Notes on Sporozoa parasitic in Tubifex. 1. The Life-history of Triactinomyxon Stole. Doris L. Mackinnon, D.Sc. D. lues Adam, B.Sc.

Transcription

1 Notes on Sporozoa parasitic in Tubifex. 1. The Life-history of Triactinomyxon Stole. By Doris L. Mackinnon, D.Sc. (Roader in Zoology, University of London, King's College), and D. lues Adam, B.Sc. With Plates 6-8 and 3 Text-figures. THE Thames mud below Waterloo Bridge is full of reddish masses of small interlaced worms ; these are the oligochaetes Tubifex tubifex (Mull.) (=T. rivulorum Lmk.) and Limnodrilus hoffmeisteri Clap. While examining their gut-contents we found in Tubifex three species of that rare and little-studied organism Triactinomyxon &tolc, and having such large quantities of the host at our very doors we continued our search in the hope of elucidating some parts of the life-history. Though we dissected 1,250 Tubifex we found the actinomyxidian only five times. This is approximately in agreement with Stoic's statement that only one worm in 300 is infected. In three worms the infection was heavy, and in the other two it was comparatively light. 1 THE GENUS TRIACTINOMYXON STOLC. GENERAL CONSIDERATIONS. The Actinomyxidia are a group of spore-forming parasites which are but little known. Pour genera have been recorded from tubificid worms, and one from a sipunculid. They are 1 In the course of our examination we saw two astomatous ciliates and two other sporozoa, Myxocystis oiliata Mrazek and a new coccidian ; we shall describe these elsewhere. NO. 270 O

2 1S8 DORIS L. MACKINNON AND D. INES ADAM classified as Cnidosporidia, among which the peculiar ternary symmetry of their spores is regarded as the chief diagnostic feature. Stole, their discoverer, considered that they are not protozoa' at all, but organisms having affinities with the so-called ' mesozoa'. We do not propose to discuss this question here ; but it seems to us that if the Actinomyxidia are protozoa, then an extension must be made of the definition usually given of that group of the animal kingdom. The genus Triactinomyxon was first described by fetolc in 1899 in a paper on the Actinomyxidia. This work Avas written in a Bohemian journal, but a summary in German was given by Mrazek in the same year. In 1904 Leger described more fully the type-species Triactinomyxon ignotum, and gave larger and more detailed drawings of the spore and its contents. He also described part of the life-history, gave figures of two developmental stages, and stated that, so far as he had been able to trace it, the process of spore-formation in T. ignotum agrees with Caullery and Mesnil's description of the development of Sphaeractinomyxon stolci. The papers by these authors to which Leger refers were preliminary notes published in 1904 ; the full account was not published until the following year. Triactinomyxon ignotum Stole is described as having spores shaped like three-rayed anchors, each crowned by a group of three polar capsules, below which is the multinucleate sporoplasm. Three parietal cells make the sporal envelope, in the upper part of which the sporoplasmic mass lies. The parietal cells are quite transparent, and their nuclei usually lie at the point where the three rays of the anchor diverge. The ripe sporozoites 1 are eight in number ; each contains a single vesicular nucleus and is shaped something like a swollen spindle markedly pointed at one pole. In the protoplasm in which these bodies lie embedded there are also two residual nuclei and a ' chromatin ball'. Neither btolc nor 1 We are making use of the terminology employed by other workers on the Actinomyxidia, though it is very doubtful whether such a word as ' sporozoite ' is applicable here.

3 SrOROZOA IN TUB1FBX 189 Leger gives the spore's dimensions ; but judging from their drawings, the magnification of which is stated, the sporoplasmic mass in this species measures from 30/i to 40/x x7/i, the axial portion of the sporal envelope is 112ja long by 17/i broad, and the branches are about 150/* long. Leger (1904b) also records another species of Triactinomyxon from Tubifex. To this he gives no name, merely remarking that it contains thirty-two sporozoites, ' en forme de navette aplatie dont rindividualisation s'effectue le plus souvent alors que la spore est deja completement formee '. x The Actinomyxidia that we have found in the gut of Tubifex all belong to the genus Triactinomyxon, but only in one instance did we find T. i g n o t u m, and that was in the last infected worm we examined. The first three infected Tubifex harboured a form with twenty-four sporozoites in the spore ; this seems to be a new species, and we have named it T. 1 e g e r i. The actinomyxidian in the fourth worm was different from this in certain respects, which we shall note later, but was essentially distinguishable from the other species in having fifty or more sporozoites to a spore; we shall describe it as T. mrazeki, n. sp. The bulk of our material, then, belongs to T. legeri, and it is this that we have most fully studied. We propose to begin, therefore, by describing the life-history of this species in as full detail as we can ; then we shall compare the other two species with it. METHODS. Our study has been based almost entirely on smears of the gut-contents of infected Tubifex. We cut sections of part of a worm infected with T. ignotum, but examination of 1 Granata (1922a) described a species of Triactinomyxon, T. magnum, with a sixteen-nucleate sporoplasm, from the bodycavity of Limnodri 1 us udekemianus, and he worked out some parts of the life-history. We regret that we did not know of his paper until our own work had gone to press. Though we are therefore unable to criticize it in detail, we shall refer to some points of interest by means of foot-notes. O 2

