can be seen through the dorsal cleft above mentioned as a knob somewhat projecting from the inner surface of the cap- sule. From this the muscular tube bends down to the lower side of the Hectocotylus and forms the penis, which here lies coiled up under the skin, instead of as usual passing forward and becoming free in the neighbourhood of the third to the fifth sucker. Where the axis bends down, the longitudinal vascular trunks pass into the penis f. Immediately beyond this flexure the ductus deferens, generally convoluted, comes from the left side to the anterior side of the penis into which it penetrates. The larger convolutions which this duct frequently makes at the base of the penis produce the transverse ridge, which may be frequently observed in many Hectocotyli between the body and the capsule upon the lower side. If now we consider the place of the cleft capsule, keeping in mind this relation of the penis to the axis, its position upon the back of the axis (where the thicker part passes into the thinner) is quite analogous to that of the membranous lobe in Hectoco- tylus Argonautce, • In one specimen the semicircular edge which usually surrounds the aper- ture behind was faintly indicated, perhaps in preparation for the subsequent exit. t Whether at an earlier period the contents of the axis also are continued into the penis, I cannot decidedly say. 73 MULLER ON THE MALE OF If then the abdominal capsule of Hectocotylus Tremoctopodis is analogous vTith this lobe, and perhaps also with the terminal vesicle of Hectocotylus Octopodis, we must give up its supposed analogy with the pigmented capsule of the two other Hecto- cotyli ; and the Hectocotylus TVemoctopodis seems to possess nothing comparable therewith. On the other hand, the presence of a free penis by no means constitutes an absolute difference from the other Hectocotyli. It is its position chiefly which distinguishes it from the appendage of Hectocotylus Argonautce, and the circumstance that the ductus deferens lies in its interior, whilst in the latter it is only attached externally to the prolongation of the axis. If it were certain that the filament which we find in the terminal vesicle of Hecto- cotylus Octopodis had the same signification, and is not a mere seminal tube, the analogy of the three Hectocotyli in this point would be complete. In the other portions of the sexual apparatus, the testis and the ductus deferens, such an agreement cannot at present be demonstrated. Kolliker has named ^ testis ' a vesicle which generally com- pletely fills the abdominal capsule. The cleft outer capsule readily separates again into two layers, the outer of which is similar to the general cutaneous investment: under the epithe- lium there is a fibrous network with numerous vessels, whose capillary loops may be seen in the delicate terminal prolonga- tion. The second layer consists, like the subcutaneous tissue of the back, of muscular bundles, which especially affect a longi- tudinal arrangement. Upon the inner surface it supports a layer of delicate polygonal cells. Below this again comes the so-called sac of the testis, which may be easily separated. Its wall has a peculiar checkered [carrirtes) appearance ; two layers of fibres are visible, which cover one another very regularly like the fibres of a tissue, at right angles, or often at somewhat oblique angles. The fibres are when isolated somewhat rigid, but often- times not dissimilar to muscular fibres. Elsewhere, however, they can be hardly separated at all, and in many places almost structureless layers occur. In the interior of this vesicle there was always contained the thread described by Kolliker, which consists almost entirely AROONAUTA ARGO AND THE HECTOCOTYLI. 73 of perfect spermatozoa. In this case a special investment is fre- quently not to be discovered, as Kolliker and Von Siebold {VergL Anat, p. 411) state; in a few instances^ however, the greater part of the thread was surrounded by a distinct structureless membrane. Its existence was clear in places where it stretched over gaps in the contents, or where it was quite empty, and also in places where it was torn, and the masses of spermatozoa had swelled up and made their way out. Whether this investment is a true membrane which subsequently disappears, or whether more probably it is rather an accidentally deposited homogeneous mass, I will not attempt to decide ; only it is to be noted that this in- vestment was not to be found in a few other portions of the same seminal cylinder. The one end of this filiform seminal mass is connected with the bulb which forms the commencement of the ductus deferens (ductus ejaculatorius, V. Siebold) described by Kolliker. One part of it lies coiled up with the seminal cylinder in the checkered vesicle, the other part stretches on into the penis. This peculiar ductus deferens* appears to consist of a sub- stance essentially identical throughout, but varying very greatly in consistence and form in different localities. It is a yellowish or colourless, sometimes tough, sometimes more brittle, but elastic mass, which frequently, e. g. in the interior of the bulb, is tolerably soft, but at other places is of almost horny hard- ness and friability. Histologically, it appears sometimes struc- tureless, sometimes marked very beautifully with parallel stria- tions. The striae are either immeasurably close or 0*004 (fre- quently 0'001-*002) of a line apart, and exhibit transitions from the most extreme delicacy to very marked lines. They have the greatest similarity with those which we see in the wall of the Echinococcus vesicles. The structureless mass appears to pass into more slightly or more strongly marked layers which determine the partly longitudinal, partly transverse striation. This often takes on very strange forms when the parts are torn, doubtless in consequence of the folding and tear- • I retain this term, although the part does not seem to be quite analogous to the organ so named in H. JrgonautcB. 74 mCller on the male of ing. The substance is little altered by a diluted solution of caustic soda. The innermost coat of the ductus deferens is generally formed by a layer which strongly refracts light, and when it is stretched looks like a tubular brittle glassy membrane. Elsewhere when collapsed it has the appearance of a longitudi- nally fibrous cord,which is easily torn transversely into fragments, and only the local dilatations (for example from 0*02 of a line in diameter to quite smooth clear vesicles of 0*2 of a line in diameter) show its tubular character. One portion is frequently invagi- nated for a certain distance in a funnel-shaped expanded portion behind it, whereby a swelling is produced. In other places the innermost layer forms the spiral band mentioned by Kolliker. This exhibits considerable elasticity, and its coils are often very close, often far separated, which depends upon the form of the penis. The band has sometimes the appearance of a spirally cut delicate tube, sometimes that of a cylinder, like the snake-toys which are cut out of horn. Next to this innermost layer the mass of the ductus deferens appears to be longitudinally striated to an irregular extent. Ex- ternally again it becomes circularly striated, and not unfre- quently some layers appear to be more sharply distinguished from the rest. In the midst some portion will frequently be found wholly structureless, and quite external to the ductus deferens ; a strongly marked layer has again all the characters of a so-called vitreous membrane. In the penis the whole ductus deferens, without reference to its contained spiral band, sometimes forms spiral coils of much greater extent, in which the outer layers of the penis take but little part. Many modifications of the ductus deferens occur which appear to indicate different stages of development. Instead of the usually tolerably solid enlargement at its commencement (Kol- liker, pi. 2. fig. 11 d), there exists at times a somewhat large pyriform body which attains a few lines in length, and the larger it is so much the softer are its contents. In its axis, however, one can already distinguish the commencement of the inner denser tube, which further on in the ductus deferens incloses or forms the spiral band. ARGONAUTA ARGO AND THE HECTOCOTYLT. J5 In one of the specimens without any externally visible penis there projected from the cleft in the back a transparent, pointed, ovate vesicle of a few lines in length, which contained merely a fluid, and at its fixed end was drawn but into a more delicate fine tube of about the same length. The latter appeared when the vesicle, in consequence of being frequently touched, detached itself, and was evidently formed analogous to the ductus deferens ; it consisted of a longitudinally striated cord (i.e. perhaps a folded tube) and an external, distant, structureless, partially laminated sheath. The larger vesicle therefore may per- haps be considered as an earlier form of development of the bulb, with which the vas deferens generally begins. Other changes in the latter and in the penis appear to belong to a later period. In two Hectocotyli which were met with in copulation upon female Tremoctopoda {vide infra), the abdominal capsule was also open and empty, probably in consequence of long lying in water. The penis, however, with the external part of the ductus deferens plainly visible in it, was distinguished on both occasions by a length of an inch and a half. It passed out of the skin, not in the middle line, but nearer to the right-hand series of suckers, and close to the penultimate pair, which plainly arose from its being torn. Its outermost third appeared to be similar to the whole free part of the penis in other cases ; the two upper thirds were thinner, as if pulled out longitudinally. After the skin was taken away from the place of exit of the penis as far as the abdominal cap- sule, the portion of the penis lying below it was seen to pass below the axis obliquely towards the last sucker of the left side, and there cease. This inner part of the penis formed a fusi- form enlargement, which appeared to be hollow. The outer- most layers of the penis passed into the surrounding fibrous tissue, but the connexion with the axis was no longer recogni- zable. In all probability these changes of form of the penis and ductus deferens are connected with the function of copulation, and a third Hectocotylus, in which the outermost portion of the penis was plainly torn off, but the inner end had the same rela- tion as in the other two, had probably already performed this act. Its abdominal capsule also was open and empty. The eversion of the ductus deferens for the purpose of ejaculation, indicated by Von Siebold (/. c. p. 411), appears to take place. 76 mCller on the male of and indeed in a peculiar manner, which reminds one of the process of eversion of the spermatophora of other Cepha- lopoda described by Milne-Edwards {Annales d. Sc. Naturelles, 1842). From the preceding facts I will merely draw the conclusion, that the sexual organs in Hectocotylus Tremoctopodis are not only more complicated than in Hectocotylus Argonautcs, but that dif- ferent stages in their development occur, although the Hectoco- tylus possesses in general the same form as that with which we are already acquainted. Our knowledge of this Hectocotylus is too imperfect to enable us to give a general interpretation of the generative organs. But its structure and the analogy with the Hectocotylus of the Argonaut lead to the supposition that the vesicle which contains the seminal coil is not the testis, but a seminal receptacle, although the mode in which the semen makes its way, and the place of its origin are less demonstrable than in the Hectocotylus of the Argonaut. Next to the generative apparatus, the most striking features in Hectocotylus Tremoctopodis are the numerous villi on each side of the back, which Kolliker has with great probability called gills *. In the living Hectocotylus the single villi are contractile, ap- parently in consequence of the fibres which form a meshwork in their interior. Independently of this movement of their sub- stance, a considerable rhythmical contraction may be observed in the efferent (venous) part of the very rich and frequently anastomosing vascular network which lies in each villus, and which passes from the finer to the coarser vessels, as was stated above to be the case in the Hectocotylus Argonautce and in the Cephalopoda in general. In one instance there were twenty-two contractions in ten minutes. Since these branchial villi might occur in different individuals in different degrees of contraction, the determination of their size in different individuals appeared to be a matter of some interest. For this purpose the largest group of vilh in each of several specimens preserved in spirit was selected. When they were very much developed, their length was about 0*6-1*2 of a ♦ Do these in any way subserve nutrition within the mantle of the female? (see Von Siebold, /. c. p. 389). ARGONAUTA AUGO AND THE 11 F.CTOCOTYLI. [77 line. The smallest were not much shorter, and some were longer. The breadth in the middle of the villi was generally 0*15, hardly under 0-12, but as much as 0*22 of a line. These were Hectocotyli with a freely projecting penis. In two others, on the other hand, whose penis was hidden, the length of the villi was but rarely over 0-6-0-7 of a line, and the most were shorter. The breadth was at the base rarely more than 01 2 of a line, and diminished rapidly to O'Ofi or 0*04 of a line. The villi namely, had here the form of a rapidly dimi- nishing and pointed cone, whilst in specimens with well- deve- loped gills the diameter of the gills in the outer half surpassed at times that of the base, and the end was more rounded than pointed. Since it may be assumed that Hectocotyli with larger gills have in general progressed further in their development than those with smaller ones, we have here a further evidence that the penis rolled up under the skin of the latter is a younger stage of de- velopment than the common form of the free penis. The Hectocotylus of the Tremoctopus, then, is very strongly distinguished from the two others by the want of a pigmented dorsal capsule ; by the position of the seminal coil in the cap- sule at the end of the body, and the peculiar structure of the ductus deferens ; lastly, by the presence of gills : and, indeed, from the great discrepancies which exist among the Hecto- co/yZw5-bearing males of Cephalopods among nearly allied spe- cies, we should be prepared for thorough differences among themselves. Upon the other hand, the analogy of the Hectocotylus of the Tremoctopus with the others in all essential points is so great, that although there is a complete want of all direct observations, we must assume it to have a similar origin, and that one day an entire male Tremoctopus belonging to this Hectocotylus will be discovered, the exact investigation of which will doubtless be still more interesting than in the Argonaut. However that may be, all three Hectocotyli must be kept in mind in attempting to determine the nature of the Hectocotyli in general. 78 MULLER ON THE MALE OF Nature of the Hectocotyli. With regard to this problem, we must consider, first, their relation to the animal which lodges them in their free condition, and, secondly, to that as whose arm they are developed. The main point as respects the former relation^ which Kolliker first demonstrated for the Hectocotyli, may be safely assumed to be — that each of the three forms of Hectocotylus, considered as one with the animal upon which it is developed *, forms the male factor with respect to a particular kind of female Cepha- lopod; Argonauta, Tremoctopus and Octopus granulosus, Lamarck, O. Carena f, Verany. The evidence of this consists in the fol- lowing facts : — 1. No other males of the Cephalopoda mentioned are known. All Argonauts { of the ordinary form with velate arms, and all individuals of Tremoctopus which have been dissected, were females with ova. To the Argonauts enumerated by Kolliker I can add fifty others of every size, and to the thirteen individuals of Tremoctopus, thirty which I have examined with reference to this matter. As has been said, nothing is known respecting the Octopus. 2. Most of the free Hectocotyli carried semen demonstrably ; and it is very probable that the others had done so. To the fifteen Hectocotyli Tremoctopodis fourteen others may be added in which this was certain, w^hilst in four the demonstration failed on account of the emptiness of the capsule. Tn Cuvier's Hec- tocotylus Octopodis Dujardin found spermatozoa. To the six Hectocotyli Argonautce enumerated by Kolliker, I can add thir- teen which all carried the white sac under the pigmented dorsal capsule, and as often as this was opened it was found to con- tain spermatozoa. To these free Hectocotyli are to be added the Argonauts which carried a Hectocotylus-axm inclosed in its sac. All the specimens, which were carefully examined, contained semen either in the sac of the Hectocotylus-arm or in the testis. Respecting these animals, indeed, doubts might be raised as to the identity of the species with the common Argonaut, on • Hectocotylus Tremoctopodis questionably. f If these be different, there will be four species. + Verany (p. 54) cites a solitary statement of Leach that he had found a male Argonaut. i; ARGONAUTA ARGO AND THE HECTOCOTYLI. 7^ account of the want of both the expanded arms and the shell. But, in the first place, the similarity of the Hectocotylus-arm with the free Hectocotylits speaks for that identity. Next, the mode in which they occur : during a few days only did I obtain the animals with the Hectocotylus-SiYm and Argonauts of the ordinary kind in any quantity together. The latter were partly larger ones with shells, partly also not more than 2-4 lines long, and these, like the males, had no shell, at least as I obtained them, while the expanded arms were already easily recognizable. The colour and remaining form of the body were, however, in such close agreement with those of the male animals *, that the presence of the expanded arms and afterwards of the shell must be considered not to be specific, but to be sexual differences. It is perhaps not without importance, that these vela with a kind of mesentery on the twisted axis of the arms occur in females of that species whose males carry an arm so excessively deve- loped, which however belongs to a different pair from the vela of the female. The same probably holds good of Tremoctopus ; for in this, as I shall elsewhere show, the two upper arms have not the shape which is commonly figured, but form, when, as rarely happens, they are well preserved, elongated lobes, which excite as much surprise by their enormous size as by the extra- ordinary magnificence of their coloration. It may be expected that the male Tremoctopus may have such a structure, as to have been described as some other kind of Octopus, Since Cuvier says nothing of any difference in those Octopods which lodged the Hectocotyli from the other which carried the Hectqcotylus as an arm, it seems that the male, as which the latter no less than Verany^s specimen of Octopus Ca- rena must be regarded, in this case is not strikingly different from the female. 3. The anatomical agreement of the suckers, &c. of each Hec- tocotylus with those of the Cephalopod female upon which it occurs, has been especially made out by KoUiker. 4. In like manner the exclusive association of each Hectoco- tylus with its kind of female only. Up to the present time free Hectocotyli have only been found in the society of female Cepha- * Even the bundles of hairs described by Kolliker {EntwicJcelungsgeschkhte d. Cephalopoden) were present as in the females of the size of a hazel-nut. In larger specimens I no longer found them. 80 mOller on the male of lopoda, and indeed, as KoUiker observes, only upon females with ripe ova. I have myself found the free Hectocotylus Argonautcs only upon the inner surface of the shell, or upon the ova, or fixed or creeping upon the animal itself, and I can affirm that among the many Argonauts, less than a nut in size, which I have examined, I never found a Hectocotylus. The Hectocotyli of the Tremoctopus were almost all fixed in the mantle-cavity ; a few crept about in its vicinity externally, or lay at the bottom of the vessel in which the Tremoctopus "was contained; since, as Kolliker has stated, they usually leave the dead animal. 5. Direct testimony that the Hectocotyli play the part of males to their female Cephalopoda is afforded by two observations of a perfect copulation in Tremoctopus. Upon the 2nd of August two large specimens of TremoC" topus were brought to me at the same time, each of which carried in its mantle-cavity a Hectocotylus, fixed as usual in the neigh- bourhood of the gills. Upon pouring water on them, it w^as seen that the penis of each was inserted far into the opening of the right oviduct. Both of the Hectocotyli moved vivaciously, and appeared to be very angry that their endeavours w^ere disturbed. Since it was late in the evening I was obliged to defer further examination until the following morning, when I found both in situ, but dead. Both Hectocotyli were distinguished by the length of the penis ; on endeavounng to draw it out of the ovi- duct it was held pretty fast, and if let go was retracted again for a certain distance ; one might so allow half an inch of the penis to glide in and out. This resulted from a very elastic filament, which, from the point of the penis, projected in deeper ; it could be drawn out for an inch from the opening of the oviduct with the penis, and then when it finally gave way it slipped back again. In both cases this thread did not exactly enter at the point of the penis, but somewhat behind ; and then in its further course it could be clearly identified as the inner part of the for- merly-described ductus deferens. The right oviduct of the Tremoctopus possessed, besides the chambered gland, two dilatations whose walls were greatly soft- ened. The external enlargement was little larger than upon the left side, and contained, together with mucus, merely a portion of the thread torn off from the penis, which has been mentioned. The second larger expansion contained the very singularly con- ARGONAUTA ARGO AND THE HECTOCOTYLI. 81 structed continuation of this ; I will only remark that there hung therein a solid white reniform body of some lines in diameter which consisted wholly of spermatozoa. These were quite similar to those which are usually found in Hectocotylits Tremoc- topodis, and it is therefore unquestionable that these Hectocotyli also serve to fecundate the female Tremoctopoda. The observation of peculiarly-formed masses of spermatozoa far back in the compartments of the oviducal gland itself, was repeated in many specimens of Tremoctopus, and it seems al- most as though this gland had at least in part the function of a spermatheca ; though indeed its relations in other Octopods do not well agree with this view. On the other side of the gland I found no semen, neither in the two copulating Tremoctopoda nor in others ; but I will so much the less question the possi- bility of its penetration as far as the ovarian capsule, as the por- tion of the oviduct between the gland and the ovary is remark- able for its conspicuous ciliary epithelium. A similar epithelium invests also the folds of the ovarian capsule itself which converge towards the internal aperture of the oviduct ; it is found there over a considerable space, and finally extends through the so- called water-canal describeci by Delle Chiaje and Krohn in Tremoctopus and Eledone, which reaches from the posterior side of the ovarian capsule towards the lateral compartment. For Argonauta I can bring forward no such complete obser- vation, yet the copulation and fecundation by the penetration of the appendage of Hectocotylus Argonautce into the female sexual aperture become very probable from the following facts. The ovarian capsule of an adult Argonaut contained a filiform body, which, from its form, from the lobe at the thicker end, and from its minuter structure, was certainly the torn- off append- age of a Hectocotylus Argonautce. Attached to it w ere very diffuse masses of spermatozoa in lively motion. In another very large Argonaut I had sought in vain for Hectocotyli. After cutting up the intestines, and especially the generative organs in many directions, I found in the water used for washing the parts, three filaments, which were the appendages of so many Hectocotyli* Functionally, then, the appendage of Hectocotylus Argotmuta is to be compared to the penis of H. Tremoctopodis, although perhaps the appendage is not always intended to reach the ova- SCIEN. MEM.— A^a^ Hist, Vol. I, Part I. 6 82 Mt^LLER ON THE MALE OP rian capsule, and may have been a result of some accident to the Hectocotylus. The state of polygamy in which many females of these Cepha- lopods live is worthy of remark. Cuvier (Laurillard) found three Hectocotyli in the mantle of an Octopus ; Kolliker, among twelve Hectocotyli of the Tremoctopus, once found three together^ and twice, two ; Von Siebold found among three, two together ; I among eighteen found four together once, and three times, two upon one specimen. 1 also met with two Hectocotyli upon one Argonaut, twice. Since it is improbable that the Hectocotyli can pass from one female to another, either the number of the males must be greater than that of the females, or many of the latter must alto- gether despair of the society of the former. On the other hand, it seems that the many Hectocotyli for one female are ovk aTrO" (jxokLot, as Homer says of the evval aOavdrcov. In the oviduct of one Tremoctopus were found two distinct, but for the rest almost identical seminal masses, each with its appended tubular filament, and many such fragments appeared to indicate some- thing more than bigamy. This is perhaps connected with the manner in which at least a portion of the Cephalopoda here referred to lay their eggs. The ova of Tremoctopus and Argonaut a are, it is well known, found in groups, each of which is attached to a delicate stalk. These stalks are in the Argonaut fastened to the convoluted por- tion of the shell ; in Tremoctopus to a principal stem some lines thick. The ova of each such single group are as a rule at about the same stage of development ; whilst the different groups vary so greatly, that in the larger bunches we frequently meet fresh- laid ova in company with perfect embryos. In this case a re- gular progression may frequently be detected, so that develop- ment gradually advances from one end to the other of the whole bunch. In one bunch of ova of Tremoctopus there was further- more this distinction between the two ends of the principal stem : that the end which carried the most fully-developed em- bryos had a brownish, wrinkled, old appearance ; while that in which the ova were undeveloped was clearer, smoother, softer, and fresh-looking. Between the two were transition-states. The size of each group which has a distinct thinner stalk answers ^ thi Kse] ARGONAUTA ARGO AND THE HECTOCOTYLI. 83 pretty closely to the quantity of ova which we often find in Trem- octopus in a dilatation of the oviduct, which, on the outer side, immediately follows the gland. The ova which we find in the part of the oviduct before the gland, as well as in Argonauta in the beginning of the oviduct, are as yet only provided with their separate proper stalks, which are delicate, but already somewhat ng ; these become united in the outer part of the oviduct into a group with a common stalk, and thence they probably remain here for a considerable period. It may therefore well be, that the different groups of a large bunch of eggs are attached at dif- ferent times, and although the duration of the intervals is wholly unknown, it may be imagined to be not very small, and perhaps too considerable to allow of the fecundation of all the ova by a single previous copulation. Such ova belonging to different periods might be fecundated by many Hectocotyli at different times. Inasmuch as it has been said that the Argonaut is hermaphro- dite, I beg expressly to observe that nothing which I have ever noticed favours this conclusion. In the male specimens the testis lay where otherwise the ovary would be found, and of the latter there was no trace ; whilst in a female of 3 lines long it was already very perceptible, and characterized microscopically by ova of 0*02 of a line in diameter. Besides, the want of the velum upon the arms of the Hectocotylus-hearers show^s that these are quite separate individuals from the females. If now with regard to two kinds of Hectocotylus the anatomical fact is established, that they are developed as arms of perfect Cephalopods, and also that all three Hectocotyli very frequently occur isolated, there is a question which promises to be one of a more general interest, viz. What is the relation of the free Hec- tocotylus to the animal from which it has detached itself ? 1. That the Hectocotylus stands still less in the relation of a parasite (Cuvier) to the animal as whose arm it is developed, than to that in whose mantle it resides, is clear. I will only call to mind, how from the very first all observers have brought forward the striking similarity to a Cephalopod-arm ; they have not, however, come to the readiest conclusion, that it is such an arm, without many deviations. k2. That Madame Power also wrongly imagined the Hectoco- 6* 84 mCller on the male of tylus Argonaut(B to be a vermiform embryo of the common Argo- nauta [vide supra) has been already shown by KolUker. 3. Kolliker brought forward formerly the view that the Hec- tocotyli as male individuals are the independent equivalents of the female Cephalopods. According to the present state of knowledge, this view can be tenable only under two suppositions. Either, one must assume an alternation of generations in its wider sense between the Hectocotylus and its previous supporter ; or, after the separation of the Hectocotylus from the rest of the, body, it must be regarded as the representative of the individu- ality, having thrown off the remainder as so much no longer useful ballast. Against the former view of a kind of alternation of genera- tions however, too many objections at once arise ; among others^ the development in the place of one of the eight typical arms ; the imperfect organization as regards the other generations; further, that the alternation would take place merely in the males; whilst the females o^ Argonaut a and Tremoctopus are known to lay eggs from which individuals similar to them pro- ceed. Lastly, the presence in Argonauta of a testis with per- fect semen, which probably passes thence into the Hectocotylus, opposes altogether the hypothesis that the latter is a male gene- ration proceeding from an asexual gemmiparous one. For the other supposition, that the Hectocotylus, together with its producer, forms only one animal, but after its separation must be regarded as a continuation of the whole, because it is the means of propagation of the species, analogous cases might be adduced of certain animals in which the organs of individual life retrograde in relation to those of the propagation of the species. It might be instanced that many Echinoderms, for example, are produced by budding from larvae, w hich then waste away ; and herewith might be compared the surprisingly great and rapid development of the Hectocotylus-arm, as contrasted with the small size of the other arms*. But before drawing such comparisons, further observations must be obtained upon the duration and mode of Hfe of the two separated moieties of the • In my specimens not more than six pairs of suckers are distinctly developed upon three arms. ARGON AUTA ARGO AND THE HECTOCOTYLI. 85 original animal. For we know as little how long the isolated Hectocotylits, as how long the seven-armed Cephalopoda lives ; whether the latter produces new Hectocotyli*, or passes through yet other metamorphoses. In this respect it is remarkable that male individuals of only a very small size relatively to the females have been observed, and that, independently of sex, they exactly resemble the very young females. The circumstance also that many of the small animals had a tolerably advanced mass of «emen in the testis, and that others already carried perfect semen in the seminal sac of the Hectocotylus-arm, rather indicates that these males do not grow large, for the ova of females of the same size are nowise developed to the same extent. If large male Argonauts occur, they have been without doubt overlooked in consequence of their wanting the expanded arms and the shell. The specimens of the Octopus described by Verany are indeed considerably larger ; and Cuvier says nothing about the size of the animals which bore the Hectocotyli either in the mantle or as an arm. Yet the case of an Octopus, brought for- ward by Verany, which, in the usual place, had merely the pe- dicle without an arm or vesicle, is the only direct evidence for the continued existence of the Cephalopod which has cast off the Hectocotylus, Until more light is thrown upon these relations, it seems unnatural to assume that all the most important organs of an animal, the central organs of the circulating and nervous systems, the apparatus of sense and digestion, and so forth, are thrown off €71 bloc, and that the remainder with the semen, which is not even produced in it, should continue to represent the individual. 4. If for the present then the Hectocotylus can hardly be considered to be an entire individual, it only remains to regard it as a separated portion of the whole, Costa has expressed the view that the Hectocotylus Argonaut the ductus deferens. In both the seminal mass is fixed to a spiral band contained in a sheath, and its uncoiling appears to be connected in the one case as in the other with the extru- sion of the semen. The substance of the spermatophora, as of the ductus deferens, is a mass which exhibits transitions from a less to a more considerable consistence, and also from a complete absence of structure to a striation, which, however, is not produced by any peculiar elementary parts. The substance of which the capsules and pedicles of the ova are formed is similar, and in the oviduct of Tremoctopus we find masses, of which it is not easy to say how much proceeds from the Hectocotylus and how much from the female herself. Should this analogy, upon which I will not enter further here, be confirmed, the Hectocotylus Trem- octopodis might at most be called a spermatophore-bearer. In any case, the Hectocotylus of the Argonaut (and probably also the two others) stands in the relation to the rest of the animal, of an arm, which is at the same time penis and ductus deferens. When separated it may be compared with any other part, which separated from a living individual, yet preserves for a certain time a given amount of vital properties. How far, as regards amount and duration, this may extend, cannot be deter- mined a priori, and the Hectocotyli may perhaps surpass idl hitherto known instances. Nothing can better illustrate the character of their movements, than that Laurillard, Delle Chiaje and KoUiker were led thereby ARGONAUTA ARGO AND THE HECTOCOTYLI. 8? to hold them for decicledl}'' independent animals, and every future observer will be unable to avoid the same impression*. The circulation of the blood of the Hectocotylus, although its course is only imperfectly known, is very lively and rhythmical. It should be observed, that in detached arms of Tremoctopus a rhythmical movement of the veins from the periphery to the centre continued for half an hour after separation from the body, although the animal had been dead for an unknown time. The protracted contractility of separated portions of the Cepha- lopoda, for example of the skin with the chromatophora, is also already known. Yet in the whole male Argonaut the Hectoco- tylus-arm was the part in which the reflex movement ceased latest, since it continued to make apparently voluntary move- ments for many hours after these had ceased in the rest of the animal. How long the movement, and indeed the existence of the Hectocotyli endure after their natural separation is altogether unknown t, but probably for a considerable time, if copulation be not effected ; if indeed they do not exist afterwards. The presence of the appendages described as gills in Hecto- cotylus Tremoctopodis is very remarkable : since they do not occur in other Cephalopod- arms and -fTec^oco^y/i, and since, like the penis, they appear to become still larger in detached Hecto- cotyli, it is to be concluded that the Hectocotylus in question is originally intended to have a longer separate existence. But the other Hectocotyli also are evidently by no means torn off accidentally, but from the manner of their occurrence as well as * Verany mentions in comparison the gill-processes of Eolidse, which, when detached, continue to move for many hours. t Since it is not easy to keep Cephalopoda with Hectocotyli in confinement long enough, it will be desirable to pay particular attention in future to their occurrence at particular seasons of the year. Verany obtained the Octopus Carena at different seasons ; Kolliker obtained the Hectocotylus Tremoctopodis in August and September pretty frequently ; that of the Argonaut again but rarely. I found most Argonauts without Hectocotyli before the end of Sep- tember, but at that time and in the beginning of October the majority of the large specimens possessed them. In the end of July and the beginning of Au- gust I obtained Tremoctopoda pretty frequently, and usually with Hectocotyli ; on one occasion there were eight of the latter in one day. Subsequently the Tremoctopoda occurred only singly, and no longer contained any Hectocotyli. Hence the deficiencies in my account of the //. Tremoctopodis, since I erro- neously expected always to obtain the same supply as at first. 88 MULLER ON THE MALE OF the structure of their place of attachment, they are intended to be detached *. In all probability, finally, the thin appendage of the Hectocotylus in Argonauta and Tremoctopus finds its way into the female sexual aperture only after the separation, for we find almost all Hectocoiyli filled with semen, and the penis o^ Hectocotylus Tremoctopodis often is in an apparently virgin condition. This is easily possible, considering the lively twisting movements which the appendage in both Hectocoiyli makes, even independently of the rest of the body, and in Hec- tocotylus Ti^emoctopodis it is rendered still more easy by the cir- cumstance that in the other portions of the penis the epitheUum forms a multitude of recurved hooks, the hinder edge of each cell riding on its neighbour. It is to be considered, however, whether preliminary acts of coition do not first determine the separation of the Hectocotylus from the rest of the animal. It is remarkable enough, anatomically, that certain Cephalopod males should be distinguished from those of the immediately allied species by the presence of the Hectocotylus-axm. ; but the facts adverted to render the relations of the detached Hectoco- tylus so peculiar, that one is forced either to remain in doubt, or to come to the conclusion that the line of demarcation between independent animated beings, and such as are not so, is by no means so distinct as the schools draw it. It is, however, hardly the time at present to draw any theo- retical conclusions, when so many matters of fact remain to be inquired into with regard to the known species of Hectocotylus (and there may be others), by which perhaps all that has been done may be upset again ; for what has been stated here can only indicate in what direction future investigations must be undertaken. I thus sum up the chief results : — 1. Perfect male Argonauts occur distinguished from the fe- males, which only have hitherto been known, by the absence of the expanded " vela '' upon the two upper arms. 2. These male Argonauts carry the Hectocotylus Argonaut ce * It is important to know whether changes in the size and form of all Nee iocotyli occur after their separation ; whether, for example, the coalescence of the everted edges of the skin in Hectocotylus Argonautce happens before or after separation from the rest of the animal. In my free specnnens the pig- mented capsule was in all cases fully formed. ARGONAUTA ARGO AND THE HECTOCOTYLI. 89 (D. Ch.) in a pedunculated sac in the place of the third arm of the left side. 3. The thick end of the Hectocotylus is fixed to the pedicle, while the thin coiled-up part is free. 4. By the bursting of the sac and the eversion of its edges, the pigmented capsule on the back of the Hectocotylus arises. 5. The testis lies in the abdomen of the whole animal, the external aperture of the ductus deferens being near the point of the HectocotyluS'2ivm, whose thin appendage has at the same time the function of a penis. 6. In the axis of the Hectocotylus there lies a chain of ganglia. 7. It is not to be supposed that the Hectocotyli are developed as vermiform embryos in especial bunches of ova. 8. The Hectocotylus Octopodis of Cuvier, which Verany has shown to be the arm of an Octopus, is mainly distinguished from the Hectocotylus Argonautce by its size, by the presence of a capsule at the free end, and by its development as the third arm on the right side of the Octopus, 9. The Hectocotylus Tremoctopodis of Kolliker is distinguished by its gills, by a peculiar structure of the ductus deferens, and by the want of a pigmented dorsal capsule ; but it possesses a ganglionic chain in its axis, its penis is a more delicate prolon- gation of this, like the appendage of the Hectocotylus Argonautce, and its cleft abdominal capsule is to be compared to the lobe upon the appendage of the latter. 10. The Hectocotylus Tremoctopodis is thence to be considered analogous to the two other Hectocotyli, although the animal as whose arm it is developed is not at present known. 11. Each Cephalopod with a Hectocotylus-^rm. is to be re- garded as the male of the corresponding female Cephalopod. 12. The Hectocotyli are intended to separate from the rest of the body, and are then received and lodged by the female. 13. In this condition they have apparently independent mo- tion and circulation ; they contain perfect semen ; and in Trem- octopus, as also probably in Argonauta, a copulation with the female animals takes place. 14. The Hectocotyli are not comparable to the spermatophora 90 MULLER ON THE MALE OF of other Cephalopoda; but perhaps the so-called ductus de- ferens in Hectocotylus Tremoctopodis is similar to these. 15. The free Hectocotyli can by no means be regarded as independent animals. Note by Professor Kolliker. I desire to take this opportunity of stating, that I have con- vinced myself of the truth of the most important of the discoveries made by M. Miiller, by examining the Cephalopods which he has brought, and that I entirely agree with his view of the re- lation of the Hectocotylus Argonaut (R to the male Argonaut. As it now appears, I was led formerly to put too great a value upon the statements of Maravigno and Madame Power, and I was thence induced to consider the Hectocotyli as male Cepha- lopoda which were developed as such in the ovum. It appears now that I was indeed right in the main, when I claimed the Hectocotyli as belonging to the Cephalopoda ; but that they are not complete animals, but only parts of them, separated indeed in a very strange manner, and by the great independence of their organization and vital manifestations forcibly resembling inde- pendent animals. Kolliker. EXPLANATION OF FIGURES 1 AND 2 OF PLATE I. (Both figures are magnified somewhat more than four times.) * Represents the natural size. Fig. 1 . The perfect male Argonaut seen from the left side : the numbers in- dicate the pairs of arms ; the second and fourth arms of the left side are thrown back, in order to show in what manner the sac containing the Hectocotylus is fixed by its pedicle in the place of the third arm. Over the exterior of the sac a ridge extends longitudinally. Fig. 2. A male Argonaut in the same position, only the Hectocotylus has come out of its sac. The sucker-bearing portion is twisted once completely round upon its own axis, so that it is seen at first from one side, then from above, then from the other side, then in the ascending portion directly from below, and again upon the same side as at first. The fixed end of the Hectocotylus is still covered by the pigmented mem- brane of the sac ; further on the latter is torn longitudinally upon the ARGONAUTA ARGO AND THE HECTOCOTYLI. 91 sucker-side towards the mouth, and so inverted in consequence of the Hectocotylus being bent back, that one looks upon what was previ- ously the inner surface of the sac : the chromatophora glimmer through only indistinctly. The margins of the cleft lie in the con- cavity of the first flexure ; one margin passes before, the other behind the thick end of the Hectocotylus ; both unite at * on the dorsal side. Between the edges and the white streak which indicate the seminal sac there is a pit, whose inner surface is formed by what was pre- viously the outer surface of the sac. Where the sucker-bearing part of the Hectocotylus passes into the filiform appendage (penis), the lobe appears on the back, and from this on each side a fold passes on to the appendage. [T H. H.] 92 SIEBOLD ON HECTOCOTYLUS. Article III. A few Remarks upon Hectocotylus. By C. Th. von Siebold, Professor at the University of Breslau. [From Siebold and Kolliker's Zeitsckrift for June 1852.] I HAVE read with the greatest interest the recent discoveries of Verany and H. Miiller as to the true nature of the Hectoco- tyli. I have now, with Kblliker, arrived at the persuasion that Madame Power, through the too great positiveness with which she described the development of the Argonauta in the egg, has partly been the cause of the hitherto erroneous views that have been entertained upon the subject. Since Maravigno, in fact, only reported upon the communications concerning Argo- nauta made by Madame Powder to the Academy of Catania, it is difficult to say for how much share in the error he is responsible by additional careless observations of his own. From the first I was desirous to have a sight of the figures which Madame Power added to her treatise, and which neither Oken, Creplin, Erichson, nor Kolliker had as yet seen. I availed myself of my last visit to Vienna to examine Madame Power's treatise in the Atti deir Accademia Gioenia di Scienze Naturali di Catania (torn, xii.), contained in the Imperial Library, and especially to convince myself of the resemblance of the figures of Argonauta embryos given by Power, with Hectocotylus, I found that the complaints made by Oken (Isis, 1845, p. 617) about the careless editing of these academical papers were fully justifiable, since even in this Viennese copy of the twelfth volume the illustrative plate of Madame Power's essay was wanting. In accidentally turning over the leaves of some of the succeeding volumes, how^- ever, I came in the fourteenth volume upon the missing plate. Figs. 1-4 represent, somewhat magnified, but very coarsely executed, a something which has a remote resemblance to a Hectocotylus ; one distinguishes an elongated clavate body, one end of which runs out into an acute point, and whose thicker end is provided laterally with a double series of indistinct pro- SIEBOLD ON IIECTOCOTYLU9. 93 minences. Figs. 1-3 exhibit five or six such elevations upon each side. Fig. 4, on the other hand, has ten upon each side. Fig. 4 then differs from the three preceding figures only by the increased number of the lateral elevations, and yet Madame Power says (see Wiegmann's Archiv, 1845, vol. i. p. 378, or Oken's Isis, 1845, p. 610) of this fourth form, which she sup- poses to be an embryo three days old, that from this stage elevations like buds gradually arise, provided with a double series of dark points, and that these are the commencement of the arms and their suckers ; where, however, in fig. 4 these com- mencements of the arms are supposed to be, I can by no means comprehend ; for this body, described and figured as an embryo three days old, reminds one only of the arm of a Cephalopod with its double row of suckers. Had Kolliker chanced to see these figures, he would certainly have still more strongly be- lieved that the Hectocotylus actually leaves the egg in its proper form. Now that Verany and H. Miiller have drawn attention to the ex- ternal sexual difFerencesof the Cephalopoda, the different accounts given by Aristotle of the sexual distinctions and functions of the Octopus acquire an especial value, since Aristotle appears to have been acquainted with the natural history and internal structure of the Cephalopoda to an extent that we must even now con- sider astonishing. From the following passages, w^hich 1 here extract verbatim from Schneider's translation {Aristotelis de Animalibus Historice, book x.), Verany and H. Miiller, who have produced a new phase in the history of Hectocotylus, will learn with astonishment, that Aristotle may fairly contest with them the priority of their discovery of the relation of the male Octopus to the HectocotyluS'QXva, In fact, in book iv. chap. 1, 6 (Joe. cit.), it is thus written : — ^' Polypus (such is Aristotle's inva- riable designation for Octopus) brachia sua ad officium cum manuum tum pedum accordat: namque duobus, quae supra OS habet, admovet ori cibum. Postremo autem omnium, est hoc inter cetera acutissimum et solum aliqua parte candidum in dorso (vocatur autem dorsum pars brachii laevis a qua prorsum acetabula collocata sunt) et in extremo bifidum hoc igitur ad coitum utitur." In the fifth book, chap. v. 1, we find further: — "Aiunt non- 94 SIEBOLD ON HECTOCOTYLUS. nulli, marem habere non nihil simile genitali in uno ex brachiis, quod duo maxima acetabula continet; id protendi quasi nervo- sum usque in medium brachium atque totum in narem (funnel) foeminae inseri." In the same book, chap. x. 1, lastly, Aristotle returns once more to the sexual distinctions of the Cephalopoda in these words : — " Differt a fcemina mas capite (abdomen) oblongiore et id quod genitale vocant piscatores habet in brachio candidum/^ The task now remains for those observers who have the op- portunity of investigating that portion of the Mediterranean which lies between Greece and Asia, to decide what species of Octopus Aristotle understood by his ^' Polypus,^' and how far his acquaintance with the sexual relations of the male Octopus coincides with the history of the Hectocotylus as it has been recently made known. [T. H. H.] H. VON MOHL. OX CELLULOSE. 96 Article IV. Investigation of the question : Does Cellulose form the basis of dll Vegetable Membranes ? By Hugo von Mohl. [From the Botanische Zeitung, vol. v. p. 497 et seq^ In a former paper* I laid down the anatomical and chemical reasons which led me to persist in maintaining the doctrine of the growth of the membrane of the elementary organs of plants propounded by myself and attacked by various writers,, and induced me to reject the view, defended by the Utrecht professors, Mulder and Harting, that the outermost layers of those mem- branes are the youngest and the inmost the oldest. Since that paper was written I have carried through a long series of new observations for the further elucidation of the conditions here in question, the results of which, so far as relates to the chemical f characters of vegetable membranes, 1 believe may be published with advantage, because they may serve to throw light upon some points as yet unknown, and to refute the chemical evidence brought forward by Harting and Mulder in favour of their view. In the Essay already referred to, I closely discussed the opposition presented by the deductions drawn by Mulder and Harting on one side, and by myself on the other, from the known reaction of cell-membranes when acted on by sulphuric acid and iodine. My opponents are of opinion that the circumstance of thin recently formed membrane being coloured blue by iodine and sulphuric acid, while in many full-grown cells only the inner layer manifests this reaction, while the outer are tinged yellow by these two substances, gives ground for the deduction that these outer layers have been formed subsequently to the others, and that the inmost layers of the full-grown cells are the same membranes which constituted alone the wall of the young cell. On the other hand, I asserted that this conclusion is too hasty, * Botanische Zeitunff, vol. iv. p. 337 et scq. Translated in the Annals of Natural History, vol. xviii. p. 145 et seq. t I shall speak of the anatomical conditions on another occasion. 96 II. VON MOHL ON CELLULOSE. since a particular layer of an elementary organ may undergo a chemical metamorphosis in the course of time, without experi- encing on that account any alteration in dimensions, or affording cause for it to be regarded as a new layer in an anatomical sense of the word : I stated that in respect to this metamorphosis we have to consider two possibilities, since it might arise either through the cellulose of which the layer was originally composed becoming dissolved and replaced by some other chemical com- pound, or through the persistent cellulose becoming saturated by another compound, and hence losing the capability of reacting with iodine and sulphuric acid. For various reasons I declared the latter view, which certainly offers the most glaring contra- diction to the views of the chemists, to be the more probable^ but I could not distinctly prove it, because I was unable at that time to extract the infiltrated matters from such membranes as offered an obstinate resistance to the action of sulphuric acid and iodine, and in which cellulose could not be demonstrated to exist by the application of those reagents; such a process of extraction being necessary to render the cellulose (which I assumed to form the basis of the membranes) accessible to the action of the iodine. Now, as the following pages will show, I have succeeded in this with all the elementary organs of vege- tables, and 1 therefore assert most distinctly that the walls of all the elementary organs of plants are composed of cellulose ; that it is quite inadmissible to draw conclusions as to the period of origin of any given layer of their walls from its chemical condi- tions, and that in regard to this question anatomical evidence alone is valid. In order to establish this proposition, I am compelled to enter somewhat minutely into the details of my investigation: if I enter into more extended explanations of the methods I followed than seems to many altogether necessary, this is to be attributed to the circumstance that I only arrived at determinate conclusions after many unsuccessful experiments, and I wish others to be able to confirm the correctness of my views. Cuticle stands first of all the structures, in which it is impos- sible to demonstrate a trace of cellulose by iodine and sulphuric acid. It either completely withstands the action of sulphuric acid, or, if it undergoes a certain degree of softening by this acid. H. VON MOHL ON CELLULOSE. 9? this never causes iodine to produce a blue colour in the substance of the membrane, but the latter is always coloured yellow or brown by the application of those reagents. The results are very different when cuticle is subjected for some time to the action of caustic potash. For this purpose a thin section of some epidermis possessing a thick cuticle, for example that of the leaf of Aloe obliqua, must be kept from 24 to 48 hours in a strong solution of caustic potash, between two slips of glass, at ordinary tempe- ratures. The solution I used was so concentrated, that crystals of hydrate of potash were formed when the temperature of the room fell to freezing-point. It is not requisite that the potash should be chemically pure. When the action of the potash on the cuticle was strong enough, the microscope revealed that numerous little drops exuded from it, of a tenacious fluid, not mixing with the solution of potash, but becoming yellow with iodine. The cuticle itself was somewhat swollen up and showed itself (like the membrane of thick-walled cells treated with sulphuric acid) to be composed of numerous superimposed lamellae, which were not continued uninterruptedly from one cell to another, and hence did not form a connected layer lying upon the epidermis and distinguishable from it, but terminated at the boundaries of contiguous epidermal cells and formed part of their walls. In most cases the epidermal cells had expanded somewhat in the direction of their breadth, and the segments of the cuticle corresponding to the individual epidermal cells had become more or less perfectly separated from each other. When a few drops of a strong tincture of iodine* are applied to the preparation, the latter dried, and then wetted with water, the cuticle acquires as bright a blue colour as the walls of the epi- dermal cells and the subjacent parenchyma. The purity of the colour is increased, as in most cases in which a cellular membrane is coloured blue by iodine without the application of sulphuric acid, by allowing the preparation to dry up once or twice after being saturated with iodine and wetted with water, and then * In these, as in all the following researches, I used a tincture of iodine, in preparing which I added an excess of iodine, so that part of this remained undissolved at the bottom of the alcoholic tincture. The attempt to apply a tincture made with sulphuric aether instead of the alcoholic, so as to gain time by the rapid drying, was not accompanied by good results, since this tincture did not wet the preparation so perfectly as that made with alcohol. SCI EN. MEM.— Aa/. Hht. Voi,. I. Part II. 7 98 H. VON MOHL ON CELLULOSE. wetting it again with water, and if requisite also a second time with tincture of iodine. A similar result is obtained by cutting off a section of the cuticle parallel to the surface of the leaf, and treating it in the same manner with potash and iodine. The cross-sections of the side-walls of the epidermal cells, composed of cuticular substance, exhibit exactly the same appearance as the other thick-walled cells, composed of a number of superimposed layers ; between them runs an external membrane common (?) to the two adjacent cells, which has frequently a yellow or greenish colour at the first wetting with water, but becomes likewise blue after a repetition of the operation. When the cells have separated from each other, this outer membrane is torn irregularly and hangs in fragments attached to one or other of the contiguous cells. The cuticle of other fleshy or leathery leaves, for instance of Aloe margaritifera, Hoy a carnosa, Hakea pachyphylla, Hakea gib- bosa, &c,, behaves exactly in the same way as that of Aloe obliqua. These statements place it beyond doubt that the cuticle of the leaves mentioned is not a homogeneous layer of substance dif- ferent from cellulose, secreted upon the surface of the epidermis, but that this membrane is composed of separate segments corre- sponding to the epidermal cells ; that it is composed of numerous superincumbent lamellae of cellulose ; and that its chemical diversity from cellulose arises from the infiltration of a substance, coloured yellow by iodine, which not only resists the action of sulphuric acid itself, but protects the cellulose which is saturated with it from the influence of sulphuric acid and iodine. The result is the demolition, in regard to the thick cuticle of thick- walled epidermal cells, of the evidence advanced on chemical grounds against the view I formerly advanced [Vermischte Schriften, p. 260) of the structure of cuticle, namely, that it is not a coating over the epidermis composed of substance excreted from the latter, but owes its origin to a metamorphosis of a portion of the outer w alls of the epidermal cells. The infiltrated substance was not totally extracted by macera- tion of the epidermal cells for 24-48 hours in solution of potash, for the addition of sulphuric acid to the preparation saturated with iodine immediately reproduced the brown colour which is caused in cuticle by these reagents before it is treated with H. VON MOHL OX CELLULOSE. 99 potash^ and the same phaenomenon is seen in the wood-cells when the solution of the infiltrated substance is imperfect. While the above-described action of potash on the cuticular layer of epidermal cells is taking place, a very thin pellicle becomes detached from the outer surface of the epidermis, either in strips, or, when the epidermal cells separate from each other, a piece of this coat remains adherent upon the outer side of each of them. This pellicle is not coloured blue by iodine, but always yellow. When the above treatment is applied to the epidermis of organs in which the outer wall of the epidermal cells is not much thicker than their side walls, and in which iodine and sul- phuric acid demonstrate only a very thin cuticle, for instance, of the leaves of Iris fimbriata^ of stems of Epiphyllum truncatuniy of the petiole of Micsa, &c., a thin yellow pellicle remains here also on the outside of epidermal cells coloured blue by iodine. If the epidermis is boiled with solution of potash, this pellicle shrinks together and by longer boiling is completely dissolved, while the epidermal cells only swell up and acquire a beautiful blue colour with iodine. This pellicle, which occurs under all circumstances upon the epidermis, whether it be a portion of its cells transformed into cuticle or not, consists, judging from its different behaviour with potash, of a substance essentially dif- ferent from the cell-membrane, and is doubtless the same mem- brane which Ad. Brongniart separated from leaves by maceration, and denominated cuticle. It has been confounded, by myself and others, with that portion of the walls of the epidermal cells which is coloured yellow by sulphuric acid and iodine, under the name of cuticle, because the methods of investigation hitherto in use afforded no means of separating these two different parts distinctly from each other. But it is evident that this membrane must be distinguished from the subjacent cells ; I therefore pro- pose to restrict the name of cuticle to it alone, and to apply the term cuticular layer to that part of the epidermal cells which becomes coloured yellow by sulphuric acid and iodine. The cuticle exists on all cells exposed to the air, without exception ; if any one choose to ascribe it to a secretion of the epidermal cells, I have no objection to offer to the notion ; but it would be difficult to bring forward any proof of its correctness ; and perhaps we ought to regard the circumstance that this cuticle 7* 100 H. VON MOIIL ON CELLULOSE. is marked with raised lines in many plants, as a proof that it is not to be considered simply as a hardened excreted fluid, since possibly those lines ought to be looked upon as a proof of definite organization. The researches of Mulder and Harting have made known that sulphuric acid and iodine do not demonstrate the existence of cellulose in cork any more than in cuticle ; any one may readily convince himself of the correctness of this statement in the cork of the cork oak, of the elder, &c. The cells also of the nascent suberous layer, in their earliest condition, w^hile still covered by the epidermis, exhibit the same yellowish brown colour on the application of the said reagents, as developed cork, even in those plants in which the cork never attains any considerable develop- ment, for instance in Cereus peruvianus. The conclusion drawn from the absence of a blue colour, that the membrane of cork- cells contains no cellulose, and is composed of a peculiar substance, is, on the other hand, just as groundless as in the preceding case ; for a thin section of the cork of the cork oak, which has been boiled in solution of potash until the brown colour it originally assumes has disappeared, acquires as bright a blue with iodine as any other membrane composed of cellulose ; in like manner, and also by the application of nitric acid in the way described below, the corks of Sambucus nigra, Acer campestre, Ulmiis campestris, and Euonymus europceus, show that their cells are composed of cellulose. It is well known that the layer to which I have applied the name of periderm, is, in anatomical respects, to be regarded as a modification of the cork-layer. This circumstance led me to conjecture that this membrane would display chemical characters similar to those of cork. This was confirmed. I subjected the periderms of the oak, of Crataegus Oxyacantha, Betula alba, and Plosslea floribunda, to the action of a boiling solution of potash, after which iodine produced the blue colour. The blue colour was quite clear in the oak and Crataegus, but in the other two less pure; the periderm of Pl'Osslea, indeed, did not require a very long boiling in solution of potash to produce the blue colour, but only isolated patches of the cells were coloured pure blue, the greater part acquiring a dirty blue tint. The periderm of the birch, which very obstinately resisted the action of the potash, H. VON MOHL ON CELLULOSE. 101 required a long continuance of the boiling before the iodine would produce the blue colour. The organs above mentioned, forming the surface of plants, especially the cuticular layer of the epidermis and the cork of the cork oak, and in a less degree the corks of the other plants enumerated, stand, in respect to the chemical properties of the substances combined with their cell-membranes, preventing the reaction of cellulose, in opposition to all those elementary organs which form the internal tissues of plants. In these also the re- action of cellulose is very frequently partially or entirely pre- vented by compounds combined with them, but caustic potash ia not the proper means of reducing the cellulose of these organs into a condition capable of the reaction. Even when, as in many cases, for example in the secondary layers of many wood-cells, such as those of the wood of Buxus, potash has this effect, the result is often waited for in vain, and in case of success, the mem- branes boiled with caustic potash but seldom acquire a pure blue colour with iodine, and a yellow or brown colour is mostly min- gled with it. On the other hand, the application of nitric acid is always attended with complete success. The effect of this acid is perhaps most perfect when the plant to be investigated is allowed to macerate for a long time in dilute acid^ at ordinary temperatures ; but since in solid woods, when even small fragments are placed in the acid, several months or even a year may easily elapse before the effect of the acid has completely developed itself, this method is scarcely applicable when a large series of observations is to be made with this agent. I therefore substitute a boiling of the substance to be examined in moderately strong acid, for the long-continued maceration ; by this means the desired effect is very rapidly obtained, only in many plants the risk is run of dissolving the cell-membrane, or at least some of its layers, by continuing the boiling too long. This inconvenience, however, may be avoided by a little care ; for in general the colour of the vegetable structure affords a mark by which we can detect whether the required effect of the acid has commenced and the boihng is to be stopped, or it is to be continued for a longer time. At first, the acid ordinarily pro- duces in the fragment of vegetable substances placed in it, immediately it is heated, a yellow or brown colour, accompanied by considerable effervescence and frequently with the formation 102 H. VON MOHL ON CELLULOSE. of vapours of nitrous acid, which colour however soon gives place to a pale yellow, or a complete bleaching of the preparation. When this bleaching takes place, the desired effect is generally attained. I then brought the preparation, if it had not already been boiled between two sHps of glass*, on to a glass slider, washed it with water, either dried it perfectly by a moderate heat, or saturated the acid with ammonia, wetted the dried pre- paration with strong tincture of iodine, let it dry up in the air, and wetted it once more with water for microscopic examination and to produce the blue colour. Sometimes it was necessary to repeat the wetting with tincture of iodine, or to moisten the pre- paration, saturated with iodine, with water, and let it dry again several times. The whole process is somewhat tedious, but time is saved, by setting to work with a number of preparations at once, and, when they are saturated with tincture of iodine, letting them dry at leisure, by which means we obtain sufficient material for the whole day's work at the microscopic investigation. I have frequently, to save time, assisted the drying of the prepara- tion saturated with iodine by artificial heat ; as a rule, however, it is more advantageous to allow the evaporation of the iodine to go on at the ordinary temperature of the room, since even a slight heating may easily cause too strong an evaporation of the iodine. It is well known that the parenchymatous cells of succulent and young organs, in which the membranes are imbued with a comparatively small quantity of the compounds coloured yellow by iodine, require no preparation to render them capable of as- suming a blue colour with iodine (see my Vermischte Schriften, p. 344). It is different with the parenchymatous cells of older structures which have become saturated with encrusting sub- stances, for instance, the cells of pith and of medullary rays, &c. Very frequently these cannot be coloured blue at all, or only very imperfectly with iodine alone, and do not assume a pure blue tint even with iodine and sulphuric acid, but are tinged with * In all cases when it is desired to boil a thin cross section of a vegetable structure in nitric acid, it is advisable to place it in a few drops of acid upon a slip of glass, to lay a thickish covering-glass over it, and then place the glass slip (to avoid cracking it) upon a metal plate which can be heated until the acid boils. The most delicate preparations may he boiled in this manner, while they almost inevitably tear up into fragments when it is attempted to boil them in a glass tube or a platinum spoon. H. VON MOHL ON CELLULOSE. 103 a dirty blue, so that it remains doubtful whether cellulose forms any considerable part of their substance, or even whether it is present at all, at least in some of the layers. These circum- stances very readily explain why Mulder, who thought he had an unerring and very delicate test for cellulose in the application of Sulphuric acid and iodine, was of opinion that the pith of Sambucus nigra, for example, was composed of cellulose only in the earHer conditions, and of a peculiar substance in the full- grown state. But the matter turns out quite differently when this pith is treated in the manner above described with boiling nitric acid, for it then assumes a beautiful indigo-blue colour with iodine. While iodine and sulphuric acid produce, if not a bright blue, at least a green colour in old parenchyma-cells, and thus the presence of cellulose is placed beyond doubt even by this method of investigation, the brown cells which surround the vascular bundles of Ferns usually resist sulphuric acid as obstinately as even cuticle, and it is absolutely impossible to demonstrate the presence of cellulose by its help. In the essay above referred to, I endeavoured to show on anatomical grounds, that this mem- brane in the Ferns is produced by the transformation of a cellu- lose membrane, but was compelled to leave undecided whether or not it still contained cellulose in its fully developed condition. The application of nitric acid affords an easy means of deciding this question, and furnishes the proof that this membrane is only prevented from reacting with iodine by its combination with some infiltrated substance. For instance, if the black membrane surrounding the vascular bundles of the petiole of Aspidium filix mas, is boiled in nitric acid until its dark brown colour is changed into bright yellow, iodine imparts a beautiful blue colour to the membrane of these cells, the texture of which is not changed in the slightest degree. In many cases, in the Ferns, other portions of their cellular tissue are so saturated with foreign compounds that they do not react with iodine and sulphuric acid. Among these, for example, are the outer layers of the dark brown petiole of Adiantum pedatum, on the cells of which the said reagents do not act at all at first ; only when the acid has been in contact with the cells for twenty-four hours does the blue colo\ir which shows itself at 104 H. VON MOHL ON CELLULOSE. the circumference of the preparation lead to the recognition of the presence of cellulose in those cells, while the membranes themselves remain yellowish brown. Here again a short boiling in nitric acid suffices to render the membranes capable of taking a very beautiful blue colour. In particular parts of the cellular tissue of Polypodium percussum, the outer coat of the parenchymatous cells acquires a yellow colour with iodine and sulphuric acid, while the inner layers swell up and become blue; in a word, they behave in this respect like the outer coat of wood-cells. In preparations boiled with nitric acid the cells are coloured blue throughout; therefore in these also cellulose is the basis of the outer layer, resisting sulphuric acid. Such cells, resisting the action of sulphuric acid, are more common than would be supposed from what has generally been stated; for many thick-walled parenchyma-cells, in the same way as many wood-cells, assume only a yellow or at most a greenish tint with the said reagents, as is the case in the parenchyma- cells of many Palm-stems, e.y, of Calamus, of Cocos botryophora, in the thick-walled pitted cells of the pith and rind of Hoya carnostty in the stony cells of the winter pear, &c. All these cells assume a bright blue colour with iodine after they have been boiled with nitric acid ; the statement of Mulder, that the thick-walled pith-cells of Hoya contain no cellulose, is conse- quently without foundation. Since nitric acid is capable of rendering the cellulose ac- cessible to the reaction of iodine in cells which more or less obstinately withstand sulphuric acid, it may readily be imagined that this acid never fails us in common parenchyma-cells, in which sulphuric acid and iodine readily produce a blue, when we desire to impart the blue colour to such cells by means of iodine. This always presents itself in the greatest purity, and without requiring the continuance of the boiling long enough to alter the texture of the cell-membrane in the shghtest degree. When it is desired to facilitate the anatomical investigation of cells by the production of this blue colour, for instance to exa- mine minutely their pits, which always appear far more distinct in the blue-coloured cells, this method is far preferable to the application of sulphuric acid, from the very fact that it does not H. VON MOHL ON CELLULOSE. 105 cause any change in the texture of the cells. The blue colour shows itself uniformly, whether we investigate thin-walled, still succulent cells, such as the rind-cells of woody plants or herba- ceous vegetables, the cells of the parenchyma of the leaf and petiole, or the dead cells of the pith or medullary rays of old wood. The walls of epidermal cells saturated with cuticular substance, and the cork and periderm of many plants, only, are inaccessible to the action of nitric acid ; in this latter respect, however, the cells of the periderm and cork of other plants form an exception, since cellulose can be demonstrated in them not only by potash but also by nitric acid, for instance in the peri- derm of Plosslea, the cork of Sambucus nigra, Acer campestre, Euonymus europceus, and Ulmus campestris. Yet in these cases it is generally necessary to boil the preparation for a long time in the acid, and the effect is mostly imperfect, as these parts do not usually become coloured blue completely after this treat- ment; there exist, however, some structures belonging to the cork system in which nitric acid is capable of producing a perfect blue colour, while only a greenish tint is obtained by the appli- cation of caustic potash, for example the spines belonging to the suberous system of Bombax and the corky rind of the rhizome of Tamus Elephantipes. The cell-membranes which assume a blue colour after the boiling with nitric acid usually form a very permanent combina- tion with iodine. While in other cases the iodine which has com- bined with a thin section of vegetable structure usually evapo- rates entirely or in great part if the preparation is exposed for a couple of days to the air, and may be extracted in a few seconds with alcohol, preparations boiled with nitric acid and saturated with iodine may often be left lying for weeks in the air without the colour becoming perceptibly paler. In particular cases the iodine combined with the membrane obstinately resisted not only exposure to considerable heat^ but the action of almost absolute alcohol heated to the boiling-point. By alkalies, on the contrary, especially by caustic ammonia, the iodine may be very quickly extracted from the membrane. Only the cells of a few of the plants which I investigated, in particular those of the petiole of Cycas revoluta, formed exceptions to this rule, that the iodine combined very firmly with the membrane. t 106 H. VON MOHL ON CELLULOSE. In all the cases which I examined, the parenchyma-cells as- sumed a pure blue colour through the entire thickness of their membrane, and no yellow -coloured outer coat could be detected at the boundary between the cross- sections of the walls of two contiguous cells. In like manner the membrane which closed the pits always exhibited a pure blue, or, in cells which were dried and then assumed a violet colour, a bright violet, and here also no trace of a yellow membrane lying between the cells could be seen, as may be very distinctly observed in the cells of the petiole of Cycas revoluta, especially on account of the large size of the pits. When such cells, coloured blue by iodine, for instance those of the pith of the elder, of the medullary rays of Buxus, the paren- chyma-cells of the stem of Calamus, the cells of the petiole of CycaSi &c., are placed in dilute sulphuric acid, their membranes swell up strongly and finally dissolve entirely, becoming more or less completely bleached in the process. Under this treatment an extremely delicate yellow pellicle comes to light at the boun- laries between the cells, to which pellicles small yellow-coloured granules (or drops of a fluid substance ?) are in most cases ad- herent. We are here reminded of an analogous structure de- scribed by Mulder and Harting under the name of the "outer cell-membrane,'^ and the ^'cuticle of the wood-cells." The question then arises whether this pellicle possessed the yellow colour at the time the cells were coloured blue with iodine, or was at first coloured blue Uke the inner layers of the cell-walls, the yellow colour having been only produced subsequently by the combined action of sulphuric acid and iodine. I regard it as more probable that the latter was the case; for if that pellicle possessed a yellow colour before the operation of the sulphuric acid, we should have an indication of its presence in the trans- verse slice of the walls of two contiguous cells, in spite of its very slight thickness, and its yellow colour would cause a greenish discoloration of the thin bright blue membrane which closes the pits. Yet all my efforts to discover even a trace of such a yellow membrane, with the application of the strongest objectives, which gave a perfectly faultless image with a great amount of light, were without the least success. Both this circumstance, and above all the observations on the outer membrane of many parenchyma- cells, to be mentioned below, lead me to conclude H. VON MOUL ON CELLULOSE. 107 that this outermost membrane of the parenchyma-cells also is composed of cellulose and is coloured blue with iodine, but that nitric acid is incapable of extracting the infiltrated matters com- pletely, whence results the insolubility of this membrane in sulphuric acid, and the yellow colour which it acquires from this acid. In this respect this membrane would bear a resemblance to the cuticular layer of the epidermal cells, in which caustic potash, like nitric acid here, is capable of freeing the cellulose from the influence of the infiltrated matters so far that it reacts with iodine, but at the same time does not extract these matters perfectly, and render the membrane saturated with them soluble in sulphuric acid. Among all the parenchyma-cells which I have investigated, those which form the outer part of the pith of a shoot, several years old, of Clematis Vitalba, are perhaps the most interesting in regard to the structure of their walls. These cells have very thick walls, and their membrane is composed of a tolerable number of layers which may be easily distinguished. It is co- loured yellow^ by iodine. Sulphuric acid causes the inner layers to swell up, and at the same time they acquire a green colour : under this operation an outer layer which is on an average t4V4*^ of a line in thickness, remains wholly unchanged. This layer therefore displays the character of Mulder^s " outer cell-mem- brane." When strong objectives are employed, a delicate line is seen running through the middle of this layer, indicating the boundaries of the contiguous cells. In a cross- section of these cells boiled with nitric acid, the inner layers are coloured deep blue by iodine, and the outer layer just described assumes, ac- cording to the amount of action exerted by the acid, a yellow, green, or blue colour. When such a preparation is wetted with dilute sulphuric acid, the inner layers swell up strongly, they are bleached and by degrees dissolved ; the outer layer also swells to some extent, but very slightly, and is bleached, remaining other- wise unaltered. If the preparation is placed in stronger acid, this outer layer is also dissolved, leaving behind an immeasurably thin pellicle (with adherent granules) which lie in the middle of it. It is therefore evident that the outer membrane of these cells, which at the first hasty glance might be taken for the outermost coat, is also composed of cellulose, but that it is in- 108f H. VON MOHL ON CELLULOSE* filtrated either with a greater quantity of that substance, ac- quiring a yellow colour with iodine, with which the inner layers are saturated, or with some compound different from this, opposing a greater resistance to sulphuric acid, which is so far extracted or altered by nitric acid, that the reaction with iodine occurs in the cellulose contained in this layer, but enough re- mains to protect this layer from the action of weak sulphuric acid, while in the outermost, immeasurably thin layer, the resist- ance to sulphuric acid is so great that the latter is incapable of dissolving it. In reference to the question whether the above-described outer brown pellicle of the parenchyma- cells contained cellulose, the behaviour of the primary membrane of the cells of the horny albumen of Sagus ttedigera seemed to me of considerable im- portance, since here I could convince myself not only of the absence of a yellow colouring of this pelHcle, as in the other parenchyma-cells, but positively of its acquiring a blue colour. When the membrane of these cells is treated with a very weak tincture of iodine, it is coloured bright yellow, and the primary membrane deep yellow ; in such a preparation very dilute sul- phuric acid produces a very bright blue in the secondary layers of the cells, and a darker blue in their primary membrane, and under these circumstances, the comparatively great transparency of the cellulose allows one to become quite certain of the purity of this colour and of the total absence of that yellow colour in the outer primary coat. If stronger sulphuric acid is added, the secondary membranes are bleached and gradually dissolved, while the primary membraneturns yellow and becomescoated with fine granules*. Under these circumstances it cannot be doubted that the primary membrane is imbued with a greater quantity of the substance coloured yellow by iodine, but that this is in- sufficient to prevent the appearance of the blue colour under the action of iodine and weak sulphuric acid; yet its presence may be the cause why a stronger acid does not dissolve this mem- brane, for it is always found that a membrane resists the action of sulphuric acid more strongly in proportion as it is more deeply coloured by it and iodine. • The granules, or minute drops (for there is no means of deciding whether they are solid or fluid), arc not separated until the sulpliuric acid begins to act. II. vox MOHL ON CELLULOSE. 