Studies on some Amoebae from the termite Mirotermes, with notes on some other Protozoa from the Termitidae

During a visit to Barro Colorado Island in the Canal Zone in the summer of 1925”, observations in many termites of the family Termitidae, most members of which according to previous examinations contain no Protozoa, resulted in the discovery of small flagellates and amoebae in several species. The most numerous and striking amoebae occurred in Mirotermes hispaniolae Banks and Mirotermes panamaensis Snyder, termites which have wood-boring and wood-feeding habits similar to those of the lower families.¹ Previously Cleveland (1923) had found Protozoa in alcoholic specimens of these two species of termites which had been collected from the same locality and preserved in the collections of the U.S. National Museum, but since the Protozoa were in poor condition he was unable to classify them. However, it is probable that they were the same as the amoebae described below.

Of these amoebae three species occurred in the intestine of every individual from one colony of Mirotermes hispaniolae which was examined, and two in the same way in four colonies of Mirotermes panamaensis. One in each termite is large in size, the others smaller, and they differ sufficiently in the two hosts to be readily distinguished, at least when stained—much more readily than their hosts. Indeed, they may serve as a means of separating the species of insects, which morphologically resemble each other closely.

Minute amoebae in such small numbers that it was necessary to search for them carefully were found in Amitermes beaumonti Banks, Amitermes médius Banks, Subu-litermes kirbyi Snyder, and Cornitermes acigna-thus Silvestri. The other species of Termitidae examined in the Canal Zone in which no amoebae were discovered were Cylindrotermes nordenskidldi Holmgren, Armitermes chagresi Snyder, Nasutitermes cornígera Motschulsky, Nasutitermes pilifrons Holmgren, Nasutitermes ephratae Holmgren, Obtusitermes pana - mae Snyder, a species of Microcerotermes, Subulitermes zeteki Snyder, Orthognathotermes wheeleri Snyder, and a species of Anoplotermes. In the last two of these, and in all those, including Mirotermes, in which amoebae occurred, minute flagellates were present in small numbers.

Several observers have examined termites of the family Termitidae for Protozoa with negative results, but I am aware of only one other positive record. From the examination of alcoholic material Cleveland (1923) recorded Protozoa in a species of Nasutitermes from the Belgian Congo, as well as in these two species of Mirotermes.

Observations of living material were made in the field by removing the intestinal contents and mixing them on a slide with a drop of diluted Locke’s solution. Besides this study, a quantity of smears on both cover-glasses and slides was prepared, fixed in Schaudinn’s, Bouin’s, or Flemming’s fluids, and preserved in 85 per cent, alcohol. This material was stained after my return in Heidenhain’s or Delafield’s haematoxylin with or without acid fuchsin, eosin, or Biebrich’s scarlet counterstain; in alum carmine; or in methyl-green or methyl-blue with eosin. Also, some intestinal canals were pulled out and fixed for later sectioning.

(Plates 22, 23 (figs. 17-26), and 24.)

This amoeba, the most abundant species present in the intestine of its host, is intermediate in size between its two associates, from which it differs not only in this respect but also in its degree of activity, in the structure of the nucleus, and in the kind of cytoplasmic inclusions. The body may be spherical or more commonly ellipsoidal in form. While most of the fixed individuals possess blunt, short protuberances over much of the surface, there are sometimes a few large, clear, lobe-like pseudopodia. Although clear enough for facile observation of the nucleus, the endoplasm is usually crowded with small particles of wood and yellow, rounded objects of unknown nature.

The nucleus, which is generally situated toward one side of the animal, is spherical in what appears to be the resting stage (figs. 1–3, Pl. 22), but sometimes there is, as shown in some of these figures, a depression at one side, which is doubtless abnormal. While it ranges in diameter from 7·5 to 9 microns, it generally measures about 8.

The structure of the resting nucleus differs from that of nuclei in most other known entozoic amoebae. Situated near the nuclear membrane, against which they may be more or less flattened, there may be several nucleoli, usually from two to four but sometimes as many as seven, which are variable in size and shape. The largest of these is sometimes 2 microns in its longer diameter. These nucleoli stain with iron haematoxylin more heavily than does the chromatin. This latter substance is aggregated in an irregular granular mass which occupies most of the inner region of the nucleus and is surrounded by a clear zone, often of width, one-third to one-half of the radius, which is apparently crossed by many radial filaments. This mass does not stain as heavily as do the chromosomes which arise later. In it the chromatin granules are small and often arranged more or less in lines. In the nucleus shown in fig. 4, Pl. 22, the chromatin granules are, as in the segmented spireme, arranged in threads, some of which extend out to the membrane.

It is remarkable that of two hundred and thirty-four individuals on seven slides, the nuclei of just half were in the form of a spindle (figs. 6, 7, 10, 11, 13, Pl. 22). This resembles a figure of mitosis, but its abundance makes it evident that it is a persistent stage. Although some spindle-shaped nuclei occur on all slides, the proportion is not always 50–50; on seven slides it varied from 72–28 to 27–66. The remaining small percentage of nuclei in the latter case were destroyed by parasites (see p. 196).

The average size of the amoebae with fusiform nuclei differs from that of those with spherical nuclei, the former being larger, but otherwise the structure of the two is similar. Measurements of two hundred and fifty individuals gave the following sizes, in microns:

The fusiform nucleus is frequently about 14 microns long. One end is sometimes more pointed than the other, and when stained with Heidenhain’s haematoxylin there is always a deeply staining substance at this pole (figs. 6, 7, Pl. 22), which usually has pushed out the membrane to a mammiform protuberance. Although with Delafield’s haematoxylin this structure does not stain, it can still be distinguished (figs. 11, 13, Pl. 22). The peripheral nucleoli, which appear distinctly both when Heidenhain’s and Delafield’s are used, vary in number, size, and position, but are always more or less flattened against the membrane. The largest of these, which sometimes appears vacuolated, is beyond the equator toward the pole opposite to that characterized by the deeply stainable substance; others are in divers peripheral positions; and sometimes there is a small nucleus near or at the above-mentioned pole.

The chromatin is collected in a fusiform aggregate at the axis of the spindle, leaving a broad, clear zone between it and the membrane. This mass may be resolved into granules of various sizes (fig. 7, Pl. 22), but, as an end view shows, it is quite compact (fig. 8, Pl. 22). It is often uneven in outline and from it fibres run irregularly toward the nuclear membrane. Especially evident are two or three strands, along which may be some granules, which extend from the chromatin mass to the mammiform pole (figs. 13, 15, Pl. 22), but similar strands do not exist at the other pole. It is evident from the above descriptions that this spindle is not symmetrical, for the two ends differ morphologically in shape, in the mammiform protuberance at one, and in the strands at this same pole, as well as in the relative position of the peripheral nucleoli.

