Scyphidia (Gerda) acanthoclini n.sp. (Pl. 10, figs. 78 and 79; Pl. 11, fig. 80).

Of the 76 examples of Acanthoclinus quadridactylus examined during these investigations, 55 were handled under field conditions. All the gill smears (Davis–Worcester fixative) from these 55 kelpfishes subsequently proved to contain trichodinids, and six of them were also positive for Scyphidia. Two of the fishes infested with the latter peritrich were from Point Jerningham, Wellington (20.12.50), one was from Otahei Bay, Ruapukapuka Island, Northland (8.2.51), while three were from Tolaga Bay, East Coast (14.2.51).

Between February and November, 1951, 21 examples of A. quadridactylus ranging from 28 mm. to 180 mm. in length were examined alive in the laboratory following their collection at Point Jerningham. The gills of all these fish, studied as whole mounts in sea water, proved to be very heavily infested with scyphidians as well as with trichodinids. However, while all cover slip smears made from these gills contained trichodinids, only those smears from crushed gills were positive for scyphidians. The reason for this is that the latter peritrichs are so firmly attached to the gill lamellae by means of the scopula, that superficial smears made in the normal manner only rarely result in their being dislodged. This fact indicates that the incidence of Scyphidia on A. quadridactylus is much higher than field results had suggested, and may perhaps help to account for the general paucity of records of members of this genus from marine fishes.

Morphology.

Living Material. Fully expanded, active examples are elongate and urn-shaped. The diameter of the peristome is equivalent to that of the body proper at its broadest part, which is some third of the total length from the anterior extremity. There is no stalk-like process between the posterior extremity of the body proper and the scopula, the attachment organelle. Twenty fully expanded individuals were measured, their dimensions being as follows:—Length, 50·4μ to 65·3μ (av. 59·7μ); greatest breadth, 30·9μ to 36·9μ (av. 31.4μ).

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Trophic individuals are themselves colourless, although they usually contain food inclusions of a greenish or reddish colour. The pellicle of the body proper is transversely annulated, the annuli, which number from 40 to 60, being so fine that they can only be made out under the most favourable lighting conditions. There are two rows of peristomal cilia. These reach the vestibule after making 1 1/4 to 1 1/2 circuits of the peristomal disc, the surface of which is raised up into the form of a dome. The nuclei cannot be made out, being obscured by food inclusions, but the vestibule, gullet and contractile vacuole are easily distinguishable.

Stained Material. Partial or complete contraction takes place during fixation. Partly contracted individuals (Figs. 78–80) range in length from 31·8μ to 60.2μ (av. 46·6μ) and in breadth from 24·4μ to 46·1μ (av. 35·8μ) (50 measured) Fully contracted ones range from 30·0μ to 45·0μ in diameter (20 measured). Although the pellicle of contracted examples is often thrown into folds (Fig. 78), annuli are never apparent in fixed and stained material. The scopula, which is clearly demarcated from the body proper, is biconcave in lateral view. It may be either cupped about the extremity of a gill filament of the host, or partly embedded (Fig. 78). The lateral portions of the scopula are often drawn out into two terminally rounded extensions equal (Fig. 78) or unequal (Fig. 80) in size.

The vestibule narrows towards its posterior extremity, where the cytopyge is located. The cytopharynx runs deep into the cytoplasm, ending blindly at a point, often marked by a food vacuole, from half to two–thirds of the total body length from the peristome. There are numerous vacuoles and food inclusions in the alveolar cytoplasm. The contractile vacuole is situated close to the vestibule, and may attain 9μ in diameter. A canal leading from this structure to the peristomal disc, as has been described in some species of Scyphidia, has not been observed.

In most individuals, the spherical micronucleus lies immediately beneath the peristomal disc (Figs. 79 and 80), although it is occasionally situated deep in the cytoplasm (Fig. 78). The micronucleus averages 1.8μ in diameter, its intensely staining and relatively large karyosome being 1·2μ in diameter. The macronucleus is ribbon-like, and has rounded extremities which are frequently somewhat inflated (Fig. 80). Its central portion coincides with, or is close to, the median longitudinal axis, and both its ends are reflected at approximately right angles to the central portion. They then follow the transverse plane, sometimes paralleling the body contour while remaining deep in the cytoplasm, and sometimes curving out towards the pellicle. The macronucleus is from one and a-half times to twice as long as the body, and varies from 1·2μ to 4·0μ in breadth.

Biology.

The cilia of the inner peristomal row are either directed vertically or else are slightly inclined over the surface of the disc, in fully expanded trophic individuals. These cilia beat more vigorously than do those of the outer row, which incline outwards. The influence of the feeding vortex extends out to a distance of some 200μ from the peristome. All floating particles entering this vortex are swept down to the disc. The organism exercises a considerable measure of selection with regard to the material ingested. Bacteria and unicellular algae are readily ingested, but blood corpuscles, always plentiful in fresh whole mounts of gill lamellae, are forcibly rejected on contacting the cilia at the entrance to the vestibule. When masses of erythrocytes accumulate about the anterior end of one of these scyphidians its peristomal cilia becomes motionless or almost so for some 30 seconds, then suddenly recommence their activity in a very vigorous manner. This procedure is repeated, often several times, until the obstruction is cleared away.

At full diastole, the contractile vacuole is some 10μ in diameter. Systole occupies approximately one second, the liquid contents of the vacuole then being expelled directly into the vestibule. Five successive cycles from the commencement of diastole to systole were timed for one active trophozoite. The durations were 76 seconds, 68 seconds, 75 seconds, 78 seconds and 73 seconds.

Partial contraction, which is induced by tactile stimuli, takes the form of an infolding of the peristome together with a broadening and shortening of the body. Fully contracted individuals are commonly observed. These are often pre-fission or pre-telotroch stages. Their peristomal cilia are completely withdrawn, the peristome is closed over, and the body is so retracted as to be ovoidal or spheroidal.

Division by binary fission, and conjugation, have been observed in living material only. Dividing individuals are not attached to the host, while macro-conjugants remain attached to the gill filaments. The small, spherical micro

conjugants are motile, and have a single circlet of locomotory cilia. The initiation of the telotroch stage was observed once only, in a preparation from a 73 mm. kelpfish collected at Point Jerningham on 19.11.51. A fully contracted scyphidian, its scopula still attached to the tip of a gill filament, was undergoing slow, pulsating movements. Although its peristomal cilia were completely withdrawn within the retracted peristome, a basal circlet of cilia was in evidence just anterior to the scopula. This organism was unfortunately lost due to the drying up of the preparation. A search of fixed and stained smears from the gills of the same kelpfish failed to result in the discovery of any further telotroch stages.

Systematic Position.