109 In reference to the characters of thin membranes, the liber- tells generally approach to the parenchyma-cells, since ordina- rily they do not possess either the great solidity and brittleness, or the dark colour which distinguishes most wood-cells. This greater resemblance between the liber-cells and parenchyma- cells presents itself also in their behaviour to sulphuric acid and iodine, since the former mostly assume a pure blue colour with these reagents. On the other hand, the liber-cells of the arbo- rescent Monocotyledons, especially of the Palms possessing hard vascular bundles, are aUied to the wood-cells of the Dico- tyledons ; they are devoid not only of the softness and flexibility which distinguish the liber-cells of many Dicotyledons, but in many cases they exhibit also a yellow colour, sometimes passing into the deepest brown. I examined the liber-cells of three species of Palm, Cocos boti^yophora, Calamvs, and the black- fibred Brazilian palms, the wood of which is used for making walking-sticks. When a cross- section of the liber-bundle of one of these is treated with iodine and sulphuric acid, the secondary layers are dissolved and the outer layer of the cells remains be- hind undissolved, with a brown colour. This outer layer, which, like the above-described outer layer of the pith-cells of Clematis, corresponds, according to the characters given by Mulder, to the " outer wood-membrane " of the w^ood-cells of the Dicoty- ledons, and would be supposed to consist of a substance different from cellulose, exhibits considerable thickness in a cross section, and in Cocos botryophora (where it is about xttW^'^ ^^ ^ ^^^^ thick) is distinctly pitted. For these two reasons we cannot regard it as the primary membrane of the cells, since in all 'cases its not inconsiderable thickness and the presence of pits, which do not perforate the membrane completely, but penetrate only to a certain distance on both sides, lead us to conclude that this layer is composed of several superincumbent lamellae. When a cross section of a vascular bundle of these plants boiled in nitric acid is treated with iodine, exactly the same phaenomena present themselves as I have described in the pith-cells of Clematis, namely, all the layers of the liber-cells, especially the outer one withstanding sulphuric acid, are coloured bright blue, which proves that this latter also is composed of cellulose. If vlilute sulphuric acid is then added to such a preparation, not 110 II. VON MOIIL ON CELLULOSE. only are the inner layers of the cells dissolved, but also the outer, which before this treatment with nitric acid were inso- luble in sulphuric acid, and on the boundary-lines between the adjacent cells, as in the parenchyma-cells, remains a pellicle of the utmost delicacy coated with fine granules. Since in these cells also, when very thin sections are coloured only pale blue by a small quantity of iodine, it is impossible to detect by a yellow colour this pellicle insoluble in sulphuric acid, I consider it pro- bable that this possesses a blue colour so long as it is not ex- posed to the action of sulphuric acid. The presence of this outer thin membrane, and the fact that it is only yellow under the simultaneous action of sulphuric acid and iodine, and blue with iodine alone, may perhaps be still more clearly proved by macerating a vascular bundle of the black palm-wood in dilute nitric acid (which however may require six to twelve months), or boiling it in this acid until the liber-cells are separable by a slight pressure. In this case isolated pieces of variable size of the outer membrane may often be found among the separate liber-cells, and we can then convince ourselves that they are coloured blue by iodine, and only assume a yellow colour when sulphuric acid is added. These are the observations which principally led me to the opinion that the outer membrane of parenchyma- and prosenchyma-cells contains cellulose, since the impossibility of seeing the yellow colour in this membrane in the cross section appeared not perfectly conclusive, on account of its very slight thickness, though at the same time it must be admitted that this circumstance is also of great importance. Passing to the prosenchymatous cells of the wood of Dicoty- ledonous plants, it is well known that cellulose may be demon- strated in their internal layer by means of sulphuric acid and iodine, but that these layers do not usually, it is true, assume a pure blue colour with these reagents, mostly acquiring only a green tint, which leads to the conclusion that cellulose does really exist in them, but that its reaction is more or less obscured by the presence of a yellow infiltrated substance. Even when, as in the wood of Taxus^ the resistance to sulphuric acid is very considerable, the presence of cellulose may be shown by applying a very strong acid, which completely destroys the texture of the cell-wall, and then, by adding tincture of iodine diluted with a H. VON MOHL ON CELLULOSE. Ill gr^at deal of water, the dissolved cellulose (which according to this experiment is dissolved as such, and not as dextrine) is pre- cipitated with a beautiful blue colour. But although these means suffice to demonstrate the presence of cellulose, the application of sulphuric acid is not adapted for the settlement of the question, whether in such solid woods the cellulose forms always the principal mass of the membranes and is only saturated with a foreign substance, or the latter is predominant and the cellulose only a very subordinate constituent. In this case the application of nitric acid removes every doubt, inasmuch as the secondary membranes of all wood-cells become bright blue through their entire thickness with iodine, when they have previously been macerated for a long time, or boiled until disorganized, in nitric acid. The compound, therefore, which Mulder termed the " in- termediate wood-substance,^^ never itself forms the intermediate layers of the wood-cells, but is a material infiltrated into them. Since this result is quite universal, I regard it as superfluous to mention particular examples, and confine myself to touching on a few points which may be doubtfuL One of these refers to the character of the internal membrane which lines the wood-cells in Taxus and Torreya, and of which the spiral fibres running in these cells form part. This inner coat, as was first shown by Prof. Hartig of Brunswick, resists the action of sulphuric acid very strongly and under its influence is coloured yellow by iodine, whence, relying upon these reagents, as Hartig did, one might be inclined to assume that this membrane was composed of a substance quite different from the interme- diate layer, and contained no cellulose. The latter is by no means the case, for the above-described treatment with nitric acid de- monstrates, by the blue colour which this membrane and its fibres assume, that these also are composed of cellulose. - The second point to be adverted to here, relates to the pits, of which it might be doubtful, from the descriptions w^hich Hartig, Harting and Mulder have given of the structure of cells and the characteristics of the " outer wood-membrane," whether the membrane which closes the outer ends of the canals of the pits is always composed of cellulose. In regard to this, there can be no doubt in preparations of the Coniferae which have been treated with nitric acid, since it is found that the pits are closed 112 II. VON MOIIL ON CELLULOSE. by blue, although not brightly coloured, membrane, as I have seen most distinctly in the wood of Taxus baccata and Abies pectinata. But I must expressly remark, that observation of this, as well as of the membrane which closes the pits of the par- enchyma-cells, requires a microscope of the highest quality, fur- nished with very strong objectives ; with objectives not of very short focus, i. e, unless at least less than a line, even when the image they give is perfectly free from error, the membrane closing the pits will be sought in vain, since the penetrating power of the microscope will be too small. The most difficult point in the investigation of- the wood-cells is that of their outermost membrane (Mulder's " outer cell-coat,'' Harting's ''cuticle of the wood-cells"). In the first place, the remark perhaps may not be superfluous in reference to this membrane, that in many wood-cells we meet with a case similar to that in the above-described pith-cells of Clematis and the liber of Calamus and Cocos botryophora, namely, that when a cross-section of the cell is soaked with iodine, layers of two kinds may be distinguished, a thick inner one, very brightly coloured, and a thinner outer one which acquires a darker yellow colour with iodine and might readily be taken for the primary membrane of the cell, e, g, in Buxus, in particular cells of the wood of Erythrina caffra and of many kinds of Ficus. This outer layer withstands the action of sulphuric acid much more strongly than the inner, so that a weaker acid suffices to cause a strong swelling up of the inner layer, bringing out a blue or green colour, while the outer layer is quite unaf- fected and remains yellowish brown. A stronger acid, however, is capable of producing a green colour in the outer layer, or at least of bleaching and dissolving it when the action is allowed to continue for some time. Treatment of a cross section with boiling nitric acid leaves no doubt of the true nature of the conditions, since in a preparation so treated both layers are coloured bright blue by iodine, and dilute sulphuric acid quickly dissolves even the outer layer, so that any confusion of this with the outer cell- membrane is out of the question. The latter exhibits the same qualities as I have described above of the outer coat of the parenchyma and liber; it is extremely thin, withstands the action of sulphuric acid, and H. VON MOHL ON CELLULOSE. 113 is coloured yellow by this and iodine. It is just as difficult with this outer coat of the wood-cells as with the outer layer of the parenchyma-cells, to answer the question whether it is coloured yellow or blue by iodine, but similar reasons render it probable here also that the latter is the case. For if a transverse section of a Dicotyledonous wood which has been boiled in nitric acid is soaked with iodine, a yellow membrane may be detected between the blue-coloured cells ; if the action of the acid has been more powerful, this yellow colour is more and more lost and passes through green into a perfectly pure bright blue, so that no trace of any yellow membrane lying between the blue cells can be seen here any more than in the parenchyma-cells. This is the behaviour, for example, of the wood of Abies pectinata, Larix europceuy Taxus baccafa, Torreya taxifolia, Buxus semper- virens. Viburnum Lantana, Viscum album, Betula alba, Fagus sylvatica, Clematis Vitalba, Erythrina caffra. When dilute sulphuric acid is added, the cell-membranes are dissolved with a slow bleaching, and there remains behind a network of immea* surably thin yellowish- brown pellicles, which correspond to the boundaries of the cells. In most, perhaps in all woods, the intercellular passages running between the wood-cells are filled up by an intercellular substance, which is coloured yellow by iodine and sulphuric acid, and not attacked by the latter, w hence one might easily be led to assume that this substance formed a common mass with that membrane of the cells which likewise acquires a yellow colour under these circumstances. But the incorrectness of such a notion is proved by the examination of the preparations boiled in nitric acid, since in these the intercellular substance retains its yellow colour on the application of iodine, while the outer cell- membrane is coloured blue. Whether the intercellular substance of Dicotyledonous woods is wholly free from cellulose, or contains it in a very strongly combined condition in which it is not acted on by iodine, I cannot yet venture to decide. The application of caustic potash to the investigation of this condi- tion entirely failed me, and the application of nitric acid fur- nished no decisive result. For if the boiling with nitric acid is stopped before the texture of the cells is attacked, the intercel- lular substance, as above mentioned, remains yellow on the SCIEN. UEU.^Nat. Hist. Vol. I. Part II. 8 114 H. VON MOHL ON CELLULOSE. application of iodine ; while if the transverse section of a wood is boiled longer than is requisite to impart to its cells the capa- bility of acquiring a blue colour with iodine, under which cir- cumstance the cells begin to separate from each other, the inter- cellular substance is no longer found, being dissolved. I hoped to find it in a transition state between these two extreme cases, and then perhaps, if it contained cellulose, to be able to detect this with iodine ; and, if I was not deceived, it indeed happened with a few woods, for example, in Buxus sempervirens and Clematis Vitalba, it assumed a bright blue colour after the yellow had disappeared. This colour however was very pale, and might possibly have depended on a bluish tinge thrown upon the bleached intercellular substance by the surrounding dark blue-coloured cells, so that I do not venture to declare the observations certain, and must leave this point undecided for the present. To mention some of the examples in which an intercellular substance with the said properties filled up the intercellular passages between the wood-cells, I may name Larix europaaf Taxus baccata, Torreya taxifolia. Viburnum Lantana, Buxus sempervirens, Clematis Vitalba, When the intercellular substance is dissolved by means of nitric acid, the cells of the wood begin to part from each other. It is difficult to trace accurately the process which occurs here, since the blue colouring of the cell-membrane by iodine, which would greatly facilitate the investigation, does not occur unless the preparation saturated with iodine is allowed to dry up ; but this drying causes a contraction and tearing of the membranes, which places the greatest difficulty in the way of detecting what goes on during the separation of the cells. The notion might readily be formed that the separation of the cells resulted from the solution of their outer membrane, as well as the intercellular substance, by the nitric acid, and that the cement which connected the cells together was thus removed. But if I have rightly understood the operation, it is something quite dif- ferent from this. The outer membrane is not dissolved, as is easily seen when such preparations, saturated with iodine (whether the cells have been separated from each other by boiling nitric acid or by long maceration at ordinary temperatures), are treated H. VON MOHL ON CELLULOSE. 115 with sulphuric acid, under which circumstances the outer mem- brane presents itself unchanged with a yellow colour, after the rest of the cell-membranes have been dissolved. The separation of the cells seemed rather to depend upon the outermost layer of the secondary membranes becoming softened into a gelatinous condition, and detached from the primary membrane. The cir- cumstance that the outer layers of the cells are caused to swell up and dissolve in strong acid sooner than the inner, is not un- commonly met with, especially in treating with sulphuric acid half gelatinous cells saturated with iodine, for example, the half collenchymatous cells of the bark, like those of Erythrina caffra. In such cases it is very common for the outer layers of the swollen cells to have a brighter blue colour than the inner, and when by a longer action of the acid these inner layers also become per- fectly blue, the outer layers are completely bleached. A similar phaenomenon sometimes presents itself most distinctly under a strongish action of nitric acid. This is especially the case with the wood-cells of Clematis Vitalba, the outer layers of which, when the boiling in acid is long continued, dissolve into an amorphous jelly, which acquires a blue colour with iodine. Such a perfect solution of the outer layers, however, is by no means necessary to bring about the separation of the cells; even a slight softening of the cell-membrane seems sufficient to separate the secondary layers from the outer membrane, and thereby the cells from each other. In favour of this, we have both the microscopic examination of cross sections which have been boiled in nitric acid till the cells have separated, and in which fragments of delicate torn membranes, but no amorphous jelly, are met with between the cells, — and also the circumstance, that in the wood of Abies pectinata, which had been macerated for about a year in dilute nitric acid, and in which the elementary organs fell apart on the slightest pressure, the canals of the pits were closed at the outer ends by a thin membrane, which could not have been the case if the outer membranes of the cells had been dissolved. The extraordinary softness which the cell- membranes had ac- quired, both in this wood and in the hard vascular bundles of the black-fibred palm above-mentioned, when treated in the [aame M-ay, was remarkable. In this separation of the cells, the outer coat never seemed to 8* 116 H. VON MOHL ON CELLULOSE. split into two lamellae remaining attached to the two adjacent cells, but the membrane situated between two cells remained un- divided, separating from both cells, or remaining attached to one of them, which of course must have been accompanied by a dis- ruption of it in other places. When exactly the right period of the action of the acid in which the wood is macerated, has been hit, comparatively large pieces of the outer coat may be often obtained isolated, by tearing up a piece of such wood with a needle, since the cells may be readily extracted, as from a shell, from the chambers formed by the cavities of the outer coat. Proceeding to the Vessels, the different layers of the walls of the forms containing spiral or annular fibres behave in the reverse way to those of cells. In the latter, namely, when they are greatly lignified, the outer layers are usually saturated most strongly with foreign compounds, and therefore offer the greatest resistance to the action of sulphuric acid, while the inner layers, as the youngest membranes, are frequently coloured a beautiful blue by iodine and sulphuric acid ; in the vessels, on the con- trary, the secondary structures (the fibres) are those which most strongly resist sulphuric acid, and only acquire a yellow or at most a green colour, while the tube on the inner wall of which the fibres are deposited may acquire a bright blue with these reagents. This difference is seen very beautifully in the vessel- like elementary organs, furnished with flat, band-like spiral fibres, of the wood of many Cactaceae, especially of the Mammillarice. When these elementary organs are treated with boiling nitric acid, both the fibre and the outer coat are coloured bright blue. In like manner the spiral fibres of the vessels of herbaceous plants, for instance of Asparagus, may be coloured bright blue after treatment with nitric acid ; in the vessels of many plants, however, especially in the spiral vessels of Sambucus nigra and the scalariform vessels of Tree Ferns, a rather long-continued boiling in the acid is requisite to destroy the green colour and bring out the pure blue. The pitted vessels of the Dicotyledons approach the wood- cells nearer than the spiral vessels in their behaviour with iodine, since it is their outer layers which are principally infiltrated with the foreign compounds coloured yellow by iodine. But treatment with boiling nitric acid also produces the blue colour in all the H. VON MOHL ON CELLULOSE. llj layers here, and not only in the thickened layers of the walls of the vessels, but in the delicate membrane which closes the pits. This was the behaviour, for example, in the vessels of Sambucus nigra, Viburnum Lantana, Asclepias syriaca, Buxus sempervirens. Clematis Vita/ba, Betula alba, Quercus Robur, and Tilia. The outermost membrane of these vessels behaves in every respect like the outer cell-coat of the prosenchymatous cells of wood, and similar reasons to those which testify that the latter contains cellulose are furnished also by the outer membranes of vessels, so that I may refrain from entering into minute details on this point. The researches of Mulder and Harting have already made known that the wall of the Milk-vessels contains cellulose ; and in like manner I only think it necessary to state briefly that the elementary organs of that part of the vascular bundles of Mono- cotyledons which 1 have described under the name o^ proper vessels {vasa propria) in the Palms, and elsewhere, acquire a beautiful blue colour with iodine after treatment with nitric acid. Looking back over the researches here described, it appears clear that the walls of all the elementary organs of vegetables may be brought, by the action of caustic potash or of nitric acid, into a condition in which they assume a blue colour with iodine, and that the only exceptions to this among all the solid structures of plants, are the cuticle, in the strictest sense of the term, and perhaps the intercellular substance of the higher plants. The action exerted by potash or by nitric acid upon vegetable membranes is not merely transitory, enduring only while the action is kept up (as Mulder assumes of the action of sulphuric acid upon cellulose), but is permanent, inasmuch as the mem- branes which have been treated in the above-described way retain the capability of taking a blue colour with iodine after the active substance has been completely removed, as when the nitric acid is neutralized by ammonia. The question now arises, whether the cellulose itself suffers a transformation by the appli- cation of these means, rendering it capable of taking a blue colour with iodine, in the manner starch does, or, whether these means act merely to extract or decompose more or lesj perfectly the foreign compounds combined with cell-wall, which acquire a yellow colour with iodine, and deprive cellulose of the 118 H. VON MOHL ON CELLULOSE. power of reacting with iodine. These questions must be answered by chemists, and not by botanists. Yet I may be permitted to allude to two points. A conversion of the cellulose into starch is out of the question, for the cell- membranes treated in the way described remained afterwards, as before, insoluble in boiling water; if therefore a transformation of the cellulose is assumed, it must be into some compound as yet unknown, the characters of which have still to be minutely investigated. For the present the assumption of such a transformation appears to me unwar- ranted, in so far that the power of the cell-membranes to acquire a blue colour with iodine is the only reason at present existing which can be urged in favour of this, while I have already shown, on previous occasions, that fresh, perfect, unaltered cell-walls of many plants, particularly of young organs, also acquire a blue colour with iodine, which indicates that cellulose possesses this quality, as well as starch, in and by itself, whenever its reaction is not prevented by other compounds united with it. For the present, and until we obtain further explanations from the chemists, it will be simplest to assume that the potash and nitric acid bring about this reaction by removing such foreign com- pounds from the encrusted membranes. Looking from anatomical and physiological points of view, and these alone I have occupied in my researches, I believe that I have shown, from the latter, that the reaction of sulphuric acid and iodine upon cellulose is altogether devoid of the trustwor- thiness ascribed to it, and that the assumption based upon this supposed trustworthiness, — that particular layers are formed of other compounds besides cellulose, in the course of the develop- ment of the elementary organs of plants, and that the chemical reaction of the different layers of a vegetable elementary organ therefore affords a certain test of the relative epoch of its deve- lopment,— is totally devoid of foundation ; — that, consequently, all the objections against my theory of the development of the walls of vegetable cells, built upon this basis, are completely un- tenable, and that in this question anatomical evidence alone can be admitted as proving anything. [A. H.] VERANY AND VOGT ON THE HECTOCOTYLI. 119 Article V. Memoir upon the Hectocotyli and the Males of certain Cephalo- pods. By MM. J. B. Verany and C. Vogt. [^Annates des Sciences Naturelles, t. xvii. No. 3, 1852.] 1 HE manner in which the fecundation of some Cephalopods, especially of the Argonaut and of the Tremoctopods, whose females only have been known up to the present time, occurs, is a zoological question of the highest interest. The very recent re- searches of MM. Kolliker and Siebold having called the atten- tion of naturalists to this point, we have neglected no opportunity of procuring fresh and living animals, by the study of which we hoped to arrive at a definite solution of the problem. We ven- ture to think that our investigations have been rewarded by complete success, at least for one species. The memoir which we now lay before the Academy relates principally to the Tremoctopus Carena (Verany) and the Hecto- cotylus which is derived from it. We shall give, first, a historical summary of the labours of our predecessors ; then the zoological description of the species which occupies us, and which was to a great extent unknown up to the present time ; lastly, we shall conclude our work by a detailed study of the reproductive organs. Historical Introduction. M. Delle Chiaje, at Naples, described and figured in the year 1825* a little animal found by him parasitic upon an Argonaut, to which he gave the name of Trichocephalus acetahularis. This animal, when once detached from the Octopus, to which it was adherent, swam and crept at the bottom of the water with uneasy movements. It appeared full of Hfe for many hours. M. Delle Chiaje did not doubt that this parasite was one of the Helminthic worms, and placed it in the genus Trichocephalus of * Memorie sulla Storia e Notomta degli Animali senxa vertehre del regno di Napoli; di Stefano Delle Chiaje. 120 VERANY AND VOGT ON THE IIECTOCOTYLI Rudolphi, although it was provided with a double series of suckers, wishing, as he says, not to burden science with a new- genus. M. Delle Chiaje only describes the exterior of the ani- mal. According to him it has a long, round, filiform, very con- tractile proboscis, attenuated towards its extremity. The body is provided with a double series of alternating and retractile pedunculated suckers, by which the animal adheres to the skin or to the shell of the Argonaut. M. Laurillard discovered at Nice, upon the Octopus granulosus of Lamarck, five specimens of a parasitic animal, which Cuvier described subsequently as the Hectocotylus Octopodis^. Among these five individuals three were found in the funnel of a female Octopus, one was discovered in the same position in another Octopus, and the fifth individual " had attached itself to an arm of the Octopus, and had transformed it into a kind of sac, into which it had introduced its head, the remainderof its body being external and free.^' In recurring to this individual, Cuvier adds, " The Hectocotylus has attached itself to one of the arms, which it has even almost destroyed, and which it seems to replace in such a manner, that at first sight it might be taken for the arm itself" With our present knowledge, we must conclude from these remarks that one of the Octopods taken by M. Lau- rillard was a male, which had just disengaged its hectocotyli- form arm from the sac in which it had been developed. Cuvier describes the external form of the body, the suckers, and the internal organization. None of his Hectocotyli had the filiform organ, which we shall call the flabellum (lefouet), everted from the sac which contains it. Cuvier found at the rounded extremity of the body an alimentary orifice leading into a sac, closed on all sides, aaid having a yellowish internal surface. This cavity, which Cuvier calls a stomach, is nothing more than the sac opening by a cleft, whose formation we shall subsequently describe. Besides this stomachal sac, the clavate extremity contains another " with stronger parietes, occupied by the innu- merable folds of a thread which has the colour and brilliancy of raw silk. One of the Hectocotijli ejected this thread very rapidly at the moment of its capture." Cuvier is disposed to look upon this thread as connected with generation. It is in fact the • Annates des Sciences Naturelles, tome xviii. 1 829. AND THE MALES OF CERTAIN CEPHALOPODS. 121 seminal thread contained in a spermatophore. Besides these organs, Cuvier describes the muscular tube forming the axis of the body, and continuous with a filament folded up in the ter- minal sac, — a filament which we have called the flabellum. According to Cuvier, the point of this flabellum is bent back into the body of the animal, and is directly continuous with the seminal thread. We know not to what circumstance we must attribute this mistake, for in all the specimens we have exa- mined, the flabellum has been found completely free at its ex- tremity. The note which was published at a later period by M. Costa of Naples, upon the Hectocotylus Argonautae^, only added to the controversy the opinion of the author, who looks upon this animal as the spermatophore of the Octopus. M. Costa^s de- scription is besides altogether incorrect, and the figure is as bad as it can be. The sac of the flabellum is represented like a pennant ; the extremity of the true spermatophore as a tenta- cular cirrhus with two points ; the convolutions of the seminal thread appeared to M. Costa to be spots formed by little spiral vessels. M. Dujardin, in arranging the Hectocotyli among the doubtful Trematoda, thus gives his opinion upon these parasites f : — " I have seen the anatomical preparations preserved in the Museum, as well as an entire specimen ; but I confess I cannot compre- hend what the thing can be ; I am only clear that it is not a Trematode worm. One might call it an arm torn from some other Cephalopod, so similar is the double series of suckers oc- cupying the ventral surface of the Hectocotylus to the larger suckers of the Octopus : the internal structure is equally mus- cular, but there is visible in the dorsal portion a long white sinuous and folded thread, which Cuvier could detect only after the action of spirit, and which therefore should proceed from the coagulation of some liquid (spermatic?) substance '' " It can only be by the study of these objects in the living state that we can decide upon their true nature, and determine if they be not the portions of some Cephalopod detached in order * Note sur le pretendu parasite de VArgonauta Argo {Ann, d. Sc, Nat, 2« s^rie, t. xvi. 1841). t Dujardin, Hutoire Naturelle des Helmlnthes ou Vers Intestinaux {Suites a Buffon, 1848). 122 VERANY AND VOGT ON THE HECTOCOTYLI to subserve fecundation. I can only state, that the long white thread described by Cuvier, whose length is more than a metre, is merely a bundle of very long and delicate, independent fila- ments, closely resembling the spermatozoa of the Cephalopoda." M. Kolliker, who during his stay at Messina in 1842 found upon many female Argonauts the Hectocotylus described by M. Delle Chiaje, originated a new epoch in the history of this ani- mal. M. Kolliker discovered in addition a new species of Hec- tocotylus upon the Tremoctopus violaceus, and obtained some fifteen specimens of it. In examining their structure he very soon convinced himself, to use his own words, that these sup- posed worms are nothing more than the males of the Cepha- lopods we have just mentioned. After having communicated his ideas upon this subject to the Italian Scientific Congress, sitting at Genoa, M. Kolliker pub- lished a note in the ' Annals of Natural History,' vol. xvi. 1845. The Manual of Comparative Anatomy, by M. von Siebold, being in course of publication at the same time, M. Kolliker entrusted three specimens of his treasure-trove to M. von Siebold, accom- panying them with a manuscript note containing the results of his observations. Upon his part M. von Siebold examined the Hectocotyli, and, while confirming many of the conclusions of M. Kolliker, he differed from him with respect to other points of structure. Notwithstanding these discrepancies, M. von Siebold adopted the views of M. Kolliker, w-ho considered these Hectocotyli to be the males of the Octopods upon which they are found. The observations of M. von Siebold are contained in his excellent ^ Comparative Anatomy of the Tnvertebrata*,' which is now also extensively circulated through France in a good translation. Lastly, M. Kolliker published a very elabo- rate memoir, accompanied with figures, in his Second Report of the Zootomical Institution at Wurzburg, for the year 1849t. M. Kolliker begins by describing at length the form of the Hectocotylus of Tremoctopus violaceus, specimens of which he met with upon almost all the females of this Cephalopod, which he foimd in August and September at Messina. He distin- * Lehrhuch der vergleichenden Anatomie der Wirhellosen Thiere ; von C. Th. von Siebold. Berlin, 1848. t Bericht von der Koniglichen zootomischen Amtalt zii Wurzburg. Leipzig, 1849. AND THE MALES OF CERTAIN CEPHALOPODS. 123 guishes externally two sets of suckers, branchiae in the form of villosities, and an oval abdomen from whose aperture the penis makes its exit. The skin is composed of two layers ; of an epi- dermis with polygonal cells, and of a corium composed of inter- woven undulated fibrils, in the midst of which contractile pig- ment-cells are deposited — chromatophora such as are met with in all Cephalopods, without exception. The muscular system is composed of bundles belonging to the suckers, and besides of a very strong muscular tube, which serves as a support to the whole body, and in the interior of which is found a cylindrical cavity almost entirely filled by a longitudinal tube which M. Kol- hker calls the intestine. The muscular tube itself is composed of three layers of fibres, the median being longitudinal, whilst the other two are circular. M. Kolliker is unable to give any complete account of the nervous system ; but having seen under the microscope a true ganglion containing six ganglionic cor- puscles, he is certain that nerves exist, and he believes that a fine white thread, which he once met with upon the upper sur- face of the intestine, is truly a nervous cord. M. von Siebold, on the other hand, remarks, that he has found in the axis of the body of the Hectocotylus of the Tremoctopus a nervous cord whose greatly developed ganglia correspond in number with the lateral suckers. M. von Siebold does not believe that the Hec- tocotyli have any digestive system. M. Kolliker speaks doubt- ingly : he describes a longitudinal tube situated in the centre of the muscular tube, which it almost wholly fills. This tube is composed of tw o membranous layers ; according to his account it is closed posteriorly, and terminates perhaps by a fine aper- ture visible only in the recent animal and situated at its ante- rior extremity. The tube contains nothing but conical masses, arranged regularly, and exactly corresponding in number to the suckers. M. Kolliker adverts also, as a supplement to this de- scription of the intestinal system, to the existence of small ellip- tical apertures arranged in line to the number of four or five, upon the ventral surface, below these conical bodies contained in the intestine. These little apertures are prolonged into as many fine canals which are directed upwards towards the central muscular tube ; but he could not decide exactly if they entered 124 VERANY AND VOGT ON THE HECTOCOTYLI into the muscular tube to communicate with the conical bodies, or if they were merely cutaneous glands. We can easily satisfy the doubts of M. KoUiker upon this matter. His so-called intestine is only a central blood-vessel ; the conoid masses seen by him in this " intestine '^ are the ganglia of the nervous cord situated below the vessel and not within it ; whilst the canals with their fine apertures are only the nerves passing from these ganglia and terminating in the skin. The branchiae described by M. KoUiker are fine fringes com- posed of an epidermis with polygonal cells ; and of a homo- geneous internal membrane, in which a simple network of ca- pillaries uniting into two very small trunks is placed. In the skin of the back are found upon each side two longitudinal vessels, w^ith whose termination M. KoUiker was not acquainted, but which appear to furnish branches for the penis also. Besides these vessels, M. KoUiker found under the microscope, in a fragment of skin with whose exact origin he was not acquainted, an oval tube which he considered to be decidedly a heart, but whose position he could not indicate. The generative organs are very much developed ; they consist of a simple testicle, of an ejaculatory duct, and of a penis. The testicle is a transpa- rent pyriform vesicle, which fills the whole of the abdomen of the Hectocotylus, In the interior of this vesicle there exists coiled up a fine cylindrical thread, without any proper enve- lope, and formed solely by filiform spermatozoa united together. Beside these spermatozoa, granular cells are found in very considerable number. This thread ends freely at its posterior extremity ; but anteriorly it is continued forwards into the effe- rent canal which commences by a clavate extremity, folds back at first in the testicular vesicle, and finally passes into the penis. The efferent canal has a very peculiar structure, for its parietes are very solid, semi-transparent, yellowish, and composed of elastic fibres. The anterior part of the efferent canal is situated in the penis itself, and is traversed in its whole length by a spiral ligament, whose nature M. KoUiker could not determine. M. von Siebold describes the posterior extremity of the Hec- tocotyli as a generative sac, in which the seminal mass, with the copulatory organ, and the efferent canal (provided with horny AND THE MALES OF CERTAIN CEPHALOPODS. 125 tubercles in its interior, which are everted during copulation), which is continued into the penis, are contained. M. KoUiker describes these tuberculosities as little conical spines. Besides the Hectocotylus Tremoctopodis, M. Kolliker describes and figures that of the Argonaut, w^hich is distinguished from the former by the absence of branchiae, of a sac-like abdomen, of a free penis, and by the presence of a filiform appendage which passes from the anterior extremity of its body. At the base of this appendage two triangular membranous lobes are placed. The anatomical structure of this animal differs only in the structure of the filiform appendage and in that of the sexual organs. The filiform appendage is the continuation of the mus- cular tube forming the axis of the body. The testicular capsule is elongated and lined with pig r.ent-cells, which are probably contractile. The testicle itself is formed by a spermatic thread coiled up, and surrounded by a structureless membranous enve- lope. The efferent duct leads into a silvery cylindrical tube, which is probably the penis, and which is certainly of a mus- cular nature. M. Kolliker adds to this description of the Hectocotyli a long dissertation, in which he sets forth the opinions of his prede- cessors, as well as his own, the latter amounting to this, — that the Hectocolyli are the stunted males of certain species of Cephalopoda. " There is no need (he says) for any lengthy evidence that the Hectocotyli are independent animals. He who has not seen, as MM. Laurillard, Delle Chiaje, and I have done, their lively, independent, and continued movements, will not take them for portions of Cephalopods, and still less for spermatophora, if he considers their complicated organization, and if he calls to mind that they have a heart with vessels, branchiae, nerves, and such well-developed generative organs." In the face of this positive assertion of M. Kolliker, we shall undertake to prove that the Hectocotyli are, however, nothing more than detached arms of Cephalopods, merely organized in a special manner. Afler having shown that we were still unacquainted with the male of the Argonaut and with that of the Tremoctopus, M. Kol- liker brings forward, first, the resemblance of structure between the Hectocotyli and the Cephalopoda ; he finds that the suckers. 126 VERANY AND VOGT ON THE HECTOCOTYLI the chromatophora, and the muscular tube of the Hectocotyli, are constructed in exactly the same manner as the corresponding portions of the Cephalopods, and that they are met with exclu- sively among the latter animals. The presence of arteries and veins, of a heart, and of branchiae, as well as the histological ele- ments, are clearly opposed to the union of the Hectocotyli with the intestinal worms. Lastly, M. Kolliker relies upon the ob- servations of Madame Power and of M, Maravigno at Catania, from which it would result that the Hectocotyli are formed as such in the q^^, and that they then resemble a little worm, pro- vided in its whole length with two series of suckers, with a fili- form appendage at one of its extremities, and a little enlarge- ment towards the other. From all this M. Kolliker concludes that the Hectocotyli are perfect animals ; and M, von Siebold has completely adopted this view, although this conscientious ob- server was able to discover neither the intestine nor the heart pointed out by M. Kolliker — essential organs, however, upon whose existence depends in great measure the opinion of these two naturalists. In the meanwhile one of us had been employed in collecting, for many years past, materials which promised another solution of the problem. In his work upon the Cephalopoda of the Mediterranean, M. Verany relates, p. 128, that in 1836 he met with an Octopus, the description of which he published under the name of Octopus Carena. The captured individual possessed, instead of the right arm of the third pair, a vesicle seated upon a little pedicle provided with some acetabula. Upon many individuals collected after this period M. Verany observed, as of constant occurrence, that this same arm was always abnormally developed, and that most frequently the pedicle carried instead of a vesicle, a very large arm terminated by an oval globe, and having the form of the Hectocotylus of Cuvier. Our friend F. de Filippi, Professor at Turin, did in fact recognize this arm to be the Hectocotylus., M. Verany having submitted to MM. Filippi and Leydig (of Wurzburg), the latter a pupil of M. Kolliker, a specimen of T. Carena preserved in spirits of wine ; this arm became detached by the least force, leaving a perfectly clean surface, whilst another arm could only be torn off by violence. The observers having opened the little terminal AND THE MALES OF CERTAIN CEPHALOPODS. 12/ sac, saw the white thread which terminates the Hectocotylus make its exit, like that of the Argonaut. M. Verany drew the following conclusions from these observations: " The Hectocotylits of the Octopus is only a caducous arm of the Cephalopod ; this arm carries the male organs, and in all probability these organs are periodically developed. The Hectocotyli of the Argonaut and of the Tremoctopus differ from that of the Octopus, The Hecto- cotyli of the Argonaut and of the Tremoctopus cannot be the arms of the Octopus which carries them, since they are infinitely smaller, and since, so far as I am aware, these Cephalopods have never been found with an arm wanting.'^ Dr. H. Miiller of Wurzburg visited the coasts of Sicily in 1850, with the intention of studying these contradictory facts. He communicated orally to one of us during his passage to Genoa, that having one day met with a very small Argonaut carrying a vesicle, he had taken this individual for an embryo still retaining its umbilical vesicle, but that on counting the arms he observed there were only seven, and that the eighth was replaced by this vesicle carried upon a little pedicle. Calling to mind the facts observed in the Octopus, Dr. H. Miiller exa- mined the animal which had come into his possession, and found that this little creature which he had taken for an embryo was in fact the perfect male of the Argonaut, and that the Hectocotylus was hidden within the pedunculated vesicle. We still await the publication of the observations collected by Dr. Miiller. As the foregoing extracts show, science has been loaded with a number of observations and opinions but little concordant with one another. Very desirous of obtaining a positive solution of a question which appeared to us to possess the highest interest, we united our efforts to procure a sufficiency of materials, and to study them during life. Fishermen were instructed by the one of us who could best make himself understood in the patois of the country, and the Cephalopods brought to the market in Nice were carefully scrutinized each day. The specimens, which were very rare, of Argonauts and Tremoctopods which fell into our hands, were the object of our minute but fruitless investi- gations. We could not find a single Hectocotylus, although one of us during a residence of two years at Nice examined seven Argonauts and three females of Tremoctopus violaceus. 128 VERANY AND VOGT ON THE HECTOCOTYLI We were in despair, when, in April of this year, the males of Tremoctopus Carena appeared all at once in very great numbers. We give in the following pages the results of our common studies, observing, however, that M. Verany especially occupied himself with the zoological portion, whilst M. Vogt principally took charge of the anatomical investigations. I. Zoology, Tremoctopus Carena. Tremoctopus Carena, Verany. Poulpe Carena, Octopus Carena, Verany. Mem, de V Academic Roy ale des Sciences de Turin, 2^ serie, t. i. pi. 2. Monographic des Cephalopodes de la Mediterranee, p. 34 & 128, pi. 14. figs. 2 & 3 ; pi. 41. figs. 1 & 2. Body sac-like, provided with a constrictor apparatus. Head compressed ; eyes large and projecting. Arms unequal : first and fourth pair longest. Acetabula pedunculated. Funnel very large ; two aquiferous apertures. Tremoctopus Carena, male (PL I. fig. 3. 4.). Body sac-like, oval, slightly acuminated behind, very smooth. Branchial aperture wide, cleft as far as the orbits. Constrictor apparatus formed by a fleshy appendage having the form of an oblique hook, situate upon each side at the base of the funnel, and a horizontal cleft like a button-hole, in the thickness of the skin upon the internal margin of the body. Head moderate, wider than deep, being compressed by the first pair of arms w^hich arise at the level of the orbits ; lateral part of the head occupied almost wholly by the orbits ; inferior portion wholly covered by the funnel. Eyes lateral ; eye-balls projecting and flattened, wholly covered by a transparent membrane which is continuous with the skin ; this membrane is pierced by a slightly contractile circular aper- ture, through which the crystalline lens presents itself wholly without covering. Arms conico-subulate, unequal ; the fourth pair is the longest and measures about two and a half times the length of the body; the first is shorter than the fourth and measures only twice the length of the body ; the second is only half as long as the fourth pair ; the left arm of the third pair is a little shorter than that of the second pair ; the right arm of the third pair is hectocotyli- AND THE MALBS OP CERTAIN CEPHALOPODS. 129 form, and has a pedicle equal in size to the base of the arm of the opposite side ; this pedicle carries a larger or smaller oval vesicle (PI. I. f. 4.), containing the hectocotyliform arm or the per- fectly developed Hectocotylus itself; this hectocotyliform arm is half as long again as the arms of the fourth pair. All the arms are provided with a double series of ace tabula. Acetabula : — these are large, cylindrical, much excavated, pe- dunculate, and distant from one another (PI. I. fig. 3.); they alter- nate from the third, enlarge up to the sixth, and afterwards de- crease progressively as far as the extremity of the arms where they are microscopic. The first pair of arms carries forty-five acetabula ; the second, thirty ; the third, thirty ; the fourth, fifty. The Hectocotylus is inserted upon a pedicle from which it is readily detached (PI. I.fig.3.) ; it is oval below and narrows towards its extremity, which appears as if truncated ; the internal sur- face is flattened, and its whole contour is beset with a very close series of pedunculated acetabula which are oblique, i. e, have an oval section opening a little upon the inner side. These acetabula are united together by a longitudinal membrane which embraces the whole peduncle, and, with the acetabula, forms a continuous line by passing over at the base upon the inner surface of the arm. Generally a little ovoid sac terminates the Hectocotylus. This sac is transparent, and allows a convoluted cord to be seen through its membrane ; the sac is frequently empty, and then the hectocotyliform arm is terminated by a filament or flabellum almost as long as the arm (PI. I. fig. 3.), and which is its conti- nuation. The Hectocotylus carries forty-seven acetabula upon each side. The dorsal part is a little convex upon its base ; this convexity is markedly distinguished by a membranous sac open below (PI. I. fig. 3.). Interbrachial membrane rudimentary, absent between the lower arms. Mouth sun-ounded by two lips, the internal ciliated, the ex- ternal very delicate. Funnel very large, extending far beyond the base of the arms, and measuring three-fourths of the length of the body. Two very large aquiferous apertures, placed at the base of the arms of the fourth pair at the point of attachment of the latero- dorsal portion of the funnel and communicating with the orbital cavity. Excluding the hectocotyliform arm, this Cephalopod is never SCIEN. MEM.— A^a/. Hist. Vol. I. Part II. 9 130 VERANY AND VOGT ON THE HECTOCOTYLI more than 0*110 (metre) in length ; taking in this arm and its fla- bellum, it measures 0*220 (metre). Colour, — During life this Cephalopod has a general transpa- rency which allows the internal organs to be seen through the body, and in the lower parts, even the chromatophora which cover the membrane investing the organs of digestion and gene- ration, and along the arms, the beaded nervous cord. When at rest, the body glistens with iridescent hues of azure, green and purple ; the pupil is of a very brilliant burnished silver, and the dorsal portion of the orbits has a brilliant blue tint shot with a metallic golden tinge. The chromatophora are visible only with a lens, and have a purplish {mauve) tint passing into violet ; when they dilate they are purple, and by the develop- ment of a great number of chromatophora the dorsal surface often passes into a velvety purple. When the creature is irri- tated or out of the water, the chromatophora take an orange-red tinge upon the lower portion ; the points are constantly larger and more scattered than upon the upper. The funnel and the membrane which invests the eyes are equally covered with them. In spirit the chromatophora have always a wine -red colour ; in a saline solution they retain their violet colour. The Hectocotylus is wholly white, without chromatophora: the pedicle which carries it and the vesicle in its interior are covered with them. Tremoctopus Carena, female. Body sac-like, oval, smooth superiorly, very slightly tubercu- lated below. Branchial aperture and constrictor apparatus as in the male. Arms conico-subulate, symmetrical; length as in the male. The first pair is provided with a longitudinal membrane upon its latero-superior portion, and the acetabula of the internal series are united together by a longitudinal membrane, as is observed also in the reticulated Tremoctopus and in the Argonaut. Acetabula moderate, cylindrical, excavato-pedunculated, but little distant from one another ; the first pair has eighty of them, the second seventy, the third sixty, the fourth eighty. Interbrachial membrane, mouth, funnel, and aquiferous aper- tures as in the male. AND THE MALES OF CERTAIN CEPHALOPODS. 131 Colour, — During life and in a state of rest the dorsal part of the animal is of a semi-transparent white colour, strongly shaded with blue ; the inferior portion is white and very iridescent ; the lateral parts of the body, of the head, of the dorsal portion of the inferior and latero-inferior arms, and the iris shine with a silvery lustre ; the internal portion of the arms is of a pale rosy colour. In this state the dorsal portion is covered with micro- scopical chromatophora, and with others which are larger and regularly diffused, of a blue or violet colour ; upon the lower portion of the body the chromatophora are disposed in the same manner, but their colour is violet passing into reddish yellow. The dorsal part of the body, of the head, and of the arms of the first pair assumes sometimes a very brilliant ultramarine colour, shaded with purple ; sometimes this passes into a very dark velvety violet ; the lower parts and the extremities of the arms are then coloured reddish yellow. Three different colours are often seen upon the back at the same time, and it is clouded with white, azure, rose-violet, yellow, and an infinity of dazzling shades produced by the mixture of these tints. Irritated, or in full vigour out of water, the animal covers itself wholly with reddish yellow chromatophora, but the dorsal portion is always strongly shaded with blue. Plunged in spirit, the skin of the lower partof the body becomes very finely reticulated, in consequence of the wrinkles formed by the subcutaneous granulated tubercles. The male Tremoctopus Carena cannot be confounded with any of the known species ; the proportions of the arms, and, above all, of the hectocotyliform arms, clearly distinguish it. The female closely approaches the reticulated Tremoctopus [T. cate- nulatus, Ver.), but the arms of T, Carena are much longer in proportion to the body, and the second and third pairs of its arms are comparatively shorter and less disproportionate to one another in size. The body is oval, while it is ovoid in T. cate- nulatus I finally, the tubercles are proportionally much smaller, more approximated and more numerous in T, Carena. This Cephalopod appears only accidentally upon our coasts, — perhaps, as M. d'Orbigny thinks with regard to all the Philo- nexid(S, because it is pelagic; we have, however, met with it in all seasons, in February, April, September and December, which proves to us that it is a denizen of our latitudes. It is always 9* 132 VERANY AND VOGT ON THE HECTOCOTYLI taken near land by means of drag-nets. Professor Bellardi of Turin, who was at Nice in April of this year, assured us that having been present at the raising of the Madrague nets*, he saw one of these Octopods attached to a dead Salpa, The Octopus evidently approaches the shore at the pairing season, and it is to this cause, as well as to the arrangements we made, that we owe the capture of twenty males and two females which were all taken last April. All the specimens have been brought to us living : a single male had lost its hectocotyliform arm ; all the others presented this arm either wholly developed or still in- cluded within its vesicle. The male by itself has been described and figured since 1846 in the Acts of the Royal Academy of Sciences of Turin, by one of us, and later in the Monographie des Cephalopodes de la Mediterranee which has just appeared. The female has perhaps been described by M. Risso under the name of Octopus tuberculatus, but the short phrase applied to this new species is so vague and diffuse, that it may equally relate to the Tremoctopus catenulatus, M. Risso having left no specimen marked by himself by which it might be decided to which of the two species his description refers, we must retain the name given by one of us in 1836; and so much the more since M. d^Orbigny has already designated Octopus tuherculatus, another species of Octopus previously described by M. de Blainville. M. d'Orbigny does not mention our T, Carena in his Mono- graph upon the Cephalopoda. It is very remarkable that this species, which this year was so frequently taken, and which be- sides does not appear to be very rare, should have escaped Delle Chiaje, who has enriched science with so many new species ; Prof. Philippi,who has so carefully explored Sicily; M. Cantraine, who has traversed a large portion of the Mediterranean ; as well as MM. Kolliker, Riippell, Krohn, and H. Miiller, who have specially occupied themselves with the Cephalopoda in Sicily. However, we have reason to believe that the Octopus granulosus, upon which Laurillard found the Hectocotyli which were de- scribed by Cuvier,is the female of the species with which we are engaged. * Large nets employed in the tunny fishery, established permanently near S.t. Hospice, a league from Nice. AND THE MALES OF CERTAIN CEPHALOPODS. 133 II. Anatomy. We shall first treat of the anatomy of the male, whose orga- nization in general does not differ, in fact, from that of other Ce- phalopods. The ventral region of the mantle is provided with a very complicated constrictor apparatus, fixed to the fibrous mem- brane which invests the intestine. After having cleft the mantle longitudinally (PI. II. fig. 1.), w^e observe the branchiae, each with a branchial heart situated at its base, and not differing in structure from those of other Octopods. During life, however, these branchiae exhibited great contractility, and the branchial hearts, whose function has been questioned, pulsated regularly. The fibrous membrane which invests all the intestines is covered, especially below, with numerous chromatophora, which contract and dilate alternately. We see across this membrane the vaguely marked contour of the last portion of the intestine, of the ink- bag and of the genitalia which occupy the posterior portion of the body. The conjoint aperture of the rectum and of the ink- bag is placed upon a little muscular tongue which ends forwards in two very fine points, and which passes so far forwards below the funnel that the posterior edge of the latter covers it com- pletely. At first we sought in vain for a long time for the aper- ture of the generative organs, which is perceived with great dif- ficulty, when an accident discovered its position. One of the specimens which we were examining presented through the mantle a streak of white matter which began to fill the branchial cavity. Having carefully opened this individual, we saw a coil of the seminal thread, w^hich we shall describe by- and- by, escape from a semilunar aperture upon the left side, by the side of the place where the branchial vessels leave the intestinal sac to pass into the branchia. It was afterwards easy to perceive this very much contracted cleft in all specimens, and to convince ourselves, by examining the internal generative organs, that this asymme- trical orifice was the sole means of discharging the contents of the generative organs into the branchial cavity. Afler having taken away the fibrous intestinal sac (PL 11. fig. 1.) we easily discover the rectum, which is very delicate, passes from below upwards from the dilatation of the large trunk, and which rises among a considerable mass of venous appendages 134 VERANY AND VOGT ON THE HECTOCOTYLI which especially invest the aortic heart and the great vessels uniting the latter with the branchiae. Below this mass of venous appendages we see a flask-shaped organ of a gelatinous trans- parency, and through which fine silvery lines may be distin- guished ; these have a very peculiar lustre, and look like white lights painted thickly and with great brilliancy upon a grayish satin ground. These lines differ much in their arrangement in different specimens ; we see, however, that they are especially numerous towards the posterior convex edge of the flask, whilst the anterior hollowed border of this organ seems to contain a more transparent substance. The bottom of the flask, whose form varies very much in consequence of the great contractility which it possesses, is turned to the right, while its narrow'cd neck passes to the left upon the great vessels of the branchia. This extremity hooks round these vessels, and at the summit of the curve is found the cleft orifice of which we have spoken above. The remainder of the generative apparatus (PL II. fig. 4.) is situated upon the dorsal surface of the intestinal sac, and is con* tinuous with the neck of the flask. It is composed of an elongated vesicle having nearly the same form as the flask, and possessing as great contractility in its transparent muscular parietes. It is divided into two portions, which however are united in a common envelope. The anterior portion we will call the cornu: it is membranous and transparent, and allows a contained whitish spiral thread to be seen through its side, much less brilliant and larger than the threads contained in the flask. At the posterior enlarged extremity of the cornu there is attached an organ having the form of a pointed apple, whose base is turned towards the cornu, its point towards the right branchia, with the heart corresponding to w hich it is in contact. This organ has a yellowish chalky tint ; its internal surface is strongly attached by cellular tissue to the base of the flask ; it is the testicle. The generative organs thus form altogether an actual ring (PI. II. fig. 3.) round the great vessels of the left branchia ; a ring which is completed behind only by the fibrous tissue uniting the envelope of the testicle to the bottom of the flask, whilst the closure of the ring in front, upon the ventral face, is produced by the union of the neck of the flask with the neck of the cornu. AND THE MALES OF CERTAIN CEPHALOPODS. 135 We have just pointed out the form of the testicle ; besides the general envelope which unites this organ to the cornu, it has a proper fibrous envelope which closely invests it. Invariably situated upon the dorsal surface, behind the branchial heart of that side, the testicle however changes its place more or less, according to the state of turgescence of the different portions of the sexual apparatus. The testicle itself is divided into two portions which differ in appearance even to the naked eye ; the pointed extre- mity being more transparent than the body of the organ which is turned towards the cornu. Fine chalky lines radiate from the pointed extremity in all directions towards the enlarged centre of the organ, where they converge again to approach that face of the testicle which is in contact with the cornu. Under the microscope these chalky lines are seen at once to be the seminiferous tubes which end in a cul de sac at the pointed extremity of the testicle, and which begin to ramify when they enter into the thick portion of this organ. These ramifications however are almost parallel with one another, so that the entire testicle seems to be composed of parallel seminiferous tubes, w^hich however converge both towards the point and towards the base of the organ. The seminiferous tubes contained, in all the individuals we examined, no trace of completely developed spermatozoa, but only granular cells and very small free granulations, which from their strongly -marked contour had the appearance of fine drops of oil ; the granule-cells also exhibited among the fine granules more strongly-marked little drops. All the individuals which we examined being in the pairing season, the spermatozoa had already passed into the ejaculatory organs, and in the semini- ferous tubes of the testicle there were found only the elements required for the reproduction of the semen. There is no direct connexion between the seminiferous tubes and the efferent canal ; we have assured ourselves of this cir- cumstance microscopically. The special envelope of the testicle (PI. III. fig. 5.) is considerably narrowed at the base of that organ, and is directly continued into the muscular envelope of the cornu. A kind of funnel is thus formed at the place where the testicle is attached to the cornu, towards which all the seminiferous tubes converge to terminate by a rounded end. The contents of these 136 VERANY AND VOGT ON THE HECTOCOTYLI tubes ought then to be poured into the space formed by this funnel, which is nothing more than the continuation of the special enve- lope of the testicle. The seminal masses would, by means of this funnel, pass freely into the bottom of the cornu, if there were not a very narrow deferent canal (PI. III. fig. 5 n)y which is affixed to the funnel like a tube, and projects freely into the cavity of the cornu. Merely looking at the cornu, this canal may be seen •applied to its dorsal surface; it is about 3 millimetres long; and even without opening the cornu, it is easy to convince one- self with a lens that its free extremity floats in the cavity of the cornu. This canal is extremely contractile and mobile, and has almost continually during life a vermicular motion. The struc- ture of the canal is very simple : it is a muscular tube, with its longitudinal fibres especially developed, united into bundles and so forming slightly projecting ridges upon the internal surface of the canal. The orifice of the canal resembles the mouth of a hydroid polype, in consequence of these radiating ridges. The whole internal surface of the canal is covered with cilia of such large size, that the single ones can be distinguished with a magnifying power of 100 diameters, whilst the movement itself and the current produced by it are very distinctly visible by the very lowest powers. It is evident, from the disposition and structure of this efferent canal, that the seminal mass, entering the funnel in consequence of the rupture of the seminal tubes, is received by the efferent canal, and conducted by the latter into the cavity of the cornu, which thus in some respects serves as a seminal reservoir. The structure of the cornu itself is very complicated, and we should have had considerable doubt as to certain points of its anatomy, if chance had not given into our hands many living animals in succession, upon which we were enabled to complete our researches. We have already mentioned the white and coiled thread which is visible across the transparent parietes of the cornu ; this coil is constantly in motion, as well in con- sequence of its own contractions as of those of the investment of the cornu, which is very solid and is formed by muscular fibres interlaced in all directions. The coil is laid bare when this contractile envelope is cut open : it swims, so to speak, in a viscous Hquid which fills all the cornu, and which glues together AND THE MALES OF CERTAIN CEPHALOPODS. 137 all the folds of the coil, and that so effectually, that there is some difficulty in disentangling them. When the attempt has been successful, we see that the coil is obviously composed of two filiform organs rolled up together, i. e, of the deferent canal and of an accessory gland, whose excretory canal is united to the deferent canal towards the apex of the neck of the cornu. The deferent canal (PI. III. fig. 5 d') begins by a trumpet- shaped orifice (PL III. fig. 5 i) folded like the orifice of the efferent canal, but wider than the latter ; this orifice is not clothed with vibratile ciha, like the efferent canal, which tends to confirm our observa- tion, that the two canals have no direct communication with one another. The plaited orifice of the deferent canal becomes conti- nued into a considerable dilatation of a pear-shape (PI. III. fig. 5^) whose wider extremity is turned towards the orifice, while the tail a little enlarged passes into the continuation of the deferent canal ; this dilatation is produced less by any enlargement of the cavity of the canal than by the very considerable development of the muscular layer, which forms very projecting longitudinal ridges at this place, converging towards the pointed extremity of the dilatation. Besides this greatly-developed muscular structure, we see, as an appendage to the pyriform dilatation, a perfectly rounded bag (PI. III. fig. 5 h) which pours its contents by means of a little excretory duct, where the plaited orifice is continuous with the muscular dilatation. We could discover no histological elements in this secretion, which was composed of a substance resembling stearine. The deferent canal is prolonged from the pyriform dila- tation as a simple cylindrical canal. It possesses contractility, owing to a muscular layer which principally forms the tube. Two clavate enlargements succeed one another at a slight distance from the pyriform dilatation, and are owing, like the latter, to a greater development of the muscular layer. The interior of the deferent canal is clothed through its whole length with a glairy secretion, and most frequently the canal appears to be wholly empty ; we however met with one individual in which a seminal mass was arrested before the first clavate enlargement. We shall return to this chance- discovery, since it throws some light upon the formation of the seminal machines in general. In passing towards the pointed extremity of the cornu, the 138 VERANY AND VOGT ON THE HECTOCOTYLI deferent canal makes a multitude of folds, united to those of the accessory gland by very fine and very elastic fibres, which while they unite them, allow of a very considerable amount of contraction to each of these canals. At the extremity of the comu the deferent canal opens into a kind of common reservoir which lies at the entrance of the comu. The accessory gland (PI. III. fig. 5 m) which is enclosed in the cavity of the cornu, with the deferent canal, is composed of two portions, which, however, in their intimate structure present no essential difference — the gland itself and the excretory canal. The gland is a flattened body rolled almost spirally, and exhibits even to the naked eye a granulated appearance, resulting from the presence of many minute points which are almost entirely opake by transmitted light. The intimate structure of this gland is very remarkable ; its parietes are very thick, and there is in the middle of the gland a cavity which is directly continuous with the excretory canal. Instead of glandular tubes, it consists of little, more or less rounded sacs (PL III. fig. 8.), which are hol- lowed in the very substance of the gland, and whose aperture is almost as large as its base. These sacs beset the whole internal surface of the gland and of its excretory duct ; they are merely deeper in the gland than in the duct, and disposed a little more obliquely with regard to the axis of the gland. They are imme- diately surrounded by a very much developed elegant capillary network. The proper walls of these sacs are very thick ; they are composed of a small number of circular fibres, and are lined internally with a considerable layer of cylindrical cells, carry- ing at their extremity very long cilia ; these are in continual movement, and they transport a viscous homogeneous liquid containing fine granules, which are here and there united into little masses. These glandular sacs, which doubtless secrete the mass of which the envelope of the spermatic machines is formed, are continued until close to the anterior extremity of the excre- tory canal ; there they disappear by degrees, and the excretory canal itself becomes considerably enlarged, to form a wide sac (PL III. fig. 5 b) into which the deferent canal also opens. This sac is on all sides attached to the parietes of the comu, so that the cavity of the latter is completely shut at this place, and thus there is no other passage from the cornu to the flask than AND THE MALES OF CERTAIN CEPHALOPODS. 139 through the sac. We see then, that in consequence of the very arrangement of the parts, the seminal masses carried by the deferent canal ought to meet in this common reservoir with the product of the secretion of the accessory gland, and to enter with it into the flask. It is probable therefore that here, in this wide sac with delicate walls, the spermatic machines are formed and pass afterwards into the flask. The latter organ (PI. III. fig. 1-4.) is situated, as we have al- ready said, upon the ventral face of the intestinal sac immediately under the fibrous envelope which surrounds the latter. The mem- brane wliich forms the flask is very delicate, very contractile, and almost impermeable to water. In its interior we see undulating brilliant white lines, such as we have described above, and which are sometimes so obvious that they might be taken for external ornaments of the envelope. The semilunar aperture situated close to the neck of the flask is generally very difficult to per- ceive, and opens only after a prolonged stay in the water to give exit to the contents of the flask, which are formed in all the in- dividuals we have examined by a unique spermatic machine, — by a single spermatophore filling the whole cavity so completely, that it is very difficult to open the latter without damaging the spermatophore contained in its interior. This enormous spermatophore (PL iV. fig. 1.), which is nearly two centimetres in length, is always folded up in the flask, so that both its extremities approach the semilunar aperture of the latter. Removed from the flask the spermatophore has the shape of a powder-horn, having one extremity pointed and elongated into a beak, the other enlarged and rounded. The beak, although the more firm and consistent portion, is however almost trans- parent, whilst the sac appears almost white, by reason of the convoluted silvery thread which it contains in its interior. The spermatophore itself is formed by a very solid membrane, per- fectly transparent, which, after having formed the sac, is con- tinued upon the beak-like prolongation, surrounding it very closely. This envelope absorbs water very quickly, and svTells out rapidly in consequence ; it becomes separated then into two layers, the exterior of which, very delicate, forms irregular folds, often so multiplied that one might imagine the beak of the sper- Imatophore to be constituted here and there by the convolutions I 140 VERANY AND VOGT ON THE IIECTOCOTYLI of a fine spiral thread. The proper coat which lies below this layer becomes distended by the action of water, very much as a gummy mass would be, and finally bursts to allow the passage of the contents of the spermatophore. These contents are composed of two very dissimilar threads ; the one belonging especially to the beak-like prolongation, the other to the sac of the spermatophore. We call this last thread the spermatic cord (PI. IV. fig. 2 c), since it is wholly made up of spermatozoa united together around a fine thread-like axis. This spermatic cord is entirely white, silvery to the naked eye, and of an even diameter throughout, except its anterior extremity, where it becomes more delicate, to attach itself finally by an ex- cessively delicate thread to the posterior extremity of the ejacu- latory cord. The spermatozoa are united together in the sper- matic cord in such a manner that their cylindrical end, which is the larger, is attached, or rather glued by a viscous liquid, to the axis of the cord, whilst their very fine caudal extremity is turned towards the periphery. We can compare this structure to nothing better than to the elongated brushes used for cleaning bottles, in which the bristles radiate upon all sides from a me- dian axis formed by an iron thread. The spermatic cord becomes distorted very soon when it is subjected to the action of water, and it is almost impossible to unroll its coils before the comple- tion of this distortion. At the very moment the spermatic cord is drawn out of its envelope, all the spermatozoa appear to be united by a glutinous liquid, which connects them so closely that it is impossible to distinguish them. By the action of w^ater, this glutinous mass at first dissolves a little ; the spermatozoa become free, and appear then to be animated by slight undu- latory movements. Little by little these motions cease, whilst the glutinous matter congeals by the operation of the water, and to such an extent that the spermatic cord soon looks like a felted mass whose elements are undistinguishable. It seems probable to us, that the irregularly nodulated appearance noticed by Von Siebold upon the spermatic cord of the Heciocotylus Trem- octopodis, proceeds from a similar alteration produced by the influence of water. We have said that the spermatic cord ends by a more delicate extremity, when, forming numerous convolutions, it approaches AND THE MALES OF CERTAIN CEPHALOPODS. 141 the anterior extremity of the sac, where the beak-Hke projection of the spermatophore commences. In fact, the cord becomes very delicate (PL IV. fig. 3 a), and finally attaches itself to the posterior extremity of the ejaculatory canal, which in its turn fills the beak-like projection of the spermatophore. This eja- culatory canal is formed by a very firm tube with thick pa- rietes, having only a very small canal in its interior. Its parietes are excessively transparent, resist pressure greatly, and do not become disfigured by the action of water. We have been quite unable to discover any further structure in this homogeneous tube, which commences in the sac of the spermatophore by a rounded extremity, slightly drawn out, and having a minute aperture into which the extremity of the axis of the spermatic cord penetrates as a very fine thread repeatedly folded up in the cavity of the ejaculatory canal. The latter widens very soon, together with its internal cavity, which is filled in the whole length of its course by a spirally folded membranous ligament. We see distinctly in the posterior extremity of the ejaculatory cord the commencement of this membrane, which clothes the in- terior of the canal of the cord, and seems at first plaited into a number of transverse folds, which by degrees assume a spiral disposition, so that at last the end of the ejaculatory canal re- sembles under the microscope the intestine of a shark with its spiral valve (PI. IV. fig. 4.). The ejaculatory canal forms at first many folds in the wider portion of the beak-like prolongation ; after which it becomes continued almost in a straight line, form- ing itself the axis of the prolongation. At the extremity of this prolongation may be distinguished the whole external envelope of the spermatophore, which becomes bent back, so to say, to- wards the interior to form the tube of the ejaculatory cord, whose canal, clothed by the spiral membrane, is continued as far as this extremity. The analogy between the structure of the spermatophore which we have just described and that of those of the other Cephalo- poda with which we have long been acquainted is evident, so that we need insist at no greater length upon this point. The envelope which swells out by the action of water ; the ejacidatory cord, with its internal spiral ligament; in all these points the resemblance is complete, with this difference only, that the semi- 142 .VERANY AND VOGT ON THE HECTOCOTYLI nal mass is not included in a sac as in the other Cephalopoda, but is formed by a long cord coiled upon itself and completely- deprived of any special envelope. The formation of these semi- nal machines, whose size is enormous relatively to that of the male animal, is pretty well explained by an observation we made upon one of the four individuals which we examined. A whitish mass (PI. III. fig. 5/) was situated in the deferent canal of this individual a little beyond the pyriform enlargement by which this canal commences. It presented itself under the form of a pear (PI. III. fig. 9.) with an elongated tail, containing internally a thousandfold coil of a fine gelatinous, very transpa- rent thread, beset on all sides with motionless spermatozoa. This thread became lost by degrees in the more enlarged portion of the mass ; and even by pressing aside the matted spermatozoa forming it, one could only distinguish here and there some traces of a similar gelatinous thread ; it was however by no means well marked. This mass of spermatozoa was not invested by any envelope, and it still retained the form of the foot-like enlarge- ment in which it had been modelled. It seemed evident to us that such a seminal accumulation is moulded by passing through the whole length of the deferent canal into an elongated thread, and that it is in the common reservoir that this thread receives at once both its envelope and the ejaculatory cord secreted by the accessory gland, which thus transforms the whole into a spermatophore. This, once formed, passes into the flask, whence it is expelled when copulation is about to occur. In the great majority of individuals which we examined, the right arm of the third pair was disproportionately developed, and had an external organization such as we have pointed out in our zoological description. In other individuals this arm was replaced by a very considerable pedunculated vesicle. We shall see by the anatomy of these parts that there exists a correlation between them — an intimate correlation — and that the formation of the vesicle must necessarily precede that of the hectocotyli- form arm. As to the latter, we have been unable to discover in its struc- ture any great differences beyond those we are about to point out, from that of an ordinary Cephalopod-arm. The axis of this arm is formed by a cylindrical muscular tube of great thickness, AND THE MALES OF CERTAIN CEPHALOPODS. 143 which is continued beyond the series of acetabula into a long filiform appendage which we name the ^' flabellum/' and which is ordinarily concealed in an ovate sac which terminates the anterior extremity of this arm. This muscular axis is formed, as M. Kolliker has already very well shown, of three different layers, one of which is longitudinal and the other two circular. In the midst there is a hollow cylindrical space, in which are situated two organs which require the more careful examination, since M. Kolliker has entirely mistaken their nature. One of these organs is a blood-vessel with very delicate and transparent parietes, which runs through the whole length of the muscular axis even in the portion which we have called the flabellum, without any great diminution of its volume. There can be no doubt as to the exact nature of this vessel, which we have many times observed by transmitted light in individuals still living, but in which the movements of the heart were too irregular to allow of our determining the direction in which the current of the blood moved. We could very well distinguish the blood- corpuscles, forming at the time of death irregular masses, which, while the heart yet beat, were agitated more or less distinctly. Unable to make any injection of the few specimens at our com- mand, we could not correctly determine the relations between this vessel occupying the central axis of the muscular cylinder, and the cutaneous vessels which appear upon the whole surface of the arm. We believe, however, that w^e have seen, towards the anterior extremity of the flabellum, this central vessel giving off lateral branches, which pierce the muscular tube to pass to the external surface and to the skin. The vessel is continued uninterruptedly into the pedicle of the arm, where we were un- able to trace it further. It seems indubitable, however, after what we have just said, that it is the central artery of this ab- normal arm, disposed like the central arteries of th^ other normal arms. The second organ worthy of remark, and which is enclosed in the muscular canal, is the nervous cord formed by as many ganglia as there are acetabula upon the whole length of the arm. The tissues of the living individuals are so transparent, that we can perfectly well distinguish the nervous cord in all the arms of these little males without any dissection, and all the ganglia 144 VERANY AND VOGT ON THE HECTOCOTYLI corresponding to the acetabula can be counted with the most scrupulous exactness. It is equally easy to see this ganglionic cord in the hectocotyliform arm, and to decide that, as M. von Siebold has well observed, there is only a single ganglion cor- responding to each acetabulum ; but the acetabula being very closely approximated, and succeeding one another alternately upon the two sides, the ganglia also are pressed against one another, so that to the naked eye, or under a simple lens, they look like the close beads of a necklace. Examined microscopi- cally, these gangha (PI. IV. fig. 6 c) all show the form of a tra- pezoid whose base is turned towards the acetabulum to which the gangHon belongs. M. KoUiker has figured the appearance of these ganglia pretty well ; but by an inconceivable mistake, he regards the ganglia as masses forming the contents of the central vessel which he makes out to be an intestine. The nerves pass- ing from these ganglia are to be seen with great difficulty, inas- much as they penetrate the muscular cylinder in such a manner that they are almost always perpendicular to the microscope. Hence M. Kolliker has described these nerves as canals ascend- ing towards the surface of the skin ; and we must confess in fact, that a little nerve viewed perpendicularly upon its axis is not unlike a fine canal with delicate sharp parietes. The chain of ganglia ends with the series of acetabula ; but some nervous threads which are very delicate might yet be per- ceived in the muscular axis of the flabellum, where they ended by becoming so fine that it was impossible for us to trace them to its extremity. We find no differences whatsoever in the structure of the rest of the arm, so far as there are suckers ; the skin, the walls of the acetabula and their whole structure appear to be entirely conformable with all that we have seen in the arms of ordinary Cephalopods : we may be allowed then to pass over these points in silence. The anterior extremity of the arm is formed by a little oval sac, which hangs between the two last acetabula, and whose wall is a direct continuation of the skin which covers the dorsal face of the arms. On carefully examining this little sac, we perceive even with the naked eye that its interior contains a spi- rally coiled thread. Between the two last acetabula there exists AND THE MALES OF CERTAIN CEPHALOPODS. 145 a little fissure which leads into the cavity of this sac, and by which the animal is enabled to evolve the spiral filament hidden in its interior. We have frequently seen instances of this oc- currence ; the filament passed by its pointed extremity out of this cleft and slowly unrolled itself; the sac all the while being agitated by repeated contractions which aided the expulsory movement. The filament itself performs very marked vermi- cular movements, which we can only compare to the motions of the tentacular filaments of certain tubicolar worms, particularly qf the Terebellce. The little sac is wholly contracted w^hen the thread has passed out of it, and it then exhibits under the mi- croscope (PI. IV. fig. 6 e^f) very numerous rugosities, presenting a sort of very pretty watered pattern. It evidently consists of two membranes, one external, which is nothing more or less than the continuation of the skin w^hich covers the whole arm, and an internal muscular layer which is continued by two muscular bundles upon the two sides of the flabellum. This sac con- tracted upon the flabellum, which it had contained, has been also seen by the authors who have written upon the Hectoco- tylus of the Argonaut', and lastly by M. Kolliker, who regards it as a membranous appendage without any other sig- nification. The flabellum itself (PI. IV. fig. 8) is composed especially, as we have said above, of the muscular cylinder of the arm, whic h occupies its centre without interruption as far as its extremity, gradually becoming smaller. This cylinder ends by a point at the very extremity of the flabellum, which is at first completely rounded, but becomes flattened little by little like a lance-head towards its end. At the distance of a few milUmetres from the extremity of the flabellum the muscular cyhnder becomes sud- denly' thickened, and takes the form of a piston (PI. IV. fig. 8 b) ; its internal cavity even is here widened. At this same place we have always seen a considerable mass of blood-corpuscles dis- tending the enlarged cavity. The median vessel which occu- pies the centre of the muscular tube, ends therefore here in a kind of reservoir, whilst the muscular axis continues as a hollow cylinder as far as the extremity. The external membrane which invests the flabellum is, especially at its base, very loose, and exhibits proper undulatory movements due to two muscular bun- SCIEN. MEM.— A^a/. Hist. Vol. I. Part II. 10 146 VERANY AND VOGT ON THE IIECTOCOTYLI dies which run along the two sides of the flabellum, and whirh may be followed even beyond the piston-like enlargement. These two cutaneous muscles are accompanied through their whole length by two venous trunks which send off numerous ramifica- tions, forming a capillary network over the whole surface of the flabellum. These ramifications are especially remarkable at the very extremity of the flabellum, where they are perceived very easily, the tissues being at this place completely transparent. The extreme mobility of the flabellum perhaps plays a part in the physiological functions of the hectocotyliform arm. We have seen this appendage continually in movement, as if it were feeling about to fix itself somewhere. It embraced the arms, and even the body of the animal to which it belonged, but it disentangled its coils again without its being possible to discover to what end all these motions tended. We shall bring forward by-and-by an observation of M. Kolliker's, which will, perhaps, set other observers upon the track of the physiological function of this part. We have yet to mention a last peculiarity of the structure of the hectocotyliform arm; it is the existence of a sac of con- siderable extent, which is visible upon the posterior or dorsal face of the arm near its base, and therefore opposite to the ace- tabula. This sac (PI. I. fig. 3 a) is 15-20 miUimetres in length, and is elongated, inasmuch as it lies along the dorsal face of the muscular cylinder of the arm. It is distinguished very easily throughout its entire extent by its deep colour, whilst the rest of the hectocotyliform arm is wholly colourless. This colora- tion is due to the chromatophora which beset the whole internal surface of the sac, and may be seen shining through it. A semi- lunar aperture, situated upon the dorsal face of the arm imme- diately above the pedicle, leads into this sac, which is a true cul- de-sac, and is closed upon all sides. We have never found any- thing in this sac, which is obviously a diverticulum, an involu- tion of the skin of the body itself, and whose formation depends, as we shall soon see, upon the mode of development of the arm. We have already many times mentioned the oval vesicle car- ried upon a delicate pedicle, which was observed in many indi- viduals in the place of the hectocotyliform arm. Upon closely examining this vesicle, we see that it is lined in its whole extent AND THE MALES OF CERTAIN CEPHALOPODS. 