At first glance it seems possible that this unusual nucleus is undergoing a transitory stage of mitosis, but further consideration renders this doubtful in view of the absence on most of the slides of any stages other than this and the spherical form, and in view of the equality in numbers of these two types. Its form could be obtained from the more typical spherical nucleus by elongation, greater concentration of chromatin, and a suitable distribution of nucleoli. Thus it may follow this earlier phase of development, but why it should be so numerous I cannot suggest.

A few stages which appear to be transitional between these two forms were found. In them the nucleus (figs. 5, 12, Pl. 22) is more broadly ellipsoidal (in some more so than in those figured); there may be several nucleoli against the membrane; and the chromatin is concentrated into a less extensive, more compact mass, leaving a broad peripheral zone crossed by some strands. However, such stages are rare.

Although karyokinetic figures were scarce in this material, I am able to give an account of this process from a study of about twenty figures from three slides, all of them stained in Delafield’s haematoxylin without counterstain, which were very clear and easily interpreted..

In earlier stages the nucleus is spindle-shaped, the poles rounded, and the chromatin arranged in tangled threads on the central core. ‘While in shape this nucleus is similar to the abundant spindles previously described, it differs in that it is symmetrical toward the two poles and possesses no nucleoli or other peripheral or polar structures. Possibly in a still earlier stage, before loss of the peripheral shape, the chromatin threads form, as in fig. 4, Pl. 22. In the mitotic spindles no centrioles can be seen, and no more darkly staining plasm at the poles, and the nucleoli have disappeared. The individual chromosomes, which are distinguishable at about this time, becomes more loosely arranged, occupying more of the intranuclear area (figs. 18-20, Pl. 28), and probably divide longitudinally, starting first at one end and before the final separation forming V’s (figs. 18-20, Pl. 23). It will be noted that in this series there is no true equatorial plate stage; intranuclear spindle-fibres are absent; and the chromosomes have varicose outlines, as if composed of rows of granules. These characteristics, as well as the structure of the resting nucleus, distinguish Endamoeba disparata from intestinal amoebae of the genus Entamoeba, which have centrioles inside of the nucleus and small chromosomes which appear to be single granules.

In the anaphase (figs. 21, 22, Pl. 23), when the chromosomes have separated and converge to the poles, twelve can often be counted in each group. These are all varicose threads, and differ from one another in length. As the nucleus elongates and constricts equatorially the poles become rounded (fig. 23, Pl. 23) until the dumb-bell form is attained, and then the middle constriction narrows to a thread which becomes longer than the original spindle before it ruptures (figs. 24–6, Pl. 23). No reorganization stages were observed.

During karyokinesis particles of food material crowd the endoplasm as during the vegetative stages.

The parasite Nucleophaga has been found destroying the nuclei in several species of both free living and parasitic amoebae. With its affinity for stains it contributes to some unusual nuclear appearances, which have sometimes been wrongly interpreted as phases of nuclear development. On the slides of Endamoeba disparata I observed numerous instances of this parasitism, some of which are shown on Pl. 24. Also on this plate is placed a figure of a normally nucleated amoeba (fig. 35, Pl. 24) for comparative purposes.

According to Dangeard (1895), the discoverer of Nucleophaga, its life-history is as follows. The parasite first appears within the nucleolus as a vacuole with a central granule, its nucleus—the clear cytoplasm of the invader appearing like a vacuole. The size increases along with multiplication of nuclei until the parasite occupies the entire interior of its host’s nucleus and possesses more than a hundred of its own. At this time, when the parasite resembles a mulberry, sporulation to form zoospores begins. The sporangia reach maturity after the death of the amoeba. Doflein (1907, quoted by Chatton and Brodsky, 1909) observed copulation of motile, flagellated zoospores.

The youngest stages of the parasites of Endamoeba blattae observed by Mercier (1910) were small, irregular, granular masses with numerous nuclei. From an account of these his description passes to the formation of spores, small spherules which stain intensely with haematoxylin. These spores possess an investing membrane, and, within this, cytoplasm enclosing one or two chromatin grains. Mercier found that after sporulation the membrane of the host’s nucleus ruptures, permitting distribution of the spores in the cytoplasm of the amoeba.

A granular body in the nucleus of a small individual of Endamoeba disparata (fig. 37, Pl. 24) resembles the earliest stages of the parasite of Endamoeba blattae observed by Mercier. This, which exists along with the chromatin and nucleoli, can only be interpreted as an invading organism, for it resembles none of the usual nuclear structures. Most of the other Nucleophaga observed filled up the interior of the host’s nucleus and possessed a large number of their own nuclei—several hundreds in the sporangia, for example, of figs. 41-3, Pl. 24. The spherules of these later stages stain intensely with iron haematoxylin and retain the stain so tenaciously that they remain coloured when all other structures have been decolourized. In some figures there are a few spherules of larger size (figs. 38, 39, Pl. 24), while in other, later stages these are more numerous and smaller (figs. 41–3, Pl. 24). The disintegrating chromatin is enclosed in the central zone and surrounded by several layers of spherules which come more and more to fill up the nucleus from the periphery inwards (fig. 40, Pl. 24). The diameter of the mature globe of spherules, which exceeds that of the original host’s nucleus, may reach 12 microns.

When Delafield’s haematoxylin or alum carmine is used the chromatin is coloured but the parasites remain mostly unstained, so that they may readily be distinguished from the components of the amoeba’s nucleus and optical sections can be observed (figs. 42–3, Pl. 24) to show the arrangements described in the previous paragraph. By the use of these stains, as well as by a more complete differentiation of Heidenhain’s haematoxylin, the spherules can be resolved into an investing envelope and a central chromatic granule (fig. 44, Pl. 24).

It is possible that figs. 36 and 46, Pl. 24, represent a different type of development of this Nucleophaga, if not a different species, as the structures drawn do not fit into the history of the form described above. The spherules of fig. 36, Pl. 24, are much too large to be nuclei in the latter series, and the distributed spherules of fig. 46, Pl. 24, are also larger and less numerous than studies of the sporangia of this would seem to indicate the spores to be. That these spherules are parasites seems probable in the latter case because there is no nucleus present in the amoeba.

On most of these slides the parasites were present in an average of 8 per cent, of the nuclei, but this percentage varied from 6 to 12 on different slides and some seemed to be without any parasites.

