The Protozoa of New Zealand Intertidal Zone Fishes

Transactions and Proceedings of the Royal Society of New Zealand, Volume 81, 1953, Page 79

The Protozoa of New Zealand Intertidal Zone Fishes Marshall Laird, M.Sc., Ph.D. Department of Zoology, Victoria University College, Wellington [Read before a meeting of the Wellington Branch, July 23, 1952; received by the Editor, August 1, 1952.] Abstract Fourteen species of Protozoa are recorded from 458 New Zealand intertidal zone fishes of 10 species. Haemogregarina bigcmina Laveran and Mesnil (Sporozoa: Coccidia) is reported from 5 new hosts, Oliverichtus melobesia (Phillipps) (Gobiesocidae), Ericentrus rubrus (Hutton), Tripterygion varium (Forster), T. medium Günther and Notoclinus fenestratus (Forster) (Blennidae). Myxidium incurvatum Thélohan (Sporozoa: Myxosporidia) is recorded from 4 new hosts, the clingfishes Diplocrepis puniceus (Richardson) and Oliverichtus melobesia, the blenny Notoclinus fenestratus, and Acanthoclinus quadridactylus (Forster) (Acanthoclinidae). Ericentrus rubrus is listed as a new host for Trypanosoma tripterygium Laird (Mastigophora: Protomonadina). Two new genera are established. The generic designation Davisia is proposed for certain Myxosporidia previously included in Ceratomyxa Thélohan and Sinuolinea Davis, with Davisia (= Ceratomyxa) spinosa (Davis, 1917) as the genotype. A unique peritrichous ciliate belonging to Kahl's family Scyphidiidae is described as the type of a new genus. Caliperia. An undetermined species of the myxosporidian genus Zschokkella is recorded from the urinary bladder of Tripterygion varium, and the following new species are described:— Haemogregarina (Hepatozoon?) acanthoclini (Sporozoa. Coccidia)—blood of Acanthoclinus quadridactylus. Leptotheca subelegans (Sporozoa: Myxosporidia)—gall bladder of Callogobius atratus (Gobiidae) and Diplocrepis puniceus. Davisia diplocrepis (Sporozoa: Myxosporidia)—urinary bladder of Diplocrepis puniceus. Sphaeromyxa tripterygii (Sporozoa: Myxosporidia)—gall bladder of Tripterygion varium and T. medium Myxosoma tripterygii (Sporozoa: Myxosporidia)—subdermal connective tissue of Tripterygion varium. Scyphidia (Gerda) acanthoclini (Ciliata: Peritricha)—gills of Acanthoclinus quadridactylus. Caliperia longipes (Ciliata: Peritricha)—gills of Oliverichtus melobesia and Ericentrus rubrus. Trichodina (Trichodina) parabranchicola (Ciliata: Peritricha)—gills of Diplocrepis puniceus, Oliverichtus melobesia, Acanthoclinus quadridactylus, Acanthoclinus trilineatus Griffin, Ericentrus rubrus, Tripterygion varium, T. medium and Notoclinus fenestratus (the dominant trichodinid on the first four hosts). Trichodina (Trichodina) multidentis (Ciliata: Peritricha)—not recorded from O. melobesia, but otherwise host list as for T. parabranchicola, over which it is dominant on the last four hosts. Trichodina (Trichodina) multidentis n.sp. is a new host for the endoparasitic suctorian Endosphaera englemanni Entz. The affinities of these protozoans are discussed, with particular regard to species already recorded from intertidal zone fishes in other parts of the world. Consideration is given to the problem of host specificity in relation to host habitat. The evidence suggests that the environmental niche occupied by fishes is a most important factor in determining the composition of their protozoan fauna.

