<'< <>;/« ;. 1 S DITED BY I KAY LANKBSTER LIBRARY UNIVERSWY OF CALIFORNIA SANTA CRU1 VJ fin BOOSS I .,.-••• -; A TKEATISE ON ZOOLOGY TEEATISE ON ZOOLOGY EDITED BY E. KAY LANKESTER M.A., LL.D., F.R.S. HONORARY FELLOW OF EXETER COLLEGE, OXFORD J CORRESPONDENT OF THE INSTITUTE OF FRANCE J DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM PART I INTRODUCTION AND PEOTOZOA SECOND FASCICLE BY J. B. FARMER, D.Sc., F.RS. PROFESSOR OF BOTANY, ROYAL COLLEGE OF SCIENCE, LONDON J. J. LISTER, M.A., RR.S. FELLOW OF ST. JOHN'S COLLEGE, CAMBRIDGE E. A, MINCHIN, M.A. PROFESSOR OF ZOOLOGY, UNIVERSITY COLLEGE, LONDON AND S. J. HICKSON, F.R.S. PROFESSOR OF ZOOLOGY, OWENS COLLEGE, MANCHESTER Reprint A. AS HER & CO. Amsterdam 1964 TEEATISE ON ZOOLOGY EDITED BY E. KAY LANKESTEK M.A., LL.D., F.R.S. HONORARY FELLOW OF EXETER COLLEGE, OXFORD J CORRESPONDENT OF THE INSTITUTE OF FRANCE J DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM PART I INTRODUCTION AND PROTOZOA SECOND FASCICLE BY J. B. FARMER, D.Sc., F.RS. PROFESSOR OF BOTANY, ROYAL COLLEGE OF SCIENCE, LONDON J. J. LISTER, M.A., F.RS. FELLOW OF ST. JOHN'S COLLEGE, CAMBRIDGE E. A. MINCHIN, M.A. PROFESSOR OF ZOOLOGY, UNIVERSITY COLLEGE, LONDON AND S. J. HICKSON, F.RS. PROFESSOR OF ZOOLOGY, OWENS COLLEGE, MANCHESTER LONDON ADAM & CHAELES BLACK 1903 Exclusive Agents for U.S.A. STECHERT-HAFNER SERVICE AGENCY, INC. 31 East 10th Street New York, New York 10003 Sole agents for India: Today & Tomorrow1 s Book Agency, 22-B/5, Original Road, Karol Bagh, New Delhi-5 PKEFACE TO SECOND FASCICLE OF PAKT I. INTEODUCTION AND PROTOZOA THE irregular publication of the parts of the Treatise on Zoology is the inevitable result of the fact that it is the work of a number of authors. I have determined not to allow Professor Minchin's most timely and valuable treatise on the Sporozoa to lie by in the printers' hands until the other sections of Part I. which logically precede it are ready for the press ; and with this I have been able to combine Dr. J. J. Lister's section on Foraminifera (which contains much that is new and original), Professor Hickson's section on Infusoria, and a section on the Structure of Animal and Vegetable Cells, with especial reference to the points which arise in the study of the Protozoa, by Professor Farmer. These four sections form the second fascicle of the First Part of this treatise ; the first fascicle, which is in preparation, will contain an Introduction and descriptions of the Proteo- myxa, Mycetozoa, Lobosa, Heliozoa, Labyrinthulidea, Eadio- laria, and Flagellata, forming sections A to G of Chapter I.— The Protozoa. The division of the work into Chapters, of which the second to the twenty-first are already published, has resulted in a somewhat awkward restriction of the Protozoa to nominally VI PREFACE one chapter, the first. This unduly large chapter is broken up into sections which serve instead of the usual division of so large a number of pages into chapters. The parts of the Treatise on Zoology dealing with the Mollusca, the Arthropoda, and the Vertebrata are in active preparation. E. RAY LANKESTER May 15, 1903. CONTENTS CHAPTER I.— PROTOZOA (continued) PAGE SECTION H. — THE STRUCTURE OF ANIMAL AND VEGETABLE CELLS . . . / . 1 „ I. — THE FORAMINIFERA „ ,«*: . . 47 K.— THE SPOROZOA . . ./ . 150 „ L. — THE INFUSORIA . . . .361 INDEX 427 CHAPTER I.— PROTOZOA (continued) SECTION H. — THE STRUCTURE OF ANIMAL AND VEGETABLE CELLS l IN reviewing the course of development of our knowledge of organic nature, there stands out one epoch-making discovery, that of the chambered structure of plants, made by Hooke in 1665, which was destined not only to profoundly modify the older con- ceptions as to the intimate organisation of animals and plants, but also to place in clear relief the fundamental unity which underlies FIG. 1. Facsimile of part of a figure by Hooke representing cells of vegetable tissues (cork). the apparently endless variety of external form. But, as in the case of most discoveries of wide-reaching import, the general recognition of the true nature of the cell did not emerge at once in its modern form, nor was it in reality the outcome of the work of any single investigator. Indeed, nearly two hundred years elapsed before the first enunciation of the doctrine of a cellular structure of plants 1 By J. B. Farmer, D.Sc., M.A., F.R.S. (1902). I THE STRUCTURE OF CELLS by Hookc became translated into a form comparable to that in which the phrase is now understood. Nevertheless to Hooke belongs the credit of having not only depicted the vesicular nature of cork and other plant-structures, but also of having designated the cavities by the name of cells. Malpighi and Grew in the succeeding century had -fully recog- nised the cellular character of plants, and even attempted a crude explanation of the origin of the cells themselves, likening them to the vesicular foam of beer. But accurate as was their portrayal of the mature structure, they nevertheless possessed no real concep- tion of the true meaning of the cell as the unit of organic life. The cells were regarded as the cavities in the matrix, not as the units which together constitute the organism, and it was to the wall that all their observations were directed. Little or no attention was seriously paid to the cell-contents. Thus, although Corti in 1772 had noticed the rotation of the viscous matter in the cells of Chara, his discovery remained without influence, and was made again, and independently, by Treviranus some forty years later. Even the discovery J of the nucleus by R. Brown in 1831-33 failed at once to excite the interest of the majority of his contemporaries, nor indeed does it appear that Brown himself at all fully grasped the signifi- cance of his discovery. Whilst in the plant-body it was the cellular structure, in the sense of Hooke, Malpighi, and Grew, which most forcibly appealed to the observer, the softer tissues composing an animal body were not so easily referable to a similar plan, although a consideration of the blood corpuscles, and of cartilage, helped to pave the way for the later generalisation. But, on the other hand, the animal body was more suited to turn the attention of the investigator upon the living substance, and the fundamental importance of the latter seems to have been first clearly appre- hended by Dujardin, who (in 1835), gave the name of Sarcode to the contractile, gelatinous, diaphanous mass constituting the bodies of the Infusoria which he was examining. He even succeeded in distinguishing some structural details, but with the lens at his command it may perhaps be doubted whether this really repre- sented more than the arrangement of granular and other inclusions in the living substance. It was, however, chiefly due to the researches of Schleiden, and especially of Schwann, which were published in 1837-38, that general interest became steadily focussed upon the cell-contents, including the nucleus which formed a cardinal point in their famous cell-theory. And it is largely to the great influence exerted by their work that the rapid advances witnessed during the next succeeding decades are legitimately to be traced. It is, of course, 1 Others, including Leeweuhoeck, had already seen nuclei in isolated cases, but their observations were quite without influence on the development of thought. THE STRUCTURE OF CELLS true that the greater part of their conclusions, especially such as are related to the genesis and growth of cells, have since turned out to have been erroneous. This is largely due, perhaps, to the weight of mistaken preconceptions on their part ; but the history of advance in any line of thought or science is full of similar examples. It is sufficient that they realised to the full the immense importance of the inquiry, and at any rate they succeeded in correlating and in co-ordinating a large mass of observations, and so became the means of immediately attracting numerous other workers into the same field. To Schwann may be conceded the merit of having first con- sciously attempted to demonstrate, in the most effective manner, the essentially similar character of the cells in plants and animals. This he did by endeavouring to follow out the origin and develop- ment of new cells in each of the two great divisions of living organisms, though how wide he was of the practical truth may be seen from the account which he gives of the process. The primordial substance out of which cells are formed consists, according to him, of a gelatinous or slimy mother- liquor, the cytollastema. In this, by a process of condensation, a nucleolus is first formed. This then grows by intussusception, and gives rise to a nucleus, in which once more a nucleolus is differentiated — itself the origin of another nucleus. Meantime, from the cyto- blastema fresh matter is deposited in the surface of the nucleus, and thus a consolidated membrane originates. This membrane, by intercalation of constantly increasing material within it, continues to grow, and ultimately it forms the wall of the new cell, the contents of which are provided for in the way just described. Thus, in the formation of cells, according to Schwann, the following stages, starting from the raw material — cytoblastema — may be distinguished. First, the condensation giving rise to the nucleolus, this in turn, by growth, produces the nucleus, and the peripheral (nuclear) wall eventually forms the wall of the new cell. At first sight it is difficult to realise how these ideas obtained the wide currency which they enjoyed, but the reason is to be sought in the fact that Schwann, like Schleiden and Nageli after him, was not fortunate in the material he selected for investigation. Cartilage, blood-corpuscles, and pollen grains were repeatedly studied, and it is perhaps not surprising that with such objects before them an incorrect conclusion was arrived at. Von Mohl, who had been engaged in studying the structure and mode of division of vegetable cells since 1835, at one time gave a true explanation of the process, but afterwards he sounded a less certain note, adhering to the view that the new cells were formed in toto within the mother cells, even in the case of algal filaments — an error which was definitely opposed by linger. Von Mohl THE STRUCTURE OF CELLS clearly recognised the importance of the formative substance of the cell, to which in 1846 he gave the name it now bears, viz. Protoplasm, the same word that had already been employed six years earlier by Pttrkinje to designate the formative substance met with in the animal embryo. Speculation was already aroused as to the possibility of instituting a comparison between vegetable protoplasm and animal sarcode, and Dutrochet had, as early as 1824, and still more definitely in 1837, indicated the general similarity which underlies the structure of animals and plants. But it was reserved for Cohn to clearly formulate in 1853 the real features of identity between them, and to express the definitely reasoned view that they were, in all essentials, composed of the same kind of substance. Cohn was well fitted for the task, by his acquaintance with the lower organisms of both animal and vegetable kingdoms. Max Schultze in 1861 further elaborated this resemblance, and his convincing demonstration at once gained the assent of all who were competent to form an opinion on the question. Moreover, Schultze clearly saw that it is the protoplasm (in the widest sense of the -term) which essentially constitutes a cell, and he, like Leydig, defined it as a mass of protoplasm containing a nucleus. About the same time also Virchow, in his celebrated aphorism, " Omnis cellula e celluia," crystallised the correct view as to the general mode of origin of new cells. But although the essential facts of cell-structure and development thus gradually emerged from the earlier and cruder notions, the finer details, and especially the relations of the nucleus, long remained obscure. The origin of this body, and its connec- tion with the rest of the cell-contents, was not understood, and a very general view was held that it disappeared (as indeed is in a certain sense correct) at each cell-division, to be formed afresh in the new daughter cells. It is true that so long ago as 1841 Remak had put forward the statement that the nucleolus and nucleus gave rise by direct fission to the corresponding structures in the daughter cells, and indeed that the whole process of cell -division was thus inaugurated ; but his views (which for a few cases are really well founded) appeared to be not generally applicable, and thus it transpired that even in the middle of the century the nucleus came to be commonly regarded as an organ of but secondary importance, and this even by so eminent an investigator as Brucke. It was not until the publication of Strasburger's mag- nificent work on the cell- and nuclear-division in 1875 that the nucleus received its proper share of attention. Strasburger, some four years later, like Virchow, in another connection, before him, defined the modern position in the phrase "Omnis nucleus e nucleo." The researches of the brothers Hertwig, Van Beneden, Flemming, THE STRUCTURE OF CELLS and others have abundantly emphasised and justified these far- reaching generalisations. But with the improvements which have been effected in technique during the last quarter of a century, new facts have come to light which have somewhat modified the conception of the cell as held by the earlier writers. It has been already seen how the centre of gravity gradually shifted from the cell-wall to the cell- contents, and that, as Max Schultze declared, the essential constituents of the cell were represented by the protoplasm and the nucleus, the wall being of altogether subordinate importance. The discovery that masses of protoplasm might contain not one but many nuclei, and that such a condition is not uncommon both amongst various groups and tissues in plants and animals, appeared to some writers to present a difficulty in accepting the cell theory as treated above, and various explanations have been offered in order to bring these cases into line with the theory as more generally understood and defined. Such organisms or tissues have been termed non-cellular — a negative and unsatisfactory expression which has been replaced by the more appropriate word syncytium or coenocyte. These words emphasise the view that, mor- phologically, the individual units, which collectively make up the syncytial tissue, are not isolated from each other by definite barriers. Sachs proposed the term energid to express the cell in Max Schultze's sense, meaning thereby the nucleus, together with the portion of protoplasm dominated by it. Essentially this is a physiological definition as contrasted with the morphological idea embodied in the word syncytium. And it is, on the whole, a legiti- mate expression, since it really does correspond to a fact. Moreover, it has the merit of being equally applicable to the cases of isolated cells as well as of those in which such limits are not structurally traceable. The objection raised to the conception. underlying it, on the ground that the nucleus of a syncytium does not always dominate the same protoplasm, is not a valid one, inasmuch as it is quite possible — perhaps even probable — that an essentially similar state of things also obtains even in those tissues in which the constituent cells are apparently isolated. For it has gradually been proved for a very large number of cases that the protoplasm of adjacent cells is in actual physical continuity through the fine pores present in the delimiting cell-walls. It is tolerably easy to observe this continuity in the epithelial cells of some amphibian tissues, and there is a con- siderable weight of evidence to show that it is far more general than was generally supposed to be the case. In plant-tissues also it has been repeatedly demonstrated since its first discovery by Tangl (in 1 8 7 9 ) in the endosperm of certain seeds. Gardiner and Russo w almost simultaneously demonstrated its existence in the tissues of several adult plants in 1883, and since that time it has been clearly proved to THE STRUCTURE OF CELLS occur throughout the tissues of the organism in those examples which have been specially investigated for the purpose. Thus it is evident that, so far from the syncytial condition presenting an exceptional case, it is in reality an extremely common one, the cell-walls merely forming a perforated skeletal framework which supports the softer parts. It is useless to argue (as has recently been done) that the pores are so fine as to be practically useless for the transference of material substances through them, since, as Horace Brown has shown, their very narrowness, taken together with the thinness of the portion of membrane on which they occur, is an important condition, for the performance of such a function without unduly weakening the framework itself. Moreover, the existence of the continuity of the protoplasm from cavity to cavity at once renders intelligible the Fio. 2. Continuity of protoplasm through vege- table cell-walls. A, cells of the pulvinus of Robinia. li, cells of the endosperm of Ueterospathf: (After Gardiner.) possibility of a transmission of stimuli from one part of the body to another, although it would perhaps be going too far to assume this as a necessary condition of transmission. Examples are known in which the stimuli appear to be less directly conveyed, as, for instance, in the case of nerve ganglia, according to Ramon y Cajal (although his results have not been upheld by some other investi- gators) ; and further, in some plant tissues, water squeezed out into the intercellular spaces has been regarded (on rather slender grounds) as being the means of exciting consecutive cells of a tissue. On the other hand, Nemec has recently shown reason for admitting, in the irritable parts of plants, the existence of specialised tracts of protoplasm which are continuous from cell to cell, and by the agency of which the stimulant impulse is conveyed to the motor executive region. Whilst the general and mutual relations of the constituent parts of the cells were being gradually elucidated, it became recog- nised that the cell-substances themselves were not composed of THE STRUCTURE OF CELLS mere structureless jelly, but possessed an organisation of their own. At first the recognition of this fact only appears in tentative sug- gestions, and hardly any serious progress was made beyond the obvious distinction of the nucleus from the protoplasm. Briicke seems to have been the first to point out the philosophical necessity of assuming an organisation in the protoplasm, but the visual per- ception of the counterpart of such a constitution hardly advanced beyond the recognition of a relatively solid mass bathed in a more fluid substance. The former was distinguished as spongioplasm, and the latter as hyaloplasm. It is significant of the difficulty experienced in arriving at a definite decision on the then available evidence that each of the two constituents has been claimed by different writers as the living substance. The views which have been put forward as to the relationships of the various substances which co-exist in the protoplasm to each other have developed in two principal directions. The earlier, his- torically speaking, was advocated by Frommann in a series of papers dating from 1864. He was led, by a study of jierves, to distinguish a reticulum, which partly corresponds with Leydig's spongioplasm. This reticulum was imbedded in a more homogeneous ground sub- stance, which, however, includes much more than spongioplasm. He extended this conception of protoplasmic structure to plant cells, and it was utilised, and in some respects modified, by Heitzmann in 1873. The views of the latter author were not so convincing as those of Frommann, for it is quite possible to identify the structures described by the latter writer in living cells, although the appear- ances are susceptible of a different interpretation from that given by him. Heitzmann's descriptions, on the other hand, are very schematic, and it is difficult to avoid the conviction that they are highly coloured by theoretical preconceptions. The phenomena of contraction and extension were brought by him in relation to the structures as described, but his views have never met with very general acceptance. A closely related hypothesis was that suggested by Flemming, who, while denying the existence of a reticulum, insisted on the presence of a fibrillar structure, the fibrils being represented as threads of irregular length (the filar elements) which were imbedded in a more fluid interfilar mass. Gradually another view of the structure of protoplasm was evolved, and which, in a measure, took account of reticular structure, but explained it differently. Strasburger in 1876 first seems to have spoken of closed protoplasmic chambers, which were filled with more fluid albumen, but he soon abandoned the idea in favour of the reticular hypothesis. But the alveolar theory thus indicated was developed and extended by Biitschli, who had, as long ago as 1873, figured in Pilidium a structure susceptible of such an explanation, though this was not given at that time. The alveolar THE STRUCTURE OF CELLS theory has exerted considerable influence upon contemporary thought, and may be given in brief outline. The whole protoplasm, including the nucleus, is conceived of as in a physical condition resembling an emulsion, the more fluid mass filling the cavities (which are very small, 1-2 p in diameter), whilst the walls are composed of a more viscous substance. Such an emulsion obeys certain well-known physical laws, and the relation of the alveolar walls both to each other and to the free outer surface can be theoretically defined. Solid heterogeneous particles enclosed in the emulsion, if too large to lie in the substance of the walls, are surrounded by a surface film in which the alveoli are arranged, as they also are on the free surface, somewhat like the cells of columnar epithelium. Move- ments may occur in the whole mass, as the result of disturbances in the surface tension of the superficial alveoli, and movements pro- duced in emulsions in this way closely resemble the streaming and other movements of protoplasm. The reticular and filar appearances are also to be attributed to disturbances, due to causes, in the interrelation of the alveoli, or they may represent the optical section of the alveolar walls themselves. Now it is quite possible to convince oneself that such an alveolar structure does actually exist in many cases, although, as Biitschli himself admit?, it is not always to be so identified. It seems probable, however, that on the whole it does represent truly the appearance of protoplasm under certain (and commonly occurring) condi- tions, but it also seems equally clear that these conditions are not necessarily always fulfilled. For it is essential for the production of such an appearance that there shall be at least two non- miscible fluids of different refractive index, and if either of these conditions is not realised, or is temporarily in abeyance, it will follow that the FI<>. 3. alveolar appearance must also be absent or dis- The foam structure appear. And we are acquainted with so many of the protoplasm of a- • ,, • ii i ..• r gre«arine. important series of changes in the relations of the various protoplasmic constituents to each other that it is hardly necessary to postulate the permanence of those conditions of which an alveolar structure is the consequence. Thus it would seem that an easy modus vivendi might be reached which would render it possible, whilst recognising the heterogeneity of the substances included under the generic term Protoplasm, to admit that at one time an alveolar, at another a filar, or a reticular appearance might occur. A fibrillar structure is certainly present during nuclear division, and although the extreme adherents of the THE STRUCTURE OF CELLS alveolar theory see in the fibrils a honeycomb structure, the cavities are generally admitted to be reduced to the vanishing point. Strasburger has attempted to utilise both the filar and the alveolar hypothesis, considering that in every cell the protoplasm outside the nucleus consists of two distinct parts. The one, which is specially nutritive and alveolar, he terms Trophoplasm ; the other, which is more closely concerned with the dynamical changes in the cell, and possesses a filar structure, he designates as Kinoplasm. The relations of the kinoplasm will be more specially considered in connection with nuclear division. Besides the protoplasm and nucleus, there are present other organised structures in the cell. The vacuoles, which have long been recognised, are cavities in the protoplasm, and lined apparently with a specialised layer of this substance. In some cases they are rhythmically contractile as in many of the Protozoa. But it is around that enigmatical body, the centrosome, that especial interest has persistently attached ever since its first definite discovery by Van Beneden in 1885. The centrosomes are minute granules, most often situated either singly or in pairs in each cell, and in close proximity to the nucleus. They are frequently con- tained in a specialised mass of protoplasm, termed by Boveri the Archoplasm. Centrosomes and their attendant structures have been differ- ently described by various observers. Van Beneden, to whom we owe the first recognition of these bodies, distinguished, in the case of Ascaris, a central granule, surrounded by two concentric areas, to which he gave the names of medullary and cortical zones respectively. Boveri described in the cells of the same animal a centrosome surrounded by a lighter zone, from which it was definitely cut off by a kind of limiting membrane. Within the centrosome he further dis- tinguished a central granule, the cen- triole. The latter body divides before the centrosome increases by fission. In still other cases (e.g. in cells of the testis of Salamander) various observers (Meves, Driiner, etc.) have distinguished FIG. 4. a whole Series Of Concentric zones Ascaris megalocepkctla. Schematic •i ,v T xv figure of diaster stage of the first around tne CentrOSOme. In the giant cleavage mitosis, c, centrosome ; m.z, cells of the spinal cord and in leuco- ggj^g, fffiei"' c°rtical Z°ne' cytes, Heidenhain has distinguished a group of granules, which replace the single or paired one more commonly met with, and in other cases again there is a reticulate sphere (echinoderms) containing a varying number of io THE STRUCTURE OF CELLS granular inclusions. In plant cells centrosomes have been far less often identified than in animals. They are more frequent, or at least more easily demonstrable, in the lower members of the vegetable kingdom than in the higher plants, in which they are probably restricted to the motile sperms. The evidence for their occurrence in angiosperms is not convincing. When present in a cell, they usually occur in the form of a small granule enclosed in a sphere, and are comparable with the centrosome and centriole of Boveri. Strasburger has proposed the convenient term of centro- sphere to designate the sphere together with its included granule, reserving the term centrosome for the latter body only. It is quite certain that the centrosphere apparatus presents itself in varied degrees of complexity, not only in different organisms, but even in different cells of the same tissues, and Strasburger's term has much to recommend it, since, in spite of the large litera- ture which has grown up around the subject, we are still mainly in the dark as to the true meaning and relations of the different parts. It seems clear, for example, that the centriole of Boveri corresponds to that which by most writers has been called the centrosome, and Boveri himself states that the division of his " centrosome " is preceded by that of the centriole. The centrosome or centrosphere is itself not unfrequently enclosed in a denser mass of protoplasm, called by Boveri the Archoplasm, and by Strasburger the Kinoplasm. Probably, how- ever, appearances denoted by these terms are not the expressions of permanent structures, but represent transient phases of cellular activity. The structures thus called into existence may, however, be, at least temporarily, very pronounced, since at least a part of the achromatic figure, which is formed during nuclear division, owes its origin to the archoplasmic mass. Nevertheless, the archo- plasm (or kinoplasm) may become absolutely indistinguishable at other periods in the life of the cell. A far more difficult question to answer than that concerned with the permanence of the archoplasmic or kinoplasmic structures refers to the centrosome itself in a similar connection. Whilst many authors have strenuously maintained its permanence from one cell-generation to another, comparing it in this respect with the nucleus itself, a considerable weight of negative evidence has never- theless accumulated in the opposite scale. The striking relations which obtain between the centrosome and the nuclear figures at phases of division naturally produce a profound subjective impres- sion upon the observer, and it has even been assumed that the centrosome still persists, even when its actual existence cannot be successfully demonstrated. There is no doubt that other granules have frequently been mistaken for centrosomes, selected because they happened to lie sufficiently near the spot where the structures THE STRUCTURE OF CELLS n in question might have been looked for, and thus no little confusion has been introduced into a subject already sufficiently bristling with difficulty. But the cases of Acanthocystis (Schaudinn) and of Actinosphaerium (R. Hertwig) show quite clearly that centro- somes may, at least in the lower animals, be certainly differentiated afresh in the cells from which they had previously been absent. /fe^C 6f -° f ^w^v&M c*< *' FlG. 5. Acanthocystis aculeata. A and B, for- mation of the centrosome from nuclear constituents in swarm-spores. C, resting cell. D, nuclear division preceding fission. (After Schaudinn.) ^tiM^jjilife •^f'.*^*-*. *';-?' ^IP^ D The so-called Blepharoplast, which is associated with the male reproductive cells of certain cycads and ferns, appears to present very strong analogies with animal centrosomes, and yet the blepharoplasts have not been seen in the antecedent cell -genera- tions of the plants in which they occur, and hence they have almost certainly been formed de now. On the whole, then, the question as to the relative permanence of the centrosomes through the series of ontogenetic cell-generations must be left an open one. Certain facts are, however, known which conclusively prove that centrosome-like structures can be formed in cells from which, under 12 THE STRUCTURE OF CELLS normal circumstances, they are absent. Morgan showed that concentrated solutions of salts could induce the appearance of centrospheres with radiations in the eggs of certain echinoderms, and Loeb further proved that by adding magnesium chloride in appropriate quantity to sea-water, the eggs of sea-urchins could be brought into such a state that when replaced in normal salt water FIG. 0. Actiiiospluurium eichornii. A and B, nuclear origin of the centrosome, which arises at one end of the nucleus. C, D, further stages in the mitosis (A-D refer to the first polar mitosis). E, diaster of a somatic mitosis, in which no centrosome is present. (After R. Hertwig.) they underwent the normal embryonic segmentation. Wilson, investigating the cytology of the process, confirmed the results, and ascertained that the treatment caused the formation of centro- spheres which seemed to direct th'e cell-divisions. And R. Hertwig long ago showed that at least the early stages of a parthenogenetic segmentation could be similarly induced by the action of strychnine. Hence there is a considerable body of evidence to show that the centrosomes are structures which, though physiologically the signs THE STRUCTURE OF CELLS 13 of important changes in the protoplasm, are not necessarily permanent organs of the cell. And a wide survey of the processes of mitosis in the lower animals and plants serves fully to confirm thiy conclusion. The Structure of the Resting Nucleus. It has already been said that it was not until the year 1875 that the nucleus was fully and universally recognised as an all- important cell organ. Even as late as the previous year, Auerbach published a treatise on its behaviour during cell-division maintaining that it completely disappeared during the process, and he gave the name of Karyolysis to the phenomenon in question. With the recognition of the complex series of changes undergone by the nucleus during division, and its obvious importance, in connection with fertilisation, also discovered in 1875, it speedily formed an object of serious study. And the investigations were not only carried on in killed cells, but its behaviour during life, as well as its chemical structure, presented attractive problems for solution. The general outcome of these investigations is as follows : — The nucleus is delimited from the cell protoplasm (the cytoplasm of Van Bambeke) by a membrane which was regarded by Schwartz as consisting of a substance called by him Amphipyrenin. In some cases, however, it appears not improbable that the membrane is at least partly produced from the cytoplasm, as a kind of precipitation membrane, whilst in other cases, as for example in some of the coccidia, Schaudinn has shown grounds for thinking that the so-called chromatin of the nucleus itself may contribute to its formation. Within the membrane the nuclear contents may be distinguished as a matrix of a substance which stains with some difficulty, and which forms a sort of meshwork within it. This is. the Linin of Schwartz, and seems to closely correspond with the plastin, distinguished chemically by Zacharias. In addition to the linin there exists a more fluid gelatinous substance, the Paralinin of Schwartz. Imbedded in the linin are a large number of granules which, by reason of their exhibiting a strong affinity for certain dyes, were termed Chromatin by Flemming. The chromatin consists of a highly complex nitrogenous substance, and always contains phosphorus. Chemically it belongs to the class of proteid compounds classed as nucleins, and by analysis can be made to yield proteids and nucleic acid. In addition to the true nuclein chromatin, there have been described other inclusions within the linin known as Lanthanin (Heidenhain) or oxy- or basi-chromatin bodies, which appear to be related to the nuclein series, and which perhaps are complex, high-graded substances which can be built still further up to true nucleins. Most nuclei contain, besides these constituents, one or more I4 THE STRUCTURE OF CELLS masses, usually of a spherical or -oval shape, known as Nucleoli. These bodies long ago attracted the attention of investigators, and it will be remembered that they were raised to a rank of con- siderable importance by both Schwann and Remak. They usually are easily stained, and thus were included amongst the chromatin bodies of the nucleus, but subsequent investigation has shown that they are, in many cases, widely different from nuclein. Two kinds of nucleoli were distinguished by Flemming under the names of eu- and pseudo-nucleoli respectively, the latter representing, at least chiefly, aggregations of a substance which closely approximates to, if it is not actually identical with, true nuclein. And more recent investigations have tended to confirm the supposition advocated by Zacharias, that the ordinary eu-nucleoli, so far from consisting of a single substance such as pyrcnin (Schwartz), are complex mixtures, or else, at any rate, bodies which readily yield, by suitable treat- ment, different substances of complex molecular composition. It is true that the author just referred to arrived at the conclusion that the nucleoli were destitute of phosphorus, but this view can hardly be maintained, at least in all cases. Investigations on the nuclei of Protozoa and of some of the lower plants seem to have shown that these nucleoli consist of at least two groups of substances, the one consisting of, or approximating to, nuclein, the other more nearly resembling the linin, or even the cytoplasm, in its staining and other reactions. At any rate, the chromatin, which forms so obvious a character in dividing nuclei, appears in some cases, e.g. Actinosphaerium, to be mainly derived from a nucleolar source. It is highly probable that these bodies are really heterogeneous, and represent reserves of complex materials which can be drawn on for various purposes during periods of nuclear activity. For at such times the nucleolus always undergoes considerable change, and is either completely used up, or its remains fragment and pass out into the cytoplasm, where their further fate is still obscure. Whilst the nucleoli .are thus losing substance they often exhibit vacuolation, and even in resting nuclei vacuoles may sometimes be detected within the nucleoli, pointing strongly to the correctness of the hypothesis as to their heterogeneous nature. A further point which deserves mention in connection with the nucleoli is the view held by some writers (e.g. R. Hertwig) that they stand in some close relation to the centrosomes, and that the existence of the latter structures may be traced back to a nucleolar origin. Further, Strasburger, in his account of kinoplasm, has suggested that the nucleolar substance may serve as the material which stirs up the dynamical and metabolic activities latent in the cell. On the whole, it is impossible, in the present state of our knowledge, to ascribe any single function to these bodies, and the THE STRUCTURE OF CELLS evidence before us seems to indicate that just as they are very diverse in structure and composition, so also they may, and almost certainly do, play very different parts in the general economy of the cytoplasm and nucleus. The resting nucleus may, then, be regarded as an organised structure containing a considerable assortment of highly complex and labile substances. But this very lability, itself a condition of the profound and important changes which succeed each other with extraordinary rapidity during the division of a nucleus, is bound up in, or at least is related to, an organisation which directs and Attraction-sphere enclosing two centrosomes. Nucleus Plasmosome or true nucleolus. Chromatin- nttwork. Linin-nctwork. • Karyosome or net-knot. Plastids lying in the S cytoplasm. Vacuole. Lifeless bodies ( meta- plasm) suspended in the cytoplasmic reticu- luuu FIG. 7. Diagrammatic representation of the structures present in a typical cell. (After Wilson.) determines that sequence of chemical and physical transformations which so strikingly accompany the whole process. Moreover, there is abundant evidence of the existence of a material exchange passing between the nucleus and the cytoplasm which becomes strongly marked at all periods of special cellular activity — such, for example, as secretion, regeneration, and the like. Nuclear and Cell Division. The multiplication of the uninucleate cell is always preceded, save in the lowest protozoa and protophyta, in which the details of the processes are still obscure owing to the absence from them of 16 THE STRUCTURE OF CELLS a well-defined nuclear body, by a division of the nucleus. This may either take place in a simple manner, as was determined by Remak, or it may only be secured as the result of a complicated rearrangement and fission of certain nuclear constituents. To the former, or direct (Flemming), method of division, the term Amitosis has been applied by Flemming, whilst the latter, or indirect (Flemming), method was also termed Mitosis by the same author. The word Karyokinesis (Schleicher) has often been substituted for mitosis, but both terms are expressive of the same phenomena. Amitosis, in the higher animals, is not of such generally widespread occurrence, but in the lower forms ^it frequently appears as an intercalated method along with a more or less complex form of ordinary karyokinesis. It is also generally met with in nutritive gland cells; thus in the follicular epithelium of the ovary, in the "foot " cells of the testis, and in the tapetal cells of the higher plants, it is not uncom- mon. In these cases it appears to be characteristic of degenerating tissues, and this explanation has been extended to amitosis gener- ally by Ziegler and vom Rath, but many instances are known in which such a view is quite untenable. Thus, there is good reason to believe, as Meves or others have shown, that in the ovary, cells which ultimately are destined to give rise to ova may multiply in this way, and Schaudinn, Siedlecki, and others have shown that in the Sporozoa such amitotic divisions often follow shortly on the act of fertilisation, and give rise directly to the new generation of parasites, and again amongst Infusoria the macronucleus seems always to in- crease in this way. Furthermore, Nathansohn proved, in the case of Spirogyra, that by appropriate treatment with anaesthetics the nuclei could be induced to divide amitotically, and that this amitotic origin in no way influenced the conditions of the subsequent develop- ment of the cells concerned, for these were capable of even pro- ceeding as far as to form sexual cells, on the restoration of a normal environment. But in comparing amitosis and mitosis together, the advantage which the latter possesses, so far as can at present be stated, seems to lie in the accurate quantivalent distribution of all the structural elements concerned in the process between the two daughter nuclei. Whether this is the only advantage, or whether perhaps some mechanical or other factors are also involved, must be left to the future to decide. In considering the general phenomena presented by karyokinesis, there are two sets of factors which, though closely interwoven in the process, may with advantage be kept as distinct as possible. For these changes involve the nucleus on the one hand and the cytoplasm on the other, and the degree of complexity which each of them may respectively assume is not necessarily invariable or correlative, either in different organisms or in different tissues of the same individual. A second, and not less important, considera- THE STRUCTURE OF CELLS tion arises in connection with the fact that nuclear- and cell- (or cytoplasmie-) division are by no means invariably associated, and that although the cytoplasm never gives rise to a number of cells in excess of the number of nuclei present, its divisions in other- respects may occur independently of that of the nuclei. This is seen in the cleavage of some animal eggs (e.g. mollusca), in the formation of endosperm in the seeds of angiosperms, in the develop- • "--V •— »sv / *& Fio. 8. Lilium martagnn, prophase of the first mitosis in the pollen-mother-cell, showing the longi- tudinal fission of the cromatin and linin. ment of the eggs of Fucus and of the spores of Mucor. In all these instances, the division of the nucleus precedes that of the cyto- plasm, which is only subsequently partitioned. The first indication of approaching karyokinesis in an ordinary somatic cell of the body of one of the higher plants or animals is usually visible in the nucleus. The chromatin granules become aggregated in lines, corresponding to a growing definiteness in the delimitation of the linin. Thus from the generally granular appear- ance, the character of a much convoluted and tangled chromatic 18 THE STRUCTURE OF CELLS skein is produced. The linin framework does not necessarily form a continuous thread. Often it is more or less broken, and it almost always shows cross-anastomoses (from which, however, in the later phases, the chromatin is commonly absent) with the neighbouring threads. This anastomosis is doubtless the expression of its segrega- tion, due to contraction ; the anastomoses themselves representing the original meshes by which the substance was formerly bound together into a coherent whole. Simultaneously the chromatin increases largely both in amount and in the intensity of its stain- ing power — a fact which may be taken to indicate a chemical or physical change in its state. The linin thread- work that contains the chromatin is often not scattered irregularly through the nucleus, but is more or less polarised, as was clearly observed by Rabl, in such a way as to converge, often with considerable distinctness, towards one point on the nucleus. This point is occupied by the centrosomes when they are present. At first usually lying in pairs, and often in a mass of archoplasm, these bodies in the simpler cases now commence to diverge, and each is either accompanied by a portion of the original archoplasm, or else the latter is differenti- ated progressively and afresh as they move apart to take up diametrically opposite positions on the periphery of the nucleus. From them radiate outwards into the protoplasm the well-known astral figures which are characteristic structures in the cell at this period, and are commonly regarded as of archoplasmic origin. Meantime within the nucleus the chromatic thread thickens and shortens. Some of its substance is probably derived, at least in many cases, from the nucleolus, which becomes vacuolated and often fragments about this stage. Finally, the thread breaks up into a number of segments which is constant for the somatic cells of the species. These segments are the Chromosomes (Waldeyer). At or immediately following this stage a fibrillar structure begins to appear within the nucleus, and as it increases the chromosomes are gradually driven to occupy an equatorial position (the equatorial plate stage) in the nucleus. What is precisely to be looked on as the origin of these fibrils (the so-called achromatic fibres which together form the achromatic spindle) is not certainly known. Some, with Strasburger, hold that they are exclusively of cytoplasmic (kinoplasmic) origin, growing inwards, as it were, from the polar centrospheres. Others again look on them as derived from nuclear substance, whilst a third view regards them as of mixed origin. Probably the last view is less open to objection than the other two. The fibrillar structures themselves are almost certainly the result of conditions of stress and strain in the viscous substances of which the cell is composed, and it would appear probable that any sub- stance capable of assuming the fibrous character might be compelled to do so. And there is abundant evidence to show that such sub- THE STRUCTURE GF CELLS stances do exist in the nucleus, as they certainly do in the cyto- plasm. For in the case of, e.g., the micronuclei of infusoria, the whole spindle is entirely intra- nuclear, the cytoplasm apparently not furnishing, at least directly, any part of it. With the congregation of the chromosomes to form the equatorial plate, the first stage or Prophase of division terminates. The equatorial plate, or aster, stage is often one of relatively long duration ; so much so that it may even happen that some of the signs of cytoplasmic activity may fall into temporary abeyance. For example, the astral radiations outwards from the centrosomes may cease to be visible at this stage (e.g. in Pellia), though they FIG. (>. Stages of the mitosis in the micronucleus of raramofcium, showing the "pole plates"; true centrosomes are not present. (After R. Hertwig.) reappear later on. The nuclear wall commonly, though not always, disappears whilst the chromosomes are collecting at the equator, and the nucleolus or its fragments, if they have not previously disintegrated, now are no longer recognisable. The individual chromosomes often, but by no means always, assume the shape of a V, with the apex turned centrally in the equatorial plane. Each one is supported by fibres of the achromatic spindle which run from the poles, and terminate on the chromosomes at the equator. The chromosomes next split longitudinally, and this partition forms the commencement of the stage known as the Metaphase. The two daughter halves rapidly diverge, being guided by the spindle fibres towards the poles. During their divergence 20 THE STRUCTURE OF CELLS Fio. 10. Fucus vesicitlosHs, stages in the first mitosis in the fertilised e;:g (oospore). A-D, Prophase ; E, coinmeucement of the Metaphaso. THE STRUCTURE OF CELLS 21 FIG. 10 (continwrf). Fuciw vmcuZosus, stages in the first mitosis in the fertilised egg (oospore). F, Metephase ; 0, H, Anaphase. 22 THE STRUCTURE OF CELLS fresh fibres are differentiated 'between the retreating groups of halved chromosomes, and form the interzonal fibres (Verbindungs- faden of the German writers). Whether these play any mechanical part in forcing the daughter chromosomes apart is uncertain, as is also the r61e assigned to the above-mentioned fibres that appear to direct the chromosomes towards the poles. Probably the latter Fio. 10 (continued). Fucus vesiculosus, stages in the first mitosis in the fertilised egg (oospore). 