Between 20 and 40 microns in longest diameter, averaging 32. Broád, blunt pseudopodia. Active in normal condition. Cytoplasm crowded with particles of wood and other material. Nucleus with distinct membrane, several peripheral nucleoli and an irregular granular mass of chromatin occupying a large part of the inner region; about 8 microns in diameter. In mitosis twelve varicose, filiform chromosomes formed from a spireme. Apparently no centrosomes or spindle-fibres and no definite equatorial plate. Xylophagous. Habitat: the intestine of Mirotermes hispaniolae Banks from the Canal Zone, Panama.

(Plate 25, figs. 47.–53.)

The largest amoeba of the three species in the intestine of Mirotermes hispaniolae is present in considerable numbers, but is much less abundant than is Endamoeba disparata. The body, which may be spherical or broadly ellipsoidal with irregular outlines, ranged in fifty individuals on eight slides from 60 × 65 to 154 × 165 microns in size, averaging 94×115 microns. Many of the smallest animals occurred on one slide. There may be short, lobose pseudopodia protruded from various parts of the body (fig. 47, Pl. 25); a delicate but distinct periplast; and a thin, clear outer layer of cytoplasm which is conspicuous only in the pseudopodia. The endoplasm is densely granular and crowded with particles of wood in gastric vacuoles; some of the particles are very small, others relatively large. This food material, which is relatively more abundant than in End amoeba disparata, stains deeply at times and obscures the nucleus, especially when iron haematoxylin is used.

The nucleus of this amoeba (figs. 48–51, Pl. 25) possesses a distinct but not especially heavy membrane. While its shape is typically that of a sphere, it is often deformed in preserved material. In diameter it ranges from 15–26, usually nearer 20 microns. The interior is occupied by an irregular reticulum which is more closely meshed toward the periphery and supports throughout some fine granules, which may be especially abundant at the outer portions. Near or at the centre is a dense aggregation of larger granules, sometimes very irregular in outline, which stains with chromatin dyes more or less deeply. The group is often quite dense and deeply staining, and presents then the appearance of a compact, granular endosome. In what seem to be early stages of karyokinetic organization these granules stain more deeply and take looser, linear arrangements, evidently forming thread-like chromosomes which become dispersed throughout the intranuclear area in the prophase, when the nucleoli, to be described below, disappear.

Irregular masses of nucleolar material are present in the region between the central chromatin aggregation and the membrane. Sometimes these are uneven, vacuolated masses which stain more intensely than does the chromatin when Delafield’s haematoxylin is used—a circumstance also true of the nucleoli of Endamoeba disparata; sometimes they are compact and rounded, or irregularly elongate; and often they appear in winding linear shapes. As is apparent from the above description, the arrangement of nucleolar material is not the same in any two nuclei. The outer covering stains with chromatin dyes (Heidenhain’s haematoxylin, Delafield’s haematoxylin, and alum carmine) while the interior does not uniformly do so, with the result that the structures often appear vacuola ted.

After suitable preparation, a relatively large granule, which is isolated by a surrounding clear area, may be seen at the centre of the chromatin mass (figs. 51–3, Pl. 25). Inasmuch as this stains with iron haematoxylin more intensely than any other structure, it may be quite striking when the nucleus is very much destained. In appearance the granule is suggestive of the small karyosomes of Entamoeba and Dientamoeba, and possibly it is homologous to these.

Two amoebae were observed in which the nucleus contained a Nucleophaga similar to that of Endamoeba disparata.

In this species, as well as in thé other amoebae of this termite, no cysts have been found in spite of much searching in many smears.

Diagnosis and Summary.

Between 65 and 165 microns in longest diameter, averaging 115 microns. Short, lobose pseudopodia often protruded from various parts of the body. Endoplasm crowded with food material, mostly fragments of wood. Nucleus between 15 and 26, usually nearer 20, microns in diameter. Interior of nucleus occupied by an irregular reticulum on which are many fine granules; a dense aggregation of chromatin granules near centre; a deeply staining granule in a clear area within this aggregation; and irregular, vacuolated masses of nucleolar material, often anguilliform, in outer zone. In division, the nucleoli disappear and the chromatin granules aggregate into threads. Xylophagous. Habitat: the intestine of Mirotermes hispaniolae Banks from the Canal Zone, Panama.

(Plate 23, figs. 27–34.)

The smallest species of amoeba present in the intestine of its host, Mirotermes hispaniolae, seems to belong to the genus Endolimax. In shape the individuals preserved on the slides are, if definite at all, spherical, ovoidal, or ellipsoidal with irregular outlines, and there are usually no marked pseudopodial protuberances. The average size is 12–5×16 microns, and the range between 9·5 × 9-5 or even smaller to 19 ×27 or rarely a little larger, as in the parasitized individual of figure 34, Pl. 23. The cytoplasm is invested by a distinct pellicle which can be readily seen in instances when it is separated from the granular plasma, as frequently occurs, though in many cases at least this appears to be a result of the preservative treatments.

The cytoplasm is usually dark, granular, and more or less vacuolated. Within many of the vacuoles are small bodies which are doubtless ingested food materials, most probably small fragments of wood, and also in the endoplasm, surrounded by vacuoles, may be elongated fragments of wood or yellow, tubular structures similar to some of those observed in Endamoeba disparata. But the food material is never in abundance, so that the granular and rather dark endoplasm is more or less uniform in appearance, with no obtrusive crowding of large food bodies as in the relatively clear endoplasm of Endamoeba disparata.

The nucleus differs greatly in appearance from that of the latter amoeba. In form it is spherical; in diameter it varies in different individuals, though apparently not in ratio to the size of the body, from 4 to 5 microns. Occupying a large portion of the inner zone is an irregularly outlined karyosome, composed apparently of plastin in which are imbedded some granules. Besides these latter there sometimes appears to be a minute central granule (fig. 32, Pl. 23), set apart by a surrounding clear space. Peripheral chromatic granules cover the inside of the nuclear membrane, so that, as observed in optical section, a deeply staining granular ring forms the boundary. Extending across the clear zone between this peripheral chromatin and the karyosome there are numerous radial filaments.

In a prophase of mitosis the peripheral granules have disappeared from the nuclear membrane, the karyosome as such is no longer apparent, and the chromatic granules are arranged in a group of threads looped toward one side of the nucleus (fig. 33, Pl. 23). An important feature of this figure is the possible existence of an intradesmose with a granule at either end. But the type of mitosis remains unknown, for the only other division stage observed was a final stage in plasmotomy in which the two nuclei were completely reorganized and the daughter individuals nearly separated.

Nucleophaga also parasitizes this amoeba, though apparently not very frequently in the specimens studied. Fig. 31, Pl. 23, seems to show a case in early parasitism in which there are only a few spherules within the nucleus, while the unusually large Endolimax termitis of fig. 34, Pl. 23, which has a cytoplasmic structure indicative of its belonging to this species, contains a large globe of spherules similar to the later stages in the development of the parasite of Endamoeba disparata.