Introduction Great numbers of parasitic and commensal protozoans have been described from marine and freshwater fishes in many parts of the world. Those of freshwater fishes include species which have been responsible for epizootics in streams and hatcheries. The prospect of gaining knowledge leading to the control of these pathogens has not only provided a stimulus to research, but has frequently induced organizations concerned with conservation for game or commercial purposes to make funds and facilities available to investigators. The nature of the marine habitat raises as yet unsurmounted difficulties in the control of pathogenic protozoans of fishes occupying it. Marine fisheries authorities have tended to concentrate their conservation research on studies designed to result in the improvement of fishing regulations. In consequence, the ecology of the Protozoa of freshwater fishes has become more completely understood than has that of the marine fish protozoans. A wealth of information regarding the Protozoa of marine fishes has nevertheless been accumulated. The majority of this, having been gathered by systematists working on their own initiative, is scattered throughout the literature in unrelated fashion. Many of the protozoans concerned have been most inadequately described, new specific names having been accorded them merely on the basis of their occurrence in new hosts or new localities. Numerous anomalies have arisen from this practice. Unfortunately, the incompleteness of the recorded information is such that it is as yet seldom possible to do more than suspect and regret the existence of synonymy. It became manifest during recent studies of the haematozoa of New Zealand fishes (Laird, 1951a, 1952) that in only few instances are published descriptions sufficiently detailed to allow of a species being confidently described either as new or as conspecific with an established species. Furthermore, much confusion is caused by our lack of knowledge concerning host specificity and host-induced morphological variation. These problems are complicated both by the plastic nature of the material and by the fact of our all but total ignorance of the transmission of marine fish haematozoa. Obviously, any attempt to monograph either the trypanosomes or the haemogregarines of marine fishes, based on the existing literature, could result in little more than the compilation of long lists of specific names embodying much suspected but undemonstrable synonymy. Equally obviously, a study of the zoogeographical distribution and host specificity of these parasites could be expected to yield highly significant results. It appeared worthwhile to conduct an introductory investigation, the scope of the project being limited in some way, and the greatest possible use being made of the literature in such a manner as to enable comparisons of zoogeographical significance to be drawn. The host preferences of haematozoans of fishes often bear little relation to the systematic positions of the hosts themselves. In the family Blenniidae, for example, the European catfish Anarhichas lupus is parasitized by Haemogregarina anarhichadis Henry, 1912, a species having close affinities with haemogregarines of certain non-blenniid pelagic fishes (Laird, 1952) but systematically remote from Haemogregarina bigemina Laveran and Mesnil, 1901, a widespread parasite of intertidal zone blennies. Again, there are species of Trypanosoma (e.g., T. blenniclini Fantham, 1930) and Haemogregarina (e.g., H. bigemina) having hosts in two or more piscine families.

In view of these facts, it was decided to make the ecology, rather than the systematic position of the host, the restrictive factor, and to limit the field of investigation to a particular biotope instead of to a particular group of hosts. The intertidal zone was selected as a suitable biotope. Here again, though, it proved that literature concerning the haematozoa of intertidal zone fishes from other parts of the world was scanty, very incomplete in coverage, and often inadequate in regard to morphological data. From time to time, however, numerous workers studying fishes of the intertidal zone in various parts of the world had dealt with protozoans of other groups besides those inhabiting blood. The pooling of the relevant papers dealing with all classes of the phylum Protozoa led to the accumulation of a sufficiently comprehensive literature to enable instructive comparisons to be made on a world-wide basis. Thus trypanosomes had been recorded from shore fishes of France by Brumpt and Lebailly (1904), of England by Henry (1910), of Italy by Neumann (1909), of South Africa by Fantham (1919, 1930), and of New Zealand by Laird (1951a). Haemogregarines of such fishes had been dealt with by Laveran and Mesnil (1901) and Brumpt and Lebailly (1904) in France, by Henry (1910, 1913) and Bentham (1917) in England, by Neumann (1908, 1909) and Kohl-Yakimoff and Yakimoff (1915) in Italy, by Fantham (1930) in South Africa, by Fantham et al. (1942) in Canada, and by Laird (1951) in Fiji and (1952) in New Zealand. A wide range of Myxosporidia had been described from intertidal zone fishes, notably by Thélohan (1892, 1895) and Georgévitch (1916, 1916a, 1917) in France and Monaco, by Awerinzew (1908-11) in Russia, by Auerbach (1909-12) in Norway, by Dunkerly (1920) and Tripathi (1948a) in England, by Doflein (1898) and Parisi (1912) in Italy, by Fantham (1919, 1930) in South Africa, by Mavor (1916) in Canada, by Ellis (1930) in Nova Scotia, and by Jameson (1929, 1931) and Noble (1938, 1939, 1941) in California. As regards ciliates, Scyphidia had been reported from European shore fishes by Fauré-Fremiet (1905) and Precht (1935); and Trichodina had been recorded