Some 30 species, several of them inadequately described, are at present included in Scyphidia Dujardin. All of these are more or less cylindrical in shape, and have an attachment organelle, the scopula, at the posterior extremity. The genus is represented in both marine and fresh water habitats. Most of the species are ecto-or endocommensals of invertebrates, while some are ecto-commensal on fishes. Kahl (1935) revised the genus, while Hirshfield (1949) followed him in recognizing the sub-genera (Scyphidia) and Gerda. In (Scyphidia) a stalk-like process is interposed between the scopula and the body proper, while in Gerda the scopula is directly and more or less broadly based on the body proper. The scyphidian of Acanthoclinus quadridactylus belongs to the latter sub-genus, as do 12 other species.

Five of the species of the sub-genus Gerda differ from the remainder, including that under consideration, in having a smooth pellicle which lacks annuli. These are S. fischeri Vayssière, S. ambigua Penard. S. patellae Cuénot, S. purniensis Ghosh, and S. ubiquita Hirshfield.

The trophozoites of three of the other species differ from those of the New Zealand one in that the central part of the body is encircled by a band of cilia. This feature is shared by S. tholiformis Surber, S. macropodia Davis and S. ameiuri Thompson et al, the band being composed of two rows of cilia in the first-named species.

The remaining four previously described species of the sub-genus Gerda differ from the present species in the form of the macronucleus. This structure, which is ribbon–like and very long in the New Zealand species, is sausage-shaped in S. annulata Edmondson and S. physarum Lachmann, and ovoidal in S. gasterostei Fauré-Fremiet and S. terebellidis Precht.

Other data, including references, concerning the above species, being available in a recent summary furnished by Hirshfield (1949), is not recapitulated herein.

Four species of Scyphidia (Gerda) have previously been recorded from fishes These are S. gasterostei, S. tholiformis, S. macropodia and S. ameiuri. All but the first of these species are found on the gills of North American fresh water fishes. S. gasterostei is the sole species of the subgenus previously known from marine fishes, although other species have been described from marine invertebrates.

Precht (1935) recorded S. gasterostei from Gasterosteus aculeatus collected from Kiel Harbour, also from the algal zone at Friedrichsort. Thirty-two examples of Gasterosteus pungiteus collected in the same habitat at Friedrichsort were negative for scyphidians. This apparent host preference may be at least

partly due to the fact that G. aculeatus is characteristically found in fresh water and brackish habitats (Yonge, 1949), while G. pungiteus is an exclusively marine fish. The New Zealand species under consideration exhibits a striking host specificity. Although it is of such high incidence on Acanthoclinus quadridactylus, I have never recorded it from blennies or other fishes sharing rock pools with this species, or for that matter from clingfishes (Diplocrepis puniceus) associated with A. quadridactylus on the rocky shore at Point Jerningham. Only one representative of the subgenus (Scyphidia, S. scorpaenae Fabre-Domergue 1888, is known from a marine fish. The host in this case is a European species of Scorpaena which, like Acanthoclinus quadridactylus but unlike Gasterosteus aculeatus, is restricted to marine habitats.

S. gasterostei differs from the parasite of A. quadridactylus not only in having an ovoid macronucleus but also in having its pellicle adorned with branching rows of small quadrangular elevations instead of with annuli (Precht, 1935). S. scorpaenae differs subgenerically from the New Zealand scyphidian. The latter species, the third of its genus and the second of its sub-genus to be described from fishes occupying a marine habitat, differs in detail from the previously known members of Gerda as detailed herein; it is hence designated Scyphidia (Gerda) acanthoclini n.sp.

The type slide has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z19), while paratypes are in the collection of the Department of Zoology, Victoria University College.

Caliperia longipes n.gen., n.sp. (Pl. 11, figs. 81–83).

This interesting peritrich was abundant on the gills of all 29 examples of Oliverichtus melobesia collected at Wellington, but was absent from those of the single example of this fish from the Bay of Islands. It was also recorded from three Ericentrus rubrus collected at Fisherman's Creek, Island Bay, Wellington.

The organism herein described belongs to the family Scyphidiidae Kahl, being a sessile peritrich lacking a lorica, an anterior neck, and a stalk. Its attachment to the host is direct, and is made by means of two long posterior processes here viewed as homologous with the scopula of scyphidians of the subgenus Gerda (Claparède and Lachmann). It is morphologically close to this subgenus and to Ellobiophrya Chatton and Lwoff as regards the body proper, and in the possession of the two posterior processes bears a strong superficial resemblance to the latter genus from which it differs generically as described hereunder.

Morphology.

Living Material. This animal displays much less activity than is exhibited by members of the genus Scyphidia, resembling in this respect Ellobiophrya donacis Chatton and Lwoff, 1923, the type and only known representative of its genus. The ovoidal body is invested by a smooth pellicle which appears to lack annuli. The macronucleus, gullet and contractile vacuole are usually partly or wholly obscured by a mass of greenish or yellowish spherules (Fig. 81). These are food particles, apparently of algal origin, and seem analogous with the type 2 cytoplasmic inclusions described from E. donacis by Chatton and Lwoff (1928). The peristomal disc, which has not been seen in a more expanded condition than that illustrated in Figs. 82 and 83, is more or less invaginated. The peristomal cilia protrude as a sheaf through a comparatively narrow apical aperture as in

E. donacis. The body proper tapers posteriorly, its cytoplasm being clearly demarcated from the attachment organelle as is that of Scyphidia (Gerda) from the scopula. This organelle tapers posteriorly from the broad base of its attachment to the body proper, for some 10 or 15 microns. It then bifurcates, to form two long, slender and flattened processes. These processes, which are usually of equal length, pass one on either side of a gill filament beneath which their distal extremities closely overlie one another. They possess no terminal connecting organelle, but are simply rounded off distally. A continuous rod–like structure, which can be seen quite plainly in the living animal (Fig. 81), runs from the distal end of one process to that of the other (Fig. 82).

Stained Material. Fixed and stained examples (Worcester's fluid/Shortt's haematoxylin/picric/eosin) show the same general contours as do living ones. The body proper ranges from 31·2μ to 68·4μ by 24·0μ to 52·6μ (av. 51·5μ by 38·8μ). The two posterior processes (which, like the stalk of vorticellids, easily become detached during the preparation of smears) are some 5μ to 6μ in breadth. They range from 3·5μ to 155·6μ in length, averaging 90μ.