In nuclear structure Endolimax termitis is similar to amoebae of the genus Endolimax, while it differs greatly from its associates in the termite.

Diagnosis and Summary.

Shape, when regular at all, spherical, ovoidal, or ellipsoidal; usually no definite pseudopodia; between 9–5 and 27 microns in longest diameter, averaging 16. Cytoplasm dark, granular; ingested particles of wood few. Nucleus 4–5 microns in diameter, with large central karyosome and many chromatic granules under the nuclear membrane. Xylophagous. Habitat: the intestine of Mirotermes hispaniolae Banks from the Canal Zone, Panama.

(Plate 26, figs. 58-67.)

The larger of the two amoebae living in the intestine of Mirotermes panamaensis is also dominant in numbers. The body form is typically that of a sphere or ellipsoid, the amoeba is not very active, and when living it possesses many short, blunt pseudopodial protuberances over much of the surface (fig. 58, Pl. 26). A thin layer of clear ectoplasm can be distinguished in living and in suitably stained material, and this is extensive in regions of pseudopodial formation, where it constitutes most of the broad, lobose extensions of the body sometimes present. Due to the large amount of ingested wood, which is, however, less abundant relatively than that of Endamoeba majestas, the amoeba is dark and the nucleus seldom can be easily seen. Fifty individuals on eight slides from three colonies ranged in size from 46 × 50 to 100 × 120 microns, averaging 65 × 77 microns, thus being smaller than End - amoeba majestas.

Though usually the nucleus is more or less obscured by the ingested fragments of wood, in some instances it appears clearly near the upper surface, and also some isolated nuclei may be found in the smears which have been released by rupture of the body, but nevertheless do not appear to be in any way abnormal. In diameter the nucleus ranges from 15 to 21, averaging 17-6 microns. It is invested by a definite but not unusually thick membrane, and the interior is differentiated into two areas (figs. 58–60, Pl. 26).

The outer of these two zones varies in width in what are apparently typical resting stages from a half to a third or even less of the radius, and appears quite clear except for alveolar walls and minute granules. It seems to consist of two morphologically distinguishable regions, an outer of many small alveoli and an inner of a single layer of large alveoli. The minute granules between the outer alveoli stain palely, so that the peripheral region is only slightly darkened. The walls of the larger inner alveoli appear like a series of radial fibres extending out from the sharply circumscribed inner mass.

While this outer zone is thus quite clear, the inner zone is occupied by an aggregation of chromatic material in which there are at least four morphologically distinguishable structures. At the periphery, in the first place, is a lightly staining, homogeneous or finely granular substance, perhaps plastin, which is probably the matrix in which the other materials are embedded and thus extends throughout the mass. In the second place, there are numerous lightly staining, moderate-sized granules distributed unevenly throughout the interior. At the inner boundary of the outer plastin zone there are sometimes numerous granules of the same size which, however, take the stain intensely. The third structure is a large, deeply staining body near the central region, which body varies greatly in form, may appear as a group of irregular masses, and often appears vacuolated or unevenly granular. Finally there is a variable number of small spheres, apparently nucleoli, situated peripherally just within the plastin boundary. When the stain is intense these may be uniformly coloured, but often only the outer shell stains so that the nucleoli appear in optical section as rings, which are, however, not perfectly outlined. While sometimes these take the stain as intensely as does the chromatin, in other cases they are pale and inconspicuous while the chromatin is deep-coloured.

Nuclei of different appearance are frequently found (figs. 62–5, Pl. 26); these are apparently undergoing karyokinetic changes. Early in this process the central aggregation expands to the membrane, so that then the outer zone is absent. The nucleoli, now situated toward the periphery of the nucleus, become pale and finally disappear. Also the granules disappear, perhaps entering into construction of the chromosomes. The chromatic body becomes more regular in outline and more evenly granular, so as to appear as a karyosome. Varicose or granular chromatic threads form, at first centrally, then more dispersed in the intranuclear region. These may number about a dozen, and there is some evidence that they undergo the early longitudinal split characteristic of many chromosomes. No later stages of mitotic division in the trophozoite were observed, so the fate of these chromosome-like structures could not be traced.

Although, in spite of extensive observations on a large amount of material, no cysts were found in any other species described in this paper, several good examples were seen in Endamoeba simulans. However, even here, judging from specimens on smears, encystment is not common. Of those observed one possessed a single dividing nucleus, one two nuclei and another four, and the cytoplasm of all these was free of food particles.

An early stage of cyst development was exhibited by a rounded amoeba, measuring 65 ×72 microns, which seemed to be investing itself with a gelatinous covering. From the body all food particles had been extruded, but the cytoplasm contained numerous chromatin granules as well as the single nucleus. The form of this nucleus was that of a spindle with broadly rounded ends; the internal structure was of an alveolar nature with a few chromatic granules unevenly distributed in it. Evidently it was in a phase of karyokinesis, but unfortunately the finer details of structure could not be accurately seen.

Each of the other two cysts, which are figured, possesses a thick hyaline wall, which in that with two nuclei is separated from the enclosed amoeba (fig. 66, Pl. 26). These have a uniform thickness of about 2·5 microns, and were covered by adherent debris. The transparency of the quadrinucleate cyst (fig. 67, Pl. 26) permitted a very clear view of the organism. Except for the complete absence of food particles, the cytoplasmic structure is similar to that of the trophozoite. Adjacent to the inner surface of the cyst wall is the investing periplast. The nuclei measure about 10 microns in longest diameter and contain numerous chromatic granules and aggregations of granules in a finely alveolar matrix. The quadrinucleate cyst measures 63 × 56 microns, the binucleate 60 × 72 microns. It is obvious from the appearance of these nuclei that the structure and type of mitosis in the nuclei of the encysted amoebae is, as in Endamoeba blattae, different from that of the trophozoites.

In the cytoplasm of this last encysted amoeba a number of granular aggregations are seen, and similar structures were seen in some of the trophozoites. Apparently these are cytoplasmic parasites,perhaps Sphaerita, which is not uncommon in the flagellates of termites as well as in many other Protozoa.

Also, as in at least two of the other species, the nucleus may be destroyed by Nucleophaga (fig. 61, Pl. 26).

Between 50 and 150 microns in longest diameter, averaging 77. Many short, blunt pseudopodia may be protruded from various parts of the body. Endoplasm contains ingested wood, but less than does Endamoeba majestas. Nucleus 15–21, averaging 17–6 microns in diameter; two intranuclear zones, outer clear, inner dark. Chromatin in an irregular, karyosomelike mass in inner zone; small, spherical nucleoli at the periphery of this zone. In mitosis inner zone expands, nucleoli disappear, chromatin granules form threads which become dispersed within the nucleus. Xylophagous. Habitat: the intestine of Mirotermes panamaensis Snyder from the Canal Zone, Panama.