from intertidal zone fishes of France by Robin (1879) and Laveran (quoted by Neumann, 1909), of Germany by Precht (1935), of England by Tripathi (1948) and of South Africa by Fantham (1930). Hyperparasitic amoebae and suctorians had also been described from marine fish trichodinids (Chatton, 1910: Padnos, 1939). Of all these authors, only Noble, in his investigations of Californian Myxosporidia, had dealt exclusively with the intertidal habitat. The others had merely described chance material from this habitat in the course of wider studies. With the exception of Fantham in South Africa, none of them had described representatives of all four of the protozoan orders commonly infesting fishes—Protomonadina, Coccidia, Myxosporidia and Peritricha—from hosts of any one particular locality. Accordingly, it was decided to analyze this literature, with the object of ascertaining such information on inter-relationships, host specificity and zoogeographical distribution as could be determined from the available data; while at the same time augmenting the data by surveying the entire protozoan faunas, hitherto quite unknown, of New Zealand intertidal zone fishes. Material The study was confined to the North Island, collections being made from tide pools, sandstone reefs and stony beaches. Most of the work was done at Welling-

ton during 1949-51, in all months of the year. Material was also gathered in the Bay of Islands, Northland (February, 1951), at Auckland (January, 1951) and at Tolaga Bay, East Coast (February, 1951). A total of 458 fishes belonging to 10 species were collected, as detailed in Table I. The identifications of these were based on descriptions published by the following authors:—Waite, 1913 (Blenniidae), Phillipps, 1927 (Oliverichtus melobesia), Hutton (1872) (Diplocrepis puniceus), and Griffin, 1933 (Callogobius atratus and Acanthoclinidae). As is evident from Table I, the dominant species in all four localities are the two blennies, Tripterygion varium and T. medium, and the kelpfish Acanthoclinus quadridactylus. All stages of the two former species occur in tide pools (particularly those with abundant coralline algae), in which they move about freely. Small to medium-sized examples of the latter species are also found in tide pools, but are of much more cryptic habit than are the blennies, remaining under cover beneath stones during the daytime. Examples of A. quadridactylus exceeding 9 cm. in length are seldom found in any but the largest rock pools briefly isolated when the tide is at its lowest. These larger stages usually remain in the kelp belt, although at Point Jerningham, within Wellington Harbour, they abound beneath loose rocks and concrete blocks at the low tide mark. It has been suggested (Oliver, 1923) that the swellings at the tips of the median rays of the dorsal and anal fins of Acanthoclinus are sensory adaptations to life in the narrow crevices beneath stones. On those parts of the coast where rocky headlands are absent but sandstone reefs occur, the habits of the blennies change accordingly. Thus at St. Heliers, Auckland, T. varium and A. quadridactylus occur together at low tide in the small pockets of water beneath loose sandstone blocks on the reef. Table I. Summary of Collection Data. Systematic Position Number Examined for Protozoa Bay of Islands Auckland Tolaga Bay Wellington Total Gobioidea Gobiidae (Gobies) Callogobius atratus Griffin, 1933 0 0 0 1 1 Xenopterygii Gobiesocidae (Clingfishes) Diplocrepis puniceus (Richardson, 1846) 0 0 0 96 96 Oliverichtus melobesia (Phillipps, 1927) 1 0 0 29 30 Blennioidea Blenniidae (Blennies) Ericentrus rubrus (Hutton, 1872) 0 0 0 20 20 Tripterygion tripenne (Forster, 1801) 0 0 0 1 1 Tripterygion varium (Forster, 1801) 5 6 9 90 110 Tripterygion medium (Günther, 1861) 36 2 6 75 119 Notoclinus fenestratus (Forster, 1801) 0 0 0 1 1 Acanthoclinidae (Kelpfishies) Acanthoclinidae quadridactylus (Forster, 1801) 5 4 11 56 76 Acanthoclinus trilineatus Griffin, 1933 0 0 0 4 4 Grand Totals 47 12 26 373 458 Species often locally abundant, but not as commonly encountered as the preceding ones, are the blenny, Ericentrus rubrus, and the clingfishes, Diplocrepis

puniceus and Oliverichtus melobesia. E. rubrus is found in the masses of brown algae in and about the last rock pools to be exposed at low tide, and was only collected on the headlands of the ocean beaches just south of Wellington. D. puniceus was also taken in the vicinity of Wellington only. This species occurs beneath stones in the surf belt (Oliver, 1923), and frequently attaches itself beneath smooth, rounded boulders in tide pools. It is particularly abundant, in association with A. quadridactylus, beneath the rocks and concrete blocks exposed at the low tide mark at Point Jerningham. O. melobesia was collected at Long Beach, Russell (Bay of Islands), and at Wellington. This species is found clinging to the lower surfaces of smoothy rounded stones, frequently ones of very small size, between tide marks. It is locally restricted in its distribution, but abounds at Princess Bay and Fisherman's Creek, Island Bay, on the coast to the south of Wellington. Acanthoclinus trilineatus, which is much less frequently encountered than is A. quadridactylus, was collected only at Island Bay, Wellington, where it inhabits the deeper rock pools and the kelp belt. Tripterygion tripenne and Notoclinus fenestratus rarely enter rock pools, although the latter blenny is said to be quite common in the masses of brown algae on the seaward faces of rocky headlands. Callogobius atratus is one of the few New Zealand gobies which dwells in the sea. It is rarely encountered, the single example found during this survey being collected under a rock exposed during a