14? with the same chroraatophora as those which are distributed over the whole surface of the body, and that it has at its base a little semilunar aperture similar to that which we have observed in the sac of the arm. Upon opening the vesicle we find in its interior the hectocotyliform arm coiled up spirally (PI. IV. fig. 7), so that its acetabula are turned towards the centre ; the dorsal face of the arm towards the periphery of the vesicle. This arm had reached its full development in the individuals in question ; but in these also the vesicle had attained its final stage of deve- lopment, since it is much smaller in a specimen described by M . Verany in his work already quoted. Whilst we were examining the arm rolled up in its vesicle, another living individual placed in sea-water unrolled little by little the arm concealed within its vesicle. The arm passed out by its base, and whilst it kept on unrolling, the vesicle was reversed by the same action, and ended by becoming the sac which we have described upon the dorsal surface of the arm. It will now be readily explicable why this sac has its internal face lined by chromatophora similar to those of the skin, the external surface of the vesicle havino* become the internal face of the sac. We can now also account for the fact, that in all the individuals we observed the hectoco- tyliform arm was always twisted and rolled up at its extremity, an arrangement which was a result of the spiral coil which it had formed in the vesicle. We may be very concise in our description of the female or- gans, which are constructed upon the same plan as in Argonauta and Tremoctopus. The simple ovary surrounded by its capsule is situate at the bottom of the intestinal sac, and communicates with two very long and convoluted oviducts, which are folded up upon the two sides of the ovary. There does not exist, as in Tremoctopus violaceus, any decided glandular enlargement in the course of the oviducts ; but in the sole female which we have had an opportunity of examining, we found in the course of the left oviduct two eggs of an oval form, which were still retained in the oviduct, and which were evidently on their way out. The two apertures of the oviducts are placed as in the two other species of Octopods which have hectocotyline generation, (i.e. Argonauta and Tremoctopus violaceus,) at the base of the branchiae; the oviducts passing under the branchial arteries and veins to 10* 148 VERANY AND YOGT ON THE HECTOCOTYLT open by two little lateral papillae into the respiratory cavity. These apertures are very distant from the funnel, and are placed altogether at the sides of the cavity. The investigations which have been set forth in the preceding pages ought, we think, to furnish a very complete solution of this important question concerning the nature of the Hectocotyli which has been agitated now for some years. Accepting the results of MM. Siebold and Kolliker, and the conclusions which they drew, one was still struck by a something surprising and in strange contradiction with certain zoological principles which were believed to be firmly established. Among these Cephalo- pods — so similar for the rest in their external and internal struc- ture— there were, according to this view, to be found kinds which could hardly be distinguished generically, and yet in which the difference between the organization of the male and that of the female was carried to its greatest extent. According to this view, females endowed with all the complicated and highly deve- loped organs of the Cephalopoda possess males which previous observ^ers had taken for intestinal worms, and w^hich in any case were so poorly organized that independent life seemed refused to them ; and so strange an exception, an example of which is hardly to be found in the animal kingdom*, was to be found in species side by side with others, in which the males were or- ganized as completely as the females. It w^as evidently a ques- tion which ought to be in the highest degree interesting to zoo- logists, and we are rejoiced to believe that the solution we offer, and which we believe to be definitive, will not remain without influence in preventing for the future similar mistakes. It would be useless to reiterate this truth, that the being dis- covered by Laurillard at Nice, and described by Cuvier, is really the detached arm of that species of Octopus which we call Trem- octopus Carena, and that the individuals which carry these deformed arms are always the males — smaller it is true than the females — but for the rest organized in as complete a manner as the other Cephalopoda. We have thought it superfluous to * The Entomostracous Crustacea, the Rotifera, and the Cirripedia, classes which present instances of disproportion between the sexes quite as remarkable as Kolliker supposed it to be in Argonauta, &c., seem to have been forgotten by MM. Verany and Vogt. — TV. AND THE MALES OF CERTAIN CEPHALOPODS. 149 give any detailed description of the nervous, alimentary, circu- latory, and respiratory systems of these animals, the high organi- zation of these systems being obvious from our figures or from the mere inspection of the animal. It. is the same for the male of the Argonaut, although the difference in shape between this and its female is still more remarkable ; so great indeed, that so scrupulous an observer as our friend M. Krohn neglected to examine these little creatures, which he took to be young just hatched, and still provided with their yelk-sac. We have already said that the discovery of these minute males of the Argonaut belongs to Dr. Miiller of Wurzburg, who will without doubt inform us that their internal organization resembles that of the other Cephalopoda. We have only been able to examine a single specimen of these minute males, long preserved in alcohol, and given to one of us by M. Krohn, and we can affirm that in point of external structure, which we have alone examined, it perfectly agrees with the type of the other Cephalopods. This is obvious from the figure w^e have given of this little male*. The male of the Tremoctopus violaceus alone is not yet known ; however, we have no fear but that further researches, in the di- rection we have taken, will discover the male which bears the Hectocotylus described by M. Kolliker, and which differs in many points from the other HectocotylL All the Uectocotyli that have been described up to the present time as perfectly independent beings are then nothing else than the detached arms of certain species of Cephalopods, in which the males, for the rest perfectly normal, are smaller than the females. All our observations prove that these arms are de- tached with extreme readiness, and that the pedicle which carries them exhibits, after its separation, a clean smooth surface, without any trace of laceration. The detached Hectocotyli prove as evidently that this separation takes place in a natural manner, for their posterior extremity never exhibits any trace of tearing nor of cicatrization. The pedicle which remains upon the body after the arm has become detached, without question re- produces this arm, and that probably by a process of budding, comparable perhaps to the reproduction of the deciduous horns of certain ruminants. Our observations have only permitted us * Omitted here. — Tr. 150 VERANY AND VOGT ON THE HECTOCOTYLI to make out the last phase of this reproduction, when the ab- normal arm, coiled up in a spiral, is enclosed within a vesicle, from which it ought soon to make its exit and unroll outwards. But it seems to us unquestionable that the arm is really repro- duced in the interior of this vesicle, by budding from the peduncle and raising up the skin covering the latter, which thus in the end forms the vesicle enclosing the arm. The vesicle which the little male of the Argonaut carries has exactly the same relations ; it also contains in its interior the spirally coiled Hectocotylus. Our observations as they stand could furnish only very scanty indications as to the physiological function of these abnormal arms of the males ; we have found in fact in these arms only the ordinary structure of a cephalopod-arm terminated by the tlabellum — a kind of tail fixed to one of the extremities, — and by a cutaneous sac at the other extremity. This sac was inva- riably empty in the individuals in which the arm was still fixed upon its pedicle; it is formed, as we have just seen, by the retroversion of the sac in which the arm is developed. But our investigations, combined with the results obtained by MM. Cuvier, Laurillard, KoUiker, and Siebold, solve the problem. All these observers have examined the Hectocotyli when de- tached ; all have noticed that in these separated individuals the sac situated at the base was by no means empty, but that it was filled by organs belonging to the generative apparatus. MM. Kblliker and Siebold have also determined that this sac contains a long folded spermatic cord, which is continued into an ejacu- latory canal, having a harder pointed extremity, which these writers call the penis, inasmuch as they believe the full sac and its contents to be a true testicle, with the excretory and copu- latory canals, which are ordinarily in connexion with this organ. Now, our observations are formally opposed to this interpre- tation. We have proved in fact that tl\e testicle is situated at the bottom of the intestinal sac, and that it is constructed upon the same type as that of other male Cephalopods. We have shown besides, that this testicle is in relation with the pecuhar excretory organs which fashion the seminal mass, so as in the end to form a seminal machine — a spermatophore — of a very complicated structure and of very considerable dimensions. AND THE MALES OF CERTAIN CEPHALOPODS. 151 Now that we have exactly described the spermatophore con- tained in the flask, let any one compare our description with that which MM. Kolliker and Siebold have given of the contents of the sac of the Hectocotylus, and the perfect agreement of the two will be obvious : the spermatophore surrounded by a trans- parent envelope, having the same properties as the envelope of the supposed testicle in the Hectocotylus : the seminal cord of the spermatophore resembling in all respects the same organ rolled up in the genital capsule of the arm : the ejaculatory cord of the seminal mass with its spiral ligament, its more solid and pointed extremity, nowise different from the deferent canal and the penis described by these authors ; — there can no longer be any doubt as to the identity of these organs. It is then evident that the seminal machine constructed in the internal organs of the male, leaves these organs to be transplanted into the sac which the hectocotyliform arm carries ; this then, loaded, becomes detached from its pedicle and affixes itself to the female, in all probability during an act of copulation or embracing, which takes place, as we know, among the other Cephalopoda. It is true that no observation at present known explains to us the mode of transport of the seminal machine from the aperture of the generative organs situated in the branchial cavity as far as the sac of the abnormal arm ; and it may be that the mobile flabelhim which terminates the Hectocotyli of the Argonaut and those of our Tremoctopus Carena, is charged with the per- formance of this removal of the seminal machine. M. KolUker has observed, in fact, that in one specimen of the Hectocotylus ArgonautcB, the anterior extremity of the flabellum was bent back into the aperture of the sac containing the seminal machine, and that this extremity was twisted around the machine. It may be that M. Kolliker surprised the Hectocotylus at the moment when the flabellum was depositing its seminal mass in the sac of the arm, though it is very possible on the other hand that this engagement of the extremity of the flabellum was simply an • accident due to the continual explorations performed by the organ. The acts of copulation and fecundation even are not yet known to occur by direct observation ; but as all the observers who have up to this time found detached Hectocotyli have found them on the arms, in the funnel, and in the respiratory cavity 152 VERANY AND VOGT ON THE HECTOCOTYLI of the females, it is probable that these charged arms creep by the aid of their numerous suckers as far as the aperture of the female generative organs, when the spermatophore then performs its office. We sum up then by offering the following conclusions : — 1. The Argonaut, the Tremoctopus violaceus, and T. Carena, have males whose structure agrees with that of the common Cephalopod-type. 2. One of the arms of these males becomes specially modified into a copulatory organ. 3. The beings known at present under the name of Hectocotyli are not separate animals, but are merely the detached copulatory arms of the males, charged with a seminal machine. 4. The copulatory arms are detached and renewed periodically. DESCRIPTION OF THE PLATES. [The great number of figures given by MM. Verany and Vogt rendered it requisite to exercise some selection, and to exclude any that were not absolutely necessary. Of their four Plates, the three last (PI. 7, 8, 9, Annales) are exactly copied in our Plates II. III. and IV. ; but only two figures of their first Plate (6, Annates) are given in our Plate I. figs. 3, 4. The others appeared to be not necessary to the comprehension of the memoir. This change has of course rendered requisite a new numbering of the figures.] PLATE I. Fig. 3. Tremoctopus Carena. Male, seen from the dorsal surface, a, aperture of the retroverted sac which contained the Hectocotylus ; b, vesicle which contained the flabellum ; c, flabellum. Fig. 4. Tremoctopus Carena (male), with the vesicle still containing its Hec- tocotylus. PLATE II. Fig. 1. Male T. Carena, seen from the ventral surface. The visceral sac is cleft longitudinally, and the left half of the mantle has been thrown back to exhibit the branchia and the aperture of the male organs at the moment of expulsion of the spermatophore. a, funnel; 6, vesicle containing the Hectocotylus ; c, eye ; d, aquiferous aperture ; e, man- tle ; /, branchia; g^ branchial heart; h, flask; i, spermatophore passing out. Fig. 2. The same. The mantle is cleft and thrown back, the funnel is taken away ; the fibrous envelope of the abdominal cavity is cut so as to AND THE MALES OF CERTAIN CEPHALOPODS. 153. exhibit the organs in their natural position. The letters a — h have the same signification as before, i, abnormal arm, i. c. the Hectoco- tylus ; k, the anus ; /, the rectum ; m, the large intestine ; w, the ink- bag; 0, the ventral aorta; p, the venous appendages. Fig. 3. The genitalia with the branchiae, the hearts, and the large intestines, wholly detached and seen from the ventral side. The letters as in the foregoing figures. I to t V : jai^ia ST ^^uaraipru ibuiuib aqx / ' ^ 216 K.E.VON BAER. — PHILOSOPHICAL FRAGMENTS. In detail it holds good as little as any other representation of organic relations upon a surface. Thus the single features ad- duced must pass for the whole characters, c. g, the formation of wings and air-sacs for the whole character of Birds. The expo- sition, again, can only be very imperfect, since for most animals the investigation has hardly been commenced. This scheme is only meant to bring clearly before the mind, how the first decisive distinction is whether the first rudiment is a true egg or a germ-granule ; how, in the germs of ova, all animals are at first alike (see above, e) ; how then the principal type becomes defined (which is called, origin of the embryo) ; whereby it remains undecided whether any radiate animal is developed from a true egg. If now the type of the vertebrate animal appears, the embryo is at first nothing but one of the Vertebrata without any particular characteristics. Chorda dor- salis, dorsal and abdominal tubes, gill-clefts, gill-vessels, and a heart with a single cavity, are formed in all. Then commences a differentiation. In a few, gill-laminae and no allantois are de- veloped ; in others, on the other hand, the gill-clefts coalesce, and an allantois buds forth. The former are aquatic animals, though not all permanently so : the others lead an aerial exist- ence. The latter all acquire lungs. Let us follow out the former series first however. The embryos for a long time retain a great similarity ; they push out long tails and scull about with them in the water. On the other hand, their extremities are developed very feebly and late, in relation to those of other em- bryos. They either never acquire true lungs, and so become fish, or else true lungs are formed. Among the latter the lungs are either feebly developed, in which case the gills are permanent and the animals become Sirenidae ; or the lungs are better formed, and the gills either remain free until they cease to act (Salamanders), or they become covered over, the tail disappears, and with it all resemblance to a fish (tail-less Batrachia). In the second series of the Vertebrata, which never has external gills, the most essential distinction is perhaps this, — that in some a simple umbilicus is formed (Reptiles and Birds), in others this umbilicus is prolonged into a cord, after, as it seems, being altogether more rapidly formed (Schol. II. h). In what manner Birds become separated from the Amphibia K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. 21? has already been shown. Probably a difference also arises very soon in the vascular system, whose metamorphoses in the Am- phibia, however, are not yet known. While the gill-clefts in the Lizards are still open, the heart has just the same appearance as in Birds at the same period. Now just as in the Bird the special characters of the family and of the genus arise, so is it in the Mammalia. The Dog and the Pig are at first very much alike, and have short human faces. Still longer does the resem- blance persist between the Pig and the Ruminant, whose lateral toes are at first almost as long as the two median ones. For the rest we are by no means sufficiently acquainted with the em- bryos of the Mammalia to state how and at what periods they become distinguishable from one another. We are best ac- quainted with the differences in the form and structure of the ova. Since these are very manifold in their form and in their relation to the parent, I have ventured, in order not to leave the Mammalia out of the Scheme, to divide them according to their ova. The embryos, in fact, may be distinguished into those which are born early and those which come into the world in a fully developed condition. Among the former the ova of the Monotremata are probably born undisturbed. In the Marsu- pialia the embryo has burst its membranes. The ova, which are retained longer, may be reduced to three principal divisions. In the first I place ova, in which the yelk-sac continues to grow for a long time. They yield Mammals with narrow hook-like nails (claws). In some the allantois is early arrested in its growth, and the placenta is limited to one spot, or two-lobed* (Rodentia). In others the allantois is developed to a moderate extent (Insectivora) : in all others it grows over the whole am- nion transversely, and the placenta is annular (Carnivora). A second division of long-retained ova is formed by those in w^hich the yelk-sac and the allantois are small; the placenta is one- sided, and is, as it would seem, in the opposite position to that of the Rodentia ; the amnion and the umbilical cord are here * I must for the present follow Cuvier in stating that in the Rodentia the allantois remains very small, since in my earlier investigations 1 did not pay sufficient attention to this point, and for the last three months 1 have endeavoured in vain to obtain pregnant Rabits. The allantois is moderately large in the Hedgehog (one of the Insectivora), as I have recently observed. 218 K. E. VON BAER. PHILOSOPHICAL FRAGMENTS; largest. These ova produce animals with flat nails and three- lobed cerebral hemispheres*. A third division has a yelk-sac which soon disappears, but an allantois which grows outimmensely at its two extremities. These ova produce ungulated and finned animals ; if the placenta is distributed over the whole ovum, but is collected in particular masses, we have animals with cleft hoofs ; if it is distributed homogeneously, we have other Ungulata and Cetaceaf. Hence the principal differences of the Mammalia are marked very early in the ovum, for according as the allantois is much developed, or otherwise, does the ovum become long or short J. In the former case, the embryo not only acquires a broader horny covering upon its fingers, but also a more com- plex stomach, and, in connexion therewith, long jaws, a flat articulation of the jaw, usually complex teeth, incapability of seizing and climbing, &c. &c. It is the plastic series among the Vertebrata. I must advert to an objection against the whole view here set forth, which may be based upon the circumstance that in some cases the embryos of nearly allied animals exhibit considerable differences at an early period. The embryos of the Ophidia, for instance, are very early rolled up, and so may be readily enough distinguished from Lizards. This plainly arises from the ex- cessive length to which in this case the vertebrate type is drawn out. Dissection, however, exhibits a great harmony in the internal structure ; and since the posterior extremity of the Lizards also forms a spiral, the difference probably lies merely in this, that the vertebrate type in the Ophidia is more elongated, and it seems, in fact, to be greater than it is, because it presents itself so nakedly. Thus also the larvaj of many families of Insects are in their external appearance very different in different families. Much probably depends in this case upon their shorter or longer sojourn in the egg. However, this objection, the only one which * It would be very interesting to know the ova of the Lemurs, so as to ascer- tain whether they are very similar to those of the Monkeys or not. t According to a letter of Rudolphi's, the chorion of the Dolphin is similar to that of the Horse. According to Bartholin it is a. placenta exilis. X Perhaps the difference may be recognized still earlier in the chorion. See Ueher die Gefiissverhindung zwischen Mutter nnd Frucht. Leipzig, 1828, fol. K. E. VON BAER.— PHILOSOPHICAL FRAGMENTS. 219 I have been able to discover against the view in question, can. have little weight so long as no internal differences in the larvae have been demonstrated. For the simple reason, that the embryo never passes from one principal type to another, it is impossible that it can pass suc- cessively through the whole animal kingdom. Our Scheme, however, shows at once that the embryo never passes through the form of any other animal, but only through the condition of indifference between its own form and others ; and the further it proceeds, the smaller are the distinctions of the forms between which the indifference lies. In fact, the Scheme shows that the embryo of a given animal is at first only an indeterminate Ver- tebrate, then an indeterminate Bird, and so forth. Since at the same time it undergoes internal modification, it becomes in the whole course of its development a more and more perfect animal. However, it may be objected here, if this be the true law of development, how comes it that so many good reasons could be adduced for that which has been previously in vogue ? This may be explained readily enough. In the first place, the difference is not so great as it looks at first sight; and in the second, I believe that an assumption was made in the latter view, and it was afterwards forgotten that it had not been demonstrated ; but especially, sufficient stress was not laid upon the distinction between type of organization and grade of development. Since, in fact, the embryo becomes gradually perfected by progressive histological and morphological differentiation, it must in this respect have the more resemblance to less perfect animals the younger it is. Furthermore, the different forms of animals are sometimes more, sometimes less remote from the principal type. The type itself never exists pure, but only under certain modifications. But it seems absolutely necessary that those forms in which animality is most highly developed should be furthest removed from the fundamental type. In all the fundamental types, in fact, if I have discovered the true ones, there exists a symmetrical {gleichmdssige) distribution of the organic elements. If now predominant central organs arise, especially a central part of the nervous system, according to which we must principally measure the extent of perfection, the 220 K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. type necessarily becomes considerably modified. The Worms, the Myriapoda, have an evenly annulated body, and are nearer the type than the Butterfly. If then the law be true, that in the course of the development of the individual the principal type appears first, and subsequently its modifications, the young Butterfly must be more similar to the perfect Scolopendra, and even to the perfect Worm, than conversely the young Scolo- pendra, or the young Worm, to the perfect Butterfly. Now if we leave out of sight the peculiarities of the Worm, the red blood, &c., which it attains at a later period, we may readily say that the Butterfly is at first a Worm. The same thing is ob- vious in the Vertebrata. Fishes are less distant from the fun- damental type than Mammalia, and especially than Man with his great brain. It is therefore very natural that the Mam- malian embryo should be more similar to the Fish than the em- bryo of the Fish to the Mammalian. Now if one sees nothing in the Fish but an imperfectly developed Vertebrate (and that is the baseless assumption to which we referred), the Mammalian must be regarded as a more highly developed Fish ; and then it is quite logical to say that the embryo of a vertebrate animal is at first a Fish. Hence it was that I asserted above (§ 1.), that the view of the uniserial progression of animals was necessarily connected with the prevailing idea as to the law of development. But the Fish is not merely an imperfect vertebrate animal ; it has besides its proper ichthyic characters, as development clearly shows. But enough I I have attempted, in embodying the course of development, to show also, that the embryo of Man is unques- tionably nearer to the Fish than conversely, since he diverges further from the fundamental type ; and upon this ground alone has much been inserted that is problematical, as the umbilical attachment of the Monotremata. In detail, this representation can as little exhibit all the relations justly, as any other repre- sentation of organic relations upon a plane surface — even if the investigation were complete, instead of being just begun. Let us sum up the contents of this section as its conclu- sion. The development of an individual of a certain animal form is determined by two conditions: — 1st, by a progressive development of the animal by increasing histological and mor- K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. 221 phological differentiation; 2ndly, by the metamorphosis of a more general form into a more special one. Corollaries to the Fifth Scholium, The history of development is the true source of light for the investigation of organized bodies. At every step it finds its ap- plication, and all the conceptions which we have of the mutual relations of organized bodies must experience the influence of our knowledge of development. It would be an almost endless task to demonstrate this for all branches of investigation. Since, however, those conceptions must spontaneously modify them- selves so soon as the course of development is otherwise under- stood, we may be permitted to bring forward a few points in order to exhibit the influence of the view here set forth, and thereby to justify the length at which it has been given. I have endeavoured also to arrange these additions or appendices, so that those which come first may contribute to the understanding of the subsequent ones ; yet I have not been able always to suc- ceed in doing this without intercalating many explanatory epi- sodes. The reader will have also to complain of repetitions. The greatest repetition of all however is, that all these considera- tions are nothing more than reflexions of the contents of this Scholium. First Corollary. — Application of this Scholium to the Doctrine of Arrests of Development, It is no longer necessary to demonstrate that monstrous growths can only be understood by knowing the normal course of development. But I may be permitted to say a word con- cerning arrests of development, since sometimes the understand- ing of these malformations has been considered to be insepa- rable from the view of the progression of the higher through the lower, forms of animals ; and it might thence be believed that a contradiction of the latter view contradicted the doctrine of arrests of development. This doctrine, however, is too well based to be shaken by any alteration in the views which are en- tertained with regard to the differences of form in the course of the development of the higher organisms. Yet these malforma- tions must not be regarded as the permanent forms of some other 222 K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. animal which the embryo had to pass through, but simply as a partial stoppage at an earlier stage of their own development. At times unquestionably there exists an obvious similarity with some permanent form in particular parts ; but it is as readily demonstrable that this similarity is not the condition of the mal- formation, but the result of other relations, either, — 1st, because that form is nearer the fundamental type, in which case any stoppage at an earlier period of development must necessarily approximate the higher form to it ; or, 2ndly, because the altered formative conditions may approximate the formative conditions of the same part in another animal. Thus, for example, the nose in Man is sometimes elongated into a proboscis, which reminds one of the snout of a Pig. But the human nose never passes through any stage of development in which it resembles the snout of a Pig ; on the other hand, the Pig^s snout, at the fourth week of embryonic life, is not only similar to that of a human embryo at an early period, but is in fact much more like the nose of the adult Man than at a later period. This relation agrees perfectly with the general law. The nose of air-breathing Mammals in general does not project beyond the jaw. Both the peculiarity of the Pig's snout, therefore, and that of the human nose, arise subsequently without any transi- tion of the one form through the other. If then a Man has the snout of a Pig, it is no arrest of development, but the conse- quence of an abnormal development, which has a result like that in the Pig, where it is normal. While we are speaking of the abnormal forms of the nose, I will call to mind the "Wolf's jaws," an indubitable arrest of development, but which is cer- tainly no stoppage at any earlier form of animal. Second Corollary. — Application of the present View to the Determination of the separate Organs in the different Forms of Animals. A closer acquaintance with the history of development will sooner or later furnish us with the sole safe grounds in the de- termination of the fitting denominations for, and in forming a just judgment of, the organic parts of the different forms of animals. At present a little may be done in this direction. Since, in fact, every organ becomes what it is only by the K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. 223 mode in which it is developed, its true import can only be re- cognised by knowing the manner in which it is formed. At present we judge for the most part in accordance with an inde- finite feeling, instead of considering every organ only as an iso- lated development of its fundamental organ, and determining from this point of view the agreements and differences among the different types. Every type has, in fact, not only its funda- mental organs, but in each these are again divided into special organs, which cannot be exactly what they are in any other type. We therefore need some complete nomenclature, which shall not merely apply the names of organs found in the verte- brate type to the organs of other types, but shall give to them special names when they have a different origin. This require- ment will hardly indeed be satisfied in a century, but it is well to call attention to it. In fact, the immediate consideration of the perfect animal has often led to the recognition of the essen- tial difference; perhaps, however, the determining conditions have been less easily comprehensible. First of all, I would refer to the question how the series of ganglia upon the abdominal side of the Articulata is to be named. They certainly do not constitute a spinal cord, since this is composed of a nervous tube, which can be produced only upon that scheme of development, which is followed in the Ver- tebrata. As little are they comparable to the sympathetic nerve of the Vertebrata, for they supply voluntary muscles, and in the Articulata the plastic nervous system lies upon the dorsal sur- face*. They are rather the terminations of the symmetrical nerves of animal life, and, on that very account, as Weber and Treviranus have already remarked, they are comparable to the nerves and ganglia called spinal, in the Vertebrata, on account of their insertion into the spinal cord. In the Articulata, how- ever, these nerves have only one series of central and peripheral * Some time ago, indeed, I was inclined to consider the so-called recurrent nerves of Articulata to be their plastic nervous system, because I had traced them a long way in the Crab ; however, I first became fully instructed on this subject by a letter from Prof. J. Miiller. Miiller, by his exactness in investi- gation and delicacy in dissection, has succeeded in following out these nerves through the whole extent of the plastic organs ; and he has had the goodness to communicate an excellent drawing of them to me. 224 K.E.VON DAER. — PHILOSOPHICAL FRAGMENTS. ends, because the whole animal part of the body is singly, and not doubly symmetrical. Whether the anterior pair of ganglia of the Articulata is to be called a brain or not, depends wholly upon the signification given to the word ' brain/ It is assuredly not the organ which we call brain among the Vertebrata, for this is the anterior extremity of that nervous tube which is absent in the Invertebrata. It is rather the most anterior pair in the series of gangha, and since these are to be compared with the spinal ganglia of the Verte- brata, the so-called brain appears to be for the longitudinal type what the Gasserian ganglion is for the Vertebrata. This also receives nerves of sense. A great importance appears to be attached to the circumstance that it lies above the oesophagus. This, however, appears to me to be an erroneous view. Properly speaking, it only lies in front of the oesophagus. If, in fact, we form a purely ideal notion of the longitudinal type, the oral aperture is not placed at the anterior extremity, but is directed downward, just as the oral aperture of the Ver- tebrata is not situated at the anterior extremity of the vertebrate type, but is placed somewhat posteriorly towards the abdominal surface, for which reason a portion of the abdominal visceral plates, the walls of the nose, lie in front of and above the oral aperture. In the Chick it is very clear that the mouth opens below. That in the Articulata the oral aperture belongs to the lower half of the simple ring, is shown very clearly by the Crustacea, and even by forms which exhibit the type in a less altered form, as the Annelida. In the Earthworm, for example, this relation of the so-called proboscis, which extends beyond the oral aperture, is obvious. It contains the most anterior im- perfectly developed rings. Now if in the Articulata the oral aperture is in fact anterior, but yet upon the lower surface, and corresponds with the most anterior extremity of the abdominal surface, a pair of nervous ganglia must necessarily lie in front of the oral aperture, and that they lie nearer the upper wall than the posterior ganglia arises partly from the passage of the mouth, partly from the very fact of its occupying the anterior extremity. Very frequently it lies actually in the same plane with the others, as in the Crustacea, where the mouth lies further back ; and in Insects, where the head, with the oral aperture, is directed more K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. 225 or less downwards. Only in the Annelida is its position de- cidedly, though only a little^ superior. — (p. 235.) p. 236. The same holds good of the cervical nervous band in the Mollusca. It is not the organ which we call brain, not even in the Cephalopoda, but in a manner the central part of a nervous system, which, in its general relations, may be compared with the plastic nervous system of the Vertebrata, but which takes on a different form, inasmuch as it is not dependent upon a pre- dominant brain and spinal cord. I can regard the so-called brain of the Cephalopoda only as the cervical nervous band of the Gasteropoda. In the former the ganglia are fused together, in the latter they are more distinct. It is a centre of the plastic nervous system, and can only be compared with the ganglia which, in the Vertebrata, give off threads to the organs of the senses and other parts of the head — here, however, possessing no predominating centre, but being subordinate to the brain. If, in the Vertebrata, we consider the ganglion maxillare, which also receives a nerve from the ear, in combination with the ganglion caroticum, petrosum, Vidianum, ciliare, and the threads which pass to the organs of sense and the pharyngeal apparatus, we have a similar ring through which the commence- ment of the digestive canal passes. That every portion is to be understood only by its relation to the type and by its development out of it, is taught still more strikingly by other parts. The tracheae of Insects are indeed aeriferous organs, but not the organ which we call the trachea in the Vertebrata, because the latter is a development of the mucous canal, while the tracheae of Insects arise either by histo- logical differentiation or by involution of the external integu- ment. Sometimes the same name has been used for different organs only from the want of some other word — the difference, however, having been universally recognised. Thus no anato- mist has regarded the wings of Insects as of the same nature as the wings of Birds. In the feet, also, the essential differences of the first joints have perhaps never been overlooked. A special name has, with justice, been applied to the antennae. They have no representative in the Vertebrata. But that they are the wings of the cephalic ring is demonstrated not merely by their position, but by their mode of development. They have SCIEN. MEM.— Nat. Hist. Vol. I. Part III. 15 226 K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. the same relative position as the wings in the pupa, with this difference only, that they come from the head. For the same reason also they agree with the lateral appendages of the Crustacea. Whatever sensitive properties then these antennae may have, they are yet never the organs of touch, smell, or hearing of Vertebrata, but sensitive cephalic wings. By these remarks I wish to render it intelligible, why every type must be studied for itself, and possesses fundamentally peculiar organs which are never to be found exactly the same in any other type. In some cases, indeed, the distinction is but small. The alimentary canal arises in all animals from the sur- face which is turned towards the yelk. We have here the smallest original difference. But in its further subdivision into organs a distinction must be discoverable for the signification of the separate organs ; it is indeed, as we know, often difficult to recognise single divisions, as the stomach, &c. We shall suc- ceed better if we determine every part only after other animals of the same type. We know, for instance, how little the sexual organs of the Massive series can be interpreted by those of the Vertebrata. Still more striking is the tentacular system with its vessels, which is found in manifold variations among the Radiata, for the ciliated bands of the Beroidae and the circular vessel of some (if not of all) Medusae, should perhaps be re- garded as modifications of this system. In the Articulata and Vertebrata, however, we are acquainted with nothing similar. It is perhaps peculiar to the peripheral type. It will suffice merely to remark how little the notion that all animals are only detached organs of Man corresponds with na- ture. A few organs of the Vertebrata may, however, certainly be said to contain within themselves the organs of the Massive and Articulate series ; at least this appears to me to be probable with regard to the organs of sense. I shall perhaps, in another treatise, show that even in the Vertebrata, development alone can guide us in interpreting the signification of the organs. Third CoroIuLAUy.— Application to the Affinities of Animals, I have above ventured to assert (Schol. V. § 1.) that the notion of a uniserial succession of animals is the prevalent one, and I K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. 22? foresee that this assertion will be regarded as too sweeping, since only a few investigators of the present day declare themselves for it openly and decidedly ; in fact, not a few have pronounced decidedly against it. I must therefore devote a few lines to show the^correctness of my statement. That doctrine has, as I believe, far more unconscious than conscious advocates. It seems to me, indeed, that a number of conceptions, proceeding from the uniserial view, and belonging to times long past, have propagated IHemselves, and without our knowledge have given a colour to our view of organic affinities, which is not the result of investigation. Are not the notions that Cephalopoda or Crustacea are allied to Fishes, or even pass into them, expressions of this fundamental view ? They could hardly have proceeded from an immediate and free comparison of their organisms. Just as incomprehensible is the alliance between Echinoderms and Mollusks. Do not these attempts to build bridges between two distant countries proceed from the endeavour to make each ajink_ in a chain? Having learnt, in fact, to understand the Crustacea according to the type to which they belong, and regarding them as the most developed forms of this type (with which I do not agree), the next endeavour was to pass from them a step further. In the same manner it was believed that a way led from the highest Radiata to other regions. If, however, in accordance with our view, we regard the separate forms or groups of forms as variations upon a theme, we shall find that the transitions are but few and isolated; consequences of the modifiability of a form, but on that very ground not in themselves necessary and determinate. We shall then not be misled into seeking agree- ments between things which are heterogeneous, since we do not regard the serial succession as the condition of the varieties of animal forms. The controversy, whether the Articulata or the Mollusca are the higher, seems to me also to depend upon this view of a uniserial development. If we properly comprehend the essence of the different types, it appears easy enough to comprehend how the plastic formations predominate in the one, in the other sensitive and motor organs. The heart and the liver of Mollusks, 15* 228 K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. as well as their glandular system in general, then will not neces- sitate us to place them higher than the Articulata. It would be almost as one-sided to raise all these above the Mollusca, although in general the greater variety in their vital manifesta- tions might give them a fair claim to such a position.** Funda- mentally, however, each of these sections of the animal kingdom has its own standard, which can only be determined by its type. The greater the histological and morphological differentiation, the higher, according to our view, is the perfection in the same type. A less morphological differentiation is, however, always an ap- proximation to the fundamental type. Thus, the Annelida, on account of the similarity of their limbs, appear to us to be of lower organization, notwithstanding their vascular system, whose limitation in Insects is readily comprehensible, as a consequence of the development of their tracheae. For us the Myriapoda do not stand much higher, their manducatory instruments being true cephalic feet and their head being but little separated from the other almost identical rings. In the Thysanura and Parasita a greater morphological differentiation has arisen. And the structure of the true Insects is in them, as it were, sketched out. Just as gradual modifications of the Annulata may be re- cognized through Myriapoda, Thysanura and Parasita, so may such be observed through the Isopoda, Amphipoda and Stoma- poda to the Decapoda, and through the Scorpionidae to the Arachnida. For what reason, hov/ever, the proper Spiders, or the Decapoda among the Crustacea, are to be reckoned as more perfect than the true Insects, is not clear. On account of their more perfect vascular system ? This is only a consequence of a less active interchange with the air, whose more powerful influence always assists the development of animal life. If, on the other hand, the Individualization of the organic constituents is to be our standard of perfection, we observe in the Decapoda, besides the slight histological differentiation which to me is obvious, a tendency to compress the organs of sense, the motor and the plastic organs, into one chief centre, — in con- sequence of which the type is greatly modified, but its essential parts become little separated ; in the Arachnida the plastic body K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. 229 is at least separated from the animal; but in Insects with metamorphosis^ sensibility, irritability and plasticity are sepa- rated, though indeed only in the perfect state. The most highly developed among these, again, appear to me to be those whose thoracic segments are not divided into many separate rings, as in the Flea, the Coleoptera, and the Orthoptera, but when they are collected into one. It is in these that the originally similar parts, such as the feet and manducatory organs, have at- tained the widest deviation. It is in these that we meet with the best developed wings and the most various manifestations of life. The Crustacea, indeed, possess an ear and a nose ; but we must not forget that the head of Insects is small enough to make such organs doubtful ; that a few investigators believe they have found them, and that, at any rate, the senses are not absent. If we have succeeded in getting rid of all preconceived notions of a gradual series, we shall consider each form to be a modifi- cation of a more general form, and that these are modifications of a fundamental type, and learn to comprehend them from this point of view. We shall then take more pains to determine the general affinities of each species, than its place in a univer- sal progressive series. If we seek, however, for the grade of development, we must look for it only according to the degree of differentiation of the parts, and within the type which belongs to the animal. That, however, in fact, the preconceived conceptions of a progressive series have led to the ordinary views, is what I shall endeavour to establish and explain by yet a few examples. We often hear of retrogressive metamorphoses of a whole animal form or of a single organ ; — can any clear conception attach to these " Retrogressions," if we do not assume that the form of some one animal is the condition of the form of some other ? This much, however, is certain, that such a view impHes the conception of a serial progression. If, in fact, we arrange together obviously allied animals, and then place them wuth their highest forms below those of another series, we shall have a retrogression. I will but shortly refer to the above (Schol. V. § 3 «) example of Fishes. In truth, the retrogressive metamorphosis of particular organs 230 K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. is spoken of, under the supposition that there is a progressive development for every organ, from the Monad to Man, and that this development is realized in accordance with the order of progression of the animals, the particular exceptions from which are now given. If, however, the organs are modifications of fundamental organs, and these differ according to the scheme of development (compare the following Corollary), there would appear to be an erroneous assumption in this very supposition, I believe, therefore, that if comparative anatomy is to be directed to the ascertainment of the laws of formation, the only proper way is, besides the constant reference to a fundamental type to which the whole animal belongs, to compare the organs by themselves in their different forms, as Burdach has attempted to do in his ' Physiology,^ without arranging the forms in the same order as the animals to which they belong would take in accord- ance with their perfection in other respects. We shall thus perceive how the general structure of the whole body of an ani- mal, or its relation to the external world, operates in modifying the form of particular organs, and so prevent ourselves from being misled by prejudices. That, however, these retrogressions in the development of organs are only an appearance which depends upon a pre- supposed uniserial development, is rendered most obvious by their disappearance, if we arrange the animals according to another organic system than that which was previously used as a base. I bring forward one example out of many. If I am persuaded that the Articulata are to be arranged in a progress- ively developed series, and dispose them according to the de- velopment of the vascular system, I may order them thus : True Insects, Myriapoda, Arachnida, Annelida. In this case the eyes are retrogressive through the series. Of the respiratory organs, and the vascular system besides, it is at once intelligible that the one appears retrogressive with respect to the other, since these systems are antagonistically related. If I consider them as modifications of a fundamental type, in which sometimes one and sometimes the other system is more changed from the simple fundamental form, then all retrogressions disappear. What 1 have here said of the Articulata, in order to select an obvious example, holds good not of them alone, nor merely of K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. 231 the antagonistic relation of respiratory organs and vascular system ; — it appears wherever there are any manifold variations. If we regard the different forms of Mammalia, we find for one series of organs other affinities than for another. If we consider the development of the animal part, which we may most readily estimate by the skeleton, the Bats are very different indeed from all the proper quadrupeds. We must suppose that they form the most aberrant order. With respect to their digestive organs, they resemble the Insectivora. Pallas, who in the Zoographia Rosso-Asiatica unites the Bats closely with the Mole, appears therefore to me to be as fully justified as Tiedemann, who about the same time, in his ' Zoology,' separated them widely. For the same reasons, Tiedemann unites the Seal with the Dugong, while by Pallas they are widely separated. The one has regarded the extremities, the other the teeth. What do such facts prove, except that the different organic systems vary in different modes ? The Mole and the Bat seek the same prey, the one in the air, the other in the earth. Their motor organs therefore differ according to the place of their residence. The Dugong and the Seal are both aquatic, and have fin-like extremities, but what they seek is totally different; so are their dentition and their stomachs. Under such circumstances, does not an approximation to Man give for every organic system a different series of animals ; and if this be so, are not " retrogressions'' without meaning? In truth, Man is only in respect of his nervous system, and of that which is connected with it, the highest form of animal. His erect progression is only a consequence of the greater develop- ment of his brain, since we find everywhere that the more the brain preponderates over the spinal cord, the more it raises itself above it. If this remark be well founded, all corporeal distinc- tions between Man and other animals may be reduced to cerebral development ; and in that case the pre-eminence of man is only a partial, although the most important one. One must be com- pletely prejudiced, in fact, not to see that the stomach of the Ruminant, which changes grass into chyle, is more perfect than the stomach of Man. — (p. 242.) 232 K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. (p. 257.) Fourth Corollary. ' nl ill ..J ■ hita i:x ^ ., , JPut from what has been said, it seems to follow that every principal type of animal organization follows a special scheme of development ; indeed nothing else could have been expected, since the mode in which the parts are united together can only be the result of the mode of development. In reality, therefore, I might have used, instead of Type and Scheme, a common term expressing both. I have only kept them separate in order so to make it obvious, that every organic form, as regards its type, becomes by the mode of its formation that which it eventually is. The scheme of development is nothing but the becoming type, and the type is the result of the scheme of formation. For that reason the type can only be wholly understood by learning the mode of its development. This introduces differences into the germs, which at first are alike in all essential points. Different conditions or formative powers must act upon the germ in order to produce this multiplicity, on which head we shall by and by raise one or two queries. Here, however, we must add the remark, that the original agreement of all animal germs does not completely disappear even in the perfect forms, and that we have to seek the most profound distinctions among animal forms, which are attainable by us in the mode of development. With respect to the original agreement, I would call to mind, that, according to the Corollary of the Second Scholium, every animal is at first a part of its mother, that it becomes indepen- dent either by the immediate development of the parent herself, or after the action of a male principle, and that then the first act of independence consists in passing into a vesicular form ; either the whole becoming the body of the new animal, or the future body (the germ) separating itself from the merely nutri- tive substance which surrounds it. Here animals and plants diverge, since the latter do not invest the nutritive matter. The vesicular form, therefore, is the most general character of the animal ; the contrast of external and internal surface is the most general, and therefore the most essential contrast in the animal. (See above, Schol. V. § 4 d.) K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. 233 There remains yet another agreement between all forms of development. In all animals, namely, which at an early period possess a germ and a yelk, the investing germ becomes divided into many layers ; that turned towards the yelk is the plastic receptive layer; that turned from it, the more animal, even though its outermost border should become merely a limiting organ, and clothe itself more or less with an excreted non-vital layer. That now the vascular system, if it is otherwise sepa- rated from the digestive cavity, is formed external to it nearer to the animal part ; that in the animal part, muscles, nerves, &c., become separated, appears also to belong to the Idea of the ani- mal in general ; and the further this histological differentiation goes, the more developed do we call an animal. Wholly different from this, however, is the relative position of the parts. This is determined by the external form of the development. We have distinguished four principal forms, or, as we have called them. Schemata of Development : — Radiate Development {Evolutio radiata), which, proceeding from a centre, repeats similar parts peripherally. Coiled Development (Evolutio contorta), in which similar parts are twisted round a cone or other space. Symmetrical Development [Evolutio gemina), in which similar parts are distributed from an axis on both sides to a sutural line opposite to the axis. Double Symmetrical Development [Evolutio bigemina), in which from an axis similar parts are disposed on both sides above and below, and are united at two sutural lines ; so that the inner layer of the germ is closed below and the upper layer above. We know that in the higher Vertebrata the Germ soon divides into two parts, — an inner, which may especially be termed the Embryo ; and an outer, which may be called the Germinal mem- brane. I have already remarked that the former is nothing but a part of the germ, which is metamorphosed according to the scheme of development peculiar to every animal, whilst the peri- pheral part remains behind in its development. In Mammalia, Birds, and Reptiles, the middle part is but small in proportion to the outer portion, and it gradually sur- 234 K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. rounds the yelk, forming the abdominal suture. In the Frog, indeed, the external part of the germ is very thick, yet, as I believe, a separation into embryo and germinal membrane is un- deniable ; for the middle portion at the period in which the back closes is still very much thicker, and the limitation is tolerably well marked between the abdominal plates and that exterior por- tion w hich I regard as the germinal membrane. The former grow together towards the suture of the abdomen. In very young Pikes, also, in which the germinal membrane is much thinner and more transparent, this appeared to me to be the case. 1 saw beside the trunk of the vertebral column a pair of very narrow dark striae, the commencing abdominal plates. It may hence, I believe, be affirmed universally of the Vertebrata, that the embryo will surround the yelk with its abdominal plates, although this has previously been invested by the germinal membrane. It is the same in the Articulata. Their lateral plates are clearly distinguished by their thickness from the proper germinal membrane. They also surround the yelk. In the Mollusks, however, the whole germ appears to change equally. It cannot therefore be said of them that the embryo grows round the yelk, but more justly, that from the moment of fecundation it remains as its investment ; for a differentiation of the germ into embryo and germinal membrane is not perceptible ; the whole germ, rather, becomes the embryo. The same would very probably also take place in the radiate type, if any animal form from this series were developed from a true ovum, of which we have no knowledge*. If they should all be developed from mere germ-granules, the relation is still more apparent; for every germ-granule, so far as we know", is developed as a whole, and is nothing but a germ without a yelk. We must not overlook here an interesting relation. In those ova in which the germ is clearly separated into an embryo and a germinal membrane, it is the animal part of the embryo which * At present hardly anything could be of more interest for the history of development than the observation of the development of the Asteridae, and after these of the Cephalopoda. According to Cavolini, the latter have a yelk- sac depending from the mouth {Abhandl. iiber die Erzeugung d. Fische und Krehse uebers. von Zimmermann, 1792, p. 54), which is difficult to understand. K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. 235 causes this separation. It is the animal part which grows so much, that we observe the marking off of the embryo from the germinal membrane. It is only after this has determined the whole form of the animal, that the plastic part appears to attain a certain degree of independence, which in the Articulata is often limited to a mere separation, the separate organs arising subse- quently, but which in the Vertebrata has sufficient power to cause a symmetrical development of the animal part. Of the action of the plastic part upon the animal, we can but detect a trace here and there. It is otherwise in the Mollusca. The plastic part becomes independent very soon, and has a decided influence upon the external form. We see how the essential character of the animal manifests itself very early, and develop- ment affords a justification for the name of Plastic animals which has been given to the Mollusca. Hence also we shall be better able to judge how far the Mollusca may be justly compared with the vegetative section of the body of the Vertebrata ; in its pre- dominant character namely, not according to the sum of all its separate parts. In the Mollusca, it is also a relatively animal part which occupies the whole periphery, and is principally developed in the foot of the Gasteropoda. Compared with other animals, they are living bellies ; but since these bellies are independently developed, without the influence of a more highly- organized animal part, they have also a part which for them is more animal, and it is that which originally formed that surface of their germ which was turned from the yelk. In all four forms the surface of the germ which is turned towards the yelk does not change its position relatively to the latter, but retains it, and becomes the digestive surface of the perfect animal. In all forms, moreover, the peripheral part of the perfect animal is the external surface of the germ, that which is turned away from the yelk. I therefore believed myself to be justified in supposing that it is the relation to the yelk which in the germ produces the primary differentiation into an animal and a plastic layer. But it is not in all animals that the whole outer layer of the germ remains exterior. In the Vertebrata, the one half of the doubly symmetrical development encloses a part of the outermost surface, and changes into the nervous tube, the spinal cord with 236 K. E.VON BAER. — PHILOSOPHICAL FRAGMENTS. the brain, parts which must therefore necessarily be absent in other types. I would here render perfectly obvious how it is the scheme of development which produces the principal cha- racter of the animal. If we suppose that, in any articulate animal which is in the earliest stage of its development, a part of the germ should become raised up on both sides, and so enclose a portion of the external surface, the enclosed part would be an animal central portion. The internal organs would all have the same relation to it as in the Vertebrata, ex- cepting the plastic nerves, which by the influence of the ani- mal nervous system appear to be approximated to this last in the Vertebrata. - In relation to the external world, however, all the internal organs would be inverted, since the central part itself would lie downwards. If we were to invert the animal, all the outer parts, the extremities and the organs of sense, would be dis- placed ; and supposing that the extensor and flexor sides had not undergone inversion by the addition of the new central part, these also. Hence we conclude, that, by the origin of a central part for the animal body, the position of the plastic organs, and their relation to the nearest animal layer, have indeed remained unchanged ; but their relation to the external world, and all which represents this relation in the body, has become inverted. In the former case, where the course of development is simply symmetrical, the central line from which it proceeds becomes the flexor side of the animal ; with a doubly symmetrical development, the side from which it proceeds becomes the extensor side. Towards the flexor side, the extremities and the feet are developed. By this it shows itself to be that which is most essentially turned towards the planet. Towards the extensor side, that turned from the ground, the organs of the senses are developed. I commenced this Corollary with the remark, that animals ought to be divided according to their mode of development, and I have shown sufficiently at length that the principal types have their own form of development. I may be permitted to point out here, in a few words, that the safest guide for further division would be found in the history of development, if we were acquainted with it with sufficient exactness, in the different K. E. VON BAER. — PHILOSOPHICAL FRAGMENTS. 237 classes and families of animals. If we keep this in view, we shall easily recognise the true Insects as a higher stage of deve- lopment than the Arachnida and Crustacea. We shall hold the Batrachia to be diflferent enough to separate them, with De Blainville, from the Reptilia as a distinct class. And, in fact, what have they in common with the latter, than that they are not Fish, nor Birds, nor Mammals ? y^ jgjfc-i(^5:tT^j / Scholium, yi^T7r,Ge«fra/i2e*w/t,,-^^^ fj^ ^^,1 , If we review the contents of all the Scholia, a universal result proceeds from them. We found that the action of reproduction consists in elevating a Part into a Whole (Schol. II.) : that, in the course of development, its independence in relation to that which is around it increases (Schol. II.), as well as the determi- nateness of its form (Schol. I.): that in the internal development, the more special parts are developed from the more general, and their speciality increases (Schol. III.) : that the individual, as the possessor of a determinate organic form, gradually passes from the more general form into the more special (Schol. V.) ; and therefore the most general result of these investigations and considerations may well be thus expressed : — TTie history of the development of the individual is the history of its increasing individuality in all respects. This general result is indeed so simple, that it would seem to need no demonstration, but to be cognizable a priori. But we believe that this simplicity is only the stamp and evidence of its truth. If the nature of the history of development had been from the first so recognized as we have just expressed it, it would and must have been a deduction thence, that the indi- vidual of any particular animal form attains it by passing from the more general into the special form. But experience every- where teaches that deductions become much more certain if their results are previously made out by observation. Man must have received a greater spiritual endowment than he ac- tually possesses for it to be otherwise. If, however, the general result which has just been expressed be well based and true, then there is one fundamental thought which runs through all forms and grades of animal development. 238 K. E. VON BAER. PHILOSOPHICAL FRAGMENTS. and regulates all their peculiar relations. It is the same thought which collected the masses scattered through space into spheres, and united them into systems of suns ; it is that which called forth into living forms the dust weathered from the surface of the metallic planet. But this thought is nothing less than Life itself, and the words and syllables in which it is expressed are the multitudinous forms of the Living. DESCRIPTION OF FIGURE. Ideal figure of the organic movements in Verte- brate. The body of the animal is suj)posed to be transparent, so that the outline only is seen. The outline of the heart is indicated (the right auricle is drawn rather too much backwards). The view is from the dorsal surface. 1'. Course of the red blood into the left ventricle. 1. Course of it out of the left ventricle. 2. Course of the venous blood from the anterior half of the body into the right auricle. 3. Course of the venous blood from the posterior half of the body into the right auricle. 4. Course of the portal blood. 6. Course of the respired air. 6. Course of the food from the pharynx into the oesophagus. 7. Course of the chyme from the stomach into the intestine. 8. Course of the faeces. 9. Course of the ova. Cr. H. H.] W. HOFMEISTER ON THE DEVELOPMENT OP ZOSTERA. 239 Article VIII. On the Development o/*Zostera. By W. Hofmeister. [From the Botanische Zeilung, Feb. 13th & 20th, 1852.] 1 HE peculiar physiological pbaenomena exhibited by Zostera, above all the unique formation of the pollen and the strange structure of the embryo^ have frequently and from an early period drawn the attention of botanists to this remarkable genus of plants. In most of the larger illustrated works devoted to the local floras of Europe, the figures oi Zoster a are accompanied by careful and more or less accurate microscopic dissections, as in Schkuhr, Hooker's Flora Londinensis, Reichenbach's Icones, Schnizlein's Iconographie der naturlichen Familien, &c. The form of the embryo has been a subject of discussion for almost every author who has studied the comparative import of the parts of the seedling plants of Monocotyledons*. Fritzschef made known the curious condition and the circulation of the contents of the pollen-cells ; GronlandJ has very recently pub- lished a contribution to the knowledge of Zostera marina, — a series of most acceptable microscopic researches on the develop- mental history: imperfect, however, in reference to several of the most interesting questions, especially to those connected with the origin of the pollen and of the embryo. The following essay, the materials for which I owe to the kindness of Prof. Nolte of Kiel§, will in some points complete, in others correct, the paper published by Gronland. The inflorescence of Zostera, like that of the nearly allied Potameae, is relatively terminal : the metamorphosed end of a * G'ax\.x\ex,De Fructibus,iA9', Bernhardi, Zi;z»^a, Bd. vii. ; Jussieu, y^wwaZf* des Sc. nat. 2me Ser. t. xi. p. 356. t Mem. de VAcad. de Si. Petershourg par div. Sav. t. iii. p. 703. pi. 3. X Bot. Zeitung, vol. ix. p. 183 (1851). § The plants which Prof. Nohe was kind enough to forward to me repeatedly, at different seasons, were perfectly fresh when they arrived at Leipsic (30or 40 hours after their removal from their native locality) ; even the adherent Chato- morphee and Polysiphonics retained their vitality unaffected. 240 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. branch. In Zoster a marina, when the formation of flowers commences, both the previously inert axillary buds and also the terminal shoot of the little-branched or even simple sterile plants, become converted into inflorescences. In Zostera minor the rule is for axillary sprouts standing far back from the sterile terminal shoot, to become transformed into blossoms. This distinction between the two species, so striking at first sight, appears to depend principally upon the fact that in Zostera marina the old portions of the stem die away rapidly, within a few months, from behind forwards, while in Zostera minor they persist for more than a year. When the terminal bud of a plant oi Zostera marina enters upon a transformation into a spadix, the ordinary series of linear stem-leaves with sheathing bases (having the divergence i), becomes interrupted by a cylindrical leaf-sheath, devoid of any indication of a lamina, the two lateral borders of which sheath, enclosing the terminal bud, adhere firmly together. The axillary sprouts which become blossoms arise with a similar sheath. The next internode of the flowering sprout bears a leaf like the stem- leaves {laub-bldtter),hMiyv\th. a comparatively shorter lamina, which, hke the leaves of the Zostera generally, has a divergence of ^ from the leaf next older than itself. The sheathing base of this leaf envelopes the inflorescence, i. e, the expanded, flat, long-protracted end of the stem, that surface of which turned away from the last leaf bears the anthers and ovaries. From the axil of the sheathing bract (vorblatt) of the fertile sprout (PI. VI. fig. 1 a) arises a shoot, resembling in all respects its parent sprout, with which it is coalescent for a considerable distance (fig. 1 b). At the place where the cohesion ceases the new sprout bears its first, sheathing leaf (fig. Id ^). This diverges J of the circumference of the stem from the bract (vorblatt) of the sprout of the preceding rank, and is parallel with the stem-leaf {laub-blatt) in the axil of which that arises. The sheathing leaf is succeeded by a stem-leaf {laub-blatt) (fig. 1 b ^), above which the sprout terminates with the inflores- cence (fig. 1 b). The axil of its lowest, sheathing leaf sends out a fructifying branch of a new rank (fig. 1 c), and so on, till the exhaustion of the growing power of the blossoming plant, in W. HOFMEISTBR ON THE DEVELOPMENT OP ZOSTERA. 241 Zostera marina, to a series of twelve, in Zostera minor only of six ranks. The rudiments of the first spadix of Zostera marina appear at the beginning of spring, those of Zostera minor in the middle of summer. From then until the period of maturation of the first seeds (in Z. marina at the beginning of July, in Z. minor the beginning of September), new flowering sprouts arise suc- cessively in the axils of the bracts {vorblatter) of their predeces- sors, so that, from the middle of May forwards, every fertile plant of the common Zostera exhibits a series of distantly situated stages of development of the floral organs*. When arising out of the axil of the sheathing leaf Of the pre- ceding sprout, the axis of a new rank, terminating in an inflores- cence, presents itself as a hemispherical papilla (fig. 1 d), the side of which facing the parent sprout coheres with the latter up to the point where the new shoot forms its first, sheathing leaf, by the simultaneous multiplication of a zone of cells situated close underneath its apex. The multiplication of the cells of both organs, axis as well as leaf, in the longitudinal direction, commences with a repeated subdivision of their apical cells — in the axis of a single one, in the leaf of an incomplete crown of similar cells — by means of alternately obliquely inclined walls. The fertile shoot forms its single stem- leaf [laub-blatt), oppo- site to the leaf-sheath, shortly aft;er the latter breaks forth. The bud of the sprout of the succeeding rank appears simultaneously in the axil of the basilar sheath. The stem-leaf {laub-blatt) makes its appearance as a flat cellular mass embracing the end of the stem so completely as to leave only a narrow slit. It first grows longitudinally, by re- * The interpretation of the series of sprouts of the flowering Zostera given above, placed beyond doubt by investigation of the development, is also the only one possible, looking exclusively at the condition when complete. Neither is the supposition admissible that the flowering sprouts, upwards from the sheath- ing leaf, are a series of equivalent secondary axes of a main axis bearing only sheathing leaves, for the sheath is opposite to the sprout terminating as an in- florescence ; — nor is that which would make the inflorescence an axillary bud of the stem-leaf {laub-blatt) which sheathes the spadix with its base, for this as- sumption would lead us to expect each couple of leaves of the main axis, a stem-leaf (laub-blatf) and a leaf-sheath, to stand regularly one above the other; a conception untenable when we recollect the ^ position of the leaves of the sterile plant of Zostera. SCIEN. MEM. -^Nat. Hist. Vol. I. Part III. 16 242 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. peated subdivision of the cells of its upper border. The multi- plication of these cells, on the side turned away from the slit, of this organ (hitherto resembling a cylinder slit up lengthways), soon outstrips that of the apical cells of the other half of the leaf: the riband-shaped lamina grows out from the sheath. The multiplication of the cells at the apex of the leaf ceases early, while a very active and long persistent increase occurs in those of the base. After the rudiment of the stem-leaf [laub-blatt) has been formed, the end of the stem above it changes its form. It be- comes widened out by a multiplication of its cells, predominantly in the direction parallel to the surface of the leaf, and it assumes the shape of a thick spatula, slightly concave at the side nearest the next (younger) sprout. The rudiment of the flattened spadix is now unmistakeable in the end of the axis. The longitudinal growth of the spadix is continued by a frequently repeated divi- sion of its apical cells by means of walls inclined alternately towards the upper and lower surfaces of the leaf-hke organ. The cells in the median line and at the margins of the spadix remain long capable of multiplication; constituting strings of cambium structure. On the other hand, two broad strips of cellular tissue, parallel to the lateral borders of the spadix, early cease to increase. With the exception of those composing the epidermis, all these cells become separated from each other at their angles; air becomes excreted in the intercellular spaces (fig. 3). The cells of both the marginal strips of cambium on the upper side of the spadix become greatly multiplied through repeated division by walls parallel to its surface. In this way are formed two swollen tracts of cellular tissue, which, in their further development, become curved strongly inwards so as to cover in a great part of the upper surface of the spadix, through the resistance which the sheath of the stem-leaf {laub-blatt), closely enveloping the inflorescence, opposes to the unfolding of them. In Zostera minor little leaf-like structures arise upon the spadix near the lateral margins of its upper face, and these lie over the surface of the spadix like clamps or claws. Zostera marina exhibits no trace of such organs. From the median streak of the spadix, which becomes greatly expanded by the active multiplica- tion of its cells, arise the anthers and ovaries, upon the upper face. W. HOPMEISTER ON THE DEVELOPMENT OF ZOSTERA. 243 arranged with strict regularity, as is well known, according to 2^, in such a manner that single anthers alternate with single ovaries in each of the two longitudinal rows of floral organs, and organs of different sex constantly stand next together horizontally. The anther first appears as a longish papilla of cellular tissue, the larger diameter of which is parallel to the longitudinal line of the spadix (fig. 4 a). The two ends of the young anther soon appear curved outward to some extent ; they rapidly increase in magnitude, and become globular protuberances (fig. 3 a, a), which gradually become spindle-shaped through continued rapid mul- tiplication of the cells. Thus each anther consists of two appa- rently independent, moderate-sized halves, which are connected together by a comparatively long riband-shaped cellular body, the altered connective, attached to the spadix by the narrow edge*. At the time when the half of the anther is passing from the globular into the ovate form, two parallel rows of cells lying in its longitudinal axis, take on a different character from the sur- rounding tissue (PI. VI. figs. 5-7). The process of division ceases in the former, while the three layers of cells of the latter continue to multiply. The cells of these parallel rows are the primary parent-cells of the pollen. Their form is pretty nearly that of a cube (fig. 7) ; in their subsequent development they become elongated into short prisms, in a direction inclined downwards from the surface of the spadix. The pollen-cells are developed from these cells through re- peated longitudinal division, by means partly of perpendicular, partly of horizontal walls (figs. 6-12). The process of their formation is thus very unlike that occurring in the great majority of Phanerogamia. There is not the slightest indica- tion of parent-cells becoming isolated, or of special parent- cells. There occurs in the primary parent-cells a series of " halvings '' or bisections in only two directions, differing in no respect from vegetative cell-multiplication ; the last genera- tion, alone, of the daughter-cells become disconnected and so form the pollen-cells. This development of the pollen oi Zoster a * This part of the development of the anther, very completely treated and correctly figured by Gronland, loc. cit. 187, is mentioned here merely for the sake of giving a connected account. 16* 244 W. HOPMEISTER ON THE DEVELOPMENT OF ZOSTERA. is not, however, altogether an isolated phaenomenon ; the earlier stages of formation of the pollen-masses of the Asclepiadaceae resemble it most perfectly. In the anther of the Asclepiadaceae, the laws of growth of which correspond on the whole to those of the rudimentary spadix of Zostera, — and which, like this, is at an early age shaped like a little sloped- off spatula with a strongly rounded back, — two groups of longitudinal rows of cells become the primary parent-cells of the pollen*. Extending considerably inwards in a direction perpendicular to the surfaces of the anthers, they acquire a different character from the sur- rounding tissue and assume the form of recumbent prisms. By a series of longitudinal and transverse divisions parallel to the long axes of these prisms, the rudiment of the pollen-mass be- comes a group of narrow cells of a length six to ten times greater than the cross diameter. At this stage of development it cor- responds perfectly to the contents of a chamber of the young anther o^ Zoster a. But in the Asclepiadaceae there now succeeds a many times repeated division of the elongated parent-cells, by means of walls perpendicular to their long axes, whereby the pollen-mass becomes transformed into a body composed of trans- verse rows of cubical cells — the special-parent-cells. These special-parent-cells become polyhedral by unequal expansion of the whole mass, and then a pollen-cell originates in each. The pollen-cell of Zostera appears at its origin as an obtuse- angled, almost cylindrical sac, the long diameter of which is three or at most four times as great as the cross diameter (figs. 11-13). The nucleus, which it is difficult to make visible, vanishes at an early period. With the formation of the pollen commences a very consider- able enlargement of the pollen-chambers (loculi), by vigorous multiplication and expansion, in a tangental direction, of their * Schacht {Das Mikroskop, Berlin 1851, p. 154) assumes that only one lon- gitudinal row of cells becomes converted into primary parent-cells. His figure (tab. iii. fig. 8) shows the cross section of the attenuated upper end of an already further developed group of such cells (a pollen-mass), where one primary parent-cell, divided into two, is visible. Longitudinal and transverse sections of earlier conditions show, beyond doubt, that several longitudinal rows of cells of the tissue of the rudimentary anther, of unlike value (in Nageli's sense, /. e. not all products of an equal number of subdivisions), take part in the forma- tion of the pollen. I shall recur to this subject in another place. W. HOPMEISTER ON THE DEVELOPMENT OP ZOSTERA. 245 cell- walls. The pollen-cells, constantly growing longer, keep pace with the enlargement of the loculi ; they thus soon attain a con- siderable (fig. 14), and finally, in proportion to the permanently small cross diameter, quite an enormous length (in the ripe anther as much as 3 lines). They retain throughout their original direction, inclined downwards from the surface of the spadix. Toward the close of the increase of length they displace the innermost of the three layers of cells which originally formed the outer wall of each loculus. A cellulose membrane becomes first distinguishable on the pollen when the long diameter is about eight times the cross diameter, and at that period it is yet extremely delicate. As the pollen approaches maturity it be^ comes tougher, but no secretion whatever of an exine occurs ; as was made known by Fritzsche. The tendency of the pollen- cell to expand its membrane often outruns the growth of the walls of the loculus, towards the approach of maturity. The pollen-cells, restrained in their natural extension, then frequently exhibit curvatures or curling of the ends and protrusions of various forms (PI. VI. fig. 15 6). As long as the longitudinal growth of the pollen lasts, the granular-mucilaginous fluid contents are uniformly diffused in the cell. Toward the time of maturity we find longish cavities in the mucilage, filled with a fluid of less refractive power, and these are finally blended into a single cavity lying in the axis. At this period we may frequently observe the active currents in the dense, mucilaginous coating (containing numerous granules) of the inside of the cell-wall ; more distinctly in proportion as the pollen-cell is riper and the temperature of the surrounding water higher. Two principal currents may be distinguished, one ascending, the other descending ; one of these is ordinarily stronger and of larger size. The moving mass splits up here and there into several arms, sometimes becoming again confluent, between which remain isolated spaces occupied by motionless and more transparent fluid (fig. 15.) The pollen-cells in which a particularly rapid circulation occurs are always greatly swollen up (lying in fresh w^ater), and ordinarily t)urst in the course of a few minutes during the observation. Change of temperature in the surrounding medium has a most decided influence on the circulation in the pollen-cell. When warmer or cooler water is applied upon the object-holder, one of the two currents is 246 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. greatly strengthened, the other weakened, sometimes even so far as to disappear. It is not improbable that the different temperatures of the ends of the filiform cell form one of the causes on which the phaenomena of motion of the granular mucilage are dependent. The investigation of the development of the pollen of Zostera is very difficult. The contents of the cells of the young anther, especially those of parent-cells of the pollen-cells, and of the latter themselves, are extremely sensitive to the action of pure water. This holds in a still higher degree of the pellicle form- ing the coat of the young pollen. When an available preparation, a very delicate longitudinal section of a young half-anther, is placed under water, the membranes of the young pollen-cells swell up, and, with their mucilaginous contents, instantly run together into a shapeless jelly. The only method of examining the young anther is to place it in a saline solution ; I employed here, as in similar cases, a saturated solution of carbonate of ammonia. No safe conclusions can be obtained regarding the structure of the anther, or the course of development of the pollen, without making sections. The connection of the different tissues is so intimate in the young anther, that it is quite im- possible to extract uninjured either the parent-cells of the pollen or the very young pollen-cells*. The first visible rudiment of the ovary is a flat papilla (fig. 3 b), composed of few cells, which, increasing in size, soon acquires the form of a horse-shoe with the convex side turned towards the median line of the inflorescence (fig. 4 b), the rudiment of a leaf attached upon the surface of the spadix. In a short time, this becomes closed in so as to form an annular wall of cellular tissue, within which appears a little, roundish mass of cellular tissue (fig. 16a), approached near to the inner border; this is the axillary bud of the carpel — the ovule. The circular wall becomes developed particularly quickly on the side next the base of the spadix. It bulges out at this place, while it grows * The globular cells which Gronland mentions as the contents of the loculi of the young anthers, and likewise figures (/oc. cit. p. 188, figs. 18, 19, 20), I was unable to find, at any stage of development of the anther. The closely packed, straight and parallel pollen-cells always completely filled up the cavity of the anther. Probably those appearances depended on some of the cells of the lax and mucilaginous innermost layer of the wall of the anther being accidentally detached in the preparation of the object. W. HOFMEISTER ON THE DEVELOPMENT OP ZOSTERA. 24? up into an obtusely conical hollow mass of cells perforated at the apex (fig. 17). The ovule unfolds itself within the bulging portion of the cavity, becoming curved downward. Simultane- ously with the sudden commencement of the very considerable elongation of the perforated mouth of the young ovary, to form the canal of the style, begins the formation of the two integuments (fig. 18). They arise close beneath the summit of the young ovule, by a pretty nearly simultaneous commencement of rapid multiplication in two zones of cells ; the development of the outer coat of the ovule begins immediately after the inner coat becomes visible. The lower part of the ovule, at this epoch by far the larger, remains uncovered by the integuments. Both integuments grow longitudinally by constantly repeated division of the crowns of cells at their summits, by walls alter- nately inclined to and from the nucleus (fig. 19). The outer integument soon acquires considerable thickness through two or three repetitions of the division of the cells of the second degree, by means of walls parallel to the free outer surface ; when the ovule is fully developed, this coat is composed of lax tissue, traversed by intercellular spaces full of air, enclosed by an epithelium formed of cells one-fourth the size and filled with watery fluid. The cells of the second degree of the inner integument, do not become multiplied in the direction of the breadth until this coat has grown up beyond the summit of the nucleus. Then, how- ever, its upper margin rapidly becomes broader and thicker through repeated division of the cells by means of walls parallel to the long axis of the ovule ; the mouth closes in, leaving only the narrow canal of the micropyle (PL VII. fig. 22) formed solely by the inner integument ; the outer coat of the ovule, which comes to a standstill when the inner begins to increase in size, becomes grown over by the thickened margin of the micropyle. At the epoch when the integuments reach the level of the summit of the nucleus, the latter is composed of an axial row consisting of a few, eight to ten cells, enclosed by a double layer of cells. A simple layer of cells covers the upper extremity of the axial string and forms the summit of the nucleus (fig. 19). As in most young organs of vegetables, the cells forming the free outer wall of its surface possess tolerable firmness, different from the gelatinous consistence of the cell-walls in the interior. 