(Plate 25, figs. 54–7.)

A smaller amoeba is also present in the intestine of Mirotermes panamaensis. From a few termites of this species which were dissected in Panama I recorded many small amoebae which moved actively, were quite clear, and were comparable in numbers and general appearance to Endamoeba disparata of Mirotermes hispaniolae. But while in some ways this amoeba described below resembles this latter species, it possesses several specific differences.

The body of the amoeba is broadly ellipsoidal, rarely spherical in shape, and generally has no pseudopodial protuberances (fig. 55, Pl. 25). In some individuals, however, a few short, broad pseudopodia have been preserved (fig. 56, Pl. 25). Fifty individuals on nine slides averaged 22 ×26 microns, ranging from 15×19 to 32×35, but the extremes are exceptional. Most individuals measure close to the average size.

Particles of ingested wood are present in the cytoplasm, but much less abundantly than in Endamoeba disparata. These are generally small, but are sometimes of relatively large size. The most characteristic feature is the large number of chromatic cytoplasmic granules (figs. 54·7, Pl. 25), which in size and form resemble the smaller mitochondria of Endamoeba gingivalis (Causey, 1925). These appear in all individuals in proportionately similar numbers, whether the material was fixed in Schaudinn’s or Bonin’s fluids; they stain intensely with iron haematoxylin but not with Delafield’s haematoxylin; and they are often especially numerous against and near the nucleus. These inclusions are evidently metabolic products, but differ from mitochondria in their behaviour toward reagents, for they are easily demonstrated in material fixed by ordinary methods. A number of smaller, non-staining, refractile, cytoplasmic granules are also present in all preparations.

The nucleus is generally spherical, 7–10 microns in diameter, and resembles in size and structure that of Endamoeba disparata (fig. 55, Pl. 25). It agrees with that of the latter species in the possession of a delicate membrane, an irregular, granular mass of chromatin nearly filling the interior, and several irregular, deeply staining nucleolar bodies peripheral to this mass. Of these last there are usually not more than four. The aggregation of chromatin granules varies in extent from not more than half the diameter of the nucleus, in which case there is a broad clear zone surrounding it, to a nearly complete occupancy of the intranuclear region. The latter condition is more common. In that the nucleolar bodies are always against this mass, never just under the nuclear membrane, the arrangement differs from that of Endamoeba disparata.

Occasionally a different type of nucleus (fig. 56, Pl. 25) may be found in individuals which otherwise resemble the usual form. This has a concentrated central mass, about half or less of the nuclear diameter in extent, surrounded by a broad zone which is clear except for indefinite lines and granules. On its periphery are several (commonly three or four) rounded nucleolar bodies, in the centre is often apparent a relatively large karyosome-like structure, and the substance is indistinctly granular and lightly staining. Sometimes (fig. 57, Pl. 25) there are two nuclei of this type applied to one another so closely that the adjacent parts are flattened. It is noteworthy in this connexion that although as many as six of the amoebae with pairs of nuclei appeared on one slide, and several on others, no binucleate amoebae with separated nuclei were found. The somewhat smaller size of the amoebae in which such nuclei occur suggests p 2 that it may follow division. The arrangement is similar to that in Sappinia diploidea (Hartmann and Nâgler, 1908).

Two poorly preserved late anaphases showed that in this amoeba there are no filiform chromosomes such as those of Endamoeba disparata, but instead small, concentrated masses of chromatin at the poles.

From 19 to 35 microns in longest diameter, averaging 26. Ingested food material not abundant. Endoplasm contains a large number of granules which stain deeply with iron haematoxylin and resemble some mitochondria in size and form; these may be especially abundant about the nucleus. Nucleus 7–10 microns in diameter, irregular mass of granular chromatin occupies most of the central zone, several rounded nucleoli, at periphery of this. In some individuals chromatin is more concentrated in nuclei which may be single or, less commonly, in couples closely applied to one another. Xylophagous. Habitat: the intestine of Mirotermes panamaensis Snyder from the Canal Zone, Panama.

From the point of view of the student of Protozoa and of those who are concerned with the host-parasite relationships in termites, these amoebae possess several features of interest. First of all they are noteworthy because of their peculiar nuclear structures and xylophagous habits; in the former of which four, in the latter all of the species differ from other entozoic amoebae which have been described, whether from invertebrates or from higher animals. Apart from brief notes concerning these forms given by myself (1925, 1926), Cleveland (1926), and Snyder (1926), I am not aware of any published records of amoebae from termites; but before I obtained this material Light had observed a small, apparently non-xylophagous species in Kalo - termes simplicicornis. In the second place, at least one of the two largest amoeba has some characteristics which indicate a relationship to Endamoeba blattae, which has hitherto been isolated.¹ Furthermore, several questions suggest themselves concerning the relation between these Protozoa and their hosts, and although these cannot be answered at present for lack of complete knowledge and experimental data, several probabilities may be outlined. Do these amoebae share in the symbiotic relationship which has been demonstrated to exist between lower termites and their intestinal flagellates ? Are they simply commensals, or do they assist the host in the digestion of wood? Whence did Mirotermes and the other Termitidae obtain their infections ? The small intestinal amoebae from four other infected species of the Termitidae will be described later.

As few cysts have been seen, and these in only one species, and as the mitotic phenomena are well known in only one case, the typical nuclear structure is the only character which can be used to advantage for comparative purposes. Each of the five has a characteristic nuclear structure which differs from that of the others. Endamoeba disparata and Endamoeba sabulosa are obviously related, Endamoeba maj estas and Endamoeba simulans less obviously so both with each other and with the previously mentioned species, and Endolimax termitis stands alone in this group.

The nucleus of the last-mentioned species, having a large, compact karyosome like that of the genus Endolimax, closely resembles the nucleus of Endolimax ranarum (Epstein and Ilovaisky, 1914) or the * limax amoeba from Típula, Vahlkampfia sp. ?’ recorded by Mackinnon (1914) and others. If the genus Endolimax is accepted, Endo-limax termitis probably belongs there. Boeck and Stiles (1923) and Craig (1926) suggest that the genus is doubtful, and should perhaps be a synonym or sub-genus of Endamoeba.