4 190 DORIS L. MACKINNON AND D. INES ADAM them did not add materially to the data we had collected from the smears. These were fixed with Schaudinn's or Bouin's fluid, and stained with Heidenhain's iron-haeniatoxylin, Delafield's haematoxylin, Ehrlich's haematoxylin, Dobell's ironhaematein, and borax-carmine. Delafield's haematoxylin counterstained with eosin gave much the best results, and all our drawings are from smears stained in this way. Triactinomyxon legeri, n. sp. As in the other Actinoniyxidia, the pan^porocyst contains eight spores (PI. 6, fig. 1). An average-sized pansporocyst has a diameter of about 60/n, and the tails of the spores are closely folded and telescoped within its narrow confines. When it ruptures, the spores escape and spread out, and then the three-rayed anchor form becomes apparent (Text-fig. 1). Neither Stole nor Leger has sufficiently emphasized the delicacy of the sporal envelope ; their drawings suggest something very conspicuous and rigid. As a matter of fact, it is of membranous consistency and the branches are easily folded and distorted ; when cleared and mounted in ordinary Canada balsam this part of the sporal apparatus is almost invisible, and we overlooked it at first. The general form of the spores is like that of the type-species ; but they vary a good deal in size. The total length to the point of trifurcation is somewhere between 90/u, and 140/u, with a width of from 11 n to 16 JU. ; the length of the branches depends on the degree of their expansion 150/x is a common length ; their maximum breadth varies from 8/x to 14/n. In the great majority the branches are imperfectly evaginated and appear to have truncate fluted ends (Text-fig. 1 a and PI. 6, fig. 2). Only a few spores were found with their tips completely extended, and they were seen to taper to fine points (Text-fig. 1 b). The sporoplasmic mass is from 15fi to 20/x in length. When ripe it contains twenty-four sporozoites, disposed, most usually, in three longitudinal rows of eight, though this arrangement is easily disturbed (PI. 6, fig. 3). The sporozoites at this stage

5 SPOROZOA IN TUBIFEX 191 are ovoid bodies (2-7/* x2-2ja is their average size), containing a relatively large nucleus (1-3 n) and a vacuole alongside it in the darkly staining cytoplasm. The nucleus has a very distinct karyosome and peripheral chromatin ; we could see nothing of the centrosomic body mentioned by Leger. The twenty-four sporozoites lie embedded in the sporoplasmic Spores of Triactinomyxon lcgori. (a) With the branches unexpanded, (6) with two of the brandies fully and the third partially expanded, x 370 approx. mass, which stains less densely than they. It contains at least three residual nuclei, rather larger and less compact than those of the sporozoites, and there is always a ' chromatin ball'. Above the sporoplasm lie the three capsulogenous cells, of which the nuclei show very plainly ; and from the polar capsules a filament may protrude.

6 192 DORIS L. MACKINNON AND D. INES ADAM This is the ripe spore. The next stage may be gone through before the spores are discharged from the pansporocyst. It consists in the fusion, in pairs, of the sporozoites (PI. 6, fig. 4). The nuclei of these remain distinct. At the end of this process there are twelve bodies in the sporoplasm : the3 r are twice the size of the uninucleate sporozoites, and each contains two nuclei, sometimes separated by a vacuole. The residual nuclei and the chromatin ball remain unchanged. While this is going on, the sporoplasm begins to slip down the sporal sheath towards the point of origin of the three rays, where, according to Leger, the aperture lies. 1 It seldom descends en masse ; more commonly it breaks into two or three portions, each containing a few of the sporozoites and at least one of the residual nuclei. PL 6, fig. 5, shows the approach of a sporoplasmic fragment to the point of exit. We have not seen these fragments in the act of escape ; but that they do get out is clear, for all around they can be seen lying free, and there are abundant empty spores. We think that the free sporoplasms become amoeboid and probably crawl about for a time, for they often show pseudopodia (PI. 6, fig. 6). How long this condition lasts we are unable to say ; but sooner or later the sporoplasm breaks down and liberates the binucleate germs ; sometimes these are all set free at once (PL 6, fig. 7) : sometimes they are deposited a few at a time. Since they tend to be liberated in batches, it naturally follows that two or three may lie together during the subsequent development ; but this development is not necessarily synchronous throughout a group. The freed germs now enlarge, their cytoplasm becomes much 1 It ma3 r be objected that, since we were working with stained material and have not observed the living sporozoites fuse together, we may have reversed the natural process and have mistaken for fusion what is really a division of twelve binucleate bodies into twenty-four uninucleate. To this we would reply that in the hundreds of sporozoites examined we never saw any trace of nuclear division, and that in this species we never found sporoplasms with uninucleate sporozoites descending the tube of the spore or outside it in the body-fluid of the worm : such stages were always binucleate.