7, Telophase. (Phil. Trans, of the Royal Society.) are actively contractile, and there is some evidence to show that the interzonal fibres are in a state of stress. In some instances, e.g. in fertilised and segmenting eggs of Fucus, the arrangement of the elongated plastids in this region plainly indicate such a condition of stress or strain. But the achromatic apparatus varies considerably in the degree of its complexity, and it probably would be unsafe to attempt to assign constant functions to its constituent parts. So much, however, may be said, that the chromosomes appear to be THE STRUCTURE OF CELLS passively moved to their respective poles, and to possess no power of automatic movement of their own. With the arrival of the chromosomes at their respective poles the Anaphase stage supervenes. This consists practically in a series of regressive changes which leads to the formation of normal resting nuclei. The chromosomes lose their sharp outline and swell up; at the same time the nucleoli once more reappear. The chromatin, or as much of it as persists, is distributed through the swollen linin just as it originally existed in the parent nucleus, and finally a wall isolates the daughter nucleus from the surrounding cytoplasm. But the cytoplasm still bears traces of recent dis- Fio Pelvetia canaliculuta, telophase ot the second mitosis in the fertilised (oospore). (Phil. Trans, of the ' Society.) turbances, and the period of gradual restoration of quiescence in it forms what is sometimes known as the Telopliase. The centrosome (when present) is often already doubled during the meta- or ana-phase, but the astral radiations frequently do not die away till much later. It is in the region of the interzonal fibres that events of the greatest interest are now proceeding. In animal tissues it very often happens that the two cells are constricted equatorially, and they may ultimately become delimited from each other, the remains of the interzonal fibres then remaining at this spot, where they may be long recognised as the Intermediate Body. In plants, owing to the existence of a cellulose skeleton, and the close adherence of the cytoplasm to its internal surface, such a con- striction does not usually arise. Instead of this the fibres increase THE STRUCTURE OF CELLS greatly in number, especially in the equatorial zone. In these fibres the primordium of the new cell-wall is laid down, in the first instance as a viscous film, but which later, by the deposition of fresh substances, becomes converted into the cellulose partition. Its mode of formation is interesting because reasons have been shown for supposing that some at least of the protoplasmic connections between adjacent cells are primarily effected by the permanence of such continuity through the membrane during its formation. Fio. 12. Diagram of the successive stages of a nuclear division. A, Hpirem, with the fission of the chromatic liiiin. II, aster. f, D, K, separation of daughter segments. F, reconstitution of daughter nuclei. (After Flemming.) It may, however, happen that no membranes are formed in the interzonal fibres, such as will serve to delimit the daughter cells from each other. A complete series can be traced between the two extremes. Thus in the first division of the spore-mother-cell of Fegatella (Fig. 13), a cell, plate is laid down, but not completed, and it is not until after the next nuclear division that this wall (which has shifted its position in the interval) becomes part of the final partitioning membranes. Again, in the endosperm of seeds, some- times the embryosac is transversely divided after the first karyo- kinesis, but far more commonly a large number of nuclei are first THE STRUCTURE OF CELLS formed. These then take up their final positions, and a new set of interzonal fibres are differentiated between them, and in the equatorial planes of these groups of fibres the cell-walls are laid down. And finally, in other cases the cytoplasm may divide into masses containing either single or several nuclei, and secrete membranes without the intervention of interzonal fibres at all. It will have become evident from the foregoing account of the relation between the nucleus and the cytoplasm, that these two principal constituents of a cell retain to a consider- able extent a separate individuality, at any rate in the higher forms. This separate nature only be- comes obscured at periods of division, but even here, as has been seen, the essential boundaries are retained through all the changes connected with fission and redistribution. Thus it is legitimate to regard the cell nucleus as an entity which does not arise de novo. The nuclei of successive cell- generations are lineal descendants of an ances- tral nucleus, just as the cells of the present day owe their being to the multiplication of ante- cedent parent cells. The nucleus, how- ever, does not stand alone amongst the cell constituents as only arising by multiplication by fission of pre- existing structures of a similar character. In the plant-cell the various plastids originate in a similar manner, arid there is no more evidence to show that they can be differentiated afresh from the general cytoplasm, than that the latter, by spontaneous generation, can arise de novo from its elemental constituents. The same is true for the curious coloured plastids known as Zoochlorella in animals, which possibly represent species of algae imprisoned in the cells of their animal hosts, or perchance, though less probably perhaps, they may be regarded as more akin to the chlorophyll corpuscles of the FIG. 13. Fegatclla conica, the division of the spore-mother-cell into four cells, showing the change in position of the first formed wall. 26 THE STRUCTURE OF CELLS plant cell. The latter hypothesis would be difficult to sustain in the absence of a series of forms through which their evolution might be traced, whilst, on the other hand, the symbiotic relation- ship existing between fungi and algal cells in a lichen strongly supports the presumption that an analogous case is furnished by the Zoochlorella organism and its host. " Reduction Divisions" Few cytological discoveries have aroused more widespread interest than that of the periodical recurrence of the so-called " Reduction Divisions," which are intercalated at some point in the cell-generations intervening between two consecutive sexual unions. Each uniting gamete or sexual cell contains in its nucleus only half the number of the chromosomes that will be characteristic of the embryo resulting from their fusion, and will be retained throughout its cell -generations up to those which lead in their turn, more or less directly, to the production of spermatozoa and mature ova. This remarkable phenomenon has been observed in all the animals and plants which have been carefully studied, with the exception of the more lowly or primitive forms in which the nuclear history is but imperfectly understood. The phenomenon in question was first made known by the investigations of Van Beneden in 1883 and 1887 on Ascaris. The choice of this animal was in some respects perhaps not very fortunate, since it does not exhibit the process in a very typical, but rather in an extreme, form, and thus a certain amount of misapprehension prevailed at one time respecting it. Since that period, however, very numerous animals and plants have been studied, with the result that the phenomenon is proved to be of very general occurrence, though differing considerably in detail in the various organisms. At first, and perhaps naturally, the view was advanced that the reduced number was secured through the mere degeneration and consequent elimination of the superfluous chromosomes, but it gradu- ally became clear that the evidence was entirely opposed to such a simple explanation, and that, on the contrary, the reduction was only arrived at after an exceedingly complex rearrangement of the nuclear constituents. It would, however, be going too far, as will subsequently appear, to deny that any nuclear substance is lost : all that can be said is that it is certain that no chromosomes, as such, are normally eliminated. In attempting to trace the sequence of events, it must be borne in mind that the process is evidently one of the highest importance, seeing that it occurs alike in animals and in plants, and this importance is increased rather than lessened by the further THE STRUCTURE OP CELLS 27 recognition of the fact that the reduction may occur at morpho- logically diverse stages in the life-history of the various organisms — a fact which clearly emphasises its profound physiological sig- nificance. But although there is no lack of hypotheses to explain it, no one as yet has given a satisfactory theory which will embrace the whole range of the phenomena concerned. In the higher organisms the process of reduction appears invariably to be closely bound up with two nuclear divisions which rapidly succeed each other, and are hence often spoken of as the Reduction Divisions. These differ in some important respects from those characteristic of other mitoses. They can only be con- sidered in outline here, and after premising the existence of a not inconsiderable diversity as to the details of tLe process in different organisms. In animals the mitoses in question only occur in direct relation to the formation of the sexual cells or gametes, but in plants it is more usual to find a greater or less number of cell- generations follow on the Reduction Divisions before the actual gametes are formed. Thus it becomes obvious that the formation of sexual cells is not a necessarily immediate consequence of the change in the nucleus. If the course of events be studied in an animal, it is seen that in the development of the spermatozoon and of the mature egg, a strictly comparable series of changes is passed through. Just as the spermatocyte gives rise, by two successive bipartitions, to four sperms, so the immature ovum, by means of two successive nuclear divisions, gives rise to four potentially fertilisable eggs, of which, however, three commonly degenerate and are known as the polar bodies. The nucleus of the spermatocyte, just as does that of the im- mature egg (which may be distinguished from the ripe egg by the name of oocyte), goes through a somewhat prolonged period of growth before entering upon the critical mitoses. As these two divisions are marked by certain peculiarities from those of the other cell-generations, it is convenient to designate them by special names. The first may be termed the Heterotype, the second the Homotype, mitosis, following the terminology introduced by Flemming. During a large part of the growth -period, leading directly to the heterotype division, the nucleus cannot be correctly described as resting, for the linin reticulum is plainly discernible, as also are the regularly arranged chromatin granules, which serve to render it distinct. In fact, this prolonged spireme is highly characteristic of the heterotype mitosis, as contrasted with those which have been previously gone through. It is during this stage that the fission of the chromatin granules occurs, as was first seen by Pfitzner in 1881. Each granule becomes drawn out into a 28 THE STRUCTURE OF CELLS dumb -bell -shaped body, and finally two rows of granules are seen to occupy the margins of the flattened linin riband. This stage is often (Salamander, Helix, Lilium, etc.) followed by a more or less complete longitudinal fission and separation of the linin riband, each half now containing, at least at first, a single row of chromatin granules. It is perhaps not improbable that a similar process also occurs during the corresponding stage of somatic 1 Fio. 14. Lilium martagon, prophase of the first mitosis in the pollen-mother-cell, showing the longi- tudinal fission of the chromatin and linin. mitoses, but the shortness of its duration in the latter renders the process difficult to observe. At or about this period a remarkable contraction of the chromatic linin filament occurs, and commonly the nucleolus is included in the tangle. To this stage the name of Synapsis (Moore) has been given, and it seems to represent an important step in the process, and one which is confined to this first (heterotype) reduction division. After the synaptic condition is over, the linin, which has been getting richer in chromatin, is usually seen to be shortening, and at the same time thickening, but THE STRUCTURE OF CELLS 29 it may happen that the subsequent events become much obscured, the filamentous arrangement of the linin and chromatin ceasing to be distinctly recognisable. Up to this period there is, on the whole, a general agreement as to the nature and sequence of events, but the subsequent changes have been very diversely described and interpreted in the case of different organisms. In the most favourable cases the parallel arrangements of the chromatin granules and of the split linin thread can be followed for some time during the shortening arid thickening of the filaments. Brauer has described, for the spermatocytes of Ascaris, a second longitudinal fission in each chromatic filament, resulting in the production of four rows of chromatin granules from the single row originally present in the primitive thread. A similar event has been stated by some to occur in the corresponding division in the pollen-mother-cell of a lily. In the majority of instances, however, the chromatic linin is seen to contract and thicken, and all traces of the fission may become unrecognisable. Finally, the chromatin comes to be aggregated in definite parts of the band, the intervening portions being occupied by colourless linin. There are often, also, cross anastomoses of the same substance between neighbouring strands. The chromatic areas in question mark the position of the developing chromosomes (Fig. 15, A\ which gradually become more definitely isolated from each other. And they are seen to be present in half the number characteristic of all the preceding nuclear divisions in the organism. Once more each chromatic band exhibits a split l along the whole or greater part of its length, and this marks the line along which, later on, the cleavage of the chromosomes in this (heterotype) mitosis will be effected. In many cases, as is especially well seen in amphibians (Fig. 15, B, (7), the fission of the young chromosome is incomplete and the sides diverge, thus causing the whole to assume the form of a closed ring. In other instances, however, the fission is completed, and the two halves, lying in close juxtaposition, may exhibit a complex series of figures which demand much care in their elucidation.2 1 It is commonly assumed that this split represents the original longitudinal fission of the linin filament. It is not, however, proved beyond doubt that this is invariably the case. 2 These appearances have, however, been differently explained by some investi- gators ; thus some have seen in them evidence of an approximation of two entire somatic chromosomes, hence when the apparent halves separate to give rise to the chromatic part of the daughter uuclt-i, it would follow that what has really occurred is the distribution of half the original entire somatic chromosomes to the daughter nuclei. That is to say, the division might be regarded as qualitative as well as merely as quantitative. And it will be evident on reflection that the same result might be reached as the result of various analogous interpretations of the fore- going processes, especially when the difficulties of investigation that occur during the synaptic tangle are borne in mind THE STRUCTURE OF CELLS But the evolution of the chromosome does not always follow along these lines. In a number of instances, exemplified by many arthropods (e.g. Cyclops), after the early chromatic fission has been B PlO. 16. Salamander, Heterotype mitosis in the spennatocytes. (After Moves.) For explanation of the figures see the text. passed through, and the number of the future chromosomes has been marked out, these bodies, it is true, may form rings (Gryllotalpa) or parallel rods (Cyclops), but the chromatin, instead of being tolerably equally distributed throughout the length of the THE STRUCTURE OF CELLS 31 longitudinal halves, becomes specially aggregated at two spots in each. Thus are formed the so-called Tetrads, to which much im- portance has been attached in the theoretical interpretation of the whole process of reduction. For it is thus seen that in the above cases the tetrad may be regarded as having originated first by a longitudinal fission of the chromosome rudiment, and then by a transversely isolated aggregation of chromatin in each half. It may also happen that tetrads are formed in a manner less easy to follow out, as in Helix and in Arion, in which the separate filaments are difficult, if not impossible, to trace in the stages immediately preceding their formation ; the chromatin thus appear- ing, so to speak, to travel to and become aggregated at definite areas, and to assume in a somewhat irregular manner the ring form of tetrad, similar to that occurring in Gryllotalpa, as described and figured by vom Rath. Still another type of tetrad formation has been described by FIG. Ui. Diagram illustrating tetrad formation. A, tlie split thread (spirem) stage. B, later stage, showing aggregation of chromatin at each end of the split bivalent chromosomes. C, fully formed tetrads, of which the one to the right represents the most typical form. Brauer as occurring in the spermatogenesis of Ascaris megalocepJiala already alluded to above. The lining filament first contains a single row of chromatin granules, each of the latter divides crosswise into four, which lie in the same transverse plane, and hence the original filament now contains four rows of chromatin granules. As the process of shortening and thickening progresses, these become, so to speak, telescoped together, and the end view of each filament exhibits four chromatin masses corresponding to the four rows just described, and which thus appear as tetrads similar to those of Cyclops, although they would appear to have a very different origin. For whereas in the latter case the single units of the tetrad have arisen as the consequence of two longitudinal fissions of the original chromatin granules, in the case of Cyclops the same appearance is apparently produced partly by a longitudinal fission, and partly by a transverse delimiting of the original granules. In the first maturation division of the egg of the same animal, each chromosome is seen to be divided completely into four segments, 32 THE STRUCTURE OF CELLS though it is not certain as to whether the details of their develop- ment are similar to that of the spermatogenetic tetrads as described by Brauer. Meanwhile, other changes have been proceeding, both within and without the nucleus. The nucleolus commonly can be seen to lose substance during the growth of the chromosomes, as is testified by its vacuolated appearance. Often it fragments into smaller particles during the prophases. But it has been definitely ascer- tained, in many cases, that some part of this body is cast out into the cytoplasm where it degenerates, and it is not improbable that this will prove to be of very general occurrence. The possible significance of this event ought not to be overlooked in view of its striking occurrence in the lower forms of life, for it is in them that the clue to the meaning of the complex changes observed in the higher animals and plants must probably be sought. In the cytoplasm, also, remarkable changes connected with the centrosome and spindle mechanism have been proceeding. The latter reaches very different degrees of completeness in different organisms, and, as has been said on a previous page, even in the different cells of the same organism. In the simplest case the two centrosomes, when present, diverge and form a spindle not dis- similar from that already described for somatic nuclear division. In other instances (e.g. in Salamander) the two centrosomes diverge tangentially to the nucleus, and the spindle is formed between them, and, in the first place, without immediate reference to the nucleus. Later on, however, from the poles of the central spindle thus differentiated not only do the radiating fibrils reach into the protoplasm, and even to the periphery of the cell, but there are also others that extend to the nucleus and become attached to the chromosomes. The latter are thus, as it were, roped up and pulled on to the periphery of the first formed spindle (Fig. 15, D, E). Almost every gradation occurs between the extreme forms here sketched, and the matter seems to be essentially one of more or less complete division of labour between the constituent parts of the achromatic spindle regarded as a whole. In the Salamander, and those other cases in which it appears in a more or less complete form, the central spindle seems to function as a sort of support to keep the two poles apart, and to serve as a sort of railroad along which the daughter chromosomes can be pulled along by the con- tracting peripheral or mantle fibrils that attach them directly to the poles. In the divisions of the higher plants, as has already been ob- served, no centrosomes have been identified with certainty, and the spindle at first starts into existence quite irregularly. It speedily, however, becomes for the most part bipolar, although not unfre- quently isolated fragments of extruded nucleolar substance exert a THE STRUCTURE OF CELLS 33 very obvious influence on the direction of individual fibrils which may deviate towards and even terminate upon them. When the chromosomes have reached the equator of the spindle they may still preserve the form of rings, tetrads, or more complex shapes, and the method of their fission which leads to the severance of the daughter chromosomes is seldom so clearly longitudinal as is the case in a somatic mitosis. The rings present the greatest difficulty, and there exists a considerable divergence of opinion as to whether their division really corresponds to a transverse or to a longitudinal fission of the whole chromosome. The answer practically turns on the conclusions arrived at as to the path of development followed by the chromosome during the earlier phases ; i.e. whether the plane of separation really corresponds to that of the cleavage of the granules, or whether it may not be related to a totally different series of events, and marks the separation of originally bivalent chromosomes as is contended by some observers. In the latter case the complete identity of all the parent somatic chromosomes would be preserved in spite of its apparent loss through the fusion of them in pairs. The separation of the daughter chromosomes would be thus interpreted as not due to the fission of one, but as the segrega- tion of each individual of a pair which had previously become temporarily united. After the separation of the daughter chromosomes and the com- pletion of the anaphase (Fig. 15, G) and telophase, the two nuclei which are thus formed commonly commence immediately to divide once more. Again the reduced number of chromosomes reappears, but the character of the process superficially resembles a somatic mitosis much more closely than the preceding heterotype division, and for this reason the name of homotype was given it by Flemming. In reality, however, there are many other points of difference, besides that of the reduced number of chromosomes. The first, and perhaps the chief, peculiarity lies in the fact that there is good reason for believing that the line of fission of the chromosomes is always predetermined during the prophases of the preceding, the heterotype divisions. This is strikingly seen in the case of Ascaris eggs, in which, during the first maturation division, two of the four rods that together represent one chromosome are distributed to the daughter nucleus as the equivalent of a single daughter chromo- some, whilst at the second mitosis each chromosome emerges as a double (not quadruple) body, and at the metaphase the two con- stituent or collective parts separate from each other as the definitive chromosomes of the next (and final) nuclear generation. Essentially the same course of events obtains in cases of tetrads. When these divide at the equatorial plate the resulting dyads thus formed retreat as the daughter chromosomes, and on the rapidly following homptype division the dyads again reappear in the early 34 THE STRUCTURE OF CELLS stages and divide at the equator, each half (monad) forming a daughter chromosome. And an essentially similar condition obtains in at least many other more obscure instances. Thus in Salamander and in Trades- cantia, during the dyaster condition of the anaphase of the hetero- type, each daughter chromosome is seen to be longitudinally divided (Fig. 15, F). This almost certainly is the result of the reopening of a split formed during the prophase of the heterotype, but which has escaped recognition in many cases owing to the great difficulties- which these earlier stages present in their investigation. And the fission, thus obvious in the dyaster of the heterotype, provides the daughter chromosomes for the next (homotype) mitosis. It is thus seen that these two mitoses, the heterotype and the homotype, which in animals are the immediate forerunners of the differentiation of the sexual cells, are clearly distinguishable from the preceding ones in several important respects. 1. The appearance of the chromosomes in the reduced number, i.e. they are only half as numerous as in the rest of the nuclei. 2. The long duration of the prophase, and the complex changes and rearrangements, including, probably universally, a preparation for the distribution of daughter chromosomes not only for this but also for the succeeding division. 3. The remarkable forms assumed by the chromosomes upon the spindle. 4. The very general extrusion of nucleolar substance, in rela- tively large quantities, from the nucleus during the prophase. Although probably of importance, it would be as yet premature to speculate on the precise weight to be attached to this phenomenon, but it is suggestive when considered in relation to the course of events described for protozoa. The accompanying figure may serve to render clearer the exact nature of the different views which have been held as to the nature of the processes which are passed through in the reduction divisions. The somatic cell (I. in the figure) is supposed to include two chromosomes, and below are represented, diagrammatically, the various alternative phases gone through during reduction, the corresponding stages being shown in any four figures on the same horizontal line. The series II. and III. represent the events which may be passed through on the assumption of the permanence' of the chromosomes, whilst the series IV. and V. correspond to those in which such a permanence is denied. In the former case the two original chromosomes, A, B, remain temporarily united, and their two methods of possible . separation are respectively shown. In IV. and V. no continuous identity is claimed for the chromosomes, and the two original ones are replaced by a single new one (C). Considerable difference of opinion exists, then, as to the real THE STRUCTURE OF CELLS 35 nature, as well as the meaning of the events which are thus bound up with the two divisions under consideration. The indications FIG. 17. Diagrams of Heterotype and Homotype mitoses to illustrate the various possible ways of interpreting the chromosome distribution. afforded by the constancy in number of the chromosomes through the cell-generations of an organism point to a morphological per- manence, and it has been argued that the same chromosomes re- peatedly are re-formed at each mitosis. 36 THE STRUCTURE OF CELLS And since, during the reduction division, there is no evidence of the elimination or degeneration of any chromosomes, it is further urged that each of the apparent units appearing in the prophase of the heterotype division are really bivalent, and represent two chromosomes joined end to end, but otherwise behaving as one. Hacker, who has ably supported this view, believes, in common with many who share it with him, that the tetrad is to be thus explained. The longitudinal fission which divides the bivalent rudiment of the chromosome is succeeded by a more or less trans- verse separation or isolation of the chromatin, which marks the fourfold character of these chromosomes. Consequently each tetrad really represents not one but two chromosomes, and whilst the first (heterotype) division corresponds to the line of longitudinal cleavage, the second (homotype) separates and distributes actual entire chromosomes. Hence it is supposed that a real distribution of entire and diverse chromosomes occurs at the homotype mitosis, whilst the heterotype is essentially similar to a somatic division, and the reduction in number (due to the coherence in pairs) of the chromosomes is only an apparent one. If it could be universally proved to be true, such an explanation would account for many of the remarkable peculiarities which, as has been seen, characterise these divisions, besides affording a very strong support to the theoretical views as to the nature of the mechanism of inheritance advanced by Weismann. But the apparently well-established belief that in other cases the preparation for the two divisions is accom- plished by means of two longitudinal divisions of the chromatic liniri militates strongly against conceding the value of a general inter- pretation to the views just sketched in outline. And moreover the facts of amitosis as known to occur in some instances, also, though in a somewhat different way, tell against the permanence of the chromosomes, and consequently against the theories which have been founded on that hypothesis. On the whole, the facts at present before us rather tend to support the view of the brothers Hertwig ; according to them the real significance of the process lies in that sudden quantitative reduction of the chromatin which is a necessary consequence of the rapid succession of the two mitoses in question. It has already been pointed out that the reduction divisions are a common feature to both animals and plants. In the latter, how- ever, there appears to exist a much greater latitude as to the point in the life- history at which they occur. In all the archegoniate series of cryptogams, which includes the mosses, hepaticae, and vascular cryptogams, as well as in all the flowering plants, the re- duction divisions are not immediately connected, as they are in animals, with the formation of the sexual cells, but with the asexual spores from which the generation bearing the sexual organs arises. Thus, after the homotype (or second) division, an indeterminate THE STRUCTURE OF CELLS 37 number, which may be very considerable, of cell-generations inter- venes between the division in question and the differentiation of the sexual cells. It is true that, as in some of the flowering plants (the embryosac development of the lily, for example), the divisions giving rise to the four spores may be omitted, but the characteristic features of the heterotype and homotype mitosis reappear, although thus postponed, in the first divisions of the nucleus of the embryosac (macrospore). This indeed is a fact of the utmost importance, as emphasising the physiological necessity of the process ; just as it in all probability (from its community to animals and plants) pre- ceded current morphological differentiation, so now if necessary it can override morphological limitations, or at any rate it is not bound up with them. Amongst the lower plants the facts have been tolerably com- pletely elucidated in the case of Fucus, an alga in which asexual spore -formation does not occur. The nuclei of the plant possess the double (somatic) number of chromosomes until the formation of the sexual organs. The oogonium gives rise to oospheres (typically eight, though some may degenerate) by these nuclear divisions. Of these, the first two are respectively heterotype and homotype, and follow on each other with great rapidity, the last mitosis not occurring till after an interval of rest. In some of the desmids, and probably also in Spirogyra, there is evidence to show that the reduction divisions, on the other' hand, occur not at the close, but at the beginning of the life-history, with the first segmentation of the fertilised oosphere. But in the majority of these lower organisms information of a precise char- acter is still lacking on the matter. And until our knowledge of the corresponding processes in the lower animals and plants becomes, much more complete than it is at present, we can scarcely expect to solve the problem as to the utility or the necessity of the complex events connected with the reduction divisions. Although the higher animals and plants exhibit considerable diversity amongst themselves in the series of changes passed through by the nucleus in division, as well as in the relationships exist- ing between the cytoplasm on the one hand and the cell-wall on the other, they nevertheless agree for the most part in the broader out- lines. The points of similarity are, on the whole, more striking than are the differences, and the latter can often be referred with- out much difficulty to unimportant deviations from a common ground- plan. But although- this is the case, the actual meaning of the phenomena, as well as their phylogenetic origin (if there be one), can hardly be grasped or explained by a reference to these forms alone. It is in the study of the lowest forms of life that the key to the solution of cytological problems may be sought for with the greatest hope of success, for amidst the striking diversity exhibited 38 THE STRUCTURE OF CELLS in them an analysis of the processes may render it possible to dis- tinguish between the essential and what is merely accessory, and may indicate the mode and directions in which the structures characterising the higher plants and animals have been elaborated. The essence of the type may perhaps be most clearly gathered from a consideration of the deviations from it. Nevertheless, one is confronted at the outset with difficulties. For although the nuclei of many protozoa are apparently extremely simple, yet in the details of their division they may exhibit considerable complexity, and this not by any means always in the direction followed by the nuclei of higher organisms. And conversely, nuclei are not seldom met with in these low organisms which surpass those of the metazoa and meta- phyta in differentiation, whilst in mitosis they are commonly simpler. The features which seem to be common to all nuclei are — (1) the existence of chromatin in some form or other ; (2) a matrix in which the chromatin is imbedded, but which in the simplest cases may be indistinguishable from the ordinary cytoplasm. Furthermore, the fission of the chromatin is a common, perhaps invariable, antecedent to nuclear division, but it is often difficult to ascertain, and may possibly be really absent in some cases ; for example, in many amitotic divisions. The subsidiary structures, amongst which the centrosome stands pre-eminent, can only be rightly appraised when their origin has been traced in the lowest forms, in which various bodies which appear to possess functions analogous with those credited to centrosomes have often been distinguished. As regards the occurrence of nuclei in the Protozoa and the simplest plants, the investigations of recent years have tended to reduce the number of those from which nuclei were formerly believed to be absent, and at the same time it has become evident that the structure in question may be present in very different degrees of completeness. Thus in Chromatium, and probably in bacteria generally, it is not possible to speak of the existence of a definite nuclear body, but granules which on good grounds have been identified with chromatin are to be distinguished in the protoplasmic framework of the cell. In the cyanophyceae also similar granules are visible, but are restricted to a definite specialised part of the cell-protoplasm, although the latter cannot be spoken of as a nucleus. In many of the forms which possess scattered chromatin granules there is visible in the cell a body which is obviously connected with the mechanism of chromatin distribution, for on cell-division the granules of the latter congre- gate around the central body, which sooner or later divides, each half carrying with it half the chromatin, in the form of attendant satellites, to each daughter cell. In some organisms, e.g. Tetramitus, the granules are distributed through the cell- THE STRUCTURE OF CELLS 39 protoplasm in the periods intervening between two fissions, and are only intimately associated with the central body at the time of cell- division ; in others again, e.g. Chilomonas, they remain constantly grouped in the vicinity of the body to which reference has just been made. No structure has been certainly made out in it, but it has often been compared to, or identified with, the central body present in many of the more highly differentiated Protozoa, such as Euglena, and it has further been likened to the nucleolus and also to the centrosphere of those cells in which these structures have FIG. 18. Tetramitus. A, resting cell. B, early phase of division, c.b, central body ; eft, chromatin granules. (After Calkins.) been found to occur. Indeed, it would seem that there is at least some justification for the latter comparison, inasmuch as it appears at least to discharge functions somewhat similar to those performed by the centrosome though in a very rudimentary degree. A distinct advance in differentiation is reached when the chromatic and other constituents of the nuclear apparatus are not only aggregated together, but are also delimited from the rest of the cytoplasm by a wall or membrane. The degree of individuality thus obtained provides a condition favourable to further special- isation, but it seems clear that at any rate the linin framework in which the chromatin is imbedded may be fairly traced back to a THE STRUCTURE OF CELLS true cytoplasmic origin, however much it may have become modi- fied or altered under more special conditions. The chromatin in these primitive nuclei is often aggregated into clumps as in Actino- sphaerium, Noctiluca, Coccidium, etc., or even concentrated into one mass as in Actinophrysand Spirogyra. These masses have been termed nucleoli by some writers, but recent investigations tend to show that they really represent composite structures which contain other substances in addition to chromatin. In coccidia, for example, Schaudinn and others have shown that, although the chromosomes are derived from them, there exists over and above the chromatin a considerable mass of substance which is left behind after its exit, and much the same is seen in Amoeba hyalina. Possibly their analogues may be sought in certain of the so-called pseudo-nucleoli of the higher organisms, or in those occurring in the nuclei of Spirogyra, where it has been repeatedly asserted that the chromo- somes originate from the nucleolus. As regards the origin generally of the chromosomes and of the peculiar features exhibited by them during mitosis, there seems but little doubt that they have arisen through stages like those seen in tetramitus and chilomonasr in which the distribution of isolated* chromatin granules can be followed. The granules first become aggregated into definite tracts, and these form the primordia of the chromosomes themselves. The actual stages passed through are obscure, and even allied species exhibit considerable differences amongst themselves. Thus in the Amoeba, amitosis seems regularly to occur in Amoeba brevipes and A. polypodia, and also in A. crystalligera ; but in A. hyalina, accord- ing to Dangeard, the chromatin separates from a central body and is differentiated to form chromosomes, whilst the remainder of the body gives rise to a rudi- mentary spindle which is- entirely intra-nuclear. In A. binudeata the two nuclei divide simultaneously in a mitotic manner, and the same is true of the colonies of the amoeba-like myxo- mycetes. When a plas- modium is about to form spores, it may be found with nuclei showing typical karyokinetic figures, all the nuclei being in the same phase. A remarkable dimor- phism occurs in the nuclei of Paramoeba eilhardi, in which Schaudinn describes one of them as resembling a normal amoeban nucleus, B Fio. 19. Amoeba binudeata. A, with resting nuclei. the two nuclei in the aster stage of mitosis. Dangeard.) (After THE STRUCTURE OF CELLS whilst the other is much poorer in chromatin. During mitosis the two act as complements, the latter nucleus furnishing the spindle apparatus, whilst the former supplies the chromatin. Much specu- lation has been built on this case, which is assumed by some writers to indicate that the centrosome or centrosphere is equivalent, phylogenetically, to a nucleus. But it may be open to doubt whether the facts in Paramoeba have really been correctly inter- «5a» d?H2a W^IICTP^ PlQ. 20. Acanthocystis aculeata. A and S, forma- tion of the centroaome from nuclear constituents in swarm-sporec. C, resting cell. D, nuclear division preceding fission. (After Schaudinn.) preted, in the sense of regarding the spindle-forming structure as a genuine nucleus. Many cases are known in which bodies which represent centro- somes originate from the nucleus, as appears certainly to be the case in Actinosphaerium, and especially in Acanthocystis, as described by Schaudinn. In the latter animal the ordinary nuclear divisions are associated with the fission of a corpuscle or centrosome occupy- ing a central position in the cell, the nucleus lying close to its side. 42 THE STRUCTURE OF CELLS But in the numerous instances in which amitosis occurs (as in budding), the " centrosome " does not divide, and the nucleus of the freshly-budded cell possesses no such structure. Soon after the complete formation of the bud, however, a dense spot is formed within the nucleus, and is then extruded into the cytoplasm, where it continues to function as a centrosphere. The evidence, as drawn from a study of the lower organisms, seems to point strongly in favour of a nuclear origin for the centrosome apparatus in many cases, although the simplest examples cited on a previous page also indicate that, in others, such a structure might be coeval with, if indeed not actually antecedent to, the primary differentiation of a true nucleus. In a considerable number of cases it seems at least clear that the body or the substance which stimulates and brings about the division of a nucleus is derived from the nucleus itself, even though it may migrate into the cytoplasm, where it may continue to exert, under appropriate conditions, that influence on the nucleus which culminates in division. Thus, in diatoms the centrosphere is found in the cytoplasm, just as it exists in many metazoon cells. The actual location of the centrosome does not necessarily, however, settle the question as to its real origin, and it may indeed assume an intra- or extra-nuclear position in closely allied forms ; as, for example, in the two varieties of Ascaris megakcepJiala, being situated within the nucleus in the variety univalens, and outside it in the variety bivakns. A remarkable side issue has been introduced into the contro- versy as to the phylogenetic origin of the centrosome by a considera- tion of the peculiar nuclear apparatus which is met with in most ciliata and suctoria. These organisms, with a few possible exceptions, possess a mega- and a micro-nucleus, the former presiding over the somatic life and divisions of the animal, the latter only becoming prominent during the phases of sexuality. Some writers have sought to derive, phylogenetically, the centrosome from the micro- nucleus, whilst they see in the meganucleus the" representative of the metazoon nucleus. But' quite apart from the fact that, as Schaudinn pointed out, centrosomes appear in the much simpler heliozoa, the fact that the micronucleus alone divides mitotically, whilst the meganucleus always does so by amitosis, seems a serious difficulty in accepting such an interpretation. Moreover, the macro- nucleus itself springs from the micronucleus after each sexual act, and only persists till the close of the sexual cycle, at which period it totally disintegrates, and thus suffers somatic extinction. It would certainly appear that at any rate it is useless to look to such a source for the origin of the centrosome, which really seems to rest on no better basis than a purely fanciful comparison. A consideration of the maturation processes which obtain in THE STRUCTURE OF CELLS 43 the lower plants and in the protozoa when more fully understood will certainly shed light on the obscure phenomena exhibited by the higher forms, and may ultimately give the clue for correctly appreciating the general significance of tire processes involved. It has already been remarked that the exact point in the life-history at which these remarkable divisions periodically recur is not identical for all organisms, whilst the universality of the process indicates clearly its great and fundamental importance. It has been urged by some that the chromosomes, which are by those writers postu- lated as permanent structures, become distributed between the daughter cells in such a manner that only half of the original number persist in each sexual cell. And in this way room is made for the new ones imported in either of the two conjugating gametes. Others again, like Oscar Her twig, regard the quantitative reducing of the chromatin (as opposed to that of the chromosomes) as the fact of prime significance. In many of the lower forms, and notably in the coccidia, the evidence tells strongly for this view ; for in them it is the definite fact that a large part of the mother nucleus is left unused when the gametes or gametocytes are produced, and thus there is a quantitative reduction of a very pronounced character. Again, in the same organisms the multiple division of the nucleus, taken together with the amitotic division of the nucleus of the zygote at segmentation, seem to tell equally against a mechanical necessity for similarity in the chromatic strands. It is not easy to believe in the permanent existence of specific chromosomes under such circumstances ; but, on the other hand, there is no doubt that if the different chromatic granules do really represent slightly different structural characters, a qualitative reduction much in the sense assumed by him may actually take place. For there can scarcely exist any doubt but that, as the result of these pro- cesses, the surviving parts of the mother nucleus do not represent (especially after a multiple division) exact images of the original nucleus from which they have sprung. But it would also seem to be clear that whilst both a quantitative and a qualitative reduction have taken place, these can hardly be regarded as direct means of ensuring that an unvarying proportion of the original chromatin shall be distributed amongst the daughter nuclei. Whether the constant proportions observed in the higher forms is to be explained as the result of a more definite constancy in the chromosomes, together with the continuous existence of these structures in the hypothetically more specialised nuclei of the higher animals and plants, is a matter upon which it is as yet impossible to make any positive statement. It may, however, be confidently asserted, having regard to the extraordinary diversity which prevails in the details of nuclear transformations in these lower forms, many of which will be found described in the present volume, that amongst them, 44 THE STRUCTURE OF CELLS if anywhere, are to be sought the clues to the complex though less variable processes which characterise the higher animals and plants. The phenomena of the sexual union of germ cells, and of their contained nuclei, for the most part are, as yet, hardly susceptible of detailed explanation, but there can be no doubt but that chemiotaxis is the proximate factor chiefly concerned. This is beautifully shown in the case of Halidrys, one of the Fucaceae, in which the large eggs attract numbers of sperms which seek to penetrate the egg. Immediately after an entrance has been effected by one of them, it is seen that the egg changes in important respects. It shrinks, and the supernumerary sperms instantly cease their endeavours to enter its substance. On the contrary, they swim rapidly away, and from the surface of the egg a substance is seen to be excreted which probably exerts the repellent influence in question. Indeed, so strong is its action that those sperms which have not quitted the surface of the eggs are rapidly paralysed and killed. The remarkable paths followed by the male nucleus in the egg has been studied by many observers, and there can be but little doubt that here also a specific attraction of some sort effects the final union. As to the significance of the fusion, the evidence at present available points in no certain direction, nor can any of the hypotheses which have as yet been advanced to explain it be regarded as affording satisfactory solutions of the problem. It has been assumed that by its means a sort of rejuvenescence is effected. But this idea, which is not very clear in itself, fails to take the many subsidiary but still general recurring circumstances into con- sideration. Moreover, it is difficult to see why a similar explana- tion should not also cover those vegetative fusions common in endosperm cells and in certain fungal hyphae, but these have never been regarded as constituting sexual acts. And indeed it would appear that the actual initiation of segment- ation in an egg is not necessarily dependent on the fusion or even the presence of two nuclei. Boveri's observations on the fertilisa- tion of enucleated fragments of eggs with sperms, and still more those of Loeb, who succeeded in causing the eggs of sea-urchins to segment parthenogenetically by treating them with magnesium chloride, indicate that the matter is of far greater complexity than a study of the normal occurrences would indicate. Again, Nathan- sohn caused parthenogenetic development of the oospheres of Marsilea to take place by keeping them at a sufficiently high temperature. This last observation seems to be one of greater importance, for it suggests that a slight modification of the metabolic processes, in this case effected by the abnormal temperature, may suffice to set the machinery of segmentation in motion ; that is, the actual THE STRUCTURE OF CELLS 45 mechanism is already present in the egg, and only an appropriate stimulus (not the importation of a missing half of the machinery) is required to set it in motion. Long ago Boveri suggested that the centrosome rather than the nucleus was the important body the introduction of which starts the process of segmentation, and it may well be that his suggestion, in a modified form perhaps, and without postulating an organisation of the specific excitatory substance in so definite a form as a centrosome, may contain a con- siderable element of truth. Amongst the lower organisms, as Klebs and others have shown, the conditions favourable to the formation of sexual cells can be largely referred to nutritive sources, and this is only another way of saying that a definite stimulus — ultimately working on the living substance of the organism itself — is responsible for the sexual reaction. But such a view of the matter leaves untouched the question of the secondary utility of sexuality as a means of ensuring variation. Indeed, this latter is perhaps best kept distinct from the primary causes and conditions which first made sex not only a possible but an inevitable incident in the life cycle of the greater part of the higher organisms. Less obscure than its relations to the phenomenon of sex are those which exist between the nucleus and the life of the cell. Gruber, Nussbaum, Verworn, and others have shown that in protoplasm which has been deprived of its nucleus the vital functions speedily become more or less deranged, and finally cease altogether. Enucleated fragments of a cell or organism fail to regenerate lost parts, and the apparent exceptions that have been met with constantly turned out, on further investigation, to be contaminated with nuclear influence. Enucleated fragments of an amoeba are unable to excrete the substance which normally enables these animals to cling to the substratum, and in other organisms, although food may be ingested, the protoplasm seems unable to digest and assimilate it. In plants Gerassimoff has shown that cells containing chlorophyll but destitute of a nucleus are usually unable to form starch, and are incapable of excreting a limiting membrane over their free surfaces. On the other hand, allusion has already been made to the transformations the nucleus may undergo in connection with the secretory activity of glandular and other cells, and in this connec- tion, no less than in that of regeneration, the nucleus may be affirmed to preside over the metabolism of the protoplasm. The peculiar processes which are extruded by the nuclei of the nutritive cells surrounding the ovum (Ophryotrocha), and the curious simulation of the initial phases of mitosis met with in secretory cells, can hardly be dissociated from the special functions discharged by their nuclei. 46 THE STRUCTURE OF CELLS Again, that frequent migration of the nucleus to the seat of special metabolic activity, so often illustrated in plants, affords additional indication of an exercise of the same influence. The developing of a lateral outgrowth on a plant hair, one-sided thickening of the cell-wall, the softening of the latter previous to its. perforation — all these are commonly preceded by the arrival of the nucleus at the part of the cell about to be affected. And the very facts of mitosis itself, with the profound chemical and physical changes which accompany it, suffice of themselves to prove that the nucleus contains substances which are capable of undergoing rapid and striking changes. And indeed it is perhaps not improbable that it is in that very lability characteristic of the constitution of the complex substances which together make up a nucleus that the supreme importance of this body to the cytoplasm, and through the latter to the organism as a whole, is to be attributed. THE PEOTOZOA (continued) SECTION I. — THE FORAMINIFERA l CLASS FORAMINIFERA Order 1. Gromiidea. 2. Astrorhizidea. 3. Lituolidea. 4. Miliolidea. 5. Textularidea. 6. Chilostomellidea. 7. Lagenidea. 8. Globigerinidea. „ 9. Rotalidea. „ 10. Nummulitidea. THE Foraminifera received their name before their nature was understood. The early anatomists, guided by the likeness of many of their tests to the nautiloid shells of the Cephalopod Mollusca, assigned them to this group, and many Foraminifera were included by Linnaeus and later writers in the genus Nautilus. D'Orbigny, in 1826, divided the "Cephalopoda," having chambered shells, into Siphoniferes, ~with a more or less tubular siphon traversing the series of chambers ; and Foraminifkres, in which the chambers are in communication by foramina. The simple character of the organ- isms which secreted these shells was first recognised (1835) by Dujardin, who placed them with Amoeba, and allied fresh -water forms in the group Rhizopoda. The limits of the group Foraminifera, as here understood, are identical with those of the Reticularia, as defined by Carpenter in his Introduction to the Study of the Foraminifera (8). It includes those Protozoa the protoplasm of which secretes a test (or shell), and is protruded in fine thread-like pseudopodia, which branch freely and anastomose with one another, and present no obvious differentiation into ectoplasm and endoplasm. The great majority of the members of the group form a well- defined assemblage of organisms, clearly allied to one another, and distinct from any other division of the Protozoa ; but we cannot at present draw with any certainty the limits between the simpler forms here included and some other simple members of the Protozoa. 1 By J. J. Lister, M.A., F.R.S., Fellow of St. John's College, Cambridge. 47 48 THE FORAMINIFERA Each of the characters by which the group is defined loses in distinctness when followed into this borderland region. The shell, which attains great complexity in the higher forms, is membranous in many of the lower, and in Lieberkuhnia, Diplophrys, and Myxotheca (Fig. 2) can hardly be said to exist. In Hyalopus the pseudopodia, though branching and pointed, do not anastomose with one another (Fig. 15), and in several of the fresh- water Gromiidae few anasto- moses are found. The filiform nature which distinguishes the pseudopodia of the Foraminifera from the blunt-ended pseudopodia of the Lobosa is a better defining character; but in view of the close parallel which, as will be explained below, the life -history of Trichosphaerium, a member of the Lobosa, appears to present with that of many Foraminifera, its importance in a natural classification may be doubted. A few of the simpler forms live in moor pools and other fresh waters, but the great majority are marine. Most of these are littoral in habitat ; many extend their range to the floors of the deep oceans ; while a small group, few in the number of species, but very abundant in individuals, lead a pelagic existence. The Protoplasm presents a uniform character without any obvious separation into an outer and inner layer of different refracting power. It is finely granular throughout, and coarse granules are usually present in the protoplasm contained within the test. The pseudopodia in some of the lower forms are long and root- like, and extend to great distances, giving off branches in their course, and to such forms Dujardin's name Rhizopoda is especially appro- priate ; but most members of the group are characterised by pseudo- podia of a different nature (Fig. 1). They are for the most part very slender, and spread from the neighbourhood of the aperture or apertures of the shell in fan-like or sheaf-like groups. The outer surface of the shell is invested by a layer of protoplasm, and from this also groups of pseudopodia originate. The pseudopodia frequently branch and anastomose in their course, and between the points of union of the reticulum which they form, they generally run straight, owing in part to the tension which they mutually exert on one another. Some of the radiating strands form broad bands, branching peripherally, but the majority are exceedingly slender, and the ultimate branches are of extreme tenuity. At the points of union broad expansions of protoplasm are often formed. The network of pseudopodia is in part projected free in the water, in part applied to surrounding objects, and serves at once for the prehension of food, as a peripheral sensory apparatus, and as a means of attachment and of locomotion. When a strand of the reticulum is attentively examined, the granules contained in the protoplasm are seen to hurry along its FIG. l. Gromia oviformis, Duj. 1, contour of protoplasm contained within the test ; 2, protoplasm reflected over the test 3. extended pseud opodia. (From Shipley and MacBride, after Max Schultze.) 