spring tide near Fisherman's Creek, Island Bay. Methods Fish were collected by hand, by netting, and by poisoning tide pools with rotenone (5 per cent.). The time elapsing between the distribution of the poison and the immobilization of the fish varies from locality to locality. It exhibits correlations with latitude, the determining factor here probably being water temperature, and with the degree of activity of the fish themselves. The rotenone used was part of the same batch employed during visits to Fiji and Campbell Island in 1949. In coral pools near Suva (18° S), fish were dying or throwing themselves out onto the reef surface four or five minutes after the water was poisoned. At Wellington (41° S) from 20 to 25 minutes, and at Campbell Island (52° S) from 40 to 45 minutes passed before the poison took effect. In all localities the more cryptic species, such as the clingfishes, were the last to be affected. In the case of those collected near Wellington, fishes were brought alive to the laboratory wherever this was possible so that their Protozoa could be studied in the living condition. On field trips, preparations from the gills, heart, urinary bladder and gall bladder were made immediately following collection. Thin smears of heart blood were made on 3″ by 1″ microscopic slides, air-dried, and subsequently fixed in absolute methyl alcohol and stained with Giemsa. All preparations from the gills and internal organs took the form of never-dried smears on 7/8″ square cover slips (No. 1). These were fixed in Schaudinn's fluid, Bouin's solution or Worcester's formol-mercuric-acetic mixture. Excellent results were obtained from the use of Davis's (1947) modification of Worcester's mixture (75cc saturated solution mercuric chloride, 20cc formaldehyde 40 per cent, 5cc glacial acetic acid), and except where otherwise stated all measurements of haematoxylin-stained protozoans given herein are derived from material fixed in this fluid. Heidenhain's iron haematoxylin and haematoxylin prepared by Shortt's (1923) rapid method were used for staining wet-fixed preparations. Gill

smears were counterstained with eosin or Bordeaux red to bring out the cilia of Peritricha. Destaining was effected with iron alum (0.5 per cent.) or a saturated solution of picric acid in distilled water, overnight treatment with the latter reagent giving particularly good results with ciliates. Either Canada balsam or Stafford Allen and Sons' neutral “Sira” medium were used as mountants. During the project, a need was felt for a technique by means of which wet-fixed material could be stained and mounted in the field—for smears on cover slips stored in liquid preservatives inevitably become more or less damaged by rubbing against one another while in course of transportation. This was found to be particularly so with gill smears, preparations from fishes heavily infested with trichodinids often having few if any of these ciliates remaining intact on return to the laboratory at the conclusion of a field trip. It was found possible to deal with a substantial amount of such material in the field, without undue loss of collecting time, by means of the following technique. Sets of 4″ by 1″ shell vials filled with iodized water, iron alum mordant, Shortt's haematoxylin, picric destaining reagent, alcoholic eosin and the various grades of alcohol and xylol required for dehydrating and clearing, are made up in the laboratory and stored in order in portable racks. Each of these vials will accommodate four 7/8″ square cover slips in a vertical series at right angles to one another. Staining may be carried out at the field base at the end of the day's collecting activities, the cover slip smears (already fixed in Worcester's mixture as modified by Davis) being immersed in the iodized water, mordant and stain for half an hour in each case, and allowed to remain in the picric acid overnight. The latter reagent is slow in action, and excessive destaining of trichodinids does not readily take place in it. Excellent results are obtained with these ciliates when so treated for from 8 to 10 hours. With practice the smears may be rapidly counterstained, dehydrated and mounted on the following morning without the necessity for microscopic checking. A further batch of smears may be mordanted overnight, stained on the following morning while the first batch are being mounted, destained during the day and mounted in the evening. Six staining sets made up as above and providing for the daily simultaneous preparation of two batches of 24 finished protozoan mounts from as many individual fishes, may be accommodated in a light plywood case measuring only 15″ by 8″ by 5″. Internal organs, to be examined later for Myxosporidia, may be dissected out in the field and preserved in bulk in formalin. In the case of small fishes, the entire contents of the gall- and urinary bladders may be mixed with polyvinyl-acetic-alcohol (P.A.A. medium) on a microscopic slide, and covered with a cover slip. This method is a particularly good one for the preservation of fully extended trophozoites, and lends itself to use in the field. It was unfortunately not tried out until recently, and was thus not employed during this investigation, but is mentioned here for the information of future workers. When dealing with fresh preparations in the laboratory, methyl green, neutral red and Lugol's iodine were used as intravitam stains. Bond's (1938) method of using 5 per cent. phenol proved most satisfactory for bringing about the extrusion of the polar filaments of myxosporidian spores. A dark ground condenser often proved of use, particularly in the study of peritrichous ciliates. All the figures were drawn with the aid of an Abbé camera lucida.