248 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. The uppermost three cells of the row occupying the longitudinal axis of the ovule are distinguishable from their neighbours, at a very early period, by a larger size and greater concentration of their contents ; this is particularly the case with the uppermost of them, which gradually grows up to be the embryo-sac, while the nucleus undergoes a profound change of structure by an active multiplication of the cells, especially of those at its lower part. The cells of the surface of the upper end of the nucleus are divided by longitudinal and transverse septa perpendicular to the outer walls ; the same division is repeated in the " daughter- cells/' The increase of this cellular hood, in length and compass, thus keeps pace at first with that of the cell growing into an embryo-sac (fig. 20), enclosed by it. A little later begins a very rapid multiplication of the cells of the lower part of the nucleus, in all three directions. In the first place the axial cells divide by perpendicular longitudinal walls; the newly-formed cells again divide by longitudinal walls standing at right angles to those last produced. Thus from this time the embryo-sac no longer rests upon a simple row of cells, forming the axis of the mass of cells below it, but on a column formed of double pairs of cells, which cells become repeatedly divided by horizontal walls (fig. 21). The multiplication of the cells of the two peri- pherical layers of the nucleus, only weak just below the growing embryo-sac, becomes greater proceeding towards the base, where the hitherto cylindrical mass of cells soon becomes strongly bulged out. The multiphcation of cells still continues in this region even after the ovule has attained its normal size, so that at the epoch of fertilization this part of the nucleus is composed of cells one-fourth the size of those lying close beneath the embryo-sac. Finally, the cells of the periphery of the nucleus all divide once oftener by longitudinal and transverse walls than those in contact with them ; they thus appear one-half the height and breadth of the latter. During the further increase of size of the embryo-sac, the cells surrounding it become gradually broken down and compressed. The firmer, free, outer walls of the superficial cells are not attacked by this softening and solution of the cellulose ; after the breaking down of the inner tissue has advanced to a certain point, these outer walls of the superficial cells form a homogeneous, connected I W. H0FMEI8TER ON THE DEVELOPMENT OP Z08TERA. 249 membrane, transparent as glass, which encloses, as a sac, the semi-fluid mass of primordial utricles set free by the solution of the cell-coats (fig. 21). The large primordial cell in the centre of this mucilaginous mass, the nascent embryo-sac, from this period begins rapidly to displace the rest, which become by degrees completely dissolved; first of all those bounding it laterally, while the cap of cells covering its summit persists for a short time longer (fig. 21). The primary nucleus of the embryo-sac is always still distinctly perceptible in its centre ; radiating threads of granular mucilage pass out from it. Three newly- formed, closely- crowded, globular cells show themselves in the chalazal extremity, while at the micropyle end appear also the now newly-formed nuclei of the germinal vesicles (fig. 21), as three brighter, globular cavities, surrounded by dense granular mucilage (the vesicles imbedded in the protoplasm wdth more transparent fluid contents). After the primordial utricle of the embryo-sac has also dis- placed that hollow conical layer of dissolving cells which covers its micropyle end, it becomes closely applied upon the now free enveloping membrane*. This adhesion is so intimate at the micropyle end, that in all subsequent stages the enveloping membrane seems at this point to have been secreted by the primordial utricle of the embryo-sac. Further down, from the region where the cavity enclosed by the inner integument and lined by the enveloping membrane (of the nucleus) becomes widened and expanded into a cyUndrical form, the growth of the primordial utricle in breadth is frequently restricted; it then runs on as a slender cylinder in the axis of the wider. The tubular space between the enveloping membrane of the nucleus and the elongated primordial utricle of the embryo-sac is filled with finely granular mucilage coloured deep brown by iodine, probably the product of the solution of the liquefied peripherical cells of the nucleus (PL VII. fig. 23). These processes are accompanied by a very considerable elon- gation of the embryo-sac and the enveloping membrane of the nucleus, which both integuments follow by a rapid multiplica- tion of their cells in the direction of the length of the ovule. • Formed by the persistent, connected outer walls of the superficial layer of cells of the nucleus (see above). — A. H. 250 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. The shape of the ovule becomes materially altered by this, passing from an ovate form into that of a cylinder with a slightly at- tenuated summit and a rather swollen base. The lower persistent part of the nucleus is now pear-shaped. Its apex, as in Crocus and elsewhere, exhibits a funnel-shaped excavation, lined by the chalazal end of the embryo-sac, in which are confined the three " enigmatical antipodes of the germinal vesicles* ^^ (fig. 23). The germinal vesicles themselves, now fully developed, pear-shaped and large, with the inside of their walls lined by a layer of protoplasm which encloses the now lenticular nucleus, bear a similar relation to the micropyle end of the embryo- sac (figs. 22-24). The primary nucleus of the latter has by this time disappeared : if the embryo-sac has become adherent at all points to the enveloping membrane of the nucleus, secondary nuclei of a lenticular form are now frequently found lying upon the inside of the wall of the former, and these are mostly the centres of slightly developed systems of radiating threads of granular mucilage (fig. 22). With the exclusion of an exceptional case occasionally occurring in Z. minor, to be mentioned hereafter, this is the only indication of a preparation for the production of endosperm met with through the w^hole existence of the ovule of Zostera. After fertilization these nuclei disappear again, without having arrived even at a transitory cell-formation. The ovule is now ready for fertilization ; during its develop- ment the mouth of the rudimentary ovary becomes prolonged into the canal of the style ; the filiform stigmas spring out later through a more active multiplication of cells at two points of the circumference of the orifice. The styles, which turn up- wards at obtuse angles, are protruded at the period of flowering from the slits of the leaf-sheath enclosing the inflorescence. The anthers burst at the same epoch ; each half-anther opens by a longitudinal slit running up over the septum dividing the two loculi. The filiform pollen-cells arrive immediately upon the arms of the stigma projecting into the burst half-anthers. They are often found, singly or several together, spirally wound round these arms. * A. Braun, Die Erscheinung der Verjungung, &c. Leipsic, 1849 (1851), p. 297. W. H0PMEI8TER ON THE DEVELOPMENT OF ZOSTERA. 251 The end of the pollen-cell penetrates into the canal of the style opening at the summit of the two arms of the stigma. I was unable to lay free, uninjured from the stigma to the micro- pyle, the pollen-tube into which the otherwise already tubular pollen-cell is doubtless converted by continued growth of one of its extremities in the longitudinal direction. Within seven hours after the dehiscence of the anther, the pollen-tube is found in the cavity of the ovary * ; adhering closely to the outer side of the pendent ovule, it grows down to its micropyle, into which it penetrates, suddenly making a sharp curve. The internal cavity of the ovary, now greatly enlarged by the expansion of its walls, is filled with transparent but tolerably firm jelly, in which may sometimes be distinguished swollen-up cells, not un- like those of the epithelium of the inside of the human mouth. They are perhaps detached cells of the conducting tissue. A little air-bubble is commonly found in the lower extremity of the cavity of the ovary, at the end opposite to the stigma. The pollen-tube is of the same diameter as the pollen-cell (figs. 24-26). The portion outside the micropyle dies away rapidly; within this it remains perceptible for a considerable time. It does not usually penetrate farther than the summit of the embryo-sac, very rarely insinuating itself for a short distance down between this and the inner wall of the second integument. After the arrival of the end of the pollen-tube at the outer wall of the embryo- sac, formed by the enveloping membrane of the nucleus, one of the germinal vesicles increases in size and its nucleus vanishes (fig. 24). The other germinal vesicles shrivel up (fig. 25) and die away ; often even before the pollen- tube has emerged from the micropyle-canal, A newly-formed, glo- bular or ellipsoidal nucleus soon makes its appearance in the lower end of the fertilized germinal vesicle, inside the mass of protoplasm accumulating there ; immediately after this comes, above the nucleus, a septum, convex on the upper side, dividing * Anthers of a plant of Zostera marina taken from the sea-water on the 4th of June 1851, opened before my eyes on the 7th of June at half-past 6 in the morning, after a land journey of more than forty hours. By one o'clock on the same day the pollen-tubes had penetrated into the micropyles of ovules of the same inflorescence. The plants had been kept ever since their removal from their habitat moderately damp (not wet), and excluded from the air. 252 W. HOFMEISTEB ON THE DEVELOPMENT OF ZOSTERA. the germinal vesicle into a small lenticular, lower cell, and a larger, expanded, upper cell (fig. 25). The latter contains no nucleus ; a thin layer of granular mucilage coats the inside of its wall, and its cavity is filled with watery fluid. This upper cell is not further developed during the subsequent completion of the seed. The little lower cell, on the contrary, at once enters upon a very rapid multiplication. It swells up into a flattened globular form, and divides by a longitudinal wall ; both the newly- formed hemispherical cells then immediately divide into two " daughter- cells,'^ having the form of quarters of a sphere, by septa stand- ing at right angles to those just before produced. These four cells, constituting the rudiment of the embryo, are each divided by a horizontal cross septum (fig. 26). By continued halving of its cells, principally in the longitudinal direction, the flat- tened globular mass of cells (fig. 26) soon becomes spherical (figs. 27, 28) ; and finally, while constantly increasing in size, ovate and compressed laterally (fig. 29). The large spherical cell which supports it is only loosely attached to the inside of the wall of the micropyle end of the embryo-sac. I more than once saw the swollen cell, and the cellular mass sprung from it, slip from the micropyle end of the embryo-sac half- way towards the other extremity, without perceptible external cause*. By the time the rudiment of the embryo has acquired the globular form, a single apical cell becomes distinctly visible (fig. 28), in the organ which in previous stages of development undoubtedly possessed four. This transformation may be aptly explained by the supposition that one of the four cells originally forming the apex divides into one inner and two outer cells; either by a longitudinal septum forming an angle of 45° with each of the two side walls of the cell, after which a radial lon- gitudinal wall appears in the outer of the newly-formed cells ; or by a longitudinal wall parallel to one of the lateral surfaces, the formation of which is immediately followed by that of an- other at right angles to it. If we further suppose that, in either case, the other three of the previously apical cells become di- vided by radial longitudinal septa, during the formation of one ♦ See the note to fig. 26 in the explanation of the figures. W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. 253 inner and two outer in the fourth cell, the final result, under either hypothesis, would be the formation of a central (apical) cell surrounded by a chaplet of eight cells. During the transformation of the smaller segment of the im- pregnated germinal vesicle into a globular cellular body, the enveloping membrane of the nucleus, which has become a tough coat of the embryo-sac, adheres most intimately to the inside of the inner integument. The edges of contact of the cells of the latter soon make marks upon the hitherto homogeneous, smooth membrane in the form of cellulose ridges running upon it ; at first extremely delicate, scarcely perceptible, but gradually more distinct, and at last so sharp, that the wall of the embryo-sac most deceptively resembles, when seen only on the surface, a layer of tubular cells filled with transparent contents. The pe- culiar conditions of this enveloping membrane, its lengthened vital activity, its energetic growth, and its nutrition by the tissues and cells of different kinds which it encloses, deserve particular attention. From one of the broad surfaces of the laterally-compressed, ovate rudiment of the embryo springs an obtusely conical pro- jection of cellular tissue, the rudiment of the future principal axis of Zosiera, which is thus a secondary axis, a lateral sprout of the leafless axis of the first rank, of the embryo. Very soon after the appearance of the new structure, the first leaf is deve- loped, a little below its apex (figs. 30, 31, 33-35). This presents itself as a little ridge nearly surrounding the end of the stem, and very rapidly grows higher at the border turned to the cha- laza, than at other parts. The leaf grows longitudinally through constant repetition of the division of a transverse row of apical cells, by walls inclined alternately towards the upper and lower surfaces of the leaf, and by division of the cells of the second degree by cross walls. The direction of its growth is parallel to the primary axis of the embryo, diverging at right angles from its parent axis. During the development of this leaf, a very considerable alter- ation occurs in the shape of that axis of the first rank : by active multiplication of its cells, predominantly in the directions of breadth and length, the ellipsoidal cellular body becomes a flat mass truncated below and gradually attenuated toward the upper 254 W. HOFMEISTER ON THE DEVELOPMENT OP ZOSTERA. and lateral borders (PI. VIII. figs. .32, 35-37). The lateral borders at the same time curve over the anterior surface, so that they finally enclose, like a hood, the secondary leaf-bearing axis attached upon that surface. The multiplication of the cells, very active at all the edges of the cellular plate, lasts longest at the upper margin, directed towards the micropyle. The large cell, to which the embryo was previously suspended, becomes pushed very much to one side by this (fig. 32). Soon compressed by the further increase of size of the primary axis, it is in a short time lost sight of altogether. The leaf-bearing axis, the lower, naked portion of which has meantime been extending upward with a considerable cur- vature, unfolds, soon after the protrusion of the first leaf, the se- cond, opposite to that ; the third becomes opposed to this higher up on the stem, and finally the fourth to the third (fig. 39). The greatly prolonged extremity of the stem grows by continued repetition of division of the single apical cell, by means of septa inclined alternately to the two surfaces of the leaf, division of the cells of the second degree by radial longitudinal walls, and so on ; corresponding to the rule of cell-multiplication in the rudiment of the fruit of Mosses, of the Marchantiea, the stem of Mosses, the stem of the Polypodiaceae, the axes of the Equisetaceae and Pilulariece, and the young embryos of the Coniferae. But, as in the Coniferae, the rule of cell-formation changes subsequently, after the germination of the seed. The one apical cell of the now obtuse, flattened terminal bud (fig. 40) divides by walls inclined successively to the four points of the compass, in agreement with the rule of cell-multiplication in the further developed rudiment of the fruit of Ant ho ceros*. Buds appear in the axils of the older leaves even before the maturity of the embryo. Subsequently also, in the fully-deve- loped, annual sterile plant, the formation of an axillary bud follows that of the leaf almost immediately (PI. VIII. fig. 40). In the foregoing I have called the cellular body, formed by the increase of the lower segment of the impregnated germinal vesicle, which is finally transformed into a hood-shaped layer of cellular tissue, the axis of the first rank, of the embryo. Older botanists regarded it as the cotyledonary leaf, — a view which • See page 6 of my Vergleichende Untersuchungen^ &c. Leipsic, 1851. W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. 255 Gronland also supports, although A. de Jussieu*, so long ago as 1838, demonstrated that this is untenable, by comparison with other Monocotyledonous embryos, and explained the organ as a transformation of the plumule {tigelle) of the embryo, with- out however recognizing the leaf-bearing stem as a secondary axis. An unprejudiced examination of the course of develop- ment completely refutes the older view. But even if it were assumed, for the sake of the pretended analogies, that the direc- tion of the main axis of the embryo underwent a division not to be detected by actual observation ; that the cell of the lateral surface of the flattened -ovate cellular body, which gives origin by its cell -multiplication to the leaf-bearing axis, was the twelfth cell of the first degree of the embryo, and that the obtuse end of that cellular mass underneath this cell was the rudiment of the cotyledon, — the cotyledon and the next succeeding leaf would be made to stand in one perpendicular line, i. e, the leaf directly over the cotyledon ; a supposition incompatible with the phaeno- mena exhibited in the after-life of Zostera. There is no example among all the Dicotyledons of the principal leaf-bearing stem of a plant being an axis of the second rank, the lateral sprout of a leafless primary axis. The develop- ment of the embryo of TropcRolum, which at first sight appears similar to that of Zostera, differs from it most essentially in the circumstance that the portion of the pro-embryo, from the end- cell of which the embryo originates, is in Tropaeolum the primary axis, only crowded to one side by the more vigorous development of the peculiar lateral sprout of the pro-embryo f. Among the Monocotyledons there is one plant, Ruppia rostellata, resembling Zostera in many characters of its sub- sequent life, which may be compared with the latter, in reference to the characters of the embryo, without straining any point. The ovule oi Ruppia agrees in structure with that oi Potamogeton, Like that, it completely resembles the ovule of Zostera in its early development, in attachment, direction and shape (fig. 41 a,b). But that large cell in the interior which becomes the embryo-sac, only displaces a moderate portion of the nucleus before impreg- nation ; the outer layers of cells of the latter persist (fig. 42). As in Pot amogeton, the previously concentrically-shaped ovule begins * j4nn, des Sc. Nat. 2 s^r., Botanique, torn. xi. t See my essay Die Ensiehung des Embryo. 256 W. HOPMKISTER ON THE DEVELOPMENT OF ZOSTER A to assume a symmetrical form shortly before impregnation. The micropyle becomes pushed downwards by active multiplication of the cells on its outer side, turned away from the contiguous angles of the four ovaries (fig. 42). A multiplication of the cells of the opposite side of the nucleus commencing at the epoch of impregnation, subsequently pushes the micropyle upwards again with a great curvature of the entire ovule (fig. 43). The integu- ments follow the increasing size and changing shape of the ovule by multiplication of their cells in the directions of length and breadth. After impregnation the process of multiplication and expansion of the cells becomes very unequal, especially on the inner coat of the ovule : much more active in the lower than in the upper half, it pushes the endostome still farther upward (fig. 44), and removes it from the endostome, which remains nearer its original place, and consequently becomes diverted to the side. The impregnated germinal vesicle of Ruppia divides, as in Zostera and Potamogeton, into a larger, upper, persistent cell and a smaller, inferior cell which undergoes rapid multiplication. During the gradual removal of the papilla (apex) of the nucleus and of the endostome from the place of the germinal vesicle, the pollen-tube, which remains outside the membrane of the embryo- sac, extends itself, inside the place where it penetrates, through the apex of the nucleus, by means of considerable longitudinal growth of its tough membrane (fig. 43 b). The lower segment of the impregnated germinal vesicle very soon becomes changed, by a series of longitudinal and transverse divisions (figs. 43 b, 44 b), into a globular and subsequently an ovate cellular body (fig. 45 b), from the side of which, below the apex, shoots forth the first leaf-bearing axis. It unfolds its first leaf close above its point of origin. After the imbedding of the base of the leaf-bearing axis in the thick, fleshy, primary axis, the first adventitious root grows out, in germination, opposite the lamina of that first leaf (the cotyledon), forming a right angle with the line of direction of the leaf. The margins of that lateral face of the primary leafless axis which bears the leafy secondary axis, rise up into a kind of collar, enclosing the base of the latter as a short, widely opened sheath (fig. 46). Irmisch*, in a clear explanation of the subsequent course of the • " On the Inflorescences of the German Poiamece,'' Flora 1851, p, 81 et seq. W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. 257 germination of Ruppia, has already drawn particular notice to the fact, that the germination not only of Potamogeton, but of all the grasses in which the first leaf stands before the scutellum, agrees in the main points with that of Zostera. The history of development of these embryos, far from weakening the founda- tion for such an interpretation of the individual parts, rather speaks more distinctly in favour of the so-called scutellum of the embryos ofZea and Sorghum being analogous to the first, leafless axis of Zostera^. The course of development of the embryo of other Mono- cotyledons, so far as known, gives still further probability to the idea that the "cotyledon^^ is an axis with limited growth, from the side of which breaks forth a sprout, the future leaf-bearing main axis of the plant. I refer to the history of development of the embryo of Cannaf given by Jussieu, with which completely agree, besides those of the plants mentioned by Jussieu himself, those of Hyacinthus comosus, Funkia ccerulea, Fritillaria impe- rialis and other Liliaceae, that of Iris and others. In all cases the first cell of the embryo (the lowest cell of the pro-embryo forming the suspensor) becomes converted by a series of divisions into a globular, ovate or spindle-shaped cellular body, the in- crease of thickness of which is arrested at a point below the summit. Thus is formed a lateral, more or less deep slit, from the bottom of which is developed the plumule, in one case {Fri- tillaria) not till after germination. No fact is less available as testimony against the axial nature of the so-called cotyledon of the Monocotyledons, than that of its base enclosing the plumule as a sheath. If the secondaiy axis originates at a point on the primary in which the growth in thickness of this latter is arrested some time before the appearance of the lateral sprout, the en- compassing of the new axis by the thickened borders of the * See figures ] 2-1 4, 32-34 of PI. xi. of my Essay Die Enstehung des Embryo der Phanerogamen. At the time when I published those researches, not yet enlightened by the study of the formation of the embryo of the Naiadeae, I thought myself obliged to adhere to the view advocated by Bernhardi, Schleiden and Jussieu, although it did no little violence to the phaenomena observed. The papilla of cellular tissue becoming the plumule, appears from the first distinctly lateral and below the apex of the nascent ** cotyledon," in these grasses, as also in the Avenacece and TriticeeB. t Loc. cit. p. 348. SCIEN. MEM.— A^a<. Hist. Vol. I. Part III. l7 258 W. HOFMEISTER OS THE DEVELOPMENT OF ZOSTERA. lateral surface of the older sprout naturally follows ; as occurs most clearly in the earliest rudiments of fronds, situated in de- pressions of the creeping main axis, of Polypodium aureum and other Ferns. But even secondary axes which become nakedly visible on the originally convex outer surface of the young main axis, may likewise become sheathed subsequently by the primary sprout, through peculiar growth of the latter ; Isolepis affords a striking proof of this. It need scarcely be explained that the discussion here promul- gated is not and could not be anything more than an invitation to repeated investigations of the different stages of development of the Monocotyledonous embryo. A complete decision in favour of one or other of the two views can only be obtained by tracing back — a difficult task indeed — the development of the seedling plant to the mode of increase of each individual cell. No question in the whole field of the Morphology of Plants has been treated by so many persons and in such varied ways, as the import of the parts of the embryo of the Monocotyle- dons ; to none better applies an often misplaced quotation from Goethe ; we may assert that " all possible combinations are ex- hausted '^ here. Thus the idea brought forward in the foregoing pages is only the repetition of an earlier one essentially similar. If it be thought better grounded than that more current among modern botanists, we find, in recompense for the wide separation it makes between the Dicotyledons and Monocotyledons, the most striking agreement of the development of the embryo of the latter, with the course of the formation of the germling of the Vascular Cryptogamia. As I have abundantly demon- strated*, the main stem of the plant of all Vascular Cryptogams is a secondary axis, arising laterally on the primary leafless axis. A second and not unimportant point of comparison then at once suggests itself. The Vascular Cryptogams may be divided into two primary groups according as the germinating plant produces the second frond below the first (below that sur- face turned away from the mouth of the archegonium), or above it. In the former case the first adventitious root appears simul- taneously with the first frond, beside its base ; in the latter case, ♦ See Botanische Zeitung, 1849, p. 797; " Vergleichende UntersuchungeUy &c." Leipsic, 1851, pp. 85, 106, and elsewhere. W. HOPMEISTER ON THE DEVELOPMENT OP 20STERA. 259 opposite to it. The Ferns and Rhizocarpeae belong to the first of these sections^ Selaginella and Iso'etes to the second. The peculiar arrangement of the roots oi Iso'etes does not follow so much from the condition just indicated, as from the phaenomena of the growth of the mantle of cambium, annually renewing the cortical tissue, surrounding the wood-structure ; combined with the suppression of the internodes. In Zostera, Ruppia, and in the Grasses, the mode of appearance of the first adventitious root is the same as in Iso'etes, It springs out opposite the first leaf, not behind it. The great majority of the Monocotyledons seem to behave in the same way. In Lemna, on the contrary, the position of the first adventitious root corresponds to that of the Ferns. To carry through the comparison of the two classes in this respect will require still more comprehensive researches. EXPLANATION OF PLATES VI, VII and VIIl. Fig. 1. Longitudinal section of the terminal bud of a fertile plant of Zostera minor, magnified 30 diam. : — a the third, b the second, c the first (youngest) spadix ; a}, b^, c\ the sheathing basal leaves (niederbldtter) of the sprouts terminating as spadices; a\ b^, c^, the stem-leaves Qaub-bldtter) of the same ; d, the new sprout arising ii? the axil of the sheathing leaf of the youngest sprout. Fig. 2. A similar preparation of Zostera minor. The section, not carried quite accurately through the slit of the sheathing leaf of the sprout of the penultimate rank, has hit the overspreading part of this sheath. Fig. 3. Apex of a spadix of Z. marina commencing a longitudinal develop- ment, seen from above. Magnified 40 diam. : — a, anthers ; 6, ovaries. Fig. 4. Young anther (a) and rudiment of an ovary (6) oi Zostera minor, seen from above. Magnified 120 diam. Fig. 5. Transverse section of a very young half-anther of Z. minor. Only the contents of the primary parent-cell of the pollen represented. Magn. 350 diam. Fig. 6. Similar preparation from a somewhat more developed anther. Magn. 350 diam. "rig. 7. Some of the primary parent-cells of the pollen, from the longitudinal section of a very young half-anther of Z. marina, Magn. 400 diam. Fig. 8. Lower portion of a longitudinal section of a half-anther of Z. marina^ parallel to the surface of the spadix ; a, cells of the septum between the two loculi. Some of the elongated parent-cells of the pollen cut across. Magn. 400 diam. Fig. 9. Portion of the longitudinal section of a half-anther of Z. marina imme- diatelv after the formation of the pollen-cells. Magn. 400 diam. 17* 260 W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. Fig. 10, 12. Transverse sections of loculi of somewhat more developed anthers of Z. minor. Magn. 200 diam. Fig. 11, A collection of young pollen-cells of ^. marina, Magn. 400 diam. Fig. 13. A few somewhat further developed pollen-cells of the same species, isolated in concentrated solution of carbonate of ammonia. Magn. 400 diam. Fig. 14. Longitudinal section of the lower part of an anther-loculus of Z. minor. Magn. 300 diam. Fig. 15, Piece of a mature pollen-cell of Z. marina. The ascending current of protoplasm is shown darker, the descending lighter. Magn. 400 diam. Fig, 15 b. Pieces of mature pollen-cells exhibiting protrusions of all kinds (see page 245). Magn. 200 diam. Fig. 16. Very young ovaries of Z. marina, seen from above; a, rudiment of the ovule. Magn. 300 diam. Fig. 17, 18. Subsequent stages of development of the ovary, in longitudinal sections. Magn. 100 diam. Fig. 19. Ovules and the surrounding parts of the walls of the ovary, in longi- tudinal section. Magn. 320 diam. The pendent ovule is here re- presented reversed and upright, as in all the succeeding figures. Fig. 20, 21. Longitudinal sections of the nuclei of more developed ovules. Magn. 300 diam. Fig. 22. Longitudinal section of an ovule ready for impregnation. The for- mation of the cells compressed into the chalazal end of the embryo- sac has not taken place here. Magn. 320 diam. Fig. 23. Longitudinal section of an ovule of the same species ripe for impreg- nation. Magn. 180 diam. Fig. 24. Micropyle and upper end of the embryo-sac of an ovule in longitudinal section, at the moment when the pollen-tube has penetrated to the membrane of the sac. Magn. 300 diam. Fig. 25. A similar preparation, a little later, with the impregnated germinal vesicle divided into an upper, large suspensor cell, and a lower, small embryonal cell. Mag. 300 diam. Fig. 26. A similar preparation with the embryonal globule, composed of 16 cells, represented in longitudinal section. During the observation the suspensor and the embryonal globule slipped from the micropyle end of the embryo-sac as far as *, without perceptible external cause. Magn. 300 diam. Fig. 27. Lateral view of a further developed embryonal globule with the suspensor. Magn, 400 diam. Fig. 28. A similar embryonal globule, in longitudinal section, at right angles to the projection of the preceding figure. Magn. 400 diam. ♦ Fig. 29. Front view of an embryo with suspensor of Z. minor, shortly before the origin of the leaf-bearing axis. Magn. 60 diam. Fig. 30. Lateral view, 60 diam. ; fig. 31. front view, of farther developed em- bryos of the same species. Magn. 40 diam. Fig. 32. Lateral view of the embryo of a half-ripe seed of the same species. The outline of the secondary axis is indicated by delicate lines, show- ing through the lateral wing of the primary axis. Magn. 40 diam. W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. 261 Fig. 33-35. Embryos of Z. marina. — Fig. 33, side view of a young embryo » fig. 34, front view of one somewhat older ; fig. 35, seen obliquely from behind. Magn. 40 diam. Fig. 36. Longitudinal section of a half-developed embryo. Magn. 30 diam. Fig. 37. Longitudinal section of an almost ripe embryo. Magn. 30 diam. Fig. 38. Ripe embryo, in front, back and lateral views. Magn. 4 diam. Fig. 39. Longitudinal section of the bud of an almost ripe embryo. Magn. 60 diam. Fig. 39 h. The end of the stem of the same. Magn. 300 diam. Fig. 40. Longitudinal section of the terminal bud of a sterile plant oiZ. marina. Fig. 41-46. Ruppia rostellata. Fig. 41 . Longitudinal section of a young ovary. Magn. 40 diam. Fig. 416, The ovule of the same. Magn. 200 diam. Fig. 42. Longitudinal section of an ovule ready for fertilization. Magn. 60 diam. Fig. 42 h. Embryo-sac of the same. Magn. 300 diam. Fig. 43. Longitudinal section of an ovule immediately after impregnation. Magn. 40 diam. Fig. 43 h. Endostome, papilla of the nucleus and micropyle-end of the embryo- sac of the same. Magn. 300 diam. Fig. 44. Longitudinal section of an ovary, some eight days after fertilization, Magn. 30 diam. Fig. 44 6. The rudimentary embryo of the same. Magn. 200 diam. Fig. 45. Longitudinal section of a farther developed ovary. Magn. 30 times. Fig. 45 b. The embryo from it. Magn. 200 diam. Fig. 46. Longitudinal section of an almost ripe embryo. Magn. 40 diam. [A. H.] 262 M. WICHURA ON THE WINDING OF LEAVES. Article IX. On the Winding of Leaves, By M. Wichura. [From the * Flora,' January and February 1852.] 1 HE leaves of plants possess in the flexibility of their tissue full capability of twisting, and it has long been known that this property must often serve to assist the efforts of plants to ap- proach towards the light. Twisting movements of this kind are readily recognizable from the circumstance that all the characters of the movement may be brought into agreement with their pur- pose, to direct the upper face of the leaf towards the light. The twisting begins directly anything turns the upper face of the leaf away from the light, and continues until that face is again directed to the full rays of the incident light. It turns indif- ferently either to the right or left, according to which path will lead quickest to its object. Its extreme degree never exceeds a half-revolution, since this suffices to turn the leaf completely over, and therefore to return the upper side again towards the light, when it has been wholly turned away from it. But there are plants the winding leaves of which present ex- actly the opposite character in all these things, and, in particu- lar, exhibit the same regularity in their lateral direction as we detect in winding stems. The winding movements of the second kind cannot be re- garded as an effect of the irritability of vegetable tissue excited by light. They are immediate expressions of the vital force ac- tive in the interior of the plant, and nearly aUied to the winding of tlie stem and tendrils. While, however, the latter have long occupied the attention of botanists, the windings of leaves have remained almost unknown, and all that we possess are a few scattered and mostly very meagre notices thereon. I myself gained my first knowledge of them through perceiving the heliacal winding of the leaves and strict regularity displayed in their direction, in some Oat and Barley plants germinating in- doors. Further researches, to which I was thereby attracted. M. WICHURA ON THE WINDING OF LEAVES. 263 made partly in the field, partly in the Botanical Gardens of Breslau and Berlin, and lastly, partly in the General Herbarium at Berlin, have revealed to me so great a number of similar phaenomena, that I venture to found upon them an attempt at a general description of the phaenomenon. I. Form and outward condition of Winding Leaves, §1. Although examples of winding leaves may be pointed out in all metamorphoses of the leaf, in the most diverse families of the vegetable kingdom, and in all the local floras of the globe, much agreement in structure is exhibited by them in spite of this wide diffusion. They are all of longish, mostly linear- lanceolate shape, have smooth entire margins, and in regard to the distribution of their vascular bundles, either belong to the parallel-nerved system, as developed most clearly in the stem- leaves of Monocotyledons, or they are totally devoid of vessels, as in the leaves of the Mosses and Liver-worts. The angular- nerved leaves of the Dicotyledons, as for instance of our fruit- trees, of the Poplar, Lime, &c., appear never to admit this winding movement. In the Dicotyledons only those leaves wind which imitate, either perfectly or at least approximatively, the parallel-nerved system of the Monocotyledons. §2. If we attempt to bring the properties of winding leaves, men- tioned in the preceding section, under a common point of view, we must say that the longitudinal growth is developed at the cost of the transverse growth in winding leaves. For, if the angular-nerved distribution of the veins in which strong branches turn off toward both sides from the midst of the leaf, and, sud- denly leaving the longitudinal direction, pass into the transverse, is to be considered, no less than the lateral subdivision of the leaf into teeth, pinnae, &c., as the expression of an active growth in the direction of the breadth, — it follows that the essential feature of leaves which neither have an angular venation, nor are toothed or pinnate, but are parallel-veined, smooth- edged and very narrow in proportion to their length, must be a 264 M. WICHURA ON THE WINDING OP LEAVES. deficiency of transverse and an excess of longitudinal growth. Winding leaves agree perfectly in this respect with winding stems and tendrils, the thread- or string-like shape of which at once reveals the predominance of longitudinal growth. §3. Another peculiarity of winding leaves might be sought in the distribution of their stomates« In the ordinary, not winding leaf, they are chiefly placed on the lower face, next the ground, in order, as is supposed, to come, in this way, into immediate and close contact with the gases and vapours ascending from beneath. A differently contrived distribution of the stomates should consequently be provided for winding leaves, which by their curvature often change the relative position of their two faces toward the vicinity. This conjecture finds a very striking confirmation in the stem-leaves of the Alstrcemerm, which turn quite over by a half-revolution, for, according to the beautiful observations of Lindley, these bear the stomates, contrary to the ordinary rule, on the upper face, turned towards the ground only by the twisting. The winding leaves of the Grasses also, as I have had an opportunity of observing, present anomalies, in so far that they are furnished with stomates almost equally on both sides of the leaf. Unfortunately I was unable to extend these somewhat ^ time-devouring ' researches to other families. I can therefore only offer a conjecture that winding leaves are distinguished from those not winding by a more or less aberrant distribution of the stomates. §4. In other respects, winding leaves present httle that is striking in their external appearance. Not twining round a support like winding stems, they complete their curvatures in the free air, and from their slender, linear form, one is at first sight inclined to ascribe their curvature to accidental effects of the wind and of desiccation. The detection of the regularity prevailing in their direction unfolds the error of the supposition, and opens to our view a rich and everywhere ready field of observation, where we previously passed by in indifference. M. WICHURA ON THE WINDING OF LEAVES. 265 II. Laws of Distribution, §5. The relation of the winding to the outer form of the leaf, mentioned in section 1, gives the key to the laws of distribution of the phaenomenon. All alterations in the form of the leaf connected with the proportion of the longitudinal to the trans- verse growth, are consequently of importance in regard to the occurrence of the winding movement. §6. Even the single leaf displays changes of shape of this kind, in its triple division into sheath, stalk, and blade. Winding occurring in one of these parts is not therefore necessarily imparted to the others, but is mostly limited within this. There are many leaves with wound petioles and straight laminae, while, on the contrary, in the stamens of Erythrcsa and other plants, the lamina metamorphosed into an anther winds, while the petiole, i. e. the filament, remains straight. Even within the lamina itself, modifications of the curving movement occur, with the expansions and contractions advancing from below upwards. In such cases only the narrow parts wind, the broader not. The broad ovate leaves of Paris quadrifolia, L., may be men- tioned as an example, being straight as far as the attenuated points which are wound slightly to the left. Leaves which wind uniformly throughout their whole extent, exhibit the corresponding perfect equality of transverse growth, from bottom to top, as for instance the leaves of Typha latifolia, L., T, an- gusti/olia, L., and of many other Monocotyledons. §7. The distribution of the winding movement in the different leaves of the same axis depends on the same principles. Here it is the metamorphoses of the leaf, known under the names of cotyledons, stem-leaves, sepals, &c., which limit the winding movement in virtue of their frequently totally different shapes. It. very frequently happens that the winding of the leaves is 266 M. WICHURA ON THE WINDING OF LEAVES. confined to ^ single metamorphosis, or that when the leaves of several metamorphoses wind, the intervening metamorphoses which do not wind, are separated from each other. In many species of Dianthus, for instance, the stem-leaves, the petals and the styles wind. Between these lie the metamorphoses of sepals and stamens, which do not betray a trace of the wind- ing movement. Winding and straight leaves are found united, even within the single metamorphosis, when the form of the leaves belonging to this metamorphosis is essentially altered from below upward, or according to the position which they occupy on the axis. Thus, for example, in Papyrus antiquorum, Willd., the lower, rather broad stem-leaves do not wind, but the upper narrow linear leaves preceding the inflorescence as bracts do. In many irregular flowers likewise, only single leaves, mostly outwardly conspicuous for prominent longitudinal growth, exhibit winding, for example the iabellum of Himantoglossum hircinum, Rich. Here belong, moreover, the so-called oblique flowers, e,g, ofHys- sopus lophanthus, L., Pedicularis palustris, L., &c., the oblique shapes of which arise from particular organs only of the flowers curving, and thus destroying the symmetry of the whole. Finally, the twisted awns of the Grasses must also be mentioned here, since in all cases they occur only in one of the paleae forming the floral envelopes. §8. The distinction in the shape of leaves belonging to relatively like metamorphoses of different individuals, is fixed by the systematic distinctions in the vegetable kingdom. The more widely two species stand apart, systematically, the larger is the sphere of possible differences in the development of the leaf, the smaller therefore becomes the probability that they will behave similarly in reference to the movement of winding. The divisions into Cellular and Vascular Plants, Monocotyledons and Dicoty- ledons, founded on the most general distinctions in the vegetable kingdom, are consequently of no particular value for our purpose, since each of these divisions includes formations of leaves of the most varied kinds. It can only be said that the stem-leaves of M. WICHURA ON THE WINDING OF LEAVES. 267 the Monocotyledons occur wound much more frequently than the stem-leaves of the Dicotyledons, on account of their structure being peculiarly suited for the movement of winding. §9. It is only when we come to the families of plants that it be- comes possible to sketch out general characteristics of the natural divisions, in reference to the movement of winding. There are families in which I have not hitherto detected a single example of a winding movement, as for instance the Umbelliferae and Cruciferae ; families in which the winding is confined to particular genera or species, as in the Compositae and Papilionaceae ; and, finally, families in which the curvature of the leaves is typical, as in the Ebenaceae, Apocynaceae and Asclepiadaceae. The same holds good of the genera, with the distinction that a complete agreement of the species included in them is much more frequent than in the families. In general, the behaviour of the individuals belonging to the same species is identical. Deviations are rare, and seem only to arise when the shape of the leaves has undergone a material change through the influence of the station. Thus I found in a hot- house a specimen of Hyacinthus orient alis, raised in rich earth, in which the stem-leaves had attained the unusual length of 2 1 feet, and were twisted, while under ordinary circumstances the leaves of this plant do not wind. § 10. Finally, peculiarities of the soil and climate also have to be mentioned among the causes of the diffusion of the phaenomenon, since these sometimes give a particular character to the habit of the plant, and especially to the prevailing form of the stem- leaves. Thus, tropical America is remarkable for an abundance of plants with broad leaves with reticular venation, in which the here incompatible movement of winding withdraws from notice, and, considered in relation to other regional floras, becomes rare. On the other hand, the floras of New Holland and the Cape of Good Hope contain principally plants with narrow, parallel- nerved stem-leaves, and it is to be regarded as a consequence of this, that Dicotyledonous plants with winding stem-leaves occur 268 M. WICHURA ON THE WINDING OF LEAVES. comparatively much more frequently in these two regions than in the other parts of the globe. III. Single and associated Winding Leaves, §11. The phaenomenon of winding is not confined to the single leaves of plants, but appears also in structures arising from the confluence of several leaves. Longish shape and parallel course of the vascular bundles are also conditions here. For example, the tubes of the corollas of several Stylideae, which may be called rather long in proportion to their diameter, those of Trifolium resupinatum, L., and certain species of Peristrophe and Hypo'estes of the family of Acanthaceae. So, moreover, the elongated buds of the Convolvulaceae and of Thevetia neriifolia, Juss. (Apocynaceae), originally composed of five separate and very narrow segments, are folded toward one side like a closed um- brella and curved spirally toward the other, &c. If the view, now continually gaining ground, that the stem is not an inde- pendent organ, but originates from the confluence of the sheathing portions of the leaves fitted into each other, could be regarded as completely demonstrated, the winding stem also would have to be included here. § 12. After these examples of winding of several confluent leaves, we may come to those which certainly wind separately, but are situated so close together that they come in contact in the move- ment of winding, and in this way enter into certain combina- tions, which, from the regularity of the movement on which they depend, assume a regular shape. § 13. Here belong : the entwined cotyledons of the Gyrocarpeas ; the awns of the panicles of Streblochate nutans, Hochst., and Andro- pogon Allionii, D.C., which are twisted together into a cord-like body ; the floral envelopes of many species of Iris and all the species of Aristea (Iridaceae) known to me, which wind round I M. WICHURA ON THE WINDING OF LEAVES. 269 spirally as the flower withers ; and lastly, what is called the " contorted aestivation/' The latter deserves a separate discus- sion, because its connection with the winding movement is not at first sight clear. § 14. The *^ aestivatio contorta '' is a structure which makes its ap- pearance at a comparatively late stage of the growth of the bud. When examined in their earliest conditions the rolled buds ex- hibit leaflets standing singly, so narrow that their margins do not touch. They become broader at a later stage, and the posi- tion of the leaves, peculiar to the twisted aestivation, arises from the mutual over- and inter-growth of their margins. Fig. 1, PI. IX. represents two horizontal sections of oppositely rolled buds. By an examination of these it is readily ascertained that the surfaces of the leaves are obliquely inclined to the centre of the flower, and in the same direction in all the leaves of the same flower. Either the leaves have been attached obliquely on the receptacle from the first ; and then the twisted aestivation is not the effect of a subsequent twisting of the separate leaflets, and does not belong here ; or the leaflets are not originally attached obliquely ; and then they can only be brought into the oblique position subsequently evident, by a slight twisting. Which of these two alternatives is the correct one, I could not ascertain by direct observations, since the oblique position of the leaflets only becomes perceptible by its effect, the growth of the borders over and between each other. It is so shght, and the curvature, if such occurs, is so lost in the earliest stages of the bud, that it sinks to a microscopical magnitude, from the transparency and delicacy of all the parts no longer recognizable by the sight. On the other hand, conclusions may be drawn from other perceptible facts, which give very reliable testimony that the twisted aestivation owes its origin to a winding of the individual leaflets. §15. In the first place, an approximative testimony is furnished by the occurrence of the twisted aestivation in gamopetalous corol- las. The history of development of these, teaches us that they 270 M. WICHURA ON THE WINDING OF LEAVES. are composed, in the earliest periods of their formation, of sepa- rate leaves which subsequently become blended together. It will be admitted that if the " aestivatio contorta " depended on an oblique attachment of the leaflets upon the receptacle, this must exist from the moment of the origin of the leaf, and conse- quently in those earliest periods of growth when the individual parts of the gamopetalous corolla were not yet confluent. If then, in the " aestivatio contorta,^^ the oblique position of the leaflets subsequently caused their borders, with the increasing breadth, not to meet together, but to pass over and under one another, this obliquity must, at the time when the confluence of the separate foliaceous elements should commence, hinder their coming into contact and consequently the confluence itself. Hence oblique attachment of the foliaceous elements and gamo- petalous growth seem to exclude each other, and if gamopetalous corollas nevertheless do occur with rolled segments, it follows that the rolled position in the bud cannot be explained by an originally oblique attachment of the leaves. Then there only remains the other alternative, that the foliaceous elements are originally evenly attached, and only after confluence has oc- curred, undergo a twisting of the free points, by which they are brought into the position suited to the formation of the contorted aestivation. §16. I regard this proof as not wholly conclusive, only because it is based upon suppositions w^hich relate to the still somewhat problematical processes of the confluence of the floral organs in the earliest stages of development of the flower. But the twist- ing of the leaflets on which the contorted aestivation depends, was completely demonstrated by an observation which I had an opportunity of making on a plant of Heticteres cultivated in the Berlin Botanical Garden. The long and narrow petals of this plant possess two opposite teeth on the margins, near the low^er part. In the bud, the teeth of the adjoining petals cover one another in the manner of the " aestivatio contorta,^^ as do like- wise the upper parts of the little petals, but in an opposite di- rection. While the right tooth of one leaf covers the left tooth of the next leaf, the left margin covers, above, the right margin M. WICHURA ON THE WINDING OF LEAVES, 271 of the next leaf. It is clear that this contrast cannot be explained by an oblique attachment of the young petals, since this can only act in one direction : we are absolutely compelled to the assumption that a curvature must have taken place here, which, as we shall see, in many cases alters its direction within the same leaf. §17- The rolled buds also betray their close relationship with the movement of winding in other respects. A comparatively very large proportion of the plants in which the floral envelopes or segments of the flower twist after the flower has opened, have rolled buds ; for example : Pay a C(srulea, Miers (Bromeliaceae), Christy a speciosa, Ward and Harv., Strophanthus diver gens, Graham (Apocynaceae), Pergularia edulis, E. M., P. accidens, Blume, Diplolepis Menziesii, R. & S., Oxypetalum riparium, H. B. K. (Asclepiadaceae), Cyclamen europceum, L., Lysimachia punctata. Wall. We likewise frequently meet with rolled buds which are at the same time twisted in one or other direction, as in ^chmea latifolia, Kth. (Bromeliaceae), Btrophanthus dicho- tomus, D.C., Echites longiflora, Desf. (Apocynaceae), Pergularia edulis, E. M., Microloma sagittata, R. Br. (Asclepiadaceae), Gil- tenia trifoliata, Moench. (Rosaceae), &c. From all these reasons, therefore, I have no hesitation in regarding the " aestivatio con- torta " as an effect of the movement of winding. IV. Heliacal shape of Winding Leaves. §18. The characteristic mark of all winding leaves is their heliacal shape, with manifold modifications, however, depending on the variability of the distance of the heliacal line from its axis, the angle of inclination, and the length of the helix. § 19. The distance between the heliacal line of the leaf and the axis of the helix may be either great or small. If it is reduced to the smallest possible amount, it coincides with the median line of the leaf itself. The fruit of Ailanthus glandulosus, L., repre- 272 M. WICHURA ON THE WINDING OF LEAVES. sented in figure 3, may serve as an example. If we draw a straight line from the extreme point of this to its stalk, it will cut through the substance of the leaf in its whole course. The parts of the fruit lying in this direction consequently form a straight line and consequently the fixed axis, around which the halves on the right and left appear wound in the shape of a screw. But if, on the contrary, the axis of the helix which the leaf describes lies outside the latter, all parts of the leaf share in the heliacal winding, and the leaf then resembles a band wound in a heliacal direction round an invisible cylinder or cone, where one side of the band is always turned towards the said cylinder or cone. The axis of the heliacal winding in this case coincides with that of the imaginary cylinder or cone. Leaves of this kind occur much more frequently than the others. Some examples are represented in figures 8 & 9. § 20. The angle of inclination of the heliacal winding, i, e. the angle formed when a line is drawn through the helix parallel to its axis, is in many leaves so small as to be imperceptible, for example in the "aestivatio contorta;" in others it rises, judging from simple inspection, which indeed leaves a wide field for error, to 30°, 40°, or even 45°. It therefore falls short of 90°, the highest degree mathematically possible, which would flatten the helix into a plane. §21. Lastly, the length of the helix is dependent on the length of the leaf, or when only a part of this winds, on the length of this portion. §22. The sum of all these elements of the heHx gives the number of revolutions. The magnitude of the angle of inclination and the length of the helix stand in direct, and its distance from the axis in inverse, proportion thereto. The greater the angle of inclination and the longer the helix, the higher the number ; the greater the distance of the helix from the axis, the smaller the number of re- volutions. Under otherwise like circumstances a broad leaf can never complete so many revolutions as a narrow one, because the M. WICHURA ON THE WINDING OF LEAVES. 273 helices which the margins of the leaf describe, at the same time as the other parts, must be further removed from the axis in the broad than in the narrow leaf. §23. Consequently the greatest number of revolutions, from twenty to twenty-five or even more, are found in the narrowest leaf- structures, e.g. in the awns of the Grasses, the leaves of many species of Gethyllis, &c., while elsewhere the most common number of revolutions is from J to 2. Although accurate pre- determination is mostly impossible here, since the length and breadth of the leaf, which influence the number of revolutions, usually vary in the same plant, still at least approximative deter- minations may be given for each species, and it is to be desired that descriptive botany should include this in the subjects of its observation. In flower-stalks which, as regards size and shape, appear always more constant than the stem-leaves, the number of revolutions of the heliacal windings admits of a tolerably exact definition. V. Rapidity of the Movement and Epoch of its Commencement, §24. The winding movement takes place either rapidly or slowly. In the appendages of the carpels which separate at the epoch of maturity in the Geraniaceae, it is so rapid as to be visible to the eye. In the other cases with which I am acquainted, we can only conclude that a movement has taken place by recognizing after long observation that a change of place has occurred. Days or even weeks may elapse before the leaf has completed even one revolution. §25. In regard to the epoch of commencement, the movement is connected with the stages of growth of the plant. It here fol- lows definite laws, which however differ very much in different plants. §26. The " aestivatio contorta " affords the earliest traces of the commencement of a movement of curvature. The extremely SCIEN. MEM.— iVa/. Hist. Vol. I. Part IV. 18 274 M. WICHURA ON THE WINDING OF LEAVES. slight winding occurs here, it must be assumed, at an epoch when the leaves are still quite imperfect, in fact scarcely visible to the naked eye. But in all other cases the winding does not commence until the leaf has attained a certain magnitude and maturity. The upper parts, as the older, wind first, the lower then gradually follow. The winding of the older part likewise precedes that of the younger in winding stems. But as in stems the lower parts are older than the upper, while in the leaf the upper parts are the older, the movement advances from below upwards in winding stems, and in the leaf on the contrary from above downwards. §27. In addition to these cases of a gradual origin and progress of the movement, there are some in which the movement begins suddenly and exhibits a more rapid course throughout its whole duration. We then constantly observe that the stage of growth which makes the commencement of the movement, is otherwise of importance in the development of the plant. Thus the rather rapid winding of the legumes of Medicago commences after fertilization has taken place, that of the anthers of many plants after dehiscence, of the petals of Cyclamen europceum, L., and a number of other plants, after the flower has opened, of the flowers of most species of the genus Aristea as soon as they begin to wither, of the awns of Avena fatua and probably all other Grasses with twisted awns, at the beginning of the maturation of the seed, and of the appendages of the carpels of the Gera- niaceae after their separation from the carpophore. § 28. The movement of winding is in all these cases a rather trans- itory phaenomenon compared with the duration of the existence of the plant, but it leaves permanent effects upon the shape of the leaf. When a twisted leaf is unrolled, it returns to its posi- tion as soon as released. It becomes "set" in the movement which has taken place in it, and an unrolling and rolling up of the heliacal winding may be produced in long dead awns of the Grasses or in the fruit-stalks of Mosses, by unequal hygroscopic expansion or contraction of their parts. M. WICHURA ON THE WINDING OF LEAVES. 275 VI. Mechanical Constituents of the Movement : Revolution round the Axis and Curvature, §29. The movement of winding is in all cases founded on a revolu- tion of the leaf around the straight line which may be imagined to pass from the point to the middle of its base. I call this the axis of the leaf, to distinguish it from the axis of the helix, which, as we have seen, may lie outside the substance of the leaf. This revolution round the axis, with the resistance, increasing from the freely moveable point down to the fixed base of the leaf, which is opposed to it within the parts of the leaf itself, causes a change in the relation of the direction of the parts of the leaf to one another. Under its influence, the substance of the leaf, originally expanded in a rectilinear plane, becomes converted into a body wound in the form of a helix, within which only the median line of the leaf, as the fixed axis around which the two laterally situated halves revolve, retains its original position. The forms of twisted leaves mentioned in § 19, in which the axis of the heliacal winding coincides with the axis of the leaf, are consequently fully explained by the mere assumption of a re- volution of the leaf round the axis. §30. Very frequently a curvature of the leaf is added to the revolu- tion round the axis. Then a mixed movement is produced, whence result the heliacally wound leaves with the axis of the helix lying outside the leaf. That it is so may be strictly proved mathematically by the aid of analysis. But a conviction of the importance of the explanation given may be also readily obtained in the empirical way, by holding a riband-shaped piece of wax by one end, and twisting it around its long axis while curving it at the same time toward one of its flat surfaces. In this way are produced heliacal bands which agree in all essential points with the forms of leaf represented in figures 8 and 9. 18* 276 M. WICHURA ON THE WINDING OF LEAVES. VII. Direction of the Moveynent of Curvature. §31. The leaf may be curved, toward either the upper or lower face. It follows from the nature of the mixed movement pro- duced by a combination of curvature and revolution around the axis, that the concave side of the curve is directed towards the inside of the heliacal winding, the convex looking outward. Wound leaves which are at the same time curved, are conse- quently divisible into those with the upper and those with the lower face turned to the inside of the winding. §32. Examples of both kinds of winding leaves occur in nature. To the first category belong, for instance, the stem- leaves of most of our Gramineae and Liliaceae, to the latter the acicular leaves of Pinus sylvestris, L., and the anthers of Erythrcea Cen- taurium, L. An individuality of a peculiar kind occurred on an Allium transplanted from the Halle Botanical Garden to that of Berlin, which was cultivated in the latter in the years 1848-49 under the name of A. simplex, and perhaps still exists there. The two alternating series of opposite leaves of this plant were curved to one side in the earliest condition, in such a manner that in one series the upper and in the other the lower sides of the leaves formed the curved surface. If revolution around the axis is subsequently added to this, we get in one series leaves with the lower faces, and in the other leaves with the upper faces turned toward the interior of the helix. The two kinds of wound leaves elsewhere only found, in the same leaf-metamor- phosis, on different species, were here combined in the same individual. VIII. Direction of the Revolution around the Axis. — Terminology, § 33. In the revolution round the axis likewise only two different directions are conceivable. They receive names from an oppo- M. WICHURA ON THE WINDING OF LEAVES. 277 sition in the dimensions of breadth, right and left. But botanists are not agreed which of the two opposite helices shall be called a right- and which a left-wound helix. According to Linna3us, the Bean winds to the right and the Hop to the left. This definition is arrived at in standing outside the winding, and in tracing the latter from below upwards with the face turned towards it. In this way both the observer and the twining stem examined retain their natural positions. DeCandolle proposed an opposite method. In defining the direction of the winding, he placed himself in the central point, and, on the contrary, called the Hod ri";ht- and the Bean left- wound. Most modern authors have followed him, because, as they say, " in every in- dependent object, rif/ht and left can only be determined accord- ing to itself, from its own above and below, front and back,^' §34. The correctness of this reason, however, and the preference herein claimed for the newer terminology, cannot be admitted. According to Kant* the right or left nature of the winding of a spiral is a distinction "which is indeed given in perception, but cannot by any means be clearly conceived, and thus cannot be made comprehensible.^' We do indeed see, when we hold two oppositely wound helices against each other, that they represent a perfect contrast in their relations to space, and in this way we acquire a conception. But the examination of either singly will not carry us to the same result. All that we attain is a percep- tion in regard to space, and thus terminology would have fully gained its object, if it succeeded in reproducing in us the im- pression of the direction in space of any helix. Both DeCandoUe's and Linnaeus's methods fulfil this, but neither of these does any more. I have therefore returned to the terminology of Linnaeus, since it is not only the older, but also has the preference over that of DeCandolle on account of greater convenience in use. § 35. Another remark which I have to make on terminology refers to the " ajstivatio contorta." Linnaeus called flower-buds rolled according to the diagram fig. 1. no. 2, right- wound, in which the * Metaiihysische Anfangsgriindc, 1787, p. 8. 278 M^ WICHURA ON THE WINDING OF LEAVES. right border of each leaf covers the left of its neighbour, and those where the opposite case occurs, as in fig. 1. no. 1, left- wound. But the diagram fig. 2 shows that the flower-buds which Linnaeus called right- wound owe their origin to a revolu- tion of the individual leaflets to the left, and vice versd, I shall therefore deviate from the terminology of Linnaeus in this point, and call buds corresponding to the diagram figs. 1 & 2. no. 1. right-wound,and those agreeing with figs. 1 & 2. no. 2. left-wound. IX. Direction of the Revolution around the Axis in reference to the Systematic Divisions of the Vegetable Kingdom. §36. The direction in which leaves wind stands, like all the other characters of plants, in a certain relation to the systematic sub- divisions of the Vegetable Kingdom. Individuals of the same species behave uniformly in the direction of the heliacally wound leaves, and there are but few exceptions to this. Thus, for instance, specimens of Medicago littoralisy Rohde, occur with the legumes wound to the right and others with them wound to the left. §37. Whole genera including only uniformly winding species are somewhat rare. When the direction changes within the same genus, the species which have like laws of direction may some- times be arranged in common subsections of the genus from agreement also in other essential characters. The genus Allium furnishes a remarkable proof of this. The Allia with left-wound stem-leaves, as A. acutangulum, Schrad., A. Moly, L., &c., with perhaps the single exception of ^. azureum, Ledeb., all have a leafless flowering stem, while right- wound stem- leaves occur only in the species with a leafy flowering stem, as A» oleraceum, L., A, sphoerocephalum, L., &c. §38. Natural families in which winding leaves are frequent, usually include species the leaves of which wind in different directions, but there are certain families in which a perfect agreement exists ; the petals of the Ebenaceac, for instance, are in all cases wound to the loll in the bud. M. WICHURA ON THE WINDING OF LEAVES. 279 X. Change of the Direction in the Winding Leaves of the same plant, §39. Plants exist in the leaves of which only one direction, to the right or left, is represented, others in which the leaves of one individual exhibit both directions. Such a change of direction in all cases presupposes a difference in the age or situation of the parts of the plant, since two heliacal lines, one turning to the right, the other to the left, form an absolute contrast, mutu- ally exclusive, and so can only occur either on different bodies, that is, in distinct places, or at different epochs of the existence of the same body. §40. The distinctions which, from the observations I have hitherto made, usually accompany the change of direction, are: 1. The different periods of age of the same leaf; 2. The differentiation of the leaf in reference to point and base; 3. The relative alti- tude of the insertions of different leaves ; 4. The dissimilar lateral insertion, both of the solitary and the verticillate leaves; and 5. The metamorphosis of the leaf. §41. Since, however, the change of direction depends, in some plants on one, in others on another of these distinctions, there exists a great variety in plants with winding leaves, which is increased moreover by the fact that the lateral direction of the helix, formed by the arrangement of the leaves, may affect in an opposite way the direction of the heliacal winding of the leaves themselves, so that the leaves are wound either in the same, or in the opposite direction, to that of the leaf-spiral. We have now to furnish some instances of the effect of those differences of the position or relative age of the plant on the direction of the heliacal winding. §42. 1 . Different periods of the age of the same Leaf, The leaves of the inner circle of the perianth oiPuya ccerulea, Micrs, P, guianensisy KL, Bilbergia zebrina — Fam. of Brome- 280 M. WICHURA ON THE WINDING OF LEAVES. liaceae ; and the petals of Chrystea speciosa, Ward et Hartw., Strophanthus divergens, Graham — Fam. of Apocynaceae ; Cycla- men eurojxBwrii L.^ Lysimachia punctata, Wallr. — Fain, of Pri- mulaceae ; — are all wound to the left in the bud, and to the right when the flower blows. The awns oi Arrhenatherum elatum, M. & K., at first wind slightly to the left below the angle ; sub- sequently, as the seed is maturing, to the right in the same situa- tion. It is probable that similar phaenomena will be shown to occur in very many Grasses with winding awns. The appendages of the seed (beak of the carpel?) of Erodium cicutarium, L^Herit., are wound to the left round the carpophore; after separating from it they wind to the right. See PI. IX. fig. 4. nos. 1 & 2. §43. 2. Differentiation of the Leaf in reference to Point and Base, Alstroemeria pelegrina — Fam. of Amaryllidaceae : — stem- leaves winding to the right at the apex and to the left near the petiole. Avena sativa, L., and the allied species Fhalaris minor, Li., &c., Lagurus ovatus, L. — Fam. of Gramineae, — and Xerotes purpurea,^r\di\. — Fam. of Juncaceae : — stem -leaves wound to the left at the points and to the right near the bases. (See fig. 5.) In like manner the awns of many grasses, especially of the genera Avena, Stipa, Danthonia, Slc, in which the angle or ' knee ' forms the boundary of the two oppositely-directed windings, so that the curvature is to the left above the knee and to the right below it. Chcotobromus Dregeanus, N. ab E., Ck. stiictus, N. ab E. : the palea runs out into two teeth above, the awn being fitted in between them ; the two teeth wind to the left, the awn to the right below and to the left above. Strophanthus dichotomus, D.C. — Fam. of Apocynaceae: — the left-wound flower- buds are slightly rolled to the left at their points and twisted to the right below. A second change of the direction occurs in the awns of Macrochloa arenaria, Koch. Immediately below the knee they wind to the right, then further down to the left, and quite at the bottom again to the right. The very long style of Protea grandijlora, Thunb., changes three times to opposite sides, but here the place which the two directions occupy is not determinate, as in the examples above mentioned, for apparently k M. WICHURA ON THE WINDING OF LEAVES. 281 the right or left direction assumes the uppermost place without any definite order. A similar change of direction is met with in the tendrils of the Passiflora and the fruit-stalks of many Mosses. In the former the relative situations of the two directions are indeterminate, as in the styles of Protea grandiflora -, in the latter they are fixed. The winding to the right is here usually uppermost, the winding to the left in the lowest place. But the Funariea form an ex- ception to this, since their fruit-stalks are wound to the left above and to the right below, §44. 3. Different relative altitude of insertion of the Leaves, Phalaris minor, L., Ph. ccerulescensy Desf., Ph. aspera, Retz., Ph. canariensis, L., Ph. paradoxa, L. — Fam. of Gramineae : — the first two leaves succeeding the cotyledon are wound to the left, the succeeding leaves wind to the left at their points and to the right below. In Avena saliva, L., and probably the allied species, the first stem-leaf, developed after the cotyledon, winds to the right, the following leaves wind to the left at the point and to the right below, and finally the uppermost stem-leaves, standing in the immediate vicinity of the inflorescence, wind to the left. In PFatsonia fulgens, Pers., and W, aletroides, Ker., — Fam. of Irideac — as well as in many plants of the family of the Grasses, e.g. Anthoxanthum odoratum, L., Calamagrostis Epigeios, L., Festuca rubra, Huds., &c., right- and left-wound leaves follow one another apparently without any fixed order. §45. 4. Unlike lateral insertion of solitary Leaves. Dichcea squarrosa, Lindl. — Fam. of Orchideae, — Eucalyptus marginata, Lk., E. stenophylla, Lk., E, micrantha, D.C. — Fam. of Myrtaceaj : — the distichously alternate leaves are wound in opposite directions on the opposite ranks. Chrysocoma Lino- syris, L. (see fig. 6), Galatella lini/olia, D.C, G. jMuctata, D.C. — Fam. of Compositae, — Andersonia prostrata, Sond., Sprengelia incarnata — Fam. of Epacrideae,— ilfe/a/ewc« stypheloides, L. — 282 M. WICHURA ON THE WINDING OF LEAVES. Fam. of the Myrtaceae — and Acacia 7nicracantha,Desv. (see fig. 7) — Fam. of the Mimoseae — have the stem-leaves vv^ound in a helix turning sometimes to the right, sometimes to the left, making the helix by connecting the points of insertion of the leaves on the stem, from below upwards, by the shortest line. In like manner are the distichous or many-ranked verticillately-arranged leaves of Pinus, e, g, P. sylvestris, L., P. Pinea, L., P. excelsa. Wall., P. MughuSy Scop., &c., determined in the direction of their winding by the direction of the helix which is described by the succession of scales at the base of each fascicle of leaves. The direction of this helix constantly agrees with the direction in which the leaves wind. The reverse occurs in the petals of Gillenia trifoliata, Moench. — Fam. of Rosaceae; the Silenece with three or several styles, e. g. Silene, Lychnis, Viscaria, Cu- cubalus ; as also in the species of Hypericum, Geranium, Linum and Oxalis, which in their contorted aestivation are wound in a direction contrary to the readily distinguishable spiral of the sepals. Probably the sometimes right-, sometimes left-wound flower-buds of Statice, Fam. of Plantagineae, — Lysinema, Fam. of Epacrideae, — Cistus, Fam. of Cistineae, — Lavradia ericoides, A. St. Hil., Fam. of Sauvagesieae, — Bombax, Helicteres, Fam. of Sterculiaceae, — Hermannia and Mahernia, Fam. of Byttne- riaceae, — Maronobea globulifera, L., Fam. of Clusiaceae, — Rici- nocarpus pinifolia, Desf., Fam. of Euphorbiaceae, — and the Malvaceae, are in like manner dependent on the direction of the forerunning leaf-spiral. There is great difficulty, however, in making out these, and I have not yet succeeded in deter- mining their direction*. §46. 5. Unlike lateral insertion of verticillate Leaves, Chironia frutescens, 1j,, Ch, grandiflora, Fam. of Gentianeae : — stem-leaves in binate whorls. The leaves of the individual whorls are wound in the same directions, the leaves of alternate * The connection between the direction of the contorted aestivation and the arrangement of the leaves was explained by Alexander Braun so long ago as the year 1838, before the Naturforscher-VersammluiKj at Freiburg, Breisgau. — See A. Braun, On regular rotation (or twisting) in the Vegetable Kingdom. Flortty 1839, i. 311 et seq. M. WICHURA ON THE WINDING OF LEAVES. 283 whorls toward opposite sides. Eucalyptus punctata, D.C., E.floribundu, Hiigel, E, corymbosa, Sm. — Fam. of Myrtaceae : — stem-leaves in binate whorls. In each individual whorl the petioles of the opposite leaves are wound towards opposite sides. In the parallel whorls, i. e. those which are separated by an in- termediate, alternating whorl, leaves wound in like directions are found upon the same sides. The same laws of direction are followed by the stem-leaves of a Podocarpus, Fam. of Coniferae, — a specimen of which, consisting of a branch without flowers, labelled " ex horto Liverpool,'^ exists in the General- Herbarium at Berlin. §47. 6. Different Metamorphoses of the Leaf, Narcissus moschatus, L. : — stem-leaves wound to the left, segments of the perianth towards the right, after flowering. ^chmea, Puya, Pitcairnia, Billbergia, Tillandsia — Fam. of Bro- meliaceae: — outer perigone wound to the right, the inner to the left, in the bud. Pavetta indica, L., P. caffra, Thunb. &c. — Fam. of RubiacejE : — segments of the perianth wound to the right in the bud, the bursting anthers to the left. Lychnis coro- naria, Lam., L, chalcedonica, L., L. flos-cuculi, L., &c. : — petals wound in the bud in the direction contrary to the spiral of the sepals, the style wound to the right. The same relation exists between the petals and the appendages of the carpels in the Geraniaceae. Chironia fruiescens, L., Ch, grandiflora, Lam. : — stem-leaves on the alternate whorls wound in opposite direc- tions ; petals in the bud and dehiscing anthers towards the lefl. Arthrosteinma Humboldtii, Fam. of Melastomaceae : — sepals woimd to the left in the bud, the points of the anthers again to the right. Cistus and Helianthemum : the three larger sepals and the petals wound in opposite directions in the bud. Con- sequently, were it correct (see § 45.) that the petals of the Cis- tincac wind in the bud, in the direction opposed to that of the leaf-spiral, the sepals must be wound in the same direction as the latter. 284 M. WICHURA ON THE WINDING OF LEAVES. XI. Regular Succession of the two Opposite Directions. §48. The examples of leaves which wind toward opposite sides, mentioned in § 12, all agree in the circumstance that the direc- tion to the left invariably precedes the subsequently occurring winding towards the right. The leaves with their points and bases wound in opposite directions exhibit a similar condition. With the single exception of Alstroemeria pelegrina, the upper part of the leaf always winds to the left, the lower part to the right. Here therefore the winding to the left likewise precedes the winding to the right, since the upper portion of the convo^ luted organs is always older than the lower. §49. The application of the same law may further be shown in organs wound only in one direction, insofar as they are to be re- garded, either wholly or in the winding part, as metamorphosed summits or bases of leaves. The summits of leaves, consequently the style, stigmas, anthers, and, above all, the petals in contorted aestivations, which wind at a period when the point only of the leaf has been protruded from the receptacle, — follow the direc- tion toward the left in a vast majority of cases. The blade of the stem-leaves, as a superior region, is most generally, although not quite so frequently, w^ound to the left, while the sepals and carpels, as metamorphosed forms of the sheathing portion of the leaf, are predominantly w^ound to the right. § 50. From this it follows that the two directions to the right and left stand in a fixed relation to the different periods of the age of the leaf, which at the same time have their material expres- sion in the upper and lower parts of these organs. The winding to the left belongs to the earlier period of growth, and occurs chiefly in the upper parts of the leaf. The winding to the right follows at a later period, and is connected principally with the upper parts of the leaf. Since the stem, considered in relation to the leaf, represents a structure standmg underneath, it is completely in consonance with this law, that, as we find, the M. WICHURA ON THE WINDING OF LEAVES. 285 winding stems are turned to the right in about the same pro- portionate majority of cases, as the petals, at the other pole, are wound to the left in the contorted aestivation. §51. The relations are somewhat different in the Mosses. As already mentioned, the winding to the right occupies the upper part, that to the left the lower part of the fruit- stalks of Mosses. The winding fruit-stalks of the Mosses are further distinguished from the leaves of Vascular plants by the circumstance that in the latter the upper winding is developed before the lower, while in the former the lower is shown before the upper. Since then the lower winding of the fruit-stalk of the Mosses is directed to the left and the upper to the right, the winding to the left precedes that towards the right here also. The only exception to this occurs in the fruit-stalks of the Funariece, which wind to the right below and to the left above ; and in these, therefore, the winding to the right precedes that towards the left. § 52. A similar relation to that just demonstrated between the winding to the left and right and the earlier and later, or upper and lower parts of the leaf, may perhaps be assumed of the w indings which derive their determination from the direction of the leaf-spiral, either in the positive or negative sense. We are led to this conjecture by the fact that leaves wound in the direc- tion opposite to the leaf-spiral occur only in the contorted aesti- vation, and leaves following the direction of the latter, only on circles of stem -leaves, and perhaps the sepals of the family of Cistineae. The contorted aestivation, on account of the wind- ings, perfectly determinate in their direction in all cases, is espe- cially important, as the principal seat of the upper winding, from the fact that the wound flower-buds take the direction toward the left in a preponderating multitude of cases. It is very probable that a similar import is to be ascribed to it, to that of windings determined in their direction by the leaf-spiral. Under this hypothesis, therefore, we should have to regard the winding contrary to the direction of the leaf-spiral as the upper or earlier, the direction of the leaf- spiral as the lower and later. 286 M. WICHURA ON THE WINDING OF LEAVES. XII. Directio7i of Winding Leaves in relation to the Direction of the Winding Stem. §53. The movement of winding displays itself in two different forms in the stems of plants. There are stems which by their move- ment twine round a support, and stems which, like winding leaves, complete their movement in freedom without twining round a prop. The remark made above, that stems, as structures situated below, wind predominantly toward the right, and thus form a contrast to the blades of the stem-leaves more frequently winding to the left, holds good, as it remarkably appears, only of winding stems of the former kind. The hitherto little-known winding stems of the latter kind, which occur not unfrequently in the Monocotyledons, wind always in the direction in which the stem-leaves of these plants are wound. Thus we find left- wound flower-stems and stem-leaves in Eleocharis palustris, R. Br., Fam. of Cyperaceae ; in all the species of Xyris, Fam. of Xyridaceae, with which I am acquainted ; in Butomus umhella- tus, Fam. of Butomeae ; Allium acutangulum, Schrd. ; All.fallax, Don; AIL ursinum, 1j. ', All. Stellerianum, WiWd.; All.flaves- cens, Bess. ; Tulbaghia alliacea, L. ; T, cepacea, L., Fam. of Liliaceae ; Leucojum cestivum, L., Fam. of Amaryllideae. — Right- wound flowering-stems and stem-leaves in Aristidea megapota- mica, Spr., Fam. of Graminese ; Papalanthus perpusillus, Kl. ; P. Ottonis, KL, Fam. of Eriocauloneae; Morcieafili/brmiSyThunh., Fam. of Irideae. Lastly, in Sisirynchium anceps, belonging to the family of the Irideae, the leaves and stems of some speci- mens wind to the right, those of others to the left, so that in spite of the change of direction occurring in the same species, leaves and stem are always wound toward the same side in the same individual. XIII. Individual observations on the Direction of the convolution round the axis, collocated in the order of the Natural Families, § 54. It is a very remarkable phaenomenon that the circularly or heliacally acting forces of nature follow an unchanging, definite M. WICHURA ON THE WINDING OF LEAVES. 