7 SPOROZOA IN TUBIFBX 193 more vacuolated and less deeply staining; and there is evidence, such as pseudopodial processes and ingested particles, that they crawl about as amoebulae (PI. 6, fig. 8). Their size is variable, but 11/x is the maximum diameter. Their nuclei also increase in size ; an amoebula of 10-5/x has two nuclei 3-25 [x in diameter. The karyosome of such nuclei is relatively large and more ' smudgy ' than in the earlier stages. The nuclei may lie close together, but they never fuse. At first they are of the same size and appearance ; but as they swell up and stain less densely one gets ahead of the other, as it were, and heteronucleate amoebulae result (PI. 6, fig. 9 a, b, and c). This is the prelude to nuclear division. The larger nucleus divides first (PL 6, fig. 10) ; the other divides soon after, and a four-nucleate condition is produced (PL 6, fig. 11 a and b). Two of these nuclei with some associated cytoplasm now take up a peripheral position, and there they form a cyst round the rest of the mass (PL 6, fig. 12). During this process the organism increases in size. The internal nuclei now swell up considerably ; they may have a diameter of 5-25/x, their karyosomes are excentric and more faintly staining than heretofore, and the rest of the chromatin is distributed in a spireme (PL 6, fig. 13). This division of the young parasite into two portions, comparable with the germcells and the soma of metazoa, has been noted by other workers on the Actinomyxidia who have seen the early stages. Nothing further happens to the cystogenous cells ; but within the central mass there is now a division of the cytoplasm, so that two cells are formed (PL 6, fig. 14). Further divisions follow (PL 6, figs. 15 and 16), and a four-cell stage, is reached. 1 Prom the time that the first two cells become binucleate there is a marked difference in their rate of division. One cell, a, divides three times while the other, j6, is dividing once. Figs. 15 to 19 show stages in this process. In fig. 16 there are three binucleate a cells (one a 1 and two a 2 ) and one binucleate /?; in fig. 17 one of the a 2 cells has again 1 A four-cell stage in which the cells are still unimicleate is shown on PI. 8, fig. 37, among the drawings of T. ignotum.

8 194 DOBIS L. MACKINNON AND D. INES ADAM divided, so that there are four small cells, two binucleate and two uninucleate (the j8 cell is still undivided); in fig. 18 there are four binucleate a 2 cells and /? has divided once. Finally we come to fig. 19 (PI. 7), where there are eight uninucleate a 3 cells and still only two uninucleate /J 1 cells. This stage is comparable to the ten-cell stage drawn by Leger ; the two large ^S 1 cells are his ' cellules meres des germes ', and the other eight are his ' cellules meres de la paroi sporale '. We have found a fair number of Actinomyxidia at this stage, and we conclude that it persists for some time ; fig. 38 (PI. 8) shows the corresponding stage in T. i g n o t u m, in which the js 1 cells lie uppermost and can be more easily seen. Leger states that the development of Triactinomyxon agrees ' dans ses grandes lignes ' with that of Sphaeractinom y x o n as described by Caullery and Mesnil; but at this point the ten-cell stage his interpretation departs notably from their final account (1905) and from that given by Ikeda (1912) of the development of Tetractinomyxon intermedium. These authors describe a sexual process, half the gametes being represented by the a 3 cells of the ten-cell stage, and the other eight being derived by further divisions of the js 1 cells. Ikeda furthermore describes and figures the reduction process in both sets of gametes, with formation of sixteen reduction bodies, which apparently take no more part in spore formation, though they tend to unite in pairs. In Sphaeractinomyxon the gametes are of different sizes; in Tetractinomyxon there is isogamy. The eight spores are developed by subsequent division of the zygotes. Leger ignores this sexual part of the story altogether. Evidently he noticed the formation of the reduction bodies, for he says, ' au cours de ce developpement de nombreuses boules chromatiques sont expulsees dans le kyste au moment des mitoses ',* but according to his account the eight /? 3 cells alone are responsible for the sporal contents, while the a 3 cells, each 1 Granata (1922a) does not describe any reduction process in Triactinomyxon magnum. The gametes are of two kinds, distinguished by a difference in chromaticity.

9 SPOROZOA IN TUBIFEX 195 dividing into six, give the spore envelope and the polar capsules. In the later stage the pansporocysts of T. legeri are relatively large and opaque, and we are unable to supply information as to every detail in the process ; but enough can be seen to convince us that Leger's interpretation was wrong, and that the spore formation in T r i a c t i n o m y x o n, as in other known Actinomyxidia, is preceded by a fusion of gametes. In interpreting a stage such as that shown in PI. 7, fig. 20, we must assume that the fi cells have divided twice, for we have in the pansporocyst at least fifteen cells of two different sizes. There has been a reduction of their nuclear substance, as is evidenced by the presence in the pansporocyst of a corresponding number of dark-staining chromatin masses, usually grouped in pairs. 1 Presumably the larger cells ( S gametes) now fuse with the smaller (a gametes). In PI. 7, fig. 20, one such fusion has already taken place, where a large and a small nucleus are lying in close apposition in the largest cell. Here then, as in S p h a e r a c t i n o m y x o n, there is anisogamy. The fate of the zygotes is as follows. The zygote nucleus divides to give two, four, six, and then seven nuclei (PI. 8, fig. 42 a, b, c, d (T. ignotum), and PI. 7, figs. 21 and 22 (T. legeri)). Three of these go to one end of the mass and, with their associated cytoplasm, become the capsulogenous cells (PI. 7, figs. 23 and 24). Three at the opposite pole belong to the cells that form the sporal envelope (PI. 7. fig. 24, sp.n.). The remaining nucleus lying in the cytoplasm between these two groups divides many times (PI. 7, fig. 24) until twenty-seven nuclei are formed, twenty-four of which are the nuclei of the sporozoites, and the other three are the residual nuclei of the sporoplasm itself. From the first the chromatin ball is present in the sporoplasm, and undoubtedly it is formed of a pair of fusing or fused reduction bodies. 2 (In 1 We have been unable to find stages showing clearly the mode of formation of the reduction bodies. 2 Granata (1922 a) notes the presence of the chromatin ball in the sporoplasm. but can say nothing as to its origin.