The Protozoa Observed Of the 458 fishes examined, only one, the solitary example of Tripterygion tripenne collected, was completely free from Protozoa. All the remainder were variously infested with representatives of three classes, four orders and 12 genera of this phylum as outlined in Table II. The incidence of Trichodina was much higher than that of any of the other protozoans, 448 (97.8 per cent.) of the 458 fishes studied being infested. Representatives of this genus were almost always confined to the branchiae of the hosts. They were sometimes found on the ventral sucking disc of Diplocrepis and Oliverichtus, and very rarely small numbers were observed moving about over the body surfaces of the hosts. Only two fishes, Callogobius atratus and Tripterygion tripenne, were recorded as negative for trichodinids, but as in each case only one fish was available for study the record is of little significance. Table II. Synonis of protozoan records Host Number, Type and Location of Infections Species No. Protomonadina Coccidia Myxosporidia Peritrioha Blood Blood Gall bladder Urinary bladder Tissues Gills Trypenosoma Haemogregarina Haemogregarina (Hepatozoon?) Letotheca Hyxidium Sphaeromyxa Davisia n. gen. Zschokkella Nyxosoma Scyphidia (Gerda) Caliperia n. gen. Trichodina (Trichodina) C. atratus 1 0 0 0 1 0 0 0 0 0 0 0 0 0 D. puniceus 96 0 0 0 5 3 0 63 0 0 0 0 94 O. melobesia 30 0 5 0 0 15 0 0 0 0 29 29 E. rubrus 20 1 16 0 0 0 0 0 0 0 0 0 0 3 19 T. tripenne 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T. verium 110 5 27 0 0 0 2 0 1 1 0 1 0 0 29 T. medium 119 3 7 0 0 0 3 0 0 0 0 0 116 N. fenestratus 1 0 0 1 1 0 0 0 0 0 0 0 1 A. qadridatylus 76 0 0 6 0 3 0 0 0 27* The actual incidence of Scyphidia is higher than this figure suggests. See text, page 114. 0 76 A. trilineatus 4 0 0 0 0 0 0 0 0 0 0 4 Grand Totals 458 9 56 6 6 22 5 63 1 1 27 32 448 Adding in the records for Scyphidia and Caliperia n.gen., the number of infestations with gill ciliates reaches 504. The number of fishes concerned remains unchanged, for all the examples of Acanthoclinus quadridactylus infested with Scyphidia and those of Ericentrus rubrus infested with Caliperia were also hosts for trichodinids; while one Oliverichtus melobesia positive for Caliperia was free from Trichodina, and another positive for Trichodina was negative for Caliperia. All the gill ciliates are new, there being two species of Trichodina and one each of Scyphidia and Caliperia. Taking double and multiple infections into consideration, 90 (19.7 per cent.) of the fishes, of seven species, were parasitized by coelozoic Myxosporidia representing five genera. This unusually high percentage of infection is largely due to

the high incidence of Davisia n.gen. in the urinary bladder of Diplocrepis puniceus. The most widespread myxosporidian was the cosmopolitan Myxidium incurvatum Thélohan, this species being recorded from 22 fishes belonging to four species. A new species each of Leptotheca and Sphaeromyxa, these having two hosts, and a single sporulating trophozoite of an undetermined species of Zschokkella, were also discovered. Haematozoa were found in 63 (13.8 per cent.) of the fishes, six of the 10 species studied acting as hosts. Haemogregarina bigemina Laveran and Mesnil accounted for the bulk of these records, infecting 56 fishes of five species. Eight of these fishes were also positive for Trypanosoma tripterygium Laird, one of the three species concerned, Ericentrus rubrus, being a new host for this flagellate. A single example of Tripterygion medium was parasitized by the trypanosome only. Six of the 11 examples of Acanthoclinus quadridactylus collected at Tolaga Bay were lightly infected with a new species of haemogregarine, possibly referable to Hepatozoon Miller, a genus occurring in reptiles, birds and mammals, but not hitherto known from fishes. A new species of Myxosoma, found in the subdermal connective tissue of a single example of Tripterygion varium collected at Wellington, was the only histozoic myxosporidian noted during the investigation. There was no outward sign of the presence of this parasite, which thus could not be conveniently searched for during field trips. Future investigations might reveal the presence of more histozoic sporozoans in the fishes dealt with than did the present studies. Although fresh rectal contents and stained gut smears were examined throughout the survey, intestinal Protozoa were never found. Previous investigators have not described such protozoans from intertidal zone fishes. Some trematodes were discovered, but these will be dealt with elsewhere. Double infections with Myxosporidia were never recorded from the same organ of any one particular fish. As regards haematozoans, Trypanosoma was in all but one instance associated with Haemogregarina. In the case of the gill protozoans, Scyphidia was only found in association with Trichodina. Caliperia n.gen., with the exception of a single pure infection from Oliverichtus melobesia, was similarly always associated with Trichodina. Double infections with two species of the latter genus were frequently encountered. All the fishes doubly parasitized by Trypanosoma and Haemogregarina were infested with Trichodina as well, and many of those having two different ciliates on the gills were also infected by a blood parasite or a myxosporidian. One example of Oliverichtus melobesia from Wellington had Haemogregarina in its blood, Myxidium in its gall bladder and Caliperia and one species of Trichodina on its gills. Obvious clinical symptoms of disease were never manifest in any of the fishes studied. Such pathogenic effects as became apparent during dissection and microscopical examination are detailed separately for each species in the following systematic account. Systematic Account Class Mastigophora Diesing Order Protomonadina Blochmann Trypanosoma tripterygium Laird, 1951. This flagellate was described from one of seven examples of Tripterygion varium and one of five of T. medium collected at Island Bay, Wellington, on