The apical aperture, through which the peristomal cilia protrude, is, like that of Ellobiophrya donacis, bordered by a sphincter-like, siderophilous rim (Figs. 82, 83). There are two rows of peristomal cilia, the disposition of which again closely parallels that of those of E. donacis. Originating from a mound towards the centre of the peristomal disc, these make approximately one and a-quarter turns of the disc in helical fashion, finally turning sharply inwards and downwards into the vestibule (Fig. 82). The latter structure is large and funnel-shaped, and terminates distally at a cytostome. At the point of junction of the vestibule and cytopharynx there is a constriction (Fig. 82), beyond which the cytopharynx swells out. This latter structure, as seen in lateral view, resembles the bulb of a pipette. It ends blindly within the posterior half of the body, usually in the immediate vicinity of the macronueleus (Fig. 82). A contractile vacuole, which attains a diameter of from 10μ to 12μ, is located behind the vestibule, into which its contents are discharged at systole.

Numerous lipoid granules of less than 1μ in diameter are distributed throughout the cytoplasm of the body proper (Fig. 83). These resemble the type 1 cytoplasmic inclusions described from E. donacis by Chatton and Lwoff (1929). From a few to many food vacuoles, containing spherical masses of homogeneous material staining a greyish colour, are usually present also. The cytoplasm in general is of a finely alveolar structure.

The macronucleus, like that of many scyphidians but unlike that of E. donacis, is more or less U-shaped and in the form of a broad band (Figs. 82, 83). Measuring up to 80μ in length (av. 66μ) and averaging 10μ in breadth, it contains many small chromatin spherules each surrounded by a clear halo. The lenticular micronucleus measures some 3·5μ by 1·5μ, and stains a uniform black. It is usually situated more or less centrally and just beneath the peristomal disc (Fig. 82).

The cytoplasm of the attachment organelle is homogeneous and stains much more lightly than does that of the body proper. The continuous rod-like structure running through the posterior processes stains a uniform greyish-black. Many small siderophilous granules are present in the cytoplasm of the two processes, but absent from that of the non-bifurcated proximal part of the attach

ment organelle. In some respects, notably in its staining reaction and in the presence of siderophilous granules which may possibly be composed of thecoplasm, the rod-like structure resembles the spasmoneme found in the stalked peritrichs. Unlike a spasmoneme, however, it is neither inserted into the base of the body proper, nor does it display contractility. Its function is probably largely a supporting one, and it is considered that it may be homologous with the rods or myonemes (probably derived from thigmotactic cilia) occurring in the scyphidian scopula. The pellicle of the posterior processes, while smooth in the living animal, may be partly contracted and thrown into folds resembling annuli in fixed and stained specimens (Fig. 82).

Systematic Position.

As already indicated, this peritrich of O. melobesia and E. rubrus is of typically scyphidian morphology. Its affinities with Ellobiophrya donacis—notably in the possession of the siderophilous rim bordering the peristomal aperture—are particularly strong. It differs from the latter species in the form of the macronucleus, which in E. donacis is elliptical and deeply invaginated (Chatton and Lwoff, 1929). In view of the wide differences in macronuclear morphology within the genus Scyphidia, this is hardly of generic significance. The attachment organelle of this animal is, however, quite unique among the known ciliates, and quite distinct from the superficially similar organelle of E. donacis.

Whereas in the organism under discussion the posterior processes are bifurcations of a process continuous with, but clearly demarcated from, the body proper, the two processes which unite to form the suspension ring of E. donacis are posterior prolongations of the body itself, from the cytoplasm of which their own cytoplasm in no way differs (Chatton and Lwoff, 1929). In the latter species, too, these structures show no traces of a central supporting rod. Instead of being flattened and strap-like, and of equal length, as in the present species, they are cylindrical in shape and of unequal length. But the most significant point of distinction lies in the fact that in E. donacis the posterior processes are swollen basally, the swellings being closely applied to one another when the processes are firmly united about a gill bar of the host (the European lamellibranch Donax vittatus da Costa) and tightly anchored together by button-like structures made up of bundles of small rods, as if by a “bouton de chemise” (Chatton and Lwoff, 1923, 1929). The union achieved by this anchoring device is so complete that in the preparation of smears it is never broken at the centre of its isthmus-the two halves of the structure, still securely fastened together, are detached with one or other of the posterior processes. In the New Zealand peritrich the posterior processes, although usually closely applied to one another distally, are not immovably united by any such device. The organism is firmly supported by its attachment organelle merely by the close application of each of the strap-like posterior processes to either side of a gill filament, the central rod-like structure probably being sufficiently elastic to enable a continual tight pressure against the gill filament to be maintained. In some of the organisms observed in vitro the posterior processes did not even meet one another distally. The transverse filaments of the gill grid of Donax vittatus, to which E. donacis becomes attached, are of course very much more slender than the gill filaments of the fishes serving as hosts to the organism under discussion—so slender, in fact, that an animal of the order of size of a peritrich could hardly be securely supported upon them by means of an organelle

establishing attachment merely by latteral pressure. The adaptation of E. donacis to the nature of the substratum takes the form of a ring, by means of which the animal hangs from a transverse filament of the gill grid of the mussel after the fashion of an ear-pendant or a padlock (Chatton and Lwoff, 1929). Being thus securely, albeit loosely, attached to the gill grid, E. donacis still has limited freedom of movement and is able to revolve about its support. The organism under discussion, on the other hand, is, like a scyphidian, so tightly anchored as to be denied all opportunity for progressional movement.

Chatton and Lwoff (1928) considered the swellings and button-like bundles of rods on the distal extremities of the posterior processes of E. donacis to be homologous with the scopula of Scyphidia. Such an organelle may perhaps have been derived as a specialization in an organism with an ancestor of the ScyAphidia(Scyphidia) type, having the body proper sharply constricted basally anterior to the point of attachment of the scopula. One can envisage the development of the organelle by the bifurcation of the scopula and of the basal constriction of the body, together with the prolongation of the latter. This hypothesis would account for the continuity of the cytoplasm of the body proper with that of the posterior processes in Ellobiophrya, as well as for the presence of the distal uniting organelle.

On the other hand, the organism under discussion has affinities not with Scyphidia (Scyphidia) but with Scyphidia (Gerda). I consider that the entire attachment organelle in this animal is homologous with the scopula of scyphidians of the latter subgenus, and that it was probably derived by the elongation of the lateral extensions of the scopula of an organism resembling S. acanthoclini n.sp., together with an elaboration of the skeletal rods to provide the central supporting rod.

It is considered that the differences between this peritrich and all other members of the Scyphidiidae, including Ellobiophry, are sufficiently great to warrant the establishing of a new genus; and it is accordingly proposed to designate it Caliperia longipes n.gen., n.sp., the generic name being in reference to the caliper-like appearance of the posterior processes and the specific one to the striking length of these processes.

The type slide is in the collection of the Dominion Museum, Wellington (catalogue number Z20), and paratypes are in my own collection and in that of the Department of Zoology, Victoria University College.