50 THE FORAMINIFERA surface, as though borne by a stream, and two streams of granules flowing in opposite directions, centrifugal and centripetal, are to be seen on opposite sides of the same filament. Sometimes a mass of protoplasm forms a swelling on the surface, and is carried along for some distance before it thins out and merges in the substance of the filament. The tip of a pseudopodium may be seen to be alternately ex- tended and retracted according as the centrifugal or centripetal stream gains the ascendency. A mass more solid than the rest of the protoplasm may be seen to be carried to the tip, turn and pass back for some distance with the return current, then to be caught in the centrifugal stream, and again carried to the tip. One is reminded of a cork at the summit of a jet of water, under the contending forces of the upward flow of the jet and of gravity. But the cause of the streaming movement in the pseudopodia of the Foraminifera, like the ultimate cause of movement in all contractile tissues, is still beyond the limits of our knowledge. The minute structure of the protoplasm has been carefully ex- amined by Biitschli, who finds in the coarser pseudopodia, and in the membranous expansions between them, the alveolar structure which is present in so many protoplasmic structures. In the fine pseudopodia, however, this is absent, and they appear under the highest powers as homogeneous threads of extreme tenuity, with the small granules scattered along their surface. From the peculiar way in which, in the living pseudopodium, the granules course along the threads, sometimes leaving them for a moment to pass apparently through the open water, Biitschli is inclined to the view, suggested by Max Schultze, that an invisible hyaline layer may invest the visible threads of the Foraminifera in the same manner as the more visible hyaline ectosarc invests the axial endosarc of the pseudopodia of the Heliozoa. In addition to this intrinsic streaming movement there is a movement of the reticulum as a whole. New pseudopodia shoot out from the central mass, others are shortened and retracted, and the whole system is in a condition of tension and constant move- ment, as becomes evident at once when an attempt is made to draw any part by camera lucida. When a strand of the reticulum gives way a momentary collapse of the neighbouring strands is seen, followed by the rapid lengthening of some strands and shortening of others, resulting in renewal of the tension. The food of Foraminifera consists largely of diatoms and algae, either alive or in a state of decay. In some cases, however, it is of animal nature, for Khumbler finds that the Globigerinae capture and digest the Copepods which abound at the sea surface, and that the pelagic Pulvinulinae contain the skeletons of Radiolaria as well as of diatoms, which have been taken as food (38, pp. 1 and 2). THE FORAMINIFERA Schaudinn also has witnessed the capture and digestion of a Copepod by Myxotheca (41, p. 25), and finds that Patellina and Discwbina feed on Copepod nauplii and Infusoria as well as on Diatoms (45, p. 182). The pseudopodia of Gromia and Poly- stomella have been seen to exercise a paralysing effect on Infusoria which come in contact with them (M. Schultze). The pseudopodia are exceedingly viscid. When an object which serves as food is entangled it becomes surrounded by protoplasm, and if it is large the strands of the reticulum between it and the shell become thicker and more numerous, and the object is drawn inwards. Whether digestion may occur in the extended protoplasm or only after the food enters the shell is uncertain. FIG. 2. Myxotheca arenilega, Schaudinn. N, nucleus ; t, the gelatinous test with embedded sand grains. (After Schaudinn, 41.) Contractile vacuoles occur in some of the fresh -water forms (Euglypha, Trinema, Cyphoderia, Microgromia, and Platoum), but they have not been seen in any of the marine genera. The granular bodies which are scattered through the protoplasm are of different kinds. Some are coloured, and confer a red, yellow, or brown colour on the protoplasm when seen in bulk. These are allied in composition to the colouring matter of diatoms (diatomin), and are probably derived directly from the food. Some are fatty ; others, apparently proteid in nature, are stained by picrocarmine, and are probably formed in the ascending metabolism of the food. In Orbitolites complanata starch grains are abundant, but their formation is probably dependent on the presence of the parasitic algae (zooxanthellae) which abound in the protoplasm in this species. These algae also live in the pelagic Gloligerinae. 52 THE FORAMIN1FERA The nuclear characters and modes of reproduction of the Foramini- fera are considered below. The Structure of the Test. — The test is found in its most rudiment- ary condition in Myxotheca, where it consists of a gelatinous layer, which may form the whole covering or may contain grains of sand. The shape follows the changing contour of the protoplasm, the pseudopodia break through at any point, and no definite and permanent orifice is formed. In Hyalopus (Fig. 15) the test is more resistent, and may have an oval form and a definite orifice ; but here again the shape varies with that of the contained proto- plasm, becoming arborescent when growing among the crowded stems of algae, and the number of orifices may be indefinitely in- creased (Schaudinn, 43). In most of the Gromiidea the test is chitinous and flexible ; but it has a definite shape, and one or two permanent orifices. Another type of test is found in some Gromiidea, and in all the Astrorhizidea and Lituolidea. The tests are here formed of foreign particles, such as fragments of sand, sponge-spicules, the shells of other Foraminifera, etc., fastened together by a cement, which may be firm or flexible, and consist of chitin or calcium carbonate or ferric oxide. In the Astrorhizidae the walls are thick and soft, con- sisting of mud, or of only slightly cemented sand (Fig. 17, a\ while in other cases, as in Saccammina (Fig. 17, 6), the particles are united into a rigid structure. In some species of Textularia, Quinqueloculina, and other genera, though the tests are chiefly calcareous, a large proportion of foreign arenaceous material is contained in them. A very remarkable feature of the tests of arenaceous Foramini- fera is the evidence they appear to offer of a selective power exercised by certain species, in collecting materials. In some cases, no doubt, the nature of the test depends on the constituents of the sea bottom in which the animal lives, but in others certain elements alone are selected. Thus in the same dredgings may be found the tests of Pilulina and Saccammina, the former composed of a close felt of siliceous sponge-spicules, laid together to form a wall of uniform thickness ; the latter of coarse grains of sand united by cement (Fig. 17, b and c). In both the test consists of a single spherical chamber, and the size attained is about the same in both. The cylindrical tests of Bathysiphon filiformis, Sars, are composed of a felt of sponge-spicules, covered externally by a layer of fine sand. The same sample of Pteropod ooze supplies representatives of species of the family Lituolidae, characterised by the coarseness of the sand grains of which the tests are composed ; and of Trochammi- nina distinguished by fine-grained tests. The type of test found in Euglypha (Fig. 3) and its allies is very exceptional. It is formed of rounded or hexagonal plates, of siliceous THE FORAM1NIFERA 53 or, in some cases, chitinous nature, which are secreted in the sub- stance of the protoplasm in the neighbourhood of the nucleus. Passing to the surface, they are there built together into a regular test. With the exception of some forms of Biloculina, which have been found when living at great depths with a siliceous shell instead of the normal calcareous one, this is almost the only instance of the secretion by the Foraminifera of a siliceous skeleton — a fact which is all the more remarkable in considera- tion of the prevalence of siliceous skeletons among the Radiolaria and He- liozoa. CtA Fio. 3. Euglypha alveolate. Two stages in the process of division. In one the nucleus is about to divide, and the plates which In most of the other orders of Foramini- fera, though a chitinous element is present in t>»A aL-AWnn if ia /™!TT In one the nucleus is about to divide, and the plates which SKClCtOn, It IS Only wm form the new shell are seen in the protoplasm ; in the the basis in which Car- other the division is nearly complete. c.v, contractile vacuole. (After Schewiakoff, 48.) Donate of lime, together with a small proportion of carbonate of magnesium and traces of other salts, are deposited.1 1 The following analyses are given by Brady (3, pp. xvii. and xxi.) of the tests of two species — one porcellanous, belonging to the Miliolidea ; the other perforate (see footnote, p. 54), a member of the Nummulitidea : — Orbitolites complanata, var. ladniata. Silica ..... ... 0*14 Carbonate of lime . . . . 88 '74 Carbonate of magnesia . . . . 9 '55 Alumina with phosphate of lime and magnesia . . \ occasional Alumina and ferric oxide . . . . / traces Silica Amphistegina lessonii. Carbonate of lime with a little organic matter Carbonate of magnesia .... Alumina with phosphates of lime and magnesia . Ferric oxide 98-43 0-30 92-85 4-90 1-95 trace 100-00 Sollas has examined the specific gravity of the shells of perforate and imperforate species of Foraminifera (66, p. 374), and finds that in perforate forms it varies from 2'626 to 2*674. In examples of Mttiola, Peneroplis, and Orbiculina it varies from 27 to 54 THE FORAMINIFERA In these the tests are rigid structures, and communicate with the exterior either by one or more large apertures exclusively, as in the majority of the Miliolidea, or by a multitude of small pores in their walls in addition to the large apertures.1 In some genera of the Miliolidea the shells have a polished white appearance re- sembling porcelain when seen by reflected light, and a yellowish brown horn colour when seen by transmitted light ; in the per- forate forms the tests are more transparent, and are in many cases as clear as glass. On this account the forms with perforate cal- careous tests are often known as the vitreous or hyaline Foraminifera. In the perforate forms the pores passing out from the chambers have, on the whole, a direction perpendicular to the walls, and transmit pseudopodia. In some cases the pores are large and com- paratively few, in others they are fine and very numerous. The Growth of the Test. — In forms such as Myxotheca and Gromia the growth of the protoplasmic body is accommodated by the simple expansion of the soft and membranous test. Among the arenaceous forms, the feebly cemented, star-shaped tests of Astrorhiza (Fig. 17, a) increase in size in part by the extension of the test along the protoplasmic trunks forming the rays of the star, but in part, doubtless, by the expansion of the central body. In the case of Saccammina (Fig. 1 7, b), however, and other forms with rigid single- •2722. The specific gravity of calcite is 272, and that of aragonite 2'97. He con- eluded that the calcareous constituent of the perforate tests is calcite, and that of the imperforate tests either aragonite or calcite together with some other and heavier substance. Cornish and Kendall (12) had previously indicated the conclusion, though without positively stating it, that the porcellanous Foraminifera were composed of aragonite — on the grounds of their opacity, and their appearance in or absence from beds coincidently with Lamellibranch and other fossils which are composed of ara- gonite. Chapman (11, p. 39) also has recently stated that the tests of the Miliolidea are of aragonite, or rather (following Miss Kelly, Mineralogical Mag. vol. xii. (1900), p. 363) conch ite. I am inclined to doubt this conclusion. It appears that the presence of the mag- nesium carbonate (specific gravity 3*056), which Brady's analysis shows is a larger constituent of the imperforate tests than of the perforate, may cause the higher specific gravity found in the former by Sollas. Moreover, Meigen (22) has recently described a chemical colour test by whicli calcite may be distinguished from ara- gonite, or from the constituent which Miss Kelly has named conchite. Tried by this test, I find that MUiolina and Orbicidina and Orbitdites do not give the colour reaction characteristic of aragonite, but agree with structures composed of calcite. I am indebted to Mr. A. Hutchinson for calling my attention to Meigen's paper. 1 In the systems of classification prepared by Reuss (1861) and by Carpenter and his colleagues (1862) the Foraminifera were divided into the two groups, the Imperforata and the Perforata. In the latter classification the group Imperfor- ata includes the Groniida, Lituolida, and Miliolida; the Perforata the Lagenida, Qlobigerinida, and Nummulitida (i.e. families 5-10 of Brady's classification, which is followed iu this article). The progress of deep-sea investigation since this date has revealed the existence of the Astrorhizidea, some of which have perforate and others imperforate tests. Moreover, even in the Miliolidea the first formed chambers may be perforate in the genus Peneroplis (Rhumbler), and in Orbicidina and Orbitolites, aa shown in this article. It thus becomes evident that the presence or absence of perforations in the shell, though perfectly characteristic of many of the orders, cannot be taken as the basis for the subdivision of the whole group. THE FORAMINIFERA 55 chambered tests, the way in which the growth of the protoplasm is accommodated is less obvious. It is probable, however, that here also, the shell, though at any particular moment rigid, is slowly moulded and expanded under the influence of the protoplasm. The small tests hitherto classified as Psammosphaera fusca, but regarded by Rhumbler (33) as the young of Saccammina sphaerica, with which he finds them to be connected by all intermediate forms, are built up of fragments of sand placed together irregularly, so that the contour of the young test is rough and uneven. In the full-grown Saccammina test the fragments are placed, as in a well-constructed stone wall, to form an even contour. If, as these facts imply, a change in position of the constituents has occurred during growth, there is no difficulty in accepting the conclusion at which Rhumbler arrives, that it has also undergone expansion as a whole. The alternate hypothesis, that in the course of growth to its full size (3*5 mm.) the protoplasm periodically discards its old shell and builds a new one, appears improbable, and the student of the growth of bone will find no a priori difficulty in admitting that a rigid structure may be the seat of profound inter- stitial changes of substance. The growth of the tests of the cylindrical forms of the Astro- rhizidea is effected by extension in a linear direction, fresh arenaceous material being incorporated at their ends. They form simple or branched (many Hyperamminae, Rhizammina), but usually un- segmented tubes. In Hyperammina subnodosa (Fig. 1 7, d) the tubes are constricted at irregular intervals, and thus present a transitional condition of structure to the definitely segmented, chambered shells of the great majority of Foraminifera. In the latter, while the growth of the protoplasm as the result of the assimilation of food is continuous, the growth of the shell is not continuous but periodic. When a new chamber is to be formed, a mass of protoplasm is protruded from the mouth of the shell, and at the surface of this the new wall is formed, by secretion in the case of calcareous shells, by cementing together of foreign elements in the arenaceous forms. In some genera the secretion of shell substance takes place only on the free surface of the protoplasm, but in others it occurs also where the protoplasm rests against the previously formed test. In such cases the septa dividing the chambers are double, and the new chamber is complete on all sides with the exception of the aperture or apertures left for communication with the exterior, or with its successor, when a new chamber shall be added. In either case the result of this periodic shell formation is the building up of a segmented test, the segments of which, the chambers, are sharply marked off from one another. Max Schultze found that in the formation of a new chamber by Polystomella, the deposit of calcareous salts began before the chamber had 56 THE FORAMINIFERA assumed its final form. The small pocket-like " retral processes," which are characteristic of this genus (see Fig. 7), were not formed until some time after a continuous wall had been secreted, so that a partial absorption and re-deposition of the lime salts must here occur (64, p. 30). Structure and Mode of Growth of the Shell Wall. — After their first formation by secretion at the surface of a mass of protoplasm, the walls increase in thickness by the addition of shell material to the outer surface. The anterior wall of each chamber is soon covered by the addition of a new chamber in front of it, and it then forms the whole, or half (as the septa in the species concerned may be single or double), of the septum dividing the chambers from one another. On the septa of the perforate forms the pores may be limited to the peripheral parts or absent altogether. The thickness attained by the septa is not great, but the part of the wall turned to the outer world continues to grow, and may attain considerable thickness. The thickening results from the addition of successive layers of material on the outer surface, and thus a laminated structure is produced ; but though laminated tangentially, the shell is built up, where it is perforated, of hexagonal prisms disposed radially to the surface, and each traversed by a pore transmitting a pseudopodium. It appears that this result may be explained as follows. The shell material is deposited by the protoplasm traversing each of the pores which perforate the wall, on the area about its orifice. It would appear that at the limits between the areas influenced by neighbouring pseudopodia there is some slight difference in the character of the material secreted, and the result is that the deposit is not quite uniform, but marked out into small hexagonal areas, with a pore at the centre of each. If this is the case, the prismatic structure results from the observance of the same limits in successive layers throughout the thickness of the shell. In the more complicated perforate Foraminifera a system of sinuses or canals is present, the main channels of which run in the substance of the shell, and are distinct from the cavities of the chambers, though communicating freely with them by branches. This is known as the canal system. It is, of course, wholly distinct from the radial pores leading direct from the chambers to the exterior. The details of its distribution in PolystomeUa are given below (cp. Fig. 9). It will be seen that in this form two main "spiral canals" run on either side of the test, parallel with the series of chambers, and give off branches, some of which run in the thickness of the septa between the chambers, while others pass direct to the exterior in the axial regions of the test. About the ultimate branches of the canal system a deposit of shell substance is laid down, which may be called the canalicular skeleton. In the test of Polystomella this skeleton is mainly limited to the axial regions, THE FORAM1NIFERA 57 but in other genera it forms extensive deposits in the interior of the septa, and on the surface of the tesf. It attains a great development in Calcarina.1 Where the canalicular skeleton comes in contact with that of the chamber walls, the two merge insensibly into one another ; the only distinction between them is that one is penetrated by the branches of the canal system, the other by radiating pores leading direct from the chambers of the test. Some confusion has arisen in the use of the terms intermediate or supplemental skeleton, and proper chamber wall. In the Introduction to the Study of the Foraminifera (p. 50) Carpenter says that the "intermediate or supplemental skeleton " is " formed by secondary or exogenous deposit " ; and further, that wherever developed to any considerable extent, " it is traversed by the canal system." The statement that the supplemental skeleton is formed by a secondary or exogenous deposit appears unfortunate, for the walls of the chambers of all the perforate calcareous forms are at first exceedingly thin, and they increase in thickness by the deposition of shell substance on their outer surface, so that the greater part of the shell in all may be said to be secondary and exogenous. The part of the shell first formed is in most, if not all, cases quite indistinguishable from that which is added later. The second character of the supplemental skeleton given by Carpenter, that it is traversed by the canal system, does, however, touch on a real distinction. Biitschli (6, pp. 26-27) calls attention to the fact that a difference between a primary shell layer (Carpenter's "proper wall") and a secondary mass is often indistinguish- able ; but he proposes to use the latter term for the outer layer of the shell, whether unperforated, perforated by radial pores, or by branches of the canal system. It appears to be more advan- tageous to distinguish the skeleton developed in relation with the canal system from that of the chamber wall, and as confusion is attached to the name supplemental skeleton, the term canalicular skeleton is used in this article for the former. Repair. — The power of re- pairing injuries is very great, and indeed a fragment may, in some cases, give rise to a new individual. This is well seen in the specimen of Orbitolites tenuissima shown in Fig. 4, 1 Cp. Carpenter, 8, p. 216. FIG. 4. Specimen of Orbitolites tenuissima in which a fragment of a test has given rise to a new disc. (From Carpenter, 9, Plate I. Fig. 7.) 58 THE FORAMINIFERA in which it is interesting to note that the centre of symmetry of the growth which occurred after the injury is entirely different from the original one. Verworn (67, p. 455) finds that when a specimen of Polystomella crispa (his observations were doubtless made on megalospheric forms, see below) is broken into fragments, several of the larger pieces remain alive and extend pseudopodia, but new shell is secreted o/er the broken surfaces by only one, and this is found on examination to be the fragment in which the nucleus is contained. The Form of the Test. — The principal forms of test met with among the Foraminifera will be considered later. We shall see that in :.ome genera a particular mode of growth, in relation to some simple symmetrical plan, whether rectilinear, piano -spiral, helicoid, o annular, is observed with perfect regularity, while in others i- symmetrical plan is only loosely followed. Many species are adherent to other objects, and in them the chambers may be " heaped " together irregularly, forming what are known as the " acervuline " tests.1 Some of the adherent forms take on an arborescent shape. Multiform Tests. — A remarkable phenomenon is presented by many genera, and that is that the plan on which the chambers are arranged in the growth of the test changes in the course of growth. In such tests the chambers which succeed the central one are arranged on a particular plan, whether piano-spiral, helicoid, or some other, and after growth has progressed on this plan for some time a change occurs, and a new plan of growth is with greater or less abruptness adopted (Figs. 24, 39, 40, 44, etc.). In some cases the plan of growth may be changed more than once before the test is completed. Thus two or more types of arrangement of the chambers are, in these genera, presented by the same test at different stages of its growth. The genus Spiropleda is an example in which the chambers are at first uniserial and arranged in a piano-spiral, and later biserial and in a rectilinear series (Fig. 44, A and B). To tests exhibiting such different modes of growth the terms dimorphic, trimorphic, and polymorphic (according to the number of forms of growth present) were originally applied, and the phenomenon of their occurrence was spoken of as dimorphism, trimorphism, or polymorphism. But it has since been discovered that two kinds of individuals occur in the life-cycle of many For- aminifera, and for this, which is, of course, an entirely distinct phenomenon, the term dimorphism has, in accordance with customary biological usage, been adopted. It has thus become necessary to find other terms to characterise tests displaying two or more modes of growth, and the adjectives biformed and informed may, as proposed by Rhumbler (36, p. 63), be con- 1 Actrvus, a heap. THE FORAMINIFERA 59 veniently used for this purpose. In what follows I have used the term multiform to cover any departure from the uniform condition of growth. It is shown below that while in some genera both forms of individual which are found in one species are alike bi- or tri-formed, FJG. 5. Block of Eocene limestone, showing the two forms of a species of Nitmmulites, constituting "a pair" ; the larger named .V. Uarritzenti*, the smaller N. guettardi, cl'Arch. Specimen from Deir en Nakhl, Egypt, in the Brady Collection, Cambridge. in others the phenomenon is exhibited by only one form (the microspheric), or exhibited by it to a greater degree. The Phenomenon of Dimorphism. This phenomenon was first recognised in the fossil immmu- lites which abound in the marine deposits of the Eocene period, and are represented by a single species living at the present day. 6o THE FORAMINIFERA They are often so abundant that, as in the rock of which some of the Egyptian pyramids are built, their coin-like shells, whole or in fragments, constitute the main part of the deposit (Fig. 5). The shells are, in reality, not flat but biconvex discs, and the chambers, arranged in a spiral, are so disposed that the greater part of the cavity of each lies in the median plane, while the shell on either side of this plane is comparatively solid. They thus readily break, as the result of weathering or by artificial means, into plano-convex halves, which display a section of all the chambers from the centre to the periphery on their broken faces. It has long been recognised that while the great majority of the specimens of nummulites occurring in a deposit attain a certain, moderate size, a few are found scattered through it, whose diameter far exceeds that of the others. On examining median sections of the smaller specimens it is usually found that the spiral series of chambers starts from a large and nearly spherical chamber readily visible to the naked eye, and occupying the centre of the shell, while in the larger specimens the spiral series is continued to the centre, where, in carefully prepared sections, it may be seen to take its origin in a spherical chamber of micro- scopic size (Fig. 6). Although the two forms were thus found to be associated in the same beds, and to agree with one another closely except in the size to which they grow and the characters of the central chambers, they were given separate specific names, and attention was called to the puzzling occurrence of these associated pairs of species, a large and a small one, in various deposits. Thus the names Nummulites laevigata, Lam., and N. lamarcki, d'Arch., have been given to two associated forms occurring in beds of the Middle Eocene formation. The former attains a diameter of 20 mm., while the latter does not exceed 3 or 4 mm. Small examples of N. laevigata are not to be distinguished by external characters from the associated form, but on splitting them open, the difference in their central chambers is at once apparent (Fig. 6). Sixteen pairs of similarly associated " species " belong- ing to the genera Nummulites and Assilina have been enumerated. The possibility that the two associated forms might belong to the same species was, however, entertained by several observers, and the acceptance of this view was accelerated by Munier- Chalmas (26), who (in 1880) definitely formulated the conclusion that the species of nummulites are dimorphic, each appearing under two forms, a large one and a small one. He also expressed the opinion that the phenomenon of dimorphism would be found to be of general occurrence among the Foraminifera. As already stated, the large forms with a small central chamber are much less abundant than the others, and it so happened that THE FORAMINIFERA 61 young individuals of this form did not come under Munier- Chalmas's observation. He was thus at first inclined to the view that the individuals of the two sets, although in some way dis- tinct in nature, began life under one form, namely, that with a large central chamber. At a certain stage, it was supposed, the growth of one set of individuals was arrested, while in the members of the other set the walls of the large central chambers were absorbed, and growth was continued not only by the addition of chambers at the periphery in continuation of the series of those already formed, but also in a centripetal spiral towards the centre FIG. (5. Nummulites laevigata, Lam. A, Central portion of a section of the megalospheric form (" N. lamarcki," d'A.) ; B, of the microspheric form. Both x 10. (After de la Harpe, 17.) of the shell, filling the space originally occupied by the large central chamber. This idea of the relationship of the two forms was controverted by de la Harpe, who pointed out, expressing his own views and those of de Hantken, that young examples of the form with a small central chamber are known to occur, and also that differ- ences may be detected not only at the central parts of the shells of the two forms of nummulites, but throughout the series of chambers. Thus it is often found to be the case that in the forms with a large central chamber (A, Fig. 6) the maximum size of the chambers subsequently added is attained early in the series of whorls, while in the others (B) the size of the chambers gradually increases to the last whorl. While the view that one form results from the modification of the other was thus shown to be untenable, it was suggested that 62 THE FORAMINIFERA they might with more reason be regarded as representing the two sexes of a species. The authors did not, however, abandon the old idea of the specific distinctness of the two forms. Investigation of other genera of Foraminifera has shown that the phenomenon of dimorphism is, as Munier-Chalmas expected, widely found among them. The relation between the two forms will be best elucidated by examining the structure and life-history of a single species. For this purpose we will select Polystomella crispa (L.), the life-history of which is most completely known. THE STRUCTURE OF POLYSTOMELLA CRISPA. — This is one of the most abundant of the littoral Foraminifera. It lives in shore pools and down to a depth of 355 fathoms, and ranges from Green- land in the north and Kerguelen Island in the south to the equator. It is very common on our own shores. The test is biconvex and symmetrical about the median plane. The chambers are arranged in a spiral series, and are equitant, i.e. they bestride the chambers of the preceding convolution and over- lap them at the sides, each being prolonged in what are known as alar prolongations, which extend towards the spiral axis of the test. Partly as the result of this overlapping, and partly because the axial region is filled in with canalicular shell substance, only the last convolution of chambers is visible externally. On the terminal face of the last chamber, where this face joins the wall of the preceding convolution, a V-shaped line of pores (Fig. 7, a) is visible. These represent the aper- ture of the test, and are the main channels of communication between the terminal chamber and the exterior. At the posterior margin (i.e. the margin remote from the terminal face) of each chamber a number of pocket -like prominences, the retral processes (Figs. 7 and 8, r), project back- wards, and are marked by ridges on the external surface. They end blindly, but are separated by pits, at the bottoms of which are the openings of branches of the canal system, which will be described later. The outer surface of Fio. 7. X about 40. , retral pro- The test of Pdystvmtlla crispa (L.> a, the line of terminal apertures ; cesses ; between them are seen the pits by which branches of the canal system communicate with the exterior. THE FORAMINIFERA 63 the shell is dotted over with minute tubercles (not visible in Fig. 7). A keel-like thickening runs round the margin of the test, and in some specimens small spines, like the points of a spur, project from it at the places where the septa join the keel. These are more frequently present in the earlier than in the later convolutions. The pores traversing the walls of the chambers are in this species exceedingly minute. They are hardly visible when the test is seen from without, but they may be detected when a broken piece of the wall is highly magnified and seen by transmitted light. On examining the external characters of the tests of a number of Polystomellas, they are found to form a uniform series, presenting such gradations of size from small to large as may be seen, for example, in a sample of the shells of any Mollusc which contains young and old. If, however, a batch of living Polystomellas is killed by some reagent which dissolves the shell but preserves the protoplasm filling its chambers, the protoplasmic casts of the shells no longer form a uniform series but fall into two sets (Fig. 8). In the great majority of them the series of chambers, when traced to the centre of the spiral, is seen to take its origin in a large spherical chamber, having a diameter generally between 60 and 100 p. In the others a small central chamber, with a dia- meter of about 10 /x,1 occupies the centre of the test, and the suc- ceeding chambers of the series are at first correspondingly small, so that for a given diameter of test these specimens have a greater number of chambers than the others. It is clear that though in Polystomella there is no marked differ- ence in the size attained by the two forms, we have here the same phenomenon of dimorphism which is exhibited in the nummulites. The large central chamber is known as the megalosphere, the small one as the microsphere, and the two sets of individuals of the species are known as the megalospheric and microspheric respectively. The numerical proportion of the two kinds of individuals probably varies with the season, but the megalospheric form is here also always the more abundant. In a large batch of several hundred specimens the megalospheric forms were found to be thirty- four times as numerous as the microspheric. In the protoplasmic casts obtained in this manner the form 1 The diameter of the microsphere varies iu the specimens of P. crispa which I have seen from 6'5 to 13 M. That of the megalosphere from 165 to 35 ^. These dimensions fall, however, a little short of the actual diameters of the chambers, owing to the shrinkage of the protoplasm produced by the reagents. When comparing the size of the microsphere in specimens preserved in this manner and in those examined by sections of the test, it is well to bear this cause of difference in mind. The number given as the diameter of a central chamber, in this article, is to be understood as the mean between the long and short diameters as presented for observation in the specimen. 64 THE FORAM1NIFERA and disposition of the chambers are well displayed. In the it Fio. 8. Polystonulla crtyxi. A, the megalospheric, B, the inicrospheric forms, decalcified, b, the central chambers of the latter more highly magnified. The canal system is omitted in these figures for the sake of clearness, r, retral processes ; «(, communications between the chambers. megalospheric form the retral processes characteristic of the genus are present in the chamber following the megalosphere (Figs. 8, A, THE FORAMINIFERA 7W.C, and 11, c). In the first convolution of chambers the alar prolonga- tions are hardly formed, but as the series is followed on, they project more and more at the sides, overlapping the chambers of the preceding convolutions ; and as they increase the number of apertures between successive chambers, single in the earlier chambers, also increases (Fig. 8, A, Fig. 9, st), so that in the terminal chamber there is, as we have seen, a V- shaped row of pores leading to the exterior. In the microspheric form the arrangement is similar in the later chambers, but in this form the retral« processes are absent from the chambers of the earlier convolutions (Fig. 8,J). The main trunks of the Canal System lie in the um- bilical region.1 They consist of a spiral canal, on either side of the test, running parallel with and just internal to the lateral margins of the cham- bers— whether these are pro- duced into alar prominences or not (sp.c, Figs. 9 and 11, c). Opposite the inter- vals between the chambers meridional canals (m.c) are given off, and run in the passing throug other chambers, in order to show the disposition ft. thickness of some little the septa, at Fio. 9. > Diagram of a section through a megalospheric example of Polystomella crispa. It is represented as jh the megalosphere but between the frrtm °f the canal system. M, the megalosphere; m.c, 1 meridionial ^ . r mpr> re'tral presses ; «j>.c, spiral their OUter margins, to meet canals ; st, protoplasm traversing the apertures be- tween the chambers. The dotted portion indicates one another in the median plane. From these numbers the protoplasm filling the chambers, but the canal system is represented as empty. The numerous minute pores leading direct from the chambers to of short Canals Dass Outwards tne exterior are omitted, arid the shell substance is , . . ' r , . ' left blank. and, in the case of the outer convolution of chambers, open to the exterior, into the pits 1 The canal system is said by Carpenter to be imperfectly developed in Polystomella crispa (8, p. 282), and Mdbius (25, p. 103) places P. craticulata, Fichte and Moll, in a new genus Helicoza, on the ground, that it possesses a branched canal system, absent in Polystomella. In decalcified specimens of P. crispa, however, whose protoplasm has been coloured dark by osmic acid, it is easy to convince oneself of the existence of the system. 66 THE FORAMINIFERA between the retral processes, before mentioned as visible on the surface of the test (Fig. 7). The short canals springing from the meridional canals of the inner convolutions open into the chambers of the convolution next external. The canal system is thus in communication with the chambers. In addition to the meridional canals, other branches spring from the spiral canals and pass to the surface in the umbilical region, traversing the thick mass of canalicular skeleton there deposited. The course of the spiral canals is in some cases irregular, and they often break up into a network of sinuses. In the small specimen shown in Fig. 11, c, the spiral canal of one side is seen close to its origin, and a meridional canal is shown, between the second and third chambers ; but I have been unable to trace the points of origin of the spiral canals. On treating a batch of decalcified specimens with a stain such as picrocarmine, the nuclei appear, and again the two sets of individuals come into marked contrast. The megalospheric form possesses a single large nucleus, while the microspheric form possesses a number of small nuclei, distributed through its chambers (Fig. 8). In Polystomella crispa, then, the megalospheric individuals are numerous, they have a large central chamber and a single large nucleus ; while the microspheric individuals are com- paratively scarce, they have a small central chamber, and many nuclei. The existence of the phenomenon of dimorphism being verified, the question arises : How are the two forms related ? For an answer to this question we turn to the life-history, and what is known on this head will now be given. LIFE -HISTORY OF POLYSTOMELLA CRISPA — The Microspheric Form. — The youngest specimens of this form that have been met with already contained many nuclei. Thus in one, described by Schaudinn (44, p. 92), with nine chambers, twenty-eight nuclei were present. The nuclei are at first homogeneous bodies, but as the animals grow nucleoli make their appearances. The nuclei are irregularly scattered through the protoplasm, though they are not found in the terminal chambers. They are often grouped in pairs, and there is good evidence that they multiply by simple division (20, p. 419). The nuclei in the larger chambers are larger than those in the small chambers near the centre, and they, may attain a diameter of 40-50 ^. In addition to these rounded nuclei, there are generally present in the protoplasm of the microspheric form abundant irregular strands of darkly staining substance, which are apparently given off by the nuclei. In some cases no definite nuclei are visible — THE FORAMINIFERA all the stained substance present being in the form of such irregular strands. Reproduction of the Microspheric Form. — The first indication of the approach of the reproductive phase as seen in the living animal Pio. 10. a-c, stages in the reproduction of the microspheric form of Pdystomdla crispa. Drawn from photographs of one specimen attached to the side of a glass vessel. is a great increase in the number of the pseudopodia. They are so abundant that when the specimen is attached to the side of a glass vessel and seen by transmitted light, they form a conspicuous milk-white halo about the brown shell (Fig. 10, a). The halo is at first composed of clear hyaline protoplasm, but in a 68 THE FORAMINIFERA short time the coarse brown granules, hitherto contained within the test, begin to pass out, and ultimately the whole of the proto- plasm, emerging from the test, is massed within the area covered by the halo and lies between the test and the supporting surface (Fig. 10, b). Here, after involved streaming movements, the proto- plasm gradually and simultaneously separates into spherical masses of uniform size. The centre of each is occupied by a nucleus, with an area of clear protoplasm immediately surrounding it. A close network of delicate pseudopodia surrounds the spheres and forms a communication between them (Fig 1 0, c). In a short time each Fio. 10 (continued). d, later stage in the reproduction of the specimen of Polystmnetta crispa represented in Fig. 10, o-c. sphere secretes a calcareous shell, a single small aperture being left by which the pseudopodia pass out. After lying in close contact for some hours, the spheres rapidly and simultaneously draw apart from one another, and within half an hour from the beginning of the movement they are dispersed over a wide area, and each becomes the centre of a system of pseudopodia of its own, though for some time they are not completely isolated (Fig. 10, d). The whole protoplasm of the parent is used up in the formation of the brood of young, the shell being left empty. The process, from the first appearance of the halo to the dispersal of the young-, is complete in about twelve hours. THE FORAMINIFERA 69 In a short time the protoplasm which lies outside the aperture of each of the spheres secretes the wall of a second chamber of characteristic form, and the young individual is then clearly recognisable in size and shape as the two-chambered young of the megalospheric form (Fig. 11, b). The nature of the parent which gives rise to this brood of megalospheric young is determined by decalcifying specimens which are entering on the reproductive phase, before the proto- plasm has left the central chambers. In upwards of fifty cases m.G. FIG 11 Young megalospheric individuals of Polystometta crispa. a, a group of six, two days after their formation. Four chambers are formed. 6, a specimen with two chambers, decalcified and stained ; N, the nucleus ; n, irregular stained mass, c, a specimen with nine chambers, simi- larly treated ; N and n as in 6 ; sp.c, spiral canal of the canal system ; in this specimen it becomes irregular near the last-formed chamber ; m.c, meridional canal. thus examined the centre was found to be occupied by a micro- sphere. We have then, in this process, a transition from the microspheric to the megalospheric form. Schaudinn (44, p. 94), has found great variation in the size of the megalospheres produced by one parent, namely, from 10 to 120 ft. He therefore holds that the two forms of Polystomella, though differing in their nuclear characters, are not always distinguishable by the size of the central chambers. He also finds (p. 93) that in some cases the shell of the young megalospheric form is not secreted until the spherical masses of protoplasm have wandered about for a long time. The details of the reproductive process given above are those which I have invariably observed in specimens which were kept in clean sea-water. 70 THE FORAMINIFERA In my first observations the water circulating in the tanks of the laboratory was used ; and though I repeatedly saw the protoplasm emerge from the parent shell and break up into spheres, the development did not pursue a normal course. Again and again the spheres, after remaining separate for some hours, fused with one another, and finally the mass broke up into irregular globular bodies of very unequal size, which remained alive for days, but did not, in most cases, secrete a shell. It was not until fresh sea-water was used in the jars that I had the pleasure of witnessing the normal process of development In view of this experience, and of the fact that though I have ex- amined some thousands of specimens of Polystomella, I have never met with a megalospheric form with a central chamber less than 34 fi in diameter, I am inclined to think that the great irregularities in size in the brood of young, and the small diameter of some of them, are the result of abnormal development. In some genera, however, as stated below, the two forms cannot always be distinguished by the size of the central chamber. Nuclear Changes. — When, in the reproduction of the microspheric form, the megalospheres first become isolated, the centre of each is occupied by a sharply -defined, round nucleus, about 7-8 ft in diameter, staining uniformly pink in picrocarmine. At this stage there is also diffused in the protoplasm a material taking a stain in this reagent ; but in a short time the stained material, pre- viously diffused, becomes aggregated in defined but irregular masses, which gradually draw together, and for a time frequently hide the nucleus from view. When two or three chambers of the young test have been formed the nucleus is again distinctly visible (Fig. 11, b, N), together with one or more irregular masses (n) formed by the closer aggregation of the previously diffused material. In many cases, though not in all, a mass apparently identical with the latter remains visible in or near the central chamber in specimens of the megalospheric form in advanced stages of growth; but it is, I believe, the round nucleus which was seen in the megalosphere at its first formation which becomes the nucleus (" principal kern ") of the megalospheric form. The relation between this nucleus and the nuclei and irregular strands of the microspheric parent has not been followed.1 Growth and Reproduction of the Megalospheric Form. — As the individuals of this form grow, and the number of their chambers is augmented, the nucleus likewise increases in size, and, leaving the megalosphere, it moves on through the chambers, becoming temporarily constricted as it passes through the narrow passages connecting them. In specimens fairly advanced in growth the 1 I have not met with any evidence in support of Schaudinn's statement (44, p. 95) that the large nucleus ("principal kern") of the megalospheric form results from the massing together of the irregular strands of the microspheric parent. THE FORAMINIFERA 71 nucleus is, as Schulze pointed out (64, a), usually found in or near the chamber which is numerically in the middle of the series. In the cells of growing vegetable tissues the nucleus moves towards that part of the cell at which growth is most actively proceeding.1 Thus in a growing root-hair the nucleus is found near the tip ; in a young stellate hair, it lies at the point of junction of the rays. Its movement towards the regions of activity results, we must suppose, from a certain force impelling it in that direction, and its position when at rest is at the point where the impelling forces, resulting from the activities of the protoplasm in different parts of the cell, are in equilibrium. In the forms of Foraminifera in which the nucleus is single it appears that its position is likewise dependent on the disposition of the protoplasm. Thus in the megalospheric forms of Cycloclypeus and Orbi- tolites complanata the nucleus is found in or near the central chamber, where, owing to the cyclical growth in these genera, the attractions, due to activity at the periphery, are in equilibrium. In the forms with spiral arrangement the nucleus moves on through the series of chambers as growth proceeds. In Polystomella, however, the nucleus of the megalospheric form always lags some distance behind the point at which, judging from the disposition of the bulk of the protoplasm, we should expect the attractive forces to be in equilibrium. At the earliest stage at which it has been recognised the nucleus appears to be a homogeneous body. As it grows, round nucleoli make their appearance, and these are comparatively large in young specimens, and decrease in size while they increase in numbers as growth proceeds. Sections through the nucleus of specimens fairly advanced in growth show a well-defined nuclear wall, a reticulum with finer or coarser meshes occupying the interior, and rounded nucleoli at the nodal points of the reticulum. Minute granules may often be detected in the strands of the reticulum. It appears that throughout the vegetative phase of the megalo- spheric form small fragments are separated off from the nucleus, and they may often be seen as irregular bodies, sometimes con- taining nucleoli, lying in the neighbourhood of the nucleus, or in the chambers through which it has passed in its passage onwards from the megalosphere. Towards the end of the vegetative phase of the life-history of the megalospheric form the nucleus loses its regular outline and its power of receiving stains, and finally disappears. In such specimens it may often be observed that additional passages have been opened up, by the dissolving action of the protoplasm on the shell substance, between adjoining chambers of inner and outer convolutions, so that the inner chambers are placed in 1 Cp. Haberlandt's researches, quoted by 0. Hertwig, Die Celle und die Gewebe i. p. 259. THE FORAMINIFERA more direct communication with the outer, and so with the exterior. Reproductive Phase of the Megalospheric Form. — In specimens in which the large nucleus has disappeared hosts of minute nuclei may be found, scattered uniformly or in groups through the protoplasm. Presumably these are derived from constituents of the principal nucleus, and possibly of the small nuclear fragments as well, but their precise relation to these bodies has not been followed. When the nuclei are evenly distributed the protoplasm breaks up into minute rounded masses, the centre of each being occupied by a nucleus. Division of these nuclei by karyokinesis now occurs, and is simultaneous or nearly so throughout the organism, and this is followed by a second division of the protoplasm to form rounded bodies about 3-4 /x in diameter, each containing one of the daughter nuclei (Fig. 12). At a later period the contents of the shell issue as active zoospores of approximately uniform size, which swim rapidly by means of flagella. It is to be noted that, as in the case of the reproduction of the microspheric form, the whole of the protoplasm of the parent is used up in the production of the zoospores. Though the zoospores have been seen issuing from the shell, their precise characters, when ripe, have not been accur- ately described ; nor have we as yet direct evidence as to their fate. That there is a close relation between the zoospore and the microsphere is suggested by the fact that there is no great differ- ence in size between these structures, the diameter of the former, before their escape from the parent shell, being 3-4 /A, and that of the latter about 10 /A. A further piece of evidence tending in this direction is furnished by an observation of Schaudinn's (44, p. 92). In an aquarium containing abundant Polystomellas, Schaudinn suspended coverslips by means of threads, so arranged that the lower borders of the coverslips were separated by about 2 cm. from the stratum covering the bottom of the aquarium. After two days young examples of the microspheric form were found on the coverslips. Now throughout the vegetative phases Fio. 12. Section through a specimen of the niegalo- spheric form of Polystomella crispa, in which the contents are divided up into zoospores. x 160. THE FORAMINIFERA 73 of its life Polystomella crawls over the surface of submerged objects, by means of its pseudopodia, but is incapable of swimming freely ; and as the depth of 2 cm. far exceeds the distance to which the pseudopodia of so young a specimen are likely to extend, the colonisation of the coverslips by the microspheric form under these conditions points to the existence of a free swimming stage prior to the vegetative stage in which the young forms were found. Such a stage is supplied by the free swimming zoospore. The results furnished by direct observations on the life-history of Polystomella may be summarised in two statements : — The microspheric form terminates its existence by becoming trans- formed into a brood of megalospheric young. The megalospheric form terminates its existence by becoming trans- formed into minute zoospores of uniform size. Before discussing the bearing of these results on the relation- ship of the microspheric and megalospheric forms, it will be con- venient to consider some facts in another life-history, that of Orbitolites complanata. LIFE -HISTORY OF ORBITOLITES COMPLANATA. — The main features of the structure of this species are described below (p. 104 and Figs. 36 and 37). For the present purpose it will be sufficient to point out that the mode of growth is, except in the early stages, cyclical, concentric rings of small chambers being added at the margin of the disc-shaped test, and that the tests are biconcave. In the microspheric form the central region is thin, being built up of the microsphere and the small chambers or " chamberlets " which succeed it. In the megalospheric form the centre of the shell is thicker owing to the presence of the " primitive disc " (Figs. 37, A, and 38). This consists of the megalosphere, a spiral passage, and of the very large crescentic chamber, which nearly surrounds the other constituents of the primitive disc. The outer margin of the crescentic chamber is perforated by pores, by which it opens into the innermost ring of chamberlets. Reproduction. — The production of megalospheric young by a microspheric parent has been repeatedly observed (4 and 20, and Fig. 36, b). In the later stages of the growth of the parent spacious brood chambers are formed at the periphery of the disc, and in the reproductive phase the protoplasm is withdrawn from the small chambers internal to them and massed in the brood chambers, where it breaks up to form the brood of young. On their escape, which is effected by the breaking down of the outer walls of the brood chambers, presumably under the dissolving action of the protoplasm of the young, the test of each young individual has already the structure of the " primitive disc " 74 THE FORAMINIFERA which is found at the centre of the megalospheric form (Figs. 37, A, and 38). This process is clearly comparable with the formation of the megalospheric young by a microspheric parent which we have followed in Polystomella, the chief difference consisting in the fact that the young are formed in peripheral brood chambers and not outside the test of the parent. By analogy with Polystomella we cannot doubt that megalo- spheric parents give rise to zoospores. Specimens of Orbitolites have, however, been found (and the corresponding phenomenon has also been observed in several other genera) in which a brood of megalospheric young occupies the peripheral chambers of a test, the centre of which is formed by a primitive disc, and which is tlerefore megalospheric (20, p. 435). Hence we must conclu-'s that in these cases the megalospheric form may be repeatec possibly more than once, before a brood of zoospores is produced. The behaviour of the nucleus under these conditions has ^ •© not been closely followed. $ Q _ Fig. 13 illustrates a similar repe- tition of the megalospheric form in •^ the case of Cornuspira inwlvens.1 ® AvSl') ] ^e ^a^on Between the Microspheric >N^-X/ find Megalospheric Forms. ® With the evidence furnished by FIG. is. the life -histories of Polystomella and Cornuspira invoivens, Reuss. The Orbitolites we may now return to the contents of a megalosphenc form have J emerged from the shell and divided up question OI the relationship 01 the into a number of young, which are also i i j v • p megalospheric. in both parent and megalospheric and microspheric forms. young the megalosphere was about rrup porliftsr VI'PW hplH nn this 80 M in diameter, x 30. Cp. Fig. 20. ' V16W subject, that they represented two species of a genus, is at once disposed of by the fact that the megalospheric form is the offspring of the microspheric. The next suggestion was that the two forms represent the two sexes. To this, two objections may be made on general grounds. (1) The difference between them is most marked at the beginning of their growth, when they consist only of the central chambers, the microsphere or the megalosphere, while males and females 1 In his preliminary paper Schaudinn states (44, p. 96) that the megalospheric generation may also be repeated (though rarely) in Polystomella, and that in such an event no principal nucleus is formed by the megalospheric parent. As Rhumbler remarks in his notice of Schaudinn's paper (ZooL Centralblatt. Jahrg. 