10 196 DORIS L. MACKINNON AND D. INES ADAM figs. 22 and 24 it is still clearly a double structure.) Except where the cyst has burst and allowed the developing spores to escape, it is impossible to make out these later details, so dense is the mass within. The chief point of interest in these observations of ours on the spore formation is our demonstration that, throughout, the sporal envelope and its contents are in intimate contact and develop together. 1 We are unable to substantiate Caullery and Mesnil's extraordinary statement, with which Leger agreed, i. e. that the germinal mass develops apart from the spore envelope, which it then enters at maturity. We think it possible that the mistake may have arisen either from the great difficulty of interpreting sections of the maturing cysts, or from their taking as normal certain appearances which are produced if the contents of the cyst have shrunk or dried a little. The portion of the developing spore that gives rise to the sporoplasm usually lies against the wall of the cyst, while the cells of the envelope and the capsules are grouped together towards the middle. If there is any shrinkage or drying the sporoplasms tend to flatten out and cleave still more closely to the cyst wall, and so delicate and transparent is the envelope at this stage that its connexion with the peripheral mass might very easily be missed. The onlj- way to demonstrate the continuitj 1 - is to examine entire sporoblasts that have escaped from broken cj^sts ; and this is what we have done. From what we have said it is clear that there must be an auto-infection of the host. Empty spores are too common to have been caused by making the smear ; the early stages are not very hard to follow, and the liberation of the sporoplasmic masses is one of the most striking features. Stole assumed that auto-infection must occur, since he could not otherwise account for the vast numbers of the parasites in an infected worm, and other observers have been of the same opinion ; Ikeda found remains of empty spores of Tetractinomyxon inside the worm's phagocytes. The most controversial point in what we have described 1 Granata's observations on this point agree with ours.

11 SPOKOZOA IN TUBIFEX 197 is of course the formation of the binucleate amoebulae. A brief review of other workers' conclusions may make this clearer. (1) Leger confessed himself unable to interpret the earliest stages in the development of Triactinomyxon ignotum. He found free uninucleate and binucleate bodies, obviously derivable from the sporozoite stage, and he suggested that the binucleate forms might have arisen from the fusion of two uninucleate sporozoites. But he did not see this happen, and he does not record anything like what we describe, i. e. the fusion within the spore of two uninucleate sporozoites and the subsequent development of these into binucleate amoebulae. (2) Caullery and Mesnil also missed the startingpoint of the development of S p h a e r a c t i n o m y x o n and were reduced to conjecture. They could not decide whether the free binucleate stage arose by division of a single nucleus or from fusion of two distinct uninucleate bodies. (3) Ikeda, whose work at this point is much the most exact, dealt with an actinomyxidian that begins with a binucleate sporoplasm ; he found that the four-celled stage was reached from the association outside the spore of two such sporoplasms. (4) Granata (1922) admitted himself very uncertain as to what interpretation he should put on the mononucleate and binucleate elements which are the initial stages in T r i - actinomyxon magnum from the body-cavity of Limnodri 1 us and in Neoactinomyxon globosurn from the intestine of the same worm. He thought that the mononucleate bodies probably divide to give binucleates ; but he gave no convincing evidence of this, and stated that the later stages leading to pansporocyst formation escaped his observation. While bearing in mind the fact that Tetractinomyxou is certainly less closely comparable with our parasite than are any of the other described genera, still it seems to us that our observations are not really contradictory to Ikeda's. In both there is an association of two individual ' germs ' ; it is merely more precocious in Triactinomyxon than in Tetractinomyxon, and in Triactinomyxon legeri

12 198 DORIS L. MACKINNON AND D. INES ADAM it goes so far as plastogamy within the spore, whereas in the other genus the associating cells remain distinct. The important point of agreement seems to lie in the association of two separate bodies to form the starting-point of the new pansporoblast; that in one the union occurs earlier and is more complete than in the other does not outweigh, we consider, the essential resemblance of the process. We are confirmed in this TEXT-FIG. 2. Spores of T r i a c t i n o m y x o n mrazeki, (a) With the branches partially and (6) fully expanded, x 370 approx. opinion by what we have seen of the corresponding stages in the two other species of Triactinomyxon. The somatic and the germ nuclei in Tetractinomyxon are said to be different from one another from the very first; in Triactinomyxon legeri these nuclei are in the beginning alike, but their difference becomes apparent very soon after liberation of the sporozoite. Here, again, it is merely a difference