Trichodina Ehrenberg.

Tripathi's (1948) summary of data on Trichodina covered 31 species, but omitted from consideration T. baltica Quennerstedt, 1869, T. (Cyclochaeta) scorpaenae (Robin, 1879), [T. (Anhymenia) scorpaenae (Fabre–Domergue, 1888)], T. entzii Breitschneider, 1935, T. scoloplontis Precht, 1935, T. terebellidis Precht, 1935, T. (Cyclochaeta) astericola (Precht, 1935) and T. urechi Noble, 1940.

Regarding T. (Cyclochaeta) scorpaenae (Robin) and T. (Anhymema) scorpaenae (Fabre–Domergue), Kahl (1935) considered that the first of these species is a Cyclochaeta, while the second is a Trichodina, both of them thus being valid species. However, Hirshfield (1949) felt that this is a case of homonymy, if not synonymy. From Robin's (1879) illustrations and brief description (reproduced by Saville-Kent, 1880–82—his Pl. XXXI, figs. 46 and 47) I fail

to see any reason for regarding his trichodinid as a Cyclochaeta. The cirri–like structures shown in the lateral view are obviously intended to portray not aboral cirri but the adoral cilia. This organism should be retained in the subgenus (Trichodina), and thus has priority over T. scorpaenae (Fabre-Domergue). Both Robin's species and the Cyclochaeta which Kahl states is also found on scorpaenid fishes are in need of further study and re-description in the light of modern knowledge.

Of the other species overlooked by Tripathi (1948), Hirshfield (1949) pointed out that two belong not to Trichodina but to Urceolaria, proposing the following new combinations:—

(Trichodina scoloplonts Precht) = Urceolaria (Leiotrocha) scoloplontis (Precht).

(Trichodina urechi Noble) = Urceolaria (Leiotrocha) urechi (Noble).

Hirshfield also indicated that Trichodina terebellidis Precht is a nomen nudem, the organism in question being recognisable from Precht's description—but only generically—as an urceolarian.

Both Tripathi and Hirshfield failed to note Trichodina baltica Quennerstedt, 1869, described from the marine gasteropod Neritina fluviatilis, and Trichodina entzii Breitschneider, 1935, an endoparasite from the urinary bladder of the edible frog Rana esculenta in Holland T. entzii may have priority over T. vesicularum, described from the urinary bladder of urodeles by Fauré-Fremiet (1943), for Fauré-Fremiet and Mugard (1946), who discussed a trichodinid from the urinary bladder of R. esculenta in France, considering it to be near to if not identical with T. vesicularum, also failed to notice Breitsehneider's paper.

Taking into account Trichodina (Cyclochaeta) tegula Hirshfield, 1949, Trichodina (Trichodina) ranae da Cunha, 1950, and the two new species described hereunder, the genus now includes 39 species—although further investigation of some inadequately described ones might well reduce this number by disclosing synonymy. Nine of these are endoparasites, of which two belong to the subgenus Vauchomia and one to Cyclochaeta. Of the 30 ectoparasitic and ectocommensal species four belong to the subgenus Cyclochaeta and the remaining 26 to the sub–genus Trichodina. All told there are thus now 32 species recognised as belonging to the latter subgenus, including the two described hereunder. As Tripathi summarized the data on 26 of these in his (1948) paper, this information is not repeated here.

Trichodina (Trichodina) parabranchicola n.sp. (Pl. 11, fig. 85; Pl. 12, figs. 86, 87, 90, 91, 93–97; Pl. 13, figs. 98–100).

This was the dominant trichodinid found on kelpfishes and clingfishes, in all the localities where collections were made and in all months of the year. All the examples of Acanthoclinus quadridactylus and of A. trilineatus, all but two of Diplocrepis puniceus, and all but one of Oliverichtus melobesia were infested. No other trichodinid was ever recorded from the latter host. The species was also collected from blennies (Ercentrus rubrus, Tripterygion varium, T. medium and Notoclinus fenestratus), on which it was always subordinate to Trichodina multidentis n.sp., wherever these fishes occurred in association with kelpfishes or clingfishes.

By far the greater numbers of the ciliates were concentrated on the branchiae of the hosts. A few were noticed outside the operculum, and fairly large numbers

were sometimes present on the ventral sucker of clingfishes. The number of trichodinids present on the branchiae usually ran into many hundreds. Fishes of all ages were infested, the examples of Acanthoclinus quadridactylus, for instance, ranging in length from 18 mm. to 180 mm.

Morphology.

Living Material. Trophozoites studied in hanging drop preparations are more or less round in surface view. Their diameter ranges from 28μ to 60μ, that of mature individuals being in the vicinity of 45μ. The height is variable, its ratio to the diameter ranging from 1:3 to 1:1. Young examples are more or less saucer-shaped, and older ones turban–shaped, in side view. Free swimming individuals move quite rapidly, always with the broad aboral end foremost. The adoral cilia are purely concerned with feeding, the aboral ciliary ring and the velum being used in locomotion. Conjugating individuals are commonly seen, the lower of the two conjugants continuing to display locomotor activity. The organism rotates continually while gliding about over the branchiae of the host, or for that matter, over a microscopic cover slip. Rotatory movement often, but by no means always, takes place during swimming. The skeletal rings and the macronucleus are prominent by dark ground illumination, and these structures show up well by bright field illumination when intra–vitally stained with methyl green.

Stained Material. Excellent permanent preparations were obtained from cover slip smears fixed in Davis's (1947) modification of Worcester's fixative, stained with Shortt's haematoxylin, destained with picric acid and counterstained with eosin.

Mounted examples in preparations made by the above technique ranged in diameter from 25·3μ to 52·3μ, the average for 50 mature individuals being 42·1μ. Their cytoplasm is granular and vacuolated, and usually contains food inclusions. That of trophozoites kept for a time in saline hanging drops together with ruptured gill filaments of the host, is often packed with ingested erythrocytes. No trace of host blood corpuscles has, however, been found in any of the thousands of trichodinids examined in smears made directly from the gills of freshly captured fishes. As in even very heavy infestations the gill epithelium of the host does not become eroded (as is sometimes the case in trichodinid infections—see Padnos and Nigrelli, 1942; Tripathi, 1948) T. parabranchicola n.sp. normally has no access to the blood of the host. This species subsists on microorganisms which it obtains from the water continually passing through the gills of the host, its relationship with which is thus a purely commensal one. Common food inclusions are small diatoms of the genus Cocconeis (see p. 132).