2 (1895), p. 453, footnote), it is difficult to see how this result can be reached, for the shell of Polystomella is too thick to allow the miclear condition to be observed during life. THE FORAMINIFERA 75 arise from eggs which are, so far as observation has yet advanced, similar, and are most differentiated when adult. (2) Although the differentiation of male and female gametes is known in several groups of the Protozoa (Sporozoa, Ciliata, Flagellata), a differentia- tion of the parent organisms which produce the gametes, which would be the phenomenon comparable with the sexual dimorphism of the Metazoa, is unknown among the Protozoa. But apart from a priori objections, it may be well to try how the ascertained facts of the life-history fit this hypothesis. The megalospheric form, producing zoospores, would evidently on this view represent the male, while the microspheric form, producing the large megalospheres, might be regarded as the female. It might be supposed, the fate of the zoospores being unknown, that they unite with and fertilise the megalospheres. But while the origin of the megalospheric form, the supposed male, is thus accounted for, that of the microspheric form, the supposed female, remains unexplained, and as the whole of the protoplasm of the parent is, to all appearance, used up in the brood of megalospheric young, the hypothesis is at fault. Moreover, in OrUtolites, Cornuspira, and other genera the megalospheric form may also give rise to a brood of megalospheric young, a proceeding foreign to the nature of a male organism. Hence neither of the forms of the species conforms to the character -assigned to it by the hypothesis : the microspheric form, the supposed female, in that it does not produce " females " ; the megalospheric form, the supposed male, in that it does in some instances produce " males." A third hypothesis is in harmony with what we know in other groups of the Protozoa, and fits the ascertained facts. It is that the two forms represent alternating or recurring generations in a life-cycle. The individuals of the microspheric form reproduce asexually by the multiple fission of their protoplasm to form broods of megalospheric young. The individuals of the megalospheric form may undergo, in some genera, a similar process, but ulti- mately a megalospheric individual is produced whose protoplasm divides into zoospores. How is the gap between the zoospore and the microsphere filled ? That they are closely related is suggested, as stated above, by their approximation in size, and by the indication, afforded by the colonisation of Schaudinn's coverslips, of a free swimming stage preceding the vegetative phase of the microspheric form. Another remarkable fact bearing on the matter is the scarcity of the micro- spheric form in comparison with the megalospheric, a scarcity all the more striking when it is borne in mind how far more numerous are the zoospores produced by one megalospheric individual than the members of a brood of megalospheres. 76 THE FORAMINIFERA The analogy of other life -histories would lead us to suppose that at some point in the cycle a sexual process, the conjugation, with nuclear fusion, of two organisms, occurs, and the life-history of Trichosphaerium sieboldi, Schn., which has lately been worked out by Schaudinn (47), to whom so much of the recent advance in our knowledge of the life -history of the Protozoa is due, appears to afford a very appropriate parallel. This form is not included in the Foraminifera, but is a somewhat aberrant member of the allied group — the Lobosa. The main features of its life -history are, however, remarkably similar to those of Polystomella. The individuals are rounded multinucleated masses of proto- plasm not contained in a definite shell, though surrounded by a gelatinous envelope. They form a dimorphic series, the members of which recur in a cycle of generations. In those of one genera- tion, which may be called by Haeckel's term Amphionts, reproduc- tion occurs by the simultaneous division of the protoplasm about the nuclei to form spherical uninucleated masses, which emerge and grow into the members of the other generation — the Mononts. These in their turn break up, after subdivision of their nuclei, into zoospores. The zoospores are biflagellate organisms, and are all alike. While the zoospores from the same parent will not unite with one another, those from different parents conjugate readily. In this process, which has been carefully followed by Schaudinn, the nuclei of the two gametes unite, their flagella drop off, and the zygote so produced, absorbing fluid, undergoes a considerable in- crease in size, so that in a few hours its diameter is more than doubled. The zygote shortly afterwards secretes a gelatinous envelope, and the characters of the full-grown individual of the amphiont generation are gradually acquired. The multinucleate condition results from successive mitotic divisions, beginning with that of the nucleus of the zygote. In Hyalopus dujardinii, which may be regarded as a member, though an aberrant one, of the Foraminifera, Schaudinn (43) has also observed the conjugation of zoospores, but in this case the •process occurred between members of the same brood. If we assume that a similar conjugation of zoospores occurs in Polystomella, the facts above alluded to are at once explained. The fusion of two zoospores (4 p. in diameter), and the subse- quent expansion of the zygote by the absorption of water before the secretion of a shell, might well form a body of the size of the microsphere (about 1 0 /*) ; the free locomotive stage prior to the settling down of the microsphere, indicated by Schaudinn's experiment, is supplied ; and the comparative rarity of the micro- spheric form is explained on the supposition that, as in Tricho- THE FORAMINIFERA 77 sphaerium, the meeting and conjugation of zoospores from different parents is necessary for the production of a microsphere. It is very desirable that the conclusion should be confirmed by direct observation, but, meanwhile, it seems not premature to admit that it is probable that the microsphere arises from the conjugation of zoospores. We may conclude, then, that the microspheric aad megalo- spheric forms of the Foraminifera represent alternating or re- curring generations in a life -cycle. While the megalospheric generation arises asexually, either from a microspheric or a megalo- spheric parent, it is probable that the microspheric generation arises sexually — i.e. by the conjugation of two similar zoospores. Representatives of the two generations have been recognised in species belonging to the following genera of Foraminifera : — ORDER l. Gromiidea (7).1 ( „ 2. Astrorhizidea.2) ( „ 3. Lituolidea.) „ 4. Miliolidea. Family MILIOLINIDAE. — Cornuspira (24, 1898, p. 612); Spiro- loculina (57, p. 201); Biloculina (27, p. 863); Sigmoilina (Planispirina, pars), Schl. (53, p. 106) ; Triloculina and Quinqueloculina (27, pp. 863 and 1598); Massilina (57, p. 218); Adelosina (52, p. 91); Idalina and Periloculina (28) ; Lacazina (27). Family HAUERINIDAE. — Planispirina, Seg. (56, p. 194). Family PENEROPLIDIDAE. — Peneroplis and Orbiculina (this article), Orbitolites (4, p. 693, and this article) ; Meandropsina (59). Family ALVEOLINIDAE. — Alveolina (51, p. 526). ORDER 5. Textularidea. Family TEXTULARINAE. — Textularia and Spiroplecta (this article). (ORDER 6. Chilostomellidea.) „ 7. Lagenidae. Family NODOSARIIDAE. — Nodosaria and Dentalina (51, p. 626); Frondicularia (15, p. 480) ; Cristellaria (?) (5, p. 45). Family POLTMORPHINIDAE. — Siphogenerina (Sagrina) (50, p. 21). (ORDER 8. Globigerinidea.) „ 9. Rotalidea. Family ROTALIDAE. — Rotalia (51, p. 526, and 20, p. 436); Truncatulina and Calcarina (20, pp. 436 and 437). Family TINOPORIDAE. — Polytrema (23). 1 The cases of Hyalopus and Mikrogromia are considered below. . 3 In the orders included in brackets I have not succeeded in finding a record of the existence of dimorphism. 78 THE FORAM1NIFERA ORDER 10. Nummulitidea. Family FUSULTOIDAE. — Fusulina (58, p. 1). Family POLYSTOMELLIDAE. — Polystomella (20, p. 415). Family NUMMULITIDAE. — Operculina (this article); Amphistegina (51, p. 526) ; Nummulites and Assilina (26, p. 300) ; Heterostegina (this article); Cycloclypeus (10, p. 21, and this article); Orbitoides (31, p. 258, and 61, p. 463) ; Miogypsina (60, p. 328). Thus the phenomenon of Dimorphism, the occurrence of mem- bers of a species under two forms — the megalospheric and micro- spheric — is widely spread among the orders of the Foraminifera, and where it is found it affords clear evidence of alternating or recurring generations in the life-history of the species exhibiting it. REVIEW OF THE STRUCTURE AND LIFE-HISTORY DISPLAYED IN THE ORDERS OF THE FORAMINIFERA. As the attention of the many workers who are occupied with this group is turned to the subject, the list of dimorphic forms will, no doubt, be greatly extended ; but there are indications in the descriptions already published of phases in the life-history of some forms, especially in that borderland occupied by the simpler Foraminifera, which depart in a greater or less degree from those described in Polystomella and Orbitolites ; and we may now take a survey of the features of life -history which have been described in the different groups, and of the more interesting modifications in the form of the test, which, as we have seen, is found to be more or less dependent on the phase of the life -history of the organism which secretes it. In chambered tests, in which the walls of the first formed chamber remain unaltered throughout growth, evidence of the mode of origin of the individual, whether as a megalosphere or a microsphere, is furnished by the structure of the test. But in the great majority of the Gromiidea and Astrorhizidea, the tests expand to accommodate the increase by growth (cp. p. 54), and all in- dications of the size of the test when it was first secreted are obliterated. Hence we are deprived in them of part of the evidence on the course of the life-history which we have in other groups, and we must rely on the characters furnished by the soft parts of preserved specimens, or on direct observation of the living animals. From the evidence which we have, however, it is not clear that the course of the life-history of some members of these orders is the same as that of the dimorphic Foraminifera above described. THE FORAMINIFERA 79 ORDER Gromiidea. In Euglypha (Fig. 3) multiplication, by division into two, occurs as follows.1 The specimen which is about to divide secretes fresh shell plates, which are at first dispersed in the protoplasm about the nucleus. The pseudopodia are with- drawn, and the protoplasm is extruded beyond the mouth of the test in a rounded mass. This grows until it assumes a size equal to that of the test from which it protrudes, and the newly-formed plates are disposed on the surface to form a new test. The nucleus divides by karyokinesis, half going to each end of the mass, and division of the protoplasm follows, one part remaining in the old shell and the other in the new one. FIG. 14. Colony of Mikrogromia sodalis in the diffused condition, a, an individual in process of multiplication by transverse fission, c.v, contractile vacuole. Two of the members of the colony are seen to be undergoing the same process. Blochmann (2) has described a process in which, after the division of the nucleus, the protoplasm was withdrawn from the newly formed shell, and this, together with the daughter nucleus remaining in it, was cast off. It is suggested that this may be comparable with the extrusion of parts of the nucleus observed in some other Protozoa and in polar-body formation. But in view of the fact that the new shell was cast off, as well as the daughter nucleus, this interpretation appears, to say the least, forced. A temporary fusion of the protoplasm of two or more indi- viduals, apparently without fusion of nuclei (plastogamy), was observed by Blochmann, and in one instance a new individual was apparently formed by the conjugation of two. In this case it 1 The process was first described by Gruber (16), and followed out in detail by Schewiakoff (48). 8o THE FORAMINIFERA was supposed that the nuclei had united (karyogamy) as the new individual was uninucleate. Encystment also occurs in Euglypha, but what the subsequent stage may be is unknown. Mikrogromia socialis, first described by Archer,1 and afterwards more fully by R. Hertwig (18), is a fresh-water form, occurring in colonies, the members of which are united by their pseudopodia. The colonies are sometimes globular and compact (Cystophrys sta^c;, sometimes diffused (Fig. 14), and in the latter condition present an interesting resemblance to a brood of young megalo- spheric individuals of Polystomella in a stage of dispersal (Fig. 10, d). The growth of the colony results from the partial longi- tudinal fission of the members into two (or three), one (or two) of the products of fission escaping, secreting a new test, and taking its place in the colony. Hertwig also observed the production of young individuals, arising by transverse fission (Fig. 1 4, a). Of the two bodies so formed one remains in the test, continuing the vegetative phase of the parent, the other becomes free, and, in some cases, swims away as a biflagellate organism. In other cases, however, the flagella were not observed, being replaced by pseudo- podia, resembling those of Actino- phrys. The further history of the young thus produced was not followed. Assuming this to be a normal phase of development of Mikrogromia, it appears to be with- out a parallel in the life-history of Polystomella. Hyalopus dujardinii, Schaudinn ( = Gromia dujardinii, M. Schultze) is a marine form distinguished by the hyaline and nongranular character of its pseudopodia, and by the absence of anastomoses between their branches. The main body of the protoplasm is covered by a chitinous envelope, and contains large brown rounded granules and many nuclei. In the condition in which it 1 Provisionally as two species, Cystophrys haeckdiana and Oromia tocialis Archer (1). Fio. 15. Hyalopus (Gromia) dujardinii. (After M. Schultze.) X 40. THE FORAMINIFERA 81 FIG. 16. 5, Shepheardella taeniformis, Siddall, x about 15. (After Siddall, Q.J.M.S vol. xx.) The nucleus is seen nearly opposite 5. 11, Amphitrema wrightianum, Archer, x^about 210. (After Archer, Q,J.M.S., N.8. vol. ix. 1869.) 82 THE FORAMIN1FERA was described by M. Schultze (64, p. 55), the shape is oval, and there is a single orifice (Fig. 15), but Schaudinn finds (43) that when living amongst the stems of algae, it loses its oval shape and assumes a branching form, new mouths being developed at the ends of the branches. Such branched forms may attain a length of 5 mm. Two modes of reproduction were observed. One is by a process of fission, the body slowly dividing into two or three parts, sometimes of unequal sizes ; the other is by the formation of zoospores. In the latter process the pseud opodia are retracted and the whole protoplasm divides up into oval or pear-shaped bodies, 5-8 /i in diameter, containing a nucleus 3-6 //. in diameter, a vacuole, and a conspicuous granule. They swim by means of a single tiagellum 30-38 /x in length. The zoospores conjugate in pairs, but in this case the conju- gation is, according to Schaudinn, between members of the same brood. The further history of the zygote could not be followed. The formation of the zoospores of Hyalopiis is evidently com- parable on the one hand with the reproduction of Trichosphaerium which gives rise to the " amphiont " generation, and on the other hand with the reproduction of the megalospheric form of Polysto- mella. A similar mode of reproduction to the slow process of fission of Hyalopus has been seen in Lieberkiihnia and Lecythium, the division of the protoplasm involving that of the envelope. Whether this is to be compared with the production of the brood of melagospheres by the multiple fission of the microspheric parent, or to the similar slow fission which occurs in Trichosphaerium in addition to the multiple fission of the "amphiont" parent, it appears to be at present impossible to decide. Many of the Gromiidea have a single orifice to the test, as in Gromia and Eugbjpha (Figs. 1 and 3). Shepheardella and Amphitrema have two orifices, situated at either end of a median axis (Fig. 16, 5 and 11). ORDER Astrorhizidea. In Saccammina, and some other members of the Astrorhizidea in which growth is accompanied (as explained above, p. 54) by expansion of the test, no evidence on the phase of life-history represented, is furnished by its structure. But in other genera, such, for example, as Hi/perammina, in which the tests grow not by expansion, but by addition, a large globular chamber is some- times found at the commencement (Figs. 17, rf, and 18). Such forms may well represent a megalospheric generation. There is, however, no evidence at present of the microspheric forms corresponding to them. Rhumbler has made a careful investigation (33) of the nuclear THE FORAM1NIFERA characters of Saccammina sphaerica (Fig. 17, b). He found that among the 286 specimens which he examined, a single nucleus was present in all but one (which had two nuclei, and was regarded as abnormal), and the phases presented by the nuclei fell into a con- tinuous series. They corre- spond with those of the nucleus of the megalospheric form of Polystomella. The nucleus in- creases in size with the growth of the organism, and the nucleoli ("binnen korper"), at first large and few, increase in number and diminish in size. Finally (PI. 23, Fig. 67) the nuclear membrane breaks, and linin threads containing chromatin grains are dispersed in the protoplasm. From these Fio. a, Astrorhiza limicola, Sandahl., x 6. I, Sacaimmiiia sphaerioa, M. Sar*, x!2. c, Piluli)ia jefreysii, Carpr., x 12. d, Hyperammina subnodota, Br., x 7. In a, b, and d the test lias been laid open. (From Brady, "Challenger" Report.) later nuclear phases it appeared that some process of reproduction was imminent, but none was observed. The formation of zoospores by the Foraminifera was at that time unrecognised, and Rhumbler THE FORAMINIFERA was surprised at finding no indication, notwithstanding the abundant material at his command, of the formation of a brood of young resembling the parent. On the analogy of the life-history of Polystomella, the absence of such indications appears in no way remarkable, for such a nuclear history is associated, as we have seen, with the production of zoospores. The only difficulty in applying this analogy arises from the fact that no indications were found of a form of Saccammina with a different nuclear history, corresponding with that of the microspheric generation of Polystomdla. It is, of course, possible that the microspheric form, although occurring in nature, did not happen to be repre- sented among the specimens examined; but however this may be, it is clear that we are not at liberty to assume the existence of a microspheric form in Saccammina. Hence, in the absence of other evi- dence bearing on the point, the Astrorhizidea cannot at present be admitted into the list of dimorphic Foramini- fera. In Haliphysema tumano- wiczii (Fig. 19) Lankester (19) described numbers of "egg- like " bodies, varying in diam- eter from T-gVfr to -5^5- inch, scattered through the proto- Fio. 18. H yperammina arborescent, Norm, o, two speci- mens growing attached to a stone, x 20 ; b, initial chamber of another specimen. (After Brady.) plasm. They appeared to be nucleated, and, in some cases, in process of division. It was surmised that they might be concerned in reproduction. Further information on the nature of these bodies would be very accept- able, but the possibility appears not to fcave been excluded that they are symbiotic or parasitic organisms similar to those which abound in the protoplasm of Orbitolites complanata. THE FORAMINIFERA ORDER Lituolidea. This order consists of arenaceous forms which are "isomorphic" with genera belonging to several of the other orders; and by many authors the order is broken up, and its genera associated 10 Fio. 19. Haliphysema tumanvwlczii. 10, part of the protoplasm stained to show the nuclei, » ; 11, living speci- men with expanded pseudopodia. (From Lankester, Art. Protozoa, Encycl. Britannica, Fig. x.) with the calcareous forms which they resemble. In some cases (e.g. Cornuspira, Nodosaria, Eotalia) the latter are, as will appear below, dimorphic, so that we should expect their " isomorphs " to be so likewise ; but though this is very probably the case, I am aware of no direct evidence on the matter. A process of reproduction is recorded by Schaudinn (42) in Ammodiscus gordialis, P. and J. The protoplasm divides within the parent test into some 50-80 young, which become invested 86 THE FORAMINIFERA with a chitinous envelope, together with siliceous particles pre- viously taken into the protoplasm. ORDER Miliolidea. On coming to the Miliolidea we have a large body of evidence on dimorphism, thanks in great measure to the careful investiga- tions of Schlumberger, by whom, either alone or in conjunction with Munier-Chalmas, the foundations of our knowledge on the dimorphism of the tests of Foraminifera have been laid. The tests will first be described, the nuclear characters and such details of the life-history as are to hand being given at the end. Family Miliolinidae. — Before considering the phenomena of dimorphism in this family, it is necessary to describe the character- istic structure of the test in certain forms. FlO. 20. Comuspira involvens, Reuss. a, the megalospheric form, x 90. b, the microspheric form, x 60. (From Brady, Parker, and Jones, Tran* Zool. Soc. vol. xii. PI. 40, Figs. 1 and 2.) FIG. 21. Spiroloculina limtxita, d'Orb., x 80. (After Brady.) The simplest type is met with in Cornuspira. The whole of the test except the central chamber (which pre- sents a well-marked difference in size in the two forms, Fig. 20) consists of a continuous tube, gradually increasing in diameter as it is followed away from the centre, but without any constrictions dividing it into separate chambers. In both forms it is disposed in a closely-wound spiral lying in one plane, so that a section in this plane would divide the test symmetrically. In the genus Spiroloculina the arrangement is somewhat similar, but here the tube is divided into distinct chambers, each of which ends in a contracted mouth with an everted lip. The chambers increase successively in length, and are so disposed that each occupies half a turn of the spiral. It results from this arrange- ment that the mouths of the chambers are directed alternately in opposite directions, and each chamber is applied to that which is next but one before it in the series. A straight line, which passes THE FORAMINIFERA through the central chamber and the mouths of all the chambers which succeed it, has been called the axis of construction. The spiral formed by the series of chambers is not quite regular, as is the case in Cornuspira ; for while each chamber is gently curved, there is a sharp bend where one chamber communicates with another. Hence the test is elongated in the axis of construction. In Spiroloculina the chambers are disposed in one plane, and the width of each is only slightly greater than that of its predecessor, so that all the chambers are exposed on the two flat faces of the test (Fig. 21). In all but the earlier chambers of the microspheric forms the arrangement characteristic of the genus Biloculina is essentially similar, but there is a marked difference in the shape and appear"- ance of the test owing to the great width of the chambers. Each Fio. 22. a, Biloculina dcpressa, rt'Orb., x 40. b, Triloeulina tricarinata, d'Orb., x 50. (After Brady, 3.) is so wide that its margins are in contact with those of its prede- cessor, and overlap them at the sides (Figs. 22, a, and 24). It results from this arrangement that the two last chambers enclose those previously formed, and they alone appear in the contour of the test. As in the preceding genus a median longitudinal section through the last chamber divides the whole series of chambers into symmetrical halves. As will appear later, the microspheric form of Biloculina departs considerably from this arrangement. In Triloculina and Quinqueloculina1 the chambers are likewise disposed about an axis of construction, and their mouths open alternately in opposite directions, but the median plane of any chamber is not that of its predecessor, but directed at a definite angle to it. It is as though in a Biloculina test, while the plane in which the new chambers are formed remains constant, the 1 These genera are now usually included in the genus MUiolina, though Schlum- berger is inclined to retain the old generic distinctions. 88 THE FORAMINIFERA Fio. 23. Quinqutlocnlina $eminulum, Linn, a-c, views of teat ; a, b, from the Bides ; c, from apertural «nd. A, section of megalospheric, B, of microspheric form. A and B from Schlumberger (57). It will be noticed that A and B represent a less flattened form of test than that seen in a-c. THE FORAMINIFERA 89 part of the test already formed were to rotate on its axis of con- struction through a definite angle in the interval between the formation of one chamber and the addition of its successor. In the genus Triloculina the rotation is through (approximately) one- third of the circumference, in Quinqueloculina through two-fifths, and the chambers are disposed in three and five radii respectively. In these genera the width of the chambers is, moreover, less than in Biloculina ; and in Triloculina three, in Quinqueloculina five chambers are exposed, at any given stage of growth on the contour of the test. We may now turn to the phenomena of dimorphism as pre- sented by members of this family. In representatives of all the genera included in the list on p. 77, a well-marked difference has been shown to exist in the size of the central chamber in the two forms of the species. Thus in Biloculina depressa the diameter of the megalosphere (M) x has been found to vary from 200 to 400 /*, and that of the micro- sphere (m)1 from 18 to 25 p (Fig. 24). In B. ringens the contrast is not so great (M = 54 /*, m = 20 /A). In Triloculina M = 204 & m= 18 p. In Sigmoilina M- 96-150 ft, m- 27-3 6 i*. In Adelosina M= 90-330 /*, m= 18 p. In Idalina M- 180-440 fly m = 12 p. In Massilina a well-marked difference is said to be present, though the actual dimensions are not recorded. Turning to the plan of growth, in Cornuspira and Quinquelocu- lina the tests are uniform, i.e. they are arranged on the same plan throughout, from the part which immediately succeeds the central chamber to the end of the test ; and this is the case in both forms of the species. Massilina has biformed tests in both megalospheric and microspheric forms, the earlier chambers being arranged on the quinqueloctiline plan, and the later on the spiro- loculine. But in several of the other genera a marked contrast is found in the arrangement of the chambers in the megalospheric and microspheric forms. In the species of these genera the tests of the megalospheric forms are, for the most part, uniform, the arrangement characteristic of the genus being followed throughout the growth of the test, while the tests of the corresponding microspheric individuals are bi- or tri-formed, the plan of growth of the chambers changing once or twice before the test is complete. Thus in many species of Biloculina the arrangement of the megalospheric form is biloculine throughout (Fig. 24, a). In the 1 It will be convenient to use the letters M and m to indicate the diameters of the megalosphere and microsphere respectively, the "diameter" being taken to imply, when the central chamber is not spherical, the mean between the long and short diameters. THE FORAMINIFERA microspheric form (Fig. 24, b) the chambers succeeding the micro- sphere are arranged on the quinqueloculine plan, and this arrange- ment is maintained during the addition of a smaller or larger number of chambers, according to the species. At a certain stage the chambers become wider, and conform to the triloculine plan. Finally, the biloculine arrangement is assumed and maintained during the remainder of growth. In like manner the megalospheric forms of Triloculina are built Fio. 24. Biloculina deprcssa, d'Orb. Transverse sections, a, of the megalospheric form, x 50. b, of the microspheric form, x 90. (Two external chambers have been omitted In b.) (After Schlumberger, 66.) I am indebted to the Cambridge Philosophical Society for permission to use the block from which these figures are prepared. up on the triloculine plan throughout, while the microspheric forms begin life on the quinqueloculine plan, though they conform early to the triloculine (Fig. 25). In some species it appears that a difference in the arrangement of the chambers is maintained throughout the growth of the test Thus in Biloculina lucernula, Schwager, the microspheric form, commencing on the quinqueloculine plan, becomes triloculine, but appears never, according to Schlumberger, to attain to the biloculine arrangement, which the megalo- spheric form follows throughout. The two forms are shown to belong to the same species by similarities in the shape of the chambers, and also THE FORAMINIFERA 91 by the presence of a thin sandy layer on the surface of the test, which none of the other species inhabiting the same locality possess (55). A difference in the arrangement of the chambers throughout growth is also said to be found in Adelosina polygonia, Schlumb., but it is in part of a different character. In this species the chambers are all arranged in Fio. 25. Sections of the test of Triloculina schreiberiana, d'Orb, x 66. a, the megalospheric form, b, the microspheric form. (After Schlumberger, 57.) a single plane in the megalospheric form, and after a preliminary quinqueloculine phase they are arranged in a single plane in the micro- spheric form. So far the arrangement agrees with that of Biloculina. But the chambers of Adelosina polygonia do not occupy a half turn of the Fio. 26. Adelosina polygonia. a, the megalospheric form ; b, the microspheric form, x about on (After Schlumberger, 54.) spiral, as do the chambers in that genus, but only one-third or a quarter of a turn. In the microspheric form each chamber occupies one quarter of a turn of the spiral, and the tests are, in consequence, quadrangular in outline. In the megalospheric form the chambers occupy each one-third of a turn, and the resulting tests are triangular (Fig. 26). In about 1 per cent of the individuals of this form, however (310 were examined), the quadrangular arrangement was adopted in the last whorl (54). 92 THE FORAM1NIFERA In the Miliolinidae, then, the tests of the megalospheric and microspheric forms of a species differ in the size of the central chambers, and also in some cases in the arrangement of the chambers which immediately succeed it. When this difference is found it depends on the fact that while the megalospheric form generally grows on a uniform plan throughout, the microspheric form 'assumes for a longer or shorter period of its growth an arrangement different from that to which it subsequently con- forms, and one which is, in many cases, characteristic of another genus of the group.1 The term Initial Polymorphism was first applied by Munier- Chalmas and Schlumberger to a varying condition with respect to the arrangement of the initial chambers observed among different individuals of the megalospheric form of the fossil Idalina ardiqua. It occurs also, as we shall see, in other genera. Fio. 27. The central regions of transverse sections of three examples of the megalospheric form of Idalina antiqua (d'Orb.). Diameter of megalosphore in a =440, b=400, and c=240 ft. After Schlumberger. (It will be observed that the magnification of a is greater than that of b and c.) In this species the megalospheric and microspheric forms are sharply contrasted in the size of their central chambers (M = 180 - 440 /A, m =? 12 /*), and the arrangement of the chambers is, in the main, that characteristic of Biloculina. The microspheric form passes through quinqueloculine and triloculine stages to a biloculine condition, which, however, is converted in this genus in the later stages of growth to a uniloculine state, by the lateral extension of each of the chambers in turn, to embrace the whole of the previously formed test. The megalospheric form begins in many cases (Fig. 27, a), on the biloculine plan, to become unilocular at a later stage, like the microspheric form. In some cases, however, the initial chambers of this form are arranged on the triloculine plan (6), and in others again on the quinqueloculine (c), though the biloculine arrangement is soon assumed, in the latter case with a brief intermediate triloculine phase. Moreover, 1 I am not, however, aware of any definite form which the microspheric tests of Adelosina polygonia resemble. THE FORAMINIFERA 93 and it is important to notice this point, the forms with the largest megalospheres are those which assume the biloculine condition directly ; those with the smaller megalospheres repeating the triloculine or the quinqueloculine arrangement. Family Hauerinidae. — The biformed genera brought together in this family are said to be characterised by the cornuspira-like or milioline (tri- or quinque-loculine) arrangement of. the chambers in the early stages of growth. In some cases, however, as in Articulina conico-articulata, Batsch, two forms of a species occur, one beginning in a large globular chamber with a short spiral passage leading to the later chambers, which are disposed in rectilinear series (Fig. 28, a), the other with a group of milioline chambers at the beginning (Fig. 28, b). It appears probable that these represent the two forms of the species, comparable with those found else- where. Family Peneroplididae. — Four genera are here included — Peneroplis, OrUculina, Orbitolites, and the fossil Meandropsina. As Carpenter pointed out, there is in this sub- family a well-marked series of forms with varying degrees of complexity of structure. Moreover, the contrast in the arrangement of the early chambers in the two forms of the species (that of the microspheric forms is here, I believe, described for the first time) appears to offer an instructive parallel to that met with in the Miliolinidae and elsewhere. Hence the group will be rather fully described. Peneroplis is represented by a single species (P. pertusus, Forsk.) presenting, within certain limits, a remarkable range of variation.1 In all cases the chambers are simple. During the earlier stages of growth they are disposed on a planospiral plan, and this may be followed until the test is complete, but more usually the terminal chambers are disposed in a rectilinear series. The width of the later chambers varies very much, as seen in the " crozier-shaped " and broad tests represented in Fig. 29. Another varying feature is the "equitant" character of the chambers, as the result of which the earlier convolutions of the spiral part of the test are overlapped and hidden in varying degrees by the alar prolongations of the chambers which succeed them. The first few chambers communicate by single apertures, but the apertures soon become compound, consisting of a single or 1 Cp. for the superficial characters of the tests, Dreyer (13). Fio. 28. pt* U'X 6> from 94 THE FORAMIN1FERA double row of pores, or, in the " dendritine " varieties, of an opening the margins of which are produced into branched . and winding recesses. The extent of the septum which forms the terminal face of each chamber, and is perforated by the apertures, varies, of course, with the shape of the chambers, being small and more or less circular in outline in the crozier-shaped forms, and much elongated I'eneroplis pcrtusus, Forsk. t4, a common flattened variety ; ift, the crozier-shaped variety (x 26). A, central part of section of inegalo- spheric form. B, of the microspheric form ; sp.p, spiral passage ( x 280). in the flattened forms. In the latter case it is markedly convex when seen from the side. In the megalospheric form the size of the oval megalosphere varies in samples from different localities. Thus, in 300 specimens from Aripo, on the coast of Ceylon, I find the average value of M to be 30 /A, and it varies in different individuals from 24-42 /x. In a batch from Watson's Bay, Port Jackson, the average value is 42 /A, and the individual variation ranges from 32-59 /A. In two specimens in a West Indian gathering, however, M= 19 and 22 /z. A narrow spiral passage leads from the megalosphere and THE FORAMIN1FERA 95 winds round half to three - quarters of its circumference before opening into the first of the spiral series of chambers. As Rhumbler has shown (35) the walls of the central chamber and the beginning of the spiral passage are traversed by minute radial perforations, so that the test formed in the earliest stage of the life of this form of Peneroplis is perforated, though, the walls formed subsequently are, at any rate usually, imperforate, in accordance with the rule in the Miliolidea.1 The microspheric form of Peneroplis is very scarce in the material which I have examined. Among 1000 specimens I have met with only five. Its proportion to the megalospheric form appears, however, to vary in different localities. Thus, in a sample of sand from the Maldive Islands, dredged in 47 fathoms by Mr. J. S. Gardiner, the proportion is 3 to 108, or 1 to 36, which is about the same as obtains in Polystomella. On the other hand, in a batch of 480 specimens from Watson's Bay, I failed to meet with a single microspheric form. The values of m in my specimens are 17, 18, 19, 22, and 24 p. On comparing these with the values of M, it will be seen that the megalosphere may, in some cases, fall below the microsphere in size. The two forms are, however, sharply separated by the fact that, as in Orbiculina and Orbitolites, the spiral passage is absent in the microspheric form (Fig. 29). Orbiculina. — All the forms of the genus are included in a single species, 0. adunca (F. and M.). Its distinctive features are the subdivision of the chambers into chamberlets, and the equitant character of the chambers in the earlier convolutions, giving rise to a prominent umbo at the centre of the flattened test. It affords a good example of the mode of occurrence of variation in one of the Foraminifera with a definitely symmetrical test. Although the tests present themselves under a great diversity of external shape, the variations from the normal are limited to certain well-marked and definite lines. Fig. 30 shows the main varieties, as they are represented in a sample of ballast sand from the West Indies, kindly sent to me by Mr. F. W. Millett. The youngest tests are uniformly nautiloid (Fig. 30, a and i), the chambers succeeding one another in a closely wound spiral. As the sections represented in Fig. 31 show, the chambers are elongated transversely to the course of the spiral ; hence we may 1 The tests of Peneroplis present throughout their growth a surface pitting, which is usually shallow, but may be so deep as to amount very nearly to perforation. Indeed I am not convinced that some forms are not completely perforated in the later as well as in the initial chambers. As we shall see below, the central chamber and spiral passage of the megalospheric form are perforated also in Orbiculina and Orbitolites marginalis. There appears therefore to be no justification for the separa- tion of Peneroplis from the other genera of this family, as proposed by Rhumbler, on account of its supposed peculiarity in this respect. THE FORAM1NIFERA speak of their two ends as inner and outer — the former directed towards the concave side of the spire, the latter towards the convex side. While they are simple at their outer ends, the PlO. 80. Tests of Orbirtt/uio rufrouu, F. and M. a-b, young speci- mens ; c-e, var. flabelliformis ; / and g, var. compressa; h, var. adunca; 13, full-grown specimen of the microspheric fomi of adunca. (g and B are magnified about 8 times, the other tests about 7 times.) chambers of young specimens of all varieties are equitant at their inner ends, being produced on either side over the surface of the already formed test, in alar prolongations (cp. Fig. 32). As growth proceeds, the chambers become more and more THE FORAMINIFERA 97 elongated, and the characters of the three main varieties become apparent, the shape ultimately assumed depending on the extent to which the successive chambers overlap their predecessors at either end. In one variety, which may be called flabelliformis l (Fig. 30, d and e), the mode of growth changes from the spiral to the recti- linear. After the nautiloid stage the chambers lose their equitant character, and a series of long chambers is formed, each of which slightly exceeds its predecessor at either end. The fan-shaped tests here figured are thus produced.2 They attain 2 mm. in their greatest diameter. The commonest variety may be called variety adunca proper (Fig. 30, h). It includes the forms 0. adunca and 0. orbiculus (F. and M. spp.), of which the latter is the young, stage of the former. Here the spiral mode of growth and the equitant character of the inner ends of the chambers are maintained until the test is complete. At their outer ends the chambers extend little, if at all, beyond the outer ends of their predecessors, and thus build up the abrupt prominence in the outline of the test, characteristic of this variety. The third main variety is compressa (the 0. compressa of d'Orbigny ). In it the chambers cease to be equitant, and increase rapidly in length at both ends, being applied to and encircling more and more of the margin of the previously formed test. It thus comes about that the two ends of the chambers meet, forming a complete annulus. Henceforth the mode of growth is continued on the annular plan, and the disc-shaped tests represented in Fig. 30, g, are produced. In the sample examined the great majority of the specimens of Orbiculina are of the variety adunca. The next commonest variety is compressa, while flabelliformis is compara- tively scarce. This sample consisted in large part of the tests, young and old, of Orbiculina, and in hunting through it a form which did not fall into one or other of these varieties in some stage of growth, was very rare, the few exceptions being of an intermediate character. The condition of the central chambers of Orbiculina can only be observed in sections. Certain large forms of the variety adunca (Fig. 30, B), attaining a diameter of 6 mm., and distinguished also by the greater thickness of the tests and the much extended alar prolongations of the 1 In the description of Brady's figures of Orbiculina the term flabelliform is applied to two varieties — one (Fig. 7) the var. adunca; the other (Fig. 8) an exceptional form .intermediate between varieties adunca and compressa. The term naturally, however, implies a bilaterally symmetrical test, and I have therefore used it as stated above. 2 The cornucopia -shaped test, represented in Plate XIV. Fig. 4 of Brady's " Challenger " Report, is a rare form of this variety in which the increase is still less. 7 98 THE FORAMINIFERA chambers, are shown by section to be microspheric, though in their younger stages the microspheric forms are not readily distinguished by external characters. In a set of 38 examples of this variety, of such a size that the microspheric forms were not externally differentiated, 34 were found, by section, to be megalospheric and 4 microspheric, a proportion of 1 to 9'5. The diameter attained by the megalospheric form of this variety is, in my examples, 4 mm. S/t./L _ _ Fio. 31. Central regions of sections of Orbiculina adiinca in the median plane. A, large, A', small «xample of the megalospheric form. B, microspheric form, x 93. b, centre of B, x 250. sp.p, spiral passage. I have failed to find a single microspheric form of the variety compressa ; 100 examples, examined in section, belong to the megalospheric form. Among 24 examples of the variety flabelli- formis, however, one appears to be microspheric. The diameters of the megalospheres in well-grown specimens of the varieties adu-nca and compressa are as follows : — Var. adunca . ,, compressa No. of Specimens examined. 28 32 Highest Value of AT. 146 M 155 /x Lowest Value of AT. 81 M 70 M Average. 117 fi 109 /* THE FORAMINIFERA 99 In flabelliformis the value of M is less, though the average in my specimens is not below 64 /z. In the microspheric form the average value of m in six cases is 18 fM, the highest being 21 /A, and the lowest 16 /x. In the megalospheric form the megalosphere is followed by a spiral passage reaching round J-f of its circumference, and both megalosphere and spiral passage frequently ex- hibit the perforated con- dition found in Peneroplis (Figs. 31, A, and 32). The first of the spiral series of chambers is usually simple, but it frequently opens into its successor by two aper- tures, and in the second Flo 34> the transverse ribs gener- OrWcnMna adunca, central part of a section of a allv maVp tlipfr arvnpar megalospheric test passing transverse to the median •<* **"? I"" plane. Af, megalosphere; sp.c, spiral passage. ance, which, becoming more marked in the following chambers, subdivide them into chamberlets. The chambers also communicate with one another by an increasing number of apertures, arranged in several rows along their peripheral walls. In the microspheric form .of Orbiculina, as in that of Peneroplis, the microsphere opens direct into the first of the spiral series of chambers, and in this form there are generally some twenty simple chambers communicating by a single aperture before the sub- division into chamberlets begins (Fig. 31, B and b). Such are the characters of well-grown specimens of Orbiculina ; but on examining the construction of the tests of small specimens, as displayed in section, a mode of variation of a different kind becomes apparent, one which illustrates the phenomenon of Initial Polymorphism described by Munier-Chalmas and Schlumberger in Idalina. In the sample of sand which furnished the varieties above described were numbers of small megalospheric specimens resembling the young of the typical forms (c and / in Fig. 30), but beginning in a megalosphere of small size. In the specimen represented in Fig. 31, A', the central part of one of these is seen in section. Here the megalosphere measures only 34 /x in diameter. Associated with- the small size of the megalosphere of these forms is a long series of single chambers before the subdivision into chamberlets begins. In both characters they thus vary in the direction of the microspheric form, though always distinguishable from it by the presence of the spiral passage. In Orbiculina then, as in Idalina, the construction of the early part of the test is ioo THE FORAMINIFERA correlated with the size of the megalosphere. If it is small the arrangement approaches that of the microspheric form, if large it departs more widely from it. Another feature met with in some of these stunted forms, though by no means in all, is that the subdivision into chamber- lets may be incomplete or wholly absent. Sometimes the sub- division dies out in the terminal chambers after becoming estab- lished in their predecessors ; in others it is absent throughout the test. I am inclined to regard these latter forms as examples of Orbiculina which have lost their secondary septation by "degenera- tion " rather than as representatives of Peneroplis, because of the existence of the intermediate forms just alluded to, in which the subdivision dies out in the terminal chambers, and also because they agree so closely in external features with small examples of " typical " Orbiculina, that they cannot be distinguished from them by the external characters of the tests.1 Four well-marked species are generally included in the genus Orbitolites, of which three — 0. marginalis, duplex, and complanata — are intimately related to one another, and form a remarkably complete series -of grades of development, while 0. tenuissima stands apart. The three former are inhabitants of the littoral zone of tropical and subtropical seas, while the last lives in the deeper parts (250-1700 fathoms) of the North Atlantic, from which it extends into the Mediterranean. In all the annular arrangement of the chambers is assumed early in life, the tests have a flattened discoidal shape, and an umbo is absent, as the chambers are not equitant at any stage of growth. All but the earliest chambers are subdivided into chamberlets. In 0. marginalis (Lamk.) the chamberlets are generally some- what quadrangular when seen on the face of the disc, and the chambers they compose have an evenly curved outline. The disc consists of a single layer of chambers, and they are throughout simply applied to the peripheral margins of their predecessors. The radial septa which divide the adjacent chamberlets of an annulus from one another are traversed at their peripheral border by a canal, which places the chamberlets in communication with one another, and the canals of any one annulus may thus be regarded (following Carpenter's nomenclature) as composing an annular canal. From the canal, as it traverses a septum, a passage leads in a radial direction and opens either to the exterior by a 1 I have not had the opportunity of examining examples of Archiacina, -but from the figure given by Schlumberger (50, Plate III. Fig. 2) it seems possible that this may be a form of variety compressa which has similarly lost the subdivision of its chambers. THE FORAMINIFEKA 101 pore at the margin of the disc, or into a chamber of the suc- ceeding annulus, as the case may be. In this species the canals all lie in one plane, which is the median plane of the disc (cp. Fig. 38, m). m FIG. 33. Orbitolites marginalis, Lamk. m, whole test (x 20) ; A, centre of megalospheric, B, of micro- spheric form, x 100 ; b, centre of latter, x 280 ; x, the outer end of the last series of chara- berlets which follows the spiral mode of growth. The figures 3 and 11 in A and 6 mark the last of the undivided chambers in the two forms respectively. The microspheric form (Fig. 33, B and b). — The microsphere opens directly, without the interposition of a spiral passage (as appears always to be the case in the Peneroplididae), into the first chamber. The chambers are arranged at first in a gradually expanding spiral, eleven to sixteen simple chambers succeeding one another as in the genus Peneroplis, but communicating by 102 THE FORAMINIFERA single apertures. As the spiral increases in width the chambers become divided into chamberlets, and the number of apertures is correspondingly increased, the arrangement at this stage repeating that which in the earlier stages of growth is common to all varieties of Orbicvtina, except that there are no alar prolonga- tions. When the stage of the spiral mode of growth is complete, the chambers become successively more and more embracing and the annular arrangement is attained. In the megalospheric form (Figs. 33, A, and 34) the megalosphere is pear-shaped. A spiral passage leads from it, and extends round about three-quarters of the circumference of the megalosphere. As in the case of Peneroplis and Orbiculina the walls of the megalosphere and the spiral passage may be perforated (Fig. 34). The single chambers which follow are usually Orily three or four in number, and beyond, the chambers become subdivided, and the arrangement resembles that of the microspheric form. The dimensions of the central chambers in the specimens which I have examined are as follows. It will be seen that as in Peneroplis those of the megalosphere vary in the samples from different localities. No. of FIG. 34. megalosphere. Megalospheric — from Aripo (Ceylon) . „ W. Indies Microspheric — from Aripo . , . W. Indies Specimens examined. 