13 SPOROZOA IN TUBlFEX 199 in the rate of development ; there seems to be no essential difference in the process itself. Triactinomyxon rnrazeki, n. sp. The pansporocyst and sporal envelopes have the same general characters as those already described for T. 1 e g e r i. The main tube of the sporal envelope is rather longer than in that species ; 150/u is a common measurement, with width up to 12/w (Text-fig. 2 a and b). The sporoplasmic mass is very much larger on the whole ; it varies in length from 25 p to as much as 65fi, and contains anything from 50 to over 100 uninucleate sporozoites (PI. 7, fig. 25). These are ovoid, very small (1-6/x. xl-3fi to 1-9/x xl-6/x), very densely staining, and all are uninucleate. The nuclei are of the type already described in the sporozoites of T. legeri, but correspondingly smaller. There are three residual nuclei in the sporoplasm, but no trace of a ohromatin ball. In this species the sporozoites do not seem to unite within the spore : the sporoplasmic mass descends and disperses them while they are still uninucleate ; but the union probably takes place soon after their emergence, for we found a number of binucleate amoebulae (PI. 7, fig. 26 a and b) comparable with those of T. legeri. The infection with this species was slight, and we were unable to trace the life-history farther. Triactinomyxon ignotum btolc. Just as we were finishing our notes and drawings of these new Actinomyxidia, we came on a worm richly infected with a Triactinomyxon that proved to be the type-species. From our study of this material we can say that the life-history is similar to that of T. legeri. We found most of the developmental stages already described and some that are supplementary to these. We propose, therefore, to make notes and drawings only of the points of difference between the species and of such additional data as may help to complete the life-history of the genus as a whole. The spores of T. ignotum differ from the other species

14 200 DORIS L. MACKINNON AND D. INKS ADAM in containing only eight sporozoites in the sporoplasm. We find that the length of the sporal envelope to the point of branching is 110/x to 165/x, with a width varying between 11/x and 15jz. The sporoplasniic mass is from 20/x to 30 /u. long (these measurements agree pretty closely with those we have deduced from Leger's figures). The sporozoites are ovoid and very much larger than those in the other species: 4-5(ix3-5^ is an average measurement (PI. 7, fig. 27). Leger states that the sporozoites become ovoid and then turn into swollen spindles with a short rostrum at one of the poles, and as such he draws them. We have often seen the sporozoites of this and of the other species assume a spindle shape in badly fixed or slightly dried spores, but we think it is quite unnatural. The size of the nucleus depends, we find, on the age of the sporozoite ; when these are first delimited from the surrounding sporoplasm, their nuclei are almost as large as they are ; but before the sporoplasm escapes they become smaller and more condensed until they are about 1-6/A in diameter. They do not show the dense peripheral chromatin and central karyosome of the other species ; 1 they stain much more faintly, and have a beautiful reticulum and a small, excentric karyosome : in fact, their resemblance to the nuclei of later stages is very close. There are two residual nuclei and a chromatin ball. All the spores contain sporozoites in the uninucleate condition ; there is no hint of any pairing of the sporozoites within the spore. PI. 7, fig. 28, shows a sporoplasm descending its envelope, and it will be seen that there is no change in the contained bodies except that their nuclei are rather smaller than in PI. 7, fig. 27. The sporoplasm seems to escape as a whole, and to break up very quickly ; the liberated sporozoites lie in groups of six or eight together (PI. 7, figs. 29 and 30). Here we are met with the difficulty that baffled Leger and 1 We think it very possible that this difference may be more apparent than real, and that it depends on the degree of differentiation of the haematoxylin preparation. Our material was not large enough to admit of our experimenting much with the stains we employed.

15 SPOROZOA IN TUBIFEX 201 Granata. But after long and careful consideration of our material we have come to the conclusion that in this species, as in T. legeri, there is certainly an association of two sporozoites to form the starting-point of each pansporocyst. This association takes place after the bodies have escaped, and it may occur (1) while they are still recognizable as sporozoites and are even yet lying in the degenerating sporoplasm, or (2) at a later stage when the development of the unmucleate bodies has gone so far that their wandering as amoebulae is over. The first of these methods by which the binucleate ' germ ' of the pansporocyst is obtained is shown in PI. 7, figs. 30 and 31 a, b, c, and d : the product is a body comparable with the binucleate amoebula of T. legeri, but showing no indication that the nuclei are of different kinds. The second method is that in PI. 7, figs. 32 and 33 a, b, where a sporozoite has remained free until it is of relatively large size, then has laid itself alongside a similar cell to initiate a pansporocyst. Obviously the condition in PI. 7, fig. 33, can be reached, and probably is reached, through the first method by division of a cell like that of PI. 7, fig. 31 d, into two ; but we are unable to distinguish the two-celled stage arrived at by (1) from those derived at by (2). The essential thing seems to be that there shall be an association of two sporozoites, and we are convinced that all the early binucleate stages arose in this way and not as the result of nuclear division within a single cell. It does not seem to matter whether the association is within the spore, as in T. legeri, or of newly escaped sporozoites still in the sporoplasm, or later still when these are free and fully grown ; nor does it seem to matter whether there is, or is not, plastogamy. What does matter is that somehow or other the new pansporocyst and its contents shall be the product of two sporozoites. The two-celled stage is followed by a four-celled (PI. 7, figs. 34 a, b), in which there are two small cells apparently cut off from two larger. This is comparable with the fournucleate condition in the amoebula of T. legeri, only there, for some reason, cell-division lags behind nuclear