The adoral groove, with its two parallel rows of cilia, makes one and a-quarter turns of the anterior part of the body before dipping into the vestibule. The latter penetrates deeply into the cytoplasm, the cytopharynx continuing on from it and spiralling towards the centre of the body. A contractile vacuole is located close to the vestibule. The macronucleus of mature examples is horseshoe-shaped as seen in surface view, and occupies a median position in the body. One end is somewhat tapered and bluntly pointed, while the other is smoothly rounded. This structure (Fig. 91) ranges up to 85μ in length (i.e., along the median line) and up to 8μ in breadth, its average in 50 mature trichodinids being 55·6μ by 6·0μ. It contains numerous chromatin spherules, each surrounded

by a clear vesicle. The spherical micronucleus is located towards the more pointed end of the macronucleus. In surface view, it frequently appears superposed on the latter structure (Fig. 91). The endosome of the micronucleus stains densely and uniformly. It averages 1·7μ in diameter, while the clear vesicle surrounding it averages 3·2μ in diameter. The macronucleus is never notched to receive the micronucleus, as is the case in some species of Trichodina.

Just anterior to the aboral disc the body is evaginated to form a thin and very plastic fold, the velum. Immediately posterior to this is a wreath of long cilia, the ciliary girdle. These structures are shown as they appear in surface view in Figs. 86 and 87. The aboral extremity itself is broad and somewhat convex. Here is located the skeletal complex, an apparatus by means of which the organism secures temporary attachment to the substratum. This complex is made up of three concentric rings. The innermost of these, the denticulate ring, is composed of from 16 to 26 separate units. Twenty-two is the mean number as well as the number most commonly present. The numbers and percentages of trichodinids having each number of denticles, derived from counts made from a random sample of 150 individuals from a heavy infestation on Acanthoclinus quadridactylus, are given in Table VIII (see also Text-Fig. 1).

Each denticle has a hollow, cone-shaped central body. The point of each cone fits into the concavity of the succeeding one (Figs 90, 98), the depth of penetration being governed by a shoulder near the base of the hook. Arising from near the base and on the inner side of each cone is a recurved ray, a slender and more or less sharply pointed structure extending towards the centre of the disc (Figs. 86, 87, 90, 98). Opposite each ray and on the outer side of the denticulate ring is a sickle–shaped hook (Figs. 86, 87, 90, 98), the concave side of which is thicker and stains much more deeply than the convex side (Figs. 86, 87, 98). The distal extremity of the hook is more or less sharply pointed in the majority of individuals (Figs. 86, 98), although in the older examples it may be bluntly rounded (Fig. 90). The diameter of the denticulate ring and the size of its component denticles increase with age (Text-Fig. 2). In the youngest post-division stages the diameter of the denticulate ring may be as little as 9·5μ, while in the largest example measured it was 30·0μ. The range in, and average of ring diameter for each denticular number are indicated in Text-Fig. 1. An average of 17·5μ was arrived at from the measurement of the diameter of the denticulate ring in 50 mature trichodinids having from 22 to 24 denticles. The average height of the denticular hooks in these same 50 examples was 4·0μ, although individuals having hooks of up to 5.4μ in height were noted. The length of the rays approximates that of the hooks.

Outside of the denticulate ring lies the striated band, the inner rim of which is anterior to the distal extremities of the denticular rays. In living examples this band curves outwards and downwards, its outer rim thus being posterior to the denticular hooks (Fig. 85). As seen when flattened and in surface view, the striated band varies in its outer diameter from 18·1μ to 43·4μ, the average

for the 50 mature examples referred to above being 28·1μ. There are from 7 to 9 striae per denticle.

Outside of the striated band, to the outer margin of which it is hinged, lies the border membrane. This membrane, which is some 1·5μ in width, is faintly striated, its striae being finer and more numerous than those of the striated band. The striae of both the striated band and the border membrane often appear in surface view to have a bead-like thickening at the distal extremity (Fig. 98). This appearance is due to the fact that the outer rims of the striated band and border membrane are frequently not completely flattened out. The border membrane covers the basal portion of the ciliary girdle aborally, the cilia of this girdle originating in a groove immediately posterior to the velum. These cilia range in length from 10μ to 20μ, and collectively form the chief locomotory organelle.

Growth and Reproduction.

Although both binary fission and conjugation have been observed for T. parabranchicola n.sp., the latter process has not been studied in detail. From the stages encountered it appears to closely parallel conjugation as described by Padnos and Nigrelli (1942) for their T. spheroidesi.

Binary fission follows the same course as in other trichodinids, the process having been described by numerous authors, including Wallengren (1897), Fulton (1923), Diller (1928), Padnos and Nigrelli (1942) and Davis (1947). Its course is briefly summarized here to provide a background for some observations concerning growth. While binary fission takes place, the organism remains in the trophic condition. The first indications of the onset of division are the appearance of a faintly staining ring on the aboral surface of the striated band, a little outside the distal extremities of the denticular hooks, and the rounding up of the macronucleus. Indentations appear in the velum on opposite sides of the skeletal complex, the striated band then splitting beneath these. The micronucleus divides mitotically while the macronucleus adopts a dumb-bell shape, dividing amitotically after the completion of micronuclear division. The denticular ring now becomes disarticulated in two places, these being located in line with the indentations in the velum and the splits in the striated band and border membrane. At this stage the whole of the skeletal complex adopts a figure of eight shape, the centre of which hinges about the articulation between the striated band and the border membrane (Fig. 98). Cytoplasmic cleavage takes place as the two halves of the denticulate ring round up into daughter rings. Each daughter ring is completed by the insertion of the point of the denticular cone at one extremity into the hollow of the cone at the other.

As division proceeds, the faintly staining ring on the striated band becomes resolved into a series of overlapping plates equal in number to the denticles of the parent ring. Once cytoplasmic cleavage has taken place these plates rapidly develop into denticles, the cones being formed first, then the hooks (Fig. 100) and finally the rays. While its units are in course of formation, the new ring lessens slightly in diameter, the denticles received from the parent meanwhile moving towards the centre of the disc. These parent denticles become disarticulated from one another (Fig. 99), then their hooks and rays are absorbed. The disarticulated cones persist for a short time on the inner side of the new ring (Figs. 93, 100), then these, too, are absorbed. Up to this stage the macronucleus

has been more or less round in shape (Fig. 93). This structure now elongates (Fig. 94), and while the daughter trichodinid continues to increase in size, it too becomes larger (Fig. 95) and finally adopts the characteristic horseshoe shape (Fig. 96). Meanwhile new striae are formed to bring the total up to that appropriate to the new denticular number, not by the formation of a new striated band, but by the development of a new striation between each two striae of the half-band acquired from the parent (Fig. 100). These new striae are faint at first and thicken rather gradually. Many individuals having a horseshoe-shaped macronucleus still show alternating thick and thin striae in the striated band, this fact indicating that they are relatively recent products of binary fission.