43 47 17 1 Highest Value of M or m. 53 n 78 At 19 A* Lowest Value of M or m. 24 M 37 A* 15 M Average Value of M or m. 36 A* 51 At 17 At 18 /t In Orbitolites duplex, Carpenter, the arrangement is at first sight similar, but the chamberlets are more elongated in a direc- tion perpendicular to the face of the disc. Here, again, they are in communication by a single annular canal, but the apertures which open out of it and lead to the chamberlets of the succeeding annulus are disposed obliquely and lie in two planes, one on either side of the median plane (cp. Fig. 38, d). There are thus in typical specimens two rows of pores at the margin of the disc. There is, moreover, a difference in the shape and arrange- ment of the chamberlets. Instead of the regularly curved series of quadrangular chamberlets which make up the well-marked annuli of 0. margiiwlis, the chambers (especially those near the THE FORAMINIFERA 103 centre of the test) are oval, being elongated in a tangential direction, and fall into lines like those on the back of a watch, making what is known as the " engine turned " pattern. Orbitolites duplex, Carpenter, m, whole test of megalospheric form, x 5 ; A A' A", central parts of three varieties of the megalospheric form; B, of microspheric form, x 100; b, centre of latter, x 280. In an example of the microspheric form the microsphere is 20 /*, in diameter, and here again the spiral canal is absent (Fig. 35, B and b). There are eleven single chambers before the subdivision into chamberlets begins, and then the orbiculine arrangement is assumed, to pass in turn into the annular, as in 0. marginalis. In the megalospheric form the megalosphere is round or pear- shaped, and has an average mean diameter of about 76 /* (the 104 THE FORAMINIFERA diameters in 108 specimens vary between 49 and 110 //,). The spiral passage almost encircles the megalosphere, and is wider than in 0. marginalis. Though it sometimes communicates with only a single chamber (Fig. 35, A"), there are usually two to five chamber- lets into which it opens directly, by as many apertures (Fig. 35, A and A'), so that the peneropline and orbiculine stages are in such cases abridged, and the annular arrangement speedily attained. In a sample of the tests of 0. duplex from Aripo, all (108 in number) were megalospheric. The specimen of the microspheric form above described is from Funafuti in the Pacific. Fio. 36. Orbitolites complanata, Lanik. The megalospheric (A, a) and microspheric (B, b) forms, whole, and in section, x 5. The primitive disc is seen at the centre of A and a, and young megalospheric individuals ( = primitive discs) may be seen at the left-hand end of b. In Orbitolites complanata, Lamk. (cp. p. 73), a much greater degree of complexity is attained, in that the chambers, which in 0. duplex are elongated in a vertical direction, are differentiated into three several layers — two layers of superficial chamberlets, one on either face of the test, and an intermediate layer of columnar spaces lying between them (Figs. 36, b, and 38, c). There are here two annular canals corresponding to each annulus of chamberlets, lying at either end of the columnar spaces in the two strata of the test between these and the superficial chamberlets. There are abundant communications between the chamberlets, and those THE FORAMINIFERA 105 at the periphery open to the exterior by vertical rows of pores at the margin of the disc.1 The Microspheric Form. — The centre of the disc of this form is much thinner than that of the megalospheric (Fig. 36, a and b). It is often the seat of secondary growth which occurs towards the end of the vegetative phase, giving rise to a button -like ex- crescence and accompanied by absorption of the original central FIG. 37. Orbitdites complanata, Lamk. Central regions of sections of the megalospheric (A) and microspheric (B) forms, in the median plane of the discs, x 100. ca.ch, circumambient chamber ; M, megalo- sphere ; p, partition ; sp.p, spiral passage ; 11, the last undivided chamber of the microspheric form. chambers. If this has not occurred an arrangement similar to that of the central regions of the microspheric forms of marginalis and duplex is revealed by section. In two specimens I find that the microsphere has a mean diameter of 17 and 18 /x, a spiral passage is absent, and seven to eleven single chambers succeed the microsphere. These are followed by subdivided chambers, continuing the spiral, and the mode of growth then changes to the cyclical as in the other species (Fig. 37, B). In some varieties, at least, of this species the microspheric form attains a much larger size than the megalospheric (Fig. 36, A and B), and the large forms with double and contorted margins, described as variety laciniata, Brady, are all, as far as my experience goe.s, microspheric. It seems, indeed, that the peculiarity of the margin of this form may be regarded as a provision for supplying a larger number of peripheral brood chambers for the accommoda- tion of the megalospheric young into which the protoplasm becomes divided. 1 For the details of the structure, cp. Carpenter's descriptions (8 and 9). 106 THE FORAMINIFERA The megalospheric form begins in a structure called by Carpenter the primitive disc (Figs. 36, 37, A, and 38). It consists of (1) the megalosphere, which is pear-shaped (about 107 /x in mean diameter) ; (2) a spiral passage leading from the megalosphere, and opening into (3) a large crescent-shaped chamber, one horn of which extends round one side of the megalosphere, and the other along the outer side of the spiral passage. It results from this arrangement that the outer wall of the latter forms a parti- tion (p) disposed perpendicularly to the flattened surface of the primitive disc, and separating the spiral passage from the crescentic chamber. The partition ends in a free border. The spiral passage ac r P D Fio. 88. Diagram representing the transition from the simple (" marginalis ") to the complex (" complanata ") type of structure in the growth of a sub-typical individual of Orbitolites com- planata. The primitive disc and half the test of a megalospheric form are represented in section. The letters P D are placed beneath the centre of the primitive disc, m, part of test formed on the marginalis type ; d, that formed on the duplex type ; /, that formed on the type of a fossil form of 0. complanata; c, that of the typical 0. complanata; ac, annular canals ; cs, columnar spaces ; mp, marginal pores ; r, radial canals ; sc, superficial chamberlets. (After Carpenter, but modified.) with the crescentic chamber together compose the circumambient chamber of Carpenter. The whole of the peripheral wall of the circumambient chamber is perforated by pores opening into the innermost chamberlets, which are thus disposed in a complete annulus from the beginning. In some primitive discs there is a single row of pores at the margin, in others there are two or three rows. In the latter case the three-layered arrangement of chamberlets characteristic of 0. complanata is assumed directly ; while in the former the region of the test immediately surround- ing the primitive disc may present varying degrees of development. In some (Fig. 38) the rings of chamberlets are at first in single series, arranged on the marginalis type, and they are succeeded by annuli on the duplex type, the three -layered character being ultimately assumed. In others the arrangement begins on the duplex type. Here, again, we have examples of initial polymorphism. On comparing the primitive disc of 0. complanata with the centre of the tests of the megalospheric forms of the other species, it appears that the crescent-shaped chamber of complanata may be THE FORAMINIFERA 107 regarded as an expansion of the end of the spiral passage. In those forms of duplex in which the spiral passage communicates with more than one chamberlet the end is somewhat expanded. An extension of this expansion round the outer side of the spiral passage would give rise to the complete crescentic chamber which we find in complanata.1 Looking back on the series of forms of Peneroplididae hitherto examined, a gradual increase in complexity of structure is to be observed. We pass from Peneroplis, with undivided chambers dis- posed at first on a spiral and often, later, on a rectilinear plan, to Orbiculina, with subdivided chambers similarly disposed, though in one variety of the megalospheric form the annular arrangement is assumed. In Orbitolites marginalis the chambers and pores are disposed in a single plane, and in the early stages of growth we find arrangements repeating in some of their features those of Peneroplis and Orbiculina before the annular arrangement which is characteristic of Orbitolites is arrived at. 0. duplex, with its double series of pores, furnishes an intermediate stage to the complex three-layered condition of 0. complanata. In his Report on the genus Orbitolites, Carpenter made this series of genera and species the subject of a " Study of the Theory of Descent," and laid stress on the remarkable manner in which the forms of the simpler members are repeated in the life-history of the more complex. When this Report was published (1883) attention had only recently been drawn to the phenomenon of dimorphism in the Foraminifera, and Carpenter does not appear to have been aware of the existence of the microspheric forms, as constituting a distinct set of individuals. On comparing the mode of growth of the microspheric and megalospheric forms, we find a contrast between them comparable to that presented by the Miliolinidae. While the microspheric forms repeat successively the shapes and arrangements of chambers which are permanent in other, and in this case, simpler, members of the group, in the megalospheric forms these stages are to a greater or less extent abridged or altogether omitted. Thus in the megalospheric form of Orbitolites marginalis the peneropline series of single chambers which succeeds the spiral canal is fewer in number than in the microspheric form, but the orbiculine arrangement is well represented. In this form of 0. duplex the peneropline condition has almost or entirely disappeared, and the orbiculine stage is much abbreviated. In 0. complanata both 1 The remarkable fossil form Meandropsina described by Schlumberger (59) appears to be related to Orbitolites, the surface of the disc being covered with a layer of chamberlets arranged in a Meandrina-like manner. Schlumberger finds both microspheric and megalospheric forms are represented in his specimens. io8 THE FORAMINIFERA peneropline and orbiculine arrangements have entirely gone in the megalospheric form. We turn now to the other species, commonly included in the genus Orbitolites, the 0. tenuissima of Carpenter (Figs. 39 and 40). The tests are exceedingly thin (^J^ inch), though they may attain 30 mm. in diameter. Fio. 39. Orlitolites tenuissima, Carp. The complete test, x about 11, from a photograph. There are undoubtedly points of similarity in structure be- tween this species and 0. marginalis, the simplest of the other members of the genus. The annular arrangement succeeds a spiral one, and the annuli are divided into chamberlets by septa disposed in a manner which is very similar to that found in 0. marginalis, especially in examples from deep water in which the radial septa are sometimes incompletely developed. Coming to the middle of the test, however, we find ourselves in new country. In five specimens a globular central chamber about 31 /u in diameter l occupies the centre, and leading from this is a succession 1 In that figured in Plate I. Fig. 1 of Carpenter's Report the central part of the test appears to have been left blank, without any intention of depicting a central chamber of the size of the blank space. The specimen here figured was obtained by the Travailleur in the Bay of Biscay, and I am indebted to the authorities of the British Museum for the opportunity of giving a photograph of it. The central chamber measures 30 x 31 /A. In the four other specimens in which I have been able to obtain evidence of the size of the central chamber, it appears to be about the same. THE FORAMINIFERA 109 of narrow elongated chambers, wound in a planospiral manner about the central chamber in some 7-8 convolutions. The lengths of the several chambers vary from 2J convolutions to \ of a con- volution of the spiral. The arrangement and mode of communi- cation of the chambers recalls the irregular spiroloculine tests of Fio. 40. Orbitolites tenuissima, Carp. Central" region of the specimen represented in Fig. 89. The figure 8 is in the eighth and last convolution of the inner series of chambers. X 80. Ophthalmidium (Fig. 41). As the five specimens have an approxi- mately similar arrangement, it is probable that the form we are dealing with is megalospheric, though the size of the megalosphere is small. When comparing the central regions of 0. marginalis (megalo- spheric) with those of 0. tenuissima, Carpenter regarded the spiral passage of the former as representing " the whole of the original 'spiroloculine' coil, drawn up into itself" (p. 24). The difficulty, however, of recognising the long spiroloculine (or Ophthalmidium- like) coil of tenuissima in any of the modifications of the spiral passage met with in the other species of the genus Orbitolites is so no THE FORAMINIFERA great that we are led to doubt whether tenuissima is really allied to them. On the other hand, the resem- blance of the inner chambers of tenuissima to Ophthalmidium,a,member of the Hauerinidae, suggests that it may be derived from this family, and have acquired the cyclical mode of growth independently. The acceptance of this view is perhaps rendered easier by the existence of another group, the Oper- culina-Cydodypeus series, in the higher mem- bers of which the annular mode of growth is likewise attained (see p. 128). It seems at any rate worth while to entertain the possibility of this explanation, before ac- cepting a conclusion so damaging to a body of evidence which may be found, if duly considered, to furnish the clue to many complicated problems of relationship.1 Family Alveolinidae. — The genus Alveolina which represents this family contains a number of recent and fossil forms which appear to branch off from the Miliolid stock in the neighbourhood FIG. 41. fi°h£kn9er" I(*vort> Tests of (a) Alveolina boscii, Def., x about 17 ; and (b) A. melo, F. and M., x about 22. of the genus Orbiculina. They are char- acterised by elongation of the chambers in a plane at right angles to that in which they are developed to form the disc -shaped tests of Orbitolites — that is, in the direction of the axis of the spire. The result is the formation of a series of oblate, spherical, ovoid (Fig. 42, b), fusiform, and cylindrical (Fig. 42, a) tests, each chamber extending beyond its predecessors laterally to a greater or less extent, and thus increasing the axial length of the test. The chambers are short in the direction of the plane of the spire, and subdivided into chamberlets by vertical septa lying parallel with that plane. In 1 It would perhaps be premature, while we are not yet acquainted with the two forms of tenuissima,, to alter its systematic position, but should this view of its re- lationship be confirmed, it must be separated as a distinct genus to which the name Cyclophthalmidium might be given. THE FORAMIN1FERA in the recent Alveolina loscii the chambers are further subdivided by horizontal septa. Schlumberger states (51) that Munier-Chalmas has recognised the phenomenon of dimorphism in a fossil Alveolina, of which the microspheric form is distinguished by a very small central chamber surrounded by five simple chambers, which are not subdivided. It would appear, therefore, that a peneropline stage is represented also in the development of the microspheric form of this genus. In specimens of the megalospheric form of A. loscii I find the central chamber to be ovoid and to measure about 150 p in lone diameter. Life-histories and nuclear characters of the Miliolidea. Direct observations on this head are very scanty. In Cornuspira, as we have seen (p. 74), the megalospheric form may give rise to megalospheric young, and the same event has occurred in a specimen of a Milioline form (? Quinqueloculind) in my possession. In this case the megalosphere of the parent was only 30 /* in diameter, and those of the young varied from 20 to 43 fM. Schlumberger (49) and Schaudinn (42) have also described the production of broods of young, which were evidently megalo- spheric, in the Miliolinidae (the latter author in Quinqueloculina seminulum, L.), but the nature of the parent is not indicated. In these Miliolinidae it appears that the division of the protoplasm to form the young may occur within the parent test or outside it. The production of megalospheric young by the breaking up, within the test, of the protoplasm of a megalospheric parent, had occurred in a specimen of Peneroplis described by Schacko (39) ; and in this case the young, consisting of the central chamber and spiral passage, resemble in size and shape the corresponding parts of the parent. Biitschli (7) has found a single nucleus in two specimens of Peneroplis, and 18-20 in another. In all three cases the parents were megalospheric, and in the last we may suppose that the division of the nucleus had occurred preparatory appar- ently to the production of a brood of megalospheric young, as in the cases of Discorbina and Patellina (see p. 123). Of the genus Orbitolites our knowledge is somewhat fuller. I have a specimen of the megalospheric form of 0. marginalis in which the two or three peripheral annuli contain young, con- sisting of the megalosphere and ^ spiral passage. The chambers containing them are in this case not different from the ordinary marginal chambers. A specimen described by Semper (63) appears to have belonged to 0. duplex, and this also is a megalospheric parent with megalo- spheric young. He mentions that the chambers containing them 112 THE FORAMIN1FERA were "ziemlich viel grosser" than those internal to them, and thus it may be the case that in this as in other characters 0. duplex is intermediate between marginalis and complanata. In 0. complanata the microspheric forms known as var. laciniata produce megalospheric young. The double convoluted margins of this variety are not completely subdivided into chamberlets as are the more central regions of the disc, but, in part at least, con- tain spacious chambers extending through the thickness of the disc, and round a large part of the periphery. Into these (as well as into similar large chambers in the secondary growths formed on the surface of the disc) the protoplasm withdraws, at the reproduc- tive phase,- from the whole central region of the original test, and becomes divided up into young megalospheric forms, which are liberated by the breaking down of the limiting walls. This mode of reproduction in 0. complanata was first described by Brady (4), though he was not aware of the full significance of his observation, and afterwards by myself (20). In Fig. 36, b, a form with a simple margin is seen bearing megalosphejic young. The megalospheric form of 0. complanata may also, as we have seen (p. 74), give rise to a brood of young of the same nature, but there can be no doubt that a phase recurs in the cycle of the life-history in which, as in Polystomella, zoospores are produced. The microspheric form of 0. complanata has, scattered through its protoplasm, large numbers of rounded nuclei, which may fre- quently be found constricted as though in process of simple division. In the megalospheric form a large nucleus may often be found throughout the greater part of the life lying in the primitive disc, and thus, as already pointed out (p. 71), at the central part of the protoplasm (20). Calciluba polymorpha appears to be a degenerate member of the Miliolid stock. Its life-history, as exhibited in aquaria, has been investigated by Schaudinn (46). It forms wide adherent expan- sions on the surface of foliaceous algae on which it feeds, spreading in irregular annular patches — like fairy rings. The colony may begin as a spherical central chamber with a spiral passage leading from it — the form which occurs so frequently at the centre of the megalospheric tests of the compact Miliolidea. From such a centre branching offsets extend in a radial direction over the algal sub- stratum, and as this is disintegrated by the organism feeding on it, the central regions, left unsupported, may fall away, while the margins spread in the annular fashion described. The portions which so fall may start a similar colony forthwith, or their proto- plasm may break up into small portions (1-20) of varying size, which at first crawl about as naked masses, and later, on initiating a new colony, may secrete the Miliolid form of test mentioned above. THE FORAM2NIFERA The walls are chitinous tubes with a calcareous deposit. They are imperfectly divided into chambers, and are not perforated. A flagellate stage did not come under observation. The protoplasm contains large numbers of small nuclei. ORDER Textularidea. This order contains a number of genera which are excellent examples of the multiform (biformed and triformed) condition of the test. The arrangement which has been regarded as typical of the genus Textularia is one with two rows of alternating chambers, but Schubert has recently drawn attention to the fact (65) that many, if not all the forms included in it, are biformed, some having the earlier Fio. 43. Spiroplecta (Textidaria) sagittula, Def. A, the megalospheric, B, the microspheric form, X 55 ; b, the earlier chambers of the latter, x 150. A and B represent specimens stained, and mounted in Canada Balsam, and show the nuclei. chambers arranged in a planospiral, others in a rotaloid, and others again in a triserial manner, before the characteristic biserial arrange- ment is assumed. Thus "Textularia" sagittula, Def., begins as Schubert states, and as I have also had occasion to observe, in a planospiral series of chambers, the arrangement being, in fact, that character- istic of the genus Spiroplecta. Out of a batch of 63 specimens of this species collected at Plymouth in the month of July, I found 57 to be megalospheric, and 6 microspheric, a proportion of 9 to 1. 8 THE FORAMINIFERA In the microspheric form (Fig. 43, B and b) m = 15-18 ^, and is followed by some five or six chambers arranged in a spiral before the alternating arrangement is assumed. One specimen contains at least 13 nuclei. In the megalospheric form (Fig. 43, A), the average mean dia- meter of the megalosphere is about 60 p. (the limits of those measured were 44 and 72 /A). The initial spiral is here somewhat shorter, consisting of four chambers. A single large nucleus is seen in these specimens some distance along the alternating set of chambers. The spiral arrangement of the early chambers is much more conspicuous in Spiropleda annectens, P. and J. (Fig. 44, A, B, and b). Fio. 44. A, B, and b, Spiroplecta annectens, P. and J ; A, the megalospheric, B, b, the microspheric form. C, Verneuuina pygmaea, Egger. D, Bigenerina robusta, Brady. B, Clavulina angularis, d'Orb. A and b, x 70, original ; B-B, from Brady (8). What appears to be the megalospheric form of this species has long been known as a Cretaceous fossil. The species occurs at the present day round the coasts of Australia, and has been recognised in sand from the Malay Archipelago by Mr. F. W. Millett (24, Part VII., 1900), to whom I am indebted for calling my attention to the evidence of dimorphism in this species, and for the oppor- tunity of examining the specimens from which the following details are given. Among six specimens of the megalospheric form (Fig. 44, A), the average value of M = 60 //, (the limits of variation are 53 and 71 /x), and one nearly complete spiral whorl of chambers intervenes before the straight and biserial part of the test begins. The microspheric form attains a larger size. Among 15 speci- mens, the average value of m= 17 /* (the limits being 11 and 20 /*), THE FORAMINIFERA and 2J-3f whorls of chambers arranged in a spiral form the earlier part of the test. The later, straight part is much longer than in the megalospheric form, and in both forms the biserial arrangement may give place to a uniserial one at the end. The characters of other genera of this family are indicated in the table of classification, and some of them are represented in Fig. 44, C-E, and Fig. 54. ORDER Chilostomellidea. I am not aware of any record of dimorphism in this order. ORDER Lagenidea. Schlumberger has found representatives of both generations in Nodosaria (Dentalind) guttifera and Nodosaria hispida, the megalo- spheric forms beginning in a large initial chamber (Fig. 45, A), larger than that which succeeds it, and having only five or six FIG. 45. Megalospheric forms of— A, Nodosaria hispida, d'Orb. B, Nodosaria (Dentalina) communis, d'Orb. C, Frondicularia alata, d'Orb. (After Brady, 3.) chambers in all ; the microspheric having a larger number of chambers, and tapering gradually to a fine point at which the little microsphere is situated. In such tests the phenomenon of dimorphism is presented in the simplest possible form. Fornasini (15) has shown that Frondicularia (Fig. 45, C) is dimorphic. The monothalamous Lagenidae often present a great resem- blance to the single chambers of Nodosaria, but the nature of the relation between the two groups is obscure. Neumayr derives the former from the latter by degeneration ; Rhumbler, by the falling u6 THE FORAMINIFERA apart of the chambers. A remarkable feature of some Lagenidae is the " entosolenian " condition in which the tubular neck is, as it were, inverted into the interior of the test (Fig. 46, b). A similar FIG. 46. a, Lagena sulcafa, W. and J., X 60. b, /,. globosa, Montagu, showing the ento- solenian neck. (After Brady.) x 80. FIG. 47. Cristettaria crepidula, F. and M., after Brady (3, PI. 68, Fig. 1), showing the pro- duction of a brood of megalospheric young, of varying size, by a megalospheric parent. X 38. inverted neck is found in Cymbalopora, and occasionally in Poly- morphina (3, pp. 558 and 638). The observations of Burrows and Holland (5) on Cristel- laria gibba, and C. platypleura appear to show (though no measurements are given) that the authors have found dimorphic forms of Cristellaria. In C. cenomana, Schacko (40) describes a form which, he suggests, is micro- spheric, having a central chamber measuring 40 /*, while M = 75 p.. In the specimen of C. crepidula shown in Fig. 47, however, the size of the young chambers (which we may suppose to be megalo- spheres) varies much, and the smallest appear to measure about 40 fj. in diameter; and as this measurement is rather large for the size of a microsphere, the mi- crospheric character of Schacko's specimen is, at least, open to doubt. The genus Polymorphina is FIG. 48. remarkable for the fistulose Polymorphina comprasa, d'Orb. ; c, the , . , simple form, x 82 ; d, the flstulose form, branching processes which are xss. (After Brady.) developed in the later stages of the growth of the test. What relations these may have to the life -history has not been determined (Fig. 48, c and d). THE FORAMIN1FKRA 117 Schlumberger has shown (50) that in Siphogenerina glair a the microspheric form tapers to a point at the initial end, and has 9-10 chambers arranged alternately before the uniserial mode of growth is assumed ; while the megalospheric form is short, begins abruptly with a large central chamber, and has only three alternat- ing chambers prior to the uniserial chambers. ORDER Globigerinidea. In this group, as defined by Brady, the tests consist of a few inflated chambers arranged in a spiral manner. The members of it inhabit the surface waters of the ocean, furnishing an important \ Fio. 49. Globigerina bulloides, d'Orb. (to left), and Orbulina universa, d'Orb. (to right) (From Rhurabler, 38.) The figure of G. biilloides represents the test as seen from the superior sur- face. The specimen departs from the normal in possessing an aperture on this aspect of the terminal chamber. constituent of the pelagic fauna ; and their empty shells, falling to the bottom, form the main constituent of the " Globigerina ooze " (see p. 138). Globigerina bulloides, d'Orb. (Fig. 49), the most abundant species of the genus, has globular chambers forming a "rotaline" test, each opening by a separate orifice into the deep umbilical space on the "inferior"1 surface. The chambers increase rapidly in size, as the series is followed, and there are three or four in the terminal convolution. The walls of the chambers are perforated by pores, and at first are thin and smooth. As the shell increases in thickness, 1 See the characters of Rotalidae, p. 145. 1 1 8 THE FORAMINIFERA it generally becomes areolated on the surface, the deposit being greatest between the pores, so that these open into cup-shaped depressions, separated by ridges. In many, but not all pelagic specimens the shell is produced on all sides into radiating cylindri- cal spines which spring from the points where the ridges meet, and may exceed the diameter of the shell in length.1 A large proportion of the individuals, which in their earlier stages conform to the type of Globigerina bulloides, complete their growth under this form ; but for others a different future is in store. Having attained a size which may be equal to that of the full-grown test of the other specimens, or may fall considerably short of it, these secrete a large spherical chamber which usually encloses the whole of the previously formed test, and is frequently more than double its diameter (Fig. 49, right-hand cut). The enclosed test is usually only connected with the investing wall at the points where its spines meet the wall and unite with it. The investing chamber is perforated by large pores, with a diameter of from 13-21 p, as well as by minute pores (5-6 p). The specimens which form the spherical chamber have been given the generic name Orbulina. It will be convenient to use the terms Orbulina chamber, and Globigerina chambers for the invest- ing and the invested chambers respectively. Unlike the Foraminifera which creep over the sea-bottom, the pelagic Globigerinae may be found invested with a vacuolated covering which is in part gelatinous (38, p. 6), though traversed by radiating pseudopodia which project beyond it. This en- velops the whole shell and the bases of the spines, and has a spherical contour. It is probable that the Orbulina chamber is secreted at the surface of this vacuolated mass. A similar cover- ing may be found investing the Orbulina shell in the later phases of the life-history. As in the free Globigerina, the outer surface of the Orbulina chamber is beset with spines, which vary greatly in length, and specimens have been found, though rarely, the surface of which is areolated by ridges as in Globigerina. These have been separated under a distinct name — 0. porosa, Terquem. Rhumbler finds that, in pelagic specimens, the Globigerina chambers are always present within the Orbulina shell, though, 1 Sir John Murray thus describes the appearance of the living animal: "In Qlobigerina bulloides (hirsuta) and aequttateralis the yellow - orange colour of the Barcode is due to the presence of numerous oval-shaped xanthidiae or 'yellow cells,' similar to those found in the Radiolaria. When the sarcode with these 'yellow cells ' flows out of the foramina, and mounts between the numerous spines outside the shell, the whole presents a very striking object under the microscope ; the trans- parent sarcode can be seen running up and down the long silk-like spines, and the 'yellow cells ' seated at the base of these spines quite obscure the body of the shell." — Nat. Science, July 1897, p. 20. THE FORAMINIFERA 119 owing probably to the solvent action of the sea- water, they are often reduced to fragments, or absent in bottom specimens. The Globigerina chambers contained in the Orbulina shell differ from the free Globigerina bulloides in no respect, except in the ex- treme thinness of their walls, and Rhumbler (38) is inclined to separate the thin-walled shells hitherto classed under that species as the young stages of Orbulina universa, d'Orb. Rhumbler also points out, however, that in Globigerina bulloides, var. triloba, Reuss, which is characterised by the large size of the three last chambers, but not by the thinness of the shell, all variations are found be- tween a terminal chamber which is folded back on its predecessor, and one which completely envelops the other chambers, as in Orbulina. The existence of these transitional forms in a variety with a shell of the usual thickness raises the question whether the Globigerina chambers enclosed in the Orbulina shell were so thin when free, or owe their thinness to the action of the protoplasm after their enclosure. However this may be, we have the fact that some specimens classed as Globigerina bulloides end their individual existence in the Globigerina form, while other specimens, little or not at all distin- guished from them in the early part of their growth, become en- veloped by an Orbulina shell. These have been classed under a separate genus as Orbulina universa. The close resemblance between these two sets of specimens in the early stages of growth, and also between the Orbulina shell and that of the free Globigerina, in the varying development of the spines and the surface sculpture, strongly suggests that there is some more intimate relationship between them than that of allied genera, but what its precise nature may be is still very obscure. A large inflated terminal chamber is also found in Cymbalopora bulloides, and in the littoral Pulvinulina lateralis, Terquem, and these, like the Orbulina chamber, are also perforated by large pores. Cymbalopora was taken in numbers by the Challenger, as a pelagic form, in the neighbourhood of coral reefs, and, according to Murray, every shell was filled with minute monadi- form bodies.1 This observation would suggest that the inflated chamber may go with the megalospheric form, bulb though Rhumbler finds a single large nucleus in all the specimens of Orbulina he examined, the same was true of the free Globigerinae. The size of the central chamber of the included test of Orbulina varies, according to Schacko (39), from 16-23 /* in diameter, while in the free Globigerina it varies from 7-20 /*. Neither in the size of the central chamber, nor in the character of the nuclei, therefore, have we at present direct evidence for dimorphism among these animals. As to the modes of reproduction of Globi- 1 Brady (3), p. 639, footnote. 120 THE FORAMIN1FERA gerina or Orbulina] almost nothing of a definite character is known.1 ORDER Eotalidea. Schlumberger has found Rotalina pleurostomata, Schlumb. ( = Pulvinulina partschiana, d'Orb.), to be dimorphic (51), and I have found the same in Rotalia beccarii, Linn. Among seven examples of this species six were megalospheric, and one micro- spheric. In these M = 55 p (limits of variation 37 and 65 /x), and FIG. 50. Rotalia bcecarii, L., seen from the superior (a) and inferior (b) surfaces, ap, aperture, x 80. m = 1 3 /x. There appears to be no difference in the mode of arrangement of the chambers in the two forms, but the nuclear characters agree with those of Polystomella. I have observed the production of a brood of megalospheric young by a microspheric parent of this species, the process agreeing with that described in Polystomella (20, p. 436). Similarly in Calcarina kispida, Brady, M = 49 /x (limits of varia- tion in twelve examples 39 and 59 /x), m= 13 /x (limits 12 and 14 /x), and here, again, I found a microspheric specimen with megalospheric young, which in this case were contained within the parent shell. The rose-coloured adherent tests of Polytrema are common on coral and other objects from tropical and sub-tropical shores. They may be depressed and encrusting, but frequently rise from an expanded base into arborescent forms. They are built up for the most part of numerous successive laminae of hard perforated shell substance, produced inwards at short intervals into hollow pillars (Fig. 51, a, p\ which are connected with the underlying shell lamina. The openings of the pillars of the superficial layer 1 The fact that the Orbulina chamber is formed in the later stages of growth of the individual, which in its earlier stages formed the enclosed Globigerina chambers, was first definitely stated by Rhumbler (34). The view had, however, been previously suggested by Major Owen (32, p. 147). THE FORAMINIFERA 121 give rise to the deep pitting of the surface from which the genus is named. Except in the early stages of growth there is no sub- division of the test into definite chambers. The protoplasm is contained between the laminae, and in irregular spaces which occupy the axes of the branches ; it communicates with the exterior by the numerous perforations in the laminae, and at the ends of the branches where the axial spaces open widely. Their mouths are often beset with sponge spicules, which appear to be used as a temporary scaffolding for the support of the extended pseudopodia, in advance of the proper wall. At the base, however, in contact with, or close to, the object to which the Polytrema adheres, a spiral group of chambers is found — the initial stage of the test (Fig. 51, a, R, and b and c). These initial chambers have the thick coarsely perforated walls, the abundant chitinous element, and the spiral arrangement character- istic of the order Rotalidea. In ten specimens of the megalospheric form, I find that M varies from 110 to 29 /A, its average value being 51 p. In these specimens there are generally three chambers, following the megalosphere, arranged in a simple spiral (Fig. 51, c); the fourth chamber usually communicates by apertures with two or more chamber's, and after this the arrangement becomes more and more irregular, all distinction between the chambers and connecting passages is gradually lost, and the laminate structure of the test is attained. I have not happened to meet with specimens of the microspheric form, but this has been described by Merkel (23) l who finds that the microsphere (size not given) is succeeded by a regular spiral of some eleven chambers, before the chambers assume the irregular transitional character. In some cases the spiral of initial chambers is separated from the supporting object by a layer of small chambers of the irregular transitional form, and Schlumberger (56) has found in sand from the Azores small examples of the megalospheric form as free globular tests, consisting of the large initial chambers invested on all sides with a layer of small ones. It is evident, therefore, that Polytrema may pass through a more or less prolonged period of free life before it becomes adherent. Merkel found the nuclear condition to agree with that of Polystomella, three examples of the megalospheric form containing a single large nucleus lying in the megalosphere or an adjoining chamber, while in one of the microspheric form four nuclei were counted. 1 In the megalospheric form, Merkel describes the megalosphere as communicat- ing directly with some three of the surrounding chambers — a condition which I have not met with. 122 THE FORAMINIFERA Polytrema was associated by the earlier naturalists with various animals classed as " zoophytes," and was included by Pallas and Fio. 51. Polytrema ininiaceum, Linn, a, section of the test passing along the axis of a branch, and through the stem of a Polyzoan (P), to which the test was adherent ; p, the hollow pillars be- tween the laminae ; B, the group of rotaloid chambers, the initial stage of the test, x 20. ft, the group of rotaloid chambers (a, R), x 78 ; s, the surface of attachment to the Polyzoan ; ch, the chitinous lining of a chamber, c, central part of the base of a young test decalcified and treated with caustic potash, seen from the surface of attachment. The chitinous walls of the chambers remain. M, the megalosphere, 1-4, the first four of the spiral series of chambers, x 78. Gmelin in the genus Millepora. Its Rhizopod affinities were first recognised by Dujardin, and its relation to Tinoporus by Carpenter. THE FORAMINIFERA 123 Max Schultze was the first to demonstrate the spiral arrangement of the early chambers, and its dimorphic character was shown, as we have seen, by Merkel. The life -histories of Patellina corrugata, Will, and Discorbina globularis, d'Orb., as exhibited by specimens living in aquaria, have been investigated by Schaudinn (45). Ordinarily the protoplasm contains many granules which, during life, obscure the nuclei, but by excluding animal food, and limiting the diet to the diatoms growing on the sides of the vessels, Schaudinn succeeded in ren- dering the nuclei visible, so that their changes could be followed in the living animal. The form of reproduction observed was that comparable with the production of broods of megalospheric young by a megalospheric parent, and Schaudinn's account of the changes which the nuclei undergo is the fullest which we yet have of their behaviour in this phase of the life-history of the Foraminifera. All the specimens which came under notice contained a single nucleus in their early stages. As the reproductive phase approached the nucleus became segregated into a number of parts (usually 7-10), which were dispersed in the protoplasm, and in some cases became subdivided by a similar process, so that there may be as many as 30 nuclei of unequal sizes. The protoplasm becomes divided up about the nuclei into masses proportional to them in size, and the young thus produced repeat in turn the same cycle of development. In Discorbina the division to form the young occurs within the parent test, from which they escape by the resorption of its walls. In Patellina it occurs in the large umbilical space, e.g. outside the parent test. Schaudinn is inclined to the con- clusion that in these species the stage in which zoospores are pro- duced has been lost from the life-history, and that reproduction takes place only in the manner described. Thus he regards these species as having been originally dimorphic, but now monomorphic. I have measured the central chambers of a number of stained and mounted specimens of Discorbina globularis collected from the seashore, and the results are shown in Fig. 52. It will be noticed that in this species the central chamber is on the whole remarkably small. In the great majority it varies in size from 12 to 31 />t, the average of 159 specimens being 19 /x. In one case it was only 9 ^ in diameter. In this species the chitinous element of the shell is very abundant, and forms an obstacle to the penetration of staining reagents, but 54 of these specimens afford an indication of the nuclear condition. In 48, including two with a central chamber 16 /x in diameter, a single nucleus is present, and in one of the remainder a large nucleus is killed in the process of breaking up 124 THE FORAM1NIFERA into fragments. In the specimen, the central chamber of which measures 9 /A in diameter, 6 nuclei are clearly seen. In the 4 remaining multinuclear specimens the mean diameters of the central chambers are 22, 18, 12, and 12 /z. Now it is possible that all these examples of Discorbina belong to a single series illustrating the phases of the life-history which Schaudinn has followed in aquaria, but the coincidence of the occurrence of the multinuclear condition with the very small central chamber, 9 p in diameter, suggests that Discorbina is, like its allies, a dimorphic form. On this view we may regard the specimens with a single nucleus as megalospheric, and the specimen 5 33 S FlO. 52. Table showing the dimensions in micromillimeters of the central chambers of 159 specimens of Discorbina globularis, d'Orb. with (at least) 6 nuclei and a central chamber 9 /LC in diameter as microspheric. The remaining multinuclear specimens may consist of megalospheric individuals, the nucleus of which is breaking up prior to reproduction, or of microspheric individuals with a larger microsphere, or, and more probably, of both kinds. On this view the form of reproduction which Schaudinn described in Discorbina is the production of megalospheric young by a megalospheric parent which is, as we have seen, of frequent occurrence in other genera. The formation of zoospores by the megalospheric parent was not observed among the specimens kept in aquaria, but we are still at liberty to suppose that this phase of the life-history may occur in the natural state. Truncatulina lobatula, \V. and J., affords another instance of the THE FORAMINIFERA 125 approximation of the megalosphere to the size of the microsphere. Thus in 13 examples I found 12 to be megalospheric and the value of M to be, on an average, 28 /x, varying from 36 to 1 5 p. The other specimen is microspheric, and m = 1 1 /A. The nuclear characters corresponded to those described in Poly- stomella (20). Plastogamy. — This remarkable and little understood process which has been observed in other groups of Protozoa was found by Schaudinn to be frequently associated with the reproduction of Patellina and Discorlina. In the former the pseudopodia of two in- Fio. 58. View from the under side of two specimens of Patelliiia corrugata, Will., which have united in plastogamy prior to the breaking-up of the united protoplasm to form a brood of young. 1, young of varying size ; 2, nucleus of a young individual ; ,3, accumulations of detritus. (After Schaudinn, 45.) dividuals that have come, apparently by chance, into juxtaposition fuse, and form a uniting band which increases in thickness until all the extruded protoplasm is involved in it, and the tests are drawn close together. The nucleus in each meanwhile divides in the manner above described. Gradually the protoplasm of both emerges into the space between the bases of the approximated tests and the surface to which they are attached, and then, as in the reproduction of a single individual, divides up about the nuclei to form a brood of young (Fig. 53), As many as five individuals may thus unite. In no case did Schaudinn observe any fusion between the original nuclei or the fragments into which they divided. He also found that the process only occurred when the nuclei of the individuals which met were in the same phase of 26 THE FORAMINIFERA development ; thus a one-nucleated individual and an individual whose nucleus had begun to divide would not unite. In Discorlina a similar process was observed ; but in this case the two individuals came together base to base, and the pair wandered about for a consider- able time before the young were produced. In some cases a deposit of lime between the opposed bases occurred in the interval, so that after the escape of the young the empty parent shells remained united together. The remarkable pairs of shells which have been observed in Discorbina, Textularia, and Bulimina (Fig. 54) are, probably, thus explained. Fio. 54. Paired tests of a species of H\di~ minn from Uelos. In n tlie paired individuals are of equal, in b of very unequal size. From speci- mens kindly given me by Mr. H. Sidebottom. ORDER Nummulitidea. The members of this order are dis- tinguished by their bilaterally sym- metrical tests, which in the early stages or throughout growth are arranged on the spiral plan ; by the double character of the septa between the chambers, containing branches of the highly developed canal system interposed between the laminae ; by their hard perforated walls; and by the slit-like aperture (subdivided in Polystomella) situated between the inner margin of the septum and the wall of the previous convolution. The structure of Polystomella is described above (pp. 62 et seq.). Nonionina is a simpler form of the same type, characterised by the scantiness of the umbilical deposit, the absence of retral processes, and the fact that the aperture is not subdivided into pores, as in Polystomella, but remains a simple slit. Amphistegina is transitional in structure between the Rotalidea and Nummulitidea ; it has simple septa, and the test is not truly symmetrical, the spire being (as in the Rotalidea) slightly helicoid and the aperture on one (the " inferior ") side of the median plane. In the marked development of the alar prolongations it approaches the genus Nummulites. In Opercvliiia (Fig. 55) the chambers are simple and disposed in an expanding spiral, some three or four convolutions completing the test. The earlier chambers are produced to a greater or less extent into alar prolongations, and thus a boss-like umbo is formed at the centre ; but the later chambers are simply applied to the THE FORAMINIFERA 127 margin of the previous convolution, and have a large radial extent. The aperture is crescentic and undivided, but is supplemented by pores distributed along the septum. The canal system is highly developed, a plexus of canals, in connection with the meri- dional vessels, running in a keel -like thickening at the outer rim of the test. A sample of sand obtained by Mr. J. Stanley Gardiner from Suvadiva in the Maldive Archipelago consists for the most part of the tests of Oper- culina complanata and Hetero- stegina depressa, d'Orb. Of the former I have found two speci- mens of the microspheric form, in both of which m = 27 ^ and the second chamber is very minute (Fig. 55, B). In five specimens of the meg- alospheric form the values of M vary from 45 to 63 /* and have an average of 54 p. I have not observed that the tests of the two forms of this species are distinguished by a difference in size. In Nummulites the chambers are of very small extent in a radial direction (Fig. 6), so that each convolution adds little to the diameter of the test, and the outline of the latter is nearly circular. The number of the convolutions is however very large. The alar prolongations, on the other hand, are highly developed, and extend nearly to the centre. As the chambers of each successive convolution are thus produced, the result is that the tests are strongly biconvex, the spiral axis measuring from one -third to three-quarters of the diameter. The alar prolongations of the chambers may be directed straight towards the spiral axis (radiate type) or take a sinuous course (sinuate type), or they may be replaced by a number of separate chamberlets forming when exposed in worn specimens a network over the surface (reticu- late type). The aperture and canal system resemble those of Operculina.' The microspheric form attains, as we have seen, a much Fio. 55. Operculina complanata, Def. in, a complete test, x 8 ; A, central part of section of megalo- spheric form ; B, of the microspheric form. (A and B, x 50.) 