16 202 DORIS L. MACKINNON ANT) D. INES ADAM division. The t'wo smaller cells make the cyst ; the two larger ones are the gametocytes (figs. 35 to 40). PI. 7, fig. 36, shows a three-celled condition of the cyst-contents, which Leger says he failed to find. Figs. 35 and 38 are comparable with the only two figures he gives of the life-cycle. In PI. 8, fig. 39, a stage similar to that in fig. 38 is given, but showing the first division of the lagging /J 1 cells ; it will be seen that there is a typical mitosis. In PI. 8, fig. 40, most of the gametes are formed, and they are of two sizes. Thirteen reduction bodies have been cast out, and apparently two of the )8 2 cells are dividing to form gametes. We noted that in T. legeri there was a tendency for the number of reduction bodies to be less than sixteen in the end, because they often fuse in pairs ; in T. ignotum, however, it is common to find twenty-four or even thirty in a ripe cyst, and we can only conclude that in this species either there are two reduction bodies from each gamete, or, as seems more probable, that these tend to divide after their formation. The gametes unite in pairs, and probably a large unites with a small, but in PL 8, fig. 41, where the eight products of this union are represented, only one shows the final zygote nucleus ; in the others there are still two nuclei, and they appear to be of equal size. The zygote nucleus divides, and PL 8, figs. 42 and 43 show different stages in this preparation of the spore nuclei. In PL 8, figs. 44 and 45, these have taken up their positions in the developing spore. SUMMARY AND CONCLUSIONS. 1. There are four known Actinomyxidia belonging to the genus Triactinomyxon Stole, parasitic in T ubif ex tubifex. 1 These are : (1) T. ignotum &tolc, with eight sporozoites in each spore. 1 To these must now be added T. magnum Granata from the bodycavity of Limnodrilus udekemianus, with a sixteen-nucleate sporoplasm and especially characterized by the great length of the branches of the sporal envelope up to 500 p.

17 SPOROZOA IN TUBIFEX 203 (2) T. legeri, n. sp., with twenty-four sporozoites in each spore. (3) T. sp. Leger, with thirty-two sporozoites in each spore. (4) T. mrazeki, n. sp., with more than fifty sporozoites in each spore. 2. We have investigated the life-histories of (1) and (2) and, in part, of (4), and we find that there is certainly autoinfection of the host. Sporoplasms can be seen escaping and lying free in the gut-contents, where apparently they creep about; and empty spores are common. 3. There is an association of two uninucleate sporozoites or their immediate derivatives to form the basis of each pansporocyst and its contents. This association may take place in the spore (T. legeri) or after the escape of the sporoplasm (T. ignotum and T. mrazeki). In T. legeri and T. mrazeki this association amounts to plastogamy ; in T. ignotum there may be plastogamy or there may not. 4. Development follows essentially on the lines of Sphaeractinomyxon stolci and Tetractinomyxon intermedium. Cystogenous and germinal cells are distinguishable at the four-cell stage in T. ignotum, but in T. legeri the nuclei divide before any cell-division takes place. The germinal cells divide by typical mitosis to give eight small and eight larger cells, which after reduction fuse in pairs, small with large. The reduction bodies tend to pair together (T. legeri) or may divide ('? T. ignotum). 5. The nucleus of a zygote divides until there are seven nuclei, three of which are the nuclei of the capsulogenous cells, three belong to the cells forming the three-rayed sporal envelope, and one is the first nucleus of the sporoplasm. This nucleus divides until there are in the sporoplasm as many nuclei as sporozoites ; there are in addition two or three residual nuclei. The chromatin ball in the sporoplasm is an engulfed reduction body (or pair of fused reduction bodies); it is absent from T. mrazeki. Throughout the develop- NO. 270 P

18 204 DORIS L. MACKINNON AND D. INES ADAM ment of the spore the sporoplasm lies within the sporal envelope. In order to make these observations of ours on the lifehistory of Triactinom.yxon more easily comparable with those by other workers on the Actinomyxidia;, we append a diagram in which the successive stages are arranged in order. There seems to be agreement among protozoologists who have studied the group that, morphologically considered, Hexactinomyxon, Synactinomyxon, Sphaeractinomyxon, and Triactinomyxon are more like one another than they are like Tetractinomyxon; indeed, in the latest classification Tetractinomyxon stands over against the other four genera as Actinomyxidia siniplicia, while they are grouped together as Actinomyxidia multiplicia. But comparison of Text-fig. 8 with Ikeda's diagram of the life-history of Tetractinomyxon shows the essential similarity between Triactinomyxon and Tetractinomyxon when they are considered from a point of view other than that of the size and shape of the spore. We think that the deviations from our scheme of the life-history which appear in the work of Caullery and Jlesnil on Sphaeractinomyxon and of Leger on Triactinomyxon (the only' genera of the Actinomyxidia multiplicia hitherto studied in any detail) are most likely to be due to the omission by these authors of some stages, which omission has led to certain misinterpretations. Our diagram is probably applicable to the Actinomyxidia generally. We are convinced from what Ave have seen that the new pansporocyst and its contents always arise from the association (more intimate in some species than in others, but never amounting to nuclear fusion) of two cells. In some Actinomyxidia these come from different spores (Tetractinomyxon), but in others (Triactinomyxon legeri and probably T. ignotum) they are derived from the same spore. Each of the associating cells, whether they combine in the uninucleate condition (Triactinomyxon) or in the