Binary fission may take place in individuals of a wide range of sizes (compare Figs. 98, 99, 100—all three of these are drawn to the same scale), although it occurs most commonly among those examples towards the lower limit of the size range. Most dividing individuals so far observed have had 18 or 20 denticles, the hooks being not much more than 2μ in height. The number of denticles formed in the new daughter ring is not necessarily quite the same as that in the parent ring. When an individual having an even number of denticles undergoes binary fission, each daughter usually receives exactly half of the parent denticles and half of the plates destined to form the components of the new ring (Fig. 98). Where the parent has an odd number of denticles, however, one of the daughters must, of course, receive an even number and the other an odd number (Padnos and Nigrelli, 1942, Text Fig. 4, G and H, of their T. spheroidesi). Furthermore, even where an equal distribution of denticles takes place, the daughters may receive different numbers of the plates from which the new denticles are to be formed. Thus the example illustrated in Fig. 100 has 10 denticles derived from the parent, but only 19 in the new ring.

As the organism grows, the diameter of its denticulate ring increases. In order to maintain full contact with one another, the individual denticles themselves become progressively larger (Text-fig 2). During growth, one (Fig. 97) or more new denticles may be formed de novo between two of those developed following binary fission. It is possible that Stein (1859-83, Abt. 2) observed this process—which later investigators seem to have overlooked—for he, being unaware of the mode of formation of the new denticulate ring following binary fission, stated that after division the normal number of denticles is gradually restored by the interpolation of new ones between those received from the parent.

Those individuals having a denticulate ring of from 9·5μ to about 21·0μ in diameter characteristically have seven striae to each denticle, while those having a ring of greater diameter than 21μ usually have eight or nine striae to each denticle. It is thus apparent that new striae are formed in the later stages of growth as well as following binary fission, although the actual process has not yet been observed.

There is no significant correlation between the size and morphology of this trichodinid and its occurrence on any particular host. The mean number of units in the denticulate ring may differ slightly among different populations of parasitized fishes, not only for fishes of different species but also for those of the same species. This might indicate the occurrence of strains or races of the trichodinid, or might merely be connected with the duration of the infestation.

Systematic Position.

The denticular number is accepted as one of the chief specific criteria in Trichodina. This number in the present species must only rarely if ever exceed 26, is characteristically 21–24, and averages 22. Those species of Trichodina (Trichodina) in which the denticular number reaches or exceeds 29, or is never as low as 24, are disregarded in the following comparison. The 14 species listed hereunder have sufficient in common with the one under discussion to bear close comparison with it (those having marine hosts are marked with an asterisk, and the bracketed numbers are the denticular numbers for each species):—

Tripathi (1948) stated that the number of denticles in T. pediculus ranges from 16 to 26, but Fulton (1923) and Mueller (1937) gave 24 to 26 as the typical number. According to Mueller, the diameter of the striated band of T. pediculus varies from 57μ to 85μ, while Saville-Kent gave the greatest total diameter of the body of this species as 1/360in. (= 70μ). T. pediculus is thus of much greater diameter than T. parabranchicola n. sp., further differing from the latter species in typically having a greater number of denticles and in occurring only in fresh water habitats.

T. steini, from European fresh water planarians, is morphologically close to T. pediculus. Claparède and Lachmann (1858) distinguished it from the latter species on the ground that its denticles lack rays. Possibly these authors based their species on young post-division forms of T. pediculus, in which the rays (the last part of the denticles to be formed) had not yet appeared. Division stages of trichodinids have more than once confused systematists. Thus Stein (1864) described as T. diplodiscus (nomen nudem) a two-ringed stage of T. pediculus, and Ariake (1929), despite the fact that good accounts of binary fission in Trichodina had already been published (Wallengren, 1897; Fulton, 1923), described no less than five new species from the various division stages of one particular trichodinid.

The endoparasitic habit of T. fariai distinguishes this species, described from the intestine of an American marine fish, Spheroides testudineus (L), from T. parabranchicola n.sp. A further point of difference is that the former species has not less than 24 denticles.

T. myakkae is an extremely small species, the diameter of the denticulate ring of mature examples ranging from 11μ to 12μ, and denticular rays being

lacking. The denticular hooks are straight and narrow (Mueller, 1937; Davis, 1947).

All the species described by Davis (1947) occur on various North American fresh water fishes. T. vallata has 18–21 denticles, the majority of individuals having 19 or 20, and there are 10 striae to each denticle. It differs from T. parabranchicola n.sp. in these features, also in that its adoral spiral is based on a prominent ridge. T. symmetrica is somewhat smaller than the New Zealand species, from which it further differs in that the usual number of denticles is 24 to 26, in having only five striae to each denticle, and in having a spindleshaped micronucleus. T. tumefaciens is close to T. parabranchicola n.sp. as regards both size and the number of denticles, but differs from the latter species in having more or less upright hooks and broad and bluntly rounded rays. Both T. bulbosa, a very small trichodinid, and T. bursiformis are unique in having spatulate denticular hooks which attain their greatest breadth near the outer end.

The remaining five species to be considered here are all marine ones. One of them, T. baltica, was described by Quennerstedt (1869) from the gasteropod Neritina fluviatilis in Sweden and briefly recorded by Precht (1935) from the same host at Kiel, while the other four are from fishes. T. baltica is inadequately described. From Quennerstedt's figures, the union of the denticles with one another is on quite a different pattern to that of other trichodinids, and the hooks themselves are very slender. These features might be attributable to Quennerstedt's having examined poorly prepared material in which the ring was not properly flattened out. T. baltica will have to be redescribed from fresh material before an adequate comparison of it with other trichodinids can be made.

T. scorpaenae (Robin), from European fishes of the genera Scorpaena and Trigla, is also in need of redescription. This species is of similar size and shape to T. parabranchicola n.sp., and was figured by Robin (1879) as having 22 denticles bearing sickle-shaped hooks. It is possible that T. labrorum described by Chatton (1910) from two European wrasses, is a synonym of T. scorpaenae (Robin). T. labrorum falls within the size range of T. parabranchicola n.sp. and has 21 denticles.

Fantham (1930) described T. clini from five species of South African kelp fishes (fam. Clinidae). These hosts are systematically close to the New Zealand acanthoclinids which are among the hosts of T. parabranchicola n.sp. Unfortunately, Fantham's description is a very brief one. His measurements for the diameter of the body of T. clini (37μ) and for that of the striated band (20μ) correspond with the equivalent measurements for the smallest stages of the New Zealand species. The fact that Fantham described the macronucleus of his species as bean–shaped suggests that he was dealing with immature trichodinids in which the macronucleus had not yet adopted the horseshoe shape characteristic of maturity. Like T. baltica, T. scorpaenae and T. labrorum, T. clini will have to be redescribed in greater detail before trichodinids from other hosts and localities can be confidently identified with it.