128 THE FORAMIN1FERA greater size than the megalospheric (Fig. 5). I am not aware of any record of the actual size of the microsphere. In the megalospheric forms M is very large, attaining in some cases 1 mm. While Operculina is closely connected on the one hand with Nummulites, it forms, on the other, the simplest term of a series — Operculina, Heterostegina, Cycloclypeus, which presents among the Nummulitidea a remarkably complete parallel to the Peneroplis, Orbiculina, Orbitolites series in the Miliolidea. In Heterostegina (Fig. 56) the arrangement of the chambers is spiral, though they become somewhat more embracing as age advances. The chambers which are first formed are simple, as in Operculina, but they soon become subdivided by partitions which are disposed perpendicularly to the plane of the spiral, and, on the whole, transverse to the long diameter of the chambers. The chambers are thus subdivided into more or less quadrangular chamberlets. As in Operculina, the chambers of the inner convolu- tions are produced into alar prolongations (which are, however, not subdivided into chamberlets), while the later chambers are simply applied to the edge of the preceding convolution. The arrangement thus presents considerable resemblance to that of Orbiculina, but in addition to the presence of the canal system, perforate walls, and double septa, there is also a marked difference in the mode of communication of the chamberlets, for here the adjacent chamberlets of a chamber do not communicate directly with one another, but each communicates as a rule with two chamberlets of the preceding and with two of the succeeding chambers. (These communications are not displayed in the sections figured, but may be readily seen in the protoplasmic casts of decalcified specimens.) The canal system is well developed and resembles that of Operculina, a marginal plexus being present here also. In the sample of sand from the Maldive Islands, above men- tioned, the great majority of the specimens of Heterostegina range from a small size up to about 4 mm. in their larger diameter (Fig. 56, A), but a few far exceed the rest, attaining a diameter of 10 or more mm. (Fig. 56, B). I am unable to recognise any differ- ence in the external appearance of the two forms, beyond that in size, and the peculiar shape of the large specimens caused by the greater width of the terminal convolution. On making sections of the tests it is found that the large specimens are microspheric and the smaller ones megalospheric. In two specimens of the former m = 27 p in both, and a spiral of some 36 simple chambers succeeds before the septa appear, dividing the chambers into chamberlets (Fig. 56, B'). In the megalospheric form M varies in four cases from 70 to THE FORAMINIFERA 129 FIG. 56. Heterosteguia depressa, d'Orb. A and B the megalospherie and microspheric forms, x 8. A', B', central regions of sections of these tests in the median plane, x 50. The figures 14 and 38 mark the first of the subdivided chambers, in the two forms respectively. The canal system is barely indicated in this figure and in figure 5.1). 1 30 THE FORAMINIFERA 80 /A, and the number of single chambers is 9 to 12. In some cases the chamber which succeeds the megalosphere is consider- ably larger than those which immediately follow (Fig. 56, A'). We thus have in this sand from the Maldives a difference in size between the two forms of Heterostegina similar to that found in the Nummulitic formations of the Eocene period.1 In the third genus of the series, represented by the species Cycloclypeus carpenteri, the great majority of the individuals do not exceed 12 mm. in diameter, but some attain the large size of 64 mm. It is very probable, from analogy with other genera, that the specimens which attain the large size are microspheric ; those of smaller size are, in the specimens which I have examined, megalo- spheric. The only specimen (Figs. 57, B, and 58) of the microspheric form which I have examined is a section.2 In it the microsphere measures 29 ^, and is followed by 9 single chambers arranged in a spiral. The chambers then become subdivided, as in Heterostegina. After being disposed at first in a spiral, they gradually extend round a larger and larger part of the circum- ference of the test until they completely encircle it, and the arrangement becomes annular. The twenty -fifth chamber from the microsphere is, in this specimen, the first to complete the circle. In the megalospheric form the centre is occupied by a structure somewhat resembling the " primitive disc " of Orbitolites complanata (p. 106). It is, however, differently constituted. The megalo- sphere is very large, its average mean diameter in 9 cases being 245 /A, and the extremes 465 and 175 /A. It communicates by a narrow neck with a large chamber which is applied to the megalo- sphere for about half its circumference, and communicates in turn with another large chamber. 1 Chapman (10, p. 19) believes that he has found the dimorphic forms of Hetero- stegina, and identifies them with the biconvex and the compressed varieties described by Brady. He finds that the size of the full-grown megalospheric test is greater than that of the microspheric, a result which he recognises as unusual. The relative sizes of M and m are said to be in the proportion of 3 : 2, and I learn from Mr. Chapman, by letter, that the actual diameters were 125 and 65 /* respectively. These results are so far at variance with the phenomena of dimorphism in general, and with my own in this species, that it appears probable that the individuals with the smaller central chamber were megalospheric specimens with rather smaller megalospheres, and that Mr. Chapman did not meet with the microspheric form. 2 I have to thank Professor J. W. Judd for the opportunity of examining and figuring this section. The specimen was obtained at Funafuti, in the Pacific, and the section was prepared by Mr. Chapman, and figured by him (10, PI. III. Fig. 2) on a small scale. In this paper the specimen is regarded as an unusual example of the megalospheric form, but I understand, by letter, from Mr. Chapman that he is now inclined to reconsider this view. Fig. 58 is prepared from a photograph of the central region, on a larger scale, and Fig. 57, B, shows the arrangement of the chambers more clearly. THE FORAMINIFERA The arrangement of the chambers which succeed varies in different specimens. In some (as in Fig. 57, A) a succession of about six subdivided " heterostegine " chambers follows, which Fio. 57. Cydodypem carpenteri, Brady, m, a complete test, x 8. A, central region of a decalcified specimen of the megalospheric form, x 85. B, central region of a section of the test of the microspheric form, x 50. The canal system is not seen in these figures. The figures 1 and 9 in B mark the first and last of the undivided chambers, 20 is a heterostegine chamber, and 25 the first of the annular chambers. become more and more embracing, until they extend completely round the previously formed chambers. The annular arrangement once attained is continued, though not without irregularities of growth, till the test is complete. The mode of communication of 132 THE FORAMIN1FERA the chamber-lets with one another is similar to that described in Heterostegina. The variations on this arrangement which occur result from a more speedy attainment, in different degrees, of the cyclical growth. Where the nuclear characters have been recog- nised, a single large nucleus was found in one of the large central chambers of the megalospheric form. Fio. 58. Photograph of the central part of the section of Cydodyptus carptnttri (microspheric) repre- sented in Fig. 57, B. Looking back on the evidence furnished by these three genera we find that Operculina is built on the same plan in both micro- spheric and megalospheric forms ; that Heterostegina repeats the operculine condition in both forms, though the number of un- divided chambers is greater in the microspheric form than in the megalospheric ; and that Cyclodypeus repeats both the operculine and heterostegine conditions in the microspheric form, while in the megalospheric the operculine stage is omitted or represented THE FORAMINIFERA 133 only by the two large chambers which succeed the megalosphere, and the heterostegine stage is considerably shortened. In fact, we find the same tendency in the megalospheric form to abridge or omit the stages repeated by the microspheric form as we have seen in other cases. The genus Fusulina is represented by a series of forms which abound in the Carboniferous and Permian rocks in Russia, North America, Sumatra, and elsewhere. By their perforate walls, their bilateral sym- metry about a median plane, and the charac- ter of the aper- ture, which is a A c Fio. 59. Forms of the genus Fusulina from the Carboniferous formation of Russia, Brady, Ann. and Mag. of Nat. Hist. 4, xviii. p. 414. slit left between the margin of the septum and the surface of the preceding convolution, they appear to belong to the Num- muline stock. Like the species of Alveolina they present varying degrees of elongation in the direction of the spiral axis from the biconvex discs of F. aequalis (Fig. 59, A) to the fusiform tests of F. cydindrica (D). The megalospheric and microspheric forms of Fusulina, have been recognised by Schlumberger (58). We may now take a brief survey of some of the main pheno- mena which have presented themselves in the several groups. Among the species of Foraminifera we meet with modifications of form of three kinds. There is the modification which occurs during the growth of an individual, producing the "multiform" condition of test. There is the difference among individuals dependent on their mode of origin, whether from a megalosphere or a microsphere, which finds its expression in dimorphism. Finally, there is the variation commonly presented to a greater or less extent by animals and plants, the departure of the individual in different degrees from the type form of the species. We may consider these three kinds of modification in the reverse order. The Variation of the Foraminifera. — It has long been recognised by systematists that in many cases the limits of the characters of the species of Foraminifera do not admit of being drawn with any 134 THE FORAMINIFERA exactness. This view was insisted on by Carpenter, who, in the " Challenger " Report on Orbitolites (p. 9), quotes with approval the doctrine that among the porcellanous and vitreous Foraminifera " everything passes into everything else." Carpenter, indeed, held (I.e. p. 8) that "the ordinary notion of species as assemblages of individuals marked out from each other by definite characters that have been genetically transmitted from original prototypes similarly distinguished, is quite inapplicable to the group of the Foraminifera." And again, in the Introduction wo read (8, p. 56) : — " The impracticability of applying the ordinary method of definition to the genera of the Foraminifera becomes an absolute impossibility in regard to species. For whether or not bhere really exist in this group generic assemblages capable of being strictly limited by well-marked boundaries, it may be affirmed with certainty that among the forms of which such assemblages are composed, it is the exception, not the rule, to find one which is so isolated from the rest by any constant and definite peculiarity, as to have the least claim -to rank as a natural species." The question, however, appears to be not whether all inter- mediate terms do or do not exist between dissimilar forms, but whether the whole body of forms, as they occur in nature, tend to group themselves, or are aggregated about certain centres. If this is the fact, and the forms, as they occur in nature, are disposed not in a continuous series, but in a discontinuous one, the large number of individuals being grouped about distinct centres, we have ,the phenomenon which is comparable with that of species in other animal* and in plants, whether the centres are or are not connected by intermediate terms. To refuse to recognise the existence of these centres, because transitional forms exist between them, is to ignore an essential fact. In a very large number of cases, at any rate, such centres do exist among the Foraminifera, as among other organised beings, and the characters of the middle individuals of them are those of the species. The dimorphism of Foraminifera depends, as we have seen, on two modes of reproduction, which recur in a cycle of generations. The megalospheric form arises by the multiple fission of a single parent, while there are strong grounds for concluding that the microspheric form arises from a zygote, formed by the conjugation of zoospores. The phenomena of dimorphism are exhibited in the size of the initial chambers, in the nuclear characters, in the mode of reproduc- tion, and, often, in the plan of growth. In most of the species of Foraminifera in which we have evidence of the sizes of the initial chambers, they are strongly contrasted in the two forms, although THE FORAMINIFERA 135 in some, as in Pcneroplis, the size of the megalosphere may, in exceptional cases, fall below that of the microsphere. In this genus, as we have seen, the microspheric form is also to be dis- tinguished by the absence of the spiral passage. In Discorbina and Truncatulina there is no such structural feature to distinguish the two forms, nor are they always to be recognised by the size of the central chambers. There is reason to believe, however, that they differ in nuclear characters, and mode of reproduction. Whether or not the two modes of reproduction prevail through- out the simpler forms of Foraminifera cannot at present be stated. The Multiform Condition. — The significance of this condition is one of the most interesting problems presented by the Foraminifera. Perhaps the simplest case of its occurrence is that of Polytrema (p. 120). We have seen that in the earliest stages of life this organism is free, and secretes a test which resembles in many of its features that typical of the Rotalidae. After it has become adherent the rotaline mode of growth is exchanged for one adapted to the attached habit, and the test assumes an encrusting or arborescent form. In the case of Polytrema, then, it seems clear that the arrange- ment of the chambers formed early in life repeats that of the rotaline stock from which it sprang, while the later chambers are disposed on a plan acquired as it has diverged from that stock. Again, the more complex members (Orbitolites and Cydodypeus) of the Peneroplis-Orbitolites and Operculina-Cydodypeus series present excellent examples of the multiform condition. The facts that each of these is a series of closely related genera, and that the simpler members of each present in a permanent form the arrange- ment which is transitory in the growth of the more complex, appear to give substantial support to the view urged by Car- penter that the stages which we have called peneropline and •orbiculine, operculine and heterostegine, in the growth of Orbitolites and Cydodypeus respectively, are, in fact, repetitions in ontogeny of a phylogenetic history. The application of this explanation to the multiform Miliolinidae appears less satisfactory because the earlier (quinqueloculine) plan of growth is somewhat more complex than the later, and we should not therefore expect it to be the more primitive. We need not assume, however, that the course of development has always been in the direction from simple to complex. Closely connected with this question is the fact that the multi- form condition is, as we have seen, much more pronounced in the microspheric than in the megalospheric form of a species. In a former paper I suggested (21) that a partial explanation of the I36 THE FORAMINIFERA contrast may be found in the difference in the mode of origin of the two forms. The life-history of the Cladoceran Leptodora hyalina, appears to offer a similar contrast. Throughout the summer months broods of young are produced, which develop parthenogenetically and are hatched in the form of the parent. The resting winter egg, on the other hand, which develops as the result of fertilisa- tion, emerges as a Nauplius larva — the form in which the members of such diverse families take their origin, and which, there is good reason to believe, repeats in several of its features the characters of the primitive Crustacea. In the case of Leptodora we see that after, and apparently as the result of, fertilisation the organism " casts back " in its develop- ment, repeating primitive features which are abbreviated or absent in the development of the form arising without a sexual process. Now although the megalosphere of the Foraminifera, the product of the multiple fission of the parent, may not be strictly comparable with the unfertilised egg of Leptodora, it has, at least, this in common with it, that it arises asexually, while it is probable that the microspheric form arises from the conjugation of gametes, a process comparable to the fertilisation of the Metazoa. In the paper referred to it was suggested that the accentuation of the multiform character of the microspheric form of the Foraminifera, as compared with the megalospheric, is likewise dependent on the process of fertilisation. It still appears to me possible that the explanation may be found in the direction indicated, but that this is not the complete solution is shown by consideration of the Initial Poly- morphism displayed by the megalospheric forms of several species. In Idalina and OrUculina we have seen that the extent to which the phases of growth which occur in the development of the microspheric form are repeated by the megalospheric form varies in different individuals, and that it is correlated with the size of the megalosphere — individuals with small megalospheres repeating these phases more completely than those with large megalospheres. What the cause of this correlation may be appears entirely obscure, but it is evident that if among the megalospheric forms, arising asexually, the completeness of the repetition of the earlier phases depends on the size of the central chamber, we are not at liberty to refer the completeness of their repetition in the microspheric form wholly to its sexual origin. In his sketch of a natural classification of the Foraminifera (36 and 37) Rhumbler takes altogether different views of the phenomena we have been considering, and the classification proposed as the result has been adopted by Lang in the new edition of his Lehrbuch. THE FORAMINIFERA 137 In Rhumbler's view " Festigkeitsauslese," the selection of the forms of test best adapted to resist mechanical stress, is regarded as the chief factor which has dominated the differentiation of the Foraminifera, and several series of genera, such as the Nodosaria-Cristellaria and the Biloculina-Quinqueloculina series, are given as examples of a " Festigkeit- skala " in which varying degrees of resisting power have been attained. In the biformed and triformed tests the early chambers are regarded as arranged on a higher (i.e. more resisting) plan than those added later, and hence it is concluded that in the ontogeny of the Foraminifera the order of the appearance of the more primitive and the later acquired characters is the reverse of that so general in the development of other animals, the earlier arrangement representing the form towards which the race is advancing, the later retaining the characters which will ulti- mately be discarded. This reversal of the usual order is attributed to the great delicacy of the young test, to compensate for which a more compact arrangement of the chambers has been acquired. In the later stages of growth, owing to the larger bulk of the protoplasm, the chamber walls can be secreted of such a thickness as to counterbalance the mechanical weakness of their arrangement. The contrast in the modes of growth of the megalospheric and micro- spheric forms is similarly explained, the small size of the latter in the early stages of growth calling for an arrangement, which is less urgently needed in the later stages, or by the megalospheric form. In the more perfected genera, however (as Quinqueloculina), the tests of the forms of both generations are moulded on the most compact type. Thus Rhumbler, like Carpenter, regards the multiform tests of Foraminifera as of great value in tracing out phylogeny, but for precisely opposite reasons, for while Carpenter considers the early phases as repre- senting a stage through which the stock has passed, Rhumbler sees in them the higher stage towards which it is advancing. As will be gathered from what has gone before, it does not appear to me that sufficient reason has been shown for discarding the view of Carpenter. Another remarkable phenomenon met with among the Fora- minifera is that of Isomorphism. It may be denned as the occurrence under similar external forms of species belonging to distinct stocks. Perhaps the most striking instance of it is presented by the Miliolidea and Nummulitidea. It has been pointed out how in the latter family the series Operculina, Heterostegina, Cydodypeus runs parallel with the Peneroplis, Orbiculina, OrUtolites series of the Miliolidea, and we have seen that in Heterostegina, as well as in Polystomella and other allied genera, the tests are to some degree extended in the spiral axis owing to the equitant character of the chambers. The resemblance between the corresponding terms in the two series is rendered all the more remarkable by considera- tion of the forms included in the genus Fusulina, which, at the 138 THE FORAM1NIFERA period when the Carboniferous and Permian rocks were deposited, had undergone elongation in the direction of the spiral axis and been differentiated into a series of forms — biconvex, obovate, spheroidal, and fusiform — closely comparable with those assumed by the species of Alveolina of the Tertiary period and the present day (cp. Figs. 42 and 59).1 Distribution. — From this point of view the Foraminifera may be divided into two classes — the attached or bottom-living and pelagic forms. While by far the greater number of genera and species belong to the former, the numbers of individuals of the latter are enormously great. The Pelagic Foraminifera belong to the genera Globigerina (with its connected form Orbulina), Hastigerina, Pullenia, SpJiaeroidina, and Candeina — forming the order Globigerinidea — and Pulvinulina and Cymbalopora among the Rotalidea. The pelagic habit of these forms, though it had previously been recognised by M 'Donald and Major Owen, was first clearly established by the naturalists of the Challenger. The species which are found at the surface extend down to considerable depths, but whether they may actually live on the bottom of the ocean is still, in spite of much discussion, undecided. They congregate at the surface at night, and partially withdraw from it during the day. It is in the equatorial and temperate regions of the ocean that they most abound, the pelagic forms being represented in the arctic and antarctic seas by small species of Globigerina : G. pachyderma in the former, G. dutertrei in the latter. Beneath this equatorial belt of warm water and its northern offset, the Gulf Stream, the empty tests of the pelagic Foramini- fera constitute the main portion of the " Globigerina ooze," which forms the ocean floor down to a depth of 3000 fathoms. As this limit is approached the thinner tests disappear, and beyond it all calcareous constituents are removed by solution.2 The species of bottom-living Foraminifera have, on the whole, a very wide distribution. Some are cosmopolitan, ranging from arctic to tropical waters and from shore pools to the bottom of the great oceans. A large proportion of the genera, however, are restricted by depth and temperature. The shallow littoral waters of the tropics contain an abundant fauna, most of the members of which do not extend to colder seas. On the other hand, genera 1 The Lituolidea are described as " isomorphoua " with various calcareous genera, but it is far from certain that the similarity in form does not depend on true affinity, in which case the term is not strictly applicable. Loftusia, among the Lituolidea, has a fusiform test, externally resembling the more elongated forms of the Fusulina and Alveolina series. a Cp. Sir John Murray, "On the Distribution of the Pelagic Foraminifera," etc. Natural Science, July 1897, pp. 17-27. THE FORAMINIFERA 139 (such as those of the Astrorhizidea) which abound in the arctic seas extend, as members of the abyssal fauna, along the ocean floor, to mingle in lower latitudes with the empty tests of the pelagic inhabitants of the warmer surface waters. Notwithstanding the wide range of many species, there is some indication of a limitation of forms to definite areas — the formation of local faunas — comparable to that met with in the distribution of other animals and of plants. Thus in the warm shallow seas of the Malay Archipelago Mr. Millett finds the forms deviate in many instances from the ordinary structure of the Foraminifera.1 In his reports hitherto published, dealing with the orders as far as and including the Lagenidea, he has described twenty-six new species and one new genus from this region. Geological Distribution. — Representatives of four orders of the Foraminifera — Textularidea, Lagenidea, Rotalidea, and Globi- gerinidea — have been recognised in the Cambrian, the oldest of the ," palaeozoic" formations. In the Carboniferous all the orders are represented except the Miliolidea — which have,however,been recog- nised in beds transitional between the Carboniferous and the Per- mian— the small and fragile Chilostomellidea, and the Gromiidea, whose slight tests we should hardly expect to find preserved. In the Carboniferous formation species of Saccammina and Fusulina give rise to extensive deposits. An abundant foraminiferal fauna has been found in many of the secondary formations, and the chalk of the later Cretaceous period is in la^ge part built up of their tests, Globigerina being an abundant form as in the oozes of the existing ocean basins. Foraminifera also enter largely into the composition of the earlier rocks of the Tertiary period. The Miliolidea here come into great prominence, and are represented by Miliolina (including Quinqueloculina and Triloculina\ and its allies Peneroplis, Orbitolites, and Alveolina. Nummulites, which had already made their appearance in the Carboniferous period, also abounded in the warm shallow Eocene seas, and the Nummulitic limestones extend across the old world from the Pyrenees to China, attaining in some places thousands of feet in thickness. It need hardly be pointed out that our knowledge of the life- history of the Foraminifera is still very far from complete. In the establishment of the prevalence throughout the higher groups of the phenomenon of dimorphism, dependent on different modes of reproduction, a substantial groundwork has been attained, but there remain many important questions of wide biological bearing on which we are very imperfectly informed. 1 Keport on the Recent Foraminifera of the Malay Archipelago collected by Mr. A. Durrand, F.R.M.S., Journ. Roy. Microscopical Soc. 1898, p. 258. 140 THE FORAMINIFERA There seems good reason to hope that the study of the plan of growth of both forms of the species during the early stages of their life-histories may throw light on the complicated problems of phylogeny. Until these early stages have received fuller attention, and we have arrived at a conclusion as to the relation of the early to the later stages of the multiform tests, efforts at forming a " natural classification " appear to be premature. The classification adopted by Brady in his Challenger Monograph is given here, with slight modifications. I have followed Neumayr (30) in placing the Astrorhizidea before the Miliolidea as they appear to be more primitive forms. The Cycloclypeinae are merged with the Nummulitidae for the reasons given above. It appears highly probable that the Lituolidea should be dis- tributed (as Biitschli has done) among the calcareous forms which they resemble, but they are here left as arranged by Brady. In conclusion I desire to express my thanks to my brother Mr. W. T. Lister, for his assistance in preparing the photographs of the shells of Foraminifera with which this article is illustrated. They were done with one of Zeiss' admirable instruments. ORDER 1. Gromiidea. Test membranous, chitinous, or siliceous ; smooth or encrusted with foreign bodies ; with one or more pseudopodial apertures. FAMILY 1. POLYSTOMATIDAE. Test with one or many openings. Genus Myxotheca, Schaudinn (Fig. 2). Test encrusted, openings many ; marine. Here may be provisionally placed Hyalopus, Schaudinn ( = Gromia dujardinii, M. Sen.). Test smooth, rounded, with one opening (Fig. 15), or branched, and with many ; pseudopodia hyaline, with few or no anastomoses ; multinucleate ; marine. FAMILY 2. MONOSTOMATIDAE. Test rounded or flask -shaped, with a single opening. (a) Test smooth. Genera — Gromia, Duj. Test chitinous, usually flexible, mouth terminal ; freshwater and marine (Fig. 1). Lieberkuhnia, Clap, and L. Test very delicate, ovoid ; mouth sub-terminal. Mikro- gromia, K. Hertw. (Cystophrys, Archer) (Fig. 14). Test small, rigid, flask- shaped, bilaterally symmetrical, not filled by the protoplasm ; pseudopodia springing from a short stalk of protoplasm ; individuals often united by their pseudopodia into colonies. Platoum, F. E. Sch. Similar to Mikro- gromia, but test more pointed. Lecythium, H. and L. Similar, but protoplasm filling the test. These four are freshwater genera. (6) Test encrusted with foreign bodies. Genera — Pseudodifflugia, Schlumb. Resembles Gromia, but test encrusted ; fresh and brackish water. Diaphoropodon, Archer. Test ovoid, built of loosely-united foreign bodies. Pseudopodia of two kinds : long, extended from the mouth ; and short, hair-like (? true pseudopodia) between the particles of the test. (c) Teat built of chitinous or siliceous plates. Genera — Euglypha, Duj. (Fig. 3). Test elliptical or pear-shaped, with terminal mouth ; built of THE FORAMINIFERA 141 circular or hexagonal siliceous plates ; pseudopodia without anastomoses ; freshwater. Trinema, Duj. Similar, but mouth on flattened lateral surface ; freshwater. Cyphoderia, Schlumb. Test flask-shaped, built of small chitinous plates ; fresh and brackish water. Campascus, Leidy Resembles Cyphoderia, but test encrusted ; freshwater. FAMILY 3. AMPHISTOMATIDAE. Test with an opening at either end. (a) Test smooth. Genera — Diplophrys, Barker. Test roundish, very delicate ; freshwater, and in manure. Ditrema, Archer. Similar, but test thicker ; freshwater. Shepheardella, Siddall (Fig 16, 5). Test chitinous, long, and tubular, contracted to an opening at either end ; marine. (6) Test encrusted. Genus — Amphitrema, Archer (Fig. 16, 1 1). Barrel- shaped, test produced to a short neck round either opening ; moor pools, Ireland. ORDER 2. Astrorhizidea. Test invariably composite, usually of large size and monothalamous ; often branched or radiate, sometimes segmented by constriction of the walla, but seldom or never truly septate ; polythalamous forms never symmetrical. FAMILY 1. ASTRORHIZIDAE. Walls thick, composed of loose sand or mud, very slightly cemented. Genera — Astrorhiza, Sandahl (Fig. 17, a). Test fusiform, or depressed and more or less stellate, and attaining a diameter of nearly one inch. Pelosina, Brady. Storthosphaera, F. E. Sch. Dendrophrya, Str. Wright. Syringammina, Brady. Test consisting of masses of branchings and anastomosing tubes. FAMILY 2. PILULINIDAE. Monothalamous, wall thick, composed chiefly of felted sponge spicules. Genera — PiluUna, Carpenter (Fig. 17, c). Nearly spherical. Technitella, Norm. Oval. Bathysiphon, Sars. Tubular. FAMILY 3. SACCAMMINIDAE. Chambers nearly spherical, walls thin, firmly cemented. Genera — Saccammina, M. Sars (Fig. 17, &). Globular, with distinct projecting aperture. Psammosphaera fusca, F. E. S., without a projecting aperture. Is regarded by Rhumbler as the young form of Saccamminina sphaerica. Sorosphaera, Brady. Many spherical adherent chambers, each with its own aperture. FAMILY 4. RHABDAMMINIDAE. Test firmly cemented, of sand, often with sponge spicules intermixed, tubular, straight, radiate or branched, rarely segmented. Genera — Jaculella, Brady. Elongate, tapering. Hyperammina, Brady (Figs. 17, d, and 18). Elongated, tubular, simple or branched, sometimes commencing in a globular chamber. Marsipella, Norman. Fusiform or cylindrical Rhabdammina, M. Sars. Rectilinear, radiate or branching, with or without a central chamber. Aschemonella, Brady. Test of inflated sacs, single or combined in series. Rhizammina, Brady. Fine chitino-arenaceous tubes, simple or branched. Sagenella, Brady. Adherent, branching tubes. Botellina, Carpenter. Test cylindrical of loose sand, with irregular cavities. Haliphysema, Bk. Test columnar, attached by a stalk, simple or branched, beset with sponge spicules (Fig. 19). 142 THE FORAMINIFERA ORDER 3. Lituolidea. Test arenaceous, usually regular in contour, chambers of the poly- thalamous forms frequently labyrinthic. Comprises sandy isomorphs of the simple porcellanous and hyaline types (Cornuspira, Peneroplis, Lagena, Nodosaria, Cristellaria, Globigerina, Rotalia, Nonionina, etc.), together with some adherent species. FAMILY 1. LITUOLIDAE. Test of coarse sand grains, rough ex- ternally ; often labyrinthic. (a) Chambers non-labyrinthic. Genera — Reophax, Montf. Test free, composed of one flask-shaped chamber, or of several united into a straight, curved, or irregular line, never spiral. Coskinolina, Stache ; Haplophragmium, Reuss. Test free, nautiloid, or crosier-shaped. Placopvilina, d'Orb. Chambers plano-convex, adherent. (6) Chambers labyrinthic. Genera — Haplostiche, Reuss. Test free, uni- serial, never spiral. Lituola, Lamk. Test free, nautiloid, or crosier-shaped. Bdelloidina, Carter. Adherent. FAMILY 2. TROCHAMMINIDAE. Test thin, composed of minute sand grains incorporated with calcareous or other cement ; smooth, often polished externally. Genera — Thurammina, Brady. Test a single sub- spherical chamber. Hippocrepindj Parker ; Hormosina, Brady. A rounded chamber, or several in a straight or curved series. Ammodiscu$, Reuss. Test non-septate, coiled in a piano-spiral (resembling Spirillina) or otherwise. Trochammina, P. and J. Free or adherent, rotaliform, nautiloid or trochoid. Carterina, Brady. Test rotaliform, constructed of fusiform spicules, said to be proper to itself. Webbina, d'Orb. One or more adherent, stoloniferous chambers. FAMILY 3. ENDOTHYRIDAE. Fossils. Test more calcareous and less sandy than in other Lituolidae, sometimes perforate. Genera — Nodosinella, Brady. Nodosariform. Polyphragma, Reuss. Involutina, Terq. Endothyra, Phil. Bradyina, Moll. Stacheia, Brady. FAMILY 4. LOFTDSIIDAE. Test relatively large, lenticular, spheri- cal or fusiform ; arranged spirally, or in concentric layers ; walls finely arenaceous and cancellated. Genera — Cyclammina, Brady. Nautiloid. Loftusia, Brady. Large, resembling Alveolina in contour. Parkeria, Carp. Large, spheroidal ORDER 4. Miliolidea. Test usually imperforate, normally calcareous and porcellanous, some- times encrusted with sand ; under starved conditions (e.g. in brackish water) becoming chitinous or chitino-arenaceous ; at abyssal depths occasionally consisting of a thin, homogeneous, imperforate, siliceous film. FAMILY 1. MILIOLINIDAE. Test of one or many chambers spirally arranged ; in the many - chambered forms there are usually not more than two chambers in each convolution, (a) Test unsegmented, plano- spiraL Genus — Cornuspira, M. Sch. (Fig. 20). (6) Test piano-spiral, two chambers to a convolution. Genera — Spiroloculina, d'Orb. (Fig. 21). All the chambers exposed on the contour. Biloculina, d'Orb. (Figs. 22, a, and 24). Chambers simple, only the last two chambers exposed on the THE FORAMINIFERA 143 contour. Fabularia, Def. Similar, but chambers subdivided in the interior. Sigmoilina (Planispirina, pars), Schlumb. (c) Three chambers exposed on the contour of the test. Genus — Triloculina, d'Orb. (Figs. 22, 6, and 25). (d) Five chambers exposed on the contour of the test. Genus — Quinqueloculina, d'Orb. (Fig. 23). (e) In the megalospheric form the second chamber completely invests the megalosphere. Genus — Adelosind, d'Orb. Chambers arranged on the quinqueloculine plan, or (in A. polygonia, Schlumb., Fig 26), plano-spirally, three or four to a convolution. (/) Earlier chambers quinqueloculine, later spiroloculine. Genus — Massilina, Schl. ( like other Protozoa, a single nucleated cell. °v Spyiofit Many of the Gregarines are quite visible to the cieus. (After van Bene- naked eve. and PorospoTd oiaantea (v. Ben.) from den, from Lankester.) "h ' 1 Vi f 1 fi t thirds of an inch (Fig. 1). In spite, however, of the extremest diversity in size, appearance, organisation, and life -history, the Sporozoa as a group possess certain very characteristic features in common — peculiarities which are clearly in direct relation with their habit of life as internal parasites. In the first place, their nutriment is always of a fluid nature, consisting of the juices of the host absorbed osmotically at the surface of the body of the parasite, and none of the special organs 152 THE SPOROZOA for ingesting or digesting solid food, so frequent in other Protozoa, are ever found in this group. Many Sporozoa possess flagella during certain phases of the life -cycle, and many exhibit the power of executing amoeboid movement and emitting pseudopodia during even the whole period of growth ; but in both cases the flagella or pseudopodia are organs of locomotion and not of nutrition, except perhaps in so far as the latter may contribute to an increase of the absorptive surface of the body. More usually all such locomotor organs are absent, and the body of the parasite has a fixed form and definite contours, limited ex- ternally by a cuticle of greater or less thickness, through which food is absorbed by diffusion. Food - vacuoles or contractile vacuoles are never found. In the second place, the Sporozoa always possess the power of rapid multiplication by sporulation that is to say, by the formation of reproductive bodies or germs, each a fragment of the parent body, in the form of a nucleated protoplasmic corpuscle, usually very minute. These germs may serve for increasing the numbers of the parasite within the same host, or may be the means of dis- seminating the species and infecting other hosts. In the latter case the germs are usually provided with protective envelopes which enable them to leave the body of the host in which they were produced and to endure for a season the vicissitudes of the outer world. In some cases the protoplasmic germs are naked gymno- sporeSj and all those derived from one parent are then enclosed in a resistent cyst, formed by the parent previous to sporulation. But in most cases the germs have their own special protective envelopes, and are then termed chlamydospores, or more usually spores simply. Within the spore-envelope a further multiplication of the germs may take place, and a cyst enclosing all the spores de- rived from a common parent may or may not be formed. Resistent spores of this kind are one of the most characteristic features of this class, as the name Sporozoa implies. Only in the com- paratively small number of cases in which infection is conveyed from one host to another by an intermediate host, are protective envelopes wanting. The bulk of our knowledge of the Sporozoa is of extremely recent date, and great advances have been made during the last ten years in the investigation of these organisms and the elucidation of obscure points in their life -history. Nevertheless, they did not entirely escape the observation of the earlier naturalists, even so far back as the eighteenth century. As might have been expected, attention was directed first to the larger forms of Gregarinea inhabiting Arthropods, especially insects, and later to the characteristic spores, often to be found in vast numbers in various animals. The first notice of a Gregarine parasite is attributed to the famous THE SPOROZOA 153 anatomist Redi (1708), but his claims to this honour are very doubtful.1 Cavolini, however, in 1787, described and figured an indubitable Gregarine (Aggregata conformis (Dies.), fide Labbe") from the glandular appendages of the stomach of Pachygrapsus marmoratus. He found conjugating individuals, and believed each such pair to be a kind of tapeworm with two segments. But the true discoverer of the group, in the scientific sense, was Le"on Dufour, who, in his researches upon insect -anatomy, became acquainted with, and described, numerous species of these parasites. He regarded them as a peculiar group of worms, allied to Trematodes, to which in 1828 he gave the generic name Gregarina. More species were subsequently made known by other authors, and in 1839 Siebold published an important work in which he described the nucleus accurately for the first time, without, however, recognising the true nature of Gregarines, which he also considered as worms, though he did not attribute to them an alimentary canal, as had been done by one of his predecessors ! Siebold also described the cysts and spores found associated with the Gregarines, and though he did not discover the connection between them, his observa- tions had the merit of drawing attention to the " pseudonavicellae " already observed by Henle (1835) and others in the sperm-sacs of the earth- worm. Contemporaneously with Siebold's work appeared the investigations of Hake upon the spores of the Coccidium of the rabbit, which, however, the author regarded as pathological products of the tissues of the host itself. In 1841 the celebrated Johannes Muller described the spores of a number of different Myxosporidia inhabiting various fishes, and termed these organisms "psorosperms,"2 a name of very frequent occurrence in sporozoan literature, applied to various kinds of spores. Muller was, however, quite in the dark as to the nature of his psorosperms, and considered them a "living seminium morbi" comparable to spermatozoa. After Muller psorosperms were studied by many observers, and generally divided into "egg-shaped psorosperms," e.g. Coccidia, and " fish - psorosperms " or " Miiller's psorosperms," the spores of Myxosporidia. Their affinities remained, however, uncertain for a very long time, and indeed the true nature of " fish-psorosperms " has only been elucidated completely in the most recent times. As long ago as 1842 Creplin compared psorosperms to pseudonavicellae, and so laid the foundation of the " Gregarine-theory " of the Myxosporidia. But this comparison was not universally accepted, although supported by Leydig, Lieberkiihn, and other observers. Many authors, on the other hand, regarded psorosperms as organisms of a vegetable nature, allied to Diatoms. A distinct epoch in our knowledge of the Sporozoa was made by Kolliker, who in 1845 and 1848 not only greatly increased our know- ledge of these parasites, and of their wide distribution and occurrence in hosts of all classes, but further expressed and maintained for the first time the opinion that Gregarines were unicellular organisms, which should be 1 See Biitschli, " Sporozoa " in Bronn's Thierreich, i. p. 480, from whom most of the historical facts here put together are taken. Labbe [4] identifies the Gregarine figured by Redi as Aggregata praemorsa (Dies.). 2 Derived, according to Balbiani, from \f/upa, mange, and crir^yia, seed. 154 THE SPOROZOA classed amongst Siebold's Protozoa, and identified Siebold's vesicle as the cell-nucleus. His views were still further borne out by the important observations of Stein, who in 1848 first demonstrated clearly the relation of the pseudonavicellae to the reproduction of the Gregarines, which he placed as a class Symphyta of the Protozoa. The views of Kolliker and Stein have gradually obtained universal assent, especially after the demon- stration by Lieberkiihn in 1855 of an amoeboid phase in the life-history, and no one now doubts the position of Sporozoa amongst the Protozoa. Nevertheless, for some years this view was energetically combated by various authors, who could not bring themselves to regard the Gregarines as adult, independent organisms. Chief amongst .the opponents of the Protozoan theory were Henle, Bruch, and Leydig, who believed that Gregarines were in some way connected with the embryonic stages of Nematodes or threadworms, and more particularly of the genus Filaria. In the course of time, and with increase of knowledge, this theory died a natural death, and it became evident that any associations of Gregarines and Nematodes, or resemblances between them, were of a purely accidental and superficial kind. Looking back, however, upon these controversies, now only of historical interest, it is not a little remarkable that in very recent times a curious nematode-like Sporozoon (Siedleckia nematoides, Caull. and Mesn.) should have been discovered, which, had it been known in the fifties, might have inclined the balance of zoological opinion strongly over to the side of the Nematode theory. A retrospect of the history of our knowledge of , Sporozoa further brings into prominence the fact that, as an obscure group of no obvious practical importance, they did not for a long time appeal to the considera- tion of the " common-sense " Englishman. Until comparatively recent times, practically the only contributions to sporozoan literature in this country were those of Lankester, who, besides other forms, discovered in 1872 the organism, parasitic in the blood of the frog, which at a subse- quent date was named by him Drepanidium ranarum. This discovery, and that of Laveran, who a few years later made known to science the malarial parasites of human blood, laid the foundations of our knowledge of the Haemosporidia, a group of such importance, from the practical point of view, that they have been the cause of focussing the attention of medical men, no less than of zoologists, in all countries upon the Sporozoa. Indeed, so great is the interest which these parasites excite at the present time, on account of their pathogenic properties in man and beast, that now scarcely a month passes without the publication of some discovery relating to them, and the study of the Sporozoa bids fair to assume in the near future a position of importance scarcely secondary to that held by the science of bacteriology. The Structure and Life-history of a Typical Sporozoon. — As an example of the Sporozoa and of the characteristic features of their life-cycle, we select for detailed description the common Monocystis agilis, Stein1 (Fig. 2), a Gregarine parasitic in the sperm-sacs 1 With regard to the proper name of this species, there ia a certain amount of confusion and uncertainty, which is none the less regrettable because of so THE SPOROZOA '55 (vesiculae seminales) of the earthworm (Lumbricus spp.). This species is not only very easily obtained, but is also a very typical example of the class ; hence in describing the various phases of its life-history it will be possible at the same time to introduce and define the terminology to which we shall adhere in the sequel Fio. 2. Trophozoites of Monocystis agilis. a and fc, young individuals showing changes of body- fonn due to contractility, c, an older in- dividual, still enveloped in a coat of sperma- tozoa, (a and b after Stein, c after Lieberkiihn, from Lankester.) Fio. 3. Trophozoites of Monocystis magna, attached to the seminal funnel of Lumbricus. a, young individual ; b, goblet - shaped epithelial cells of the seminal funnel, in which the extremity of the parasite is inserted. (After Blitschli, from Lankes- ter.) for the corresponding stages of other Sporo- zoa. It should be understood, however, that the form which can be selected as most typical of a group is not necessarily the most primitive of its members. From the type chosen we shall have to work backwards to simpler forms, as well as forwards to more complex. The earthworm is infested by various species or varieties of Monocystis; according to Cue" not [13] by no less than seven or eight species, of which four are stated to be of common occurrence, namely, M. magna, A. Schmidt ; M. lumbrici (Henle) ( = M. agilis, Stein) ; M. pilosa, Cue"not ; and M. porrecta, A. Schmidt. The specific distinctness of all these forms cannot be unhesitatingly conceded, but at least two distinct species, probably with several varieties, are generally recognised, and are to be found in almost every worm, viz. M. magna and M. agilis. The two species differ in size and in other specific details of character. M. magna (Fig. 3) is the larger of the two, and occurs attached by one extremity of the elongated body to the epithelium of the seminal funnel, only quitting this situa- tion at the period of conjugation, when it drops off into the sperm-sac. M. agilis (Fig. 2) is found in the interior of the clumps of developing spermatozoa, or floating freely in the sperm - frequent occurrence in the zoological nomenclature of just the commonest or most familiar forms of life, particularly amongst the Sporozoa. According to Labbe [4] the species under consideration should be called Monocystis tenax (Dujardin) ; according to Cue"not [13] its proper designation is M. lumbrici (Henle). We refer to these authors for a discussion of these knotty questions, and retain here the name most generally employed, in this country at least, for the species. 156 THE SPOROZOA sacs. But all essential details of the life-history are quite similar in the two species. The earliest known stage of Monocystis agilis is a minute protoplasmic body, with a distinct nucleus, lodged in one of the " sperm-morulae " floating in the sperm-sac. As is well known, each sperm-morula of the earthworm gives rise to a cluster of spermatozoa, attached by their heads to a central residual mass of protoplasm termed the " sporophore." The young Monocystis is found within the sporophore and grows at its expense and at that of the attached spermatozoa. This stage of the parasite, during which it is absorbing nutriment from its host and growing rapidly, may be termed the trophic stage, and each individual parasite during this stage may be termed a trophozoite. The parasite soon becomes elongated in one direction. It assumes first an oval contour and becomes later more or less vermiform. As it grows it destroys the sperm-cluster in which it is lodged, and in later stages it is found enveloped in an adventitious coat or fur composed of the tails of the degenerated spermatozoa, giving the appearance of a ciliated covering, which is thrown off in the final stages of growth (Figs. 2, c, and 4, a). The full-grown trophozoite (Fig. 2) is still a single cell, with a single nucleus. The body is limited by a distinct cuticle, within which the protoplasm is differentiated into an external clear cortical layer or ectoplasm, and an internal granular medullary layer or endoplasm. The ectoplasm is the seat of contractility, and contains in its deepest part a layer of fine contractile fibres, the so-called myocyte-fibrillae. The endoplasm lodges the nucleus, and contains numerous coarse granules representing nutriment held in reserve for impending reproductive and developmental processes. The nucleus is a clear spherical body, in the form of a vesicle limited by a delicate membrane, containing fluid in which float one or more nuclear corpuscles or karyosomes. Each karyosome is a small globule, resembling in appearance the nucleolus of a tissue-cell, but differing from it in containing a certain amount of chromatin in its substance. The karyosomes usually have a vacuolated structure. The trophozoite is actively motile, as the specific name implies. In Monocystis the movements consist chiefly of changes of form brought about by the contractility of the myocyte-fibrillae, whereby the body may be bent or contracted as a whole, or may exhibit ring -like constrictions in different parts. After the trophic stage, which is a period of purely vegetative growth, the parasite enters upon the reproductive phase of its life-history, a period in which two distinct events follow each other; first, the formation of gametes or conjugating individuals, which pair with one another and unite to form zygotes ; secondly, THE SPOROZOA 157 the formation from the zygotes of the resistent spores, by which the parasite is disseminated (see Fig. 29, p. 185). The adult trophozoite, when it is ripe for reproduction, is commonly known as a sporont, but may be better termed a gametocyte, since it gives rise to the gametes. Two gametocytes come together and become very closely apposed to form a spherical body (Fig. 4), the two individuals remaining, however, perfectly PlO. 4. Association and encystation of Monocystis magnet,, a, a couple attached to the ciliated epithelium of the seminal funnel of the earth- worm. The two gregarines are covered by a furry coat of adherent spermatozoa, b, a couple detached from the seminal funnel and en- veloped by a common cyst-mem- brane. s.f, seminal funnel ; n, nucleus of the parasite ; c.m, double cyst - membrane. (After Cuenot.) a, x37; b, slightly less. distinct from one another, each forming one hemisphere of the common mass. This union of the two gametocytes must not be confounded with the true conjugation : the two individuals are merely in association • they are keeping company, as it were, as a preliminary to the formation of gametes. The two associated gametocytes, which often exhibit a slow rotatory movement, now become surrounded by a common envelope or cyst (Fig. 4, b, c.m), secreted by them in two layers ; first a rigid external epicyst, then a thin internal endocyst.1 Meanwhile important changes are going 1 According to Cecconi [11], in M. ay His the sporonts first become encysted singly, and two such cysts then approach each other and join together. This prob- ably applies only to the first signs of cyst-formation, as two completely encysted gregarines can hardly be suiliciently motile to admit of their travelling towards one another. 158 THE SPOROZOA D. T dir. £fl FIG. 5. Formation of the segmentation- nucleus, and its subsequent division, in sporonts of Monocystis. a, portion of a section through two encysted associated sporonts, separated by their cuticular body-walls (c.w). In the lower half of the section is seen a nucleus in which the karyosomes (ky) are breaking up, while in the nuclear sap near them a number of granules of chromatin have appeared to form the segmentation - nucleus (n.seg). Outside the nuclear membrane a patch of archoplasm (arch) has appeared. In the upper half the archoplasm has divided to form an achromatic spindle (ac.sp), in the middle of which are seen the chromatin granules of the segmentation -nucleus; two karyosomes are also seen, one showing a vacuolated structure. 