287 lateral direction in their course. In cosmical nature the planets describe heliacal lines winding to the right in space, by virtue of their circulation from west to east, since this is combined with the advance, in company with the sun, toward a point in the northern hemisphere. In the department of physics we meet with allied phaenomena in the circular polarization of light and in the course of electro-magnetic spirals. Organic life exhibits the same laws in the circulation of the blood, in all cases starting from the left side of the animal body, and in the heliacal windings of the shells of Mollusks, which follow a direction determinate for every species. But plants above all give evi- dence of a wonderful obedience to such laws, in the direction of the spiral vessels, the heliacally winding trunks of trees, winding stems and leaves, and probably also in the circulation of their saps. In regard to the conceptions formed by the in- tellect, it makes no difference at all in our views of the essential nature of any force working in lateral direction, whether it sets out towards the right or the left. In the rare cases in which the human heart has been placed in the right side instead of the left, it has been demonstrated that this anomaly did not interfere with the vital powers of the body. Nevertheless, Nature, as though dealing here with the most important business, displays the most surprising regularity in this very lateral direction. This apparent contradiction between our ideas and the phaeno- mena of Nature has something in it so remarkable, that a rather detailed account of observations on this particular point may seem desirable. I shall therefore here subjoin my observations on the direction peculiar to winding leaves, singly, in the order of the Natural Families. This collocation will at the same time serve to furnish a summary of the material of which I made use. Jungermanniaceae. — Jungermannia Trichomanis, Dicks. : de- hiscent valves of the fruit wound to the left. §56. Musci frondosi. — Cinclidotus fontinaloides ; Dichelyma falca- turn ; Bartramia pomiformis, Hedw. ; Tor tula ruralis, Swartz : stem-leaves wound to the right. Cinclidotus, Barbula, and 288 M. WICHURA ON THE WINDING OF LEAVES. Syntrichia : teeth of the capsules, to the right. Encalypta strep- tocarpa, Hedw. ; Barbula rigida : walls of the capsules, right. Barbula anomala, Br. & Sch., furnishes a solitary example of a Moss with the teeth of the capsule winding to the left. From the notice on this plant in the ' Europaischen Laubmoose,' however, " Peristomii denies complures ad dextrum [i. e. to the left in the Linnaean sense) convoluti," it would appear that in- stances of the opposite winding also occur, and it would be interesting to discover whether the fruit-stalks also are wound in the contrary direction when the teeth of the capsule are. I have not seen this plant. §57. Filices. — Lygodium circinatum, Swtz.; L. salicifolium, Presl; L. polymorphum, H. B. & K. : midrib of the frond without defi- nite order, sometimes to the right, sometimes to the left. — Palm* saw two stems of Ophioglossurn japonicum (with which I am unacquainted) wound round one another toward the right, that is, to the left in our terminology. §58. Lycopodiaceaj. — Lycopodiwn contiguum, Kl. ; L. mendiocca- num, Raddi ; L. inundatum : stem-leaves slightly wound, with prevailing but not exclusive direction toward the right. § 59. Graminaceae f. 1. Oryzece. — Caryochloa chilensis, Spr. : awns right. 2. Phalaridece. — Phleum, Alopecurus, and Beckmannia : stem- leaves constantly left. Baldingera arundinacea, Gaert. ; Holcus lana^us, Li.; H. mollis, L. ; Hierochloa borealis, R. et S. : stem- leaves right. Anthoxanthum avenaceum^ Retz. : awns right at the lower part. — For the species of Phalaris see § 44. 3. Panicece. — Setaria verticillata, Beauv. ; S. viridis, Eeauv. ; S. glauca, Beauv. ; Pennisetum fasciculaium, Trin. : stem-leaves ♦ Ueher das W'mden der Pfianxen. Stuttgart, 1827, pp. 41, 42. t See Alex. Braun on the Italian Rye-grass, ^ Flora,' 1834, Jalirg. xvii. Bd. 1 . 262, 2G3, whose attention was already directed to the property of Grass leaves and the awns, of curling in definite directions. M. WICHURA ON THE WINDING OF LEAVES. 289 right. Apparently the winding of the leaves is more rare in this section, inclined to a broader development of the leaves. 4. Stipacece. — Stipa mongolica, Turcz. ; Lasiagrostis Calama- grostis, Lk. ; Aristida meg apot arnica, Spr. : stem-leaves right. Aristida vestita, Thunb. ; A. chapadensis, Trin. ; A, Kotschyi, Hochst. ; A. divaricata, H. B.W. ; A. Sieberiana, Trin., &c. : lower paleae of the flowers right. 5. Agrostidece, — Polypogon monspeliensis, Desf. ; P. mewicanus, Spr. ; Agrostis canina, L. ; A. vulgaris, Withering : stem-leaves right. Apera Spica-venti, Beauv. : stem-leaves left. 6. Arundinacece. — Calamagrostis Halleriana,T).C.; C.LangS' dorfii, Trin. ; C. littorea, D.C. ; Arundo poceformis, Labil. : stem- leaves right. Desyeuxia retrofracta, Kth. ; Lachnagrostis splen- dens, Trin. : stem-leaves left. Calamagrostis sylvatica, R. Br., and Desyeuocia eriantha, H. B. K. : awns right at the lower part. For the stem-leaves of Calamagrostis Epigeios, Roth, see § 44. 7. Chloridea. — Chloris petrma, Sw. ; Diplachne poceformis, Hochst. : stem-leaves right. 8. Avenacece, — Aira do/^mc«, Wahlenb. ; Avena flavescens, L. : stem-leaves right. Trisetum neglectum, R. & S. ; TV. Alopecurus, R. & S. : stem-leaves left. For the doubly wound stem-leaves and awns of the Avenaceae, as also the doubly wound awns of the Grasses generally, see §§ 42, 43. 9. FestucacecB. — Poa, Bromus, Festuca Myuros, L. : stem- leaves left. Festuca alplna, Sut. ; F, arundinacea, Liljebl. ; F, pratensis, Huds. ; Bromus cmspitosus. Host. ; Br, rupestris. Host.; Brachypodium ramosum, R. & S.; Cynosurus echinatus, L.5 Briza media, L. ; Melica pyramidalis. Lam. : stem-leaves right. As the genera Bromus and Festuca are so extremely nearly aUied, it is important for their distinction that the former are predomi- nantly wound to the left, the latter to the right. For the stem- leaves of Festuca rubra, Huds., and Melica altissima, see § 44. Bromus madritensis, L. ; Br. rubens, L. ; Br. scaberrimus, Te- nore : awns slightly left. Bromus confertus, M. B. ; Br. wol- gensis, Spr. ; Br. patulus, M. & K. ; Br. lanceolatus. Roth : awns slightly left. 10. Hordeaceae. — Lolium: stem-leaves right. Triticum, Se- cale, Elymus, jEgilops: stem-leaves left. Elymus Caput-Me- dus(R, L. ; E. platantherus, L. : awns left. SCIEN. MEM.— iV^a«. Hist. Vol. I. Part IV. 19 290 M. WICHURA ON THE WINDING OP LEAVES. 11. RottboelliacecB, — Lepturus incurvatus, Trin. : stem-leaves right. Hemarthria fasciculata, Kth. : stem-leaves left. 12. Andropogonea. — TricholcBna tonsa^ N. ab E. ; Elionurus argenteus, N. ab E. : stem-leaves right. Saccharum cylindricum : stem-leaves left. Apluda microstachya, N. ab E. : stem torn at the apex into filiform strips which wind to the left. Anthistiria Wightii, N. ab E. ; A, tremula, N. ab E.; A. abyssinica, Hochst. : awns right at their lower part. Erianthus contortus, Elliot : awns left, §60. Cyperacec^. — Cyperus refleams, Vahl ; Cyp, difformis, L. ; Cyp. auricomus, Sieb. : stem-leaves right. Papyrus antiquorum, Willd. : the narrow glumes enveloping the inflorescence right. Chrysithrix capensis, L. ; Fimbristylis torfa, Kth. ; Eleocharis palustris, R. Br. : stem-leaves left. Carex hirta, L. ; C. ampul- lacea, Good. ; Rhyncospora alba, Vahl ; Schoenus nigricans, L. ; Fimbristylis ferruginea,\sLh\; F. junciformis, Kth. ; Androtri- chum polycephalum, Kth. ; Ficinia striata, Kth. ; Eleocharis glaucescens, Schult. ; E. palustris, R. Br. ; E. uniglumis, Lk. ; Scirpiis maritimus, K. ; Scirpns glaucus, N. ab E. ; Eriophorum gracile, Koch ; E, latifolium, Hoppe : anthers left when de- hiscent. §61. Eriocaulon€(B, — Pcepalanthus perpusillus, K\,; P, Ottonis, Kl. : stem-leaves right. §62. Xyridece. — Xyris communis, Kth. ; X. subulata, R. & P. ; X. arescens, Kth. ; X, str obilifera, Kth. ; X metallica, Kl. : stem- leaves left. §63. Commelinacece.—TradescantiaLyoni: stem-leaves left ; I must remark, however, that I have seen but one specimen of this plant, on the " Pfauen-InseV^ near Potsdam. § 64. Juncaginece. — Scheuchzeria palustris, L, This plant presents individuals with right- and others with left-wound leaves. Leaves of the same specimen wind constantly in the same M. WICHURA ON THE WINDING OP LEAVES. 291 direction. Perhaps the direction of the leaf-spiral has an influ- ence here in the direction of the winding, producing, by its difference in different specimens, a corresponding difference in the direction of the winding leaves. I have no exact observa- tions on this point. Butomacece, — Butomus umbellatus, L. : stem-leaves left. §65. Juncacea. — Xerotes mucronata, R. Br. ; X. cemula, R. Br. ; X effusa, Endl. ; X. suaveolens, Endl. ; X, rupestris, Endl. ; X, micrantha, Endl.; X, flexifolia, R. Br. ; X rigida, R. Br. ; X longifolia, R. Br. : stem-leaves lefl. Xerotes umbrosa, Endl., and two species from Major Mitchell's Expedition, distributed unnamed under the Nos. 542 and 564 : stem-leaves right On the doubly-wound leaves oi Xerotes purpurea, Endl., see § 43. The stem- leaves of some species of Juncus, in particular Juncus compressus, Jacq., seem to behave like the stem-leaves of Scheuch- zeria. In many specimens the leaves wind to the left, in others to the right; the right- winding predominate. Lmzula albida, D.C. ; L. campestris, D.C. ; anthers after dehiscence lefl. The stigmas of all the species of Juncus and Luzula with which I am acquainted wind to the left before the flower opens. § 65*. Melanthaceae, — Tofieldia calyculata, Wahlenb. ; Colchicum Szovitsii, Fisch. et Meyer ; C. autumnale, L. ; C. variegatum, Bivon : stem-leaves left. Merendera caucasica, M. B. : stem- leaves sHghtly right. §66. Liliacea, — Omithogalum umbellatum, L. ; O. Gussonii, Ten. ; O. lanceolatum, Labill. ; Bulbine annua, Willd. ; Johnsonia mu- cronata, Endl., as well as all the numerous species of the genera Lilium, Eremurus and Asphodelus I have seen : stem-leaves right. Hyacinthus orientalis, L. ; Thysanotus proliferus, Lindl. ; Tut- baghia alliacea, L. ; T, cepacea, L. ; AcanthocarpusPreissii, Lehm. : stem-leaves left. In the genus Fritillaria occur species with lefl- and with right-winding leaves. To the former belong, for example, F. imperialis, L. ; F. Meleagris, L., &c. ; to the latter, 19* 292 M.WICHURA ON THE WINDING OF LEAVES. F, pyrenaica, L., and F, montana, Hoppe. Between many of the oppositely winding species there exists so close an affinity that hybridizing would be most probably successful. It would be interesting to discover the direction in which the leaves of such a hybrid would wind. — For the winding leaves of the Allia see § 37. Gagea minima, Salisb. : segments of the peri- anth left during the withering of the flower. Massonia latifolia, L. : the filiform bracts of the flowers right. §67. Smilacem. — Streptopus perfoliatus, L. : segments of the peri- anth left after the flower has opened. Paris quadrifolia, L. : the attenuated points of the four stem-leaves, the leaflets of the perigone, both those of the inner and the outer, wind to the left in the bud in the contorted aestivation, and likewise after the flower has opened ; finally also the dehiscing anthers and the style, all left. § 68. Hamadoracece, — Conostylis bracteata, Hcemodorumplanifolium: stem-leaves right. Barbacenia tricolor, Mart. : stem-leaves rolled together like ringlets, without definable order, sometimes left, sometimes right. Vellozia variabilis, Mart. : the dehiscing anthers left. §69. Hypoxidece, — Hypooois gracilis, Lehm. ; Curculigo brevifolia : stem-leaves right. § 70. IridecB. — Morcea filiformis, Thunb. ; Tritonia lineata, Ker. ; all the species I know of Ixia and Gladiolus, together with the species of Iris with straight, narrow, linear leaves, e. g. Iris si- birica, L. ; /. Pallasii, Goldb. ; Trichonema chloroleucum, Ker., and several species of Tritonia which were temporarily culti- vated without names in the Berlin Botanic Garden in 1848: stem-leaves left. In the species of Iris with broad, sabre- shaped leaves, e. g. I. lurida, Ait., /. germanica, L., and /. Pseud- Acorus, the attenuated points wind sometimes right and some- times left. For Watsonia fulgens, Pers., and W, aletroides, Ker., M.WICHURA ON THE WINDING OF LEAVES, 293 see § 44; for Sisyrinchium anceps, § 53. The petals of the spe- cies of Iris are constantly wound to the left in the contorted aestivation. Aristea cyanea, Ait. ; A. bracteata, Pers. ; A, ca- pitata, Gawl. ; A. ccsrulea, Vahl ; A. spiralis, Vahl : petals wind- ing to the left as the flowers wither. The withering flowers of Iris exhibit the same twistings, apparently however without de- finite direction, §71. Amaryllidea.— Hitherto I have only met with left-wound stem-leaves in this family. The species examined were : Leu- cojum (Bstivum, L. ; Sternbergia colchiciflora, W. K. ; Brunsvi- gia JosephincBy Ker. ; a Cyrtanthus, the leaf of which is repre- sented in fig. 8 ; Zephyranthes Candida, Herb., several species of Narcissus, Gethyllis and Alstroemeria, For the doubly- wound stem-leaves of Alstroemeria pelegrina, see § 43. Narcis- sus moschatus, L. : petals right soon after the flower opens. Pancratium patens, Red. ; P. expansum, W. Herb. ; P. lobatum, Kl. : dehiscing anthers right. §72. Bromeliacece. — See § 4? and § 42, for the winding of the outer and inner perianth in the bud, and of the leaves of the inner perigone after the flower is open. We also observe a constant winding towards the left in the style. §73. Orchidea, — Epidendrum macrochilum, Hook.; Scaphy glottis violacea, Cyrtochilum flavescens, Diuris filifolia, Lindl. : stem- leaves left. Ornithidium album. Hook. ; Cymbidium giganteum ; Disa Zeyheri, Sond. ; D, tenella, Sw. : stem-leaves right. For Dichaa squarrosa, Lindl., see § 45. Epidendrum cochleatum, Li, ; Trichopilia tortilis, Lindl. : perigonial leaflets left, with the exception of the labellum which does not wind. Orchis hircina. Scop. : lobes of the labellum, especially the long middle lobes, left. Cypripedium Calceolus, L. : the pair of corresponding segments of the outer perigone right. Every one is aware of the half revolution of the germen of many Orchideae, by which the position of the parts of the flower in relation to the axis of 294 M. WICHURA ON THE WINDING OF LEAVES. the inflorescence appears reversed, so that the labeUum comes to the front, while in its original position it is turned towards the axis. No definite direction can be detected in this winding ; right- and left-wound ovaries succeed one another quite irregu- larly. It is possible that this revolution is a mere effect of irritability, caused by the effort of the labellum to assume an external position, corresponding to its greater proportionate de- velopment. §74. Aroidets, — Acorus Calamus, L. : stem-leaves left. §75. TyphacecB. — Typha latifolia, L. ; T. angustifoUa, L. ; T. ste- nophylla ; Sparganium ramosum, Huds. : stem-leaves left. Spar- ganium natans, L. : stem-leaves right. The only example I know of a plant winding under water. The winding may be frequently prevented, or its effects rendered imperceptible even by slow flowing of the water. But I observed it very clearly in July 1847, on the heath behind Moabit, near Berlin, in little pools where the undisturbed development of the leaves was favoured by the calm and clearness of the water. §76. Palmce. — Maximiliana insignisy Mart. : anthers mostly right-, but occasionally left-wound. §77. Coniferce. — For the acicular leaves of the species of Pinus, see § 45. — Podocarpus elongatus, L^Herit. ; P. macrophyllus, Wall.; P.falcatus, L'Herit., and P. latifolius, Wall.: stalks of the stem-leaves and lowest part of the blades right. §78. CasuarinecE. — Casuarina suberosa, Otto and Dietr., and a Casuarina from Cuming, discovered in the Philippines and distributed as No. 730 : sheath-teeth right. Casuarina humilis, Otto and Dietr. ; C. quadrivalvis, Labill., and C. suberosa, Otto and Dietr. : style left. M. WICHURA ON THE WINDING OP LEAVES. 295 §79. Salicinece. — Saliie stipularis, Sm. : branches of the stigma right when dried up. §80. Nyctaginecs, — Bugainvillea spectabilis : withering flowers right. Pisonia grandis, R. Br. ; a plant which I have never seen my- self, but which is represented with right-wound styles in Endli- cher's ' Iconographia ' (Vindob. 1838). §81. Gyrocarpece, — According to Nees v. Esenbeck, — Wallich, Plantae Asiaticae rariores, ii. 68, — the cotyledons are wound heliacally round the two-leaved plumule : I have had no oppor- tunity of observing for myself. The direction of the winding is unknown. §82. Santalacea. — Thesium ramosum, Hayne : stem-leaves left. The stem-leaves of Thesium pratense, Ehrh., and Th, ebractea- turn, Hayne, are also often wound, but not in any definite direc- tion. Santalum cognatum, Miquel : petioles left, not quite \. §83. Daphnoidea. — Passerina pectinata: stem-leaves slightly to the lefl. The same in Daphne olecefolia, Lam. §84. Proteacece. — Persoonia quinquenervia, Hook. ; Hakea steno- carpa, R. Br. ; H, latifolia, Loddig. ; Protea longiflora, Lam. ; Pr. inflexa, Willd. ; almost all species of Leucadendrum, e. g. L. ascendens, R. Br. ; L, diversifolium, R. Br. ; L. tortum, R. Br. ; L, cuspidatum, KL, &c. : stem-leaves left ; in Hakea latifolia, Loddig., it is the attenuated points of the leaves which wind, in the rest the lower parts of the leaves which run down into an imperfect petiole. Conospermum capitatum, R. Br. : the narrow stem-leaves, nearly a foot long, right. Conospermum teretifolium, R. Br. : petals in the bud twisted like a cord to the left. For the winding style of Protea longiflora, see § 43. 296 M. WICHURA ON THE WINDING OF LEAVES. §85. Plumbaginece. — Armeria, in all the species with which I am acquainted : petals wound to the left in the contorted aestivation. The genus Armeria is essentially distinguished by this point from the allied genus Statice, in the distichous inflorescence of which right- and left-wound buds alternate regularly. But this law seems only to be manifested clearly in the blue-flowered species of the genus Statice. Two red-flowered species, one of which is St. acicularis, exhibited a predominance of right-wound flower-buds, with the intermixture of a small number of left- wound. §86. Composites, — Liatris punctata^ Hook. ; L. cylindraceay Mx. ; L. squarrosa, L. ciliata ; Eclopes parallelinervis, Lessing ; Gero- pogon glaber : stem-leaves right. Cephalophora aromatica, Schrad. ; Calycadenia villosa, D.C. ; C, cephalotes, D.C.; Me- talasia, R. Br. ; Stabe, Less. — almost all the species of these genera — Perotriche tortilis, Cass. ; Erythropogon umbellatum, D.C; Elytropappus glandulosus. Less.; E. ambiguus, D.C; Disparago ericoides, Gaertn.; J), lasiocarpa, Cass.; Centaurea depressa, M. B. : stem-leaves left. For Galatella and Chryso- coma, see § 45. Podolepis subulata, Steetz: points of the membranous leaves of the anthodium not quite \ right. Prenanthes purpurea, L. : radiant florets thrown back toward the left. R. Brown has directed attention to the w indings of the style to the extent of about \ of its circumference. The arms of the stigma which stand front and back in reference to the axis of the inflorescence, are thus brought into a lateral position. I have observed this phaenomenon, not in all, but in a great multitude of Compositae, and have never detected a de- finite law of direction. §87. Campanulacece, — Campanula Loreyi, Pollin. : apex of the calyx left. Campanula patula, L. ; C. Rapunculus, L. ; C, Lcefflingii, Brot. ; C. persicifolia, L. ; C. cana, Wall. ; Phyteuma canescens, W. et K. : dehiscing anthers to the right. Campanula rapun- culoides, L. ; C, Trachelium, L. ; C glomerata, L. ; C. barbata, M. WICHURA ON THE WINDING OF LEAVES. 29? L. ; C. Loreyiy Pollin. ; C Medium, L. : dehiscing anthers left. The species in which the anthers wind to the same side, exhibit a certain agreement in their habit, w^hich indicates that the two groups which are distinguished from each other by the opposite winding of their anthers may also be separated by other cha- racters. But this requires a more exact knowledge of this large genus than I possess. §88. Stylidece. — A slight revolution of the corolla brings the anterior, odd petal into a lateral position. I have only examined one species of this genus, which was cultivated in the Berlin Garden under the name of Stylidium suffruticosum. In this the winding of the corollas was without exception directed to the left. §89. Rubiacea. — Pavetta indica, L. ; P. Cqffra, Thunb. ; P, lan^ ceolattty Ecklon ; P. manitensis, Walpers ; P. longissima, Kl. ; Ixora Bondhuca, Rxbrg. ; /. acuminata, Rxbrg. ; /. parviflora, Vahl ; /. nigricans, R. Br. ; Wendlandia coriacea, Vahl ; Augusta lanceolata, Pohl; Randiaferoa?,D.C,; Gardenia latifolia, 'Rxbrg,: segments of the flowers right on the contorted aestivation. Ea^o- stemma maynense, Poepp. : segments of the flower left in the bud. Coffea arabica, L. ; Gardenia latifolia, Rxbrg. : dehiscing anthers right. Pavetta (the above-mentioned species) : dehiscing anthers left. The anthers of Gardenia latifotia, Rxbrg., and Pavetta, con- sequently follow opposite directions in their winding, while the segments of the flowers of both are turned in the same direction. §90. Jasminece, — Jasminum officinale, L. : segments of the flower left, in the contorted aestivation, which however not unfrequently passes over into the imbricated aestivation. Jasminum hirsutum, Hoffmsgg. : segments of the flower predominantly right, in the contorted aestivation. § 91. Oleacece, — Fraxinus heterophylla, Vahl: fruit slightly right during the ripening, §92. Apocynacece, — Segments of the flower wound in the bud, in 298 M. WICHURA ON THE WINDING OP LEAVES. certain species right, in others left. Alphonse DeCandolle*' has used the direction of the rolled buds in the descriptions of the generic characters. His very accurate statements are made use of in the following summary : — 1. Parsonsieee, Alph. D.C., and Ec/dtece, Don: seeds with a C07na above. Segments of the flower left in the bud. The genera Neriandra, Alph. D.C., and Ecdysanthera, Hook., funu exceptions with right-wound segments. Since, however, there has not yet been any possibility of examining their seeds, it is doubtful whether they actually belong here. 2. Wrightiece, Don : seeds with a coma below. Segments of the flower right in the bud. 3. Alstoniece, Don : seeds with a coma above and below. Direction of the segments of the flower variable. In Blaberopus, Alph. D.C., they wind to the right ; in Adenium, Roem. et Schult., left. The genus Alstonia, R. Br., includes several species with left- wound segments of the flowers, and one other species in which the segments wind to the right, — Alstonia scholaris, R. Br. 4. Plumeinem, Alph. D.C. : seeds naked, two separate ovarian segments of the flower wound to the right in the bud in most of the genera. They are wound to the left only in Kopsia, Blume ; Odontadenia, Benth. ; Malonetia, Alph. D.C. ; Anisolobus, Alph. D.C, and Thyrsanthus, Benth. 5. WillughbiecB, Alph. D.C. ; Carissecs, Alph. D.C. : seeds naked ; ovary simple ; segments of the flower right in the bud. A few other movements of winding of the parts of the flower occur more isolated : — Thevetia neriifolia, Juss. : funnel-shaped corolla with five- toothed border. In the bud it is weakly folded towards the left, and at the same time wound heliacally in the opposite direction, that is, to the right. Echites longiflora, Desf. : the left-rolled buds slightly twisted together to the left above. For Christya spinosa, Ward et Harvey; Strophanthus divergens, Graham, and Str» dichotomus, D.C, see §§ 42, 43. Parsonsia spiralis, Wall. ; P. oblonga, Wall. ; P. Cumingiana, Alph. D.C. ; P. heli- candra, Hook, et Arn. : filaments left. Nerium Oleander, L. : the filiform points of the anthers left. * DeCandolIe, Prodr. Syst. Natural, viii. 317-489. M. WICHURA ON THE WINDING OP LEAVES. 299 §93. Asclepiadece, — Segments of the flower predominantly left in the bud. I have not observed exceptions to this rule myself. But according to Decaisne*, the segments of the flower are wound to the right in the genera Secamone ; Toxocarpus, W. et A. ; Calostigma and Oxypetalum, R. Br. Microloma sagittata, R. Br. : points of the flower-buds twisted to the right. Pergu- laria accidens, Blume; P. edulis, E. M.: buds twisted to the left at the points, the unfolded segments of the flower wound to the left. Diplolepis Menziesii, R. et S. ; Oxypetalum ripa^ rium, H. B. K. : segments of the flower left after the flower opens. Probably the genus Pycnoneuron, Dene., also belongs here, since Decaisne says of it, " Laciniis post anthesin superne dextrorsum contortis," i. e. according to our terminology, to the §94. Gentianece. — For the winding of the stem-leaves of Chironia see § 46. Segments of the corolla of all the Gentianeae, per- haps with the exception only oiMenyanthes and Villarsia, wound in the bud, and everywhere toward the left. The tubes of the corollas of Gentiana are in addition laid in folds towards the right in the bud. Chironia, Erythroea, Plocandra palustris, Griseb. ; PL Krebsii, Griseb. ; Sabbatia corymbosa, Ell.: de- hiscing anthers to the left. Lisiarithus uliginosa, Griseb. : style left. Sabbatia corymbosa, Griseb. : style right. §95. Labiatae, — Burgsdorfia montana, Hort. Bot. Berol.: stem-leaves to the left at the base. Hyssopus lophanthus, L. ; Lophanthus chinensis, Benth. : tube of corollas right. The revolution amounts to about ^, so that the lip of the flower is carried to the upper side. Bentham, in his description of the Labiatae, explains this resupination incorrectly, by a revolution of the flower-stalk {. §96. Cordiacece, — Cordia Geraschanthus, L. ; C. complicata, Ruiz; C. alliodora, Ruiz : segments of the corolla left in the bud. * D.C. Prodromtis, I. c. 501, 504, 580, 581. f Ibid. p. 582. J Tom. i. Introduction, p. xxiii. 30b M. WICHURA ON THE WINDING OF LEAVES. §97. Asperifolice, — Pulmonaria maritima, and all the species of Myosotis I am acquainted with : segments of the corolla left in the bud. Trichodesma africanum, Lehra. ; TV. indicum, R. Br. ; TV. zeylanicum, R. Br, : points of the anthers wound together towards the left. §98. Convolvulaceae, — The corollas of Convolvulus and the allied ge- nera are folded to the right and at the same time wound helia- cally to the left in the bud. See § 92. for the contrast in Thevetia neriifolia, Juss., with which, in other respects, the folding and twisting of the corolla before the opening of the flower exactly agree. §99. Hydrophyllece. — Nemophila, all the species as yet known to me, and Hydrophyllum virginianumx segments of the corolla right in the bud. § 100. Scrophularinece. — Pedicularis : the tube of the corolla under- goes a revolution to the left, seldom exceeding \, during the opening of the flower, whereby the parts of the flower acquire an oblique position. Apparently this revolution commences at that part of the tube which is continued above into the lower lip. This is always most strongly twisted, and it may be clearly detected that the winding of the part of the tube of corolla be- longing to it commences independently and alone before the flower opens, and in a comparatively very young condition. Then, as the part of the tube belonging to the upper lip remains in its position, and the part belonging to the lower lip undergoes a curvature to the left, the left side of the lower lip becomes approximated to the upper lip in the same degree as it is removed from it on the right. The result of this is, that the left side of the lower lip becomes more intimately blended with the upper lip than the right, hence the sinus which separates the upper and lower lips on either side is always cut in deeper on the right than on the left. Pedicularis palustris, L. ; P. asplenifolia, Florke ; P. canadensis, L. ; P. contorta, and P. lanceolata, Mx., are species M, WICHURA ON THE WINDING OP LEAVES. 301 in which all these phaenomena are exhibited with especial clear- ness. Probably traces of them might be discovered in all spe- cies. Anthocercis littorea : the narrow, lanceolate segments of the corolla to the right after the flower is open. § 101. Acanthacece, — Gendarussa orchioides, N. ab E. : stem-leaves to the left. Peristrophe caulopsilay N. ab E. ; P. Kotschyana, N. ab E. ; P. speciosa, N. ab E. ; Hypoestes aristata, R. Br. ; H, polymorpha, N. ab E. : tube of corolla right. § 102. Primulacece, — Lysimachia punctata^ Wall. : sepals convoluted to the left at their tops in the bud. Lysimachia, Lubinia, Ana- gallis, Cyclamen, Trientalis, and Samolus littoralis, R. Br. : segments of corolla left in the bud. Cyclamen europceum, L. ; Lysimachia punctata, Wallr. : segments of the corolla right after the flower is open. §103. Gesneracece, — Dorcoceras hygrometrica, Bge. : the walls of the dehiscent capsule left. § 104. SapotacecB, — Mimusops dissecta, Spr. (Herb. Gen. Berol.) : the linear-lanceolate acuminated segments of the flower left, after the flower is open. § 105. Myrsinacecp, — Segments of the flower frequently wound in the bud, and then constantly towards the left. " Lobis petalisve, aestivatione varia, sa3pius sinistrorsum contorta/^ says Alph. DeCandolle* of this family. According to his observations, Ardisia Pickeringla, Torr. & Gray, has right-wound petals. The segments of the calyx are said to wind to the right. I have not investigated this family myself. I Prod. viii. p. 75. 302 M. WICHURA ON THE WINDING OF LEAVES. §106. jEgiceracece, — "Calyx 5-partitus, lobis sinistrorsum imbricato- convolutis Corolla lobis aestivatione sinistrorsum imbri- cato-convolutis/' Alph. DeCandolle*. This family also is un- known to me. § 107. Ebenacece, — Segments of the flower left, in the bud, in all Bpecies. §108. Epacridea. — On the winding of the stem-leaves oiAndersonia prostrata, Sond., and Sprengelia incarnata, see § 45 ; also, as to the winding of the segments of the flower in this family, and in the family of the Rosaceae, Caryophylleae, Hypericineae, Gera- niaceae, Lineae, Oxalideae, Cistineae, Sauvagesieae, Malvaceae, Sterculiaceae, Byttneriaceae, Clusiaceae, and Euphorbiaceae. § 109. Ericaceae, — In the great majority of species of the genus Erica the segments of the limb of the corolla are wound, and constantly towards the left. § 110. Saxifragece, — Vahlia capensis, Thunb. : stem-leaves left. §111. Ranunculacece. — Nigella sativa, L. ; iV. damascena, L. ; N, hispanica, L. : style to the left. Clematis azurea : appendages of the carpels wound to the right. § 112. Papaveracece, — Platystemon californicum^ Benth. : style of the ripening fruit left. §113. Mesembryanthece. — Mesembryanthemum aureum, L. : petals convoluted to the left in the budi * Prod. viii. p. 141. M. WICHURA ON THE WINDING OP LEAVES. 303 § 114. SilenecB. — Dianthus diutinus, Kit. ; D. atrorubens. All. ; D. tri- fasciculatus, W. et K. (fig. 9) : stem-leaves left. Dianthus, Gyp- sopldla, Saponaria, Vaccaria, and generally, apparently, in the Silenece with only two styles: petals to the left in the bud. The peculiar twisting of the petals in a direction contrary to that of the spiral of the sepals peculiar to the Silenece with three or five styles, and already spoken of above, does not present itself distinctly in all the species. Thus, for example, I observed in Lychnis chalcedonica, L., Silene Armeria^ L., and many other Silenece, that the left-wound buds predominate; while, if the petals of every bud were wound contrary to the direction of the spiral of the sepals, right- and left- wound flower-buds ought to exist in equal numbers, since the leaf- and consequently the sepal-spirals of each pair of opposite branches of the inflorescence of this family are wound in contrary directions. Such examples of plants with prevailing but not exclusively left- wound flower- buds form the transition from the Silenece with petals wound contrary to the direction of the sepal-spiral, — in which, there- fore, on account of the antidromy of opposite flower-branches, right- and left- wound flower-buds alternate regularly, — to the two-styled Silenece with only left-wound buds. Agrostemma Githayo is a plant in which the sepal-spiral may be particularly easily made out from the imbrication of the teeth of the calyx, and I recommend it for investigation of this point. It shows the law prevailing in the three- or many- styled Silenece with almost unalterable regularity. Dianthus deltoides, L. ; Lychnis coronaria, Lam. ; L, chalcedonica, L. ; L, Flos-cuculi, L. ; Silene repens, Patrin., &c. ; style right. §115. Frankeniacece, — Frankenia campestris, Schauer. ; F.pulveru- lenta, L. : tube of calyx right. § 116. Loasece, — Loasa lateritia. Gill, et Hook.: capsules mostly wound to the right, more rarely to the left. Blumenbachia in- signis, iSchrad. : capsules right. ■304 M. WICHURA ON THE WINDING OF LEAVES. §117. Sterculiacece, — For Helicteres, and its contorted aestivation, see § 16. Helicteres guazumcefolitty H. et B. : fruits of the two- flowered cymes probably wound towards opposite sides on ac- count of the antidromy of their leaf-spirals. I have only been able to examine the fruits of this one species, but think the rest would exhibit a similar character. §118. Xanthoxylece. — Ailanthus glandulosus, Desf. : fruits wound to the right (fig. 3) at the tips. §119. Oxalide(B. — Besides the winding of the petals in the bud, already mentioned, I observed a very distinct and constant winding of the stem-leaves to the right, in a species of Oralis which was cultivated in 1848, in Decker^s Nurseries at Berlin, under the doubtful name of Oralis palmata. § 120. Onagrarieae. — {(Enotherece) : segments of the corolla left in the bud. Lopezia : the upper part of the solitary existing sta- men to the right. §121. Combretaceae. — Combretum purpureum, Vahl : petals in the bud left. §122. PhiladelphecB, — Philadelphus coronariusy L. ; P. Gordonianus, Lindl. : petals in the bud right. Philadelphus hirmtus : petals in the bud left. §123. Melasto7nace(B, — Petals in the bud left, probably in all the Melastomaceae. For the winding of the sepals and anthers in Arthrostemma Humboldtii, see § 47. §124. MyrtacecB, — Callistemon and Metrosideros : stalks of the stem- leaves about \ towards the left. For the winding of the stem- leaves of Eucalyptus and Melaleuca, see §§45 and 46. k M.WICHURA ON THE WINDING OF LEAVES. 305 § 125. Papilionaceae. — Dillwynia ericifolia, Sm.; D, glaberrima, Sm. ; D, parvifolia, R. Br. ; Z). rudis, Sbr. ; D, laxiflora, Benth. ; D, pinea, Sbr. ; Coslidium ciliare, Vogel ; Amphithalea, Eckl. et Zeyh. : stem-leaves right. Trifolium circumdatum, Kze. ; TV. resupinatum, L. : the resupination of the flowers of these two plants is produced by a half revolution of the tube of the co- rolla toward the right. Similar windings of the tube of the corolla are met with also in Trifolium fragiferum. But the winding is much slighter here, inconstant in its direction, and amounts scarcely to a quarter of the circumference. Medicago the legumes of most of the species wind to the left. The fol lowing have right-wound legumes : — M, tuberculata, Willd. M. tribuloideSf Lam. ; M, rigidula, Lam. ; M, striata. For Me dicago littoralis, Rohde, see § 36*. Sesbania cegyptiaca, Pers. the very long linear articulated pod left. Crotalaria retusa, L. Cr, verrucosa, L. ; Cr. ovalis, Pursh ; Cr. quinquefolia, L. ; Cr. carinata, Steud. : the dried-up style right. The twisted styles of some species of Lathyrus, e. g. L, roiundifolius, Willd., and of Phaseolus, follow the same direction. The revolution, however, begins here in the bud, and is imparted to the keel tightly en- closing the style, the former thus Kkewise acquiring an hehacal winding. Petalostemon candidum, Mx. : drying-up style wound in many flowers to the right, in others to the left. Dalea bra- chyptera, Kunze : style slightly left. For the revolution of the phyllodia in Acacia micracantha, Desf., Fam. of Mimosece, see § 45. XIV. Cause of the Curvature of Wound Leaves. § 126. The curvature of the winding leaf depends either on an un- equal tension of its margin in relation to the axis, or on an in- equality of length of its two surfaces. The specific gravity, which otherwise is certainly to be enumerated among the causes of the curvature of the leaf, comes but little or not at all into * Sec also A, Braun on " Regular Revolutions in the Vegetable Kingdom," Flora, 1839, i. 313, in which the heliacal winding of the legumes of the species o{ Medicago is described correctly according to DeCandolle's method of de- fining it. SCIEN. MEM.— iV«/. HisL Vol. I. Part IV. 20 306 M. WICHURA ON THE WINDING OP LEAVES. account in the winding leaf. For since the weight always draws downward, if it were this that compelled the winding leaf to curve, the latter must, in every half revolution, according as it turned this or that side downward, curve, alternately, at one time to this, at another to that side. It must consequently — see § 31. — turn sometimes one, sometimes the other side to the interior of the helix. But it is just one peculiarity of wind- ing leaves, that they always turn one side to the interior of the helix, no matter how many revolutions the leaf may make. The cause of the curvature can only be sought therefore in the leaf itself, and not in any one-sided force acting upon it from without. §127. The unequal tension of the margins of a leaf in relation to the axis, and the parts lying next to it, is produced by the revolu- tion around the axis. It lies in the nature of the curve, as a bent line, that the heliacally wound lateral parts of a leaf with a straight axis must have passed over a longer course, and there- fore must be longer, than this axis itself, which travels through about the same distance by a straight path. If a leaf winding round its straight axis is unrolled, the margins are thrown into waves and folds, and thus present clearly to the eye the over- plus of longitudinal development which was expended in the formation of the heliacal curve. But such forms of the leaf are comparatively rare. In the great majority of leaves the borders and axis exhibit a perfect uniform longitudinal development, and no obstacle is offered to their parts being spread out in one plane. Hence it is clear, that, when such a normally formed leaf be- comes affected by revolution of its axis, its first effort must be directed to supply the want of predominant longitudinal develop- ment by elongation of the lateral part of the leaf. The extensi- bility of vegetable fibre is certainly capable of this up to a certain point. Vegetable fibre is, however, at the same time elastic, and thus the force which expands it is met at once by another force, which strives to draw it back again into its former dimensions. Only a part of this form can be taken off by the curvature of the axis, since the two forces do not act in diametrical opposition, but obliquely against each other. Another part of the force M. WICHURA ON THE WINDING OF LEAVES. 307 remains over, and draws the leaf downward from the point where the revolution of the axis takes place. The results are various, according as the leaf is strong enough to oppose a resistance to the pressure burdening it, or not. In the former case the re- volution of the axis is limited to the degree admitted by the extensibility of the fibre, which is so slight that the movement of winding possible under such conditions will remain almost invisible to our eyes. But if the leaf gives, and curves, there arises, as we have seen above, § 30, from the combination of revolution of the axis and curvature, a new movement, which is very essentially distinguished from mere revolution of the axis, insomuch that now all parts of the leaf, its axis included, take share in the heliacal winding around an ideal axis lying outside the substance of the leaf. The contrast which appears between axis and margins in a leaf wound round its own axis, is com- pletely removed, and it no longer requires a predominating longitudinal development of the lateral parts of the leaf to render the heliacal winding possible. In this way the obstacles opposed to revolution of the axis are removed by the curvature of the leaf, after the revolution of the leaf having itself previously pro- duced the curvature. The two movements then act as reciprocal causes, and the result is that revolution of the axis and curvature of the leaf are mostly combined ; winding leaves not curled are, on the contrary, very rare. § 128. At the same time it is true, that in the curled winding leaf, two other contrasts, likewise accompanied by an unequal tension in the substance of the leaf, present themselves in the place of the contrast between the axis and margins of the leaf, removed by the curvature. Under supposed equal longitudinal develop- ment of both surfaces of the leaf, in the first place, the surface turned outwards must be rendered more tense, since it describes a larger arc, than the surface turned to the interior of the helix, and this in the greater proportion as the leaf is thicker. Se- condly, from mechanical causes which it would occupy too much space to discuss here, one margin of the leaf, — namely in left- wound leaves with the upper face turned inwards, and in right- wound leaves with the lower surface turned inwards, the right 20* 308 M. WICHURA ON THE WINDING OF LEAVES. margin — as seen from below ; and in left-wound leaves with the lower face turned inwards, and also in right- wound leaves with the upper face inwards, the left margin — must be rendered more tense than the other, and this the more, the broader the leaf is in proportion to its length. But the leaves which, as we here pre- suppose, are curved by the pressure arising from the revolution of the axis, cannot, for that very reason, be very broad or very thick, since otherwise they would not give to the certainly very shght pressure acting upon them. The very slight breadth, in proportion to the length, is, as already remarked, a characteristic peculiarity of all winding leaves. The unequal tension of the surfaces and borders of the leaf caused by curled wound move- ment, is consequently really so inconsiderable, that it is doubt- less mostly compensated by the extensibility of vegetable tissue. §129. The unequal length of the two surfaces, and the curvature of the leaf caused by this, is the effect of a hygroscopic difference of these surfaces, on account of which their relative lengths alter in the drying-up of the leaf: whether this happens through an elongation or shortening of one of the two faces, I must leave undecided, from the want of accurate observations. This hy- groscopic curvature of the leaf acquires especial importance in regard to the revolution of the axis, when this requires a curva- ture of the leaf to bring it to light, not being strong enough in itself to produce this. The heliacal winding of the leaf is thus first made possible through the addition of the hygroscopic cur- vature to the revolution of the axis. Since the former, however, depends upon the desiccation of the leaves, only making its ap- pearance towards the end of the life of the plant, the revolution of the axis in this case makes its first appearance in the latest stages of growth as the last expression of an expiring vitality. § 130. Such winding leaf-structures, which are very remarkable, like the awns of the Grasses, the appendages of the carpels of the Geraniece, the carpels of Dorcoceras hygrometrica, Bunge, &c., possess the property of recovering the originally equal length of M. WICHURA ON THE WINDING OF LEAVES. 309 their two surfaces, in contact with moisture. The curvature is thus removed, and in the same proportion as the leaf stretches itself straight, the heliacal winding only possible under the hy- pothesis of a certain degree of curvature, is unrolled. As soon as the leaf dries and again curls, under the influence of the again in- creasing difference of length of its two surfaces, the heliacal wind- ing is restored. The whole of this movement depends therefore upon a play of mere mechanical forms, totally distinct from that originally causing the revolution of the axis. They nevertheless have been confounded in many cases. The heliacal winding of awns, the fruit-stalks of Mosses, &c., has been generally looked upon as an effect of their hygroscopic nature, without separating the two different movements contained in it, and the mechanical explanation of the whole process following as a consequence of this view, may have contributed to have drawn away the atten- tion of botanists from the striking physiological phaenomenon of revolution of the axis, which is exhibited so clearly in these very structures. XV. Causes of the Revolution of the Axis. §131. To cause the revolution of the axis, w^e must imagine a force circulating in a direction perpendicular to the longitudinal di- rection of growth. It is an immediate revelation of the vital force acting in the interior of the plant, and therefore stands in the closest connexion with the growth as its regular expression. Such a connexion may be detected in some cases by the cir- cumstance, that in the plants mentioned in § 45, the lateral direction in which solitary leaves follow one another in a spiral line, acts in determining the direction of the heliacal winding of the leaf, in other cases it becomes observable in the remarkable inverse proportion between transverse growth and revolution of the axis, through which leaves in which the growth in breadth has become developed in the angle-nerved venation, exhibit no revolution of the axis, while parallel-nerved leaves in which growth in breadth is deficient, develope that force rotating at right angles to the direction of longitudinal growth, requisite for the revolution of the axis. That this inverse proportion between growth in breadth and revolution actually does exist, and that the parallel- nerved leaves do not exclusively wind, merely 310 M. WICHUKA ON THE WINDING OF LEAVES. because the usually narrow form opposes the least resistance to the rotating force, perhaps present as in other leaves and only restrained in its effects, is testified by the occurrence of angle- nerved leaves of very long and narrow shape, which, never- theless, do not wind. So far as regards length and breadth of their parts, these stand in exact agreement with parallel-nerved leaves ; indeed the lobes and pinnae into which they frequently divide are sometimes much finer and more delicate than those of winding parallel-nerved leaves. If, therefore, the forces moving the parallel-nerved leaves were active in these, they ought to exhibit the same effect, and it should produce an heli- acal winding of the angle-nerved leaves in a determinate direc- tion. But we do not see this ; — proof, consequently, that the angle- nerved leaves are devoid of the rotating force, and that the cause of the absence of revolution of the axis from them does not arise from their merely opposing a greater resistance than the parallel-nerved leaves to a rotating force existing in them. Hence latitudinal growth of the leaf and rotation of the axis are really mutually exclusive. When the former is deve- loped, the latter is wanting, and so it seems to follow from this that one and the same force lies at the foundation of both, manifesting itself either in the transversely directed activity of growth, or, when this is absent, in a lateral movement of the leaf. § 132. While attention was paid merely to winding stems, it was possible to suppose the object of this motion to be that of ob- taining attachment to outward objects. We are now aware that a great proportion of winding stems, and winding leaves in general, never coil round a support nor contribute to the fixing of a plant. The latter, therefore, appears to be merely an isolated application of the movement of winding in the ceconomy of nature, and we must confess that we are totally ignorant of its peculiar relation to the vital purposes of the in- dividual, or the propagation of the species. This movement is, at the sa^e time, the most frequent of all those detected in the outward organs of plants. The number of plants with winding stems alone, particularly if we include the winding flowering stem of the Monocotyledons and the winding fruit-stalks of the M. WICHURA ON THE WINDING OF LEAVES. 311 Mosses, is very considerable. Still larger is the number of the plants with winding leaves enumerated in §§ 55-125. And it must not be forgotten that, according to § 12?, there must be movements of rotation which escape notice on account of their slight degree ; as indeed, for example, the rotation of the indi- vidual leaflets which causes the contorted aestivation is so slight, that it would be wholly concealed were not its traces preserved by the regular overlapping of the margins. In the face of such considerations, the character of external fixedness, which it has hitherto been customary to connect with the conception of a plant, vanishes, and we rather arrive at the conviction that plants, too, possess, in the rotation of the axis, a motion peculiar to them, which reveals itself everywhere that the moving force has not been already exhausted in latitudinal growth, and there found a different expression proportionate to it. EXPLANATION OF PLATE IX. Figs. 1 & 2. Each of these figures represent horizontal sections of two penta- [)hyllous buds rolled to opposite sides. Fig. 2. is intended to render clear the direction followed in the rotation. If the line a 6 is to be removed, by a rotation round the point x, the point of section of the axis of the leaf, into the situation marked by the line c d, this takes place by a rotation which is continued to things standing outside the rotation, as the arrows indicate, in No. 1. from left to right, in No. 2. from right to left. The comparison of the two figures 1 & 2, shows us that the aestivation fig. 1. No. 1. is produced by the former movement, the aestivation fig. 1 . No. 2. by the latter. Therefore, in logical application of the Linnsean termi- nology of the aestivation, we call the arrangement fig. 1. No. 1. right-wound, that of fig. 1. No. 2. left-wound. Fig. 3. Fruit oi Ailanthns glandulosa, L., wound to the right at the summit. Fig. 4. No. 1. Fruit of Erodium Cicutarium, L., on an enlarged scale. The appendages of the carpels are wound to the left round the carpophore. No. 2. A carpel separated from the carpophore, also magnified ; the appendages wind to the right. Fig. 5. Leaf of an Avena, which is wound to the right below, and to the left above. Fig. 6. Branch of Chrysocoma Linosyris, L. Leaves wound to the right in agreement with the direction of the leaf-spiral. Fig. 7. Branch of Acacia micracantha, Desv. Phyllodia wound to the left in correspondence with the direction of the leaf-spiral. Fig. 8. A left-wound and at the same time strongly curved leaf of Chrysanthus angustifolia. Fig. 9. Branch of Dianthus trifasciatiis, W. & K., with left-wound leaves. [A. H.] 312 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. Article X. On the Development of the Ascidians. By A. Krohn. [From Miiller's Archiv, 1852-1853.] In setting forth the results of my observations on the develop- ment of the Ascidians, it is nowise my intention to lay before the reader a connected account of all the appearances which they present. Considering the earlier works in this department, I shall endeavour to exhibit only the most important steps, and in connexion therewith, as occasion may serve, to consider some- times one, sometimes another organ, whose mode of development may possess a special interest. I will premise here that the facts I am about to communicate all relate to the development of the Phallusia mammillata of Cuvier, and have been obtained by the employment of artificial fecundation *. By this method I have succeeded in following out their progress step by step for three months, until the young Ascidians had attained the size of a line or more. I commence with the unfecundated ovum, inasmuch as its structure is distinguished by many im- portant peculiarities. 1. Unfecundated ovum. — The fully formed ova which fill the oviduct consist, most externally, of a villous investment, within which is the proper vitellary membrane. Under the vitellary membrane there lies a glassy layer, in which roundish green structures of a peculiar kind are imbedded, and which surrounds the colourless yelk. The germinal vesicle and spot, both of which are still readily distinguishable in the ovarian ova, are already wanting f. The investing membrane is thin, and beset superficially by * The first successfully conducted attempts to fecundate the ova of the PhallusicB were made by Von Baer. t The same structure is exhibited by the unfecundated ova of the Phal- lusice in general, as well as by those of the genus ClaveUna. A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. 313 numerous short, blunt-pointed processes. Each villus consists of an aggregation of round, transparent, vesicles or cells, without nuclei. The glassy layer which lies upon the yelk is altogether homo- geneous, but is otherwise remarkable from the first-mentioned structures which are imbedded in it*. Each of these bodies consists of densely aggregated vesicles or cells. We find these structures sometimes solitary, sometimes united into groups of different forms and sizes, in the glassy layer. The latter is, however, nothing else than the primitive rudiment of the mantle, which is already present in the unfecundated ovum of the future Phallusice. The green structures, which persist un- changed during the whole of larval life, change, after metamor- phosis has taken place into the granules, which are contained in abundance in the mantle of the adult animal. This layer has already been recognized by Milne-Edwards (Obs. sur les Ascid. composees des Cotes de la Manche, p. 26. pi. 4. fig. 4) in the ova of Amouroucium proliferum, and has been quite justly interpreted to be the future test (Mantelschicht), as the following passage especially shows, p. 36 : — " La couche tegumentaire (i. e. the test) est dans le principe la couche gela- tineuse (glassy layer) qui dans Poeuf revet en dehors la masse vitelline.^^ According to this, however, Kolliker's view (An- nates d. Sc. Nat. 1846, t. v. p. 218), that the test in Amourou- cium Nordmanni and Aplidium gihhulosum^ arises only after the cleavage of the yelk, is confuted ; and I cannot doubt that it is this layer, which has been regarded Iby Van Beneden (Mem. sur FEmbryogenie, &c., des Ascidies, in the Mem. de PAcad. Roy. de Bruxelles, t. xx. p. 37) as the albumen of the ^^'gy although Milne-Edwards expressly draws attention to the pos- %sibility of falling into this error. 2. Yelk Cleavage and ^m^r^o.— Without doubt the ova, as iCuvier already supposed, and Von Baer has endeavoured to [demonstrate with more conclusive arguments, are fecundated in [the so-called cloaca, into which the seminal canal and the ovi- * The remarkable green or yellowish green colour of the ovarium and of |the oviduct, in many Phallusice, arises entirely from these green structures. pVheii the latter are colourless, as in Clavelina for example, the parts in question are without colour. 314 A. KROHN ON THE DEVELOPMENT OP THE ASOIDIANS. duct open, close by one another. Since, however, in the Phal- lusia, we never meet in this cavity with ova whose development has commenced, as is commonly the case in the compound Ascidians, perhaps nothing is more probable than that the ova are cast forth soon after fecundation, and that their development begins first, external to the parent. The cleavage of the yelk commences, about two or three hours after the semen has been brought into contact with the ova. It proceeds for the first stages, at least, in a very regular progression. I believe I have made out with tolerable certainty that every cleavage mass, after the completion of its division, is surrounded by an excessively delicate membrane. On the ad- dition of water impregnated with acetic acid, we see this invest- ment gradually raised up from the vitelline contents, which con- tract into a smaller space, as an independent membrane. As regards, however, the clear vesicular nuclei within the cleavage masses, I believe I have arrived at the result, that they dis- appear when a new division impends, and only after this has come to an end do they reappear again, newly formed. In their place there may be observed, in any yelk-mass about to divide, a very peculiar arrangement of the viteUine molecules. The latter are in fact disposed in thick striae, which radiating from the centre, are directed on all sides towards the lighter periphery of the mass, and seem to spread from two centres of irradiation. When, after the completion of division, the nuclei appear within the new cleavage masses, the radiate striation has also dis- appeared, and the yelk-granules are seen, lying close to one another in the masses, without any definite arrangement. All these appearances agree, however, with that later view of the cleavage process, which was first taken by Riechert (Mull. Archiv, 1846, p. 196). The glassy layer upon the yelk, which we have just described as the original rudiment of the test present in the unfecundated ovum, takes not the smallest share in these thorough changes of the yelk; it remains, without undergoing any alteration, closely apposed to the vitelline membrane. Before the expiration of the first four-and-twenty hours after fecundation, we find in most ova that the embryo has assumed its well-known cercarian form, with a more or less developed A. KROHN ON THE DEVELOPMENT OF THE ASOIDIANS. 315 tail. It is invested by the test which contains the, as yet wholly unchanged, green structures, and is separated from the vitellary membrane by a space filled with fluid. The substance of the body and of the tail consists of cells ; at least the latter are clearly distinguishable on the surface of these parts. The cells are polygonal, contain granules, and in addi- tion a central nucleus ; the axis of the tail is composed of larger rectangular cells, simply disposed in series, one behind the other, and also provided with a central nucleus, thereby attain- ing a transversely striated or articulated appearance *. There are various views as to the mode in which the tail is formed. According to Milne-Edwards the peripheral portion of the embryo, as it were in one piece, becomes constricted off from the body. KoUiker follows this opinion when he states that the tail grows out, not in the manner of a process, but, as a determinate portion of the blastema, becomes separate all at once in its whole length from the body. According to Van Beneden, on the other hand, the tail buds forth as a short process, which elongates by degrees. I can but confirm this last opinion. In fact, according to my observations, the tail appears in the younger embryos, very short and thick in rela- tion to the body, although it is already slightly curved. After- wards it becomes continually longer and more slender, embracing the body in a continually larger arc, until at last it has grown up to such an extent that its terminal portion enfolds the anterior part of the body of the embryo. A short time before the completion of the development of the larva, the tail undergoes remarkable changes. The axis becomes hollowed out, its whole cellular substance disappearing into a canal. This excavating process, which is accompanied by a contemporaneous liquefaction of the cell- contents, appears always to proceed from the two walls which are in contact with one another, of every pair of cells, and indeed in many places at the same time, extending further and further, until at last, by the flowing into one another of the separate cavities, the canal in question is formed through the whole length of the axis. In * This composition of the axis of the tail of large rectangular cells was first demonstrated by Kblliker in the embryos of Amouroucium Nordmanni and Aplidium {I. c. p. 22L fig. 43). 316 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. the meanwhile, however, the superficial layer of the axis of the tail, which is composed of smaller cells, appears to have become metamorphosed into a muscular layer consisting of longitudinal fibres, by whose energy those active movements of the tail ex- hibited by the young larvae after birth are performed. At present these movements are manifested only by the occasional twitches of the tail, which become more and more frequent in the last period of larval development. In consequence of these twitchings, the vitelline membrane eventually becomes torn, and thus, towards about the thirtieth hour after fecundation, the larva escapes. One point must not escape notice ; with regard to those two dark pigment spots which have been observed on the back of the larval embryos, but which have been, as it seems to me, too hastily taken for eyes. At first we see only a single pigment spot exactly in the middle line of the back. Behind this and more laterally a second, larger one, soon becomes visible (PI. XII. B. fig. 1. e). I have never succeeded in discovering any refractive medium in these pigment spots ; but I could, as little, satisfac- torily interpret the optical expression of the immediate neigh- bourhood of these spots, which I have endeavoured to repro- duce truly in the figure just cited. This much is certain, that both pigment spots, which in the larva remain always separate from one another (see fig. 2), approach quite closely in the course of metamorphosis and, during the development of the young Ascidian, continue to appear for a long time as an appa- rently single mass, lying close below the nervous ganglion. In the end, however, this mass breaks up into the two original portions, or even into many pieces, and so passes into the current of the blood, in which they may be seen driven up and down for a time, until at last they are wholly dissolved and dis- appear. This long continuance of the pigment spots, extending far beyond larval life, appears to me not to be in accordance with the function ascribed to them. For the present, therefore, their true import remains doubtful. 3. Larva (see fig. 2). — The body of the hatched larvae is elongated, has two slightly convex, lateral surfaces and is pro- vided at its anterior extremity with three very short processes, as it would seem excavated into suckers. Two of these lie A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. 317 higher and laterally opposite to one another ; the third, which springs from the middle of the anterior part of the body, is more inferior. By means of these prolongations, which are seen as conical processes even in the embryos (fig. 1, c, c), the larva attaches itself to some appropriate place in order to go through its metamorphoses. These processes, first seen by Milne- Edwards, and first justly interpreted by him, and whose ex- istence has been confirmed by KoUiker, have been, it would seem, quite overlooked by Van Beneden. With regard to the portion of the test which invests the tail, it may be stated here, that it terminates in a fin-like, expanded, probably horizontal appendage. The true composition of this appendage is, at first, the more difficult to perceive, as it can rarely be got under the microscope with its whole breadth at once. Therefore, it seems to me, as if the flabelliform process, into which, according to Van Beneden, the test of the tail in the larvae of Ascid, ampulloides (I, c. pi. 2) runs out, were a similar appendage, only more strongly developed, and seen from the edge. 4. Metamorphosis and Development, — The changes which the tail undergoes, immediately after the attachment of the larva, have been already described by Milne-Edwards in a manner, in the main, in accordance with nature. According to this observer, the contractile central portion or axis of the tail becomes gra- dually retracted from its investment and passes at last into the body of the larva, so that the investment remains behind, as an empty sheath, which subsequently falls off. Milne-Edwards' investigations give no account of the further fate of the retracted axis. According to what I have seen, the loosening and retraction of the axis of the tail, which, as we saw, sinks deep into the body of the larva, are only the forerunners of the retrogressive metamorphosis which it soon undergoes. Immediately after its retraction, we find the tail still in good preservation in the posterior division of the now enlarged body. Here it lies rolled up spirally into a coil, which may be pressed out of the body or unrolled by the use of carefully increased compression. Whilst the development of the young Ascidian commences, the coil breaks up first into many closely appressed lobes, and now 318 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. withers away by degrees in such a manner that the lobes become smaller and fewer, until they are but minute remnants. In the end these also disappear. When the coil begins to diminish, it occupies, as before, the whole posterior division of the body of the developing Ascidian ; but subsequently, when it has become visibly smaller, it lies more on the left side near the oesophagus (PL XII. B. fig. 3). From what has been said then, it results, that Van Beneden has not rightly comprehended the appearances during the extri- cation of the tail from its investment, since he considers this act, which for the rest his figures depict very well, to depend upon an absorption of the tail. Had Van Beneden recognized all the phases of the retrogressive metamorphosis of the tail, he could not long have remained in doubt as to the true nature of the undefined organ (/. c. pi. 3. figs. 11 & 12 e) met with in the posterior part of the body of Ascidia ampulloides during the first periods of development. He would have at once seen it to be the coiled-up and already lobulated tail. Very soon after the tail has been withdrawn into the body, and the latter has thence even to some extent increased in size, the three processes of attachment of the larva disappear, while in the meanwhile the mantle of the developing animal is fastened to the ground by its whole lower surface. From the mass of the body, and indeed from the midst of the abdominal surface, three processes of another nature now grow forth, and pene- trating deeper and deeper into the mantle, attain its surface. These hollow processes are the first indications of that dichoto- mously branched vascular system, with distinct walls, which pe- netrates the mantle of all Phallusice^. These processes are soon seen to become longer and to divide dichotomously into the first branches, whose extremities appear clavate; the dichotomous branching goes on further and further, and the terminal branches * According to Kollikev's more exact investigations (/. c), this vascular system consists everywhere of double vessels, which running close together ramify correspondingly, and accompany one another to the very finest ter- minal ramifications, when a mutual anastomosis takes place. According to my observations, this anastomosis occurs in such a manner that the terminal twigs of the two vessels bend round into one another like a loop. All these observations are also, as we shall see, perfectly confirmed by the study of deve- lopment. A. KROHN ON THE DEVELOPMENT OP THE ASCIDIANS. 319 are always found enlarged in the mode described. In young developing Ascidians which have fixed themselves normally, that is, by their whole under surface, we see subsequently, when the vessels of the test have branched out extensively, that all the branches and twigs extend perpendicularly and radially, on all sides, towards the circumference of the mantle. In those indivi- duals, on the other hand, which have been unable to find any, or no convenient, place for fixing themselves, while still larvae, but whose development nevertheless goes on*, the original vessels of the test, namely the three above-mentioned processes, as well as their subsequent ramifications, take the most various direc- tions, curving downwards for the most part. It would seem, therefore, as though a speedy termination were put to the further ramification by the above-mentioned unfavourable circum- stances ; such individuals probably die prematurely. It is, however, so much the easier to mistake the true nature of the vessels of the test, during the early stages, since even when the heart has appeared and the circulation has become established, we see not the slightest trace of any current of blood in them. In addition, the mantle is so transparent that its contours are easily overlooked, and the idea may readily arise, that the much more sharply defined vessels extend beyond its boundaries, whereas they are invested by it. It is therefore difficult, at first, to avoid considering the vessels of the test as stolons or runners, by the aid of which the young Ascidian seeks to fix itself more and more strongly to its residence. This view has the greater plau- sibility because msiny A.scidians{Cynthiapapillata for example) do, in fact, attach themselves to foreign bodies by means of branched stolons ; however, the establishment of the circulation in these vessels of the test at once releases us from this error. At first one sees nothing but a column of blood containing merely a few granules, irregularly oscillating. Afterwards, when the ramifi- cations have extended, the blood is more rich in granules and the circulation goes on more rapidly ; the oscillation is converted * A yet more striking example how little the course of development allows itself to be arrested, is presented by certain larva? which have been unable to burst their envelopes, but whose metamorphosis has nevertheless commenced, as we see by the coiled tail and the commencing formation of the vessels of the test. 320 A. KROHN ON THE DEVELOPMENT OP THE ASCIDIANS. into a regular current ; at the same time the, originally, quite simple main trunks and brancheshave become double, whilst their globularly enlarged, terminal twigs, directed towards the peri- phery of the mantle, are still simple tubes. In the larger vessels, which accompany one another and branch out in common, just as in the adult animal, the blood now streams in two opposite directions, in the one vessel towards the periphery, in the other towards the heart. Shortly before the division of the last branches into the terminal ramuscules, we see these two streams pass into one another by a loop ; on the other hand, in the terminal branches no continuous current of the blood is as yet to be ob- served; the blood-corpuscles indeed penetrate into them, but they frequently stagnate and sometimes become excessively accu- mulated. They are only occasionally seen to enter the current and return in one or other of the streams we have mentioned. All this continues until the doubling has extended as far as the terminal twigs. The cause of the two opposite currents is readily discovered, inasmuch as observation shows that each of the double vessels opens into opposite ends of the heart. Since, furthermore, the heart, even soon after its appearance, pulsates periodically in two opposite directions, at every change, we see the current of blood in each vessel reversed. Van Beneden has described processes very analogous to the rudiments of the vessels of the test, growing forth in the Ascidia ampulloides, soon after the retraction of the tail ; after a short existence, however, they disappear. If these last processes are identical with the vessels of the test of the developing Phallusiw^ the latter statement contradicts the observations communicated above. Ascidia ampulloides is very probably a Cynthia. The cartilaginous test, the plaited respiratory sac, and the double reproductive organs, arranged exactly as in the Cynthia, speak for this view. In the mantle of the CynthicR, so far as I know, no vessels have as yet been demonstrated. If they be absent, then the processes indicated by Van Beneden must have some other import ; if they be present, the statement that the pro- cesses soon disappear is probably erroneous*. * The budding vessels of the test will not easily be confounded with the puzzling diverticula of the test, which, Proteus-like, sometimes protrude, some- times are again retracted and disappear wholly, observed by Milne-Edwards A. KROHN ON THE DEVELOPMENT OP THE ASCIDIANS. 321 With respect to the first stages of development, I have not obtained all the information I desired, on account of the many difficulties in the way of a satisfactory examination. Never- theless I will not hesitate to communicate, in a condensed form, the observations which belong here, imperfect as they may be. It will be observed that as regards many points, they harmonize with the views of Van Beneden. Soon after the three hollow processes, the rudiments of the future vessels of the test, have appeared, we distinguish in the body, besides those parts which belonged to the larva, viz. the two pigment spots, now closely approximated, and the coiled tail, a cavity which represents the future respiratory or gill-sac. Immediately behind this appears the first indication of the ali- mentary canal, which appears in the form of a canal, everywhere of even diameter and bent round into the form of a loop. The whole posterior division of the body is filled by the coil of the tail. Subsequently one may observe on the dorsal surface, on the second layer of the body, which lies under the mantle and is clearly distinguished, three apertures, over which the mantle, still imperforate, passes. The one lies exactly in the middle at the anterior extremity of the body, the two others are alto- gether lateral and diametrically opposite to one another, in the middle portion of the body. The anterior, somewhat larger, corresponds with the future ingestive or respiratory aperture ; the two posterior are intended to take the place of the future cloacal aperture, which arises, at a much later period, by their coalescence. At this time the nervous ganglion also has become developed, and may be readily distinguished as an elongated structure in the middle of the dorsal surface and close above the two pigment spots. Near it we see the first indications of the future interwoven muscular cords of the body ; the abdo- minal furrow is already indicated; the nutritive canal has become further developed ; it lies now, for the most part, above the respiratory sac and describes a curve, commencing at the bottom of the latter and passing round to the immediate neigh- bourhood of the left excretory aperture. Three divisions may during the development of Amouroucium proUferumy and which he has fully described. SCIEN. MEM.— iV^a<. Hist. Vol. I. Part IV. 21 322 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. be distinguished in it, the oesophagus opening into the respi- ratory sac, the stomach, and the intestine. Soon after this there appear in the wall of the respiratory sac the first respiratory or gill-clefts {stigmates branchiaux, M. Edw.), in the form of four round apertures, provided with vibrating cilia. Their distribution is perfectly symmetrical, each pair lying upon the two opposite sides of the respiratory sac, one close behind the other and below their respective excretory apertures. Van Beneden (/. c, p. 44. pi. 3. fig. 11^.) has ob- served these apertures and discerned their true import. The heart appears to be formed last of all. It represents at first, a very short sac lying on the right side, near the stomach, or rather near the abdominal furrow, distinguishable by its un- dulating but, at present, very slow movement. It is distended with blood-corpuscles, which are driven alternately hither and thither. Before long the test gives way, over the three apertures of the body, whose margins have, in the meantime, become notched, and thus the young Ascidian enters into relation with the external world, since, from this time forth, it is enabled to take in nutritive matter and the water necessary for respiration. Before this, however, the second layer of the body on each side, in the entire neighbourhood occupied by the two apertures of the gills, has become detached from the respiratory sac, which elsewhere lies in close apposition with it, in the form of an externally convex roof. This roof therefore includes a cavity, which communicates on the one hand, by means of the respiratory apertures with the branchial sac, on the other by means of the corresponding ex- cretory apertures, which are seated upon the summit of the roof, with the exterior ; the terminal portion of the intestine, which has in the meanwhile become longer, curves at this time round the bottom of the respiratory sac upwards, and opens in the end, by means of the anus into the left space above mentioned*. In this way it becomes possible that the water drawn in through the anterior aperture of the body into the respiratory sac, should pass through the branchial apertures into the two spaces, and become emptied out again through the excretory apertures, whilst * We shall subsequently recur once more to these two spaces. A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. 323 the undigested remains of the food merely pass into the left space and are extruded through its aperture. We soon see clearly that the heart pulsates more swiftly, and in addition in alternately opposite directions, and that the blood, which already had begun to oscillate in various parts of the body, now follows a determinate, though very simple course ; this consists of a dorsal and an abdominal current, united by two transverse currents on each side, the one of which may be seen round the anterior aperture of the body, the other within the bridge formed by the wall of the respiratory sac between the two respective branchial apertures. As regards the test, it has already been stated that the green aggregated vesicles originally deposited in it become afterwards changed into its granules; this change commences only sub- sequently to metamorphosis, and consists in the green struc- tures gradually losing their cellularity and becoming smaller, less coloured, and more angular, and so by degrees assuming their ultimate form. It is only at a later period that those large, rounded, thin-walled cell-spaces appear, which, as is well known, occupy, closely appressed, the substance of the mantle in the full-grown animal (comp. Kolliker, /. c). At first their number is but small, but afterwards they multiply, increasing in size at the same time, so that in this w^ay the finer structure of the mantle becomes more and more perfect*. In accordance wdth my plan I will now treat only of those organs whose further development is either interesting in itself, or has not yet been sufficiently investigated, and upon whose structure the study of development is calculated to throw some light. We may consider, first, the respiratory sac, concerning the minute structure of which, in the adult animal, the following points may be called to mind. It is well known that its inner surface is divided into rectangular compartments by longitudinal and transverse ridges, which intersect one another at right angles. It is, besides, beset with numerous ciliated papillae, one of which * It is remarkable that the same metamorphosis of the green structures occurs also in the test investing the tail, which has been left behind as an empty sheath. 21* 324 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. is placed at every crossing point of the ridges ; the bottom of each compartment is pierced by from four to six narrow longi- tudinal clefts, the branchial apertures {stigmates brancMaux, Milne- Edwards), around whose edges is a fringe of numerous vibratile cilia. The bridges between these clefts are hollow, and receive the blood requiring aeration, brought to them on one side by the ridges, which are also hollowed, and carried away, on the other, after it has undergone the necessary change, to be re- distributed by other ridges. The play of the cilia round the edges of the clefts drives the fresh water, continually drawn into the branchial sac, past the bridges. As we know, however, the apertures lead into a large internal space {chambre thoracique et cloaque, Milne-Edwards) between the respiratory sac and the second layer of the body, which opens externally by the excretory or so-called anal siphon. Of all those parts which contribute to the complicated orga- nization of the respiratory sac, the most important, that is, the gill-clefts and the intervening bridges, are formed first ; for we have seen the apertures appearing in the very earliest stages of development, as two rounded openings on each side. It was stated also that a current of blood was already established in the bridges. For a long time the respiratory sac possesses only these two pairs of apertures, which in the meantime elongate and be- come more and more cleft-like. At last there appear two new apertures on each side in the bridge between them, and soon after- wards another is added behind the originally posterior cleft, so that there are now altogether five openings. In this manner, successively and in serial order, a multitude of openings soon rapidly appear and pierce the two sides of the respiratory sac. At first, a second series is formed close to and above the first, then above this, a third and so more and more series, until the middle line of the back is reached. In similar succession rows of aper- tures appear, one below the other, between the first series and the abdominal surface. Like the original openings, all these ap- pear, originally, as small gaps beset with vibratile cilia, and during their gradual enlargement undergo similar alterations in form, becoming more and more lengthened into clefts. It is easily comprehended that with the progressive increase of the respira- tory sac, the number of the clefts also subsequently increases in A. KROHN ON THE DEVELOPMENT OF THE A8CIDIANS. 325 each series, and that new series continually appear between the old ones. Soon after the respiratory sac has become thus perforated in its whole extent, we find the clefts, which previously had their long diameters disposed transversely to the axis of the branchial sac, arranged as in the adult animal with their long axes longi- tudinal. Many of the papillae, covered with vibrating cilia, have at the same time become developed, upon the inner surface of the transverse and longitudinal bridges, between the clefts. The blood streams rapidly and abundantly through all the bridges. I have been unable to trace the development of the respiratory sac any further. In its present, by no means complete, struc- ture, however, it singularly resembles the respiratory sac of the compound Ascidians. In this case we have again an example of the often enunciated and repeatedly confirmed law, that the higher forms of any given order of animals, at least as respects particular organs, exhibit transitorily, relations of form and struc- ture which occur permanently, that is, in the adult state of the lower forms. It will be recollected, that the three apertures of the body which were at first covered by the test, had opened externally after the appearance of the first gill-clefts and of the heart, in consequence of which the young Ascidian was enabled to take in nutritive matters and the water required for respiration. These apertures now become prolonged into short projecting tubes, the siphons. The aperture of the respiratory siphon, which from the beginning surpasses the two posterior or excre- tory siphons in circumference, soon appears fringed by eight lobes, as in the adult animal. In all the siphons the circular fibres, by whose action they are occasionally closed, may now be very readily distinguished under the test. Externally, above the circular bundles, there appear also a few longitudinally fibrous bundles, processes of the muscular bundles which, already inter- laced, are disposed over the second layer of the body. The ultimate fate of the two posterior siphons is particularly interesting, since, after persisting for a long time, they fuse together into the single siphon present in the adult. In fact, during the period in which the rapid multiplication of the respiratory apertures in the branchial sac is taking place, we see 326 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. the two siphons gradually approximate towards the middle line of the dorsal surface and at last come so close, that they are separated only by a small bridge. When the respiratory sac is completely perforated, this also disappears. Instead of the pre^ vious two excretory siphons, we now find a single one situated exactly in the middle of the dorsal surface, close behind the nervous ganglion ; its lip is already, as in the adult animal, pro- vided with six lobes. The presence of two excretory siphons, at an earlier period, need not be at all surprising, if we recollect what has been pre- viously stated respecting the two spaces which were formed over the original gill-clefts ; it appears, in fact, to be almost a neces- sary consequence. It was shown that the two spaces are deve- loped by the raising up of the second layer of the body, in the region of the apertures, from the respiratory sac, which is else- where closely applied to it. This mutual separation goes on to a greater and greater extent, during the further develop- ment of new clefts in the respiratory sac. The consequence is, that the boundaries of the two spaces extend further and further, until at last, w^hen the branchial sac is everywhere perforated, they coalesce upon the dorsal side, and in this manner consti- tute that inner space, between the branchial sac and the second layer of the body, to which we have already referred in the adult animal. When this has taken place there is no longer any need of more than one siphon, and, in fact, the formation of the single siphon is contemporaneous with that of the space in question. I pass now to an organ, which in all the Phallusiae surrounds the whole nutritive canal, from the mouth to the anus, and has the appearance of a compact honey-yellow mass, dotted over with chalky-white points. By the majority of zoologists, among whom w^e find a great authority (V. Siebold, Vergleichende Anatomie, p. 269), this structure has been regarded as a liver ; a view which, to me, appears so much the more doubtful, as an organ which will subsequently be described, and which has hitherto escaped notice, probably has a greater right to this title. The organ now in question consists of obvious, clear, round, tolerably dense walled vesicles, the supposed simple glandular utricles in which the bile is prepared. Every vesicle is apparently distended by a transparent fluid. A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. 327 in the centre of which a solid concretion, in the form of a nucleus, may be remarked. In the larger vesicles the white chalky nucleus appears to be composed of two other rounded segments in close apposition, in the smaller it is perfectly spherical*. According to my observations the vesicles are all completely unconnected with one another; sometimes I thought I could make out a finely reticulated network upon the wall of some of them, without being able to come to any clear conclusion upon the matter. It is the more difficult to decide upon the purpose of the whole structure, as its mode of development, as we shall see immediately, does not throw the desired light upon it. One is tempted to consider it as a purifying organ, a kidney, which would agree very well with the deposits in the vesicles f. In that case, however, we ought to be able to discover excretory ducts, of which I could find not the least trace. The first indication of this organ appears when the residue of the tail of the larva, reduced to a very few lobes, is rapidly dis- appearing. Close to these lobes, we see first a small, round, transparent vesicle, already precisely similar in structure to those of the adult organ. The round chalk-white nucleus is already visible in its centre. The vesicle now grows more and more, and at last, when the residue of the tail disappears, it takes its place close to the oesophagus on the left side. Subsequently a second vesicle is formed close to the first, after this a third, and so more and more, until the whole posterior space of the body between the oesophagus, the stomach, and the intestine, is filled up by a mass of such isolated vesicles, in every stage of develop- ment. Afterwards vesicles are seen arising singly and remote from the great mass near certain parts of the alimentary canal. In a few vesicles the nucleus appears double, in others it is pro- vided with the previously mentioned mass of laminated crystals. Whether the vesicles arise at the same time with the nuclei, as we have here described it, or whether the nuclei, as would seem * In Phallusia monachus the round nuclei exhibit a close concentric stria- tion, which seems to indicate a deposition in layers. In many vesicles a mass of laminated or acicular crystals lies upon the nucleus ; sometimes one vesicle or another contains, instead of the nucleus, a great prismatic crystal with pyramidally pointed ends. t Delle Chiaje {Animali invertehrati d. Sicilia citeriore, t. iii.), to whom these concretions were not unknown, appears to hint at this interpretation. 