The four species which have been placed in the genus End - amoeba (with the type Endamoeba blattae, not corresponding to that of Boeck and Stiles) seem to be more closely related to the type species of that genus than to entozoic amoebae of other well-established genera. Two of them resemble in nuclear structure Branchipocola (Breindl, 1925) more than others, but not sufficiently to warrant an assignment to that genus, established for an amoeba from the crustacean Branchipus grubyi. The alternative would be to give them distinct generic names, but it does not seem wise to multiply genera for these closely associated forms. If the definition of the genus Endamoeba can be extended to include such forms, perhaps they should at least be given sub-generic distinction.

English workers have pointed out the error in confusing the genus Entamoeba of Casagrandi and Barbagallo (type: Entamoeba coli) with the genus Endamoeba of Leidy (type: Endamoeba blattae). Because these type species are generally different, Endamoeba stands for Endamoeba blattae and cannot be used for the others, so that the entozoic amoebae of the coli type must be given another name. Perhaps the use of Entamoeba is not justified, in view of its similarity to the earlier name; then it must be something else (as Poneramoeba Lühe)-¹

Several amoebae of the Entamoeba coli type have been found in insects; as Lôschia hartmanni Mackinnon (1914) from larvae of crane flies; Entamoeba belostomae Brug (1922) from the water bug; Entamoeba apis Fantham and Porter (1911) from the bee; Entamoeba sp. from the cockroach, Lucas.

Since Endamoeba blattae has an important place in this discussion, I shall give a brief description of it from published accounts. Among those recently published, Wenyon’s summary of the morphology and life-history is probably available to all interested; Kudo (1926) has given a description of division; and Thomas and Lucas (1926) have published a brief account corroborating Mercier (1910). Inasmuch as Endamoeba blattae is a comparatively well-known species, I shall limit this sketch chiefly to a comparison of the occurrence, size, nuclear structure, and division with the same phenomena in Endamoeba maj estas and Endamoeba simulans.

Records of infection of cockroaches with Endamoeba blattae vary from 4 to 50 per cent. This incidence of infection differs from that of the amoebae from termites, which seem to occur in all individuals—a result of the perfect transmission by proctodaeal feeding.

Since reproduction by cysts plays only a minor part in the life cycle of these amoebae of termites, if, indeed, cysts occur at all in Endamoeba maj estas, the variation in size is not so great as in Endamoeba blattae. A comparison of size should therefore take into consideration only the larger vegetative individuals of Endamoeba blattae, for those hatched from cysts are, of course, much smaller than they later grow to be. These larger individuals are smaller than Endamoeba maj estas, but are comparable in size to Endamoeba simulans, averaging above 50 microns and ranging up to 130 microns. The diameter of the nucleus is also similar to that of this latter species.

The nuclear membrane of Endamoeba blattae is remarkably thick, much thicker than in our forms. Within this two zones may be distinguished, the outer dark and the inner clear. In life the outer zone contains numerous refractile granules which dissolve on fixation and expose a finely reticular structure on which are numerous chromatin granules. Whether or not these refractile granules exist in the termite amoebae is unknown, for I was unable to study the living nucleus with these in mind. The inner zone is coarsely reticulated or alveolar, free from chromatin granules, and otherwise, according to Kudo (1926), shows no particular structure. In Janicki’s previous work (1909) chromatin granules were found in the inner zone, and a large karyosome, according to Kudo the ‘separated central reticulum ‘, was described, but these observations have been contradicted by later workers. Between these two zones occur numerous nucleoli similar in form and arrangement to those of Endamoeba simulans. Wenyon states that these are large chromatin granules, but Janicki, Mercier, and Kudo regard them as nucleoli, as I have done. In probably degenerating nuclei Kudo finds that these nucleoli may fuse together in threads or masses, when they resemble those of Endamoeba maj estas. In the latter, however, as regards the form of the nucleoli, there is no question of degeneration.

However, the two zones in the nucleus of Endamoeba simulans are not comparable with those of Endamoeba blattae, for in the former the outer zone is clear and the inner, darker zone contains the chromatin, mostly in an irregular karyosome-like structure. The most striking resemblance is in the arrangement and form of the nucleoli. Endamoeba majestas differs still more, resembling the type only in the arrangement of differently formed nucleoli, a form which may sometimes appear in Endamoeba blattae.

After the structure of the resting nucleus, we have next to consider the changes taking place during karyokinesis. As in the amoebae from termites the nucleoli of Endamoeba blattae disappear early in this process and their place is taken by fine chromatin granulations, which I do not believe to be related to the former structures. Chromatin granules then appear in the central reticulum, enlarge and organize into a thread which later breaks into a number of segments. This course of events is very similar to that of Endamoeba simulans and also of Endamoeba majestas, where the nucleoli disappear and filiform chromatic bodies appear in the central region. Mercier (1910) calls these pseudochromosomes, while Kudo (1926) finds them to be linear groups of chromatin granules. In Endamoeba blattae these become arranged into two groups and move toward the poles, and apparently no centrosomes or spindle structures have been found in the vegetative amoebae since the doubtful observations of Janicki.

Furthermore, the cysts of Endamoeba simulans somewhat resemble those of Endamoeba blattae. Those of the latter measure from 30 to 50 microns in diameter and possess thick, hyaline walls; while those of the former, with similar walls, are somewhat larger. Since only three cysts of Endamoeba.simulans were found, little is known of the number of nuclei, but morphologically these seem to resemble those of the other species in the irregular distribution of a limited number of very large chromatin granules on a reticular substratum, and sometimes a denser aggregation perhaps comparable to what Mercier interprets as a karyosome containing a centriole.

Inasmuch as Endamoeba simulans and Endamoeba blattae, besides resembling each other in size and encystment, possess certain nuclear features in common, it seems defensible to conclude that the two are related, though possibly they differ sufficiently to warrant sub-generic distinction. While Endamoeba majestas, with the small central karyosome surrounded by a dense aggregation of chromatin granules and a clear alveolar peripheral zone in which are situated the elongate nucleoli, shows few similarities during the resting stage, it shows some during organization of the nuclear material prior to mitosis.

Endamoeba disparata can be more definitely placed, for its manner of nuclear division is known. Both in the structures of the resting and dividing nucleus it stands apart from most intestinal amoebae hitherto described. It is evident that in its peripheral nucleoli and large, granular mass of chromatin, with no karyosome; in the curious spindle form of the nucleus which is as prevalent as the other type; and in the formation of filiform chromosomes during division it is unique among intestinal amoebae. In this last point, it seems to accord in some degree with the type species of the genus. It also differs from many other entozoic amoebae in the absence of intranuclear centrosomes and spindle-fibres. The type of division makes an interesting addition to the cytology of this group of Protozoa.