July 30, 1949. Both infections were light ones, thin smears of the heart blood containing from four to seven trypanosomes each. Ericentrus rubrus (Hutton) is now given as an additional host of T. tripterygium, a 25 mm. example collected at Island Bay on March 15, 1951, being lightly infected. Four more examples of T. varium and two more of T. medium were found to be parasitized by this flagellate. All the infected fishes were collected on the sea coast to the south of Wellington, from Island Bay to Moa Point, Lyall Bay, during August, 1949, and December, 1950. The parasite rate in a 64 mm. T. medium from Moa Point (December 25, 1950) was appreciably higher than usual, a single thin smear of heart blood containing 59 trypanosomes (av. 3 per 10,000 erythrocytes). This trypanosome having only recently been described and figured, its morphology will not be fully detailed here. It shows marked polymorphism, small slender forms, intermediate forms and stout, degenerating forms occurring. The average dimensions of the intermediate form, which is the commonest one, are given below, as published in the original description. Length of free flagellum 14.0μ Length of body 79.1μ Total length 93.1μ Width of body at centre of nucleus 5.3μ Width of undulating membrane 0.8μ Diameter of nucleus 4.5μ Diameter of karyosome 3.8μ The extremes of the various dimensions for all forms of T. tripterygium are:— Length of free flagellum 3.3μ–16.8μ Length of body 36.3μ–85.1μ Overall length 50.9μ–100.1μ Width of body at centre of nucleus 1.3μ–7.4μ Width of undulating membrane 0.5μ–1.8μ Length of oval type of nucleus 2.4μ Width of oval type of nucleus 1.1μ–1.7μ Diameter of round type of nucleus 2.4μ–4.9μ Diameter of karyosome 1.7μ- 4.3μ Small slender forms of T. tripterygium have an oval nucleus situated some 28 per cent. of the total body length from the anterior extremity and occupying the entire width of the body. As the flagellate increases in size its nucleus becomes rounded, distinctly karyosomatic and more posteriorly situated. In the common intermediate form the nucleus is from 40–50 per cent. of the total body length from the anterior extremity. Although the actual distance between the kinetoplast and the posterior extremity of the body is approximately the same as in the smaller forms, this distance expressed as a percentage of the total body length is considerably reduced (from 17–24 per cent. to 9 per cent.). Similarly, the actual length of the free flagellum is substantially unchanged although it is relatively much shorter in proportion to the length of the body. Stout, degenerating trypanosomes may attain 7.4μ in breadth. These have a very short free flagellum, in consequence of a marked anterior extension of the cytoplasm along this organelle. This extension also results in the nucleus becoming centrally, or even somewhat posteriorly, situated. Myonemes are conspicuous in the species, five or six being clearly evident in the cytoplasm of the intermediate form.

Class Sporozoa Leuckart Order Coccidia Leuckart Haemogregarina bigemina Laveran and Mesnil, 1901 (Pl. 7, figs. 1–43). Type Hosts: Blennius pholis L., B. gattorugine Bloch (Blenniidae). Type Locality: Cap de la Hague, Channel coast of France. Additional Locality Records from the Type Hosts: Mediterranean coast of France (Laveran and Mesnil, 1902), Naples, Italy (B. pholis only) (Neumann, 1909), Isle of Man, England (Henry, 1913). Additional Hosts and Localities: Chirolophis galerita (L.) = (Blennius montagui) (Blenniidae)—Italy (Neumann, 1909); Zoarces viviparus (L.) (Zoarcidae)—England (Bentham, 1917); Zoarces anguillaris Peck (“Zoarces angularis”)—Atlantic coast of Canada (Fantham et al. 1942). H. bigemina belongs to the group for which Henry (1912) coined the name of “schizohaemogregarines”. This name refers to the unique course of gametocyte formation in these interesting parasites. In the typical haemogregarine life cycle, as elucidated for Haemogregarina stepanowi Danilewsky of chelonians by Reichenow (1910), asexual multiplication takes place in the bone marrow or internal organs of the vertebrate host. Repeated schizogonies culminate in the production of merozoites larger than those concerned with the continuation of the asexual cycle. Each of these large merozoites invades a red corpuscle in the circulating blood, and there develops into a micro- or macrogametocyte. The microscopic examination of smears of infected blood seldom reveals the presence of any stages other than gametocytes, and it is often the case that the two sexes cannot be satisfactorily differentiated from one another. In the schizohaemogregarines, the merozoite destined to develop into the sexual stage does not do so directly. Having entered a red corpuscle in the circulating blood it divides again, once in the case of H. bigemina and allied species, twice in H. quadrigemina Brumpt and Lebailly, 1904, and related forms. Each of the products of division then develops into a gametocyte, so that each parasitized erythrocyte finally contains two or four