19 Diagrammatic representation of the life-history of Triactinomyxon. (Black nuclei are those destined for the cells forming the pansporooyst; the reduction bodies and chromatin balls are also in black. Nuclei of the a gametes are shown thus 0, and of the j8 gametes thus Q ; nuclei of the sparal envelope, capsulogenous cells, and the residual nuclei thus O-) 1, ' sporozoite'; 2 to 6, formation of pansporocyst (3, 4, and 5 are stages in T. ignotum; 3', 4', and 5' are stages in T. 1 e g e T i and are given here as an alternative); 7 to 11, formation of the gametes and appearance of their reduction bodies ; 12, the copulae, their nuclei not yet fused ; 13, a zygote, and above it a representation of the subsequent nuclear divisions (r.n., residual nuclei; g.n., nuclei of the sporoplasmic mass ; p.c.n., nuclei of the polar capsule cells; sp.n., nuclei of the sporal envelope). These nuclear divisions are related to figs. 14 and 15, which show a young spore and a ripe spore respectively (T. i gnotum). P 2

20 206 DORIS L. MACKINNON AND D. INES ADAM binucleate (Tetractinomyxon), contributes equally to the pansporocyst and to its contents, i. e. the gametes. Since half the gametes arise from the ' germinal' portion of one associate and half from the other, Hartmann's statement that the sexual process in the Actinomyxidia is a true autogamy seems to be without foundation. We offer our thanks to the Department of Scientific and Industrial Eesearch for a grant received by one of us (D.I.A.) during the early part of our work. We also wish to thank Professor J. Stephenson, to whom we have sent a number of oligochaetes for identification, and Miss Beatrice Sorby for taking a share in the routine examinations and section-cutting. We are especially grateful to Mr. Clifford Dobell, who gave us much useful criticism and advice. REFERENCES. Oaullery, M., et Mesnil, F. (1904 a). " Sur un type nouveau (Sphaeractinomyxon stolci, n.g., n.sp.) d'actinornyxidies et son developpement", ' C. B. Soc. Biol.', Paris, torn, lvi, pp (19046). " Sur les affinites des Actinomyxidies ", ibid., pp (1905). " Recherches sur les Actinomyxidies. I. Sphaeractinomyxon stolci", 'Arch. f. Protistenk.', Bd. 6, pp , 1 pi. Granata, L. (1922 a). " Ricerche sugli Attinomissidi. I. Triactinomyxon magnum, n. sp.", ' Monitore Zool. Ital.', xxxiii, pp (1922 b). ' Ricerche sugli Attinomissidi. II. Neoactinomyxum globosum, n. gen., n. sp.", ibid., pp , 2 Text-figs. Ikeda, I. (1912). " Studies on some sporozoan parasites of Sipunculoids. I. The life-history of a new Aetinomyxidian, Tetractinomyxon intermedium, g. et sp. nov.", 'Arch. f. Protistenk.', Bd. 26, pp , 1 pi. and 1 text-fig. Leger, L. (1904a). " Sur la sporulation du Triactinomyxon ", ' C. R. Soc. Biol.', Paris, torn, lvi, pp , 4 figs. (19046). "Considerations sur le genre Triactinomyxon et les Actinomyxidies ", ibid., pp Mrazek, A. (1899). Referat on Stoic's work, " Aetinomyxidie, eine neue Gruppe der Mesozoa, den Myxosporidien verwandt ", ' Zool. Centralbl.', Bd. 7, pp Stole, A. (1899). " Actinomyxidia, nova skupina Mesozon pribuzna Myxosporidium", in ' Rosp. leske Ak. Ces. Frant. Jos. II', vol. 8, 12 pp., 3 pis.

21 SPOROZOA IN TUBIFBX 207 EXPLANATION O PLATES 6 TO 8. All drawings except fig. 1 were made with the help of the camera lucida from material fixed with Bourn's or Schaudinn's fluid and stained with Delafield's haematoxylin and eosin. The lenses used were objective ^T oil imm. of Leitz with ocular 4 (Leitz) and compensating oculars 8 and 12 (Zeiss). The magnification, unless otherwise stated, is 2,000 diameters. LETTERING. a, gametooyte producing small gametes ; a 1, daughter cell from first division of a ; a' 2, daughter eel lfrom division of a 1 ; a 3, daughter cell from division of a" (= small gamete); /5, gametocyte producing large gametes ; f) 1, daughter cell from first division of /?; /? 2, daughter cell from division of /J 1 ; js 3, daughter cell from division of /3 2 ( = large gamete); c.c, capsulogenous cell; c.6., chromatin ball; f.v., food vacuole; p.c, polar capsule; p.n., nucleus of pansporocyst; ps,, pseudopodium; r.b., reduction body; rm., residual nucleus; s., sporozoite; $p., spore; sp.n., nuclei of sporal envelope. PLATE 6. Pigs. 1 to 24. Triactinomyxon legeri, n. sp. Fig. 1. Living pansporocyst containing eight spores. X750. Fig. 2. Upper end of ripe spore with twenty-four uninucleate sporozoites. Fig. 3. Invaginated tips of branches of the sporal envelope. Fig. 4. Sporoplasm with, six binucleate sporozoites descending the tube of the spore. At <i> two sporozoites are fusing. Fig. 5. Sporoplasm with six binucleate sporozoites Hearing the point of exit from the spore. Fig. 6. Liberated sporoplasm containing four binucleate sporozoites, two residual nuclei, and a chromatin ball. Fig. 7. Disintegrating sporoplasm depositing nine binucleate bodies. Fig. 8. Binucleate amoebula with like nuclei. Fig. 9. a, 6, c. Heteronucleate amoebulae ; three stages. Fig. 10. Amoebula with three nuclei. Fig. 11. a, b. Amoebulae with four nuclei. Fig. 12. Pansporocyst with binucleate contents. Alongside is a small binucleate amoebula, presumably from the same batch of sporozoites. Fig. 13. Later stage of the same. The nuclei of the central mass have enlarged. Fig. 14. Pansporocyst containing two nninucleate cells.