It has already been stated that in the structure of the resting nucleus, as well as in some general characters of the body, Endamoeba sabulosa shows some affinities to Endamoeba disparata, but in the absence of knowledge of the nuclear division, apart from two fragmentary observations of telophases which seem to show some differences from the latter species, the closeness of this relationship is uncertain. The significance of the other type of nucleus in this species, and of the peculiar S a p p i n i a-like binucleate condition, is at present unknown.

One would expect to find cysts, since they are the propagative stages of most entozoic amoebae. Such intestinal amoebae as Endamoeba blattae and Entamoeba coli could not often, if ever, infect new hosts without the aid of cysts. But, when a more direct method of infection is provided, as in the case of Endamoeba gingivalis, cysts may not be formed. The habit of termites in taking proctodaeal food directly from one another makes the transmission of active Protozoa simple, so that encystment may not always take place. However, although no cysts were seen on these slides except in the case of Endamoeba simulans, it is not certain that there are none in the life-histories of the other amoebae, for, owing to the death of the amoebae in the transported termites, careful investigation of faecal deposits was not possible.¹

Finally, the parasites of nucleus and cytoplasm may be mentioned. Nucleophaga infests the nuclei of four species: Endamoeba disparata, Endamoeba majestas, Endamoeba simulans, and Endolimax termitis, most commonly in the first; and Sphaerita is not uncommon in the cytoplasm of Endamoeba simulans.

There remains for consideration the problem of the relation of these amoebae to their hosts. This has already been briefly discussed by Cleveland (1926) and Snyder (1926). Although abundant, the amoebae do not occur in the closely packed condition usual among the flagellates in the Kalotermitidae and Rhinotermitidae, where a section through the intestine will show it to be completely filled with Protozoa. Because of the small numbers, it does not seem probable that they could prepare all the food used by the host, as is supposed to be the case with the true symbionts. Also, Mirotermes will live at least for many weeks without the amoebae. On the other hand, certain considerations render it very possible that the amoebae may be at least of some service to the host for the same reason which gives the flagellates this credit. The wood on which this termite feeds seems to be in the same condition as that eaten by lower termites, and probably is acted upon by the amoebae in the same manner as it is by the symbiotic flagellates.

Infection may have been carried on for ages from individual to individual and colony to colony. But whence did it originally come ? Was it obtained during development from lower termites most of the nearer relatives of which have since lost their amoebae, for, as already stated, none of those examined, except Kalotermes simplicicornis, contain any; or has it been acquired later by the Termitidae, which have lost the large flagellates ? The possibility of obtaining infection from other coexisting species of termites seems remote, because of the isolated habits of termite colonies; and also improbable, for the same reason, seems the possibility of infection by cysts from other animals. If there is any real affinity between Endamoeba simulans and the others and Endamoeba blattae, as well as between Endamoeba disparata and the smaller amoeba of the cockroach, the line of development must lead to the ancestors common both to the termite and cockroach.

In conclusion it may be well to reiterate what Snyder (1926) has already mentioned; that is, that the statements sometimes encountered that the Termitidae contain no Protozoa are unwarranted. It is true that at least most of those examined contain no faunas comparable to those of other families, but these records from Mirotermes evince the possibility of abundant infections being found in some of the Termitidae which are yet to be studied. Others may contain small amoebae and flagellates which are few in number, inconspicuous and difficult to study, but nevertheless they are of considerable interest.

Thanks are due to Dr. L. R. Cleveland, who enabled me to obtain this material while collecting for him through a grant given to him by the Bache Fund of the National Academy of Sciences. I am also indebted to Dr. T. E. Snyder for identification of the termites and to Mr. J. E. Zetek, resident custodian of the Barro Colorado Island Biological Station, for assistance in collecting.¹

For the sake of completeness, a brief note concerning a gregarine found in Mirotermes panamaensis may be given. Gregarines have been found infrequently in the intestines of certain termites. Leidy (1881) first recorded Gregarina termitis from Reticulitermes flavipes of the eastern United States: Ellis (1913) observed a larger form in western termites, probably Reticulitermes hesperus, in which I have often seen them; they are sometimes abundant in Termopsis; and Bequaert (1925) mentions having found a gregarinid ‘apparently of the genus Stylocephalus(Stylorhynchus) ‘lining the intestinal wall of Neo termes castaneus in Brazil. This last was ‘provided with a long, snout-like appendage or epimerite, fixed in the epithelium ‘. It seems possible, however, that this was Oxymonas, which I have found in Neotermes holmgreni as well as in all other members of the sub-family Kalotermitidae which I have examined. Oxymonas, which frequently adheres in large numbers to the intestinal wall by means of the long ‘proboscis superficially resembles in general form some stages in the lifehistory of Stylocephalus.

To these records I may add a description of the gregarine found on slides prepared from Mirotermes panamaensis. It obviously differs from Gregarina termitis, but in the absence of gametocysts or any stages other than the free gre-garines, it is impossible to define it completely. However, it is undoubtedly a new species, and may be named Gregarina mirotermitis.

Each of the two pairs observed consisted of two individuals in syzygy. One of these is figured on fig. 68, Pl. 26. In this the primite is larger than the satellite, which is attached to it by a deep cup in the protomerite receiving the narrowed posterior end of the anterior individual. Measurements from fixed material will be of value, though they cannot be accurate. The total length of the primite is 94 microns, of which the deutomerite is 79; the width of the protomerite is 16, of the deutomerite 39. The total length of the satellite is 65 microns, of which the deutomerite is 59; the width of the protomerite is 14, of the deutomerite 27. Thus this gregarine may be of about the same length as Gregarina termitis (60 microns), or may be larger, but it is proportionately narrower and the protomerite is relatively much smaller. The septa are distinct, and the surface of the body is delicately striated longitudinally. Host: Mirotermes panamaensis Snyder from the Canal Zone, Panama.

All figures were drawn with a 2 mm. oil-immersion objective, occulars 10 or 20, and a camera lucida at magnification 1,440 or 2,740. The material on Pls. 22, 23, 24 was fixed in Schaudinn’s fluid; the rest in Schaudinn’s or Bouin’s as stated. The staining of each figure is indicated by abbreviation in the legend, as follows: H., Heidenhain’s iron haematoxylin; D., Delafield’s haematoxylin; AC., alum carmine; M., methyl-blue and eosin.

Figs. 1–16. Endamoeba disparata sp. nov. from Mirotermes hispaniolae Banks. × 2,000. All figures may be measured by the scale of microns.

Figs. 1–2.—Entire amoebae showing food particles and resting nuclei with peripheral nucleoli and granular chromatin mass. The nuclei are somewhat deformed. H.

Fig. 3.—Resting nucleus showing linear arrangement of granules. D.

Fig. 4.—Nucleus showing arrangement of granules in definite threads. H.