328 A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. to be theoretically more probable, are deposited from the fluid of the vesicles at a subsequent period, must be left undecided. I have but rarely seen the nucleus wanting at the first appear- ance of the vesicles. The organ could be traced no further in its development, which is, however, clear enough from what has just been stated. Finally, it only remains that I should describe an organ as yet unknown, which is so hidden, as only to be partially visible in the full-grown animal. It consists of a system of fine canals, which are distributed over the whole intestine, commencing in caecal extremities, which are partly cylindrical, partly, and for the most part, enlarged or clavate, and often locally sacculated ; these enter into manifold anastomoses with one another, and thus enclose the wall of the intestine with a close network ; at last they unite into twigs and branches. The canals are found, most abundantly, within the strong projection which runs along the inner surface of the intestine from the stomach almost to the anus, and which is besides filled with the vesicles of the organ which has been already described. The ramification of the branches and twigs, which frequently, during their course, per- fonn arched curvatures somewhat in the manner of the vasa vorticosa of the eye, and in some places dilate into ampullae, is dichotomous ; the contents of the caeca and canals are pellucid. So much for the relations of this organ in the adult animal ; development affords us the following further information. Very early, even before the appearance of the first gill-clefts and of the heart, there appears at the beginning of the intestine immediately behind the stomach, a cylindrical process, alto- gether homogeneous, and with a somewhat enlarged, clavate, free extremity, which passes transversely, on the left side, to the vici- nity of the terminal portion of the intestine. Without percep- tible alteration, this process continues slowly to grow, until at last we see it divide at its extremity into many branches, which pass to and embrace the intestine. Very soon these branches also divide, dichotomously, into twigs, which by their frequent anastomosis form a network enclo- sing the intestine. In this manner the ramification proceeds, the network upon the intestine becomes closer, and at last the caeca above referred to make their appearance upon it. The branches A. KROHN ON THE DEVELOPMENT OF THE ASCIDIANS. 329 now exhibit their characteristically wavy course, and in some places the ampulliform expansions. It is worthy of being re- marked further, that at this time, when the organ, in fact, is but sketched out in its first rudiments, the main trunk of the canals, i. e, the original process, as well as its branches, are still homo- geneous solid cords, whilst in the caeca and the twigs which have anastomosed into networks, the wall and the cavity are already distinguishable. It results hence, that the organ which consists of branched canals opens by means of an excretory duct into the intestine, as an appendage of which it at first appears. Its whole struc- ture would lead one to consider it to be a gland, whose secretion, prepared in the caeca, would seem, if we may judge by the place of insertion of the duct into the gland, to be accessory to diges- tion. Whether the watery secretion is bile and the gland there- fore is a liver, must for the present be left undecided. EXPLANATION OF PLATE XII. B. Fig. 1 . Embryo in a late stage. Two pigment spots already exist, a, body ; b, commencement of the tail ; c, c, the two upper processes of attach- ment of the larva commencing their development; d, anterior pigment spot ; e, posterior pigment spot. Fig. 2. Larva lying upon its side, a, a, test with the imbedded and as yet un- changed green structures ; &, b, axis of the tail ; c, c, hollow canal of the axis; d, the fin-like horizontal appendage into which the test covering the tail becomes expanded ; e, anterior pigment spot ; f, posterior pigment spot; g, upper right process of attachment; h, lower process of attachment. Fig. 3. A developing Phallusia of the stage when the three apertures of the body have become already elongated into very short siphons, a, re- spiratory siphon wide open ; b, b, the two posterior or excretory siphons in their contracted state ; c, nervous ganglion with four nervous trunks ; d, oesophagus ; e, stomach ; /, intestine (its terminal portion curves round the bottom of the respiratory sac upwards towards the left excretory siphon) ; g, coil of the larval tail breaking up into lobes and disappearing ; h, h, A, A, the four first gill- apertures with their intermediate bridges; i, the dark, apparently single mass of pigment under the ganglion ; kj abdominal furrow ; Z, /, test. [J. H. H.] 330 KOBEN AND DANIELSSEN ON THE Article XI. Observations on the Development of the Pectinibranchiata. By MM. KoREN and Danielssen. [From the Annates des Sciences Naturelles, t. xviii. No. 5, and t. xix.] ^ \. On the Development of Buccinum undatum, Linn. Although, since we first made known our investigations into the development of the MoUusca, in the Magazin for Naturvidenskaberne, many remarkable memoirs have appeared upon the same subject (those of Nordmann*, of Vogtf? of Qua- trefages J, of Loven§, of Reid ||, and of Leydig^f), yet we think it advisable to set forth our observations here in full, and we trust that they will not be read without interest — particularly as we have been able, in two genera of Gasteropods, to follow step by step the development of the embryo from its first appearance to its complete formation, and as we shall have to bring forward facts new to science. On the 1st of March, 1851, we procured many oviparous capsules of Buccinum undatum. It is well known that these capsules, spherical or ovoid in form, are most usually found united into racemes nearly as large as the fist. They are attached to various bodies, such as rocks, bits of old wood, marine plants, &c. In many of these racemes, the capsules contained incom- plete young, in others ova. We possessed then all the materials required for our researches. To judge from the internal condition * Versuch einer Monographic von Tergipes Edwardsii. f Recherches sur VEmhryogenie de V Action {Annates des Sciences Naturelles, 3rd ser. t. vi. 1846, p. 1). X Arenales des Sciences Naturelles, 3rd s6r, t. iv. p. 33. § Bitrag til Koennedomen om Utwecklingen af Mollusca Acephala lamelli- hranchiata (Kongl. Vet. AJcad. Handl. 1848). II Ueber die Entwickelung der Eier der Mollusca nudihranchiata (Froriep's Jahresherichte, Januar 1850). ^ Ueher Paludina vivipara {Zeitschrift fur wissenschaftliche Zoologie. Leipzig, 1850). DEVELOPMENT OF THE PECTINIBRANCHIATA. 331 of the ova and from other evidence which we have obtained, it would appear that Buccinum undatum lays its eggs from the beginning of January to the end of April. Many oviparous capsules were opened, and their contents subjected to microscopical examination. Each capsule enclosed a limpid pellucid viscous liquid, resembHng white of egg ; in this viscous humour and at the bottom of the capsule there lay a large number of ova (6-800), which could with difficulty be detached from one another ; each egg was composed of a delicate transparent membrane (chorion), in the interior of which was another, still more delicate, enveloping the yelk. The latter, spherical in shape, was composed of a viscous fluid, in which was dispersed a multitude of rounded granules, variable in size and of a clear yellow colour (PI. X. fig. 2). No trace of either germinal vesicle or germinal spot was visible ; the diameter of the ovum varied between 0*257 and 0*264 of a millimetre. On the 8th, 13th, 15 th and 20th of March we re-examined capsules of all the racemes ; the ova had remained, to our great astonishment, spherical and without division*. On the 24th of March the ova were still undivided, but in- stead of being scattered as before, they had approximated to one another. The chorion had begun to disappear ; the larger por- tion of the yelk was diffused {s^etait epanchee), and invested by the viscous albuminous fluid which surrounded the viteUine membrane. Some days afterwards, the ova had become agglo- merated into a single mass, which at its surface was divided into many portions, in which each ovum could be distinguished by the naked eye. In general the number of the ova constituting each of these portions varied from six to sixteen. On the 29th of March again, we examined many capsules ; the isolated groups were more sharply defined and each mass had become ovoid or reniform ; it was observable also that these groups were united together. On the 1st of April we examined many capsules ; one of them enclosed twelve embryos ; they were ovoid or reniform, and provided with two rounded lobes (vela) and a foot (figs. 3. 4 c & 5 e.). The fluid contained in this capsule was as clear as * We have at times, in a few ova, seen a small conical eminence upon the vitelhis. 332 KOREN AND DANIELSSEN ON THE water and perfectly fluid, so that the embryos could easily be extracted from it. In another capsule there were but six em- bryos, four of which were well formed. We began then to form a conception of the import of all these facts ; but they seemed to us to be so extraordinary, that for a long time we did not venture to put confidence in our observa- tions— so different were these results from all that had hitherto been seen with respect to the development of Mollusks, and even from all known physiological phaenomena. But the attentive study of what went on under our eyes at last cleared away all our doubts, and led us to dispute a law generally admitted and based upon a great number of facts. In fact, we had before us a mode of development which required for its proper under- standing much new investigation ; and it appeared to us to be indispensable to follow out these singular phaenomena in neighbouring genera, in order to see if, towards the limits of this zoological group, we should find them disappearing ; an anticipa- tion which, as we shall see by and by, has been fully justified. But let us first trace out these phaenomena in the Buccinum, and we shall find them surprising enough ; for we shall look in vain in a fecundated ovum, for any of those changes which in ordinary cases take place during its development, in virtue of the well-known law of yelk-division. We see no grooving — no cell appears ; in a word, the interior of the ovum remains without alteration ; on the other hand, external to it we observe active alteration. The excessively viscous albuminous fluid collects and, so to speak, glues together the ova, which were primarily separate, uniting them into bunches ; subsequently the mass in which they were immersed, from being viscous, be- comes as liquid as water ; and it is then that we first perceive traces of activity in the ovum itself. Its external membrane bursts in certain places — the yelk becomes diffused, and a mem- brane, destined to circumscribe the developing individual, is produced around each bunch of ova; between these bunches we see many isolated ova, which seem excluded from this organic process which gives rise to embryos. We do not know if these isolated ova die at once, or undergo a further imperfect develop- ment ; but in any case they are beings, whose existence is but ephemeral. As soon as the grouped ova are invested by their DEVELOPMENT OF THE PECTINIBRANCHIATA. 333 membrane, the formation of the embryo begins ; a very clear, finely granular and viscous liquid is deposited at first on the external surface of the ovum. In this plastic mass we see, in some localities, cells developed, in others muscular tubules, ac- cording to the nature of the organ which is to be produced ; thus, the first signs of activity visible about the ova, after they are laid, are observed in the very viscous albuminous mucus which surrounds them, and this activity is manifested by the mere union of a certain number of ova into more intimate rela- tions. We imagined at first, that these phaenomena of agglo- meration might represent yelk-division ; but we soon gave up this idea, which seems to us altogether incorrect*. Besides, the development o^ Purpura lapillus, which we have subsequently examined, and in which division and the phaenomena of agglo- meration go on at the same time, has also contributed to our rejection of this notion. We come therefore to the conclu- sion, that yelk-division is not always indispensable for the production of the embryo ; but the fact of the agglomeration of some fifty or more perfectly formed ova, into a single individual, is certainly strange enough. Where does the formative principle exist? Is it enclosed in an isolated ovum? Or is it extended over the whole mass ? and is it the common power which be- comes then the organizing force of the substance ? We have seen the isolated ova undergoing a certain amount of develop- ment, but the being which results from them is very incomplete and becomes very rapidly destroyed. It would seem to want the materials requisite for its permanent existence. We shall return to this subject in treating of the development of Purpura lapillus, and we shall confine ourselves here to adding, that this ♦ In publishing here the observations of MM. Koren and Danielssen, which are highly interesting in themselves, apart from the interpretation which may be put upon them, we consider ourselves bound to remark, that the rounded masses of vitelline matter, regarded by these authors to be simple ova, seem to us to be merely vitelline spheres, whose utriculiform envelope presents a little more consistence than ordinarily, and that therefore the aggregation from whence the body of the embryo arises, is the result of the grouping together of the vitelline spheres of a single ovum, and not the product of the union of many primitively distinct ova. What these authors remark with respect to the phaeno- mena of yelk-division, which are also sometimes observed in these spheres, would in nowise oppose this interpretation.— (iVo/? by M. Milne-Edwards.) 334 KOREN AND DANIELSSEN ON THE mode of development appears to us to be of great physiological importance and will assuredly be confirmed by further obser- vations. With regard to the number of ova which are grouped together for the formation of a single embryo, there is consider- able variation, and it seemed to us to be the greater, the greater the number of the ova in the same capsule. In the mean, each capsule contains from six to sixteen embryos ; however, we have found as many as thirty- six. The more ova there are in a capsule, the more of them there are in each mass, whence we sometimes meet with embryos which have at their first appearance a very considerable size, — attaining in fact as much as 1| millimetre. The number of ova united into one embryo varies from 40 to 60 ; we have often, however, met with as many as about 130. Gray long since observed*, that a capsule enclosing more than a hundred ova yielded only four or five embryos. The English physiologist explains this fact by the law of atrophy, that is to say, by supposing that the great increase of some ova hinders the development and ultimately effects the destruction of the others. It remains to be seen whether Gray deceived himself, or whether it is we who have fallen into error. We believe that he was on the right track, but did not attain to the exact truth. After having seen that the ova group themselves together, to form the embryo, and become coated by a delicate and trans- parent pellicle, we now proceed to explain the manner in which the various organs make their appearance. The first step is the exudation of a clear, finely granular substance over the surface of the ova, which then begin to appear more trans- parent. We perceive in this mass, a multitude of cells which continually multiply ; it then takes on a distinct form and be- comes bilobed (fig. 4 c). These lobes are gradually provided with cilia, and the first movements are exhibited. The foot, which is formed in a corresponding manner, appears as an eminence, distinct from the remainder of the body, and pre- senting cilia ; the embryo then turns upon itself in an extremely slow manner (fig. 5e). Scattered cirrhif now make their ap- * Annates des Sciences Naturelles, 2 S^r. torn. vii. p. 375. f Sars drew a distinction between the long cirrhi which are found on tlie lobes and the cilia, and proposes to call them natatory hairs (Svoemmehoar). Subsequently many authors have called them cirrhi, retaining the name of cilia for the very short, fine hairs. DEVELOPMENT OF THE PECTINIBRANCHIATA. 335 pearance here and there upon the upper edges of the rounded lobes, and in a short time spread over their whole surface (figs.5 &, 6) ; the lobes are then rounded and provided with cilia, as well as with cirrhi (fig. 15). After the formation of the lobes and of the foot, we see a semi-transparent granular substance between the grouped ova and the membrane which invests them, exuding and becoming the rudiment of the mantle, covering itself superficially with a membrane which occupies a greater or less space and takes on a determinate structure (fig. 4). Sub- sequently we observe, on the lowest part of the mantle, a hemi- spherical transparent body, which is the rudiment of the shell (figs. 5, 6, 7«)» The foot increases in size and takes on a more rounded form, and at its base we distinguish early the two audi- tory organs (figs. 7? 8, 9^) . They are constituted by two spherical vesicles, pellucid as water, filled with a perfectly transparent and colourless liquid, and early exhibiting a double contour. Each vesicle encloses but a single otolith. When the animal is placed under the compressorium these organs become very di- stinct, and by increasing the pressure, each otolith readily be- comes separated into four regular segments. The vibratory movement which most authors have observed in these otoliths, has not been observed by us in those of Buccinum undatum, neither have we been able to distinguish any cilia upon the internal surface of the vesicle, although we have employed the strongest magnifying powers. The eyes probably appear at the same time as the auditory organs, for we have never seen the latter without being able to observe the former. Leydig has remarked, that the rudimentary eye is a vesicle which exists at the root of the tentacles. We have confirmed the exactness of this observation; the internal sur- face of this vesicle is always provided with cilia (fig. 10 a). It is filled with a liquid, enclosing a quantity of pigment- granules, of a strong, clear yellow colour, which are invested by an excessively fine pellicle (fig. 10 b). At this period the vibrating cilia com- municate a rotatory movement to these granules. We have been unable to distinguish any lens at this period ; it is only subsequently developed. At the same epoch as that in which the eyes appear, we also found the two conical tentacles and the rudiments of the salivary glands ; the latter always make 336 KOREN AND DANIELSSEN ON THE their appearance as two pyriform organs, composed of rounded cells (figs. 8 & 9 A) ; their lower extremity is the thicker, and their centre is filled by a mass of deeply coloured pigment-granules. The keel- shaped heart appears at the same time, as well as the mouth and the rudiments of the proboscis. Sars, Loven, Nord- mann and Vogt have been unable to discover any heart in the first periods of development, and we have ourselves sought it in vain in many Nudibranchiate genera ; we are therefore led to conclude that the first period of development passes by without the existence of this organ. It is not the same with the Pectinibranchiata, at least with Buccinum undatum and Purpura lapillus. In these the heart has already appeared in the interval between the twenty-third and the twenty-eighth day. Grant* was the first to observe the heart in Buccinum undatum, and to draw attention to its strong pulsations. To the same observer we owe the discovery of the two rounded organs provided with vibratile cirrhi upon the sides of the head, by the aid of which the embryos move about in Purpura, Trochus, Nerita, Doris, and Eolis. Subsequently, Lund, Sars, Loven, Nordmann and other naturalists, have confirmed the exact- ness of these observations. The heart lies upon the dorsal side, a little to the left, is a little twisted and completely un- covered at its first appearance ; for different organs which cover it afterwards, such as the roof of the branchial cavity, are still rudimentary. In the place where the heart is about to appear, a grayish, transparent, finely granular mass, having a rounded form, arises ; it is attached to the lobes and the foot, and pre- sents no sensible movement. Soon, slight contractions are per- ceptible ; the form becomes more and more marked, and at last it resembles a great clear vesicle. Two or three muscular tubes, of extreme delicacy, are now traceable upon its parietes and effect its contractions. In this primitive state, the heart possesses the form which it will subsequently retain (fig. 6 e). In propor- tion as its parietes increase in thickness and magnitude, the muscular tubes are multiplied, transverse tubes appearing and be- coming filled with a transparent and colourless liquid (figs. 7? 8 & 9 e) ; we have often counted the pulsations, and have found that their rate varies ; on the average, there were some forty to * Edinburgh Philosophical Journal, 1827» vol.vii. p. 121. DEVELOPMENT OF THE PECTINIBRANCHIATA. 337 fifty a minute ; they are not always regular and frequently change their character; strong pulsations occur among the weaker pulsations. It frequently happens that the heart sud- denly ceases to contract, remaining at rest for some time ; and it is nowise uncommon to observe strong pulsations succeeding intervals of rest. The cavity of the primitive heart is cylin- drical, and is divided by a single partition. Its parietes are excessively delicate, very transparent, and refract light in an altogether different manner from the other organs (fig. Se) ; we have never met with any liquid nor with any cellular structure in them. We have also observed similar muscular tubes in the two rounded lobes*; but here we see that many are super- imposed and frequently ramify. This branching becomes more and more abundant as we approach the periphery of the lobes, where the finer ramifications frequently interlace and a muscular network results (fig. 15 a, i) which serves to move the two rounded lobes in all directions. It may be, that a vascular network exists in some parts of this muscular reticulation, for little gra- nules, refracting the light strongly, are to be met with dispersed through the mass. Since the two distingiushed French natu- ralists, MM. Milne-Edwards and Valenciennes, have pointed out the greater or less incompleteness of the vascular system of Mollusks, it has appeared to us to be important to observe the condition of the circulation in the young Buccinum, but not- withstanding all our efforts, we have not been sufficiently for- tunate to detect the slightest trace of a circulating current f. We have already stated that the proboscis is one of the organs that appears early ; it is recognizable by its cylindrical form and * Leydig, /. c, describes the structure of the muscles of a great number of Gasteropoda, and remarks that the result of his researches upon this subject does not altogether agree with that at which MM. Lebert and Robin have arrived. We believe that those of Leydig are the more exact in all essential points, inasmuch as they agree better with our own observations. f We cannot help remarking that there is something rery anomalous about this supposed " heart," and we would venture to suggest that its structure and position indicate rather, that it is no heart at all, but a *' contractile vesicle,** similar to that which exists in the corresponding locality in the pulmonate Gasteropoda, and which, acting in concert with the caudal vesicle, perfornvs the functions of a heart before the true heart makes its appearance. See Gegenbaur, Beitrdge zur Entwickelungsgeschichte der Land Gasteropoden^ Siebold und Koiliker's Zeitschrift, 1852. [T. H. H.] SCIEN. UEU.—Nai. Hist. Vol. I. Part IV. 22 338 ROREN AND DANIELSSEN ON THE by the strong muscular contractions which it presents. A little later, the stomach and the oesophagus appear ; the latter pre- sents itself under the form of a hollow cylinder attached to the proboscis, and presenting upon its parietes, which are of excess- ive thinness, a certain number of clear striae, the rudiments of the future muscular tubes. As soon as the oesophagus becomes free from the proboscis, it bends a little backwards and upwards, running along the lower surface of this organ for a short distance ; then, having undergone a flexion in the opposite direction, it directs itself a little to the left to terminate in the slightly elon- gated portion of the stomach (fig. 12 n). It is extremely difficult to follow the early phases of the development of the oesophagus ; for it is surrounded by the proboscis, whose parietes are both thicker and more opake than its own ; in some places it is en- tirely covered by it. It is for this reason, that we have been un- able to determine if the oesophagus is formed, at once, in all its length, or if a portion appears at first, and elongates until it be- comes united with the stomach. The latter organ is found a little to the left, and, at first, is almost spherical. It appears to arise thus : a grayish semitransparent mass exudes from a single yelk, soon becoming invested by a delicate pellicle, which completely encloses this same yelk (fig. 9 m). This membrane at first stretches upwards to become united with the oesophagus, and afterwards downwards to give rise to the intestine, which turns towards the right side of the body (fig. 11 o). It is by reason of this origin that we see the stomach always filled, so to speak, with yelk-granules, which are constantly agitated by the vibrating cilia with which the parietes of the stomach are pro- vided. These cilia are found, not only on the internal surface of the latter organ, but W'e have also observed them in the oesophagus and in some parts of the intestine. As we have been unable to follow the intestine for more than a small portion of its course (fig. ] 1 e), we have not made out the anus. It is at this period that the first outlines of the nervous system are distinguishable: they are tv\'o compact, ovoid and yellow bodies (cerebroid ganglia), which surround the oesophagus ; and about the same time, the rudiments of the two pedal ganglia, placed side by side, having a strong yellow hue and more or less ovoid in form, make their appearance. DEVELOPMENT OF THE PECTINIBRANCIIIATA. 339 After the edges of the mantle have developed themselves into a covering over the back of the animal, a cavity clothed with fine cilia is formed, in which the heart and branchiae lie. The first tract s of branchiae which we observed were two indistinctly marked bands arising from the edge of the mantle, and diverging at intervals to unite again and so form loops. We have found, at a more advanced stage, that these bands are tubes taking a very sinuous course, whence they have a certain resemblance to a corkscrew. These sinuosities were better marked above and be- low than in the middle, where they were wider and more closely appressed; an active ciliary motion was distinguishable upon their internal edge (11 j»). Loven has shown, in his excellent memoir (op, citJ), that there is an extreme resemblance, in respect of development, between the Gasteropoda and Acephala ; he has described the mode of formation of the branchiae in the latter, and we have observed even that the production of these organs takes place in a similar manner in Buccinum undatum and Fur- pura lapillus. About the period of the formation of the branchiae we ob- served, below, at their union, a vesicle which is formed and de- veloped like the heart. It is ovoid or slightly pyriform, ends below in a very long tube which follows the course of the intes- tinal canal, and eventually, like the latter, becomes lost in the obscure mass of the yelk (11, 12 q). Its walls are thin, semi- transparent, and provided with a multitude of varicose muscular tubes which take a direction both longitudinal and transverse to the vesicle. These tubes have smaller dimensions than those of the heart, and for that reason it is necessary, in order to distin- guish them readily with the microscope, to employ a much higher power than that required to examine the muscles of the heart. The contractions of the bladder are strong, and take place from above downwards, whilst the heart contracts from one side to the other. When this vesicle dilates it becomes filled with a clear liquid in which a great number of obscure molecules are distinguish- able. We can regard this organ as nothing but the kidneys. The period for the development of new organs has now gone by, but all those already formed, become gradually perfected. 22* 340 KOREN AND DANIELSSEN ON THE The head and the back of the animal are now distinct, and are provided with fine ciHa and with much elongated tentacles, re- sembling cilia. The eyes are more conical, and their lenses are distinctly perceptible. The buccal aperture appears in the form of a transverse cleft. The proboscis is perfectly developed. The tongue appears in it with its armature such as Lebert and Loven have described. The salivary glands are very voluminous, and their excretory canals may be readily traced passing along the sides of the oesophagus. The siphon is already much developed, and provided with cilia (fig. 13 ^). The form of the foot is modi- fied, being considerably elongated; in addition, two rounded lobes have arisen in its upper part; its surface is ciliated all over (11,12/). As regards its structure, the foot is composed of a multitude of primitive muscular tubules — some cylindrical, others varicose — running in all directions, without however uniting into bundles. We have never been able to detect either granules or cells in these tubes. At this period the nervous system has become very distinct. We see that the two ganglia placed at the sides of the oesophagus (cerebroid ganglia, 11 ^,12/?, 14/) are united by a commissure; from each of these ganglia a very thick commissure passes (12^, 14 i) to unite them to the pedal ganglia (12r, 14 A:), which are ovoid, with their small extremities turned towards the cerebroid ganglia, and furnish a great number of branches for the foot, (fig. 14 /.) We see besides, in the lobes of the foot, two smaller, also ovoid ganglia (125, 14m) which send branches to the lobes. Two commis- sures (14 w), unite them tothe pedal ganglia: the cerebroid ganglia distribute a nervous twig to each eye and to each auditory or- gan {I4g8ih). We have observed that one of the pedal ganglia sends a nerve to the intestinal mass (fig. 14: p). The description given by Cuvier of the nervous system of this mollusk* differs in many points from what our observations have shown us. It is certain, that the nervous mass which Cuvier has called the cerebrum is a pedal ganglion ; for we have seen, in the adult animal, that the true cerebroid ganglion which sur- rounds the oesophagus, and which, without doubt, escaped his notice, is situated above this latter ganglion. ♦ Memoires pour servir a IHistoire de VAnatomie des Mollusques. Paris, 817. DEVELOPMENT OF THE PECTINIBRANCHIATA. 341 The shell, which is membranous during the first periods of the development of the embryo, excessively delicate and ovate or reniform, subsequently takes on the form of that of the Nau- tilus (fig. 1 1 a), but subsequently becomes, — by degrees, more ovoid. The calcareous particles then begin to be deposited in considerable numbers, forming very obvious longitudinal and transverse striae, and the shell is much less transparent than be- fore. However, it is still possible to distinguish the internal organs : the heart and the bladder have divided into two cham- bers, the upper of which is the smaller. When the auricle con- tracts, the ventricle dilates, and vice versd. A very strong muscle is also observable, arising from the internal surface of the shell and passing to the foot (fig. 1 1 *) ; its office is to retract the ani- mal into its shell. Finally, the liver appears upon the under surface of the stomach ; it is ovoid, and is formed of a mass of granules containing yellow pigment (fig. 11 r). On the inner surface of the mantle a series of folds may be observed {feuillets muqueux, Cuvier), in which lie a mass of mucous crypts. Subsequently the little animal continues to grow ; more and more calcareous particles are deposited in the shell : the mantle thickens, and it becomes almost impossible to distinguish the internal organs. The two rounded lobes have disappeared, but behind the tentacles we see a linear eminence indicating the place they occupied. The shell has assumed a horn-yellow co- lour; it becomes hard, brittle, and only semitransparent. It was commonly in this state that the young left its capsule, after a residence of at least eight weeks there, creeping around the vessel in w^hich it had been kept, with tentacles, foot, and siphon protruded. The little w^helks are now distinguishable from the adults, only by their shell having not more than one or two spiral turns. We may add, that we have found no traces of generative organs in these young. The ova, grouped together in considerable numbers, occupy the posterior part of the shell. § 2. Purpura lapillus. The ovigerous capsules are not unlike a little flask whose rounded end is directed upwards — and its delicate neck, by the extremity of which it is attached to stones or other bodies, down- wards (PI. XI. fig. 1). Each capsule is hermetically closed, and 342 KOREN AND DANIELSSEN ON THE contains a transparent, pellucid, viscous liquid, like the white of egg — investing a mass of ova (60 or more) . A number of the ova were placed under the microscope, after freeing them from the viscous humour in which they had been enclosed, and we saw that they possessed a delicate chorion, a vitelline membrane, and a yelk consisting of a liquid, with many contained granules. We were unable to distinguish either ger- minal vesicle or spot. Each ovum was 0*194 millim. in dia- meter. After some days had elapsed, we opened another capsule, and observed, in the greater number of ova, the commencement of a cleavage which appeared to be altogether irregular. In fact, the number of spheres indicated by this cleavage was very variable; and some of the ova, which, we may add, were all provided with a chorion, had assumed an ovoid form (figs. 3, 4, 5, 6, T, 8, 9). The cleavage masses were all dark, and unpro- vided with any nucleus. M. Nordmann has been equally unable to observe any nucleus in Tergipes, Rissoa, or Littorina, The clear body which MM. Van Beneden, Nordmann, H. Rathke, F. Miiller, Loven, and other authors, have seen passing from the interior of the vitellus to its surface (to which F. Miiller and Loven attribute the power of determining the cleavage), has not appeared in the ova we are describing, although we have taken great pains to look for it*. Some days later we examined many other capsules. The viscous albuminous humour had undergone no appreciable change. However, the ova were not so scattered as before, and had approached one another. Examining them microscopically, we observed that some had undergone no division ; others had remained in a state of incipient division, — while around these imperfectly developed ova there were a great number of others in which division was more advanced. We see then, in this case, that the ova have a disposition to collect together, and that although enclosed in the same capsule, they exhibited a great difference in the extent to which the cleavage process had taken * In a very recently published Appendix to this Memoir (Supplement til Pectinibranchiernes Udviklings-historie, af J. Koren og D. C. Danielssen), llie authors give an account of new observations on this point in Buccinum unda- turn. In examining the youngest ova they have seen the clear (oily) body pass out, and they describe, at length, the mode in which this takes place. DEVELOPMENT OF THE PECTINIBRANCHIATA. 343 place. In these ova, 2, 4, 6, 7? 9^ 10, and even 18, cleavage masses might be counted ; the contents of all were obscure, and without any nucleus. Even at this period, we thought we no- ticed a tendency to agglomerate the ova in the viscous liquid, analogous to that whose effect we observed in Buccinum unda- tuniy but it was far from being well marked, and the commencing cleavage threw great obscurity upon what was going on. But all our doubts were completely dissipated on the twelfth day, when the phaenomenon which had been exhibited by Buccinum unda- turn was repeated by Purpura lapillus. In fact, the ova were agglomerated and formed a compact mass; the viscous and albuminous liquid had at the same time become as clear as water, and could be separated from the conglomerate with great ease. Examining the latter with attention we observed that it was composed of many groups of various extent and without any determinate form ; these groups, under the microscope, ap- peared to be composed of ova, the greater number of which had undergone cleavage, while others had not (PL XI. fig. 24). On the sixteenth day we re-examined many capsules. All the ova were agglomerated, but the conglomerate was a little altered, inasmuch as certain groups had become more distinct, more sharply circumscribed, and projected more from the common mass. Some were cylindrical, others pyriform, but they were all terminated by a peduncle which connected them with the common mass (fig. 27). The microscopical examination of each group showed it to be formed by the union of ova imbedded in a very viscid mass, and invested by a delicate membrane, which soon became covered with very fine cilia (fig. 27). The ova themselves had undergone no further cleavage, and it ap- peared to us that this process had stopped as soon as the agglo- meration commenced. Soon after, we observed a well-marked grayish, semitransparent, finely granular matter, exuding from the sides of the above-mentioned peduncle, and becoming, at a later period, covered with vibratile cilia (the foot) (figs. 26, 27 & 28 A). We also observed a similar mass becoming developed, in the same manner, at the base of the peduncle, and giving origin to two lobes, which afterwards increased and acquired fine cilia at their edges (fig. 7 d). The embryo thus constituted now begins to move a little by the help of its cilia. In fact, it was observed to make feeble efforts in various directions, as if it 344 KOREN AND DANIELSSEN ON THE sought to detach itself from the common mass — and having, at length, after many fruitless attempts, succeeded in so doing, it immediately began to rotate upon itself. We have, in this manner, observed all the individuals of a conglomerate, be- coming detached and separated, one after the other, and when all the embryos were developed, the mass had totally disappeared. It would seem that in this animal, as in the Buccinum, the number of ova which are grouped together to form the future embryo is altogether variable and fortuitous, for not only is there no discoverable law regulating their union, but these conglome- rates are constituted by very different numbers of ova. Thus we have observed in the same capsule some embryos resulting from the union of three or four ova, while sixty or more, had contributed to form most of the others. The different size of the individuals depended on the same cause. This difference of size was very considerable, for we observed swimming in the liquid contained in the capsule, some embryos of I miUim. in diameter, and others of as much as 1^ millim. The number of the embryos in a given capsule varied as much as their size ; — depending in the same manner on the greater or less number of the ova which had united to form each individual. On the average we found from twenty to forty, rarely more. Having now become acquainted with the mode of formation of the embryo in Purpura lapillus, let us turn to another phae- nomenon, one of the most surprising which is to be met with in the development of this moUusk, and which will help to explain the singularity in the development of Buccinum, to which we have already referred. It will be remembered that in the latter animal, many of the ova took no share in the act of conglomera- tion (probably in consequence of accidental obstacles), and that their ova soon died, or rather became developed in an excessively incomplete manner. Something similar occurs in Purpura; and as we have had better opportunities for observing this pecu- liarity in the latter moUusk, we are enabled to give a fuller ac- count of it. We have alw^ays found in each capsule an ovum undergoing all stages of cleavage, and which was composed, until the end, of a peripheral layer of clear, and a central mass of dark cells (figs. 10 & 11). A membrane then became rapidly developed around the yelk, and acquired exceedingly fine cilia ; at the upper part of the peripheral layer there were also visible DEVELOPMENT OF THE PECTINIBRANCHIATA. 345 he rudiments of the two rounded lobes (velum) with the foot ^figs. 12 a^bjC. 13 & 14 ^, c). Cilia quickly made their appear- ance both on the foot and on the lobes : subsequently, scattered cirrhi could be observed upon the lobes, and then the embryo began to rotate. Later still the lobes and the foot increased in size (figs. 15, 16 b, c), and the rudiments of the auditory organs (fig. 16 d) appeared at their base, — the membranes of the mantle became more and more thickened, the shell began to be formed on its most sloping part,, and calcareous particles to be deposited in it (figs. 14, 1 5, 16 a). The embryos whose first development we have just been following were true monsters, and subsequently assumed such various and whimsical forms, that no one could have imagined them to be individuals of the same species. In a few we have seen the salivary gland appear (fig. 16 e), but it was the only new structure which appeared after the formation of the external organs, and these beings permanently remained in the same state of arrest of development. Finally, until eight weeks had elapsed, this monstrous embryo was always to be met with in the capsule. We have already stated that an ovum of this nature always existed in each capsule, and its embryo was known at once by its small size and the excessive vivacity of its motions. We have sought for them in vain in the capsule after eight weeks, and we suppose that they had all perished. When our attention was first directed to these simple ova, which had regularly undergone the cleavage process, we imagined that their development had taken place in a normal manner; but, far from this, it was, in fact, an abortion. For the viability of the indivi- dual organized, more than one ovum is necessary : and despite the regularity and the vivacity observable in the young product of the single ovum, we see that its development remains in the highest degree incomplete. This single ovum had in fact under- gone all the stages of cleavage, and to all appearance united all the anatomical and physiological conditions necessary to its complete development, — while, on the other hand, it appears to us to be incontestable, that it had never been in possession of the materials requisite for the formation of organs. Without doubt there is much that is obscure in these ideas ; we shall endeavour by and by to throw as much light as we can upon them. Having described the monstrous embryo resulting from a 346 KOREN AND DANIELSSEN ON THE single ovum, we may return to the embryos formed by multiple ova, and describe at length their further development. We have already remarked, that after the formation of the ciliary membrane, the foot and the two rounded lobes are the organs first developed. At about the same time we perceive a transparent mass between the membranes and the conglomerate ova (fig. 28 d). Cells are developed in this mass in layers, and give rise to the mantle (figs. 29, 30^). The most sloping part of the latter secretes a very clear and viscid humour, which increases by degrees and forms the rudiment of the shell, which, when it first appears, resembles a clear and gelatinous pellicle, w^herein subsequently calcareous particles are deposited (fig. 29 a). These particles afterwards increase in number and impede the exami- nation of embryos which are a little older. The lobes are small at first, but their volume rapidly increases; a multitude of cilia appear on the surface, cirrhi are developed from their upper edge and produce much more vivacious move- ments (figs. 29 6^, 30 e). The foot becomes strongly separated on the ventral surface and thus forms a transverse eminence (fig. 28 b), which rapidly increases in volume and exhibits the first rudiments of the auditory organs at its base, developed as in Buccinum undatum (figs. 29 e, 30/). At the same time as the auditory organs, the rudiments of the tentacles, of the eyes and of the salivary gland, appear. The tentacles commence as two conical eminences, at the base of which the eyes are visible as two rounded vesicles, filled internally with a pellucid liquid 5 dark pigment granules may be seen in them (fig. 31 1,m). We have been unable, at this stage of development, to discover any lens, nor have we met with any cilia on the internal parietes of the vesicle. The first trace of the salivary glands which manifests itself, is a mass of rounded cells upon each side of the base of the foot, which are usually nucleated. These cells soon acquire a delicate membrane, which afterwards elongates to meet the future oeso- phagus, whose outlines are not yet distinguishable. In pro- portion as the salivary glands are developed, these cells become more and more multiplied in their interior and are closely dis- posed in elongated lines ; we see also, in the widest portion of this organ, a mass of dark yellow pigment granules. In its more DEVELOPMENT OF THE PECTINIBRANCHIATA. 34? delicate part, directed towards the oesophagus, the excretory duct of the gland is indicated, and elongates to meet this por- tion of the intestinal canal (PI. XL figs. 30, 31 g, PL XII. fig. 4). The salivary gland forms but a single conglomerate mass in the adult, but its double excretory duct shows clearly enough, that it was divided into two parts in the very youngest state. The heart was seen on the twenty-third day. The mode of its development is analogous to that which occurs in Buccinum. It is situated upon the dorsal side, presents the form of a bladder, and is directed from above downwards and from left to right ; it contracted in this direction, at the rate of forty to fifty pulsations a minute. It possesses primitive muscular fibres, having the form of longitudinal tubes, undivided above. We have met with neither cells nor nuclei in these tubes (PL XII. figs. 1,3^). In this stage of development, the branchial cavity is not deep enough to contain the whole of the heart, a considerable portion of which extends beyond the edge of the mantle. Subsequently, when the mantle elongates and covers in the back of the animal, its edge is directed more outwards and detaches itself from the body, so that the branchial cavity becomes deeper and wider and encloses the heart completely. We have as yet been unable to observe the circulatory current in this mollusk. It is only after the formation of these organs that the buccal aperture, the proboscis and the oesophagus, are perceptible. The proboscis is exceedingly short and its parietes are very thick, so that it is easily seen through the oesophagus (fig. 31i). The latter is cylindrical and takes a course beneath the stomach (figs. 31, 32 k). The latter hes to the left, it is small and oval, and a long and delicate intestine passes out of it, which turns to the right, bends back afterwards to the opposite side in a curved direction, and finally terminates in the branchial cavity by a projecting anus (fig. 32 /, m, n). The oesophagus, the stomach and the intestine are ciliated upon their inner surface. It is not until a somewhat later stage of development that the nervous system is distinctly discoverable. It is composed of two cerebroid ganglia upon each side of the oesophagus (figs. 31 w. 348 KOREN AND DANIELSSEN ON THE 32 q) ; these ganglia are united by means of a commissure, and give rise to two other commissures (fig. 31 7i, s), which connect them with the pedal ganglia. They are oval, are distinguished by their clear yellow colour, and send a great number of branches to the foot (fig. 32 s). We have been unable to trace the nervous system further, all parts of the body having rapidly become opake. It is also about the time of the appearance of the nervous system, that the first traces of the branchiae, of the siphon and of the retractor muscles of the foot, are discoverable. The branchiae spring from the edge of the mantle and then form a hollow cylinder, which is tuisted into loops; fine cilia appear upon its inner edge. Subsequently it becomes a little flattened and consider- ably spread out. In its parietes, longitudinal and transverse fibres are discoverable, which we regard as muscular tubules. The cilia which exist in the middle of each loop have an extra- ordinary length (PL XII. fig. 8 b, c). After the development of the branchiae, it becomes exceed- ingly difficult to make out the formation of the other organs ; on the one hand, because the animal rarely protrudes from the shell sufficiently to show these parts, and on the other, because the mantle is greatly thickened and a large quantity of calca- reous matter has been deposited in the shell. The latter has taken the shape of that of a Nautilus, and when it is placed under a strong magnifying power, it is observable that the cal- careous matter is deposited in the form of a network with fine meshes (fig. 2). The two rounded lobes diminish in volume (fig. 5). The foot, lobed above, takes on more and more the shape of that of the adult animal, and the operculum which closes the aperture of the shell is completely developed (fig. 6). The heart is, in this stage, divided into two chambers, whence the great vessels arise. The lenses of the eyes are clearly di- stinguishable, and we have frequently met with a single eye presenting two streaks of pigment, but never with more than one lens. The branchial cavity, whose internal surface is covered with cilia, has become, at this period of development, deep enough to enclose the heart completely. The edge of the mantle, which divaricates further and farther from the body of the animal, is ciliated, and at the bottom of the branchial cavity DEVELOPMENT OF THE PECTINIBRANCHIATA, 349 we observe, for the first time, a contractile vesicle similar to that which exists in Buccinum undatum. After the lapse of eight weeks the young had not yet left the capsule, but when one was taken out, it began to creep like the adult animal, with foot, tentacles and siphon stretched out. It is then distinguished from the adult by the incomplete dis- appearance of the lobes, by the shell being still soft, and by the spire having only one or two turns (fig. 6). From the ninth and tenth weeks the young begin to leave the capsules, the rounded lobes disappear, and behind the tentacles an elevated line may be seen occupying their previous place (fig. 7) ; the shell has elongated and approximates more nearly that of the adult ; it is hard, frangible, and almost opake ; the last turn of the spire, however, is not yet formed. We shall not describe the development of the organs of Purpura at greater length, because it does not differ from what we see in Buccinum undatum. We may call the attention of the reader to the interesting investigations of Kolliker and Siebold* on Actinophrys Sol and on Diplozoon paradoxum', for, perhaps, something approxima- ting to what we have just described may be found in their ob- servations. In conclusion, we may add, that we entertain a strong desire to have the opportunity of continuing our investigations upon other genera allied to Buccinum and Purpura, for assuredly, with some slight modifications, all these Mollusks are developed in the same way. Summary, For the more ready comprehension of the history of the de- velopment of Buccinum undatum and of Purpura lapilluSy we here shortly sketch the most essential points. Buccinum undatum. 1. The ovigerous capsule is filled with a transparent, colour- less viscous liquid which resembles white of egg. Each capsule encloses a mass of ova (of 6-8 centimetres). ♦ Zcitschrift fur wiss. Zoologie, torn. i. p. 198, and tom. iii. p. 62. 350 KOREN AND DANIELSSEN ON THE 2. Each ovum consists of a chorion and albumen, of a vitel- laiy membrane, and of a vitelhis composed of larger or smaller globules. Its diameter varies from 0*25 7-0*264 millim. In the eggy when laid, we have never been able to observe either germinal vesicle or spot. 3. The cleavage which occurs in other mollusks does not take place in these animals. 4. The ova begin to approximate towards the eighteenth day, and the chorion is detached. The yelk, more or less laid bare, invested only by its very firm membrane, is enveloped by the viscous albumen-like liquid. 5. Some days later, the ova, even those which were most distant, approximate and form only a single mass, whose dif- ferent portions, larger or smaller, have become grouped ; so that each group, usually composed of from six to sixteen ova, is distinguishable by the naked eye. 6. On the twenty-third day, these groups are still more distinctly marked out, and are invested by a very delicate mem- brane peculiar to each group, which has by this time taken on an oval or reniform shape : the ova are connected together, and the liquid which enclosed them has lost its viscosity. 7. Towards the twenty-fifth day the groups possess a more decided membrane and boundaries. Many of the ova which have remained isolated and simple appear as embryos, whilst the others are attached together. 8. The embryo thus formed consists of a delicate membrane enclosing many ova. 9. The number of the ova grouped to form one embryo varies greatly, amounting in some cases to as many as a hundred and more. 10. The number of embryos in different capsules varies ; commonly it is from six to sixteen. 11. The first organs which are formed after the membrane in question are the rounded lobes with cilia and cirrhi. [The em- bryo then begins to move.] At a later period, the foot, the mantle, the shell, the auditory organs, the proboscis, the eyes, the salivary gland, the heart, and the contractile bladder appear ; still later, the digestive and the nervous systems and the branchiae, are developed. DEVELOPMENT OF THE PECTINIBRANCHIATA. 351 12. After an interval of at least eight weeks, we see the young leave their capsule ; the shell is a little more elongated (about 2 millim. long), hard, frangible and semitransparent. The lobes have disappeared, and the young animal creeps like the adult ; it is still distinguishable by the number of the turns of its spire (there are only one or two). We should observe, that we have hitherto found no traces of the generative organs in these young. 13. Finally, the grouped ova are sufficiently numerous to fill the posterior part of the shell. Purpura lapillus, 1. The ova, dispersed through an extraordinarily thick and viscous humour, fill the ovigerous capsules of the animal, which are flask-shaped. 2. The size of the ovum is about 0*194 millim. This ovum is composed of a delicate chorion, of an albumen, of a vitelline membrane and yelk. 3. The vitellus undergoes a very irregular cleavage. The cleavage masses have no nuclei. 4. After a certain progress of the cleavage, the ova begin to be grouped together. 5. On the twelfth and thirteenth days the ova have become, so to say, a compact mass, subdivided into many masses of ova, disposed in bunches. 6. On the sixteenth day the separate groups were more sharply circumscribed, and projected from the remainder of the mass. These projecting groups soon took a cylindrical or pyri- form shape, and were fixed to the rest by means of a peduncle. The microscope showed them to be composed of a delicate ciliated membrane, and that they enclosed a mass of ova; a transparent substance exuded from the two sides of the peduncle, upon which fine cilia appeared (the foot) ; and at the base of this same peduncle the first traces of the lobes were distinguish- able. Finally, many of these pyriform bodies became detached from the mass and rotated upon themselves; these were the embryos. 7. It is impossible to determine the exact number of the ova which go to form one embryo, as it varies greatly. In each capsule 352 KOREN AND DANIBLSSEX ON THE an embryo proceeding from a single ovum is constantly met with; but this embryo never attains to a complete development. 8. The number and size of the embryos vary in different cap- sules ; the average number is from 20-40. The largest embryo was 1^ millim. in diameter. 9. The first organs w^hich are formed after the tegumentary membrane are the foot with its vibratile cilia and the two rounded lobes, w hich are both ciliated and provided with cirrhi ; then the mantle, the shell and the auditory organs, the salivary glands, the heart (on the twenty-third day), the eyes and the tentacles. The digestive apparatus, the nervous system, the branchiae, the siphon, and the retractor muscles of the foot, ap- peared later still. Subsequently the heart divides into two chambers, the shell presents one or two spiral turns, and it is only after all these changes that the contractile bladder appears. After eight weeks the young had not yet left the capsules ; and when one was extracted in this stage of development, it began to creep like the adult animal, from which however it was distinguished by the lobes, which had not wholly disap- peared, and by the shell, which still had only one or two spiral turns. 10. Towards the ninth or tenth week, the young leaves the capsule ; the lobes have disappeared ; the shell has become fra- gile and opake. EXPLANATION OF THE PLATES. Buccinum undatum. Plate X. Fig. 1 . An egg from the oviduct of Buccinum, x 200. a, yelk ; h, germinal vesicle ; c, germinal spot. Fig. 2. Egg from a capsule, X 200. a, chorion ; h, vitellary membrane ; c, yelk. Figs. 3 & 4. Embryos, a, chorion ; 6, yelk with its membrane ; c, rudi- ments of the two lobes. Fig. 5. Embryo seen from the side, a, membranous shell ; 6, mantle ; c, yelk ; d, lobes ; e, foot. Fig. 6. Embryo. Dorsal view, f, heart. Fig. 7. Embryo. Ventral view. /, foot. Fig. 8. Embryo. Dorsal view. A, salivary glands. Fig. 9. Embryo. Ventral view. », tentacles; /.-, proboscis; /, oesophagus; OT, stomach. DEVELOPMENT OF THE PECTINIBRANCHIATA. 353 Fig. 10. Embryo from the side, a-g, as in the foregoing figures; h, eyes; i, tentacles ; k, proboscis. Fig. 11. Young Buccinum seen laterally, a, shell; b, mantle; \ IS' n & (J rw m f\ 13 I ^u;a:: IS (i\ 18 v^. y^- J'. £a^tr& sc . ScUtl. JiUrrv- Ifevt. Mist . VoLl. PI. YDl J!^^l. J^asirfi .V Scim,.Menv^NatJIist. VoLI. 7?.X. J.Baswe.- Sciai. J4£fiu^ Nat. Hut. \'ol. l.n.Jl. # HO C!> J.Sasire . Saen.Mm. . N^at2I{stY6ll.Fi:XlL. 'b f\Ac^. SCIENTIFIC MEMOIRS, SELECTED FROM THE TRANSACTIONS OF FOREIGN ACADEMIES OF SCIENCE, AND FROM rOEEIGN JOURNALS. NATURAL PHILOSOPHY. EDITED BY JOHN TYNDALL, Ph.D., F.R.S., / , AND WILLIAM FRANCIS, Ph.D., F.R.A.S., F.L.S. '| LONDON: TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET. Printers and Publishers to the University of London. 1853. \/ FLAMMAM, " Every translator ought to regard himself as a broker in the great intellectual traffic of the world, and to consider it his business to promote the barter of the pro- duce of mind. For, whatever people may say of the inadequacy of translation, it is, and must ever be, one of the most important and meritorious occupations in the great commerce of the human race." — Goethe, Kumt und Mterthum. CONTENTS. PART I. Page Art. I. — On the Mechanical Equivalent of an Electric Discharge, and the Heating of the Conducting Wire which accompanies it. By R. Clausius I Art. II. — On the Processes which have taken place during the formation of the Volcanic Rocks of Iceland. By R. Bunsen 33 Art. III.— On the Dependence of Radiant Heat in its passage through Crystals upon the direction of transmission. By H. Knoblauch, Professor of Natural Philosophy in the University of Marhurg , 99 PART II. Art. III. — On the Dependence of Radiant Heat in its passage through Crystals upon the direction of transmission. By H. Knoblauch, Professor of Natural Philosophy in the University of Marburg {continued) 101 Art. IV. — On the Conservation of Force. By Dr. H. Helmholtz 114 Art. V. — On the Connexion of Diamaguetism with Magnetism and Electricity. By M. W. Weber 163 Art. VI. — On the Work performed and the Heat generated in a Conductor by a Stationary Electric Current. By R. Clausius 200 PART III. Art. VI. — On the Work performed and the Heat generated in a Conductor by a Stationary Electric Current. By R. Clausius {continued) 201 IV CONTENTS. Art. VII. — On the Deviation of Projectiles ; and on a remarkable Phsenomenon of Rotating Bodies. By G. Magnus 210 Art. VIII. — On the Motion of Liquids in a closed Galvanic Circuit. By G. Wiedemann 232 Art. IX. — Researches upon the Optical Characters of Double Re- fraction in Isomorphous Substances. By M. H. de Senar- MONT 2.57 PART IV. Art. IX. — Researches upon the Optical Characters of Double Re- fraction in Isomorphous Substances. By M. H. de Senar- MONT (continued) 281 Art. X. — On the Phaenomena of closed Electro-Magnets. By J. C. POGGENDORFF 294 Art. XI. — On the Nature of those Constituents of the Atmo- sphere by which the Reflexion of the Light within it is effected. By R. Clausius 303 Art. XII. — On the Blue Colour of the Sky and the Morning and Evening Red. By R. Clausius 32G Art. XIII. — On the Theory of Diamagnetism, the Explanation of the transition from the Magnetic to the Diamagnetic deport- ment, and the Mathematical treatment of the Phaenomena observed in Crystals. By M. Plucker, Professor of Natural Philosophy in the University of Bonn 332 FIVE PLATES. SCIENTIFIC MEMOIRS. NATURAL PHILOSOPHY. Article I. On the Mechanical Equivalent of an Electric Discharge, and the Heating of the Conducting Wire which accompanies it. By R. Clausius. [From Poggendorff's Annalen, No. 7, 1852.] oY the application of heat we are able to obtain mechanical work, and it is known that electric currents are also capable of producing various mechanical effects, and also of producing heat. All those phaenomena possess in themselves great inter- est, and this interest is considerably increased by the practical applications which have been made of these forces, and which may possibly yet be made of them. To this is added the cir- cumstance that these actions are capable of strict mathematical treatment, and are therefore peculiarly suited to the investiga- tion of the relation which subsists between them and their causes, and also of the varied influences which they exert upon each other. They have indeed already been made the subject of various investigations. Particular attention has been heretofore devoted to galvanic electricity and to electro-magnetism, inasmuch as SCIEN. MEM.— AW. Phil. Vol. I. Part I. B 2 CLAUSIUS ON THE MECHANICAL EQUIVALENT these have been most frequently employed and upon the largest scale. In such an investigation, however, it appears to me more suitable to commence with machine electricity ; although this subject presents greater difficulties to mathematical treat- merit, still in principle it is the most simple ; for here we have to do with electricity alone, unaccompanied by the incidental actions of chemistry and magnetism. In the following pages I have endeavoured to reduce the effect produced by an electric discharge to a definite unit derived from the principles of mechanics, and have compared in certain simple cases the result arrived at with those of experiment. The coincidence has, as will be observed, been so satisfactory, that in my opinion the deduced result is not only an undoubted law of electric action, but also a new corroboration of the me- chanical theory of heat. Let us imagine a system of material points possessing the masses m, m', W, &c., and having, in a rectangular system of coordinates, at a certain point of time /, the coordinates x,y, Z', a^, ?/', z^ ; a?", 2/", z", &c. Let these masses be acted upon by a system of given forces, and let the components of the total force which acts upon m be X, Y, Z, and for m', X', Y', Z', &c. Let the points be either quite free to move, or else limited in their motions, which last will of course be the case when the points are in any way connected together, as also when certain exte- rior conditions of motion exist ; for example, if one of the points should be compelled to move in a certain surface, or along a cer- tain line. The conditions, however, must not be such, that by them alone, and without the exercise of the given forces, motion can take place ; w hich, for instance, would be the case if the surface or line just supposed, and which the point cannot forsake, were itself in a state of motion, dr such that the motion of the given masses should be able to set other masses in motion which are not embraced in the system. That is, in other words, the moving forces and the masses moved by them must be given expli- citly. Finally, let the velocities of the masses m, m', m", &c. at the time t, be denoted by v, v\ ?;", &e., we then obtain the fol- lowing general expression, ^^md{f)z:zl,{:^dx^Xdy^r'Ldz), . . . . (1) OP AN ELECTRIC DISCHARGE. 3 in which the signs of summation S refer to all the masses of the system, and to their coordinates, forces, and velocities. In this equation the left-hand side is immediately integrable, and so is the right side, if we regard the quantities x, y, z; a/, y' z', &c., not as independent variables, but as w hat they really are, all of them functions of one and the same variable, namely, of the time t. We thus obtain ^-^mv^=fX{Xdx-\-^dy-\-Zdz)+con?^i.', or, if we call the velocities at any time of commencement what- ever Vq, Vq', Vq, &c., and take this time as the lower limit of the integral, \y,mv^-\Zmv\=r*L{Xdx^-Ydy + Zdz), . . (2) The quantities which appear in these equations are of very common occurrence in mechanics, and have on this account re- ceived particular names. With regard to that on the left-hand side, it is known that the product mv^ is called the vis viva of the mass m ; and according to this ^nvif^ would express the vis viva of the entire system. Here, however, only the half of this quantity makes its appearance ; and as this will always be the case in the following pages, and likewise occurs so in most me- chanical investigations, we propose to call, as other authors also have done, the quantity -'Xmv'^ the vis viva of the system. The quantity on the right-hand side can, in the first place, be brought to a more simple form. Let ds represent the element of its path which the mass m describes during the time dt, and of which dx, dy, dz are the three projections on the three axes of coordinates; further, let S be the component of the total force acting upon m in this direction, we then have ^ds^r^Xdx + Ydy + Zdz', the quantity ^ds is that which in treatises on mechanics is called the quantity of work produced by the force acting upon m during the time dt, bearing in mind that this work must be regarded as positive or negative, according as the element ds is in the same direction as S or opposed to it. Applying these names, the meaning of the above equation may B 2 4 CLAUSIUS ON THE MECHANICAL EQU IVALENT be thus expressed: the increase of via viva in the system during any given time is equal to the quantity of mechanical work pro- duced during the, same time in the system. The determination of the work may be much simplified in particular cases \Yhich very often present themselves. For instance, suppose a portion of the given forces to consist of attractions and repulsions which remain equal throughout the whole time, these forces being either such as might be ex- erted by other points upon the given ones, in which case the former must remain motionless, or such as the given points may exercise upon each other; let the intensity of each of these forces depend solely on the distance, and not upon the position which the attracting or repelling points may occupy ; they may be any function whatever of the distance. The portion of the total sum due to these forces will then be expressed by l.^[Xdx + Ydy + Zdz)', and not only is this a complete differential, because all the quan- tities which appear in it are functions of one and the same va- riable t, but it also remains so if the single quantities oc^y, z-y a?, ?/', ^ 'o while the member to the left expresses the increase of the vis viva of the entire system due to the discharge. This last can also be of a twofold nature. In the first place, under favourable circumstances, actual visible motions may be produced in the system by the electric attractions and repulsions ; secondly, and more especially, heat will be excited by the current in the con- ductors. Let us suppose all the above-mentioned quantities of negative work to be positive and brought over to the left-hand side of the last equation, so that only the increase of the potential remains to the right ; the proposition contained in the equation can be expressed in a very complete manner as follows : — The Sum of all the effects produced by an electric discharge is equal to the increase of the potential of the entire electricity upon itself, in which place under electric discharge is meant every alteration in the arrangement of the electricity, by which the electric states of the various portions of a system of conducting bodies, in which the earth itself may be included, entirely or partially neutralize each other. We will now apply this general proposition, which for the sake of brevity will be hereafter called the principal proposition, to the special case of a Leyden jar, or to a battery composed of a number of such jars; this case is of peculiar interest on account of its frequent occurrence, and presents the best oppor- tunity of comparing the results of theory with those of ex- perience. With regard to the latter, the investigations of Riess, which are carried out with the greatest care and ability, furnish us with a fund of materials in which the greatest confidence may be placed; the comparison of the same with our principal pro- position is rendered the more easy by the circumstance that Riess himself has deduced from the facts which he has observed determinate laws, which he has expressed in strict mathematical formulas. OP AN ELECTRIC DISCHARGE. 9 In the first place, let it be required to determine the value of the potential in the case of a charged Leyden jar or battery. In order to present the matter more simply to the mind, we will commence with the special case of a jar of particular form, and apply the expression thus found to the deduction of the general expression. .We will first choose a form without indeed an actual existence, but which, however, in all essential points must be subject to the same laws as the common Leyden jar, and which leads to results of extraordinary simplicity. The glass vessel shall form a closed hollow sphere, possessing at all points the same thickness, and completely covered with tinfoil both inside and out. Let us suppose a quantity of electricity Q to be imparted by some means or other to the interior sur- face, where we assume as unit such a quantity of positive elec- tricity as exerts upon an equal quantity of the same electricity the unit of force at the unit of distance. Let the outer surface stand in connexion with the earth, and let the quantity of elec- tricity which it receives from the latter be denoted by QJ, In this case it is evident that Q and Q! must spread them- selves uniformly over both the respective surfaces, and in con- sequence of this the determination of the potential function and of the potential is greatly facilitated. The potential function V of any quantity of electricity Qi at any point O, will, in general, be determined by the equation *dq ^'-f- v where dq denotes an element of electricity, and r its distance from the point O, the integral extending over the entire quan- tity. For the particular case however where Ql is spread uni- formly over a spherical surface we do not need this general equation, but can apply the following two propositions: — 1. Within the sphere the potential function is everywhere the same, that is, if r be the radius of the sphere, r 2. Without the sphere, and for a distance R from its centre, the potential function is V — « R* 10 CLAUSIUS ON THE MECHANICAL EQUIVALENT On the surface both expressions give the same value, and hence both propositions in this case hold good simultaneously. In the instance before us the tinfoil coatings form two con- centric spherical surfaces, whose radii may be called a and a + c, c being the thickness of the glass. Let us first consider a point on the interior coating ; in this case we can apply the first pro- position to both the spherical surfaces ; connecting the potential functions of both quantities of electricity Q and Q! together, we obtain V=-5_-«L (5) a a-{-c For a point on the outer coating, on the contrary, the second proposition is to be applied, and if in this case we call the po- tential function V, we have (6) a + c a-\-c Through the condition that the exterior surface is in commu- nication with the earth, we possess a means of estimating the quantity of electricity Of. It is known that when several con- ducting bodies are connected together, that the equilibrium of the electricity is so established that the potential function in the interior of the entire system possesses the same value. Hence, as in the earth, where in general equal quantities of positive and negative electricity are present, the potential function is zero; this must also be the case for the exterior covering. We have therefore V' = 0; and hence, according to (6), and thus (5) becomes V=-Q_£_; (7) a{a-i-c) when the fraction is developed according to C, the superficial content of the inner coating, that is 4aV, being denoted by S, we have (i-£+l,-&c.). . . . (7«) We are now in a condition easily to determine the potential OF AN ELECTRIC DISCHARGE. 11 of the entire electricity upon itself. The potential W of a given quantity of electricity upon itself is, in general, where dq and d(i represent any two elements of the electricity, r their distance asunder, and both the integrals extend over the entire quantity, the factor - must be introduced, because in the double integral every combination of every two elements dq and d(^ appears twice. Now as -/ -*_v, r instead of the foregoing expression we can write y^=\fydq (8) Now, as before mentioned, in every connected conducting body the potential function is constant, and may therefore be taken from under the sign of integration ; the integral that remains represents simply the quantity of electricity distributed over the body. If we apply this to the two coatings of a Ley den jar with the potential functions V and V, we obtain as the total potential of both the quantities Ql and Q! upon themselves, W=l(Q.V + a'.V); (9) and if for our special case we set V'=0, and in the place of V set the value of it, as found by (7) or (7«)^ we obtain the re- quired potential in the case of a charged spherical jar, W=-Q^— £ ,, (10) or W=-^^27^cfl-f-f ^-&c.V . . (10«) S \ a a^ J The plate of Franklin with circular coatings may be regarded as the next simplest form of the Leyden jar. In a former me- moir I have devoted especial attention to this form, and from the results there obtained I will introduce but one here, which corresponds to the foregoing example. This is the case in which one of the metallic coverings is supposed to be connected with w= 12 CLAUSIUS ON THE MECHANICAL EQUIVALENT the earth, while a quantity of electricity Q is communicated to the other. On the former the value of the potential function is of course zero, while on the latter, when c denotes the thickness of the glass, and a the radius of the circle, the value is for which, when the superficial content of one of the coatings, that is arir, is denoted by S, we can write To obtain from this the potential of the total electricity upon itself, it is only necessary to multiply by |Q, hence IM'-4("«^'«)]- • ■<-' Comparing the expressions (11) and (12) with (7a) and (10a), we observe that the principal members — — 47rc and o — -^27rc coincide in both cases, and the divergence first appears IS in those members which, in regard to c, are of a higher order than the first, and hence if c is small compared with the dimen- sions of the coatings, they may be neglected. The same may be proved in general for every other Leyden jar in which the glass is of a uniform thickness, that is, where c is constant. Neglecting therefore the members of a higher order, we obtain the equations V=-|4,rc, (13) W=-^2^C, (14) by means of which jars of various forms, magnitude, and thick- ness of glass may be compared witli each other, with this con- dition only, that the thickness c is constant in each, which it must be confessed is only partially the case with jars actually applied. When instead of different jars, a battery of jars, all of the same size, is applied — a case of frequent occurrence in practice — and that the superficial content is changed solely by increasing or decreasing the number of jars, the various cases that then pre- OF AN ELECTRIC DISCHARGE. 13 sent themselves can be compared with each other, without neglecting the members of a higher order, and without the limiting assumption made with respect to the thickness of the glass. Whatever the nature of the jar may be, when s is the surface of its interior coating, and q the quantity of electricity upon it, we can set V=-|A, (15) where A: is a constant dependent on the nature of the jar, which, though not known, is the same for equal jars, and besides this is of the first order in respect to the mean thickness of the glass. Now let n such jars be taken, and having been all charged to the same degree, let all the inner and all the outer coatings be connected together. Then if we do not take into account the influence exerted by the jars upon each other, in case they stand too near, nor the influence of the electricity distributed over the connecting pieces, then as regards the potential function no change will be effected by the combination, but the potential, on the contrary, in the entire battery will have an n times greater value than in every single jar. Hence calling the surface of the common interior coating S, and the total electricity distributed over it Gl, to obtain the quantities V and W for the entire bat- tery it is only necessary to set in the foregoing equations in the Q, S place of q its value - , and for s its value -, and to multiply the second equation by n. In this case n disappears in both equa- tions, and we again obtain v--|j (in w=-f.|. ,.„ Regarding then the potential of a charged Leyden jar or battery as known, the increase of the potential by the act of discharge, and hence the work produced by the electricity are given. If the 14 CLAUSIUS ON THE MECHANICAL EQUIVALENT discharge take place only partially, and the potential of the re- sidue be denoted by Wp then the work produced will be =W,-W, (19) which, as W and W, are always negative, and Wj has an abso- lutely smaller value than W, is a positive quantity. If, on the contrary, a complete discharge takes place, we must set Wi = 0, and in this case the work produced is = -W (20) We will now consider the effects which are due to the dis- charge. Let us suppose the discharge to be occasioned by connecting one coating wath the other by conducting bodies, whose ends are brought either so near that a spark springs from one to the other, or else brought into actual contact. In this case, during the act of approximation, an electric action takes place, inasmuch as the approximated ends of the connecting arc attract each other, and thus render their approximation easier. In our case, however, where the greatest portion of the electricity is bound to the coatings, and hence cannot contribute to this attraction, the latter must be so trifling that it may be neglected altogether. Further, to simplify the matter, we will for the present not take into account the excitation of induced currents, nor any permanent changes which are due to magnetical or chemical action ; we w ill assume that the work expended at the places where the connexion is interrvpted, and where a spark must spring across, together with the heat generated in the system, are the only ac- tions w hich present themselves. Then, in accordance with our principal proposition, the sum of these both must be equal to the increase of the potential. Let it be next assumed that in a series of experiments the strength of the discharge, that is, the increase of the potential, re- mains the same, but that the connexion is altered, in this case the sum of the actions must be constant. With regard to the development of heat, we possess, in refer- ence to its dependence on the nature of the connexion, the fol- lowing two important laws established by Riess*: — * Pogg. Ann. vol. xliii. and xlv. OF AN ELECTRIC DISCHARGE. 15 1. The quantity of heat developed in two different wires, of which the circuit is composed, is directly proportional to their reduced lengths; when under the term reduced length the quantity -2<2? is understood, in which \ signifies the actual length, p the radius, and x a quantity depending on the nature of the wire, which Riess has named its retarding force, and which corresponds to the inverse value of the conductive power. 2. When, other circumstances remaining the same, the circuit is lengthened, by introducing into it a wire of the reduced length \, the effect is that the heating of another wire which forms a por- tion of the circuit is lessened in the proportion of l + h\: 1, where b compresses a constant to be determined by experiment. Both propositions may be expressed by the following equa- tion*— c=OT-^' • ^''^ where /' denotes the length of the portion of wire under consi- deration, and C the quantity of heat excited in it, while b and / retain the signification already given to them, and A is a quantity dependent on the strength of the charge, which in our present case, inasmuch as we have to deal with equal discharges only, is constant. This equation contains a corroboration of the conclusion already drawn. The wire / will of course be likewise warmed by the discharge ; according to the foregoing equation the quantity of heat developed in it will be - — j-. A. The consequence of this will be, if the total sum of the actions remains constant, a decrease of the remaining actions, which is indeed proved by Riess^s second proposition, and by the equation. With this general coincidence we must rest satisfied for the present. An exact quantitative investigation whether the decrease of all the remaining actions taken together is actually equal to the quan- tity of heat expressed by A, seems to me incapable of accomplishment without new data of observation. Vorsselman de Heer has, it is true, deduced a general propo- • Pogg. Ann. vol. xlv. p. 23, 16 CLAUSIUS ON THE MECHANICAL EQUIVALENT sition from the equation (21), which might at first sight be taken for a complete corroboration of our conclusion. Accord- ing to this proposition, the total heat which is excited in the entire circuit by an electric discharge is independent of the nature of the circuit^. This proposition is indeed brought forward by Helmholtz as coincident with the theory f; I do not think, how- ever, that it is quite applicable, inasmuch as it contains many inaccuracies. In the first place, Vorsselman de Heer limited the considera- tion expressly to " the wire which unites both the coatings of the battery J." The development of heat extends however to the other bodies of the system, a portion of it indeed is excited within the battery itself; and another portion, in case the battery and connecting wire are not insulated but are connected with the earth, within the conducting wire and the earth. The latter portion will in general be trifling, as the excess alone of the one or the other electricity escapes to the earth, and this, in com- parison with the total quantity of electricity, is very small ; the same may perhaps be assumed with regard to the former portion, under the condition that the connecting wire possesses a con- siderable reduced length. With very short connecting wires, on the contrary, such an assumption would be unwarrantable, and this portion must, at all events for the present, be regarded as unknown. Further, he treats the connecting wire so as if it were one continuous piece. In this respect Riess § has already drawn at- tention to the fact that his connexion was established by several pieces of wire attached together ; and as his experiments refer only to the continuous wires, and not to the heat developed at the points of junction, so his proposition, so far as it is a conclu- sion from his experiments, was refuted. From a theoretic point of view, indeed, those places of junction where actual metallic contact takes place, and where the electricity in its passage gives rise to no mechanical alterations, would be embraced in the general proposition, without its being necessary to know the quantity of heat developed in each singly. It is otherwise, how- ever, with the places where an interruption occurs, where a * Pogg. Ann. vol. xlviii. p. 298. % Pogg- -^^^^^ vol. xlviii. p. 297. f His memoir, p. 44. § l*ogg. Ann, vol. xlviii. p. 320. OP AN ELECTRIC DISCHARGE. 17 spark springs over. An exterior mechanical action takes place here, which, as so much work consumed, must be deducted from the total action, in order to obtain that portion which is changed into heat within the system of bodies under consideration. With respect to the magnitude of the work here expended, and its influence on the development of heat, I am enabled to adduce an experimental corroboration of the theory. It is, in the first place, clear that the consumption of work must depend on the resistance presented to the passage of tTie electri- city by the non-conducting layer, and hence that it will be more considerable when the ends of the connecting wires are separated by a non-conducting solid body than when merely air intervenes between them. From this it follows, that in the first case an air-thermometer placed in some other portion of the circuit must be less warmed than in the latter case; this indeed has been established by a series of experiments carried out by Riess. At the place of interruption were placed either two small discs, or two small balls, or two points opposite to each other, at a distance of 0*2 of a line apart in each case. Between these the bodies mentioned in the first column of the following table were introduced one after the other; and the quantities of heat developed under precisely the same conditions are contained in the following columns. Where Riess has given several numbers I have given the mean : — Body introduced. Layer of air A card Two cards with tinfoil between them Two cards A plate of mica Heat developed in Air-thermometer on the passage of the spark. Between the discs. 15-9 n-7 9-7 8-0 6-8 Between the balls. 15-4 12-0 9-3 8-8 4-7 Between the points. 151 11-6 10-4 4-8 The influence of the solidity of the bodies penetrated by the spark is very plainly visible here*; and it is at the same time * Only in the case where two cards with tinfoil between them were applied do we find an exception, inasmuch as those three bodies exercised a less influ- ence than the two cards alone. According to this we must assume that the plate of tinfoil, although it also was penetrated, caused, not an increased con- SCIEN. MEM.— i^a^ Phil, Vol. I. Part I. C 18 CLAUSIUS ON THE MECHANICAL EQUIVALENT observable, from the great difference in the numbers, how con- siderable the quantity of work expended by the spark, in cases when difficulties are presented to its passage, may be. An exact measure of this work cannot be deduced from the experiments ; and such a measure, in my opinion, we do not yet possess, even for the simplest and most important case where the spark passes through air. At first sight it might be imagined that the quantity of work ex- pended must, when the density of the air is uniform, he proportional to the thickness of the layer. When, however, the nature of the circuit or of the charge is changed, the distance across which the spark passes remaining unaltered, such differences are observed, even in the exterior manifestations of the spark, its noise and magnitude, that such sparks cannot be regarded as alike, in respect to the quantity of work which they have consumed. Further, it might be perhaps concluded from some of the ob- servations communicated by Riess*, that the work consumed by a spark in passing through the air is so small that it might be neglected. Riess has conducted the experiments with the little discs and balls before mentioned, bringing them first into con- tact, and then within different distances of each other, so that in the first case the electricity passed without a spark, and in the latter case with a spark ; for each of these cases he observed the quantity of heat developed in the connecting wire under the same conditions of experiment. In these instances the quantity of heat developed when the discs were separated by an interval of air, was only a little less than by actual contact ; and in some cases it was actually greater, which was the more extraordinary, inasmuch as a residue here remained in the battery, while in those cases where contact was established a complete discharge took place. I believe, however, that these observations would not justify the foregoing conclusion. sumption of work, but the contrary; and this seems to involve a contradiction. I believe, however, that this assumption, even although it is as yet incapable of being firmly established, is not to be regarded as contrary to common sense. For with regard to the quantity of work expended, we have not only to take into account the bodies which were penetrated, but also the manner in which the penetration is effected, and this latter will certainly be affected by the con- ducting body introduced between the cards. * Fogg. uinn. vol. xliii. p. 78. OF AN ELECTRIC DISCHARGE. 