mature gametocytes. The schizogony leading to the formation of the pre-gametocytic merozoites takes place in the circulating blood. Neumann (1909) described the development of up to 16 such merozoites of his H. polypartita in the red corpuscles of the European Gobius paganellus Gmelin. Merozoite formation in H. bigemina, described for the first time hereunder, takes place not in red corpuscles but in leucocytes. Nothing is yet known of the means of transmission of schizohaemogregarines, or indeed of piscine haemogregarines in general. In the present incomplete state of our knowledge it would be premature to accord them separate generic status, although what little is known of their development suggests that it may ultimately prove necessary to do so. It is better for the present that they be retained in the genus Haemogregarina Danilewsky which, as Reichenow (1927-29) aptly stated, is largely a repository for insufficiently known forms. Much confusion exists with regard to this genus, within which numerous species—including many schizohaemogregarines—have been established merely on the grounds of their occurrence in hosts from which haemogregarines had not previously been described. As Minchin (1907) remarked, many of these names “must not be taken as necessarily denoting distinct natural species … but merely, so

to speak, as labels affixed to certain classes of objects, whereby they become represented by parts of speech and can be referred to briefly.” H. bigemina has not previously been reported from the Southern Hemisphere, although limited material of haemogregarines obviously close to this species and possibly referable to it has been described from blennies in South Africa (Fantham, 1930) and Fiji (Laird, 1951). During the present studies it was found in the blood of five new hosts:— Gobiesocidae. Oliverichtus melobesia (Phillipps)—1'/1, Russell (February 6, 1951); 4/29, Wellington (July 6, September 14, 1951). Blenniidae. Ericentrus rubrus (Hutton)—16/20, Wellington (September 17, December 2, 1950; March 15, May 6, July 6, 1951). Tripterygion varium (Forster)—2/5, Russell (February 6, 1951); 2/6, Auckland (January 12, 1951); 2/9, Tolaga Bay (February 15, 1951); 21/90, Wellington (July 30, August 8, 1949, September 1, October 15, December 26, 1950, March 3, June 6, July 6, September 14 and 29, 1951). Tripterygion medium Günther—1/36, Russell (February 8, 1951); 6/75, Wellington (July 30, August 7, 1949, October 8, December 25, 1950). Notoclinus fenestratus (Forster)—1/1, Wellington (October 15, 1950). The length range of the infected fishes was as follows, the bracketed figures representing the overall length range noted for each species:— O. melobesia : 19–36 mm. (19–39 mm.) E. rubrus : 25–51 mm. (25–53 mm.) T. varium : 21–99 mm. (21–99 mm.) T. medium : 22–73 mm. (20–91.) N. fenestratus : 41 mm. Although a precise study of the seasonal incidence of H. bigemina was not undertaken, the parasite was recorded throughout the year except in April and November, in which months very few fish were collected. The haemogregarine was found in fishes of practically all sizes, the heaviest infections being found as a rule in the younger examples. Its common occurrence in very young fishes of about 2 cm. in length is of considerable interest as regards transmission. Leeches are the vectors of some, if not all, of the haemogregarines of aquatic reptiles and amphibians, but these have not been recorded from any of the New Zealand hosts of H. bigemina. Neither Laveran and Mesnil (1902) nor Neumann (1909) found leeches on the European blennies infected with H. bigeminà, although Abranchus blennii has recently been described from Blennius pholis L. in Wales by Jones (1940). Laveran and Mesnil (1902) recorded leeches (Hemibdella soleae) from soles at Roscoff infected with schizohaemogregarines (H. simondi Lav. and Mes., 1901), but Lebailly (quoted by Neumann, 1909), found only unaltered examples of H. simondi in the gut of the leech Platybdella solea. With the exception of the peritrichous ciliates described herein, the only evidence of ectoparasites discovered in the present studies took the form of some detached legs of an isopod found beneath the operculum of a 45 mm. example of Ericentrus rubrus. Laveran and Mesnil (1902) and Neumann (1909) recorded isopods from European shore fishes, but the latter author could find no trace of haemogregarines in these. At all events, if H. bigemina does

indeed have a blood-sucking invertebrate as its intermediate host, this not only must be of very small size to be able to draw blood—and that without leaving any external sign of its attachment—from fishes only 2 cm. in length, but must also remain attached for only brief periods. Perhaps, as Neumann (1909) suggested, transmission may be effected by way of the intestinal tract, as has since been proved to be the case for some of the haemogregarines of lizards. The incidence of infection in Ericentrus rubrus was extremely high (80 per cent.). It was 24.5 percent. in Tripterygion varium, 16.7 per cent. in Oliverichtus melobesia, and only 5.9 per cent. in Tripterygion medium. Laveran and Mesnil (1901) published only a brief account of H. bigemina, their description being confined to intraerythrocytic schizonts and gametocytes. They described and figured the division forms as cylindrical or globose, nuclear enlargement and division followed by cytoplasmic cleavage resulting in the formation of two rounded or pyriform bodies (the rounded ones illustrated in their Fig. 5 being of unequal