T. branchicola, which Tripathi (1948) described from various English marine fishes, including sculpins, blennies, rocklings and sticklebacks from the intertidal zone, is one of the few really adequately described species of the genus. As in T. parabranchicola n.sp., the shape of this species is extremely variable. Its

diameter (30μ to 53μ) and height (22μ to 36μ.) are close to those of the New Zealand species, as is the length of the aboral cilia (15μ to 20μ). As in the latter species the characteristic number of denticles is 22. The skeletal complex of T. branchicola is, however, both actually and relatively smaller than that of the species under discussion. The diameter of the denticulate ring of the former species does not exceed 19μ, while that of the striated band does not exceed 33μ. Tripathi does not record any examples of his species having less than 20 denticles, and the denticular hooks of T. branchicola, being only 2·3μ, in length, are very much smaller than are those of mature examples of T. parabranchicola n.sp. Finally, the micronucleus of the former ciliate, unlike that of the latter, is located in a pocket on the outer edge of the macronucleus.

The species under discussion, then, differs in detail from other members of its subgenus as described herein, although it is apparently close to three inadequately described species, T. scorpaenae Robin, T. labrorum Chatton and T. clini Fantham. In so far as can be gathered from the literature its closest relative is undoubtedly T. branchicola Tripathi, and it is accordingly designated Trichodina (Trichodina) parabranchicola n.sp.

A slide designated as the type of T. parabranchicola n.sp. has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z21), and paratypes are in my own collection and in that of the Department of Zoology, Victoria University College.

Trichodina (Trichodina) multidentis n.sp. (Pl. 12, figs. 88, 89, 92; Pl. 13, fig. 101).

All but one example of Ericentrus rubrus, all but one of Tripterygion varium, all but three of T. medium, and the solitary example of Notoclinus fenestratus were infested with this ciliate, which was frequently associated with, but always dominant over, T. parabranchicola n.sp. T. multidentis n.sp. was frequently subordinate to the latter species in mixed infections on the gills of Diplocrepis puniceus and kelpfishes living in association with blennies. This species was never collected from Oliverichtus melobesia. It, too, is always present in greatest abundance on the branchiae of its hosts, and infests fishes of all ages.

Morphology.

Living Material. This species is appreciably larger than T. parabranchicola n.sp., which it otherwise resembles as regards shape and feeding and locomotor activity.

stained Material. The macronucleus, as seen in haematoxylin preparations, resembles that of T. parabranchicola n.sp. in shape but is of greater size, ranging up to 121·2μ in length and 8·5μ in breadth, with an average of 96·8μ by 7·1μ (50 mature examples measured). The spherical micronueleus is usually located between the incurved extremities of the macronucleus.

The three concentric rings of the skeletal complex attain greater diameter than do those of T. parabranchicola n.sp. The units of the denticulate ring exceed those of the latter species in number (although there is a slight overlap at the lower extremity of the size range—see Text–fig 1) and markedly differ from them in shape. Mature examples characteristically have from 30 to 37 denticles, the mean number for the species being 32. The range in denticular number is very considerable (23–45), as is shown by the counts from a random sample of 150 examples of all ages in Table IX (see also Text-fig. 1).

Instead of being slender and sharply pointed as are the denticular rays of T. parabranchicola n.sp., those of mature examples of T. multidentis n.sp. are thick and either bluntly pointed or rounded proximally (Figs. 88, 89). As Fig. 89 shows, there may be considerable variation in the shape of the rays of a single individual—this is particularly so in the larger and older examples. The denticular hooks are never sickle–shaped as in T. parabranchicola n.sp., and in the older examples are quite upright (Fig. 89). The hooks of normal mature examples are rather longer than those of the latter species, averaging 4.5μ in height and on occasion attaining as much as 6μ. Once again, the length of the denticular rays approximates that of the hooks. Frequently, and more especially in the smaller examples of this trichodinid, the skeletal complex is not flattened out in fixed and stained material. When this is the case the hooks are viewed obliquely from the surface aspect, and then appear to be slender, sharply pointed, and of triangular shape. The rays, being more or less vertically oriented in surface view, appear to be very short or even lacking. This condition very closely approximates that described as characteristic of T. myakkae Mueller, by Mueller (1937) and Davis (1947).

The diameter of the denticulate ring ranges from 10·8μ to 45.4μ the average for 50 mature individuals being 34.8μ. This average is some 70 per cent. higher again than that for the ring of T. parabranchicola n.sp. and exceeds the maximum diameter attained in the latter species.

The outer diameter of the striated band ranges from 25.3μ to 67.5μ, with an average for 50 mature examples of 53.2μ. This average is some 65 per cent. higher again than that for the band of T. parabranchicola n.sp. As in the latter species there are 7 to 9 striae per denticle, the smallest division stages having 7 striae. The number increases with age, the largest and oldest examples almost always having 9 striae per denticle.

The border membrane is some 2μ in width, and the aboral cilia range from 11μ to 26μ in length.

Growth and Reproduction.

Binary fission follows the same course as in T. parabranchicola n.sp., the most outstanding point of difference (apart from the denticular number) being that in the present species the macronucleus re–assumes the characteristic horseshoe shape at a much earlier stage, before the denticles derived from the parent have been absorbed. The denticles themselves undergo a more rapid initial development, the hooks of post-division individuals in which the parent denticles have been absorbed never being less than 3μ in height, those of T. parabranchicola n.sp. at the equivalent stage of development being not much over 1μ in height (Text-Fig. 2).

The following combinations of denticles have been noted in young daughter individuals in which the old ring has not yet been absorbed, the first figure

representing the number of denticles in the new ring and the bracketed one being that in the inner ring of denticles derived from the parent:—

Binary fission in this species is most common among individuals towards the lower limit of the size range, having 28 denticles and a ring of from 15μ to 25μ in diameter. The diameter of the denticulate ring increases relatively rapidly up to the 32-denticulate stage, the increment thereafter being more gradual (Text–fig. 2).

Conjugation has been observed, but has not been studied in detail.

Examples of from small to average size have been found on all the host species, and stages in binary fission have been collected from all the blennies and from Acanthoclinus quadridactylus. Individuals having up to 35 denticles have been found on all the hosts. One with 37 denticles was collected from Diplocrepis puniceus, and several with 39 denticles from Tripterygion medium. T. varium was the only host from which trichodinids having 40 or more denticles were secured.

Biology.