19 With regard, in the first place, to th^ fact that where a spark occurred a residue remained in the battery, the influence of this residue must not be rated too high. This residue, even in the extreme case that the distance between the discs w^as equal to 2 the length of the stroke of the battery, could only be about — J. o of the entire charge*. By this, however, the total action of the 2 discharge would not be diminished by — of the whole, but only (2 \^ 1 — ) , or — . The total action is, according to our princi- pal proposition, equal to the increase of the potential, and the potential in the case of any given battery is proportional to the square of the quantity of electricity. Hence, if we denote the quantity of the entire charge by Q, and that of the residue by Qj, the increase of the potential is Wi~W = A(Q2-a^2)^ where A is a number independent of the quantity of electricity. This gives, according as we set, 2 Gli = 0, or =T^Gl, W,-W=Aa2,or =^^'[l-(f3)]- Further, besides those sparks which were intentionally called forth by placing the discs or balls at a distance from each other, w^e must also take those into account which w^ere the necessary accompaniment of the mode of experiment. Riess, in order to make the discharges as regular as possible, obtained them by a peculiar apparatus constructed by himselff, which was so ar- ranged that at every discharge two sparks sprang over. Now it may be deduced from other experiments by RiessJ, that by ma- king an interruption anywhere in the circuit, the distance across which the spark will spring in any other portion of the circuit will be diminished, and hence in the case before us, at the same time that they create the new spark between the balls or discs, the two other sparks must themselves become shortened, from which we can infer a partial compensation as regards the work con- • See Riess, Pogg. ^nn. vol. liii. p. 11. t Pogg. /inn. vol. xl. p. 339. J Fogg. Ann. vol. liii. p. 11. c 2 20 CLAUSIUS ON THE MECHANICAL EQUIVALENT sumed. In many cases the two latter sparks indeed disappeared totally, inasmuch as the discharge did not take place until the two. balls of the apparatus were in contact*. Hence one new spark had made its appearance, and for this two old ones had disappeared, which justifies the expectation that a lessening of the quantity of work consumed, and a corresponding increase of heat would be the result. It was, in point of fact, in these cases that Riess observed the increased temperature of the connecting wire. We thus see that in explaining these phaenomena it is not necessary to assume that the work consumed by a spark is very small ; and to me it appears that the experiments do not furnish ground for a safe conclusion as to the magnitude of the same quantity of work. Although it thus appears that, on account of the unknown quantities w^hich enter into the production of the total action, it is impossible to demonstrate an exact quantitative coincidence between the equation (21) and our principal proposition, we could perhaps try the inverse method, and assuming both, and their connexion with each other, to determine the said unknow'n quantities, or at least their sum ; indeed the form of the equa- tion seems to invite such an effort. It must however be remem- bered that the equation itself, being an empirical one, must not have absolute accuracy ascribed to it, as is also shown by the numbers of Riess. In two series of experiments in which he introduced w ires of different lengths and thicknesses, thus chan- ging only the quantity / in the expression on the right side of equation (21), he determined from each observation the con- stant b. The values so found diverge from each other in the first series between 0*01358 and 0*01101, and in the second series between 0*00000926 and 0*00000840 1; and although these differences, considering the great variety of the wires introduced, cannot be regarded as considerable, still they are worthy of re- gard, inasmuch as a certain regularity is observed in them. In both series, when the reduced length I of the wire is increased, the corresponding worth b is in general smaller. We will therefore follow this subject no further, and turn now * Pogg. Ann. vol. xliii. p. 79. \ Pogg. Ann. vol. xliii. pp. 68 and 73. The great difference between the numbers of the first and second series is due to a difference in the unit made use of. OF AN ELECTRIC DISCHARGE. 21 to the second point of comparison between theory and expe- rience, namely, to the case where the connecting wire remains the same, but the magnitude of the battery and the quantity of electricity distributed over it are changed. Here also the difficulty spoken of lies in our way. As we are ignorant of the nature of a portion of the action of the discharge, we are unable to say how this portion changes with the mag- nitude of the battery and the quantity of electricity, and cannot, therefore, from the effects observed at a particular place in the circuit, conclude with security as to the totality of the action. Only in regard to the heat developed in the continuous portion of the connecting wire can we with certainty assume that every alteration observed in any one portion is proportional to those which take place in all other portions. Now when the connecting w4re has a great reduced length, it may certainly be assumed that the greatest part of the total action is applied to heating it; and in this case also, if a diver- gence from the said proportionality should exhibit itself, the differences caused by it will be comparatively small, so that without any considerable inaccuracy we may assume that the heating observed at any portion is proportional to the corre- sponding total action. Now the total action, according to equations (18) and (19), for an incomplete discharge is represented by ^ . const., and for a complete discharge, such as took place in the expe- riments of Riess, according to (18) and (20), by — __. const.; b and this is the exact expression which Riess established experi- mentally for the heating in the connecting wire, inasmuch as the equation (21) completely expressed*, is "-^uv <-' where a is a constant quantity f. * Pogg. Ann. vol. xlv. p. 23. t This coincidence between theory and experiment is also adduced by Helmholtz (his Memoir, p. 43) ; the development of his formula is not however 22 CLAUSIUS ON THE MECHANICAL EQUIVALENT The cases heretofore considered referred to the incomplete discharge of an ordinary battery. We will examine only two other cases, namely, the incomplete discharge of an ordinary bat- tery, and the complete discharge of the so-named cascade battery of Franklin, With reference to the first, we possess experimental measure- ments by Riess*j who partially discharged a charged battery by bringing both its coatings into connexion with one another ; so that the electricity which was formerly confined to one, now distributed itself over two. He changed the experiment by making both batteries to consist of different numbers of jars, and observed each time the heating of the one or the other con- necting wire. The jars of each battery were all equal, but un- fortunately the jars of one battery were different in size from those of the other. As the result of his observations, he states that all the quantities of heat generated at a constant place in the inner and in the outer connecting arc are completely ex- pressed by the following formula f : — for the sake of more easy comparison with my other formulas, I have here chosen different letters from those chosen by Riess. C expresses the observed heat, Q, the quantity of elec- tricity applied, s the superficial content of the inner coating of a jar of the first battery, and n the number of these jars ; 5' and n^ the same quantities for the other battery, and finally a a con- stant, which is somewhat larger for the inner connecting arc than for the outer ; this is explained by the fact that a little more elec- tricity was distributed over the inner coating than over the outer. We will now compare this heating with the increase of the potential. From equation (1 8) we derive as potential of the first battery before the discharge, when the quantity of electricity is denoted quite clear to me, inasmuch as he introduces a quantity which he names the quantity of work, and concerning which he says that it is proportional to the coated surface of the battery, without however explaining its meaning further, or giving any reason for this proportionality. * P"gg' •^"M- vol. Ixxx. p. 214. t l*ogg- Ann, vol. Ixxx. p, 217. OF AN ELECTRIC DISCHARGE. 23 by Q, and when for the total superficial 'content S its value ns is set, the expression W=-^'-^ (23) ns 2 ^ ^ In order to determine how the entire quantity of electricity Q is distributed by the discharge over both batteries, we know the condition, that on the coatings which are connected together the potential functions must be equal. Let Vj and V/ be the po- tential functions on the inner coatings after the discharge, and Qj, Q/ the quantity of electricity upon them, which it is our object to ascertain; we have then, according to (17)> v.'^ n's' ' where k' is the same magnitude for the second battery that k is for the first. If these expressions be set down as equal, and bearing in mind that ns we obtain ns n's (24) Q,' = . T + From this we obtain further, when W^ denotes the entire po- tential of both batteries after the discharge. ia« k k' and thus we obtain the increase of the potential, 1^2 s' Wi-W=: 2k''s .Q^ (>?7h (25) (26) The quantity -^ .- is constant for the entire series of expe- ji K S 24 CLAU8IUS ON THB MECHANICAL EQUIVALENT riments, and hence we can write If this expression be compared with that found by Riess for the heating (22), it is seen that to make both proportional it is only necessary to assume that in the jars of both batteries, although they were not equal, the quantities k and k' possess nearly the same value ; and this assumption is peculiarly jus- tified by the fact that Riess found by direct measurement, that when they were combined the electricity distributed itself over the batteries in proportion to their surfaces, and this, according to (24), could only be the case when k was =A;'*. Riess altered his experiments so that he increased the length of the connecting wire, and observed the consequent diminution of heat at a certain place. The results of these observations coincide in general with those already mentioned, and we will therefore pass over them here, and also some other experiments described in the same paper. In regard to the cascade battery, we possess experiments from Dovef and Riess J. It consists, as is known, of a number of single jars or complete batteries, insulated and so connected with each other, that the exterior coating of the first is connected with the interior coating of the second, the exterior of the second with the interior of the third, and so forth. Only the inner ♦ As the quantities k and k', according to the above, depend principally upon the thickness of the glass of both batteries, it appeared to me of interest to ascertain these thicknesses ; while this Memoir was still in the press, 1 therefore requested M. Riess to make the measurement, and am indebted to his kindness for the following communication. In the small jars (those of the second bat- tery) the thickness of the glass varies considerably, its mean thickness being 1^ Paris line. The large jars (those of the first battery) he was unable him- self to measure as they were closed above, and he has therefore measured two othorg which he had made at the same time with the former, and with the in- tention of being used along with them. The glass in these is nearly imiform and 14 line thick. As an absolute equality of the glass thicknesses was not to be expected under the circumstances, and as it is not a necessary consequence of the assumed equality of k and W, inasmuch as the latter are also to a certain, although subordinate extent, dependent on the shape and magnitude of the" jars which are different in both batteries, I believe that the coincidence of the numbers 1^ and 1^ is sufficiently exact. f Pogg. Ann. vol. Ixxii. p. 406. \ Pogg. Ann, vol. Ixxx. p. 349. OP AN ELECTRIC DISCHARGE. 25 coating of the first and the outer coating of the last are free ; and, in charging, these are treated as if both surfaces formed a single battery. The principles according to which the charge of such a com- pound battery must be determined, are for the most part already given by Green*. Let the quantities of electricity distributed over the two coatings of the single batteries, and the correspond- ing potential functions be expressed in order by — Qi, Q\ ; Q2, Q'2; Q3, Q'3 &c,l . . Vi,V\; V„V',; V3,V'3&c.J • • • ^ ^ Now if positive electricity be communicated by a conductor to the inner coating of the first battery, the outer coating of the same can only receive its negative charge from the interior of the second battery, which will thus become positively charged ; we therefore have Q',= -Q,; and further, as in two bodies which are connected by a con- ductor the potential functions are equal, we have for the same two coatings and two equations exactly similar apply to every other pair of connected coatings, so that the following series of equations is given, Q\= -Q„ a',= -Q3, Q'3= ~Q, &C.-1 . v\= v„v',= V3,v'3= v,&c.J • ^'^r Besides this, the four quantities Q, Q', V and V, stand for each of the batteries in such a relation to each other that from every two of the same the two others may be determined. In accordance with the remark at the conclusion of a former paper of mine, if we denote the said four quantities by q, g', V and V, the equation (15), whieh applies to the case where V'=0, can, if we neglect the members of a higher order in respect to k, be changed into the following more general one, V-V=-^£=*:, ...... (30) where instead of ^-— ^ we may also write q^ or — tj. * An Essay on the Application of Mathematical Analysis to the tbeoriea of Electricity and Magnetism, kx\, 8. 26 CLAUSIUS ON THE MECHANICAL EQUIVALENT Further, we have quite generally the equation j + g'=-(aV + ^V'), (31) wherein a and j3 are two positive constant quantities, which, like k, depend upon the nature of the jar, and hence in the case be- fore us, where we have simply to do with equal surfaces, possess throughout the same value. For a battery of n jars, and with the quantities Q and Gl' of electricity, both equations pass into V-V'=-l®^' ...... (32) ns 2 ^ Q + Q'=-w(aV + ^V') (33) By means of these equations, in connexion with equation (29), if any two of the quantities (28) be given, all the others may be found. The experiments conducted by Dove and Riess consist in both cases of two different series. In the first, the number of jars in all the connected batteries was the same, but the number of batteries was varied ; in the second, on the contrary, the number of the batteries made use of remained constant, namely, always two, but in each of these batteries the number of jars was varied. Both series of experiments present many difficulties to a com- parison with the theory. In the second series, however, these are less considerable than in the first, and we will therefore com- mence with the consideration of the second. The experiments were arranged as follows: — the batteries were both insulated, and the inner coating of the first w^as connected with the conductor of the electric machine, the outer of the second with a Lane^s unit jar. The number of sparks from the unit jar gave the quantity of electricity communicated to the second outer coating, and we can set the potential function upon this coating after the passage of each spark equal to zero, the potential function of the quantity remaining in the unit jar being neglected. There are therefore, as above required, two of the quantities (18) known, and to deduce the others from these we can proceed from the second outer coating to the second inner one, to the first outer, and finally to the first inner coating. If the quantity of electricity as measured by the unit jar be — Q, and thenum- OF AN ELECTRIC DISCHARGE. 27 ber of jars be denoted by n^ and n^^ all members which, in regard to k, are of a higher order than the first being neglected, we obtain the following series of expressions : — Q'=-Q V'=0 ■<2— ^ " 2" ^(34) a^=-(i+/-)a Y^=-±a The potential of the whole compound battery is W=i(Q,V, + Q',V', + Q5V, + Q',Vg; . . (35) and this gives W = -{l..[2«+(« + /3)g-*}(i-fi-)±Q- . (36) or, neglecting the member of the second order in respect to k, W — {iyn)>' ^''^ As after the discharge the potential is equal to zero, — W is the increase of the potential due to the discharge ; and if, as before, we assume that, other things being the same, the heating at any single place in the connecting wire is proportional to the total action, we can write '' trachyte ... J Silica Alumina and prot- oxide of iron. . . Lime Magnesia Potash Soda 0*000 76-67 14-23 1-44 0-28 3-20 4 76-00 14'6l 1 0-43 3-14 4-14 0-106 74-00 15-73 2-43 0-91 2-96 3-97 100-0 0*199 0-310 72-00 16-86 3*17 1-38 2-78 3-81 70-00 18-00 3-90 1 84 100-0 100-0 I 0-444 0-609 68-00 1913 4-65 2-31 2-42 3-49 66-00 20-26 5-38 2-78 2-23 3-36 64-00 21-39 6-13 3-25 2-05 3*18 100-0 1-084 1-446,1*959 -00,60-00.58-00 22-52 23-65 24-78 6-87 3-72 1-87 3-02 100-0 7-61 4-19 1-69 4-65 1-51 2-86, 2-71 ,100-0 lOO'O I I 2-745 4-099 -00 54-00 52-00 50-00 25*92 27-04 '28- 17 29*29 9*08 9'83 10-5711-31 17-43 5-12' 5-69 1*33| 1*15 2-55 2-39 6-06 6-53 0-96 0-79 2*24 2-08 100-0 1000 100-0 100-0 I I I 48-47 30-16 11-87 6- 0- 196 100-0 If these numbers should be found to correspond with the composition of all the original varieties of Icelandic rocks which have not yet undergone a chemical metamorphism, it may be considered as proved that they are all either merely mixtures of the above-mentioned acid and basic extreme members, or those members themselves ; and that the great mineralogical and pe- trographic ditferences which these rocks present, are only con- sequences of the varying proportions in which the mixture took place, and of the prevailing physical conditions under which the rocks reached their present situation and assumed their present form. Among the large number of analyses of Icelandic rocks which have been carried out in my laboratory, there is not one in which the results differ from the calculated composition ob- tained in the above-mentioned manner, to a greater extent than might reasonably be expected in such a calculation, based merely upon the mean results of analyses. The following experimental and calculated results were obtained for a granular whitish gray trachyte, containing hornblende and orthoclase crystals from Oexnadalr in the Norden Islands ; — 42 BUNSEN ON THE FORMATION OF 31. Found. Calculated. Silica 73-57 73-57 Alumina and protoxide of iron . 17*19 15-89 Lime 1-41 2-58 Magnesia 0-81 I'Ol Potash 2-19 2-92 Soda 4-83 3-93 10000 10-00 The calculation requires that there should be 0-1325 pyroxenic mass to 1-0 of trachyte. Another compact gray pyroxenic rock from Reyjadalr Foss, a small waterfall which is formed by the Nordhra near Hvammr, shows a still greater correspondence. This rock was either massive or separated by columnar fissures, and contained small zeolitic druses filled with green earth. 32. Found. Calculated. Silica 51-75 51-75 Alumina and protoxide of iron . 28-39 28-31 Lime 10-49 10*65 Magnesia 5-90 6*13 Potash 1-01 0-96 Soda 2-46 2-20 100-00 100-00 The calculation requires 7*597 pyroxenic mass to 1*0 trachyte. A compact blackish gray rock, with an indistinct angular frac- ture in small pieces, and a conchoidal fracture in large pieces, presenting a very uniform crystalline granular appearance, and occurring in the neighbourhood of Kalmanstijnga, which affords particularly good opportunities for the study of the mixed rocks, consisted of — 33. Found. Calculated. Silica 53-08 53-08 Alumina and protoxide of iron . 28-70 27*57 Lime 9*92 10'16 Magnesia 5-32 5-81 Potash 0-61 1*06 Soda 2-37 2-32 100-00 100-00 THE VOLCANIC ROCKS OF ICELAND. 43 The calculated relation of trachyte and pyroxenic masses in this rock is 1*0 to 5-117. Experiment and calculation gave for a trap (clinkstone ?) from EskiQordhr, analysed by Damour, 34. Found. Calculated. Silica 66-12 66*12 Alumina and protoxide of iron , 24*05 20*20 Lime 3*24 5*34 Magnesia 0*46 2*76 Potash 1*29 2*24 Soda 4*84 3*34 100*00 100*00 This rock represents a mixture of 1*0 trachyte and 0*5991 trap mass. A clinkstone from Klettaberg near Kalmanstunga, in sepa- rate slabs, consisted of — 35. Found. Calculated. Silica 73-37 73*3? Alumina and protoxide of iron . 17*26 16*09 Lime 2*49 2*66 Magnesia 1*52 1*05 Potash 3-01 2*90 Soda 2*35 3*93 100*00 100*00 The relation of the normal trachytic and pyroxenic masses is in this instance 1 : 0*1325. A great number of other analyses of this kind, which some- times showed a greater difference and sometimes a still greater correspondence between the experimental and calculated results, lead to the same conclusion. But even if these examples should leave any doubts of the existence of an intimate connexion between the various unal- tered rocks occurring in Iceland, such doubts must be entirely removed by the evidence presented in favour of such connexion by a great number of trachytic and basaltic dykes. It is here very frequently possible to detect the intermixture of the extreme acid and basic members, not only by means of chemical analysis, but by direct observation. In one of the south-eastern valleys of 44 BUNSEN ON THE FORMATION OF the Esja mountains, opposite Mossfell, the conglomeratic pyro- xenic rock is traversed by a trachyte dyke, which consists in its interior of a pure white mass, becoming gradually darker and more ferruginous towards the surrounding rock. This is but one among many examples which might be brought forward. The analysis of the mass from this dyke, without taking into ac- count the water which it contained, in consequence of especial circumstances, gave for the part in the centre of the dyke (36), for that situated nearer to the saalband (37), and for that part of the rock immediately surrounding the dyke (38) : — 36. 37. 38. Silica 78-95 66-18 50*25 Alumina .... 7*71 9*74 1255 Protoxide of iron . 4*32 1205 16-13 Lime 155 449 11-10 Magnesia .... 0*42 3*04 7*59 Potash 2-48 0*94 0*34 Soda 4-57 3-56 2-04 10000 100-00 100-00 A comparison of these results with the above average compo- tions, shows that the interior mass of the dyke has almost ex- actly the composition of the pure normal trachytic rocks, while, on the contrary, the surrounding mass has the composition of the normal pyroxenic rocks ; and the part situated near to the surrounding rock is an admixture of both in the proportion of 0*5923 pyroxenic mass to 1*0 trachyte, for which the following calculated composition closely corresponds with that found by actual analysis : — ^ 39. Found. Calculated. Silica 66-18 66*18 Alumina and protoxide of iron . 21*79 20*15 Lime -K^> 4*49 5-32 Magnesia ; 3*04 2*74 Potash 0-94 2*25 Soda 3-56 3*36 100-00 100-00 Upon closer examination, indeed, the most distinct indications are found of a lateral fusion of the neighbouring rock and the THE VOLCANIC ROCKS OF ICELAND. 45 substance of the dyke, gradually decreasing from the surface of contact towards the central part of the dyke, and to which they do not extend. These observations and experiments prove that it is not necessary to regard the chemically intermediate mem- bers between the extreme acid and basic members of the Ice- landic group of rocks as resulting from simultaneous eruptions from the two great plutonic foci, but that even the already formed rocks may, during eruptions or injections, have furnished one part of the material for the formation of mixed rocks. A confirmation of this law is likewise met with in applying it to the rock formations upon the lava streams of the Icelandic volcanoes. The final members, and the intermediate rocks re- sulting from their mixture, are not absent even among these most modern formations. The enormous lava streams which have rent the palagonitic ranges of hills of Krafla and Leirh- niikr, afford the most striking evidence of this, and Hecla not less so. However, the law is in this instance not so distinctly recognizable, because the volcanic mountains penetrated by these lava streams consist of metamorphic products of the pala- gonitic tuff, which, although it corresponds in composition with the normal trap mass, is subject to greater variation in the pro- portions of its constituents. The discrepancies accounted for by this circumstance were rendered still greater by the impos- sibility of obtaining for examination any other than hand spe- cimens, which contained separate and not very uniformly mixed felspathic masses. But notwithstanding these disadvantages, the law is still found to obtain. The blackish gray stony lava of Hrafntinnuhryggr (40) on the south-eastern foot of the Krafla, and the obsidian (41) al- ternating within beds, have exactly the composition as the normal trachytic mass. 40. 41. Silica .... 75-12 75-28 Alumina . . . 11-34 10-22 Protoxide of iron 3-92 4-24 Lime .... 1-73 1-81 Magnesia . . . 0-39 0-25 Potash . . . 1-85 2-44 Soda .... 4-39 5-53 Water .... 0-41 0-23 100-00 100-00 46 BUNSEN ON THE FORMATION OF The streams of obsidian occurring on the north-eastern de- clivity of Hecla approach very closely in composition to the above. They represent a mixture of 1*0 trachyte and 0*2325 pyroxenic substance. 42. Found. Calculated. Silica 71-35 71*35 Alumina and protoxide of iron . 17'33 17*32 Lime 1-24 3-41 Magnesia 0*19 1*53 Potash 4-23 2*73 Soda 5-66 3'75 100-00 100-00 The extreme basic member is represented in the substance of the great lava streams which has flowed from this volcano in a W.N.W. direction to the banks of the Thjorsa. 43. Found. Calculated. Silica 49-60 48-47 Alumina and protoxide of iron . 28-81 30-16 Lime 13*07 11*87 Magnesia 7-56 6-89 Potash 0-20 0-65 Soda 1*24 1-96 100-00 100-00 The lavas from the western foot of Hecla, examined in my laboratory by Dr. Genth, are represented as resulting from the mixture of the two extreme members by the following analyses: Lava at Hals. Efrahvols lava. Hecla ash of 1845. 44. 45. 46. Found. Calc. Found. Calc. Found. Calc. Silica .... 55-95 55-95 59*45 59-45 56-76 56*76 Alumina and prot- oxide of iron .29-83 25-93 27-68 23*96 27*47 25*48 Lime .... 6-54 9*10 5-50 7*80 6*75 8*79 Magnesia . . . 4*21 5*14 2*38 4*32 4*04 4*95 Potash .... 0-96 1*33 1*48 1*65 2*63 1*40 Soda .... 2-51 2-55 3-56 2*82 2*35 2*62 100*00 100-00 100-00 100*00 100*00 100*00 1 trachyte+2-77 1 trachyte+ 1*568 1 trachyte-|-2-402 pyroxenic mass. pyroxenic mass. pyroxenic mass. THE VOLCANIC ROCKS OF ICELAND. 4? The order of the analyses corresponds with the succession of the streams examined, from the more ancient to the younger. They show that the flow from the two volcanic foci which main- tain the activity of the volcano, is as irregular as this activity itself. On the formation of the lava streams of Thjorsa the pyroxenic focus alone acted, and in the formation of that at Hals the action of the pyroxenic focus preponderated over that of the trachytic. The contrary relation is observed in the Efrahvols lava; and during the most recent eruption in the year 1845, the action of the pyroxenic focus again preponderated over that of the trachytic. The great resemblance which is everywhere found to exist between the rocky masses poured out from volcanoes, jus- tifies the conjecture that these processes of rock formation are not limited to Iceland alone. A number of analyses of analo- gous kinds of rocks from other localities, executed in my labo- ratory, prove that this is really the case. The volcanic system of the high table-land of Armenia is in this respect especially remarkable. The mean results of analyses of rocks from that locality, for which we are indebted to the excellent researches of Abich, likewise give for the extreme acid member occurring there precisely the same composition as the Icelandic. Normal trachytic The same from the high table- substance of lands of Armenia *. -^ Iceland. 47. 48. 49. 50. Silica 76-67 77*27 77*60 77*42 76-66 Alumina and prot- oxide of iron Lime . Magnesia Potash . . Soda . . . Loss by ignition This resemblance is recognizable, however, not only in the extreme acid member of the series, but is still more distinct 14-23 14-14 13-72 14-83 15*17 1-44 1-31 1-40 2-73 1-25 0-28 ? ? ? ? 3-20 2-44 2-301 . .^^ 2*94 ^\ 4-32 4-18 4-15 4-21J 3-52 000 0-51 0-61 0-57 1-12 * Ueber die geolog. Natur des Armen. Hochlandes, von Dr. H. Abich, Dorpat, 184.3. 47. Brown obsidian from the Lesser Ararat. 48. Obsidian porphj'ry from the Greater Ararat. 49. Obsidian from Kiotangdag. 50. Dioritic porphyry from Besobdal. 48 BUNSEN ON THE FORMATION OF in the mixed rocks. A compact black basalt from the source of the Euphrates, which was analysed by M. Jackel in my labora- tory, represents a mixture of 1 trachytic and 0'7332 pyroxenic substance. 51. Found. Calculated. Silica 64-76 6474 Alumina and protoxide of iron . 22*31 20*99 Lime 5*13 5*84 Magnesia 1*91 3*07 Potash 2-51 2-12 Soda 3-38 3*24 100-00 100-00 All the other analyses which Abich gives in his paper likewise correspond in a surprising manner with the law : — Rock at the summit of the Greater Ararat, 52. Found. Calculated. Silica 70-14 70-14 Alumina and protoxide of iron . 18*17 17'92 Lime 4-72 3*85 Magnesia . l-OO 1-81 Potash 1-47 2-61 Soda 4*50 3*67 100-00 100*00 1 trachyte and 0'3013 pyroxenic mass. Dark Gray Rock from the summit of the Kasbeck. 53. Found. Calculated. Silica 69-77 69*77 Alumina and protoxide of iron . 18*27 18-13 Lime 5-13 3*99 Magnesia 1*65 1-89 Potash and Soda 5*18 6*22 100-00 100-00 V , 1 trachyte and 0*3239 pyroxenic mass. THE VOLCANIC lloCKS OF ICELAND. 49 Reddish-brown Rock from the summit of the Kasbeck, 54. Found. Calculated. Silica 70-97 70*97 Alumina and protoxide of iron . 18*13 17*45 Lime 4*24 3*55 Magnesia 1*54 1*62 Potash and Soda 5*12 6*41 100-00 100-00 * ^ / I trachyte and 0*2533 pyroxenic mass. Rock from the summit of the Elbruz, 55. Found. Calculated, Silica . 69-65 6963 Alumina and protoxide of iron . 19-85 18*25 Lime 4-40 4*02 Magnesia 2-27 1*92 Potash and Soda 383 6-18 100-00 10000 1 trachyte and 0*3314 pyroxenic mass. Rock from Ararat, 56. Found. Calculated. Silica 65-96 65-96 Alumina and protoxide of iron . 22*18 20*28 Lime 4*27 5 39 Magnesia 2*13 2*79 Potash 1*34 2-23 Soda 4*12 3*34 10000 10000 *. ^ ' 1 trachyte and 0*6124 pyroxenic mass. Compact Porphyritic Rock from Ai'arat, 57. Found. Calculated. Silica 65-27 65*27 Alumina and protoxide of iron . 20*88 20-67 Lime 657 565 Magnesia 3-47 2*96 Potash and Soda ...... 3*81 5*45 100*00 100-00 V I 1 trachyte and 0*6786 pyroxenic mass. SCIEN. UEM.^Nat. Phil Vol. I. Part I. E 50 BUNSEN ON THE FORMATION OF A somewhat decomposed porous Rockfro^n Ararat. 58. Found. Calculated. Silica 65-39 65 39 Alumina and protoxide of iron . 22*65 20*61 Lime 7*40 5*61 Magnesia 3*00 2-92 Potash and Soda 1*56 5*47 100*00 100-00 V ^ > 1 trachyte and 0*6666 pyroxenic mass. Unknown Crystalline Rock from between Keschet and Kobi, 59. Found. Calculated. Silica 61*25 61*25 Alumina and protoxide of iron . 25*72 22*95 Lime 6*27 7*14 Magnesia . 3*77 3*89 Potash and Soda 2*99 4*77 100*00 10000 V ^ > 1 trachyte and 1-207 pyroxenic mass. Abich mentions only one variety of rock which approximates nearer in its composition to the normal pyroxenic substance. This is an amygdaloid basalt from Ararat, which may be regarded as consisting of 1 trachyte and 3*427 pyroxenic substance, thus: 60. Found. Calculated. Silica 54*84 54*84 Alumina and protoxide of iron . 27'96 26*56 Lime 9*28 9*51 Magnesia 3'72 5*39 Potash and Soda * 4*20 3*70 100*00 100*00 While, therefore, on the one hand, the composition of the Cau- casian rocks may be deduced (their per-centage of silica being known) from the composition of the Icelandic rocks, the Cau- • The alkalies are here estimated from the deficiency. THE VOLCANIC ROCKS OF ICELAND. 51 casian rocks, on the other hand, afford a possibility of ascertain- ing, by means of calculation from the extreme acid members which concur in conjunction with the accompanying basic sub- stances, the composition of the pyroxenic substance which filled the non-trachy tic focus of volcanic rock-formations in Ice- land. The various constituents of such a substance may be easily ascertained by means of the following equation derived from (2), a -Pny in which A„ represents, in per cents, the separate constituents of a mixed rock. This formula gives, for reasons which may readily be understood, results which are more accurate, in proportion as An differs more from the numerical value of the normal tra- chytic composition. This is the case only with the analysis (60). If therefore from this the value of Pn is calculated, we obtain — Caucasus. Iceland. Calculated. Found. Silica 48-47 48-47 Alumina and protoxide of iron . 31*97 30*16 Lime 11*56 11*87 Magnesia 4*72 6*89 AlkaUes 3-28 2*61 100-00 10000 It cannot therefore be doubted that both the extensive vol- canic elevations constituting the high table-land of Armenia and the island Iceland have flowed from sources which were chemi- cally identical. The idea that perhaps all the volcanic forma- tions on the earth's surface have originated from the same source, or even, indeed, that it is from this very source that all the ferrugi- nous and non-ferruginous plutonic rocks have originated by fu- sion together, is rendered the less improbable by the fact that the mineralogical differences existing between those Caucasian and Icelandic rocks which present the same mean composition, are not less marked than those observed among other ferruginous rocks of plutonic origin. It would therefore be very interesting to trace the genetic relations existing among the more ancient E2 52 BUNSEN OiN THE FORMATION OF formations, in a manner similar to that which I have adopted with regard to the volcanic rocks. It would be necessary to select for this purpose only the more extensive and largely deve- loped rock systems, in order to avoid those disturbing influences which may have resulted from the contact of chemically meta- morphosed exogenous rocks with those of plutonic origin. These disturbing influences manifest themselves even where volcanic rocks have penetrated in isolated elevations, through calcareous or siliceous sedimentary beds. It still remains, after the above consideration of this subject, to point out the connexion which exists between the mode of formation of Icelandic and Caucasian rocks just described, and their mineralogical composition. This investigation affords the most interesting results, but it would not be compatible with the limits of this abstract to trace these various relations. Their consideration must therefore be deferred for the present. II. Genetic Relations of the Metamorphic Rocks, 1. Palagonitic Rocks. The metamorphic rocks, which constitute a by no means in- considerable part of Iceland, are far more interesting than the unaltered rocks. The most remarkable member of this group is certainly the palagonitic tuff, which consists of a mixture of anhydrous and hydrated silicates. The anhydrous silicates be- long exclusively to the pyroxenic rocks treated of above, and are never accompanied by trachytic masses, or even replaced by them ; the hydrated silicates, which generally cement together the fragmentary rock, may, on the contrary, be regarded as a mixture or combination of two silicates, of which one is repre- sented by the formula 3R0 2SiO^ + Aq, and the other by 3 AP O^, SiO^ + Aq. These silicates appear to combine together in definite proportions ; at least I consider it necessary to regard palagonite, for which I have proposed the formula 3R0 2SiO^ -|-2APO^SiO^ + Aq, as well as a mineral occurring in the tuffs of Chatham island, one of the Galapagos, which I found to con- sist of 3RO 2Si03 + A12 O^ Si08 -f Aq, as being such com- binations. THE VOLCANIC ROCKS OF ICELAND. 53 Without here entering into any special examination of the various members of the tulF rocks, it may be as well to mention that the palagonitic substance appears to occur everywhere as a characteristic constituent of these formations when the pyro- xenic rocks of the volcanic period are especially well developed. Besides Iceland, where it occurs in extensive masses, it is met with in the more considerable basaltic elevations of Germany and France, in the Euganean islands, on Etna, the Azores, Canaries, Cape Verde Islands and the Galapagos Islands, pro- bably also upon the volcanic islands of the South Sea. The following analyses will give an idea of the degree of uniformity in the chemical composition of this very generally occurring cementing substance of the volcanic tuff: — Iceland^ Seljadalr. Silica . , . . Alumina . . . Protoxide of iron Lime . . Magnesia . Potash . . Soda . . Water . . Foreign residue Oxygen found. 19-43 9-47 5-12 Oxygen calculated. 19-44 9-72 4-86 100-17 Iceland, Trollkonugil near Hecla, 62. Oxygen found. Oxygen calculated. Silica .... 39-98 '21-16 19-96 Alumina ... 8*26 1 ^ Protoxide of iron 17*65 J 9-98 Lime .... 8*48 "^ Magnesia . . . 4*45 1 ^ Potash .... 0-43 [ ' 4-99 Soda .... 0*61 ^ Water .... 18-25 Foreign residue . 1-89 100-00 54 BUNSEN ON THE FORMATION OP Iceland, Palagonitic Sandstone near Reykjahildh. 63. Oxygen found. Oxygen calculated. . . 35-09 18-57 18-31 . . 10-60 Silica .... Alumina . . . Protoxide of iron 13*65 Lime .... 4-83 Magnesia . . . 7*07 Potash .... 0-25 Soda . . ^ . 0-50 Water .... 17-25 Foreign residue . 11-13 } 9-05 4-11 9-16 4-58 100-00 Iceland, Laugarvatnshellir, 64. Oxygen found. Oxygen calculated Silica . . . 40-38 21-37 20-38 Alumina . . Protoxide of iron • 1^-^^ \ 9-10 13-52 J 1019 Lime . . . . 8-56 ^ Magnesia . . Potash . . . . 6-35 . 0-64 519 5 09 Soda . . . 0-61 J Water . . . 16-98 Foreign residue 2-32 100-16 Iceland, Krisuvik, 65. Oxygen found. Oxygen calculated Silica . . . 37-95 20-09 20-60 Alumina . . Protoxide of iron ■ S?l } '»■« 10-30 Lime . . 6-48 -| Magnesia . . Potash . . . ^•^^ I 5-17* 0-42 1 ' 515 Soda . . . 1-72 - Phosphoric acid 0-43 Water . . . 12-68 Foreign residue 7-25 100-419 After deducting the oxygen belonging to the phosphates and carbonates. THE VOLCANIC ROCKS OF ICELAND. 55 Silica . . . . Alumina . . . Protoxide of iron Lime . . Magnesia . Potash . . Soda . . Water . . Foreign residue Iceland, Naefrholt near Hecla, Oxygen found. 17-39 8-47 5-03 Oxygen calculated. 17*65 8-83 4-41 100-42 Iceland Tuff, containing Fossils, from Folsvogr, 67. Oxygen found. Oxyg en calculated Silica . . . . 28-53 15-03 15-20 Alumina . . . Protoxide of iron rs } '- 7-60 Lime . . . . 6-02 ^ Magnesia . . . Potash .... 5-60 0-96 >- 4-34 - 3-80 Soda . . . . Water .... 0-84 -> 7-61 Foreign residue . 31-05 99-27 Iceland Detritus from the Lava near Hruni, 68. Oxygen found. Oxygen calculated Silica . . . 37*11 19-64 18-74 Alumina . . . 9*78 "1 Protoxide of iron 14-67 J 8-97 9-37 Lime . . . 4-99 -^ Magnesia . . 5-61 > 3-93 4-68 Potash . . . 1-57 Soda . . . . 0-00 Water . . . . 14-04 Foreign residi le . 12-24 10000 56 BUNSKN OX THE FORMATION OF Galapagos, Dyke Mass. 69. Oxygen found. Oxygen calculated. Silica .... 37-83 1964 19-24 Alumina . . . 12-95 \ ^.^3 g.^g Protoxide of iron 9-93 J Lime .... 7-49 -| Magnesia . . . 6-54 I ^,^ ^.^^ Potash .... 0-94 f Soda .... 070 J Water .... 23-00 Foreign residue . 0-96 100-.34 Galapagos, Rock forming a Crater. Silica . . . . 70. 3615 Oxygen found. 18-77 Oxygen calculated. 18-31 Alumina . . . Protoxide of iron 11-^1 1 8-43 10-48 J 915 Lime . . . . 7-78 ^ Magnesia . Potash . . . . 6-14 0-76 '> 4-85 4-58 Soda . . . . 0-54 J Water . . . . 24-69 Foreign residue . 219 100-00 The analysis of the cementing matter of tuffs, occurring upon the Cape Verde Islands, the Azores, Canaries, and in the basaltic formations of Germany, yielded results corresponding perfectly with the above. The frequent occurrence of these rocks, and their alternation with the unaltered volcanic rocks by which they have been penetrated, indicates that the palagonitic tuff rocks have essen- tially contributed to the formation of the trachyto-pyroxenic mixed members of this group, and a closer examination of the phaenomena of plutonic contact proves that such is really the case. The influence of the tuffs upon the formation of these rocks consequently demands an especial investigation. The fol- lowing analyses give the composition of the varieties of tuff THE VOLCANIC ROCKS OF ICELAND. 57 examined^ calculated for anhydrous substance and protoxide of iron : — Laugar- Selia- Troll- Gala- Reykm- vatn- Krfsu- Naefr- Fosb- dair. konugil. pagos. hildh. shellir. vik. holt. vogr. Laxa. Silica 48-29 M-20 5018 4969 6071 4763 46-29 47-78 51-36 Alumina 14-41 10-58 1718 15-01 13-55 1708 10-30 15-55 13-53 Protoxide of iron... 16-47 2034 11-85 1740 1544 15-53 21-30 1417 18-27 Lime 11-31 1085 9-94 1001 1075 814 958 1009 6-91 Magnesia 7-79 5-70 8-68 6-83 789 8-95 8-64 9-39 776 Potash 0-88 0-55 0-93 0-35 0-81 0-52 111 1-61 217 Soda 0-85 0-78 1-24 0-71 076 2-15 2-78 1-41 000 100-00 10000 100-00 100-00 100-00 100-00 100-00 10000 100 00 Oxygen in SiO^ and the bases 3:2-032 3: 1-703 3 : r992 3: 1-906 3:1-835 3:2-178 3:2-039 3:2*099 3:1-746 Silica 48-29 51-20 5018 49-69 50-71 47-63 4629 4778 51-36 Alumina and prot- oxide of iron 30-88 30-92 29-03 32-41 28-99 32-61 3160 29 72 31-80 Lime 11-31 10-85 9-94 1001 10-75 814 958 1009 6-91 Magnesia 779 5-70 8-68 6-83 7*98 8-95 8-64 939 7'76 Potash 0-88 0-55 0-93 0-35 0-81 0-52 Ml 161 217 Soda 0-85 0-78 1-24 0-71 0-76 215 2-78 1-41 000 100-00 100-00 100-00 100-00 100-00 10000 100-00 100-00 100-00 These analyses lead to the unexpected result, that the sub- stance of the melted palagonitic silicates almost exactly corre- sponds in its composition with the normal pyroxenic rocks, with the single exception of a somewhat greater variation here and there in the relative proportions of the several constituents. In fact, the following mean composition, deduced from all these ana- lyses, scarcely shows any perceptible difference from that of the normal pyroxenic rocks, when we overlook the trifling excess in the quantity of magnesia and the somewhat smaller quantity of lime : — Normal pyroxenic Palagonite. substance. Silica 49-24 48-47 Alumina and protoxide of iron . 30-82 30-16 Lime 9-73 11-87 Magnesia 7*97 6-89 Potash 0-99 065 Soda 1-34 1-96 100-00 10000 The proportion of the oxygen in the acid to that in the bases is for palagonite as 3 : 1*948, and for the pyroxenic rock as 58 BUNSEN ON THE FORMATION OF 3 : 1'998, which may be regarded as identical ; and these numbers certainly differ less than the results of the several analyses from which these mean values are deduced. For this reason an almost uniform composition, frequently corresponding more closely with the calculation than when the normal pyroxenic substance is applied to this purpose, is obtained when the normal palago- nitic substance is made the basis of the calculation. But the cir- cumstance that the composition of these palagonitic substances deviates rather more from the mean results of analysis, explains at the same time in the most simple manner the slight discre- pancies observed in some few of the lavas occurring in the pala- gonitic district of Iceland, the quantity of lime corresponding to the palagonitic composition decreasing slightly in proportion to the increase of magnesia. Even the almost perfect identity of the constitution of the palagonitic tuffs with that of the normal pyroxenic rocks, the absence of trachytic admixtures, and still more their gradual transition into the almost anhydrous sub- stance of the pyroxenic rocks, which may be observed on a large scale, as well as in the separately-imbedded masses, render it in the highest degree probable that the formation of tuff is most intimately connected with the formation of the pyroxenic rocks. Observations which Darwin made in the Cape Verde Islands, and my own examination of specimens of the rocks to which they refer, for which I am indebted to the kindness of that distin- guished investigator, have chiefly contributed to give me a clue to the remarkable processes concerned in the formation of pala- gonitic tuff. In the neighbourhood of Porto Praja, a basaltic lava occurs, which has been poured out over a recent deposit of limestone. It may there be seen that the lava has, while in a liquid state, acted upon the underlying limestone over which it flowed, and become loaded with fragments of it. The product of this mutual action is a breccia-like conglomerate, in which the altered lava is mixed with a very pure mass of carbonate of lime. A closer examination of this mixture, which even in its exterior has quite the appearance of having been kneaded together in a pasty con- dition, excludes every possibihty of supposing that the fragments of limestone accompanying the lava have originated from sub- sequent infiltration. The chemical change which has resulted THE VOLCANIC ROCKS OF ICELAND. 59 from the contact of the limestone with the lava does not leave any doubt as to the nature of the process by which the palago- nite has been formed. Wherever the lava is in contact with the Hmestone, it is converted into a mass presenting all the mineral- ogical characters and chemical reactions of palagonite ; and this metamorphism^ characterized by a gradual transition into the unaltered rock, is more fully developed where the calcareous substance preponderates over the other constituent of the mass. The analysis of this metamorphic lava showed it to have the following composition, which scarcely differs from that of the pure palagonite : — Silica . . . . 71. 26-21 Oxygen found. 13-87 Oxygen calculated. 14-58 Alumina . . . Protoxide of iron 8-62 1 10-96 J 7-31 7-29 Lime . . . . 4-79 -^ Magnesia . Potash. . . . 9-44 1-81 ^ 4-33* 3-65 Soda . . . . 2-85 J Carbonic acid 5-10 Water . . . . 14-62 Residue . . . 15-65 I have been able to recognize perfectly analogous, although not identical relations, by observation and experiment on basaltic dykes traversing sedimentary limestone. Direct observation is therefore sufficient to prove that a palagonitic substance may result from the action of Hme upon pyroxenic rocks at a high temperature. And, in fact, a number of tuffs from our basaltic formations, and from the volcanoes of the Galapagos islands, which I have had an opportunity of examining, prove in the most decisive manner that some masses of this kind have actually been formed in this way. The following analysis of a tuff from Chatham island, mixed in the most intimate manner with car- bonate of lime, and forming a crater, furnishes a proof of this statement : — * After deducting the oxygen of the carbonates. 60 BUXSEN OX THE FORMATION OF 72. Silica .... .S4-516 Oxygen found. 18-27 Oxygen calculated 17-54 Protoxide of iron 10-400 \ Alumina . . . 10-33S J 7-95 8-77 Magnesia . . 7*801 " Lime .... 4788 Potash . . . 1-644 > 4-49* 4-39 Soda .... 1-525 ^ Phosphate of lime 0-336 Carbonate of Hme 4-320 ' Water . . . 18-140 Pyroxenic subst. 6-476 If the formation of palagonite really takes place in the way pointed out, it must be possible to obtain this mineral artificially by similar means. And, in fact, this m^ay easily be done by igniting an intimate mixture of 1 part finely-powdered basalt and 13 parts of slaked lime and elutriating the mass thus ob- tained with water. The product consists of a mixture of lime and palagonite, which may be recognized under the microscope by its peculiar characteristics. Nevertheless we are obliged to infer that the greater number of palagonites, and especially those of Iceland, have not originated from such a reaction of pyroxenic rocks and limestone, from the fact that carbonate of lime scarcely ever occurs as a constituent of the undecomposed palagonites of Iceland, and also because the per-centage of lime in the mineral itself, calculated for anhydrous substance, does not equal the per-centage of the same constituent in the normal pyroxenic rocks ; for which reason it is not well possible to imagine that palagonite, poor in lime, can originate from pyroxenic rocks rich in lime, in consequence of a further addition of this con- stituent. On the contrary, it might be expected that alkalies would give rise to the formation of palagonite more readily than alkaline earths, as they do not necessarily require to alter the proportion of the constituents in palagonitic pyroxenic rocks. Experiment fully justifies this conjecture. The most beautiful palagonite powder, possessing all the mineralogical and chemical characters of the Icelandic, is obtained when finely-powdered * After deducting the oxygen belonging to the ])hosphates and carbonates. THE VOLCANIC ROCKS OF ICELAND. 61 basalt is mixed with a large excess of fused hydrate of potash, and the basic alkaline silicate formed removed by means of water. The substance obtained after washing and elutriation is hydrated, and pulverulent when dry ; it gelatinizes even with the weakest acids, is readily decomposed by carbonic acid and sulphuretted hydrogen, and has the following composition, cor- responding with that of the purest Icelandic palagonite : — 73. Oxygen found. Oxyg. ;n calculated. Si03 . . . .30-764 16-28 17-1 Fe^O^ . A12 03 . , . 20-497 1 . 4-273 J 9-15 8-5 CaO . . . . 8-016 ->! MgO KaO . . , . 4-600 1 , . 1-826 1 4-56 4-2 NaO . . . 0-532 - H^O. . , . 30-047 26-70 The residue after elutriation consists of a mixture of silicates, whose mean composition differs from that of palagonite only by about a one-fifth smaller per-centage of silica, which remains in the alkaline water used for washing the substance, from which distinctly-formed zeolitic crystals are sometimes separated having the composition SCaO 2Si03-f Aq. I shall again refer to these further on. This conversion of pyroxenic rocks into palagonite is accom- panied by a very remarkable phaenomenon. A considerable quantity of pure hydrogen is disengaged, resulting from the oxidation of protosilicates of iron to persilicates taking place at the cost of the equivalent of water contained in the hydrate of potash. The consequence of this is, that there is no trace of protoxide of iron in the palagonites, and the protoxide of iron in the pyroxenic rocks always appears as peroxide in the palago- nitic tuffs. The protoxide and peroxide of manganese behave in a manner similar to the protoxide of iron, either free or com- bined with silica, passing, with disengagement of hydrogen, into manganate of potash, which reaction may be partly the cause of the frequent incrustations and dendritic deposits which are met with in the palagonitic rocks. The occurrence of metallic cop- per, evidently reduced from chloride of copper, in the palagonite 62 BUNSEN ON THE FORMATION OP tuffs of the Faroe Islands, and elsewhere, admits of the most simple explanation by the reducing action of this hydrogen. It is not very easy at first sight to conceive from whence the alkali might have been derived which could thus give rise to the formation of tuff in Iceland. Considering the variable per-cen- tage of alkali in the pyroxenic rocks compared with the more constant proportions of their other constituents, we might be in- duced to assume that a separation of these alkalies took place in the rocks while in a state of igneous fusion. The idea of such a separation of alkalies is indeed not to be unconditionally rejected. It is well known that most salts are decomposed at high tempera- tures. If their acids are considerably more volatile than the bases, either basic salts are formed, or the bases are liberated while the acid is volatilized. The compounds of sulphuric, car- bonic, nitric, arsenious acids, &c. undergo, with few exceptions, this decomposition. If, on the contrary, the acid is less volatile than the base or the salt itself, as is the case with the ammonia- cal salts, then it is the base which is volatilized while the acid remains behind. The alkaline silicates may very probably suf- fer the latter kind of decomposition ; for if silica and caustic pot- ash or soda are heated together upon a platinum wire by means of a galvanic current to near the melting-point of the metal, the alkali is volatilized by a temperature at which the silica does not even begin to melt. It would appear therefore that in lavas which are heated to a temperature so elevated as to become liquid enough to be shot out in large parabolic curves from the opening of the crater, it is not only possible, but even probable, that a separation of alkalies by heat takes place, the more espe- cially if it is remembered that carbonic acid or aqueous vapour, never wanting in such volcanic processes, must have given rise to the formation of hydrated alkalies and carbonates, compounds which are so volatile that their sublimation may be directly ob- served in some technical operations. There are indeed instances in which the separation and volatilization of the alkaline consti- tuents even of silicates may be detected. In the coal and flux used in the blast furnaces in England, which are remarkable for their intense heat, the whole quantity of the alkali present is in combination with silica. Nevertheless, at the foot of these fur- naces, where for years together a temperature prevails nearly THE VOLCANIC ROCKS OF ICELAND. 63 equal to the melting-point of platinum, such a sublimation of car- bonated alkali, together with cyanide of potassium, takes place, that these products are in some instances collected by hundred weights. Unless this separation and sublimation of alkalies be attributed exclusively to the reducing influence of the coal, it would not be inadmissible to assume that similar processes take place in the immediate neighbourhood of those volcanic foci which are surrounded by the substance of pyroxenic rocks in a state of igneous liquidity. Such conditions therefore may pos- sibly have contributed largely to the formation of palagonite. However, the enormous extent of the Icelandic tuff- rocks renders it very improbable that we have here to do with a mode of origin vvhich would in any case bear the character of a merely local one. It is therefore certainly more in accordance with the strict principles of scientific inquiry to dispense with any merely hy- pothetical explanation of this question, and to rest satisfied with the assumption, justified both by experiment and observation, that in the volcanic periods there was, besides the trachytic ftnd pyroxenic foci, a third focus in a state of activity, which has now become extinct, and whose contents consisted of siHcates rich in alkalies and sufficiently superbasicto break up under the influence of water into palagonitic substance and soluble com- pounds, which were washed away. The occurrence of the pala- gonitic cementing substance, scarcely ever absent in Iceland from among the conglomerates and masses of fragmentary rock desti- tute of fossils and accompanying eruptive rocks, is readily ac- counted for by means of this assumption. It is indeed a direct and necessary consequence of the pouring out of such masses of rock over the surface ; and the palagonitic tuff's containing fossils are nothing more than the products of submarine accumulations of sediment, which received the material for their palagonitic cement from the above-mentioned eruptive silicates, rich in alka- lies, while in a state of metamorphism. 2. Zeolitic Formations. The zeolitic amygdaloid rocks are most intimately connected with those belonging to the palagonitic and pyroxenic classes. They constitute, indeed, the intermediate metamorphic members which connect together these two classes. As the relative mean 64 BUNSEN ON THE FORMATION OF compositions of the latter are almost exactly one and the same, the chemical relation of the zeolitic rocks to the form which they originated, can no longer be ascertained by means of calcu- lation. However, even a superficial observation of the geogno- stic relations in which they occur, leaves'no doubt as to the nature of their formation. Near Silfrastadir, as well as at innumerable other places in Iceland, the zeolitic amygdaloids, having some- what the character of conglomerates, are seen to pass through gradual transitions on the one side into solid trap rock, and on the other into palagonitic tuff, so intimately blended indeed that the concretions and fissures may be traced from the trap through the amygdaloid and into the tuff rocks. At Silfrastadir, where the trap, rising in rocky precipices above the beds of tuff, admits of a closer examination of these relations, the formation of zeo- lites appears most perfectly developed at the contact of the two kinds of rock characterized by this gradual transition, and de- creases towards the compact rock in proportion as the visible traces of a mutual action disappear more and more, so that finally it is only in fissures and isolated cavities that the fine druses of chabasite are found which are characteristic of the amygdaloid formations* of that locality. This phaenomenon occurs every- where in Iceland. It may be observed even in the most recently erupted lavas. One of the most remarkable examples of this kind occurs on the Krafla. The frequently more arenaceous than tuffaceous beds of this volcano, if indeed a range of tuff moun- tains penetrated by craters and lava and traversed by furma- roles may be so termed, are intersected upon the north-western declivity of the mountain by what appears to be a very recent lava, which has issued not from the opening of a crater, but from horizontal fissures in beds. The neighbouring palagonitic rock has, from the point of contact with these beds of lava, suffered a most remarkable metamorphism, which may be best observed under a microscope of thirty or forty-fold magnifying power. The substance of the anhydrous rock, without exactly having been melted, is separated into two siliceous masses, of which one is dark and ferruginous, and the other brilliantly white and free from iron. The former appears as a homogeneous matrix in which the latter is imbedded ; both are amorphous. Nearer towards the lava, where the igneous action was more consider- b THE VOLCANIC ROCKS OF ICELAND. 65 able, the rock assumes the character of the porous basaltic amyg- daloid conglomerates, which are so frequently met with in Ice- land as intermediate members between the palagonitic and pyroxenic rocks. The darker ferruginous matrix, which under the microscope has entirely the appearance of green-bottle glass, but when examined by the naked eye in large pieces appears more to resemble certain conglomeritic pyroxenic rocks, presents sjiheroidal cavities with smooth surfaces, either empty or filled wath granules of the non-ferruginous silicate. When this latter crystalline mass, consisting of zeolitic substance, does not com- pletely fill the cavities, druses of zeolitic crystals have been formed, or isolated zeolitic crystals. This separation into ferruginous and non-ferruginous silicates may be effected artificially in the palagonite and palagonitic tuffs by the most simple means. When pieces of these sub- stances about the size of hazel-nuts are rapidly heated in the flame of a spirit-lamp, or before the blow-pipe, until they are red-hot externally, all the phases of this metamorphism may be most distinctly traced by means of a microscope of forty-fold magnifying power, from the exterior vitrified crust to the interior scarcely decomposed nucleus. In one zone, which, even by its vitrified state, presents the most evident indications of previous strong ignition, there may frequently be recognized a mass filled with amygdaloid and drusy cavities which resembles most per- fectly the basaltic amygdaloid rocks which dip beneath the trap on the Esja near Hruni, and at numerous other parts of Iceland. This resemblance extends so far, that even the crusts covering these artificial crystal druses are identical in outward appearance with those of the natural rocks. Even the manner in which the crystals are situated upon the walls of the druses is precisely the same. Sometimes brilliant chabasite crystals, with the striation peculiar to this mineral, are observed in the ignited mass sepa- rated from the uncrystalline matrix by a crystalline mass of chabasite, followed by a crust resembling a saalband. The extraordinary abundance of zeolitic amygdaloid rocks in Iceland is accounted for in the most simple manner by these experiments, for the conditions requisite for their formation could scarcely be found anywhere combined in a manner more favourable than they are there. Even a hasty glance at the SCIEN. MEM.^Nat, Phil. Vol. I. Part I. F 66 BUNSEN ON THE FORMATION OF lofty perpendicular cliffs of the mountains on the coast, con- sisting chiefly of pyroxenic substance, furnishes a distinct picture of this stupendous metamorphism. Trap dykes of more than a thousand feet in height are not unfrequently seen there tra- versing the entire rocks, sometimes massive, sometimes stra- tified, and branching out through the enormous horizontally extended trap beds in such a way as to leave not the slightest doubt that these masses, intersecting and covering the tuff rocks, are no other than the product of melted matter which was erupted through these dykes. The results of the igneous action of these injected beds of trap stand in a most intimate relation to the magnitude of the heating and heated beds, when especial influences have not prevailed to modify them. In the amygdaloid rocks resulting from the metamorphism of tuff beds, whose original aggregation may frequently be di- stinctly enough recognized in the sometimes angular, sometimes rounded, imbedded masses, such a gradual transition into the unaltered trap rocks may in some instances be traced, that there is an absolute want of any line of demarcation between the two. The passage of a crumbling hydrated rock into one which is perfectly anhydrous, with all those characteristic stages of zeo- lite formation presented by a fragment of ignited palagonite, may here be observed on a large scale. It consequently cannot admit of a doubt that it was neither neptunian nor plutonic agencies alone which gave rise to the formation of zeolitic rocks in Ice- land. On the contrary, we have here to do with a long series of different phases of metamorphic development, the products of which present themselves in the amygdaloid rocks. A truly plutonic rock possessing a superbasic composition, suffers a neptunian metamorphism into palagonite and palagonitic tuff, either at the place of its original eruption, or during the trans- port of its mechanically broken fragments. New plutonic masses break through this altered rock frequently after a long period of rest, and by a second act of metamorphism, now plutonic, con- vert it into zeolitic amygdaloid. Finally, a third neptunian metamorphosis, caused by gaseous exhalations and aqueous va- pour, results from this change, and of which, as the last stage of all these processes, I shall speak subsequently. However simple and inteUigible these phaenomena may be as regards the forma- THE VOLCANIC ROCKS OF ICELAND. 6? tion of zeolitic amygdaloids, the origin of the zeoUtic minerals imbedded in the substance of the trap beds and the more com- pact basalt, where it is obviously necessary to assume the former prevalence of very much more elevated temperatures than those which even, according to the above experiment, are compatible with the formation of zeolites from palagonite, must still appear very enigmatical. But even this phaenomenon admits of being accounted for by means of an experiment, which would appear completely to clear up the mystery attaching to the origin of hydrated siliceous minerals imbedded in plutonic rocks. When, for instance, a finely powdered mixture of 0*2 part lime and 1*0 silica is introduced into 9 parts of caustic potash, melted in a silver capsule, and the whole allowed to cool slowly, after having been kept for some time at a strong red heat in a muffle, upon treating the mass with water, a network of prismatic crystals, frequently 4 to 5 lines in length, is found partly attached to the sides of the capsules. These crystals are hydrated silicate of lime mixed with some carbonate of lime, and represented by the formula 3CaO 2Si03 + Aq :— 74 Silica 27-215 Lime 22*241 Potash 0733 Water separated at 228°-2 . . 36*915 Water separated by ignition . 9*508 Carbonate of lime .... 2*603 99*215 The artificial preparation of this beautifully crystallized hy- drated silicate, and still more the altogether uncommon mode of its formation, are of very great interest in a geological point of view. We have here to consider, not merely the fact that a hy- drated silicate is formed at a red heat, but the still more import- ant fact, that it is not destroyed again under this condition, although, after it has once been separated from the surrounding mass, it loses four-fifths of its water at 228°*2, and the re- mainder at a temperature below red heat. It most undoubtedly follows from this, of its kind isolated, phaenomenon, the fur- ther discussion of which, with all its experimental and theoreti- F2 68 BUNSEN ON THE FORMATION OF cal consequences, I must defer for the more extended considera- tion of this subject, that the palagonitic and zeolitic metamor- phism, which we generally find to succeed one another, may even take place at the most elevated temperatures under the simultaneous and subsequent influence of water. For it is only necessary to add to the red-hot superbasic mixture of caustic potash some powdered basalt, in order to obtain by subsequent treatment with water a mixture of palagonitic substance with those zeolitic crystals of hydrated silicate of lime. And indeed zeolitic druses, traversed by a palagonitic tuff, presenting none of the indications of such a second plutonic metamorphism, are actually found in Iceland, and very abundantly in the Faroe Islands. I have, for instance, a specimen of this kind from the Faroe Islands, which consists of concentric radiated desmin sub- stance, enclosing a nucleus of unaltered palagonitic tuff, and surrounded externally by unaltered tuff. After these experiments and observations, the occurrence of olivine and augite in sharply defined crystals, together with zeolitic minerals in the midst of a hydrated palagonitic mass, is easily accounted for. Those anhydrous minerals are crystalline plutonic products, which were not influenced in their composi- tion by the subsequent neptunian metamorphism. They are consequently found in their original form, together with the zeolitic and palagonitic products of this metamorphism. Simi- lar phaenomena present themselves at the Pferdekopf in the Rhon mountains, with the single exception, that there the pro- ducts of the last stage of rock formation, caused by aqueous vapour and volcanic gases, predominate. The formation of zeo- litic minerals in pyroxenic rocks becomes equally intelligible through the aid of these experiments and observations. They may be produced in the melted rock when this is sufficiently alkaline and superbasic, as in their artificial preparation. And, in fact, the palagonitic constituent so characteristic of the meta- morphosis of superbasic silicates is scarcely ever wanting in the zeolitic pyroxenic rocks. It is indeed the amorphous part of basalt which gelatinizes with acids, and which it is customary to regard as the zeolitic substance of this species of rock. THE VOLCANIC ROCKS OF ICELAND. 69 3. Formation of Rocks by Pneumatolytic Metamorphism. The processes considered under this head refer to manifold products which result from the action of volcanic gases and vapours upon the rocks which have been already treated of. They are of no less importance than the latter, and acquire for geologists an especial interest from the fact, that the processes to which they owe their origin admit of direct observation. In order to comprehend the various reactions which take place during these processes, it must be remembered that the mass of most rocks consists of a mechanical mixture of solid rock with water ; and that the action of the melted masses coming in con- tact with these beds impregnated with water was twofold, and took place at two distinctly different periods. It must have commenced with a vaporization of the water ; and it was not until after this had ceased that the hitherto prevailing temperature, determined by the existing pressure, could increase and be raised so high that the thorough plutonic reaction between the heating and heated rocks commenced. All those apparent contradictions which are presented by the pha^nomena of plutonic contact ad- mit of easy explanation by the aid of this physical necessity. For this purpose three possible cases must be kept in view. The first is when a melted eruptive mass, possessing the lowest pos- sible temperature and moving very slowly, comes in contact with a rock which receives a very rapid supply of water from fissures and spring strata. Here all the conditions are combined, which would tend to prevent any trace of a direct igneous action upon the substance of the adjoining rock. The first result of the con* tact is the formation of a solid crust, frequently of a glassy, sco- riaceous and basaltic character, such as may be observed in many places where basalt has penetrated, and especially when in the form of dykes. This solid crust may be compared to a badly conducting furnace-wall, through which the equalization of tem- perature in the adjoining rock, maintained constantly at the boiling temperature of water, can only take place slowly. A plutonic metamorphism or interfusion of the adjoining rock would here be a physical impossibility. The second case, fre- quent in Iceland, but more rare in our basalts, is that in which the last-mentioned conditions ai'e either absent or comparatively 70 BUNSEN ON THE FORMATION OF insignificant in amount, and its former existence is indicated by- all the signs of a previous state of ignition. The nature of this action is in part essentially determined by the substance of the rock heated. The readily fusible palagonitic tuff is converted into basalt and zeolitic conglomerate, limestone to a superbasic silicate — the material of the palagonitic formations — sandstone to a vitrified mass resembling hornstone, in which the eruptive rock frequently pours through capillary rents and fissures in the form of an extravasated interfused mass. Besides these two modes of action, there is a third, that, namely, in which the boiling-point of the water is raised by enormous pressure, even to a red heat, in consequence of which direct interfusion products of red-hot melted rock with red-hot liquid water are formed. There are, indeed, phaenomena in Iceland w^hich appear scarcely to admit of any other explanation than this. I shall return to the consideration of this point at some future opportunity. These remarks will suffice to show that the greater number of processes of plutonic interfusion and metamorphosis must be preceded by a vaporization of water. The mechanical effects of this change present themselves in the volcanic concussions and eruptions, the chemical effects in the manifold exercise of fumarole action. Consequently the study of this action, and the products resulting from it, is of especial interest for the theory of volcanoes. The volcanic actions continuing after eruptions and manifesting themselves in solfataras, geysers, and thermal springs, furnish the material, which renders it possible, by means of direct observation and experiment, to penetrate to the source of all these phaenomena connected in the most intimate manner with internal volcanic activity. The most important data for such an investigation consist in a knowledge of the composition of the exhalations, which, as con- sequences of the great volcanic catastrophes, issue from the earth in fumaroles. Besides aqueous vapour, which constitutes the chief part of these exhalations, they contain as gaseous constituents only car- bonic acid, hydrochloric acid, sulphur vapour, sulphuretted hy- drogen, sulphurous acid, free hydrogen, and, together with these as foreign admixtures not properly volcanic, nitrogen, oxygen THE VOLCANIC ROCKS OF ICELAND. 7l and ammonia. I have never been able to detect the minutest trace of carbonic oxide or carburetted hydrogens, although I was enabled to employ methods of investigation which would have detected the presence of only a few thousandths of those gases. The chemical character of the fumarole action depends upon the preponderance of one or other of these gaseous constituents. The hydrochloric acid fumaroles, which not unfrequently occur on a large scale near the Italian volcanoes, and are then gene- rally accompanied by a very considerable sublimation of chloride of sodium, appear to be of less importance in Iceland. I was only able to detect traces of hydrochloric acid in a free state in the crater fumaroles a few months old, which owed their origin to the last eruption of Hecla, as well as in the exhalations of vapour from the lava which was then erupted. As the volcano when I visited it- shortly after the last eruption was already so far in a state of rest that there were no violent exhalations of vapour, the gases could only be drawn up by means of an air- pump from the quietly smoking crater fissures, which were tole- rably open to the access of air. The aqueous vapour condensed while drawing up the gases always contained sensible quantities of free hydrochloric acid, which is put down as an uncertain quantity in the following analyses : — 75. 76. 77. 1. Fumaxole 2. Fumarole 3. Fumarole of in the great in the great the lava stream crater of Hecla. crater of Hecla. of 1845. Nitrogen 81-81 82-58 78-90 Oxygen 14-21 16-86 20-09 Carbonic acid .... 2-44 0-56 101 Sulphuretted hydrogen . 0-00 0-00 0-00 Sulphurous acid . . . 1-54 0-00 0-00 Uncertain quantity of hy- drochloric acid . . . Carbonic oxide .... O'OO 0-00 0-00 Carburetted hydrogen 0-00 0-00 0-00 100-00 100-00 100-00 It follows from these analyses that carbonic and hydrochloric acids constitute one part of these crater gases, and that they are sometimes accompanied by sulphurous acid, which decreases so 72 BUNSEN ON THE FORMATION OF much in the fumaroles of the lava streams that it can no longer be detected in the gases, and scarcely in the condensed vapours. The composition of the solid and liquid products of the fuma- roles in the newly-erupted crater of Hecla in 1845, and that of the lava stream which flowed from the lowest crater, correspond perfectly with this fact. The moist gravel surrounding the melted masses of sulphur in the interior of the uppermost and largest crater had the following composition ; — 78. Sulphur . 58-272 Sulphate of lime . . . . 0-796 2AP03 + APCF . . . . 0-425 Protochloride of iron . . . 0-282 Chloride of calcium . . 0-650 Chloride of magnesium . . 0-056 Chloride of potassium . . 0-452 Chloride of sodium . . 0-024 Chloride of ammonium . 0-005 Water 9-402 Decomposed lava gravel 29-636 100-000 With the exception of the ammoniacal salts derived from the atmosphere, these are the same products which may be arti- ficially obtained by the humid reaction of the sulphurous and hydrochloric acids present in the crater fumaroles and the rock of the crater. On the other hand, the only incrustation of salt scantily spread over the bottom of the highest crater, and which, from the unaltered condition of the rock upon w^hich it w^as de- posited, could only have been formed by sublimation, possessed an entirely different composition. Such an incrustation consisted of— Chloride of sodium . . Sulphate of lime . . Sulphate of magnesia . Sulphate of soda . . . Sulphate of potash . . 79. . 5-65 . 63-41 . . 12-68 . 16-78 . . 0-88 99-40 THE VOLCANIC BOCKS OF ICELAND. 73 Since the sulphates as such are not capable of sublimation, it must be assumed that this saline incrustation was originally volatilized in the form of chlorine compounds, which were sub- sequently converted into sulphates by means of sulphurous acid in the presence of aqueous vapour and atmospheric air. In the fumaroles of the lower lava streams, on the contrary, characterized by only a small per-centage of sulphurous acid, the chlorine compounds again predominate, as is shown by the following analyses of products collected there a few months after the last eruption : — Chloride of ammonium 4Fe2 03 + Fe2CP . . 4AP03-fAPCF . . Chloride of magnesium Chloride of calcium . Chloride of sodium Chloride of potassium Silica Water and stony residue 80. 81. 81-68 74-32 5-04 6-75 3-73 0-28 1-69 5-45 0-53 4-63 1*73 2-33 0-53 0-70 0-95 0-25 3-12 5-29 99-00 100-00 The mode in which this mixture of salts was formed is the same as that of the crater products just treated of, except that in this instance it was not the atmospheric air alone which yielded the ammonia on the formation of chloride of ammonium, but, as I have shown* in another paper, likewise the vegetation t par- tially covering the ground overflowed by lava. With regard to the origin of the hydrochloric acid in the crater gases there can be no doubt. Chloride of sodium, which occurs so frequently as a sublimate in volcanoes, is decomposed at very high temperatures, in the presence of aqueous vapours, by silicates into that acid and soda, which unites with the silicates causing the decomposition. It is not necessary to assume that the chlorine compound suffers this decomposition apart from the lava. * Liebig's Annalen, vol. Ixv. p. 70. t According to my experiments, a square metre of meadow land yields on dry distillation a quantity of ammonia corresponding to 223*3 grms. of chloride of ammonium. 74 BUNSEN ON THE FORMATION OF It is very easy to ascertain that the rock formed from the lava which flowed in 1845 from Hecla^ and which yielded the subli- mates frequently impregnated with traces of free hydrochloric acid, itself contains a sensible quantity of basic chlorine com- pounds in its mass. 100 parts of lava from the eruption crater contained in fact 0*246 chlorine, and the same quantity taken from the end of the stream 0*447. From the nature of the origin of hydrochloric acid fumaroles, it necessarily follows that they present the character of perma- nent phaenomena only when the direct volcanic activity, together with the degree of temperature indispensably necessary for this processf*has not receded to any considerable depth below the surface ; for otherwise the hydrochloric acid acting so powerfully upon the substance of the rocks, would very soon give rise to the formation of chlorides which are destitute of that degree of vola- tility which would enable them to reach the surface under the influence of a moderate temperature. It is for this reason that fumaroles are observed as the direct results of great volcanic eruptions soon after they cease, and continuing in a permanent state of activity only where lava has reached the surface in the form of scoriaceous ejections, which have lasted for a consider- able time. If these conditions are wanting, the seat of the change recedes further below the surface, where it may long continue in a state of activity in foci of mineral waters, as may even be inferred from the composition of thermal spring waters of Iceland, with regard to whose formation there cannot for one moment be a doubt, as they may easily be made artificially by the action of the volcanic gases upon the Icelandic rocks. The sulphuretted fumaroles, on the contrary, have an entirely different origin, and are far less ephemeral, lasting sometimes for centuries afler the great volcanic eruptions. Iceland presents, in its stupendous solfataras and geysers, which are the two prin- cipal stages of this fumarole process, the most abundant material for its thorough investigation. Nevertheless, I must in this place restrict myself to the statement of those results of my in- vestigations upon this subject which relate to the general con- nexion of these phaenomena with the original processes of vol- canic activity. THE VOLCANIC ROCKS OP ICELAND. 75 The gases which penetrate forcibly through the muddy soil of the solfatara fields, or break through the more solid rock in vio- lent streams of vapour, must in this case likewise form the start- ing-point of the investigation. The solfataras of Krisuvik present the most considerable of all the gaseous eruptions of this kind. That jet of vapour, which issues from the loose strong ground of the upper ridge some few hundred feet above the principal group of springs in the valley, pours forth with a hissing noise an immense stream of vapour, the tension of which is sufficient to project stones the size of the hand to a height of several feet. This stream of vapour contains 82*30 steam and 17'20 of gas, having the following composition : — 82. Carbonic acid 87*43 Sulphuretted hydrogen . . 6*60 Hydrogen 4*30 Nitrogen 1-67 Carbonic oxide .... 0*00 Carburetted hydrogen . . 0*00 190-00 The composition of this vapour as it issues is — 83. Aqueous vapour .... 82*30 Carbonic acid 15*47 Sulphuretted hydrogen . 1*17 Hydrogen 0*76 Nitrogen ...... 0*30 100*00 According to a measurement, which however must only be regarded as in the highest degree approximative, this spring alone yields, during twenty-four hours, 223 cubic metres of sul- phuretted hydrogen, 12 cubic metres of pure hydrogen, and a quantity of steam, the total effect of which is equal to the power of thirty horses. Close to this spring there is another, scarcely inferior in mag- nitude, and exhaling gas, having almost exactly the same com- position as the last, namely, — 76 BUNSEN ON THE FORMATION OF 84. Carbonic acid .... 88*24 Sulphuretted hydrogen . 6*97 Hydrogen 4*10 Nitrogen 0*69 Carbonic oxide .... 0*00 Carburetted hydrogen . . 0*00 100-00 ' About a quarter of an hour's walk from this place, at the spot where in coming from Reykjavik the first large exhalation of vapour occurs in the bottom of the valley itself, there are at the edge of a piece of meadow ground, generally used by travellers for pitching their tents, a number of large pools of boiling mud, between which vapour is observed to break out with remarkable violence. Although the small space of solid ground then sur- rounding it was continually covered with hot clouds of vapour, it was possible to penetrate to the opening from which the va- pour issued by means of the crusts of gypsum which had been formed between the boiling pools, and to collect the gas for analysis with an appropriate apparatus. Its composition was — 85. Nitrogen 0*50 Carbonic acid .... 79*07 Sulphuretted hydrogen . 15*71 Hydrogen 4*72 Carbonic oxide .... 0*00 Carburetted hydrogen . . 0*00 loo-oo The enormous force with which these gases accompanied by masses of steam issue, would make it appear that these springs are the principal openings of the fissures and channels from out of which the fumarole gases are diffused through the surround- ing rock, causing its metamorphism. As the dissolved products of this metamorphosis possess, as a predominating character, an acid reaction owing to the formation of sulphuric acid, no traces of carbonate of lime or of silica present themselves among the products of the decomposition of the rocks in which the solfa- taras occur. Since the carbonic acid takes no part in these de- THE VOLCANIC ROCKS OF ICELAND. 77 compositions, it is in this instance exclusively the sulphuretted hydrogen and sulphurous acid which, in the presence of heated water, bring about all those alterations of rocks, the most re- markable of which I have already pointed out in a previous paper on the pseudo-volcanic phaenomena of Iceland *. The ana- lysis of the gases which are evolved from the smoking ground in the fumarole district, or from small pools of water and mud, affords the most striking proof of this exclusive action of sulphu- retted hydrogen. Thus, for instance, the considerable per-cent- age of carbonic acid is not diminished, while the quantity of sulphuretted hydrogen, in proportion to that of free hydrogen, decreases more and more. The following analyses of the gases taken from different small pools of boihng water situated in the midst of the Solfatara of Krisuvik, most distinctly show this de- crease of sulphuretted hydrogen : — 86. 87. Nitrogen 1-80 1-44 Carbonic acid 88*54 86-92 Sulphuretted hydrogen . . 1*79 3-28 Hydrogen 7*87 8-36 Carbonic oxide O'OO 0*00 Carburetted hydrogen . . Q-QO 0*00 100-00 100-00 It may be as well, for the sake of completeness, to add here the analysis of such a gas from Reykjahlidh in the far north of Iceland, which was drawn from the smoking muddy soil of a large fumarole by means of an artificial stream of vapour, and was remarkable for its unusually large quantity of hydrogen : — 88. Nitrogen 0*72 Carbonic acid .... 30-00 Sulphuretted hydrogen . 24*12 Hydrogen 25-14 Carbonic oxide .... 0*00 Carburetted hydrogen . . O'OO 100-00 It is evident from these experiments how little ground there is for denying the presence of combustible gases in the exhala- * Liebig's Annalen, vol. l.xii. p. 1. 5^8 BUNSEN ON THE FORMATION OF tions of volcanoes. The objections which are supposed to be fatal to the old volcanic theory of Davy entirely lose their value after these results. For if, in the spirit of this theory, it is assumed that the lavas, and the pha3nomena of ignition accom- panying them, result from an oxidation of alkaline and earthy metals determined by a decomposition of water, it admits of being proved, quite in contradiction of the views which have hitherto been entertained, that the quantity of the hydrogen evolved from volcanoes bears a perfect relation to the magnitude of the streams of lava formed. A. single one of the vapour springs of Krisuvik yields, according to the measurement quoted above, about 12 cubic metres of hydrogen in twenty-four hours. Assuming, then, that the remaining innumerable springs, toge- ther with the large fumaroles occurring there, yield together ^ quantity only one hundred times as great — which may safely bq regarded as far less than the quantity of this gas which is ac- tually evolved — we may by means of this assumption and simple calculation show, that the formation of lava which would be equivalent to such an evolution of gas within the period which elapses between two great eruptions, is sufficient to produce im- mense streams of lava. Nor is it any longer possible to attach any importance to the second of the principal objections which have been made to Davy's hypothesis, namely, that it is unusual to observe any sensible appearance of flames during great vol- canic eruptions. For if, from the known composition of the first-mentioned fumarole gas, we estimate the temperature of its flame, we find it to be 305°*6 ; consequently a temperature which is far below the point of ignition of hydrogen. These gases are therefore combustible only at a red heat, and even under the most favourable circumstances can only produce by such a combustion an increase of temperature amounting to 305°*6, which in a red heat must necessarily escape altogether observation by the eye. Since, as I have already pointed out, the palagonitic meta- morphism is likewise accompanied by a disengagement of hydro- gen, it would appear almost impossible that there should be any doubt as to the source of this gas. However, the constitution of the mixture of gas examined affords in itself a direct proof that neither the formation of palagonite nor a decomposition of water THE VOLCANIC ROCKS OF ICELAND. 