Fig. 5.—An enlarged nucleus in entire amoeba showing concentration of chromatin mass. D.

Figs. 6–7.—Spindle-formed nuclei, which occur in half the amoebae. Note mammiform protuberance at one pole and granular chromatin mass. H.

Fig. 8.—An optical transverse section of a fusiform nucleus. D.

Fig. 9.—Fusiform nucleus in entire amoeba. Dense nucleolar substance, which always appears in iron-haematoxylin material, at one pole. H.

Figs. 10, 11, 13.—Spindle-formed nuclei stained in Delafield’s. Note the different staining of the polar nucleoli.

Fig. 12.—Nucleus showing concentration of chromatin mass, flattening of nucleolus, and enlargement. Possibly transitional to spindle shape. Compare with fig. 5. D.

Fig. 14.—Chromatin granules arranged in lines in fusiform nucleus. D.

Fig. 15.—An unusually large spindle-formed nucleus, with a large, flattened nucleolus. D.

Fig. 16.—A smaller amoeba. A.C.

Figs. 17–26. Endamoeba disparata sp. nov. Kary okinesis.

Figs. 27–34. Endolimax termitis sp. nov. from Mirotermes hispaniolae Banks. × 2,000. All figures may be measured by the scale of microns.

Fig. 17.—Spindle nucleus in prophase; chromosomes arranged in close mass which possibly follows fig. 14, Pl. 22. D.

Fig. 18.—End view of an early prophase stage in which there is some evidence of a V-like separation of chromosomes. D.

Fig. 19.—Lateral view of a similar stage, also some evidence of V-like separation of chromosomes. D.

Fig. 20.—Chromosomes loosely arranged in late prophase. D.

Fig. 21.—Entire amoeba, nucleus in anaphase. Food particles in body. D.

Fig. 22.—Anaphase. There are twelve chromosomes in each group. D.

Fig. 23.—A later anaphase in which the poles are becoming rounded. D.

Figs. 24-6.—Telophases showing elongation and disappearance of the strand connecting the two nuclei. Twelve chromosomes may be counted in each of these. D.

Fig. 27.—Endolimax termitis showing vacuolization of cytoplasm, karyosome, peripheral chromatin, and radial filaments in nucleus. D.

Fig. 28.—A similar stage, differently formed body. D.

Fig. 29.—Fewer peripheral granules. A.C.

Fig. 30.—Granules in karyosome, as also indicated in previous figures. D.

Fig. 31.—Possibly early stage of Nucleophaga in nucleus. H.

Fig. 32.—Little cytoplasm surrounding nucleus, which shows granulation of karyosome and central granule. H.

Fig. 33.—A stage of early organization for division. Chromatin threads and possibly intradesmose. D.

Fig. 34.—A large sporulation body of Nucleophaga. D.

Figs. 35–46. Endamoeba disparata sp. nov. Nucleophaga. ×2,000. All figures may be measured by the scale of microns.

Fig. 35.—A normal amoeba with the usual form of nucleus, inserted here for comparison with parasitized nuclei. H.

Fig. 36.—Large spheres in nucleus. Early in development of Nucleophaga, possibly a different form from most others on this plate. H.

Fig. 37.—An early stage in parasitism by Nucleophaga. Spherical, granular mass in nucleus. H.

Fig. 38.—Sporulation of Nucleophaga. A large mass of deeply staining spherules. H.

Fig. 39.—A similar stage in entire amoeba. H.

Fig. 40.—An optical section of a parasitized nucleus’showing degenerating chromatin in centre. H.

Fig. 41.—Surface view of a similar stage. H.

Figs. 42–3.—Similar stages stained in Delafield’s showing the pale stain of Nucleophaga.

Fig. 44.—A mass of developing spores showing chromatic granules within an envelope. H. Counterstain acid fuchsin.

Fig. 45.—Looser, more irregular arrangement of spherules. The amoeba is ingesting a piece of wood. H.

Fig. 46.—Thirty-eight large spherules which have perhaps ruptured the nucleus, for none is present. D.

Figs. 47–53. Endamoeba majestas sp. nov. from Mirotermes hispaniolae Banks.

Figs. 54–7. Endamoeba sabulosa sp. nov. from Mirotermes panamaensis Snyder.

Magnification as stated. Fixation of figs. 54–7 Bonin’s, others Schaudinn’s. Fig. 47.—Entire individual showing body-form and wood. D. X 1,056. Figs. 48–50.—Nuclei showing granulated reticulum, central mass of chromatin, anguilliform nucleoli, karyosome in fig. 50. A.C. × 2,000.

Fig. 51.—Central karyosome and irregular nucleoli. H. × 2,000.

Fig. 52.—Entire individual containing large pieces of wood. Note that this is much larger than the individual of fig. 47. H. × 575.

Fig. 53.—A single nucleus showing irregular nucleoli, karyosome, and granular chromatin. H. × 1,056.

Fig. 54.—A large individual of Endamoeba sabulosa. Note nucleus, small amount of wood in body, many mitochondria-like granules. B.H. × 2,000.

Fig. 55.—A smaller individual, similar structure. B.H. × 2,000.

Fig. 56.—A different form of nucleus, more concentrated. B.H. × 2,000.

Fig. 57.—A binucleate individual, with the nuclei closely applied to one another. This is not a cyst. B.H. × 2,000.

Figs. 58-67. Endamoeba simulans sp. nov. from Mirotermes panamaensis Snyder.

Fig. 68. Gregarina mirotermitis sp. nov. from Mirotermes panamaensis Snyder.

Magnification, fixation, and staining as stated. All figures except those of individual nuclei may be measured by scale.

Fig. 58.—Entire individual showing pseudopodial protuberances and nucleus in resting condition. B.H. × 1,056.

Figs. 59-60.—Isolated nuclei showing two zones, nucleoli, granular karyosome-like mass of chromatin, &c. B.H. × 2,000.

Fig. 61.—Nucleus parasitized by Nucleophaga. S.M. × 1,056.

Fig. 62.—Outer zone suppressed, nucleoli pale, chromatin in threads. B.H. × 2,000. ‘

Fig. 63.—A similar stage to previous one; the threads seem to be splitting. S.H. × 2,000.

Fig. 64.—A similar stage to fig. 62. Nucleoli seem to be absent. B.H. × 2,000.

Fig. 65.—A large individual, chromatin threads formed in nucleus. B.H. × 1,056.

Fig. 66.—A binucleate cyst. B.H. × 1,056.

Fig. 67.—A quadrinucleate cyst. B.H. × 1,056.

Fig. 68.—Gregarina mirotermitis. Two individuals in syzygy. B.H. × 1,056.