6, section through one sporont of an associated couple in a cyst, showing the segmentation-nucleus in the diaster stage. The karyokinetic spindle (ac.sp) stretches across the whole body. The chromosomes (chr) form two groups, c, section through a couple of encysted sporonts, showing in the lower one two resting nuclei preparing for division, and the remains of a karyosome in the cytoplasm ; in the upper one two nuclei in the diaster stage, and two karyosomes ; ep, epicyst ; en, endocyst. d, section through one sporont of a couple showing seven resting nuclei (n), one dividing nucleus (?uc), and two karyosomes (ky). (After Cuenot.) a, x 1180 ; b and ) is break- ing up into separate masses, in some of which degenerating residual nuclei (T.JI) are to be found; n/, cyst -en- velope, o, the zygotes (definitive spm-o- blasts)have begun to secrete sporooysts (.«/>.<•), within which the sporophism (s;>.jp) is becoming contracted ; a lew sporoblasts degenerate (*;/.W). Other letters as before. (After Cuenot.) x 1180. THE SPOROZOA 161 ("falciform body," " Sichelkeim "). The protoplasm of the sporo- zoites is finely granular, and when they are formed, a surplus of coarsely granular protoplasm is left over from the sporoplasm as the sporal residuum ("reliquat sporal"). The fully-formed spore has in Monocystis the form shown in Fig. 7, C ; it is more or less boat-shaped, and resembles a diatom of the genus Navicella, whence is derived the name pseudonavicella, by which Gregarine spores have long been known. The sporocyst is slightly thickened at each Fio. 7. Development of the spore of Mono- cystis, A, oval sporoblast with single nucleus (a). B, the sporoblast has secreted the sporocyst at its surface, and the sporoplasm within it has become contracted and diminished in volume. C, ripe spore with eight sporozoites and residual protoplasm (6). D, diagrammatic cross-section to show the arrangement of the sporo- zoites round the central residual pro- toplasm. (After Biitschli, from Lankester.) pole, and within this very resistent and impervious envelope the eight sporozoites are packed lengthways round the centrally placed sporal residuum. During the formation of the spores the cystal residuum is slowly absorbed, and the ripe cyst contains only a great number of the pseudonavicellae, not arranged in any definite pattern (Fig. 8). The above account of the conjugation and spore - formation is that given recently by Cuenot, whose researches confirm the discoveries of Siedlecki with regard to an allied form Lankesteria ascidiae (Lank.), and are in harmony with the still more recent account given by Leger [23] for Stylorhynchus. Cuenot's description of the spore-formation and the events antecedent to it is confirmed in all essential details by Cecconi [11] and Prowazek [2 5 a]. Previous to Cuenot the reproduction of Monocystis had only been studied by Wolters [29], whose description of the process is very different. According to Wolters the association of the two full- grown trophozoites or sporonts within the cyst is a true conjugation, similar in its details to that known in Actinophrys from the researches of Schaudinn. Wolters describes the nucleus of each sporont as dividing mitotically to form two nuclei, one of which is given off in a polar body, while the other remains as a pronucleus. The two nuclei are then stated to travel towards the septum formed by the apposition of the cuticular body-walls of the two sporonts, and at one point the septum becomes dissolved, permitting the fusion of the pronuclei into a single nucleus. After a time the fusion -nucleus divides into two nuclei, which then rapidly divide up to form numerous small nuclei, round which the proto- plasm of the sporonts becomes segmented to form the sporoblasts. From the sporoblasts the spores are formed as above described. It is unfortunate that these statements of Wolters, which seem to be ii 162 THE SPOROZOA totally erroneous, have remained uncontradicted for ten years, during which time they have got into numerous text -books and have been generally accepted and taught The spores of Monocystis do not appear to be able to develop further in the earthworm, but require to be transferred to a fresh host before they can germinate. How the infection is effected has not yet been ascertained in the case of the type Fio. 8. Ripe cyst of Monbcystis, showing the numerous spores (pseudonavicellae) scattered within the cyst, without any cystal residuum present. (From Lankester.) that has been selected for description, and the course of events can only be conjectured by analogy from what is known to take place in other Sporozoa. It is highly probable that the spores pass to the exterior and are scattered broadcast in the earth, and that they are then swallowed accidentally by an earth- worm with its food, and so pass into its digestive tract. The action of the digestive juices upon the spores has the effect of causing the sporocysts to burst open, setting free the sporozoites, which are actively motile and possess the power of boring their way through cells and tissues. In this way the sporozoites prob- ably traverse the wall of the earthworm's intestine and reach the reproductive organs, where each one attacks a sperm-mother-cell THE SPOROZOA 163 and there develops into the minute trophozoite which is formed later within the sperm-morula. With this stage the life-cycle has come round again to the point at which the description of it was commenced. A few points with regard to the life-cycle require brief further dis- cussion. It has been suggested that the spores may sometimes germinate in the host in which they are formed, and so increase the numbers of the parasite within it. But the improbability of this occurring is very great, as was pointed out by Butschli, in view of the relatively small number of parasites in the trophic stage which are met with, as com- pared with the vast number of spores. Thus in a given earthworm there will be found in the sperm -sacs perhaps a dozen trophozoites and as many ripe cysts. Each of the latter contains, however, at a low esti- mate about fifty spores, and each spore eight sporozoites. A single cyst contains, therefore, about four hundred individuals, more or less, and if it were a frequent occurrence for the spores to germinate in the same host, the number of trophozoites in each earthworm might be expected to be vastly greater than is usually the case. Another question which may be raised is whether the Monocystis has any method of multiplication during the trophic phase, that is to say, in the period from sporozoite to sporont. It has sometimes been stated that the trophozoites multiply by division during the earlier stages of growth. From what is known of other Sporozoa, there is nothing inherently improbable in this view, but it has not been proved satisfactorily that such multiplication can take place in Monocystis, and the above- mentioned paucity of the trophozoites is an argument against its occurrence. With regard to the passage of the spores to the exterior, precise information is lacking as to how this is effected. In Sporozoa generally we find one of two conditions. In some cases the spores are produced in a position where they can leave the body by natural channels, as in the numerous instances of sporozoan parasites lodged in the digestive tract, when the cysts and spores are cast out with the faeces. In other cases the spores cannot pass out by natural channels, and are set free either by provoking suppuration or other organic disturbance, or by the death and break-up of the host. In the case of the Monocystis of the earthworm, the spores could only be discharged from the body, in the ordinary course of events, by passing out of the sperm-sac with the sperm at copulation. They would then be transferred to the spermathecae or receptacula semiuis of another worm, and would pass ultimately into" the cocoon in which the eggs are laid ; but there is no record of their occur- rence in either of these situations. It seems more probable that spores are set free by the dissolution of their host. Very possibly birds or some other of the numerous creatures which prey upon worms are the agents by which the dissemination is effected. If a bird swallowed an earthworm containing spores of Monocystis, from which very few worms are free, the spores would probably pass unaltered thorough the bird's digestive tract. Uninjured spores of Gregarines have been observed by 1 64 THE SPOROZOA L. Pfeiffer in the intestines and faeces of various birds.1 A parallel case is that of the Coccidium infecting the centipede Lithobius, the spores of which, if swallowed by another animal, such as a wood-louse, pass unaltered through it (see p. 221). If this suggestion be correct, it is easy to under- stand that any worm-eating bird would be continually scattering spores of Monocystis on the ground, where they would wash down into the soil and be swallowed very easily by worms again. There is, however, no direct evidence bearing upon the mode of dissemination of the spores, and the above suggestion must be regarded merely as a more or less probable surmise. When the spores have reached the digestive tract of their new host, and the sporozoites have been liberated there, the question arises how they reach the sperm-sacs. This problem, however difficult to solve, is by no means one peculiar to the Monocystis of the earthworm. In many other Sporozoa we have instances of parasites affecting some particular organ, which invade the body in the first place from the digestive tract. It must be assumed that the sporozoites have in some way the power of selecting the particular organ they affect, and of migrating through the body of the host in order to reach their specific habitat. Probably they make use of vascular or lymphatic channels in order to arrive at their destination. General Characters of the Sporozoa. — From the above account of Monocystis it is seen that the life-history of a typical Sporozoon is a single cycle, which may be summed up in the following way : 2— Sporozoite-->-Trophozoite-^-Gametocyte (Sporont) x n Gametes \ , Sporozoite->-Trophozoite->-Gametocyte (Sporont) x n Gametes / T =n Zygotes (Sporoblasts)-^-?i Spores x Sn Sporozoites. The life-cycle may further be divided into three main periods. First, the period of growth, during which the minute sporo- zoite grows by absorption of nutriment from the host into the sporont. Secondly, the period of proliferation, accompanied by conjuga- tion, and resulting in the formation of a large number of germs, destined to spread the species. Thirdly, the period of rest, during which the parasitic germs pass out from the host into the outer world, to effect, if fortune favour them, the passive infection of a new host. In Sporozoa, considered generally, the life-history is similar in the main to that described above, but exhibits, in different forms, variations of every kind, in the direction either of greater or of less complexity. The deviations from the selected type may 1 Fide Wasielewski [7], p. 26. 3 In this and in all subsequent formulae of sporozoan life-histories au arrow is used to mean "becomes" or "grows into" ; the sign x to indicate a distinct cell- generation, a multiplication of individuals of any kind ; and a bracket with the sign + to denote the occurrence of zygosis or true conjugation Mid fusion of gametes. THE SPOROZOA 165 be considered from two points of view, according as they affect the characters (I.) of the individual stages or (II.) of the whole life-cycle. I. Each phase of the life-history may be varied or modified in structural or other details, in accordance with the special environ- ment and conditions of life to which a given species of these parasites is adapted. The modifications that occur under this head will receive detailed treatment in due course in the systematic review of the orders, families, and genera of Sporozoa in the sequel, but a few of the more important variations and simplifications may be considered here. The trophozoites have commonly, as in Monocystis, a definite body-form, limited by a cuticle ; but in many forms the protoplasm is naked, and the body is amoeboid, and of indefinite and changeable form. In Monocystis the gametes are not differentiated and the conjugation is isogamous, but in other types there may be anisogamous conjugation between sharply differentiated male and female gametes. The greatest variation, however, is seen in the spores. The number of the sporozoites is usually eight in Gregarines, but may be greater or less in other types. Hence the spores are distinguished as monozoic, dizoic, tetrazoic, polyzoic, and so forth, according as they contain one, two, four, or many sporozoites. In the monozoic condition there is no secondary multiplication within the sporocyst, but each sporoblast simply becomes a sporozoite. In many such cases the sporoblast secretes no sporocyst, but becomes a naked gymnospore, resembling a free sporozoite. These gymnospores may be formed within a resistent cyst secreted round the sporonts, or the cyst may be entirely absent. And further, the spore -formation may be preceded by conjugation of gametes, or the spores may be produced asexually, by the segmentation of a single sporont. Hence in an ideally primitive type of sporozoan development, the full-grown trophozoite or sporont simply breaks up, without previous conjugation or encystment, into a number of naked gymnospores, and each of them becomes a trophozoite which is similar to its parent, and repeats the process in due course. II. The plan and character of the life-cycle as a whole may be greatly varied, and secondary modifications or complications of various kinds introduced into it. The more important of these variations will be briefly described. (1) In Monocystis it has been seen that the period of growth and the period of proliferation are sharply separated, the latter fol- lowing upon the former. The same is the case in the whole order Gregarinida, to which Monocystis belongs, and also in two other orders, the Coccidiidea and the Haemosporidia. On the other hand, in the two orders known as Myxosporidia and Sarcosporidia, spore-formation commences at an early stage in the growth of the 1 66 THE SPOROZOA trophozoite, and spores are formed continually during the trophic stage, so that there is no distinction between trophozoite and sporont. Hence it has been proposed by Schaudinn to group the Gregarinida, Coccidiidea, and Haemosporidia together as a sub- class Telosporidia, contrasting them with a sub -class Neosporidia comprising the Myxosporidia and Sarcosporidia. The Telosporidia are Sporozoa in which the reproductive phases follow completion and cessation of growth ; the Neosporidia are Sporozoa in which growth and reproduction go on at the same time. It is probable that this distinction indicates the deepest phylogenetic cleft in this class of Protozoa. (2) In Monocystis the whole life -history is a single cycle, adapted entirely to spreading the infection amongst new hosts; it is, in fact, monogenetic. But in many other Sporozoa, belonging to either of the two sub-classes recognised above, the parasite may be capable of rapid multiplication within the body of its host, which it thus completely overruns in many instances. In such cases the life-cycle becomes digenetic, that is to say, it is differentiated into two distinct generations or series of generations, the one endo- genous or self-infective, the other exogenous or cross-infective. In the endogenous generations the reproductive processes are usually of a primitive type, taking place by binary or multiple fission, or by a simple form of sporulation, known as schizogony, in which a trophozoite, without encystation, breaks up into numerous gymno- spores, implanted on a certain amount of residual protoplasm. The sporulating individuals in this case are termed schizonts, and the gymnospores are known as merozoites, to distinguish them from the sporozoites of the exogenous generation. After a number of endogenous generations, the parasite soon or later reproduces itself by exogenous generation or sporogony, with the formation from sporonts of resistent spores that can be disseminated outside the body of the host. In monogenetic types the life-cycle consists of sporogony alone. (3) Considerable differences are seen in the manner in which the infection of a new host is brought about. The vast majority of Sporozoa appear to be disseminated passively, and the spores are taken up directly, in an accidental manner, by another host of the same kind as that from which they came. Should the spores chance to be devoured by an animal of another species, they will either be digested completely or will pass through its body un- altered. Only in their proper host do the digestive juices have the effect of liberating the sporozoites without harming them. In some cases, however, especially amongst Sporozoa parasitic in the blood, an intermediate host has been acquired, and is the agent by which the parasite is disseminated. The best-known instance of this is found in the malarial parasites, and is fully described THE SPOROZOA 167 below. Here the endogenous generations multiply by schizogony in the blood of a vertebrate host, until sporonts are formed, which must be taken up by a blood-sucking insect, such as a mosquito, in order to develop further. In the invertebrate host the exo- genous generation takes place, and the sporonts give rise, by sporogony after conjugation, to a number of gymnospores or sporozoites, with which the vertebrate host is again inoculated. In Sporozoa up to the present three modes of infection have been observed. The first and commonest method may be termed casual infection, where there is no intermediate host, and the infection is acquired by swallowing spores accidentally with the food. In the malarial parasites the infection is effected by the inoculative method, through the agency of an intermediate host. The third method is that of hereditary infection, a rare type, but known in at least one instance, the silkworm-disease produced by the myxosporidian parasite Glugea lonibycis (see p. 290), and possibly occurring also in the tick- fever parasites of cattle and other mammals (p. 262). In the first of these two instances the parasites penetrate the ovum and produce spores there, which germinate and infect the next generation of the host. It is possible that to these three modes a fourth should be added, which may be termed the contagious method, seen in the parasites which cause certain human skin-diseases (p. 238), but the sporozoan nature of these bodies is by no means demonstrated with certainty. Classification of the Sporozoa. — At least five well-established orders of Sporozoa are generally recognised — the Gregarinida, Coccidiidea, Haemo- sporidia, Myxosporidia, and Sarcosporidia. In addition, there are three orders which are at present less well known and of very uncertain value — the Haplosporidia, Serosporidia, and Exosporidia. The organisms formerly known as Amoebosporidia must now be included in the Gregarinida. Many ways of grouping these orders into higher subdivisions or sub- classes of the Sporozoa have been proposed. Labbe" set up two sub- classes : (1) Cytosporidia, in which the trophozoite is intracellular, either throughout the trophic period or at least in the earlier stages of growth ; (2) Histosporidia, in which the trophozoite is an intercellular tissue- parasite. The Cytosporidia comprise the Gregarinida, Coccidiidea, and Haemosporidia ; the Histosporidia include the Myxosporidia and Sarco- sporidia. The grouping proposed is a natural one, but the distinctions on which Labbe founded it have not the value which he attributed to them, since the young stages of the Histosporidia are intracellular as often as they are intercellular. Labbe now [4] subdivides the class into Cytosporidia, defined as having "no spore, or a simple spore without polar capsules," and Myxosporidia, having " the spore furnished with polar capsules containing an evaginable filament," while the Sarcosporidia are relegated to the Sporozoa incertae sedis. 1 68 THE SPOROZOA Delage and He"rouard [2] in their classification made use of the char- acter of the sporozoite or protoplasmic germ within the spore, and divided the class into (1) Rhabdogeniae, " with sporozoite of definite form, generally falciform (arqude) " (Gregarinida, Coccidiidea, Haemosporidia, and Sarco- sporidia) ; and (2) Amoebogeniae, " with amoeboid sporozoite " (Myxo- sporidia). This classification has the disadvantage, however, of separating the two nearly allied groups Myxosporidia and Sarcosporidia, and it has not been followed by any subsequent writers. Mesnil [6], making use of names invented by Metschnikoff, divides the Sporozoa into Ectospora (Gregarinida, Coccidiidea, and Haemosporidia) and Endospora (Myxosporidia, Sarcosporidia, and Haplosporidia). In the Ectospora, the sporulation takes place at the close of the trophic period, and the spore-mother-cells (sporoblasts) are formed at the periphery of the sporont ; in the Endospora, spore-formation goes on during the growth of the trophozoite, and the spore-mother-cells (pansporoblasts) are cut off in the interior of the body (p. 283). In the sequel the classification of Schaudinn into two groups, Telo- sporidia and Neosporidia, as defined above (p. 166), is followed. It is seen that, as compared with the classifications of Labbe" and Mesnil, the dis- tinction depends rather on the mode of defining the two subdivisions than on essential differences in the plan of grouping the orders. SYSTEMATIC REVIEW OF THE SPOROZOA SUB-CLASS TELOSPORIDIA. Sporozoa in which the reproductive phase of the life-cycle is distinct from, and follows after, the trophic phase. ORDER 1. Gregarinida. The Gregarinida, commonly known as Gregarines, are an order of the Sporozoa remarkable for the degree to which structural complexity of the individual, and adaptive specialisation of the species, are carried. On the other hand, the life-cycle is usually extremely simple. It might, in fact, be said, speaking generally, that the Gregarines are the highest of the Sporozoa from the standpoint of morphology, and the most differentiated from the point of view of taxonomy, but are at the same tinle amongst the simplest as regards reproductive phenomena. Their distinctive characters are as follows : — The trophozoite commences its growth typically as an intracellular parasite, usually, if not always, of an epithelial cell ; never of a blood-corpuscle. It soon outgrows the host-cell, and bulges from it, and finally drops out into an internal cavity of the host, usually the digestive tract, but often the THE SPOROZOA 169 haemocoele (blood-vessels or body-cavity), and sometimes the true coelom. Here it continues to grow, absorbing nutriment from its host, until it becomes a ripe, full-fed sporont. It then encysts, with or without previous association with another of its kind, and the process of spore-formation or sporogony commences. Sporo- blasts are formed which usually secrete sporocysts and give rise to spores, and within the spore-envelope the sporoplasm breaks up into sporozoites, eight in number as a general rule. In a few rare instances there is endogenous reproduction by schizogony, in addition to the ordinary sporogony. As has been said above, the Gregarines were the earliest Sporozoa to be observed and studied, on account of their large and conspicuous size. The history of the group may be said to commence, for all practical purposes, with the founding of the genus Gregarina by Dufour in 1828. From that time onwards numerous observers, amongst whom Aimd Schneider and L4ger deserve special mention, have added to our know- ledge of the abundance of genera and species of these parasites, or have studied the details of their life -history and development. Nevertheless, it is only in the most recent times, practically in the new-born twentieth century, that the facts concerning the conjugative processes have become accurately known, largely in consequence of renewed investigations upon them stimulated by the interesting dis- coveries made in other orders of Sporozoa. Occurrence, Habitat, etc. — The Gregarines are confined for the most part to invertebrate hosts, and have never yet been found in any craniate vertebrate. The great majority of them lead blame- less lives in the interiors of various arthropods, to which the segmented forms comprised in the sub-order Cephalina are almost confined. The unsegmented forms are found commonly, however, in other groups also, especially in echinoderms, in annelids, including gephyrea and hirudinea, and in tunicata. A few have been recorded from turbellaria, nemertines, and enteropneusta, and a doubtful species is known from Amphioxus. In molluscs, however, they are almost unknown, the single recorded instance being a species from the body -cavity of Pterotrachea. Thus the Gregarines are to a large extent the opposite to the Coccidia in the matter of the hosts they affect, the arthropods alone being ground common to the two orders. The infection of the host is probably effected in all cases by way of the digestive tract, and the sporozoites, when liberated there, proceed to attack the lining epithelium. In some cases the sporozoite traverses the epithelium without stopping, passing on into the haemocoele, as in Diplocystis of the cricket, or into the coelom or one of its subdivisions, as in Mo'nocystis. In other cases it remains attached to an epithelial cell only by a small THE SPOROZOA portion of the body ; the intracellular stage or " Coccidian phase" being practically suppressed. In other cases, again, a large portion of the body, containing the nucleus, is imbedded in the epithelial cell, while the rest of the body projects freely from the host -cell. But in typical cases the youngest trophozoites are found as intracellular parasites, completely enclosed by a cell of the epithelium, either of the gut or some of its diverticula. In this situation the parasite grows rapidly, and soon becomes larger than the host-cell. The trophozoite then falls out of the exhausted cell, usually passing inwards towards the lumen of the gut, sometimes, however, outwards into the vascular system or body-cavity. Gregarines in the latter situation are commonly termed "coelomic," without distin- guishing whether the body-cavity in which they lie is a true coelomic space or a part of the haemocoele. Coelomic Gregarines, in the latter sense, occur very frequently in insects (Fig. 9), and in many cases a Gregarine may occupy different situations at different periods of the life of its host. It commonly happens that a Gregarine inhabiting the digestive tract of an insect -larva passes through the wall of the gut at the metamorphosis, and so becomes a coelomic Gregarine in the imago. The young trophozoites have been shown to have remarkable effects upon the cells in which Larva of Tipuia oie- ^ey are parasitic. The infected host-cell passes racea, opened to show through two successive phases — first one of the gut covered with , , , . * , m, - , coelomic Gregarine hypertrophy, then of atrophy. The facts have leCki, affiJugS)19" been investigated by Laveran and Mesnil [16], and still more recently by Siedlecki [28], in several species. The youngest trophozoites of Lankesteria ascidiae, studied by Siedlecki, place themselves deep in the basal portion of the epithelial cell, the region where the protoplasm of the cell is least differentiated for secretion (Fig. 10, a). The nucleus of the host-cell soon begins to appear swollen, its chromatin network becomes loose and stains in a diffuse manner, and its nucleolus increases greatly beyond the normal size, acquiring irregular contours and often dividing into several parts. Hyper- trophy of the nucleus is soon followed by that of the cytoplasm, which appears clearer than in the adjacent cells, apparently as the result of a sort of liquefaction. The protoplasm becomes difficult to fix, and always stains much more feebly than the protoplasm of THE SPOROZOA 171 neighbouring cells. The Gregarine meanwhile is also increasing in size, and its rate of growth exceeds that of the infected cell. When it is large enough to fill the hypertrophied host -cell, degeneration of the latter commences (Fig. 10, b). Its nucleus shrinks, becomes crescent-shaped, and finally becomes a flattened corpuscle which stains strongly and consists of debris of chromatin. At the same time the cytoplasm is absorbed until it forms a thin skin enclosing the Gregarine, and is finally cast off with it from the epithelium. It is remarkable that in other cases a similar series of changes may be provoked in the epithelial cell after the Fio. 10. Intracellular stages of Lankesteria ascidiae (Lank.) (par. Ciona intestinalis) in the intestinal epithelium, a, young stages showing the hypertrophy of the epithelial cells induced by the parasites at an early stage. 6, older stage showing very great hypertrophy of the epithelial cell, with atrophy of its nucleus, ep, normal epithelial cell ; ep', hypertrophied epithelial cell con- taining (G) the young Gregarine : n, nucleus of normal cell ; n', nucleus of infected cell. (After Siedlecki, x 750.) Gregarine has grown out from it, and is attached to the cell only by a minute point of contact (Fig. 11). In many cases a Gregarine, after having been set free from an epithelial cell which it has destroyed, may secondarily attach itself again to the epithelium (Figg. 12 and 13). Although such second- ary attachment may be exceedingly complicated, and may affect a large number of epithelial cells, as in Pterocephalus (Fig. 12), never- theless it only produces mechanical alterations in the cells, and never has the marked effects which result from the primary attachment. It is thus seen that the Gregarine destroys completely the cell in which it is parasitic at the commencement of its career. Never- theless Gregarines appear to be extremely innocuous to their hosts. Since they do not reproduce themselves by schizogony, except in a very few instances, they do not overrun their host in the way 172 THE SPOROZOA that the Coccidia or Haemosporidia do. Though a given host often contains a considerable number of Gregarines, it must be sup- posed that they represent simply the batches of sporozoites derived from several distinct infections. The epithelial cell that each individual Gregarine has destroyed is not missed, and the injury FIG. 11. Effects produced on epithelial cells by the trophozoites of a Gregarine (Pyxinia frenzeli, Lav. et Mesn.) (par. Attagemis pellio, larva). A , hypertrophy of the cell (first stage). B, atrophy (second stage). Combined after figures by Laveran and Mesnil. Fio. 12. Portion of a section through the apparatus of fixation of a Pterocephalus, showing root - like processes extending from the Gregarine between the epithelial cells, which are not modified or altered in any way, but appear to be under the influence of traction exerted by the Gregarine. g, head of the Gregarine ; r, root-like processes ; ep, epithelial cells. (After Siedlecki, x 500.) Fio. 13. Trophozoite of iMnkes- teria ascidlae (Lank.) (par. Ciona intestinalis), attached by an anterior pseudopodium-like pro- cess to an epithelial cell (ep'), which is withered and apparentlytiestroyed by it. ep, normal epithe- lial cells. (After Sied- lecki, x 500.) is easily repaired. The nutriment that the Gregarines absorb in the gut of the host seems to be a tax lightly borne. There is, in short, no record of any pathological effects produced by these parasites beyond those already noted in the case of the host-cell. Morphology and Life-history. — Since a typical Gregarine has already been described in Monocystis, it is only necessary to review THE SPOROZOA 173 briefly the variations exhibited by this order as compared with the type selected for description. The body -form and external characters of the trophozoite furnish sharp distinctions for classificatory purposes. The funda- mental type of body-form may be described as a sphere or ovoid. In many species this type of form is very nearly retained (Fig. 22), especially in the non-motile coelomic forms, which often have a great resemblance to ova. More usually, however, the body becomes strongly elongated in one direction, a mode of growth correlated either with attachment by one pole or with forward Epimerite. Protomerite. Deutomerite. FIG. 14. Scheme of development of a Gregarine from a sporozoite. a, free sporozoite ; b and c, stage in the growth of the parasite within an epithelial cell ; d, the Gregarine beginning to protrude from the cell ; e, segmentation of the body and emigration outwards of the nucleus, the iiitra- cellular portion of the body remaining as the epimerite ; /, adult Gregarine with three- chambered body. (From Wasielewski, after Schneider.) movement in a definite direction. In many cases the body is extremely drawn out and attenuated, becoming vermiform in character (Fig. 1). In the sub-order Acephalina, of which Monocystis is an example, the body remains simple and is not subdivided into different regions, whatever its form. The Cephalina, however, are, with few exceptions, septate, that is to say, the body is divided by septa or partitions into distinct chambers or segments, usually three in number. The septate condition is brought about in the following way. The sporozoite penetrates an epithelial cell (Fig. 14, b) and grows within it into an oval body, which at an early stage cannot be distinguished in any way from a young Mono- cystid, or even from a Coccidian parasite (Fig. 14, c). Very soon, 174 THE SPOROZOA however, the young trophozoite grows out from the host-cell, and its nucleus travels out into that portion of the body which projects from the cell (Fig. 1 4, d, e). The free extremity of the Gregarine body continues to grow, while the intracellular portion becomes cuticularised and forms simply an organ of fixation, commonly called the epimerite. The extracellular Gregarine body becomes now divided by a septum into two chambers, one smaller proximal (i.e. nearer the host-cell and the epimerite), termed the protomerite, Fio. 15. Cephalont of Pyxlnia rubecula, Hamm. (par. Der- mestes spp.) still attached by its epimerite to a de- tached epithelial cell. (From Wasielewski, after Leger.) Fio. 16. Corycdla armata, Leger (par. Gyrinus natator, larva), a, cephalont ; b, epimerite in the host-cell, magnified ; c, sporont. (From Wasielewski, after Leger.) and one larger distal, termed the deutomerite, which usually contains the nucleus (Fig. 14, /); abnormal forms are sometimes found, however, in which the nucleus is lodged in the protomerite, owing apparently to precocious formation of the septum, before the nucleus had reached its distal position. The young trophozoite in the Cephalina remains attached to the host-cell for some time by its epimerite (Fig. 15). In this condition it is known as a cephalont. Soon it becomes detached and set free by a rupture of the junction between the epimerite and protomerite (Fig. 16). The epimerite remains sticking in the THE SPOROZOA '75 withered remains of the host-cell, and the Gregarine body, com- posed of protomerite and deutomerite, is free in the gut, where it continues its growth and further development. The free Gregarines are commonly termed sporonts. The epimerites of Gregarines show every variety of size, shape, and pattern, and may be ornamented with hooks, spines, and other appendages (Fig. 17). They function as organs of attachment, as has heen said, and probably also as organs of nutrition, since Laveran and Mesnil [16] have Fio. 17. Epimerites of various Gregarines. a, Gregarina longa (Leger), (par. Tipula sp., larva) ; 6, Sycia inopinata, Leger (par. Audouinia sp.) ; c, Pileocephalus heerii (Roll.), (par. Fhryganea, larva) ; d, Stylorhynchus longicollis, Stein (par. Blaps mortisagd) ; e, Bdoides firmus (Leger), (par. Der- inestes lardarius, larva) ; /, Cometoides crinitus (L^ger), (par. Hydrobius sp., larva) ; g, Geneio- rhynchus monnieri, A. Schn. (par. Libellula, larva) ; h, Echinomera hispida (A. Schn.), (par. Lithobius forficatus) ; i, Pterocephalus nobilis, A. Schn. (par. Scolopendra spp.). (From Wasielewski, after Leger.) shown that they evoke changes in the host-cell which cannot be explained as the result simply of mechanical irritation. The possession of an epimerite is a feature which is used for classifying the Gregarines, and the legion Eugregarinae is separated into the two sub-orders Cephalina and Acephalina, according to the presence or absence of this appendage. Aa a general rule the forms which possess an epimerite have the body behind it divided into protomerite and deutomerite by a septum, and have hence been termed Polycystida seu Septata (Lank.), while those without an epimerite are also without a septum ; hence Monocystida seu Haplocyta (Lank.). But in one family, Doliocystidae, Labbe, an epimerite is present, and may attain a considerable size, as in Doliocystis (Monocystis) aphroditae (E. R L.), without any septum dividing the rest of 1 76 THE SPOROZOA the body (Fig. 1 9). It is purely a matter of definition whether these forms be considered as Cephalina without a septum, or as Monocystida with an epimerite. The Cephalina in which the body is non -septate are some- times distinguished as Dicystida from those in which there is a distinct protomerite and deutomerite (Tricystida). These terms are to be under- stood, however, in a purely descriptive sense, and cannot be used for Fio. 18. Three specimens of Schneideria mucronata, Leger (par. Bibio marci, larva), o, young cephalont, attached to a host-cell, b, older cephalont. c, sporont, showing traces of a protomerite. (From Wasielewski, after Leger.) C classificatory purposes, as there is no doubt that many dicystid species are derived from tricystid forms secondarily, by obliteration of the protomerite (Fig. 18). On the other hand, such forms as the Doliocystidae (Fig. 19) and Selenid.ium (Fig. 46) appear to be truly and primitively dicystid, and are to be regarded as intermediate forms transitional from Acephalina to Cephalina. In the aberrant forms comprising the legion Schizogregarinae, the unsegmented body grows out into irregular processes, which give it an amoeboid appearance, whence these forms obtained their older name, THE SPOROZOA 177 Amoebosporidia. Recent observations have shown, however, that these processes are not pseudopodia, but are stiff outgrowths of the body, clothed FIG. 19. Doliocystis aphroditae (Lank.) (par. Aphrodite aculeatd), a non - septate Oregarine with a distinct epimerite. (After Lan- kester.) Fio. 20. Associations of Gonospora sparsa, Leger, from the gut of Glycera. (From Wasielewski, after Leger.) by cuticle (Fig. 21), so that the name Amoebosporidia rests upon a mis- conception and must be abolished. The genus Pterospora is also remarkable for the possession of retractile processes, resembling tentacles (Fig. 37). A curious feature of Gregarines, and one which has a marked influence in many cases on their external form and appear- ance, is their tendency to form associations during the trophic period, a peculiarity from which the type-genus Gregarina probably derives its name. In Monocystis it has been ji. seen that two individuals come together when full- FIO. 21. grown and become associ- ?ort.ion °f » •ecJJ°n of a Maipighian tubule magica infested by Ophryocystis schneideri, showing three ated to form a Cyst in individuals of the latter species (G), one of them with T ^4.1. o, n two nuclei, attached by stiff processes (the pseudopodia Common. In Other Greg- Of Schneider) to the wall of the tubule. pt syncytial arnes association may take place at a much earlier stage in the development of the individual (Fig. 20). 12 THE SPOROZOA In the Diplocystis found in the body -cavity of the cricket, young trophozoites become associated in couples almost immedi- ately after leaving the host-cell, and, according to Cu6not, no solitary individuals are to be found above a certain size, since all the old maids die off. In Diplocystis major the two associates retain their distinctness, but in Diplocystis minor each couple becomes surrounded by a common membrane (Fig. 22). In Cystobia holothuriae early association has still more far-reaching results, since a fusion, complete except as regards the nuclei, of the two trophozoites takes place, so that the appearance of a single Gregarine with two nuclei is produced, with no trace of FIG. 22. Precocious association in Diplocystis minor, Cuen., of the cricket, m, common membrane uniting the two associates ; g, grains of albuminoid re- serve material. (After Cuenot, x about 120.) Fio. 23. Adult trophozoite of Cystobia holo- thuriae (Ant. Schn.) (par. Holothuria tubulosa), showing the two nuclei, derived from the fusion of two indi- viduals, but not separated by any septum. (After Minchin.) any septum between them (Fig. 23). While in the cases mentioned the association is undoubtedly a preliminary to the conjugation of gametes, it is more difficult to interpret the peculiar aggregations known as syzygies commonly seen in many species, especially amongst Cephalina. Free Gregarine individuals become attached to one another, the anterior extremity of one adhering to the posterior end of the other (Fig. 24, a). Usually such a syzygy consists of two individuals, but may be composed of a chain of half-a-dozen or more (Fig. 24, c). The most anterior individual is termed the primite, those behind it the satellites. The latter are always individuals which have lost their epimerites, if they belong to the Cephalina. The syzygy does not necessarily take the form of a simple chain. Two satellites may be attached side by side to the hinder end of the individual, primite or satellite, in front of them (Fig. 24, b). In some cases the individuals com- posing a syzygy are loosely attached and easily separated from one another, and the members of it are not modified in any way. In other cases the association is more intimate, and the satellite THE SPOROZQA 179 or satellites may become modified in structure. In segmented forms this alteration affects chiefly the protomerite, which may be reduced or even absent. Thus in the genus Didymophyes syzygies of two individuals are formed in which the satellite loses its protomerite entirely, so that the resulting combination looks like a three - chambered Gregarine with two nuclei (Fig. 25, a). Although in many cases the syzygies appear to be tem- porary attachments which have no con- nection with the sub- sequent reproductive phenomena, it is prob- able that as a general rule they represent associations of indi- viduals destined to form conjugating gametes as described for Monocystis, especi- ally in those cases where the union of FIG. 24. a, Eirmocystis ventricosa, Leger (par. Tipula spp.). b and c, E. polymorpha, Leger (par. Lim- ia, larva). Associations of two, three, and five Gregarines. p, primite ; s, satellites. (From Wasielewski, after Leger.) c primite and satellites is an intimate one, as in Didymophyes and others. Having regard to the manner in which conjugation takes place, there is no reason why any number of sporonts or. gametocytes should not come to- gether to form gametes within a common cyst, and the presence in a cyst of more than two sporonts appears to be of frequent occurrence in some species. The body of a Gregarine trophozoite always consists of cuticle, ectoplasm, and endoplasm containing a nucleus, but each of these parts are subject to considerable variation in structure. The cuticle or epicyte is a membrane secreted by the ectoplasm, usually of some thickness, and appearing doubly contoured in optical section (Fig. 26, c). Sometimes it can be broken up into fine vertical lamellae corresponding to the ridges presently to be described on the external surface. As has been said above, the cuticle is often produced into hooks or spines or other organs of fixation, especially on the epimerite. On the other hand, all such i8o THE SPOROZOA processes may be wanting entirely. The surface of the cuticle is not smooth, however, but has a delicately ribbed or fluted structure, producing fine striations which run in a meridional direction from pole to pole. As a rule there are no openings or visible pores of any kind in the cuticle, but, according to Siedlecki, a pore exists at the anterior end of Lan- Jcesteria ascidiae, from which is protruded a minute pseudopodium-like process which serves for the secondary attachment of the trophozoite to an intes- tinal cell of the host, and pores are stated to exist in the longitudinal furrows of the cuticle, as described below. The ectoplasm is a clear, hyaline layer of tougher protoplasm which in the motionless forms shows no special differen- tiation. But in most Gregarines, correlated with rmea, each of which gives rise to very numerous sporozoites grouped round a central mass of protoplasm, THE SPOROZOA 189 and each such cluster of sporozoites resembles to a certain extent an ordinary Gregarine spore, but has no enveloping membrane or sporocyst (Fig. 41). Aggregata and Porospora are grouped together as a tribe Gymnosporea in distinction to the ordinary Gregarines, the Angiosporea, in which a chlamydospore is formed. The proto- plasm of the sporonts may, in some cases, be entirely used up to form sporoblasts, in which the sporulation is said to be complete, but more often it is incomplete, with a more or less considerable mass of residual protoplasm. The typical Gregarine spore contains eight sporozoites, and is therefore said to be octozoic, but exceptionally only four sporozoites are formed, as in Selenidium, hence tetrazoic. The sporozoites are grouped in very various ways round a granular mass of residual protoplasm, which contains the last remnants of the reserve nutrition stored up by the sporont. The protoplasm of the sporozoites is clear and finely granulated. Each sporozoite is typically sickle-shaped with the nucleus in the middle. Some- times the nucleus is at one extremity, and the sporozoite then has a form more resembling a tadpole (Fig. 38). The spore -envelope or sporocyst consists of two layers, an outer clear and delicate epispore, and an inner refringent and tough endospore. Sometimes these two layers are quite separate, or, on the other hand, they may be intimately united. In external characters the spores show the greatest possible variety of form and pattern, and are frequently ornamented with long tails or processes, which may vary considerably even in closely allied species, as in the species of Cystobia infesting Holothurians (Fig. 38). Another remarkable feature seen in some genera is the union of the spores by their sporocysts to form strings or ropes ("spores en chapelet") (Fig. 34,/). With regard to the dissemination of the spores, and the manner in which they infect new hosts, there is nothing to add to what has been stated above with re- gard to Monocystis (p. 163) and Sporozoa generally (p. 166). In no case is a true intermediate host known to occur. The life-cycle of Gregarines is, in the vast majority of cases, monogenetic, and consists of sporogony only, FIG. 33. Spores of Pyxiniarubecvla, Hamm. (par. Dermestes spp.). a, a ripe spore showing distinct epispore and endospore. 6, the endospore set free after bursting of the epispore. After extrusion of two polar spheres (n), the sporozoites (s) escape from the spore. (From Wasielewski, after Leger.) 190 THE SPOROZOA as described above for Monocystis. The cases in which proliferation, by other methods than the usual spore-formation, has been alleged, are for the most part very doubtful, and, though not incredible, are highly improbable for reasons already put forward in dealing with Monocystis. There are, however, a few well-attested cases of Fio. 34. Spores of various Gregarines. a, simple oval spore, type of Eirmocystis, Sphaerocystis, etc. (Gregarinidae). b, cylindrical spore, type of Echinomera, Da'ctylophonis, Pterocephalus, etc. (Doc- tylophoridae). c, barrel-shaped spore, type of Gregarina and other Grega- rinidae with sporoducts. d, navicular spore of lieloides (Actinocephalidae). e, biconical spore with spines ol Ancyrophora (Acanthosporidae). / purse -shaped spore typical ol Stylorhynchidae. g, crescent-shaped spore typical of Menosporidae. h. sac -like spore of Gonospora tere bellae (Koll.). i, tailed spore o: Ceratospora. j, tailed spore of Uro spora synaptae, Cu4n. (Froir Wasielewski, after Leger.) schizogony, which is correlated in Eugregarinae, as shown by Caullery and Mesnil [10], with an intracellular stage of long dura- tion, and takes place during this phase of the life-history. Thus in Gonospora longissima, Caull. et Mesn., from the Annelid Dodecacerio concharum, the nucleus of the intracellular trophozoite multiplies by division, and the body divides into six or eight merozoites arranged as a "corps en barillet " (see p. 222). The merozoites then separate, escape from the host- cell, and develop into the intercellular sporonts. Another and similar case has been THE SPOROZOA 191 observed by Caullery and Mesnil [Sb] l in a species of Selenidium from Spio fuliginosa. In the Schizogregarinae, on the other hand, schizogony is of constant occurrence, as their name implies, and takes the form of multiple fission during the free extracellular phases of the life-history. CLASSIFICATION. The systematic arrangement of the Gregarinida that follows is taken from Labbe's " Sporozoa " [4], for the most part, but with some additions or modifications necessitated by recent advances in our knowledge of the group. SUB -ORDER I. SCHIZOGREGARINAE, Le'ger (Amoebosporidia auct). Gregarinida in which schizogonic reproduction takes place during the extracellular phase of the trophozoite, in addition to the ordinary sporogony. The forms composing this sub-order have been regarded until recently as a very problematic group. Their position in the Sporozoa and their affinities with other members of the class have been considered doubtful and altogether uncertain. Up to the end of the nineteenth century the group was represented only by two species of Ophryocystis, except for the fact that the supposed cancer-parasite has been referred to it by some authorities. The misconception which has prevailed with regard to the natural position of these forms appears to be largely due to the fact that the species of Ophryocystis were originally described as amoeboid, and this character was supposed to be diagnostic of the order represented by them, hence termed Amoebosporidia. The recent investigations of Leger [20, 21, 25], however, have not only made known an allied form, Schizocystis, which has a fixed body-form like other Gregarines, but have demonstrated that even Ophryocystis is not amoeboid, as originally described, but has a definite orientation of the body, the apparent pseudopodia being merely stiff processes of attach- ment (Fig. 21). There can be no question that the natural position of the group is amongst the Gregarines ; indeed it is difficult to find any constant diagnostic character, except the mode of reproduction, separating Ophryocystis and Schizocystis, the only two genera known at present, from the rest of the order. Genus 1. Schizocystis, »Leger, 1900, for 8. gregarinoides, Leger, from the intestine of a dipterous larva, Geratopogon sp. The trophozoites are cylindrical and elongated, about 150 //. in length, with an anterior clearer region, and occur fixed to depressions in the intestinal wall. They resemble a Monocystid in general appearance, but while uninucleate in the youngest stages, the full-sized individuals may have as many as sixty nuclei. The body then divides up to form a number of merozoites, which become trophozoites of the second generation. The latter are uninucleate, and when full -sized they associate and become encysted, giving rise to gametes which conjugate and produce octozoic spores, exactly after the 1 Whose figures, however, are far from convincing. 192 THE SPOROZOA pattern of Monocystia. Genus 2. Ophryocystis, A. Schneider, 1884. Several species are known, all from the Malpighian tubules of beetles ; type- species 0. btitschlii, A. Schn., from Blaps mortisaga (Fig. 31). The body of the trophozoite is of irregular form, with pseudopodium-like processes (Figg. 21, 31, and 35). The trophozoites of the first generation have several nuclei (apparently not more than six or eight) when full-sized, and then divide up to form as many small individuals of the second generation, which often remain connected together for some time (Fig. Fio. 85. Stages in the life-history of Ophryocystis francisci, A. Schn. (par. Akis spp.). o, rosette of six small uninucleate individuals produced by division of a schizont ; b and c, individuals detached from a rosette, in 6 still showing the process of attachment ; d, rosette of four indi- viduals ; e, sporont ; /, association of two sporonts ; g, cyst with single spore and two masses of residual protoplasm. (From Wasielewski, after A. Schneider.) 35, a, d), but ultimately separate and become the uninucleate trophozoites of the second generation. The latter, when adult, associate and become encysted, and then give rise to a single octozoic spore, after elimination of nuclei and conjugation of a surviving pair, as described above (p. 188). SUB-ORDER II. EUGREGARINAE, L6ger. Gregarinida in which schizogonic reproduction is of very exceptional occurrence, and takes place only during the intracellular phase, if at all Spores octozoic with the rarest exceptions. TRIBE 1. ACEPHALINA, Kolliker (Monocystidea, Stein). Eugregarinae in which the body is non-septate, and without an epimerite at any stage. Chiefly " coelomic " parasites. THE SPO&OZOA 193 Genus 3. Monocystis, Stein, 1848. Trophozoites characterised by considerable contractility, and consequent changeability of body -form. Spores navicular. Several species from Oligochaetes, one from Clymenella torquata, and one from Diap- tomus and Cyclops, all inhabit- ing the vesiculae seminales or general body-cavities of their hosts. Type M. agilis, Stein (Figg. 2-8). The genus Sper- matophagus, Labbe, 1899 (noin. nov. for Spermatobium, Eisen, 1895, preoccupied), for two species parasitic in the vesi- culae seminales of earthworms, is apparently a synonym of Monocystis. Genus 4. Zygocystis, Stein, 1848. Adult tropho- zoites generally piriform, always found associated in couples or threes (Fig. 36) ; Fio. 36. spores biconical. Type Z. Zygocystit cometa, Stein (par. Lvmbricus eommunis .