The feeding habits of T. multidentis n.sp. resemble those of T. parabranchicola n.sp. Individuals kept in hanging drop preparations of gill scrapings rapidly become engorged with erythrocytes. None of those collected in the field and immediately fixed on cover slips, however, have been observed to contain ingested blood cells. The only food inclusions identified with certainty in such trichodinids have been diatoms of the sub–order Monoraphidineae. All the diatoms seen have had elliptical and dissimilar valves, the upper valve having transverse, punctate striae. They are identified from the keys of Boyer (1927) and Hendey (1937) as belonging to the genus Cocconeis. The members of this genus are epiphytic, the frustules being attached to the sub–stratum by the lower valve. In all the ingested examples seen in T. multidentis the upper valve alone is intact (Fig. 92). This Cocconeis ranges in length from 14.0μ to 20.3μ, and in breadth from 8.0μ to 10·9μ. Its average number of punctate striae per 10μ of valve-length is 9. Large examples of T. multidentis containing up to 5 ingested Cocconeis frustules have been seen.

The Cocconeis concerned is epiphytic on the coralline algae among which blennies browse (and among which kelp fishes feed at night-time). Detached from the host plants by the movements of the fishes, the frustules, or at least their upper valves, pass via the mouth through the gills where they are seized upon by the ectocommensal trichodinids.

Systematic Position.

The 15 species of Trichodina (Trichodina) listed hereunder bear comparison with T. multidentis n.sp. (those having marine hosts are marked with an asterisk, and the bracketed numbers are the denticular numbers for each species).

T. urinicola, T. okajimae, T. vesicularum and T. ranae have host relationships quite different from those of the species under discussion, being endoparasitic in the urinary bladder of amphibians. All four of these species are of much smaller average size than T. multidentis n.sp.

T. truttae is the largest known trichodinid, the diameter of the denticulate ring attaining 85μ, and that of the striated band 125μ, in full grown individuals (Davis, 1947). Apart from its larger size and the smaller range of its denticular number, this species differs from T. multidentis n.sp. in having 20 striae to each denticle.

Davis's species are all from North American fresh water fishes. Although up to 30 denticles occur in T. discoidea, the usual number is 20 to 25. Further points of difference between this species and T. multidentis n.sp. lie in the form of the hooks and rays. The hooks of T. discoidea appear to be truncated distally, and the rays, unlike those of T. multidentis n.sp. are long, slender, and sharply pointed. T. platyformis is close in size to the latter species, from which it differs in having 10 striae to each denticle and usually only 28 to 31 denticles. The denticular hook of T. platyformis is of similar shape to that of the New Zealand species, but the denticular ray of the former trichodinid is very distinct, being 10μ in length, slender, and sharply pointed. T. fultoni is much larger than T. multidentis n.sp., the diameter of the denticulate ring ranging from 50μ to 58μ, and that of the striated band from 75μ to 90μ, in mature individuals. The usual number of denticles in T. fultoni is only 27 to 29, and the hooks are of the strongly recurved type typical of T. parabranchicola n.sp. T. californica is of smaller size than T. multidentis n.sp., which it resembles in being more plastic than is usual in trichodinids, but from which it differs in usually having 26 to 28 denticles, the hooks of which are strongly recurved and the rays straight and pointed.

T. spheroidesi and T. halli were described by Padnos and Nigrelli (1942) from a marine fish, Spheroides maculatus (Bloch and Schneider), from the coasts of New York and New Jersey. Both these species, like T. parabranchicola and T. multidentis n.sp., were most abundant on the gills of their hosts. T. halli is a much smaller species than T. multidentis n.sp., the maximum diameter of its denticulate ring and striated band not much exceeding the minimum diameter of these structures in the latter species. The maximum number of denticles in

T. halli is only 31; and the minimum number, 21, is below the minimum for T. multidentis n.sp. From Padnos and Nigrelli's Text–Fig. 1, it is apparent that the denticular hooks of T. halli are strongly recurved, and the denticular rays in this species are slender and pointed and thus quite unlike those of the New Zealand one. T. halli is a larger species than T. multidentis n.sp., the diameter of its denticulate ring ranging from 30μ to 54μ and that of its striated band from 41μ to 81μ. Despite the greater diameter of its denticulate ring T. halli exhibits less variation as regards the denticular number, the maximum number of denticles formed, 34, only reaching the mean number and falling far short of the maximum in T. multidentis n.sp.

It appears that Fantham (1930) was, although unaware of the fact at the time, dealing mainly with immature trichodinids when he described T. blennii, T. mugilis and T. chelidonichthyos from South African marine fishes. Thus he stated that T. blennii has two forms of macronucleus, the one being rounded or ovoid and the other horseshoe shaped, and suggested that this might indicate sexual dimorphism. T. mugilis was described as having a macronucleus varying in form from round through dumb–bell shape to horseshoe shape; and T. chelidonichthyos as having only the oval type of macronueleus. No indication of a range in denticular number was published for the latter two species, T. mugilis being stated to have 32 denticles and T. chelidonichthyos 30. Examples of T. multidentis n.sp. having from 30 to 32 denticles are of approximately equal size to these South African trichodinids. The diameter of the striated band (Fantham's “ciliated disc”) ranges from 23μ to 28μ in T. mugilis and from 19μ to 32μ in T. chelidonichthyos; and relatively few examples of T. multidentis n.sp. having a striated band of up to and including 32μ in diameter have more than 32 denticles. Examples of the latter species of this order of size, being within the range in which binary fission takes place, may have the rounded or ovoidal type of macronucleus described by Fantham for his South African species. T. blennii has from 24 to 32 denticles, and falls within the size range of T. multidentis n.sp. The former species was described as having only five striae per denticle, but this, too, might be attributable to the examination of immature forms in which the new striae had not yet made their appearance.

T. tenuidens, described by Fauré–Fremiet (1943) from Gasterosteus aculeatus L., a European stickleback equally at home in fresh or brackish water or in the sea, appears to be close to T. multidentis n.sp. Like the latter species, T. tenuidens occurs in association with another trichodinid—T. (Cyclochaeta) domerguei (Wallengren)—from which it is readily distinguishable biometrically. T. tenuidens has from 27 to 37 denticles, in this respect resembling T. multidentis n.sp., but as Fauré–Fremiet omitted to give details of the diameter of the body or of the various parts of the skeletal complex it is unfortunately impossible to make a closer comparison of the two species.

The present species, in so far as can be gathered from the literature, differs in detail from other members of Trichodina (Trichodina) as described above. Further studies of Fantham's trichodinids from South African marine fishes, and of T. tenuidens Fauré–Fremiet, may reveal closer affinities between these species and the New Zealand one than can be drawn at present. There are only three records in the literature of trichodinids having 40 or more denticles. Wallengren (1897) gave the number of denticles of his T. (Cyclochaeta) domerguei as 19 to 25, with one atypical example having 49. More recent studies on this