22 208 DORIS L. MACKINNON AND D. INES ADAM Kg. 15. Pansporoeyst containing two binuoleate cells, a and /3. Fig. 16. Pansporocyst containing four binucleate cells, two a 2 cells, one a 1, and one j8 cell. Fig. 17. Pansporocyst containing five cells one binucleate fl cell, two binucleate a 2, and two uninucleate a 3. Fig. 18. Pansporoc3'st with six cells. The first 0 cell has divided into two, and there are four binucleate a 2 cells. PLATE 7. Fig. 19. Ten-celled stage. There are eight uninucleate a 3 cells and two uninucleate p l cells. (A similar stage (T. ignotum), with the /3 1 cells uppermost, is shown in fig. 38.) Fig. 20. Pansporocyst containing the ripe gametes, a a and 3. At ^ana gamete has fused with a js, and their nuclei are in apposition. Fig. 21. Portion of a pansporocyst with three developing spores indicated, two of them in detail. They each contain six nuclei and a chromatin ball. Fig. 22. Later stage of the same : each young spore has now seven nuclei. In the upper spore the chromatin ball is seen to be a double structure. Fig. 23. Young spore with seven nuclei. Three of these have gone to one pole, and the sporal envelope is beginning to form. Fig. 2-i. Later stage, with the nuclei of the three capsulogenous cells at one pole and of the three sporogenous cells at the other. Between these is the sporoplasm with thirteen nuclei and a chromatin ball. Figs. 25 and 2G. Triactinomyxon mrazeki, n. sp. Fig. 25. Upper end of a ripe spore with sixty-four sporozoites and three residual nuclei. Fig. 26. a, b. Binucleate amoebulae. In b some engulfed food particles are shown. Figs. 27 to 45. Triactinomyxon ignotum Stole. Fig. 27. Upper end of spore showing eight sporozoites in the sporoplasm ; there are two residual nuclei and a chromatin ball. Fig. 28. Sporoplasm slipping down the tube of the spore ; the sporozoites are still uninucleate. Fig. 29. Group of eight liberated sporozoites. Fig. 30. Liberated sporoplasm with five sporozoites, two residual nuclei, and a chromatin ball. At <i> is a binucleate body formed of two fused sporozoites. Fig. 31. a, b, c, d. Binucleate bodies, probably developed from the binucleate stage in fig. 30.

23 SPOEOZOA IN TUBIFEX 209 Fig. 32. Uninucleate body, probably developed direct from a uninucleate sporozoite. Fig. 33. a, 6. Two-celled stage, reached either by division of binucleates into two, or by association, of two uninucleates of the type in fig. 32. Fig. 34. a, b. Four-celled stage. Each of the cells in fig. 33 has cut off a small cell at one pole. Fig. 35. The two small cells in fig. 34 have become the cystogenous cells of a pansporocyst containing the two larger cells (gametocytes). Fig. 36. Pansporoc3 r st containing three cells, two a 1 and one B. PLATE 8. Fig. 37. Pansporocyst with four contained cells, two a 1 and two 8 l. Fig. 38: Pansporocyst with ten cells, eight a 3 and two j8 l. Fig. 39. Pansporocyst at the ten-cell stage, but with the nuclei of the two j3 l cells dividing. (No clear details are given in these cells only.) Fig. 40. Pansporoc3 r st containing fourteen cells. Of these eight are a gametes with reduction bodies already formed ; the other six are p cells, two of which (j3 2 ) are undergoing the last division. Some of the /3 3 cells have extruded their reduction bodies. (The cells that are not dividing are shown in outline only.) Fig. 41. Pansporocyst with eight cells, four of which are binucleate. Since the reduction bodies are formed, these cells represent either the fusing gametes or the product of the first nuclear division of zygotes. At «fr is probably an undivided zygote. Fig. 42. a, b, c, d. Four stages in multiplication of nucleiinthe developing spore : (a) with two nuclei, (6) with four, (c) with four, one of which is beginning to divide, (d) with seven nuclei, three of which are in division. All but (c) show a chromatin ball caught up in the spore. Fig. 43. Sporoplasm with sporozoites still undiflerentiated. Fig. 44. Young spore with capsulogenous and sporogenous nuclei differentiated. The polar capsules are forming. Fig. 45. Young spoce with polar capsules and envelope forming. The nucleus of the sporoplasm is still undivided.

24 D.LM. del.

25 Quart. Journ. Micr. Sci Vol. 68, N. S., PL 6

26 \ c.b D.L.M. del.

27 Quart. Journ. Micr. Scl Vol. 68, N. 8., PI, 7

28 Quart. Journ. Micr. Sci. Vol. 68, N. S., PI 8 p.c-- ~.S P.n. s P. n.-- D.L.M. del.