79 by means of alkaline or earthy metals, can have had any share in the generation of the volcanic hydrogen ; for both of these processes presuppose the prevalence of a temperature in which carbonic acid cannot exist in contact with hydrogen without suffering a partial reduction to carbonic oxide. But not the minutest trace of this gas is found in the volcanic exhalations. Besides sulphuretted hydrogen, it is especially sulphurous acid which determines the character of the action of solfataras. This gas likewise always appears associated with aqueous vapour. On account of the great readiness with which it dissolves in the condensed vapour, it cannot be collected in the gaseous form. But even the smell of the water condensed from such vapour springs, and the reaction which it gives with iodine, testify to the presence of a considerable quantity of this gas. As sul- phuretted hydrogen and sulphurous acid mutually decompose each other with a separation of sulphur, the two gases can never present themselves together. However, they frequently occur in the same fumarole district, close together. It has been attempted to connect the formation of volcanic gases, in which the presence of hydrogen has hitherto been en- tirely overlooked, in part with the decomposition of organic substances. But the gases which result from the spontaneous decomposition or dry distillation of organic remains do not pre- sent the remotest similarity to these exhalations. To prove this statement, it will be sufficient to bring forward here a few ana- lyses which I have made of coal-gas, and of some natural exha- lations of combustible gases, with regard to whose organic origin there can be no question. Marsh gas from a Pond in the Botanical Garden at Marburg, 89. Nitrogen . . Oxygen . . Carbonic acid Marsh gas . . Hydrogen . . Carbonic oxide Olefiant gas . Summer. Winter. 49-39 18-03 0-17 0-00 3-10 5-36 47-37 76-61 0-00 0-00 0-00 0-00 0-00 0-00 ] 00-00 100-00 80 BUNSEN ON THE FORMATION OF Gas from a Brine-spring yielding Mineral Oil near Hanover, 90. Nitrogen .... 25-12 Oxygen 0-00 Carbonic acid . . . 14-41 Sulphuretted hydrogen 3-18 Marsh gas .... 56-61 Mineral oil vapour 0-68 Hydrogen .... O'OO Carbonic oxide . . 0-00 100-00 Gas from the detonating Salt of Wieliczka. 91. Nitrogen . . . . 10-35 Oxygen . . . . . 2-98 Carbonic acid . . . 2-00 Marsh gas . . . . 84-60 Hydrogen . . . 000 Carbonic oxide . 000 Olefiant gas . . . 0-00 98-93 Marsh gas from the Coal-seams at Obernkirchen^ , 92. Nitrogen 7*16 Oxygen 0-45 Carbonic acid . . . . 2-61 Marsh gas 97-53 [87*53 ?] Hydrogen 0-00 Carbonic oxide .... 0*00 Olefiant gas 0*00 10000 * From the same boring from which BischofF took gas for analysis. THE VOLCANIC ROCKS OF ICELAND. 81 Gas from the Thermal Springs at Aix-la-Chapelle. 93. 94. Kaiserquelle. Corneliusquelle. ( • > ■■ " ' Gas Gas diffused Gas Gas diffused evolved through evolved through freely. water. freely. water. Nitrogen 66*98 900 81-68 779 Oxygen O'OO 1-23 0*00 0*00 Carbonic acid . . . 30*89 89*40 17*60 92-21 Sulphuretted hydrogen 0-31 0*00 0*00 0*00 Marsh gas ... . 1-82 0-37 0-72 O'OO Hydrogen .... O'OO 0*00 O'OO O'OO Carbonic oxide . . . 0*00 0*00 0*00 0*00 Olefiantgas .... 0*00 0*00 0-00 0-00 100-00 100-00 100-00 100-00 95. 96. Quirinusbad. Rosenquellc. Nitrogen 6*41 9-14 Oxygen 0*08 000 Carbonic acid .... 93-25 90*31 Sulphuretted hydrogen 0*00 O'OO Marsh gas 0*26 0*55 Hydrogen 0*00 0*00 Carbonic oxide . . . 000 O'OO Olefiant gas ... . 0*00 0*00 100-00 10000 Gas diffused in the Sulphuretted Water of Nenndorf, 97. 98. 99. Medicinal Spring under Bath spring. the roof. spring. Nitrogen 17*30 19-71 23*91 Oxygen O'OO 0-00 0*00 Carbonic acid .... 69-38 68-29 72*63 Sulphuretted hydrogen . 11*86 11*72 3*29 Marsh gas ... * . 1*46 0*28 O'l? Hydrogen 0*00 000 0*00 Carbonic oxide . . . 0*00 0*00 0*00 Olefiantgas .... 000 0*00 0*00 100*00 100*00 100*00 SCIEN. MEM.— Nat. Phil. Vol. I. Part I. G 82 BUNSEN ON THE FORMATION OP Purified Coal-gas from an English Gas-work. 100. Nitrogen 1*89 Oxygen 0*00 Carbonic acid 2-83 Sulphuretted hydrogen . . . trace Marsh gas 2691 Hydrogen 3513 Carbonic oxide 5*11 Elayle gas 2-70 Ditetryle gas 2-28 100-00 These analyses, to which I could add a great number of others, sufficiently show that the volcanic gases are characterized by the absence of all combustible carbonaceous substances, while in the gaseous products of dry distillation, or spontaneous de- composition of organic remains, they are scarcely ever wanting. If, in accordance with these facts, the solfatara gases can in no way be of organic origin, still it does not require any special hypothesis to account for their formation. The most simple experiment shows that when sulphur and aqueous vapour come into contact with heated pyroxenic rocks, all the conditions necessary for their formation are present. When the vapour of sulphur is passed over basalt, or any other of the pyroxenic rocks treated of above, at a red heat, a partial decomposition of the peroxide of iron in these rocks takes place, the sulphur being divided between its constituents. The oxygen of the oxide escapes in the form of sulphurous acid, and the metal remains in the rock as sulphuret. If afterwards steam is passed over the sulphuretted rock, still at a red heat, an abundance of sulphu- retted hydrogen is disengaged, and magnetic oxide of iron is formed. If the temperature exceeds only very slightly a red heat, a part of this sulphuretted hydrogen is decomposed into its elements, and a sensible quantity of free hydrogen is found mixed with the sulphur vapour. Fragments of basalt, from the Stempelskopf near Marburg, heated to redness in sulphuretted hydrogen, and then treated with steam at a higher temperature, yielded a mixture of gases having the following composition : — THE VOLCANIC ROCKS OF ICELAND. 83 Sulphuretted hydrogen . . . 93*99 Hydrogen 6*01 100-00 The phaenomena upon which depends the activity of solfataras become, after these experiments, very easily intelligible. It is well known that almost all volcanic eruptions are accompanied by sublimation of sulphur. Consequently the zone from whence the sulphurous acid originates is situated where such masses of sulphur, whose occurrence may be easily accounted for by the action of volcanic heat upon decomposable sulphur compounds, come into contact in the state of vapour with red-hot pyroxenic rocks. When, at a subsequent period, the temperature commu- nicated to this zone by the volcanic action sinks, a new phase of chemical action commences. The sulphur compounds of iron, and perhaps likewise of alkaline and earthy metals, which have been formed there, commence their action upon the aqueous vapour, and from this mutual action result sulphuretted hydro- gen and the products of its decomposition — free hydrogen and sulphur vapour. It is therefore evident that these two processes are blended in each other, and meet in such a manner as neces- sarily to determine the irregular simultaneous exhalation of those gases at spots but little distant from each other in the same fumarole district. These processes likewise afford an explana* tion of the chronological course of the fumarole action. The sul- phurous acid, whose appearance alone characterizes the initial stage of all these phaenomena, is accompanied after a while by sulphuretted hydrogen, which by its reaction with the former gas gives rise to that succession of decompositions which cha- racterize the true solfataras. Acid liquids saturate the rocks traversed by separated sulphur and torn up by aqueous vapour, converting these rocks, as I have already shown, whether they belong to the pyroxenic or trachytic group, into clay, by ex- tracting from the silicates potash, soda, magnesia, lime, protoxide of iron, and frequently a part of the alumina, as sulphates. This phase of destructive action is followed, in course of time, by a productive one, which increases in proportion as the source of sulphurous acid becomes extinct, and the gradually decreasing evolution of sulphuretted hydrogen recedes to greater depths. G2 84 BUXSEN ON THE FORMATION OF In consequence of these changes, the acid reaction of the water with which the rocks are impregnated is converted into an alka- line reaction, resulting from the formation of alkaline sulphurets at the cost of the now alone-acting sulphuretted hydrogen. Simultaneously with the disappearance of the acid reaction, commences the action of the free carbonic acid upon the rocks, and the alkaline bicarbonates resulting from this action provide a solvent for the silica, by which means, and in accordance with the most simple laws, which I have already described, those wonderful geyser structures are formed, giving rise to the stu- pendous phaenomena of the Icelandic eruptive springs. The springs of carbonic acid finally make their appearance as the terminal stage of all this series of phaenomena. They gene- rally survive the plutonic catastrophe the longest, and appear to be exclusively limited to the Western Islands. The fumaroles on the craters of Hecla were, when I had an opportunity of examining them more carefully shortly after the eruption in 1845, in that state which I have described as being the first stage of the secondary volcanic action. Not the slightest trace of sulphuretted hydrogen could then be detected either by the smell or by means of reagents, while, together with the abundant sublimation of sulphur, the presence of sulphurous acid could be recognized by its smell at a considerable distance from the craters. On approaching a lighted cigar to the fuma- roles, those thick clouds of smoke were indeed observed which Piria has pointed out to be an indication of the presence of very minute traces of sulphuretted hydrogen. However, as it is very easy to ascertain that sulphur itself will cause the same phaeno- menon when subliming with aqueous vapour, it is doubtful whether even a trace of sulphuretted hydrogen accompanied the exhalations from the crater at that time. In the year 1843, I observed exactly the same phaenomena in the crater of Vesuvius, when, after a long period of rest, it again began to show signs of activity, and lava, solidifying to scoriae, commenced to pour forth from the crater-cone with periodic explosions of vapour. The last eruptions of the Krafla and Leiihnukr, during the last century, likewise appear, according to the certainly very scanty records preserved, to have been accompanied by such phae- nomena. THfi VOLCANIC ROCKS OF ICELAND. 85 In the extensive solfataras surrounding these latter two vol- canoes, the second phase of the secondary volcanic action pre- sents itself at the present time on the most stupendous scale, while the volcanic activity is already in a state of considerable decrease. As at Krisuvik, in the south-west of Iceland, the ex- halations of sulphurous acid are very much inferior in magnitude to the sulphuretted hydrogen, which here escapes from the smoking clay soil and the boihng pools of mud in preponderating quantity. With regard to the third phase of these operations manifesting itself in the phaenomena of geysers, it can scarcely anywhere be studied in its chronological relations more favourably than at the famous springs of Haukadalr, which bear the name of the Great Geysers. The crater of these springs, which has acquired the material of its coating of sihceous tuff, like most of these springs, from the readily decomposable palagonitic tuff, rests upon a surface which is still in a state of fumarole action, situated at the north-western edge of the spring cone, a section of which is shown by a ravine. The streams of vapour, which there burst through the fumarole clay, coincide most perfectly, in their ex- ternal appearance and in their actions, with the springs which are met with in the solfataras of Krisuvik and Reykjahlidh, with the single exception that at the geyser there is no trace of sul- phurous acid or any sensible deposition of sulphur. A glance at the following composition of the gases, taken from this fuma- role district, must, in fact, remove every doubt as to the identity of the origin of all these phaenomena : — 101. Nitrogen 84*11 Carbonic acid 8*92 Hydrogen . 6*59 Sulphuretted hydrogen . . . 0'38 Carbonic oxide 0*00 Marsh gas 0*00 Oxygen 0*00 100-00 The proportion of hydrogen to sulphuretted hydrogen and to free carbonic acid, furnishes here also a standard for estimating the consumption of the two latter gases, which, as may be seen. 86 BUNSEN ON THE FORMATION OF is far greater in this instance than in the solfataras. These few simple and easily-intelHgible processes of volcanic gas formation contain the key to a whole series of metamorphic changes, which may be comprised under the name of the pneumatolytic, as a uni- versal and frequently-occurring class of phaenomena. The com- position and mode of formation of the acid and alkaline thermal waters of Iceland, which may easily be ascertained even by direct experiment, follow as simple consequences from these processes ; and the argillaceous formations, which are found on the saal- bands of the trachytic and pyroxenic dykes, but on the most extended scale in the beds of conglomeritic amygdaloids and tuffs, penetrated by plutonic rocks, have in part resulted only from a stupendous repetition of those very same processes of decomposition, which we may daily observe going on at the surface of the Icelandic solfataras. Without here referring at length to the less interesting modes of rock formation, which are determined by the separate action of aqueous vapour and hydro- chloric acid, I will restrict myself, in order not to exceed the limits of this paper, to a brief statement of some few alterations of rocks which take place under the influence of the solfatara gases, and with which the long series of rock metamorphisms concludes. While the palagonitic metamorphosis gives rise to the formation of hydrated products, in w^hich the relative pro- portion of the normal pyroxenic mass appears scarcely altered, the pneumatolytic metamorphism is accompanied by a loss of substance, by the decomposed rocks extending in the first in- stance to the alkalies and alkaline earths, and then to the oxides of iron and the silica. The influence of the fumarole gases, which no volcanic rock is able to withstand, not even the most acid trachyte, may be traced through all the phases of a pro- gressive decomposition in the rocks of the solfatara districts. The first commencement of this action manifests itself as a bleaching of the rock ; the dull appearance thus acquired is then followed by a disintegration of the mass, which advances until the rock becomes easily pulverizable ; and finally, when the action is completed, there remains a plastic argillaceous mass almost entirely free from iron, which perfectly resists any further change, and after drying presents a very friable character, and gives a bright streak. The characteristic collateral products of THE VOLCANIC ROCKS OF ICELAND. 87 this process, whose special modes of formation I have already fully discussed in another place*, are crystals of iron pyrites, silica in the form of hyalite, hydrated oxide of iron and anhydrous oxide of iron, formed from it by long- continued boiling with water, as well as in some instances carbonate of lime and gypsum. The springs of the Great Geyser have formed their siliceous incrustations over a surface of palagonitic tuff, which, as I have already mentioned, is penetrated by solfatara gases. From this palagonitic bed, covered with siliceous tuff and boulders, rises the small trachytic chain of mountains called Laugarfjall, which extends along by the side of the springs in a north-eastern direc- tion. The traces of a not inconsiderable geyser action may be followed on the declivities of these hills as far as their trachytic ridge of rocks, and which, with the exception of a few isolated and unimportant springs of vapour, is almost entirely extinct. Below these is one spring, issuing from the trachytic rock itself, which it has converted into a white, earthy, friable mass, of a dull appearance, or even into a plastic clay. The following analyses of the original and the decomposed trachyte, show it was especially the alkalies which were extracted from the rock, while water was taken up : — 102. 103. Undecomposed trachyte. Decomposed trachyte. Silica 75-48 75*84 Alumina .... 12-97 13-71 Protoxide of iron . 2*61 Peroxide of iron . 3*21 Lime 1-01 0*70 Magnesia. . . . 0-03 0-14 Potash .... 5-43 1*24 Soda 2-72 1-94 Water 0-32 2-18 100-57 98-96 The earthy mass passes finally under the increased action of the fumarole gases into a fat, pliant pipe-clay, in which the iron of the original rock is found in the form of small crystals of iron pyrites, the formation of which, as I have already shown, de- pends upon a very simple process of decomposition. The com- * Liebig's Annalen, vol. Ixi. p. 1. 88 BUNSEN ON THE FORMATION OP mencement of this pneumatolytic process is not exclusively limited to the more concentrated evolutions of vapour from the solfataras ; on the contrary, it extends not unfrequently beyond the widely-spread masses of trachytic rocks. It is more espe- cially where these rocks penetrate the pyroxenic rocks, or are penetrated by the latter, consequently in the nearest proximity to the foci from whence, as I have shown above, the solfatara gases originate, that all those characteristic indications of the commencement of such fumarole actions make their appear- ance. The yellowish or bluish gray colour of the trachyte is replaced by a white ; the rock assumes a duller appearance ; and even if the decomposition does not usually advance so far that a perceptible loss of alkalies manifests itself, still a number of small and chiefly microscopic crystals of iron pyrites and a not inconsiderable per-centage of water may be detected in the rock as characteristic indications of the already commenced ac- tion of the solfatara gases. But these indications present them- selves much more frequently, and in a more marked manner, in the saalband of the trachytic dykes, where the metamorphoses and abundant formation of iron pyrites in the adjoining rock point out the course which was taken by these gases, whose evolution was the secondary result of the great elevations of trachyte. The basic palagonites and the pyroxenic rocks are still more readily decomposed under the influence of the heated water and the gases which it contains in solution than the acid trachytic rocks are. The dark substance of the rock assumes in this case also a lighter colour in the first instance, and breaks up into an earthy mass, which gradually becomes richer in water and poorer in alkaline bases and protoxide of iron, until finally it is entirely converted into a white, bluish gray, yellow or red clay, filled with small crystals of sulphur in beds, and not unfrequently con- taining admixtures of gypsum. It sometimes happens that all the stages of this metamorphism may be observed in one frag- ment of pyroxenic rock taken from the spot where the solfatara occurs. The frequently quite unaltered nucleus is found to pass towards the exterior into a plastic argillaceous mass, which con- sists of separate layers of white, gray, yellow, or brownish red colour alternating with each other, sometimes containing iron THE VOLCANIC ROCKS OF ICELAND. 89 pyrites, and sometimes free from it. The recently mentioned palagonitic substance of the tuff surrounding the solfatara of Krisuvik consists of — Silica .... Alumina . . . Peroxide of iron Lime . . . . Magnesia . . Potash . . . Soda . . . . Water . : . Phosphoric acid Residue . 104. 37-95 13-61 13*75 6-48 7*13 0-42 1-72 12-68 0-43 7-25 101-42 } Oxygen found. 20-09 10-49 5-47 Oxygen calculated. 20-60 10-30 5-15 It may be seen that the composition corresponds almost com- pletely with the formula proposed for palagonite, 3RO 2Si 0^4- '2R^ O^ SiO^ + Aq, and consequently approximates very closely to the composition of the pyroxenic rocks. A fumarole clay formed from this tuff contained nearly 30 per cent, of very finely developed crystals of iron pyrites ; and the argillaceous matter separated from them consisted of — SiUca . . . . 105. 49-84 Oxygen found. 26-38 Oxygen calculated. 25-75 Alumina . . . 26-78 12-52 12-88 Lime . . . 0-38 ->! Magnesia Potash . . 1-09 . 0-26 1-89 2-15 Soda . . . . 010 ^ Protoxide of iroi 1 5-73 Sulphuret of iro Gypsum . . Water . . n 1-53 . 0-55 . 14-95 101-22 The mixture of which the argillaceous mass consists corresponds with the formula RO 2Si03 + 2R203 SiO^ + Aq. Exactly the same decomposition is observed in the pyroxenic rocks. At the north-eastern foot of the Namarfjall, near Reykjahlidh, there are 90 BUNSEN ON THE FORMATION OF a number of large boiling pools of mud, which throw up their blackish gray argillaceous contents to a height of 10 or 15 feet, and heap it up in circular crater-like enclosures. One of these pools still shows at some depth traces of the adjacent pyroxenic lava stream, which bounds the solfataras on the north-east, and whose decomposed substance forms the argillaceous mud which fills the pool. The following analysis of this mud shows that lava rocks suffer the same decomposition as the palagonite when exposed to the action of the solfatara gases : — 106. Silica 55-62 Alumina 12*77 Peroxide of iron . . 1*91 Lime 1*56 Magnesia .... 036 Potash 0-43 Soda 1-18 Water 5*53 Sulphur 0-92 Sulphate of lime . . 3*45 Iron pyrites . . . 16*27 100-00 When we find the solfatara action lasting for centuries afler the comparatively insignificant eruptions of modern volcanoes, we may certainly expect to meet with traces of similar processes resulting from the far more stupendous catastrophes which gave rise to the formation of the older trachytic and pyroxenic rocks. And in reality these may be observed, with all their characteristic peculiarities, in the plutonic dykes or injected masses occurring in Iceland, and traced from thence under circumstances which do not leave the least possibility of entertaining the remotest idea of their being the results of changes which rocks suffer under the influence of a simple process of extraction by surface water. The pneumatolytic changes, indeed, frequently take place without any removal of constituents from the altered rocks. The pyroxenic dyke masses, passing in their saalbands into a clay mixed with pyrites, carbonate of lime and gypsum, present themselves in a state which does not admit of their being distin- THE VOLCANIC ROCKS OF ICELAND. 91 guished either mineralogically or chemically from those products of decomposition which are observed in the rocks of the solfa- taras, and in mere appearance they are totally different from those decomposition crusts presented by fissures in the interior of the same rock, and communicating with spring strata, whether the decomposition may be less advanced, or still more so than in these dykes. A basaltic dyke, which has penetrated through the adjoining older trap on the north-eastern coast of Vidhey in the harbour of Reykjavik, presents a fine example of this kind. The trap rock contains a considerable per-centage of water, and possesses exactly the composition of the normal pyroxenic mass. Where it is in contact with the basalt dyke, a crust resembling trachyte is observable, which has a pitchy-black appearance, conchoidal fracture, and glassy lustre. Dried in the air at the ordinary temperature, it contains only 0*84 per cent, of water, and presents entirely the appearance of a fresh obsidian-like glass. The following analysis shows that this crust is actually nothing more than melted normal pyroxenic substance, having the same composition as other trap, from the lateral fusion and rapid cooling of which it has originated : — 107. Silica 47-58 Protoxide of iron .. 17*51 Alumina .... 13-78 Lime 11-36 Magnesia . , . . 6*48 Soda 2-90 Potash 0-60 Water 0-84 101-05 The blackish gray substance of the basalt dyke itself has a dull earthy appearance, contains pyrites, together with calcareous spar disseminated through its compact mass, and resembles in the most deceptive manner the rocky nuclei which are sometimes found in imbedded pyroxenic masses which have been already converted by lying in the solfataras into pyritic clay on their surface, while in the centre the decomposition is only in its initial stage. The composition of this altered basaltic mass gives 92 BUNSEN ON THE FORMATION OF a full explanation of the nature of the metamorphosis which has taken place. 108. Silica 46-47 Alumina .... 14*71 Protoxide of iron . . 14*29 Lime 8-04 Magnesia .... 4*98 Soda 1-53 Potash 0*87 Carbonate of lime . 5*35 Pyrites 1*04 Water 3*58 Gypsum trace 100-86 We here observe the remarkable fact, that when the per-centage of lime and protoxide of iron, which has suffered from the me- tamorphic action, is again added to the original silicate, we obtain almost exactly the composition of the normal pyroxenic substance, thus : — 109. Silica 49-17 Alumina .... 15-57 ^ ^i-^'i } Protoxide of iron . 15-78 Lime 11*68 Magnesia .... 5*27 Soda 1-62 Potash 0*92 100*00 The composition, which, as in the normal pyroxenic substance, corresponds to the relation between the quantities of oxygen in the acid and bases of 3 : 1*936, proves that the change which has taken place in the rock cannot have been connected with any sensible extraction, and that consequently it was not cur- rents of water, but the gases and vapours dissolved in the water which attacked and caused the metamorphism of the rock. The remarkable trachyte dyke at the south-eastern foot of the Esja- mountains, whose interfusion with the adjoining pyroxenic rock I THE VOLCANIC ROCKS OF ICELAND. 93 have already described above, has exercised precisely the same influence upon the surrounding rock, which has thus for some con- siderable distance round been converted into a pitchy black mass, nearly resembling obsidian, sometimes dull and sometimes almost brilliant in appearance, containing calcareous spar and zeolites intimately disseminated throughout, and which, to judge from its zeolitic admixture and frequently still recognizable imbedded earthy fragments of pyroxenic rock, is the product of the meta- morphosis of a palagonitic tuff. Analysis gave the following mean composition for this rock : — Silica 47-47 Alumina . . . . 11-85 Protoxide of iron . 15-24 Lime . . . * . 5-76 Magnesia . . . M7 Soda 1-93 Potash .... 0-32 Carbonate of lime 8-45 Water .... 2-61 The absence of pyrites and gypsum proves that it was car- bonic acid alone which caused the metamorphism in the moist- ened rock. Here likewise, if the per-centage of lime contained in the original rock is restored from the disseminated calcareous spar, we obtain as the composition of the anhydrous rock, — 111. Sihca 50-25 Alumina 12*54 Protoxide of iron . . 16*15 Lime 11-09 Magnesia .... 7*59 Soda 2*04 Potash 0-34 100-00 By this means we again obtain a composition for the original rock, which is almost exactly the same as that of the normal py- roxenic substance, and indeed with the oxygen relation 3 : 1-81, 94 BUNSEN ON THE FORMATION OF which approximates remarkably near to the corresponding rela- tion in palagonite, 3 : 1-95. The palagonite rock of Laugar- vatnshellir possesses almost exactly the same composition, cal- culated for anhydrous substance, as this rock, which is so totally different from it in all its characters : — 112. Silica 50-71 Alumina 13'55 Protoxide of iron , . 15*44 Lime 10*75 Magnesia .... 7*98 Soda 0-76 Potash 0*81 100-00 If this carbonate of lime, amounting to more than 8 per cent., had been deposited by percolating calcareous water, and did not result from the penetration of carbonic acid into the moist rock, it would be inconceivable why the per-centage of lime in the dis- seminated calcareous spar should give almost exactly the com- position of the original rock ; and further, why such an enor- mous process of extraction should have failed to exercise any influence upon the so readily decomposable alkalies ; and finally, how such a deposition of lime could have taken place at all in the compact rock, entirely destitute of cavities. Without at- tempting here to follow up these relations more closely, it may suffice to bring forward a few more examples of these fumarole actions which are frequently met with in Iceland, without any connexion with the course of the spring strata, and com- mencing from the plutonic dykes spread out into the sur- rounding rocks. At the spot where, on the road from Hruni to Storinupr, the Laxa is first met with, there rises on the south-eastern bank of this river the trachytic precipice Arnarhnipa, which is traversed by a pyroxenic dyke, whose edges are converted by fumarole action into a soft lavender -blue clay containing pyrites and carbonate of lime, and possessing the following composi- tion • — THE VOLCANIC ROCKS OF ICELAND. 5 113. Silica ...... 47*05 Alumina 10-91 Protoxide of iron . . 12*66 Lime 11*79 Magnesia .... 7 '73 Soda 1-23 Potash 0*75 Carbonate of lime. . 1*01 Pyrites 0*20 Water 6*67 Gypsum trace 10000 On account of the unequal distribution of the pyrites in layers, it is not possible to compare this product of decomposition with the original rock from which it is derived. It perfectly resembles, not only in its composition, but likewise in its natural appearance, a rock decomposed by solfatara action, while it does not present the slightest similarity to the decomposed mass resulting from the mere action of water upon the fissures of pyroxenic rocks. Further towards the centre of the dyke the rock is more difficult of fracture, has a less decomposed appearance, and is impregnated throughout with calcareous spar. The composition of this sub- stance was found to be — 114. Silica 50-82 Alumina 11*10 Protoxide of iron . . 12*97 Lime 4-34 Magnesia .... 3*90 Soda 1-93 Potash 0*31 Carbonate of lime . . 8*16 Pyrites 0*26 Water 5*05 Gypsum trace 100-00 In this instance, likewise, we meet with the remarkable fact that the metamorphic substance of the dyke has not lost any of 96 BUN8EN ON THE FORMATION OF its constituents by extraction, and that the calcareous spar and pyrites present in it have not been introduced from outwards, but were formed by simple re-ariangement of the constituents upon the spot. For if the radicals contained in the calcareous spar and the pyrites are restored to the rock as lime and prot- oxide of iron, we obtain the composition of a tracheo-pyroxenic rock perfectly corresponding with theory ; — 115. Found. Calculated. Silica 56-48 56-48 Alumina and protoxide of iron . 26*62 25*65 Lime 9*78 8*91 Magnesia 4*33 4*99 Potash 0*34 1*38 Soda 2*15 2*59 100*00 100*00 This constitution corresponds to a mixture of 1 part trachyte and 2*521 pyroxenic substance. The penetrated rock has there- fore here also mixed with the penetrating mass. These exam- ples, to which I could add a great number of others, may suffice to explain the fact, that the gases penetrating the rocks toge- ther with water and aqueous vapour, are capable of metamor- phosing their substance on the spot without any removal of the products of decomposition. It only remains for me at the close of this general view, to mention the phaenomena which present themselves where the pneumatolytic and zeolitic metamorphoses take place together. The result of this state of things is the production of argilla- ceous amygdaloid conglomerates, extremely rich in zeolites, and which alternate in irregular order as thick beds or masses with the pyroxenic eruptive rocks. They may be regarded as pyro- caustic products of an immense subterranean fumarole action, in which sulphuretted hydrogen or sulphurous acid were less active than aqueous vapour, either alone or mixed with carbonic acid. The already-mentioned districts round Silfrastadir are especially remarkable in this respect. This small spot is situated in the valley of the Heradsvotn, which opens into that of the SkagaQordhr, and is enclosed by rocky precipices, consisting of THE VOLCANIC ROCKS OP ICELAND. 97 a compact trap rock alternating -with zeolitic, amygdaloid and palagonitic tuff. The trap, which belongs to the gray coarse- grained variety occurring all over Iceland, passes very gradually into a tough, bluish-gray, zeolitic amygdaloid, presenting an almost earthy appearance, and whose substance consists, to the amount of almost one-third, of crystal druses and rough masses of chabasite. No perceptible line of separation between the two rocks can be detected in the fissures and imbedded masses. When the homogeneous mixture of this amygdaloid admits of an average composition being ascertained by analysis, it is found to have exactly the constitution of the trap into which it passes ; and it follows from the accompanying analyses, calculated for anhydrous substance, that both rocks consist of pure normal pyroxenic or palagonitic substance. 116. 117. Trap. Amygdaloid. Silica 49-87 49-60 Alumina 14-66 13-98 Protoxide of iron . . 13*57 14*60 Lime 12-56 11*78 Magnesia 6-55 6*90 Potash ...... 0-42 0*22 Soda 2*37 2-92 100-00 100-00 An equally gradual transition may be observed from the amygdaloid into a red, friable, argillaceous bed, only a few feet thick, but extending for miles round, the substance of which, under the microscope, appears as an altered amygdaloid, in which the unaltered chabasite is found with all its characteristic peculiarities of distribution and concretion, surrounded by the decomposed rock. The succeeding beds of palagonitic tuff are again connected by gradual transitions throughout all the phases of a progressive decomposition in the most intimate manner with the above argillaceous bed, in such a manner, indeed, that separate concretions may be traced from the sohd trap rock down into the unaltered beds of tuff. This metamorphosis, which may be imitated on a small scale with any piece of palagonite, has here been effected on a large scale in Nature, at the time when the erupted melted trap caused the fusion of the underlying pala- SCIEN. MEM.— Aa^ PhJL Vol. I. Part I. H 98 BUNSEN ON THE VOLCANIC ROCKS OF ICELAND. gonitic beds, and by a vaporization of water gave rise to the formation of the red pneumatolytic amygdaloid. Here, likewise, we find the metamorphism commencing as a physical necessity from the melted trap masses terminating in a fumarole action, whose products most perfectly resemble those of the solfatara localities. These phaenomena present themselves very frequently in Iceland, and still more frequently in the palagonitic trap for- mations of the Faroe Islands, so distinguished for their abun- dance of zeolites. It follows, with the most indisputable necessity, from these last-mentioned facts, that this formation of zeolites does not depend upon a removal and deposition of substances, but results solely from a transformation of palagonitic rocks taking place upon the spot. 99 Article III. On the Dependence of Radiant Heat in its passage throughCry- stals upon the direction of transmission. By H. Knoblauch, Professor of Natural Philosophy in the University of Marburg, [From PoggendorfTs Annalen, 1852, No. 2.] 1 HE quantity of radiant heat which passes through diather- manous bodies, depends, as is known, upon the nature of the latter. Melloni was the first to raise the question, whether, in one and the same body — in a crystal for instance — the quantity transmitted was different along the different axes. In the results arrived at by Melloni and myself while experimenting with transparent rock-crystal, and with calcareous spar, no difference of the kind was detected. Notwithstanding this, I have been induced to extend the inquiry to other crystals. A brown specimen of quartz was first subjected to experiment. It was a complete cube, having four of its sides parallel to the axis of the crystal, and the remaining two perpendicular to the same. The sun^s rays appeared to be peculiarly well adapted to experiments of this nature, inasmuch as they offer the double advantage of being parallel among them- selves and of possessing the highest possible intensity. By means of the reflector of a heliostat, a sheaf of rays was thrown into a dark room ; the direction of the sheaf, notwith- standing the changing position of the sun, did not materially alter during the time of observation. The rock-crystal was placed in the path of the rays, being fixed upon a disc which permitted of a rotation of 90° in a horizontal plane. The heat, after having passed through the cube, was permitted to fall upon the square forward surface of a thermo-electric pile, and the consequent excitation was measured by a multiplier connected with the pile. It is scarcely necessary to remark, that all inci- dental heat was carefully shut out, and those rays only were allowed to fall upon the pile which had first traversed the mass of the crystal. H2 100 KNOBLAUCH ON THE TRANSMISSION Under these circumstances, it was observed that the calorific rays, when transmitted parallel to the axis of the crystal, pro- duced a deflection of the needle of the multiplying galvanometer amounting to 34°-05*. When, however, the rays were trans- mitted in a line perpendicular to the axis of the crystal (the latter having been turned 90° round), the deflection produced was only 31°'19. The ratio of these numbers, which, as in most of the cases hereafter cited, are the arithmetic means of twelve different ob- servations, is as 100 : 92. The quantity of heat, therefore, transmitted by a brown rock- crystal, in a line perpendicular to its axis, is less than the quan- tity transmitted in a direction parallel to the same. There is no reason whatever to suppose that this difference is in any way attributable to accidental imperfections on the part of the crystal. The same experiment was made with a cube of beryl', here also a weaker power of transmission was exhibited perpendicular to the axis than parallel to the same. If, as in the former case, we express by 100 the quantity of heat transmitted parallel to the axis of the beryl, then the quantity transmitted perpendicular to the axis, as measured by the multiplier, will be expressed by the number 54. This result * When accurate determinations with the thermo-electric pile are required, it is not sufficient to simply read off the deflection ; for this is not due to the temperature which is to be measured alone, but also to accidental differences of temperature between both faces of the pile. The amount contributed by the latter source to the total effect must therefore be ascertained ; a metallic screen is placed in the path of the rays which totally intercepts them, without altering the accidental difference spoken of. The deviation of the needle which remains, indicates the true zero, from which we must reckon in the subsequent deter- mination. If, for example, the total deflection were 35°05, and the needle, after the introduction of the metallic screen, remained standing at 1°, the action to be attributed to the incident rays would amount to .34°'0o. All the values contained in the present memoir have been determined in this way. Each of the numbers whose arithmetic mean we have taken is thus the result of two readings of the thermo-multiplier, one of which was simply to ascertain the true zero for the observation made at the time. The deflections of the needle are to be considered as the measure of the quantity of heat as within the limit of 35° (according to a special examination of the instrument); the deflections were proportional to the forces which pro- duce them. OF RADIAN'P HEAT THROUGH CRYSTALS. 101 does not however possess the same demonstrative value as that obtained in the case of rock-crystal, inasmuch as the mass of the beryl was in some measure unhomogeneous. A cube of tourmaline cut, as regards its axis, exactly as the former, and subjected to the same conditions of experiment, showed an opposite deportment. In this case the quantity of heat transmitted perpendicular to the axis was greater than the quantity transmitted parallel to it. The deflections of the galvanometer by which these quantities were measured are to each other in the proportion of 100: 158, the former number denoting the quantity transmitted along the axis, and the latter the quantity transmitted in a line perpendi- cular to the same. The homogeneous character of the specimen experimented with, does not permit of the idea that the observed difference is due to any accidental irregularity of the mass of the crystal. The foregoing experiments, therefore, prove that, in certain crystals, radiant heat is transmitted in unequal quantities in dif- ferent directions. It seemed to me of interest to pursue this inquiry further, by applying polarized calorific rays* instead of the natural ones. For this purpose the rays were caused to pass through a Nichol's prism, previous to falling upon the cube of rock-crystal. The prism was first so placed that the principal section, passing through its obtuse angles, was vertical, and hence the plane of polarization of the issuing rays horizontal. When the calorific rays thus polarized were permitted to traverse the cube in the line of the crystallographic axis, a de- flection of 8°*1 was produced. When the line of transmission was perpendicular to the axis, the deflection was 8°* 3 ; the dif- ference between the«e two values lies within the limits of the error of observation, and hence they must be regarded as equal. Hence when heat, polarized in the manner indicated, is ap- • Strictly speaking, the heat as applied in the foregoing cases was not in a natural state, but polarized elliptically by reflexion from the steel mirror of the heliostat. But the peculiarity communicated in this way (the angle of incidence being so small), compared with the linear polarization to which our experiments refer, is of so vanishing a character, that, as far as our results are concerned, the light may be regarded as being in its natural state. SCIEN. MEM.^Nat, Phil. Vol. I. Part II. I 102 KNOBLAUCH ON THE TRANSMISSION plied, the differences observed in the case of the natural rays disappear. The deportment of beryl was similar to that of rock-crystal. True, in the case of beryl a greater quantity was transmitted along the crystallographic axis than perpendicular to it; but the proportion, which in the case of the natural rays was as 54 : 100, became, when polarized heat was applied, as 70 : 100. The difference between these two numbers is to be ascribed to the irregularity of the mass already alluded to. In the case of tourmaline, no difference was observable in the quantity of heat transmitted. The thermoscope exhibited the same deflection, whether the rays were transmitted parallel to the axis or perpendicular to it. From this we may conclude, that polarized calorific rays, whose plane of polarization coincides with the axes of the crystals men- Honed, are transmitted in equal proportions in all directions through the mass. The Nichol's prism was next turned so that the plane of its principal section lay horizontal, the rays passing through it being therefore polarized in a vertical plane. The deflection produced by these rays, after having traversed the brown rock-crystal in a line parallel to its axis, was 9°'68. When however the line of transmission was perpendicular to the axis, the observed deflection was only 7°'03, — the ratio of these numbers is that of 100 : 7^. Hence the difference already noticed with the unpolarized rays, not only exhibits itself here once more, but is shown in an in- creased degree. In the case of beryl, the ratio of the quantity transmitted, under the same conditions, parallel to the axis to that transmitted perpendicular to the same, was as 100 : 21. In the case of tourmaline, the proportion was as 1 00 : 2 1 9. The following table contains the results of these and of the former observations : — Name of crystal. Heat polarized in I MQ+„«ai ra-ne I Heat polarized in horizontal plane. | J^iaturai rays. | vertical plane. Proportionate quantities of heat transmitted through the crystal — "«' I 'IK to the horizontal cry- stallographic axis. to the horizontal cry- stallographic axis. to the horizontal cry- stallographic axis. Brown rock-crvstal. Beryl .' Tourmaline 100 : 100 100 : 70* 100 : 100 100 : 92 100 : 54 100 : 158 100 : 73 100 : 21 100 : 219 Difference due to the accidental irregularities of the specimen examined. OF RADIANT HEAT THROUGH CRYSTALS. 103 A comparison of these numbers leads us to the conclusion, tJiat the differences between the quantities transmitted are greatest when the heat is polarized in the manner last described, less in the case of the natural rays, and vanish altogether when the rays are polarized in the manner first described. The essential difference between the two cases w^here polarized heat was applied is to be traced to the fact, that in the first case the plane of polarization coincides with the axis of the crystal, whereas in the second case the axis was perpendicular to the plane of polarization. The question now arises, whether dif- ferences exist in the quantities transmitted along the axis, when the heat is polarized in different planes. To solve this point the crystal was placed with its axis in the line of transmission, the principal section of the Nichol was set successively vertical and horizontal, and the quantities of heat transmitted in both cases, respectively, were measured and compared. No difference which w^ould permit of the conclusion that the quantity transmitted along the axis was at all affected by the position of the plane of polarization, was observed*. When the transmissive power in different directions perpendi- cular to the axis of the crystal was determined, no difference was perceptible, it mattered not whether the light was in its natural or in a polarized condition. For all these directions the same result was obtained as that already communicated for one of them. The foregoing results rendered another question of interest to me, and that is, whether the calorific rays whose quantitative dif- ferences were exhibited in the foregoing experiments, have also qualitative differences impressed upon them by passing,in the one or the other direction, through the crystals above-mentioned f. * Slight differences in the thermoscopic results were sometimes observed according as the principal section of the Nichol was vertical or horizontal ; but this was solely due to the slight disturbing influence caused by the elliptic polarization of the rays already noticed. Even without the rock-crystal, beryl, or tourmaline, these differences showed themselves when the Nichol was turned ; and the ratios of the observed values were not in the least altered by the intro- duction of the crystal in the manner indicated, — a result established in the most decided manner by exact measurements in the case of each. t In my earlier experiments with transparent calcareous spar no such dif- ference was observable. — Pogg. Ann. vol. Ixxiv. pp. 185, 18G. 12 104 KNOBLAUCH ON THE TRANSMISSION To ascertain this, I made use of a process which on former occasions enabled me to detect very small differences in the properties of radiant heat. The principle of the procedure consists simply in ascertaining whether the calorific rays whose qualities were to be compared, possessed the same power of transmission through diathermanous substances. The experiments commenced, as before, with brown rock- crystal. The deflection produced by the heat which traversed the crystal parallel to its axis, was first observed and found to be 34°*7l' A diathermanous body, a plate of blue glass for ex- ample, was then introduced between the crystal and the pile ; a portion of the heat was thus held back, the pile being excited by those rays only whic!i had passed through the blue glass. The deflection became reduced in consequence to 17°' 19. Hence the quantity of heat transmitted through the crystal and falling on the forward surfaces of the glass, is to the quantity which emanates from the latter in the proportion of 34°' 71 : 17^*19? or of 100:51. The calorific rays were now permitted to traverse the crystal in a line perpendicular to its axis. Their direct action upon the thermoscope produced a deflection of 33°'58. When the blue glass was introduced the deflection was reduced to 17°*21. Ac- cording to this, the quantity of heat falling on the forward sur- face of the glass, is to the quantity transmitted through it in the ratio of 33*58 : 17*21, or as 100 : 5 1, exactly as in the former case. In the same manner the influence of other diathermanous bodies, of yellow, red and green glass, was ascertained. The re- sults of these experiments are contained in the following table ; the quantity of heat transmitted through the rock-crystal and falling upon the diathermanous plate is, as before, expressed by 100, and each number is the mean of about six observations. Diathennanous bodies. Ratio of the quantity of heat which falls upon the diather- manous body to the quantity which passes through the latter, the heat having been first transmitted through rock-crystal Parallel | Perpendicular to the crystallographic axis. Blue glass 100 : 51 100 : 79 100 : 65 100 : 14 100 : 51 100 : 81 100 : 64 100 : 13 Yellow glass Red glass Green glass In none of these cases does the observed difference exceed OF ifADIANT HEAT THROUGH CRYSTALS. 105 what may be attributed to errors of observation. It remained thus doubtful whether the calorific rays which penetrate the rock-crystal parallel and perpendicular to the axis are to be re- garded as qualitatively alike, or whether their differences had not merely escaped observation. For the further examination of this point, I made use of heat similarly polarized to that which in the former experiments ex- hibited the greatest quantitative differences. The Nichol's prism was set in the path of the rays with its principal section horizontal, and the heat, thus polarized in a vertical plane, was permitted to traverse the crystal. Having observed the deflection produced under these conditions, one of the glass plates already made use of was introduced, as before, between the rock-crystal and the thermic pile, and the new de- flection was observed. The question to be decided was, whether the heat, under these circumstances, would exhibit any essential differences as to its capability of transmission through the before-named diathermanous substances, according as the line of transmission through the crystal was parallel or perpendicular to its axis. The following table exhibits the results of the ob- servations. Here, as in the former instances, the ratio of the quantity of heat falling on the diathermanous body to the quan- tity transmitted through it, is given ; the former being expressed by 100. Every number is the result of about thirteen observations. Diathermanous bodies. Ratio of the quantity of heat which, polarized in a vertical plane, falls upon the diathermanous body, to that which passes through the latter, the incident heat having tirst passed through a cube of rock-cr) stal Parallel | Perpendicular to the horizontal crystallographic axis. IJlue class 100 : 40 ICO : 73 100 : 42 100 : 7 100 : 39 100 : 78 100 : 45 100 : 11 Yellow glass Kpf] fflass .. ... Green glass Here, comparing the first series of ratios with the second, differences are observed which lie without the limits of error due to observation. Hence calorific rays which have passed through the brown rock-crystal in different directions possess, in different degrees, the power of penetrating one and the same diathermanous body. Thus, for example, the heat which has passed through the rock-crystal perpendicular to the axis, passes in greater quantity through the green glass than that which has traversed the crystal 106 KNOBLAUCH ON THE TRANSMISSION parallel to its axis ; for while in the former case 1 1 out of every 100 incident rays are transmitted, in the latter case only 7 per cent, pass through. It is thus proved that the rays in question differ from each other in respect to quality. On the other hand, no perceptible difference is exhibited when the heat, instead of being polarized in a vertical plane, is polarized in a horizontal one. Experiments conducted exactly as the foregoing, and differing only in the circumstance that the principal section of the Nichol instead of being horizontal was vertical, gave the following values, each of which is the result of seven observations : — Diathermanous bodies. Ratio of the quantity of heat which, polarized in a hori- zontal plane, falls upon the diathermanous body, to that which passes through the latter, the incident heat having first passed through a cube of rock-crystal Parallel | Perpendicular to the horizontal crystallographic axis. Blue glass 100 : 43 100 : 77 100 : 50 100 : 10 100 : 42 100 : 77 100 : 50 100 : 9* Yellow glass Red glass Green glass The numbers of the two columns differ only within the limits of the error of observation. Hence we conclude that calorific rays whose plane of polarization is parallel to the axis of the rock-crystal are homogeneous, whatever be the direction in which they have traversed the crystal. Just as little do differences in the properties of the calorific rays which have traversed the crystal parallel to its axis exhibit them- selves, whether the said rays have been polarized in a vertical or in a horizontal plane. This is proved by the following table ; the values contained in it are each the result of four single determinations. DiathermanouK bodies. Ratio of the quantity of beat which falls upon a diather- manous body, to that which passes through the latter, being first transmitted along the axis of the rock-crystal Polarized in vertical plane. | Polarized in horizontal plane. Blue glass Yellow glass Red crlass... 100 : 40 100 : 71 100 : 53 100 : 10 100 : 40 100 : 71 100 : 53 100 : 10 • The different tables cannot be compared in every respect with each other, as they are the results of observations made on different days, on which it was impossible to preserve external conditions perfectly constant. OF RADIANT HEAT THROUGH CRYSTALS. 10? The equality of these ratios proves that the rays in both cases possess the same capabiUty of transmission through the same diathermanous substance. As long as the direction of the rays was perpendicular to the axis of the crystal^ there was no difference observed, it mattered not how the direction was otherwise altered. The result true for one of these directions, and mentioned in its proper place, was true for all of them. Beryl was next submitted to precisely the same examination as that carried out in the case of rock-crystal. In the first instance, I permitted the natural rays to traverse the crystal parallel and perpendicular to the axis, and tested their deportment in the manner already described. After the full description given of the mode of experiment with rock-crystal, no obscurity can exist about it in the present in- stance. From the observations with beryl, we obtain the fol- lowing values, each of which is the mean of five distinct obser- vations : — Diathermanous bodies. Ratio of the quantity of heat which falls upon the diather- manous body, to that which passes through it, the heat having previously traversed a cube of beryl Parallel | Perpendicular to the crystallographic axis. Blue crlass 100 : 28 100 : 56 100 : 29 100 : 50 Yellow glass From these numbers we draw the inference, that the heat which has passed through the beryl in the di7'ection of its axis is trans- mitted through a diathermanous body in a different proportion from that which has traversed the crystal perpendicular to its axis. For, of the first-mentioned class of rays, 56 out of 100 are transmitted through the yellow glass, while of the latter class only 50 rays out of 100 pass through the same medium. The calorific rays, in this instance, are therefore to be regarded as qualitatively unlike. The difference exhibited itself in a still more striking manner, where the heat, before it entered the crystal (the axis of which, as before, was horizontal), had been polarized in a vertical plane. The following table contains the results of this observation : — 108 KNOBLAUCH ON THE TRANSMISSION I Ratio of the quantity of heat which, polarized in a vertical plane, falls upon the diathernianous body, to that which Diathernian(ni8 bodies. P'""''^ througli the latter, the heat having previously tra- I versed a cube of beryl Parallel | Perpendicular to the horizontal crystallographic axis. Blue glass..., Yellow glass 100 : 39 100 : 54 100 : 50 100 : 30 In this instance, the difference, as regards the transmissive power of both groups of rays, exhibits itself very distinctly. One group (that parallel to the axis of the crystal) is transmitted in less quantity through the blue glass than the other, while the latter is transmitted through the yellow glass more sparinglj^ than the former. The differences in the present case are, in them- selves, greater than when unpolarized heat was made use of. Thus, in the case of unpolarized heat, out of 100 rays incident on the yellovv glass, 56 were in one position transmitted, and 50 in the other position ; but in the present case the numbers corresponding to these positions are 54 and 30. All differences of this kind again disappear when the heat, instead of being polarized in a vertical plane, is polarized in a horizontal one. This is manifest from the following numbers, which differ from each other, in both compartments, only within the limits of the error of observation. Diathermanous bodies. Ratio of the quantity of heat which, polarized in a hori- zontal plane, falls upon the diathermanous body, to that which passes through the latter, the heat having previously traversed the beryl Parallel I Perpendicular to the horizontal crystallographic axis. Blue glass Yellow glass 100 : 43 100 : 63 100 : 43 100 : 61 Hence in the case of beryl also, we find that the calorific rays whose plane of polarization coincides with the axis of the crystal are homogeneous, although they may have passed through the crystal in different directions. The same similarity of deportment is exhibited by all rays which have traversed the crystal parallel to its axis, whatever be their plane of polarization, Tlie following table contains the observations which refer to this point : — OF RADIANT HEAT THROUGH CRYSTALS. 109 Diatherraanous bodies. Ratio of the quantity of heat which, polarized in a vertical | horizontal plane, falls upon the diatherraanous body, to that which passes through the latter, having previously traversed the beryl along its horizontal axis. 100 : 35 100 : 42 100 : 35 100 : 40 Yellow glass Those rays which pass through the beryl perpendicular to the axis exhibit no differences, however the direction may be varied in other respects. The difference of deportment between such rays and those which pass along the axis, has already been pointed out. To the experiments heretofore described, another was added, by which the results already arrived at were partly corroborated and in part expanded. Two cubes of beryl, both cut in the manner already indicated, were so placed one after the other, that in one instance their axes were parallel and in another per- pendicular to each other. In the former case both axes were vertical, in the latter one was vertical and the other horizontal. The horizontal calorific rays proceeding from the heliostat, tra- versed both crystals successively, forming an angle of 90° with each of the axes. Behind these cubes diathermanous bodies were introduced, as before, in order to ascertain the transmissive power of the rays which had passed through the cubes. The following table exhibits the ratios in which they passed through the blue or yellow glass. The rays incident upon the plates are, as in the former instances, set equal to 100, and to this standard the quantity transmitted is referred. Every number is the arithmetic mean of eight distinct observations. Diatherraanous bodies. Blue glass .. Yellow glass Ratio of the quantity of heat which falls upon the dia- thermanous body, to that which passes through the lat- ter, having previously traversed ooth the cubes of beryl with their axes Parallel. Perpendicular. 100 : 7 100 : 26 100 : 20 100 : 10 The differences here exhibited are too considerable to need particular discussion. Thus in this example also we find essential differences between those calorific rays which are permitted to pass through the cubes with their axes parallel, and those which traverse the cubes when the axes are perpendicular to each other. 110 KNOBLAUCH ON THE TRANSMISSION The examination of tourmaline led to the same results as those observed with the other crystals. The results of the observations with natural rays are contained in the following table: each number is the mean of five distinct determinations : — Diathennanous bodies. Ratio of the quantity of heat which falls upon the dia- thermanous body, to that which passes through the same, having previously passed through the cube of tourmaline Parallel | Perpendicular to the crystallographic axis. Red glass Green glass 100 : 62 100 : 29 100 : 43 100 : 36 The comparison of both series shows plainly in how different degrees the rays which pass parallel to the axis, and those which pass perpendicular thereto, penetrate the same diathermanous body. The former rays pass through the red glass in greater quantity than the latter ; while, in the case of green glass, the latter rays have the preponderance. In the same manner, but in an increased degree, these differ- ences show themselves w^hen the calorific rays are polarized in a vertical plane and transmitted through the tourmaline with its axis horizontal. The following greatly divergent values place this beyond all doubt : — Diathermanous bodies. Ratio of the quantity of heat which, polarized in a vertical plane, falls upon the diathermanous body, to that which passes through the latter, having previously traversed the tourmaline Parallel | Perpendicular to the horizontal crystallographic axis. 100 : 14 100 : 19 100 : 6 100 : 41 Green &rlass The differences disappear in this case also when the plane of polarization is horizontal, that is, when it is parallel to the cry- stallographic axis. This is shown by the following numbers, whose difi'erences lie within the limits of the error of obser- vation : — • Diathermanous bodies. Red glass .. Green glass Ratio of the quantity of heat which, polarized in a hori- zontal plane, falls upon the diathermanous body, to that which passes through the latter, having previously traversed the tourmaline Parallel | Perpendicular to the horizontal crystallographic axis. 100 : 19 100 : 9 100 : 16 100 : 10 OF RADIANT HEAT THROUGH CRYSTALS. Ill Just as little are the calorific rays which pass along the axis of the tourmaline distinguishable from each other, whatever may have been their plane of polarization. Nor are differences ob- served among those which pass perpendicular to the axis, whatever be their direction in other respects. But these two groups, when compared with each other, exhibit, as already shown, striking qualitative differences. Experiments were next conducted with two plates of tour- maline, through which the calorific rays were permitted to pass, first when the axes of the plates were parallel, and afterwards with the axes perpendicular to each other. The following num- bers exhibit the proportions in which the rays, in both cases, re- spectively, pass a diathermanous body. Each number is the arithmetic mean of eight distinct determinations : — Diathermanous bodies. Ratio of the quantity of heat which falls upon the diather- manous body, to that which passes through the latter, having previously traversed the tourmaline with their axes Parallel. | Perpendicular. Red glass 100 : 30 100 : 19 100 : 42 100 : 11 Green glass The comparison of both series permits no doubt to exist as to the different capacities of transmission of both groups of rays. They are therefore to be regarded as unlike. Besides calcareous spar, rock-crystal, beryl and tourmaline, dichroite was examined : so far as the examination of this crystal extends, qualitative differences, dependent upon the direction of transmission, have also been observed. The principal results of this investigation may be thus summed up:— 1. Radiant heat passes through certain crystals, as brown rock-crystal, beryl, tourmaline and dichroite, in unequal quan- tities in different directions, and shows itself (for example in its deportment towards diathermanous bodies) to be of a different quality, according as it has traversed the crystal in one or an- other direction. These differences are connected with the polar- ization of the calorific rays, and in connexion with this point we find— 2. That calorific rays whose direction is perpendicular to the 112 KNOBLAUCH ON THE TRANSMISSION axis of the brown rock-crystal, beryl or tourmaline, when their plane of polarization forms an angle of 90° with the axis of the crystal, are transmitted in quite different proportions from those observed when the line of transmission is parallel to the axis. The rays traverse the crystal in all directions alike when their plane of polarization coincides with the axis of the crystal. 3. In the first case the greatest qualitative differences manifest themselves, in the second case no such differences have an existence. 4. When the rays are transmitted along the axis, no differences are observed either as regards quality or quantity, whatever may be the position of the plane of polarization. 5. Compared among themselves, the different directions per- pendicnlar to the axis exhibit no differences of transmission in the three crystals above-mentioned. When we reflect that (according to former investigations) the natural calorific rays, on passing through the crystals perpendi- cular to their axes, are decomposed by double refraction into two groups, and so polarized that the plane of polarization of the one group coincides with the axis of the crystal, while that of the other group is perpendicular to the same ; further, that this does not take place when the direction of transmission is along the axis ; this, in connexion with what has gone before, explains the deportment observed in the case of the natural calorific rays. Those rays which proceed parallel to the axis of the crystal undergo a certain absorption which depends upon the nature of the crystals, and determines the quantity and character of the transmitted rays, — determines, for instance, their deportment towards diathermanous bodies. In passing perpendicular to the axis of the crystal, the two groups of rays are absorbed in dif- ferent proportions. Those whose plane of polarization coincides with the crystallographic axis suffer the same absorption as the rays which pass along the axis ; the other rays, however, will be absorbed either in a greater degree, as in the case of brown rock- crystal and beryl, or in a less degree, as in the case of tourmaline. Further, as this absorption is elective, that is to say, directed in different degrees towards the different rays of one and the same group, the composition of such a group will be modified, and hence its capacity to penetrate a diathermanous body ^\ill be OP RADIANT HEAT THROUGH CRYSTALS. 113 altered. In this respect also the rays whose plane of polarization coincides with the axis are not to be distinguished from those which pass parallel to the axis. The others however exhibit essential differences. If the two groups of rays remain unse- parated, their total action being alone examined, as in the expe- riments with the natural rays, we must arrive at results which, as regards quantity and quality, lie between those obtained from observations made with the separate polarized groups, where in one instance no difference, and in another the maximum differ- ence is exhibited. The values above quoted furnish abundant proof of this. [J-T.] 114 Article IV. On the Conservation of Force ; a Physical Memoir. By Dr. H. Helmholtz. [Read before the Physical Society of Berlin on the 23rd of July, 1847. Berlin. G. Reimer.] Contents. Introduction. I. The principle of the Conservation of vis viva, II. The principle of the Conservation of Force. III. The application of the principle in Mechanical Theorems. IV. The Force-equivalent of Heat. V. The Force-equivalent of the Electric Processes. VI. The Force-equivalent of Magnetism and Electro-magnetism. Introduction. 1 HE principal contents of the present memoir show it to be addressed to physicists chiefly, and I have therefore thought it judicious to lay down its fundamental principles purely in the form of a physical premise, and independent of metaphysical considerations, — to developethe consequences of these principles, and to submit them to a comparison with what experience has established in the various branches of physics. The deduction of the propositions contained in the memoir may be based on either of two maxims ; either on the maxim that it is not possible by any combination whatever of natural bodies to derive an unlimited amount of mechanical force, or on the assumption that all actions in nature can be ultimately referred to attractive or repulsive forces, the intensity of which depends solely upon the distances between the points by which the forces are exerted. That both these propositions are identical is shown at the com- mencement of the memoir itself. Meanwhile the important bearing which they have upon the final aim of the physical sciences may with propriety be made the subject of a special introduction. The problem of the sciences just alluded to is, in the first place, to seek the laws by which the particular processes of HELMHOLTZ ON THE CONSERVATION OF FORCE. 115 nature may be referred to, and deduced from, general rules. These rules, — for example, the law of the reflexion and refraction of light, the law of Mariotte and Gay-Lussac regarding the volumes of gases, — are evidently nothing more than general ideas by which the various phaenomena which belong to them are con- nected together. The finding out of these is the office of the experimental portion of our science. The theoretic portion seeks, on the contrary, to evolve the unknown causes of the processes from the visible actions which they present ; it seeks to compre- hend these processes according to the laws of causality. We are justified, and indeed impelled in this proceeding, by the con- viction that every change in nature must have a sufficient cause. The proximate causes to which we refer phaenomena may, in themselves, be either variable or invariable ; in the former case the above conviction impels us to seek for causes to account for the change, and thus we proceed until we at length arrive at final causes which are unchangeable, and which therefore must, in all cases where the exterior conditions are the same, produce the same invariable effects. The final aim of the theoretic natural sciences is therefore to discover the ultimate and un- changeable causes of natural phaenomena. Whether all the pro- cesses of nature be actually referrible to such, — whether nature is capable of being completely comprehended, or whether changes occur which are not subject to the laws of necessary causation, but spring from spontaneity or freedom, this is not the place to decide ; it is at all events clear that the science whose object it is to comprehend nature must proceed from the assumption that it is comprehensible, and in accordance with this assumption investigate and conclude until, perhaps, she is at length admo- nished by irrefragable facts that there are limits beyond which she cannot proceed. Science regards the phaenomena of the exterior world according to two processes of abstraction : in the first place it looks upon them as simple existences, without regard to their action upon our organs of sense or upon each other; in this aspect they are named matter. The existence of matter in itself is to us some- thing tranquil and devoid of action : in it we distinguish merely the relations of space and of quantity (mass), which is assumed to be eternally unchangeable. To matter, thus regarded, we 116 HKLMHOLTZ OX THE CONSERVATION OF FORCE. must not ascribe qualitative differences, for when we speak of different kinds of matter we refer to differences of action, that is, to differences in the forces of matter. Matter in itself can there- fore partake of one change only, — a change which has reference to space, that is, motion. Natural objects are not, however, thus passive ; in fact we come to a knowledge of their existence solely from their actions upon our organs of sense, and infer from these actions a something which acts. When, therefore, we wish to make actual application of our idea of matter, we can only do it by means of a second abstraction, and ascribe to it properties which in the first case were excluded from our idea, namely the capability of producing effects, or, in other words, of exerting force. It is evident that in the application of the ideas of matter and force to nature the two former should never be separated : a mass of pure matter would, as far as we and nature are con- cerned, be a nullity, inasmuch as no action could be wrought by it either upon our organs of sense or upon the remaining portion of nature. A pure force would be something which must have a basis, and yet which has no basis, for the basis we name matter. It would be just as erroneous to define matter as something which has an actual existence, and force as an idea which has no corre- sponding reality. Both, on the contrary, are abstractions from the actual, formed in precisely similar ways. Matter is only discernible by its forces, and not by itself. We have seen above that the problem before us is to refer back the phaenomena of nature to unchangeable final causes. This requirement may now be expressed by saying that for final causes unchangeable forces must be found. Bodies with un- changeable forces have been named in science (chemistry) ele- ments. Let us suppose the universe decomposed into elements possessing unchangeable qualities, the only alteration possible to such a system is an alteration of position, that is, motion ; hence, the forces can be only moving forces dependent in their action upon conditions of space. To speak more particularly: the phaenomena of nature are to be referred back to motions of material particles possessing unchangeable moving forces, which are dependent upon con- ditions of space alone. Motion is the alteration of the conditions of space. Motion, HELMHOLTZ ON THE CONSERVATION OF FORCE. 117 as a matter of experience, can only appear as a change in the relative position of at least two material bodies. Force, which originates motion, can only be conceived of as referring to the relation of at least two material bodies towards each other ; it is therefore to be defined as the endeavour of two masses to alter their relative position. But the force which two masses exert upon each other must be resolved into those exerted by all their particles upon each other; hence in mechanics we go back to forces exerted by material points. The relation of one point to another, as regards space, has reference solely to their distance apart : a moving force, therefore, exerted by each upon the other, can only act so as to cause an alteration of their distance, that is, it must be either attractive or repulsive. Finally, therefore, we discover the problem of physical natural science to be, to refer natural phaenomena back to un- changeable attractive and repulsive forces, whose intensity depends solely upon distance. The solvability of this problem is the condition of the complete comprehensibility of nature. In mechanical calculations this limitation of the idea of moving force has not yet been assumed : a great number, however, of general principles referring to the motion of compound systems of bodies are only valid for the case that these bodies operate upon each other by unchangeable attractive or repulsive forces ; for example, the principle of virtual velocities ; the conservation of the motion of the centre of gravity ; the conservation of the principal plane of rotation ; of the moment of rotation of free systems, and the conservation of vis viva. In terrestrial matters application is made chiefly of the first and last of these prin- ciples, inasmuch as the others refer to systems which are sup- posed to be completely free ; we shall however show that the first is only a special case of the last, which therefore must be regarded as the most general and important consequence of the deduction which we have made. Theoretical natural science therefore, if she does not rest con-^ tented with half views of things, must bring her notions into harmony with the expressed requirements as to the nature of simple forces, and with the consequences which flow from them. Her vocation will be ended as soon as the reduction of natural phaenomena to simple forces is complete, and the proof given SCIEN. MEM.— .Va^ Phil. Vol. I. Part II. K 118 HELMHOLTZ ON THE CONSERVATION OF FORCE. that this is the only reduction of which the phaenomena are capable. I. The principle of the Conservation of vis viva. We will set out with the assumption that it is impossible, by any combination whatever of natural bodies, to produce force continually from nothing. By this proposition Carnot and Clapeyron have deduced theoretically a series of laws, part of which are proved by experiment and part not yet submitted to this test, regarding the latent and specific heats of various natural bodies. The object of the present memoir is to carry the same principle, in the same manner, through all branches of physics ; partly for the purpose of showing its applicability in all those cases where the laws of the phaenomena have been sufficiently investigated, partly, supported by the manifold analogies of the known cases, to draw further conclusions regarding laws which are as yet but imperfectly known, and thus to indicate the course which the experimenter must pursue. The principle mentioned can be represented in the following manner : — Let us imagine a system of natural bodies occupying certain relative positions towards each other, operated upon by forces mutually exerted among themselves, and caused to move until another definite position is attained ; we can regard the velocities thus acquired as a certain mechanical work and trans- late them into such. If now we wish the same forces to act a second time, so as to produce again the same quantity of work, we must, in some way, by means of other forces placed at our disposal, bring the bodies back to their original position, and in effecting this a certain quantity of the latter forces will be con- sumed. In this case our principle requires that the quantity of work gained by the passage of the system from the first position to the second, and the quantity lost by the passage of the system from the second position back again to the first, are always equal, it matters not in what way or at what velocity the change has been effected. For were the quantity of work greater in one way than another, we might use the former for the pro- duction of work and the latter to carry the bodies back to their primitive positions, and in this way procure an indefinite amount of mechanical force. We should thus have built a perpetuum b HELMHOLTZ ON THE CONSERVATION OF FORCE. 119 mobile which could not only impart motion to itself, but also to exterior bodies. If we inquire after the mathematical expression of this prin- ciple, we shall find it in the known law of the conservation of vis viva. The quantity of work which is produced and consumed may, as is known, be expressed by a w^eight m, which is raised to a certain height h-, it is then mgh, where g represents the force of gravity. To rise perpendicularly to the height h, the body m requires the velocity v= \^2gh, and attains the same by falling through the same height. Hence we have -rrw^^mgh; and hence we can set the half of the product mv^, which is known in mechanics under the name of the vis viva of the body m, in the place of the quantity of work. For the sake of better agreement with the customary manner of measuring the intensity of forces, I propose calling the quantity -mv'^ the quantity of vis viva, by which it is rendered identical with the quantity of work. For the applications of the doctrine of vis viva which have been hitherto made this alteration is of no import- ance, but we shall derive much advantage from it in the following. The principle of the conservation of vis viva, as is known, de- clares that when any number whatever of material points are set in motion, solely by such forces as they exert upon each other, or as are directed against fixed centres, the total sum of the vires vivce, at all times when the points occupy the same relative position, is the same, whatever may have been their paths or their velocities during the intervening times. Let us suppose the vires vivce applied to raise the parts of the system or their equivalent masses to a certain height, it follows from what has just been shown, that the quantities of work, which are repre- sented in a similar manner, must also be equal under the con- ditions mentioned. This principle however is not applicable to all possible kinds of forces ; in mechanics it is generally derived from the principle of virtual velocities, and the latter can only be proved in the case of material points endowed with attractive or repulsive forces. We will now show that the principle of the conservation of vis viva is alone valid where the forces in action may be resolved into those of material points which act K 2 120 HELMHOLTZ ON THE CONSERVATION OF FORCE. in the direction of the lines which unite them, and the intensity of which depends only upon the distance. In mechanics such forces are generally named central forces. Hence, conversely, it follows that in all actions of natural bodies upon each other, w here the above principle is capable of general application, even to the ultimate particles of these bodies, such central forces must be regarded as the simplest fundamental ones. Let us consider the case of a material point with the mass m, which moves under the influence of several forces which are united together in a fixed system A; by mechanics we are enabled to determine the velocity and position of this point at any given time. We should therefore regard the time t as pri- mitive variable, and render dependent upon it, — the ordinates 37, y, ;2 of m in a system of coordinates, definite as regards A, the tangential velocity g, the components of the latter parallel to the axes, «<= — , '^=-^9 ^=777? ^^^ finally the components of the acting forces V du ^^ dv rw dw ■^^^-JJ^ Y=m-7-, Z=m-7:. dt' dt^ dt Now according to our principle - m (f, and hence also g^ must be always the same when m occupies the same position relative to A ; it is not therefore to be regarded merely as a function of the primitive variable /, but also as a function of the coordinates 07, y, z only ; so that ^<.-)=f-'^-f-'*4f- ■•■<■) As 5'2 _- ^2 _f. ^2 ^ yji^ ^g jj^^^g ^ ^^2j _ 2 udu + 2vdv + 2 wdw. Instead of u let us substitute its value ^, and instead of du its dt value -^, the corresponding values of v and w being also used, we have dr (2) Let us regard this equation more closely ; we find at the left- hand side the difference of the vires vivcB possessed by m at two different distances. To understand the import of the quantity R / (f>dr, let us suppose the intensities of (ft which belong to dif- ferent points of the connecting line ma erected as ordinates at these points, then the above quantity would denote the super- ficial content of the space enclosed between the two ordinates r and R. As this surface may be regarded as the sum of the infinite number of ordinates which lie between r and R, it therefore represents the sum of the intensities of the forces which act at all distances between R and r. Calling the forces which tend to move the point m^ before the motion has actually taken place, tensions, in opposition to that w^hich in mechanics is named vis viva, then the quantity / dr would be the sum of the tensions between the distances R and r, and the above law would be thus expressed : — The increase of vis viva of a material point during its motion under the influence of a central force is equal to the sum of the tensions which correspond to the alteration of its distance. Let us suppose the case of two points operated upon by an attractive force, at the distance R ; by the action of the force they will be drawn to less distances r, their velocity, and conse- quently vis viva, will be increased ; but if they should be driven to greater distances 7-, their vis viva must diminish and must finally be quite consumed. We can therefore distinguish, in the case of attractive forces, the sum of the tensions for the distances between r =0 and r = R, / <^dr, as those which yet remain, but HELMHOLTZ ON THE CONSERVATION OF FORCE. 123 those between r = R and r=oo as those already consumed; the former can immediately act, the latter can only be called into action by an equivalent loss of vis viva. It is the reverse with repulsive forces. If the points are situate at the distance R, as the distance becomes greater vis viva will be gained, and the still existing tensions are those between r=^R and r^oo, those lost are between r = 0 and r=R. To carry our law through in quite a general manner, let us suppose any number whatever of material points with the masses nil, ^2J ^39 ^^' denoted generally by m^ ; let the components of the forces which act upon these parallel to the axes be X„, Y„, Z^, the components of the velocities along the same axes u^, v^, w^, the tangential velocity q^ ; let the distance between m^ and nif, be 7'ad) the central force between both being c^a^. For the single point nin we have, analogous to equation (1), V vf/ \(l>anl dVn Y„=SL(y„-2,„)-J=m„^, where the sign of summation X includes all members which are obtained by putting in the place of the index a the separate in- dices 1, 2, 3, &c., with the exception of w. Multiplying the first equation by dx^ =u^dt, the second by dyn—Vndt, the third by dZn=Wndt, and supposing the three equa- tions thus obtained to be formed for every single point of m*, as it is already done for m„ ; adding all together, we obtain 2[(z.-z,)rfz,g] =2gm,rf(c«,\)]. The members of the left-hand series will be obtained by placing instead of a all the single indices 1, 2, 3, &c., and in each case 12 i HELMHOLTZ ON THE CONSERVATION OF FORCE. for b also all the values of b, which are greater or smaller than a already possesses. The sums divide themselves therefore into two portions, in one of which a is always greater than b, and in the other always smaller, and it is clear that for every member of the one portion a member rpq must appear in the other portion : adding both together, we obtain -{^p-x,)(,h^-dx,)^: rpq drawing the sums thus together, adding all three and setting \d^{xa-.x,f + {y^-y,Y^ (Za-z,y^ =r«,c/r„„ we obtain --S\j>a,dr^'\ = ^\^mj{q\)\, (3) or -2[^j'"Va»<^'-»»]=2[i™,Q\]-2[^'m,g\], . (4) where R and Q, as well as r and q, denote contemporaneous values. We have here at the left-hand side again the sum of the tensions consumed, on the right the vis viva of the entire system, and we can now express the law as follows : — In all cases of the motion of free material points under the influence of their at- tractive and repulsive forces, whose intensity depends solely upon distance, the loss in tension is always equal to the gain in vis viva, and the gain in the former equal to the loss in the latter. Hence the sum of the existing tensions and vires vivce is always constant. In this most general form we can distinguish our law as the principle of the conservation of force. In the deduction of the law as given above, nothing is changed if a number of the points, which we will denote generally by the HELMHOLTZ ON THE CONSERVATION OF FORCE. 125 letter d, are supposed to be fixed, so that q^ is constantly =0; the form -of the law will then be S [atdrab'] + 2 [^„d^r J = - 2 [^^.^^(g^,)] . . . (5) It remains to be shown in what relation the principle of the conservation of force stands to the most general law of statics, the so-called principle of virtual velocities. This follows imme- diately from equations (3) and (5). If equilibrium is to set in when a certain arrangement of the points ma takes place, that is, if in case these points come to rest, hence 5'a = 0, they remain at rest, hence dqa=0, it follows from equation (3), ^ 2[<^a6flfr«J=0; . / (6) or in case that forces act upon them from points m^ without the system, by equation (5), S[(^„At/r«,]+5;[<^arf^rJ=0 (7) In these equations under dr are understood alterations of distance consequent on the small displacements of the point rria, which are permitted by the conditions of the system. We have seen, in the former deductions, that an increase of vis viva, hence a transition from rest to motion, can only be effected by an ex- penditure of tension ; in correspondence with this, the last equations declare that in cases where in no single one of the possible directions of motion tension in the first moment is con- sumed, the system once at rest must remain so. It is known that all the laws of statics may be deduced from the above equations. The most important consequence as regards the nature of the acting forces is this : instead of the arbitrary small displacements of the points m, let us suppose such intro- duced as might take place were the system in itself firmly united, so that in equation (7) every rfr„4 = 0, it follows singly, 2[(/)„6fl?raJ=0, and ^[