£>,,.. , • -I [=herculeus ? ]). a. syzycy of two Individuals' b of co7neto,Stem,from the vesiculae three. (After stein, Vaso.) seminales and general body- cavity of Lumbricus agricola. Two other species are known. Genus 5. Zygosoma, Labbe", 1899 (nom. nov. for Conorhynchus, Greeff, 1880, preoccu- pied). Trophozoites pear-shaped, the entire body covered with finger-like processes, the endoplasm filled with vacuoles ; always associated in couples FIG. 37. Pterospora maldaneorum, Rac. et Lab. (par. Liocephalus llopygus and Clymene luinibritalit). a, two associated trophozpites ; the individual on the left is fully expanded, that on the right is commencing to retract its processes, b, spore showing the sporozoites coiled spirally round the central mass of granular residual protoplasm, c, transverse section of a spore showing the three wing-like processes and the sporozoites (four of them) in the section round the central residual protoplasm. when full-grown. Sporulation unknown. Unique species Z. gibbosum (Greeff) from the gut of Echiurus pallasii. Genus 6. Pterotpora, Eacovitia and Labbe", 1896. Trophozoites pear-shaped, the smaller extremity bearing two groups of finger-shaped retractile processes, four in each group ; always found associated in couples. Spores with dissimilar poles, the epiapore prolonged into three lateral wing -like expansions. Unique species P. '3 I94 THE SPOROZOA maldaneorum, R and L. (Fig. 3 7), from the coelom of Liocephalus liopygus and Glymene lumbricalis. Genus 7. Cystobia, Mingazzini, 1891. Trophozoites large, oval or irregular in form, with two nuclei, resulting probably from early fusion of associated individuals (Fig. 23). Spores with dissimilar poles, the epispore form- ing a funnel-like projection at one pole, some- times also a tail-like expansion at the other (Fig. 38). Parasites of Holothurians occurring in the blood-vessels, whence the cysts dehisce into the coelom. Type C. holothuriae (Ant. Schn.) from Holothuria tubulosa. Genus 8. Lithocystis, Giard, 1876. Trophozoites large, ovoid or vermiform, with the endoplasm filled with crystals of calcium oxalate. Spores with Flo 8g long tubular processes of the epispore at one Spores of o,'c**>Ma irregu- P°le' Uni(lUe 8PeCieS L' «*>™d**, Giard (Fig. (Minchin), (par. Holothuria 39), from the coelom of various Echinids. ' Jl Genus 9. Ceratospora, Leger, 1892. Tropho- losa). (After Minchin.) zoites of elongated conical form, associating by their truncated extremity, and giving rise to spores without encystment and without change of external form. Spores oval, with a collar-like expansion at one extremity, and two long rigid filaments at the other (Fig. 34, i). Unique species C. mirabilis, Le"g., from the body-cavity of Glycera sp. Genus 10. Urospora, A. Schneider, 1875. Trophozoites large, spores oval, with a caudal filament at one pole (Fig. 34, j). U. saenuridis (Koll.), from the vesiculae seminales and coelom of Tubifex rivulorum. Other species from Nemertines, Sipunculus, Synapta, etc. Genus 11. Gonospora, A. Schneider, 1875. Trophozoites ovoid, piriform, or vermiform (Fig. 20). Spores with dissimilar poles, rounded at one extremity, bearing one or more tooth-like processes at the other (Fig. 34, h). Four species, all from the coelomic cavities of Polychaeta. Type G. terebellae (Koll.), from Terebella, etc. Genus 12. Syncystis, A. Schneider, 1886. Trophozoites ovoid or piriform. Spores navicular with four divergent bristles at each extremity. Unique species S. mirabilis, A. Schn., from the body-cavity and fat-body of Nepa cinerea. Genus 13. Diplocystis, Kiinstler, 1887. Trophozoites of " coelomic " habitat, associat- ing precociously to form spherical masses. Spores spherical or oblong. D. schneideri, Kiinst., from the body-cavity of Periplaneta americana. D. major , Cue"n., and D. minor, Cue"n. (Fig. 22), from the common cricket. Genus 14. Lankesteria, Mingazzini, 1891. Trophozoites more or less spatulate (Fig. 13). Spores oval (compare Fig. 29, /). Type L. asddiae (Lank.), from the gut of Ciona intestinalis. The sporozoan parasite described by Pollard,1 from the intestinal epithelium of Amphioxus, is identified by Labbe* as the intracellular stage of a Gregarine belonging to this genus. Genus 15. Gallyntrochlamys, Frenzel, 1885. Trophozoites with the body constricted into two regions not separated by any septum, and with a fur-like covering of rods, resembling cilia, clothing the surface of the 1 Quart. Journ. Micr. Sci., N.S. xxxiv. p. 811. THE SPOROZOA '95 body except at the anterior extremity. Spores not described. Type C. phronimae, Frenz., from the gut of Phronima sedentaria. Genus 16. Ancora, Labbe", 1899 (nom. nov. for Anchorina, Ming., 1891, preoccupied). Trophozoite anchor-shaped, with two anterior lateral prolongations of the FIG. 39. Lithocystis schneideri, Giard (par. EcJiinocardium, etc.). a, an association of two of the extremely lively trophozoites, which attach themselves loosely to one another in pairs, keeping up at the same timo very active movements. 6, two trophozoites (sporonts) about to become encysted ; the bodies have contracted into compact motionless masses, and in each individual vacuoles have appeared containing clinorhombic crystals of calcium oxalate ; the whole mass is surrounded by a coat of amoebocytes from the coelomic fluid of the host, c, an unripe spore, before formation of sporozoites, highly magnified, d, a ripe spore, c and d also show the differences between the two kinds of spores ; c is a microspore, d, a macrospore. n, nucleus ; am.c, investment of amoebocytes ; cr, crystals ; /, funnel-like prolongation of the epispore, through which the sporozoites pass out ; t, tail-like process of the epispore, tubular in the unripe, filamentous in the ripe spore ; sp.z, sporozoites. (After Leger.) body. Spores unknown. Unique species A. sagittata (Leuck.), from the gut of Capitella capitata. The following genera of Acephalina, known only in the trophozoite phase, are insufficiently characterised : — Genus 17. Pleurozyga, Mingazzini, 1891. Trophozoites more or less claviform, associating laterally. Three species from Ascidians. Genus 18. 196 THE SPOROZOA Ophioidina, Mingazzini, 1891. Trophozoites elongated, vermiform, the body cylindrical, blunt at one end, pointed at the other. 0. bonelliae, Frenz., from the gut of Bonellia viridis. Two other species are also referred to this genus. Genus 19. KolliJcerclla, Labbe", 1899 (nom. nov. for Kollikeria, Ming., 1893, preoccupied). Trophozoites of rhomboidal form, the anterior extremity rounded and forming a sort of head, separated by a constriction from the rest of the body. Unique species K. staurocephali (Ming.), from Staurocephalus rudolphii. Genus 20. Lobian- chella, Mingazzini, 1891. Trophozoites of elongated form with the anterior end rounded. Unique species L. bdoneides, Ming., from the coelom of Alciope sp. TRIBE 2. CEPHALINA, Delage (^Polycystidea auct + Doliocyttidae). Eugregarinae which always possess an epimerite, which may be present only in the young stages or may be a permanent organ. The body is divided, typically, by a septum into protomerite and deutomerite, but may be simple, non-septate. Parasites chiefly of Arthropods, usually occurring in the gut. (a) SUB-TRIBE GYMNOSPORIA, Le"ger. The cyst contains naked gymnospores (sporozoites) not enveloped in sporocysts to form spores. FAMILY 1. AGGREGATIDAE, Labbe". Sporonts septate, forming associa- tions of two or more individuals. Sporozoites grouped irregularly round a number of residual masses (Fig. 32). Genus 21. Aggregate, Frenzel, 1885. Trophozoites elongated, cylindrical. A. portunidarwm, Frenz., from the intestine of Carcinu* maena* and Portunus arcuatus, and several other species from other Crustacean hosts. FAMILY 2. POROSPORIDAE, Labbe. Each sporoblast gives rise to numerous sporozoites grouped round a residual mass, but the " spores " so formed are not enveloped in sporocysts (Fig. 41). Genus 22. Porospora, A. Schneider, 1875. Epimerite minute, button-like. Trophozoites large, septate, usually solitary (Fig. 1), some- times associated (Fig. 40). Unique species P. gigantea (E. v. Ben.), from the gut of the lobster. (6) SUB-TRIBE ANGIOSPOREA, L6ger. Spores well developed, with double sporocysts composed of epispore and endospore. FAMILY 3. GREGARINIDAE, Labbe* (Clepsydrinidae, Le"ger). Tropho- zoites with simple epimerites (Fig. 17, a). Cysts with or without sporo- ducts. Spores oval, in forms without sporoducts (Fig. 34, a), but in forms with sporoducts they are barrel-shaped (Fig. 34, c) and united in strings by their flattened ends. Genus 23. Gregarina, Dufour, 1828 (Clepsydrina, Hammer- schmidt, 1838). Epimerite conical or knobbed, rarely large. Cysts spherical or oval with sporoducts (Figg. 28 and 42). Spores barrel-shaped (Fig. 34, c). G. blattarum, Sieb. (Fig. 42), from the common cockroach Periplaneta orientalis ; G. ovata, Duf., from the earwig Forficula auri- cularia ; G. polymorpha (Hamm.), from the meal-worm ; and numerous other species, parasitic in the intestinal tracts of insects. Genus 24. Gamocystis, A. Schneider, 1875. Trophozoite with transitory protom., resembling a monotystid. Cyst with sporoducts. Spores elongated, THE SPOROZOA 197 FIG. 41. Two spores of Porospora gigantea, y. Ben. (par. Homarus vulgaris), show- ing the numerous sporozoites planted round a central mass of residual pro- toplasm. (From Lankester.) FIG. 42. Gregarina blattarum, Sieb. (par. Periplaneta orientalis). 1, a syzygy of two sporonts; a, nucleus. 2, ripe cyst, partially emptied ; a, channels leading to the sporoducts ; b, the re- maining spores ; c, endocyst ; d, the everted sporoducts ; e, the gelatinous epicyst. (From Lankester.) cylindrical. G. tenax, A. Schn., from the gut of the cockroach Ectobia lapponica ; G. ephemerae, Frantz, from the intestine of Ephemera, larva. Genus 25. Eirmocystis, Leger, 1892 (Hirmocystis, Labbe\ 1899). Epim. a conical knob. Sporonts forming syzygies of numerous individuals (Fig. 24). Cysts without sporoducts. Spores oval in form (Fig. 34, a). E. polymorpha, Leger, from the intestine of Limnobia, larva, and other species from the digestive tracts of insects. Genus 26. Hyalospora, A. Schneider, 1875. Cysts without sporoducts. Spores ellipsoidal, pointed at the ends, bulging in the middle. H. roscoviana, A. Schn., type-species, from the gut of Petrobius maritimus, and two other species. Genus 27. Euspora, A. Schneider, 1875. Cysts without sporoducts. Spores prismatic. Unique species E. fallax, A. Schn., from the gut of Rhizotrogu* aestivus. Genus 28. Sphaerocystis, Le"ger, 1892. Body of trophozoite spheroidal, with transitory Porospora gigantea, v. Ben. protom. Cysts without sporoducts. Spores oval ffidffin? to*^ £*?£ m form (Fig. 34, a). Unique species 8. simplex, of which the hindermost has Le"g., from the gut of Cyphon pallidus, larva. no obvious protomerite. (From ~ ° nt\ n -j AOTL -j Wasieiewski, after Leger.) Genus 29. Cnemidospora, A. Schneider, 1882. FIG. 40. I98 THE SPOROZOA Epim. large, lancet-shaped. Sporonta solitary, the body elongated and cylindrical in form, with globular protom. Cysts without sporoducts. Spores ellipsoidal with thick sporocysts. Unique species C. lutea, A. Schn., from digestive tract of Glomeris sp. Genus 30. Stenophora, Labbe", 1899 (nom. nov. for Stenocephalus, A. Schneider, 1875, preoccupied). Sporont oval, obese, with small conical protomerite. Cysts without sporoducte. Spores fusiform with a dark equatorial line. Unique species S. juli (Frantz.), from the digestive tract of millepedes, Julus sabulosus and terrestris, Spirobolus marginatus. FAMILY 4. DIDYMOPHYIDAE, Le*ger. Sporonts always associated in pairs, one behind the other, in such a way that the protomerite of the satellite disappears, and each syzygy resembles an individual with three chambers and two nuclei (Fig. 25, a). Genus 31. Didymophyes, Stein, 1848. Epim. in form of a cylindro- conical spike (Fig. 25, 6). Cysts dehiscing by simple rupture. Spores oval. D. paradoxa, St, from the intestine of Geotrupes stercorarius ; and three other species. FAMILY 5. DACTYLOPHORIDAE, Leger. Epimerite asymmetrical, irregular, bearing digitiform or root-like prolongations (Fig. 17, h, i). The dehiscence of the cysts is effected by simple rupture or by means of a pseudocyst (p. 183) placed laterally. Spores elongated, cylindrical (Fig. 34, 6). Genus 32. Rhopalonia, Leger, 1893. Epim. a subspherical knob bearing flexible digitiform processes. Trophozoites solitary, the conical body not septate, but with an indication of the protomerite. Unique species R. geophili, Le"g., from digestive tract of Geophilidae and of Stigmatogaster grarilis. Genus 33. Echinomera, Labbe, 1899 (nom. nov. for EchinocepJialus, A. Schneider, 1875). Trophozoite of oval or subconical contour, massive ; epim. persistent, spiked, the point furnished with small digitiform appendages which are not persistent, the whole forming with the protomerite a cone with summit displaced and slightly excentric (Fig. 1 7, h). Cyst dehiscing by simple rupture. Spores cylindrical with rounded bases, usually in strings (Fig. 34, 6). Unique species E. hispida (A. Schn.), from the gut of Litliobius forficatus. Genus 34. Trichorhynchus, A. Schneider, 1882. Cephalont with cylindrical or truncated protom., bearing an elongated conical rostrum. Cysts oblong with wart-like emi- nences ; dehiscence by means of a lateral pseudocyst. Spores cylindrical or ellipsoidal, not in strings. Unique species T. pukher, A. Schn., from the digestive tract of Scutigera. Genus 35. PterocephaluSj A. Schneider, 1887. Trophozoite with bilaterally symmetrical protom., divided into two lobes bearing spines or root-like processes, the two lobes united at one of their extremities to form a coiled horn (Fig. 17, i). Spores oval in form, connected obliquely into strings. Unique species P. nobilis, A. Schn., from the gut of Scolopendra spp. Genus 36. Dactylophorus, Balbiani, 1889. Protom. expanded excentrically, and carrying the digitiform processes of the epim. Sporonts solitary, of elongated form. Cysts dehiscing by means of a lateral pseudocyst. Spores cylindrical. Unique species D. robustus. Leg., from the gut of Cryptops hortensis. FAMILY 6. ACTINOCEPHALIDAE, Le"ger. Sporonts always solitary ; epim. symmetrical, simple or with appendages. Cysts dehiscing by THE SPOROZOA 199 simple rupture. Spores navicular, biconical, or cylindrical with conical extremities (Fig. 34, d). Parasitic for the most part in the guts of carnivorous Arthropods. (1) SUB-FAMILY SCIADOPHORINAE, Labbe. Protom. umbrella-shaped, with radiating ridges terminating posteriorly in recurved spines. Spores biconical, the epispore with equatorial, the endospore with polar, dehiscence. Genus 37. Sciadophora, Labbe, 1899 (nom. nov. for Lycosella, Leger, 1896, preoccupied), with the characters of the sub-family. S. phalangii (Leg.), from the gut of Phalangium crassum and P. cornutum ; two other species, also from Phalangidae. (2) SUB -FAMILY ANTHORHYNCHINAE, Labbe". Spores ovoid with pointed ends, joined in strings by an equatorial suture. Genus 38. Anthorhynchus, Labb£, 1899 (nom. nov. for Anthocephalus, A. Schneider, 1887). Epim. a large grooved knob. Unique species A. sophiae (A. Schn.), from the gut of Phalangium opilio. (3) SUB-FAMILY PILEOCEPHALINAE, Labbe*. Epim. simple, conical, or lance-like. Spores usually biconical. Genus 39. Pileocephalus, A. Schneider, 1875. Epim. shaped like a lance-head (Fig. 17, c). P. heerii (Koll.), from the gut of Phryganid larvae and P. chinensis, A. Schn., from the gut of Mystacid larvae ; two other doubtful species. Genus 40. Amphoroides, Labbe, 1899 (nom. nov. for Amphorella, Leger, 1892, preoccupied). Epim. spiked or globular; sporonts solitary ; protom. very short, compressed, hollowed out into a cup. Spores biconical. Unique species, A. polydesmi (Le"g.), from the gut of Polydesmus complanatus. Genus 41. Discorhynchus, Labbe", 1899 (nom. nov. for Discocephalus, Leger, 1892, preoccupied). Epim. large, in the form of boss surrounded by a thick ring ; protom. globular, larger than the deutom. Spores biconical, obese. Unique species, D. truncatus (Leg.), from the gut of Sericostoma sp., larva. (4) SUB-FAMILY STICTOSPORINAE, Labbe". Spores biconical, with slightly curved, pointed terminations ; the endospore with numerous little papilliform elevations. Genus 42. Stictospora, Leger, 1893. Epim. with a globular head, depressed ventrally and covered with projecting ribs terminating posteriorly in spikes. Spores biconical, the points slightly curved inwards. Unique species, 8. provincialis, Leg., from the gut of the larva of Melolontha and Rhizotrogus. (5) SUB-FAMILY ACTINOCEPHALINAE, Labbe". Epim. with appendages (except in Stylocystis). Spores symmetrical, navicular, biconical, or cylindrical with pointed ends. Genus 43. Schneideria, Leger, 1892. Trophozoite non-septate ; epim. a thick plate bordered by a rim composed of rib-like thickenings (Fig. 18). Spores smooth, obese, biconical. S. mucronata, Leg., from the larva of Bibio m,arcit and S. caudata (Sieb.), from the larva of Sciara nitidicollis. Genus 44. Stylocystis, Le"ger, 1899. Trophozoite non-septate; epim. in the form of a pointed bristle or sharp spine, usually curved. Sporonts solitary. Spores biconical. Unique species S. praecox, Le"ger, from the gut of the larva of Tanypus sp. (Diptera). While evidently closely allied to the foregoing genus, the epimerite lacks the appendages characteristic of the 200 THE SPOROZOA sub-family, and from the morphological point of view is intermediate between the simple epimerite of Doliocystidae and the more complex epimerites of Actinocephalidae (L^ger). Genus 45. Asterophora, L£ger, 1892. Epim. composed of a circular ridge with radiating rib -like thickenings, surrounding a prominent central papilla. Protom. ordinarily larger than the deutom. Sporonts solitary, of elongated form. Spores cylindrical with conical extremities. A. mucronata, L£g., from the gut of the larva of Rhyacophila, and A. elegans, L£g., from the digestive tract of the larvae of Phryganea grandis and Sericostoma sp. Genus 46. Stephano- phora, Le"ger, 1892. Epim. large, in form of a convex disc bearing a crown of finger-shaped tentacles. Spores as in the last. Unique species S. lucani (Stein), from the gut of Dorcus parallelipipedus. Genus 47. Bothriopsis, A. Schneider, 1875. Epim. in form of a lenticular knob bearing long flexible non- motile filaments. Sporonts solitary, with protom. greatly developed and very mobile. Spores biconical, obese. Unique species B. histrio, A. Schn., from the gut of Hydaticus sp. Genus 48. Goleorhynchus, Labbe", 1899 (nom. nov. for Coleophora, A. Schneider, 1885, preoccupied). , Sporont with protom. in the form of a sucker or strawberry, extending over the deutom. ; septum convex, projecting into the protom. ; deutom. subspherical or cylindrical. Spores navicular. Unique species C. heros (A. Schn.), from the gut of Nepa dnerea. Genus 49. Legeria, LabbR, 1899 (nom. nov. for Dufouria, A. Schneider, 1875). Protom. dilated, club-shaped ; septum convex, pro- jecting into protom. Spores subnavicular, with thick sporocysts. Unique species L. agio's (A. Schn.), gut of larva of Colymbetes sp. Genus 50. Phialoides, Labbe", 1899 (nom. nov. for Phialis, Le"ger, 1892, preoccupied). Epim. in the form of a retractile boss, surrounded by a circular ridge and a collar-like membrane with pleats terminated by triangular teeth. Sporonts massive, solitary. Spores biconical, obese. Unique species P. ornata (Le"g.), from the gut of the larva of Hydrophilus piceus. Genus 51. Geneiorhynchus, A. Schneider, 1875. Epim. in the form of a disc bristling with fine pointed teeth, carried on a very elongated neck (Fig. 1 7, g). Spores subnavicular. Unique species G. monnieri, A. Schn., from the gut of the nymph of Libellula. Genus 52. Actinocephalus, Stein, 1848.1 Epim. sessile or on a well-marked neck, and provided with hooks or spines. Spores biconical. A. stelliformis, A. Schn., from the gut of Ocypus olens and other beetles ; and other species. Genus 53. Pyxinia, Hammerschmidt, 1838. Epim. in the form of a cup or saucer with fringed rim surrounding a central spike (Fig. 15). P. rubecula, Hamm. (Figg. 15 and 33), from the gut of Dermestes lardarius and D. vulpinus ; and other species. Genus 54. Beloides, Labbe", 1899 (nom. nov. for Xipho- rhynchus, Le"ger, 1892). Epim. in form of disc or knob furnished with about ten teeth, and bearing in the centre a long spike (Fig. 17, «). Spores 1 The two species mentioned by Stein (Mailer's Archiv, 1848) under the genus Actinocephalus were A. acus. Stein, from Carabus glabratus and A. lucani from Lucamts (Dorcus) parallelipipedus. The former of these is not mentioned in Labb^'a Sporozoa, the latter is placed, following L£ger, under the genus Stepkanophora, L£ger. These facts may necessitate a revision of the nomenclature of the genera of Actino- ctphalidae, since the genus Actinocephalus in Das Thierreich does not contain either of the species placed in it by the founder of the genus. THE SPOROZOA 201 elongate-oval or boat-shaped. B. firmus (Le"g.), (Fig. 1 7, e\ and B. tennis (Le"g.), both from the intestines of larvae of Dermestes spp. FAMILY 7. ACANTHOSPORIDAE, L6ger. Sporonts always solitary. Epim. symmetrical, simple or with appendages. Cysts dehiscing by simple rupture. Spores garnished with bristles at the poles or equator (Fig. 34, «). Parasites of carnivorous insects. Genus 55. Corycella, Leger, 1892. Protom. spherical, more or less dilated ; epim. in form of a knob bearing a crown of eight large hooks (Fig. 16). Unique species C. armata, Le"g., from the gut of Gyrinus natator, larva. Genus 56. Acanthospora, L£ger, 1892. Sporonts solitary, of elongate oval form. Epim. in form of a conical obtuse knob. Spores oval, with a tuft of four bristles at each pole and an equatorial circlet of sharp spines. A. pileata, Le"g., from the gut of Omoplus sp., larva, and two other species. Genus 57. Ancyrophora, Leger, 1892. Sporonts solitary, the posterior end pointed. Epim. a knob bearing flexible or rigid appendages in the form of recurved hooks. Spores biconicaL, with polar tufts and six equatorial bristles (Fig. 34, e). A. gracilis, L4g., from Carabus auratus, 0. violaceus, and Silpha thoratica ; A. uncinata, L£g., from the larvae of Dytiscus, Colymbetes, Sericostoma, and Limnophilus rhombicits. Genus 58. Cometoides, Labbe", 1899 (noni. nov. for Pogonites, Le"ger, 1892, preoccupied). Epim. a spherical knob, flattened centrally, bearing a circlet of flexible slender filaments (Fig. 17,/). Spores with a tuft of bristles at each pole and two circlets of equatorial bristles. C. crinitus (Le"g.), from Hydrobius, larva ; G. capitatus (Le"g.), from Hydrous, larva, gut. FAMILY 8. MENOSPORIDAE, Le"ger. Sporonts solitary. Epim. sym- metrical, with appendages, and connected to protom. by a long neck. Cysts spherical, dehiscing by simple rupture. Spores in the form of crescents more or less curved (Fig. 34, 48> membranous rim with vesicular appendages. Protom. Hopiorhynchus oligc.- depressed. Cysts irregular, subspherical with areolar canthus (Sieb.), from the Vr- • r-'-/»eii_\ larva of Calopteryx. eminences. Unique species, L. insrgms (A. Schn.), (Prom Lankester.) from gut of Helops striatus (Fig. 44). Genus 62. Cystocephalus, A. Schneider, 1886. Epim. vesicular, with short narrow 202 THE SPOROZOA neck. Unique species, C. algierianus, A. Schn., from the gut of Pimelia sp. Genus 63. Oocephalus, A. Schneider, 1886. Epim. in the form of a rounded knob carried by a short conical neck. Unique species 0. hispanus, A. Schn., from the gut of Morica sp. Genus 64. Sphaerorhynchus, Labbe", 1899 (nom. nov. for Sphaerocephalu*, A. Schneider, 1886). Epim. small, spherical or oval, borne on a long, broad cylindrical neck, sharply constricted below the epim. Unique species S. ophioides (A. Schn.), from the gut of Akis. Genus 65. Stylorhynchus, Stein, 1848. The protom. of the cephalont is prolonged into a cylindrical, elongated rostrum, carrying at its termination the FiO. 44. Lophocephalua insignis (A. Schn.), (par. Helops striotus), showing the large epimerite and the nucleus with a band - shaped karyo- some. (From Wasielewski, after Leger.) FIG. 45. Stylorhynchus longicollis, Stein (par. Blaps morti' saga). On the left a cepha- lont, with a long epime- rite (a) attached to the protomerite (6). On the right a sporont, the epi- merite having been cast off. (From Lankester.) small knob-shaped epim. (Fig. 17, d)-} protom. of the sporont rounded ; the deutom. very elongated. S. longicollis. Stein, from the gut of Blaps mortisaga (Fig. 45) ; and other species. FAMILY 10. DOLIOCYSTIDAE. Epim. symmetrical, simple. Body non-septate. Spores oval. Parasites of marine Annelids. Genus 66. Doliocystis, Leger, 1893. Body showing no trace of protom. or septum. Spores oval, with a thickening of the sporocyst at one pole. D. pellucida (Roll.), from the gut of Nereis cultrifera and THE SPOROZOA 203 N. beaucoudrayi ; D. aphroditae (Lank.), from the gut of Aphrodite (Fig. 19) ; and other species. The following genera of Cephalina are of uncertain position : — Genus 67. Nematoides, Mingazzini, 1891. Trophozoite vermiform, without septum ; epim. in form of a fork or pair of pincers, borne on an elongated neck. Unique species N. fusiformis, Ming., from the gut of Bal- FlO. 46. Selcnidium, various species, a, comma -shaped species (" Selenidium en virgule") from Cirratulus cirratus ; ep, miaute epimerite, afterwards thrown off. b, semicolon-like species ("S. en point et virgule") from the same host, with very large epimerite. c-f, stages of S. echinatum, C. and M. (par Dodecaceria concharum). c, free sporont ; d, syzygy of two sporonts ; e, a spore, external view, showing the spiny surface ; /, a spore in section, showing the tour sporozoites. a and 5, x 500 ; c and d, x 300 ; e and/, x 850. (After Caullery and Mesml.) anus perforatus and Pollicipes cornucopia. Genus 68. Ulivina, Mingazzini, 1891. Body of elliptical form, protom. a quarter the length of the body. The external membrane forms a continuous sac round the animal. U. elliptica, Ming., from the gut of Audouinia filigera. Genus 69. Syria, Le"ger, 1892. Epim. knobbed, bordered by a thick ring (Fig. 17, 6). Protom. subspherical ; deutom. conical ; with numerous enclosures. S. inopinata, 204 THE SPOROZOA from the gut of Audouinia sp. Genus 70. Selenidium, Giard, 1884, emend. Caullery et Mesnil, 1899, incl. Esarhabdina, Mingazzini, 1891, Polyrhabdina, Ming., 1891, and Platycyttis, Leger, 1892. Body attenuated, vermiform, without septum, showing longitudinal striations due to myocyte fibrillae at the surface. Epimerite slender, conical, or large and globular (Fig. 46, a, 6). Spores spherical, spined, and exceptional amongst Greg- arines in being tetrazoic (Fig. 46, e,f). Parasites of Polychaetes. Type S. pcndula, Giard, from the body-cavity of Nerine ; S. echinatum, C. et M., from the gut of Dodecaceria concharum; other species from Scololepi$ fuli- ginosa, Cirrhatului cirratu$, and other marine Annelids. See especially Caullery and Mesnil [86]. ORDER 2. Coccidiidea The Coccidiidea are an order of the Telosporidia (p. 166), characterised by the following distinctive features. They are cell- parasites, attacking tissue - cells, and especially epithelial cells, rarely other forms of tissue, and never blood-cells. The trophozoite grows within the infected cell into an oval or spherical body, with great resemblance to an ovum ; it is quite motionless, never at any period amoeboid, and remains intracellular during at least the whole trophic stage. The dissemination of the parasite is always accomplished by means of resistent oocysts, the formation of which is preceded by the conjugation of differentiated gametes in all cases that have been thoroughly investigated. Within the oocyst the zygote breaks up into sporoblasts (archispores), which either become converted into naked sporozoites (gymnospores), or into spores (chlamydospores), each containing from one to four sporozoites, seldom more. In addition to this exogenous method of reproduc- tion, or sporogony, by means of durable cysts, the life- cycle is often complicated by endogenous multiplication, or schizogony, serving for the increase of the parasites within the host. The schizogony is not preceded by conjugation, and is not accompanied by formation of any oocysts or sporocysts. The Coccidia have attracted the attention of naturalists and medical men for a long time past, by their frequent occurrence in the rabbit and other Vertebrates, in which they may be present in such masses that their presence cannot fail to be detected by simple inspection when the host is dissected. Earlier observers often held, however, very erroneous views as to the nature of these parasites. Hake, who in 1839 was the first to describe Coccidia, regarded them as pathological products of the diseased animal — in fact, as a form of pus - corpuscles ; and similar views were held by many subsequent writers. On the other hand, a number of authorities in the forties and fifties believed Coccidia to be eggs of parasitic worms. Remak, in 1845, was the first to point out their relations to Midler's " psorosperms," and in 1854 Lieberkiihn insisted THE SPOROZOA 205 upon their affinities with Gregarines. A year later Kloss gave the first thorough account of the life-cycle, in the case of the form infesting the snail, subsequently named Klossia helicina by A. Schneider. Kloss's work was also the first proof of the existence of these parasites in Inverte- brates. The endogenous life-cycle was first described by Eimer in 1870, in the form infesting the mouse, termed by him Gregarina falciformit; but later (1875) made by Schneider the type of a new genus, Evmeria. Henceforth these organisms became known as " egg -shaped psoro- sperms " (eiformige Psorospermien, Psorospermies oviformes), and their affinities with the Gregarines received general recognition. In 1879 Leuckart greatly increased our knowledge of the Coccidian parasites of the rabbit, and introduced for them the new generic name Coccidium, in the second edition of his well-known treatise upon human parasites. From this time onwards these parasites were commonly known as " Coccidia " — a word often used in an extremely vague sense by writers whose zoological knowledge is defective, and by whom it is sometimes employed in a sense practically synonymous with the older word " psoro- sperms." In the eighties our knowledge of the forms of Coccidia and their life cycles was steadily increased, chiefly by the labours of Aime" Schneider, and in more recent years by Labbe\ In the last decade of the nineteenth century a vast amount has been written about Coccidia on account of the connection suspected to exist between them and cancer, but this work has been for the most part barren of results, contributing little to extend our knowledge either of cancer or of Coccidia. It is in this period, however, that the complete life-cycle haa been gradually worked out by a number of observers. An alternation of generations was first suggested by L. and R. Pfeiffer, whose ideas met with the most vigorous criticism, but a double life -cycle has now been demonstrated to be of almost universal occurrence amongst Coccidia. Towards the end of the nine- teenth century, also, sexual reproduction has been observed and accu- rately studied in a number of forms. The new century commences with an exhaustive monograph by Schaudinn upon the complete life-history of the forms infesting the centipede Lithobius, a publication which marks an epoch in the investigation not only of Coccidia but of Sporozoa generally, and completes our knowledge of a most fascinating chapter in natural history. (a) Occurrence, Habitat, Effects on their Hosts, etc. — The Coccidia are an abundant group of the Sporozoa, but appear to be confined, in the matter of hosts, to three great phyla — the Arthropoda, Mollusca, and Vertebrata.1 In the last named they are found 1 Exceptions are the Coccidian parasites discovered by Caullery and Mesnil in the gut of Capitetta capitata [126] and other Polychaete worms [129a]. Since only the schizogony was observed, the systematic position of these forms could not be deter- mined ; they remain for the present, therefore, without any generic or specific designation. On the other hand, these authors are of opinion that the alleged Coccidian parasites in Perichaeta, described by Beddard (Ann. Mag. Nat. Hist. (6), ii. 1888, p. 433), are nothing more than segmenting eggs of Nematodes. It is 206 THE SPOROZOA more commonly than any other Sporozoa, and have long been familiar on account of their frequent occurrence amongst domestic animals, both in birds and mammals, and even in man. They are met with in all the five classes of Vertebrates more or less commonly, and the very numerous species of the type -genus Coccidium are almost confined to Vertebrate hosts.1 In Mollusca, Coccidian parasites are very common in Gastro- poda and Cephalopoda, and Hyaloklossia pekeneeri, Leger, occurs in the kidneys of the Lamellibranch Tellina ; a species, of position as yet doubtful, occurring also in the kidneys of Donax (L4ger [41]). In Arthropods, Coccidia occur sparingly in Insects, more abundantly in Myria- pods, but have not been found as yet in either Arachnida or Crustacea. The Coccidia are chiefly parasites of epithelial cells, and since the infection of the host appears to take place in all cases by way of the digestive tract, it is the epithelium of the gut or of its appendages, such as the liver (Fig. 47), that is most often the seat of the parasite. In a considerable number of cases, however, the parasitic germs, after entering the system by way of the gut, go further afield before settling down. Passing through the gut-wall, the parasites are transported, probably, by the circulation of the blood or lymph, to their specific habitat. In those cases in which the vascular system forms the general body-cavity (haemocoele), we find occasionally, though very rarely, what is so common in the Gregarines, namely, Coccidia as "coelomic" parasites.2 possible that some of the supposed Coccidia seen in Polychaeta are really intra- cellular stages of Gregarines ; but a genuine Coccidian, Caryotropha mesnilii (Fig. 67) has recently been described by Siedlecki [55a] from Polymnia nebulosa. 1 Exceptions are two species found in Lithobius forficatus, viz. Coccidium lacazei (Labbe), and C. schubergi, Schaudinn. 3 An example is Adelea mesnili, Perez, 1899 [50J from the body-carity of Fio. 47. Section of rabbit's liver infected with Coccidium oviforme, Leuck. After Balbiani, from Wasielewski. THE SPOROZOA 207 As a general rule, however, the parasite selects some particular organ, most often the excretory organs.1 In Molluscs, especially, the kidneys are the seat of these parasites more often than any other organ (Fig. 48). In Arthropods this is less frequently the case, but Eimeria nova, A. Schn., is found in the Malpighian tubules of ....K FIG. 48. Klossia helicina, A. Schn., from the kidney of Helix hortensis, after Balbiani, from Wasie- lewski. a, portion of a section of the kidney, showing normal epithelial cells containing con- cretions (C), and enlarged epithelial cells containing the parasite (K) in various stages, b, cyst of the Klossia containing sporoblasts. c, cyst with ripe spores, each enclosing four sporo- zoites and a patch of residual protoplasm. Glomeris. In Vertebrates again the kidney is very often attacked ; in other cases amongst this phylum it is not infrequently the spleen, and even in a few instances the testis, which is selected by Coccidian parasites — never, however, the ovary ; so that in this the moth Tineola biseliella ; it attacks chiefly the fat -body, but may overrun also the pericardial cells, oenocytes, Malpighian tubules, muscles, and epidermis ; it is never found, however, in the gut - epithelium, nor does it penetrate the nervous system, gonads, or imagiual discs. Another example is Adelea akidium, Leger, parasitic upon various beetles (Akis spp. ; Olocrates abbreviate) ; it also attacks the fat-body and the pericardial cells, but not any other organs. 1 With regard to the question of the transport of the parasites within the body of the host, Laveran [35] has drawn attention to an association between Coccidium metchnikovi, Lav., and a Myxosporidian, Myxobolus ovtformis, Thel., in the gudgeon. The Myxobolus in the liver, spleen, and kidney is found containing the Coccidium in various phases of development, especially in the stage of cysts with spores, in which case the Myxobolus usually contains no spores of its own. Free Coccidia not contained in Myxosporidia are found only in the intestine. Laveran believes that the Coccidia penetrate the Myxosporidia in the intestine, and that the latter then invade the organs they affect, and transport the Coccidia with them. This view is contested by Blanch ard ([30], p. 161). 208 THE SPOROZOA respect the predilections of Coccidia are the opposite of those of Myxosporidia, which frequently attack the ovary but never the testis. A given species of Coccidian parasite may confine its attentions entirely to some particular organ, or it may attack several organs, as for example Coccidium minutum, Thel., found in the liver, spleen, and kidney of the tench; but as a rule it is rare for a form infesting the epithelium of the digestive tract to attack other internal organs as well. Coccidia during the trophic stage are always intracellular parasites,1 and each trophozoite destroys completely the cell which harbours it. As a rule the trophozoite lies in the cytoplasm and does not attack the nucleus directly, but pushes it to one side, often indenting or compressing it. The first effect of the extranuclear parasite is to produce a considerable hypertrophy of the host-cell, especially of its nucleus. The cell is stimulated to increased metabolism, shown not only in rapid growth, but also in the formation, by it of fatty substances, which serve as nutriment for the parasite and are consumed by it (Schaudinn [51]). The effects of the parasite are not confined to the cell which harbours it, but may extend to the surrounding tissues ; in Helix hortensis attacked by Klossia helicina the neighbouring epithelial cells of the kidney are stimulated to karyokinesis and multiplication, and a proliferation of the cells of the connective tissue is induced, leading to the formation of a fibrous envelope round the masses of Coccidia as a healing process on the part of the host (Laveran [38]). Ultimately, however, the infected cell is so weakened that it can no longer assimilate, but dies and is finally absorbed by the parasite, only a compact lump of chromatin and a small quantity of protoplasm remaining. The parasite then passes into the reproductive stage, either still enclosed by the remnants of the cell it has destroyed, in schizogony, or freed from the cell, in sporogony. A certain number of Coccidia occur, on the other hand, as intranuclear parasites. The schizogonous generations of certain species of Coccidium occurring in Amphibia (frog, salamander, newt) commonly attack the nucleus itself of the infected cell, and have hence been described by Steinhaus under the generic name Karyo- phagus. The recently described Cyclospora caryolytica, Schaud., parasite of the intestinal epithelium of the mole, owes its specific name also to its intranuclear habitat, which in this case seems to be an invariable characteristic of the parasite. The effects of this intranuclear parasitism have recently been studied by Schaudinn [5 la] and Dormoy [33], and are seen chiefly in an enlargement of 1 Very recently Laveran and Mesnil have described a species under the name Coccidium mitrarium (see p. 233), which, according to these authors, is unique amongst Coccidia in having an extracellular development like a Gregarine. THE SPOROZOA 209 the nucleus, accompanied by absorption of its contents. The linin framework is broken up, vacuoles are formed in it, and the chromatin fuses into irregular lumps and strands. The nucleus becomes enlarged to six or even ten times its normal diameter by absorption of fluid from the cell. The chromatic substance is forced out, by growth of the parasite, to the periphery of the nucleus, and ultimately disappears, so that "the entire nucleus is transformed into a gigantic vacuole, in the interior of which the parasite floats" (Schaudinn). The cytoplasm of the cell, on the other hand, is absorbed and shrivels up rapidly as the nucleus enlarges, without going through any stage of hypertrophy such as results from extranuclear parasitism. Each individual trophozoite in this way brings about the destruc- tion of a cell, but of one only, in its host. Nevertheless, the parasites are often present in such vast numbers that the epithelium of the organ affected may be completely destroyed, and the host itself killed or reduced to the last extremity. In centipedes experi- mented upon by Schaudinn, the faeces became milk-white during the acute stage of the Coccidiosis, and consisted entirely of epithelial remains and Coccidian cysts. The intestine may be so stripped of its epithelium that the young sporozoites are unable to find an epithelial cell to infect, in which case they may attack a full- grown Coccidian of another species, but never of their own kind (Schaudinn [51]). In the mole, Cydospora caryolytica is the cause of a pernicious form of enteritis accompanied by violent diarrhoea, which is generally fatal to the host (Schaudinn [5 la]). In rabbits young animals are often killed by the attacks of Coccidia infesting the epithelium of the bile -ducts, and similar cases are known in the human species.1 The liver is greatly enlarged, and its blood-vessels compressed, leading to functional derangements ; the secretion of bile is reduced to a minimum ; the blood becomes pale and watery, as in pernicious anaemia ; the respiration becomes gasping, and the animal finally dies in convulsions. In all these cases the destructive power of the parasite varies directly as its power of multiplying by schizogony, and so overrunning the tissues which it attacks ; and it is a very interesting and important fact, that in no case, apparently, can the schizogony continue indefinitely, but has its own natural, intrinsic limit, after which conjugation, with consequent sporogony, is necessary for the recuperation of the parasite and the continuance of its race. If, therefore, the patient can safely pass the acute stage, the disease heals itself through the failing reproductive powers of the parasite, on the one hand, and the regenerative capacity of the epithelium on the other. The injury inflicted on the host is repaired more or less completely; 1 For a full account of the pathology of Coccidiosis, with special reference to man, see Blanchard [30]. 14 210 THE SPOROZOA but the patient is by no means immune against the consequences of a fresh infection from without. In other Coccidia the schizogony may be wanting altogether, or be more limited in its duration, and in such cases the parasites are very harmless and inflict little or no injury upon their hosts. This is especially true of those found in Mollusca, commonly infesting, as has been said, the kidneys in these animals. (b) Morphology and Evolution. — The complete life-cycle of Coccidium schubergi has been worked out so thoroughly and in such detail by Schaudinn, that it may serve very well as a type of the whole order, the chief variations that are known to occur being specified afterwards.1 Coccidium schubergi is parasitic in the intestinal epithelium of Lithobius forficafus, where it is commonly found in company with two other species, Coccidium lacazci (Labbe), and Added orata, A. Schn. The infection of the centipede is started by its accidentally swallowing cysts with its food. The cyst-wall is then dissolved by the digestive fluids, the four spores each split lengthways, and the sporozoites, of which two are contained in each spore, are liberated in the digestive tract. Each sporozoite proceeds at once to attack, and to penetrate within, an epithelial cell of the host. The free sporozoite is a minute, sickle-shaped body 15-20/x in length, 4-6 /A in breadth (Fig. 49, a and b, and Fig. 50, a). The anterior extremity is more pointed and refringent, the posterior end more rounded. The finely-granulated protoplasm, which is not limited by any distinct cuticle, contains a spherical nucleus placed near the middle of the body, visible in life as a clear spot, and showing after preservation and staining a number of chromatin granules, lodged in an alveolar linin framework, but no special central corpuscle, nucleolus, or karyosome. The sporozoite performs active movements of various kinds. In the first place, it changes its form, as a whole, either by bending the body like a bow, and then straightening it out again, or by ring-like constrictions of the body 1 In the following account of the life-histories of Coccidia, the terminology employed for the various stages is that which has been gradually evolved by numerous authors during recent years, and to which Schauil>er>-/i, Schaud. (par. Lithobius forficatus). After Schaudinn [51]. <>, 1>, forward progression of a sporozoite by secretion of a gelatinous thread (g.s), which is attached to foreign objects, and pushes the little creature forwards. In b the portion of the thread between two foreign bodies has snapped and shrivelled up. c, a merozoite in forward progression. The arrow on the left shows the direction in which the merozoite is moving ; those on the right, the direction in which the gelatinous substance secreted by it is flowing backwards to form a filament, d-y, penetration of an epithelial cell by a sporozoite. H.C, host-cell ; N, its nucleus ; sp.z, sporozoite. from five to seven times its own length, it comes to a stop, bends its body three or four times, and starts again. Thus the sporozoite greatly resembles in its movements and general appearance a minute Gregarine. By means of its progression the sporozoite reaches an epithelial cell, and presses its anterior pointed end into it (Fig. 49, d). The opening is widened by its euglenoid contractions, and is still further increased by its movements of flexion and extension. In five or ten minutes it has worked its way into the cell (Fig. 49, e-g). Its movements then slowly cease, and it comes to rest near the nucleus, but sometimes a sporozoite traverses four or five epithelial cells before settling down. Within the epithelial cell the sporozoite becomes a motionless oval body, which absorbs the fatty nutriment provided for it by the cell (see above, p. 208), without, however, forming any fat-granules 212 THE SPOROZOA in its own substance or laying up any kind of reserve nutriment. It grows rapidly, becoming in twenty-four hours a full-sized, spherical trophozoite. Most remarkable are the changes which take place in the nucleus during the growth of the trophozoite. Larger frag- ments of the chromatin, which was at first scattered evenly in the nuclear framework, collect gradually towards the centre of the nucleus, where they soon appear imbedded in a diffuse, feebly- refractile substance, apparently allied to plastin in nature (Fig. 50, b). The pieces of chromatin fuse with the plastin matrix to form a solid spherical body, homogeneous in appearance, except for a few vacuoles of nuclear sap (Fig. 50, c-e). The body thus formed resembles the nucleolus of Metazoan cells in its appearance and relations, but differs in containing chromatin. It has therefore received the distinctive name of karyosome. The karyosome lies Fio. 50. Development of a sporozoite into a schizont, showing the formation of the karyosome, in Coccidium schubergi, Schaud. (par. Lithobius forflcatus). After Schaudinn f51J. a, sporozoite with a granular chromatic nucleus (n.sp.z) but no karyosome. 6, larger granules of chromatin appear towards the centre of the nucleus, c, the larger granules become more concentrated. d, they become united by a ground - substance into a central corpuscle or karyosome. e, schizont, with a large nucleus (n.szt) containing the karyosome (ky). towards the centre of the nucleus, or slightly excentrically. The rest of the nuclear framework retains its finely meshed condition, and lodges very minute chromatin -granules. The karyosome is retained through all the stages of schizogony, and its presence is absolutely distinctive of the schizogonous generations, but of them alone. When the trophozoite is full-grown and has exhausted the host- cell, it proceeds to reproduce itself by schizogony (Fig. 51, 1-IV), and is hence termed a schizont. The schizogony goes on within the host- cell, the withered remains of which form an envelope to the schizont, no cyst or protective membrane being formed by the parasite itself. The schizonts are distinguished by their coarsely alveolar or vacuo- lated protoplasm containing very few granular enclosures, if any. The nucleus of each schizont divides to form a number of daughter nuclei, which travel to the periphery and are scattered at more or less regular intervals at the surface of the cell-body. The proto- plasm adjacent to each nucleus then commences to grow out and THE SPOROZOA 213 project above the surface of the schizont, taking the nucleus with it. Thus are formed a number of club-shaped bodies, each very similar to a sporozoite, but differing from it in certain points of structural detail as well as in origin, and hence distinguished as a inerozoite (Figg. 51, IV, and 49, c). The parent schizont, which drops out of the host-cell at this stage, is not converted entirely into mero- zoites, but a certain amount of residual protoplasm is left, destined ultimately to be cast off and to die and break up. The schizogony here described takes place in a similar manner in many Coccidia, and has been frequently observed since it was first described by Eimer for the Coccidium falciforme of the mouse in 1870 ; but until recently the connection between the different parts of the life- cycle were not understood, and the schizogonous generations were con- sidered as representing a distinct generic type, to which A. Schneider in 1876 gave the name of Eimeria. Hence this portion of the life-history is often termed the Eimerian phase (" Cycle Eimerien "). The division of the nucleus of the schizont in the process of schizogony sketched above does not always follow the same method in all Coccidia, not even in the three species inhabiting Lithobius. In Adelea ovata and Coccidium lacazei it takes place by a multiple fragmentation of the nucleus and karyosome, the fragments coming together again at the periphery in patches to form daughter nuclei, each with a central karyosome. But in C. schubergi the nucleus divides by repeated binary fission (Fig. 52, a-e). The karyosome divides first in all cases, and then the chromatin forms two masses round each of the daughter karyosomes, which play a part in the division similar to that performed by the nucleolo-centrosome in Euglena and Paramoeba. The process is one more akin to direct nuclear division than to mitosis, and current descriptions, showing beautiful nuclear spindles, are inaccurate and imaginative (Siedlecki, Schaudinn). The number of merozoites formed is very variable, and is probably directly related to the nutrition furnished by the host -cell. Usually about thirty or forty, apparently, the number may sink as low as four. Simple binary fission of merozoites or schizonts never occurs however, since in all cases of schizogony, however much reduced, there is always left a residuary mass of protoplasm on which the merozoites are implanted all round, if numerous, or only on one side, if few. The merozoites, at first connected by a stalk with the residual protoplasm of the schizont, soon begin to exhibit active movements and wriggle themselves free. Each merozoite resembles a sporozoite in its movements and general appearance, and differs chiefly in being more club-shaped and in possessing a distinct karyosome. The mero- zoites proceed to seek out and to attack fresh epithelial cells, as did the sporozoites before them, and in a similar way each merozoite grows into a trophozoite which becomes a schizont, and breaks up in its turn into a fresh generation of merozoites. In this way schizogony may proceed merrily for many genera- tions, and the numbers of the parasite increase by geometrical 214 THE SPOROZOA rpsp THE SPOROZOA 215 progression within the host, until almost the entire epithelium of the digestive tract may be destroyed. Sooner or later (in C. schubergi after about five days) a limit is reached both of the nutritive capacity of the host and of the reproductive power of the parasite. Schizogony is then replaced by sporogony, a process always initiated by the production of sexually differentiated con- jugating individuals or gametes. Merozoites, descended from a long succession of maiden schizonts, infect epithelial cells and become Fio. 51. The life-cycle of Coccidium schuleryi, Schaud. (par. Lithobivs forjhxttus), represented in all its principal stages, combined into a single diagram, after Schaudinn [51]. I-IV represents the schizogony, commencing with infection of an epithelial cell by a sporozoite or merozoite. After stage IV the development may start again at stage I, as indicated by the arrows ; or it may go on to the formation of gametocytes (V). V-VIII represent the sexual generation. The line of development, hitherto single (I-IV), becomes split into two lines— male (VI <5, VII