sizes). These authors stated that the division products develop into “Hémogrégarines adultes” (recognized as gametocytes by Neumann, 1909), elongated and somewhat crescentic bodies swollen anteriorly and tapering posteriorly and attaining 12μ in length by 1.5μ–2μ in breadth. Merezoite formation in H. bigemina closely resembles this process as described for H. polypartita of the European Gobius paganellus Gmelin by Neumann (1909), with the important difference that it takes place in cells devoid of haemoglobin instead of in red corpuscles as in the latter species. In the New Zealand material, schizogony leading to the production of merozoites was observed to take place in basophil erythroblasts, small and large lymphocytes, and monocytes. This process was seldom recorded in the blood of the older fishes, these having chronic infections characterized by the presence of very scanty numbers of gametocytes only. Intraleucocytic schizonts and merozoites far out-numbered schizonts and gametocytes in cells containing haemoglobin, in the blood of young fishes having recently-acquired infections. Table III. Schizont and Merozoite Counts from 100 Successive Small Lymphoèytes and Monocytes of E. rubrus parasitized by H. bigemina. Type of Leucocyte Number of Schizonts Number of Merozoites 1 2 3 4 5 2 4 6 8 10 Small lymphocytes 10 32 6 3 0 7 34 6 2 0 Monocytes 0 37 5 8 2 0 30 6 11 1 A 37 mm. example of Ericentrus rubrus collected at Moa Point, Lyall Bay (Wellington) on September 17, 1950, had by far the heaviest infection noted during this investigation. Fifty per cent. of its small lymphocytes and 85 per cent. of its large lymphocytes and monocytes contained schizonts or merozoites, while 4 per cent. of its erythroblasts and erythrocytes held schizonts of the final series or gametocytes. Table III details the results of counts of 100 successive parasitized cells made in the case of small lymphocytes and monocytes respectively, to determine the incidence of the various stages present. The earliest schizogonies of haemogregarines are those resulting from the initial invasion of various cells of the vertebrate host by sporozoites derived from the invertebrate host. These schizogonies result in the formation of large num-

bers of small merozòites. Some later schizogonies result in the production of smaller numbers of larger merozoites, which invade cells of the circulating blood and there develop into gametocytes. The earliest stages of H. bigemina encountered were small, ovoid or reniform merozoites occurring free in the plasma of the host and measuring some 3μ by 2μ (Fig. 1). As seen in Giemsa-stained smears, these have a relatively small amount of cytoplasm staining a very pale blue. The relatively large nucleus, which has an irregular outline and measures some 2μ in diameter, is quite rich in chromatin. These merozoites are equivalent to those formed in the earlier schizogonies of typical haemogregarines. The location of the schizonts from which they are derived was not discovered. Invading various white cells (Figs. 2, 3), the small merozoites develop into schizonts of a pyriform (Fig. 4) or rounded shape.* These schizonts closely resemble certain of the life-history stages of a protozoan recorded by Neumann (1909) from the mononuclear or endothelial cells of Gadus aeglifinus L, and described by that author as representing a new genus, Globidium (this name being preoccupied, it was changed by Brumpt, 1913, to Globidiellum). They also have affinities with a haemosporidian described by Henry (1913a) from Scomber scombrus L. Henry (1913b) brought forward evidence suggesting that Globidiellum is an intraleucocytic stage in the life-history of his Haemogregarina aeglifini. At this stage the nucleus of the parasite is pinkish staining and rather diffuse. The cytoplasm is whitish blue and, like that of the equivalent stages of some other sporozoans found in white cells (e.g., the saurian Plasmodium mexicanum Thompson and Huff, 1944), is often so masked by the deeply staining cytoplasm of the host cell that its outline is very difficult to distinguish. The nucleus of the host cell is usually (Fig. 4), though not always, indented by the haemogregarine. Frequently, especially when within the larger leucocytes, the parasite completely penetrates and becomes surrounded by the nucleus of the host cell (Fig. 9). Binary fission now takes place (Figs. 6, 9), the rounded or oval bodies so formed ranging in size from 2.7μ to 4.8μ by 2.4μ to 4.0μ (av. for 50 examples, 3 7μ, by 3.2μ). Two such schizonts are usually present in parasitized cells (Table III), each of them giving rise to two merozoites, the dominant number of merozoites per cell thus being four. In some small lymphocytes, the bodies resulting from the initial binary fission, instead of rounding up to become schizonts in their turn, develop directly into merozoites. In such cases only two merozoites are formed in the host cell concerned (Table III; Fig. 5). A further binary fission of one or both of the schizonts of the second series (Fig. 10) may lead to the production of six (Figs. 11, 12) or eight merozoites, or a fourfold division of the initial schizont (Fig. 8) may result in the formation of four schizonts instead of two in the second series (Fig. 13) and consequently of eight merozoites (Fig. 14). Eight merozoites are comparatively seldom formed in a small lymphocyte, although they are quite commonly found in large lymphocytes and monocytes (Table III). Sometimes, but only in the larger host cells, one of the four schizonts formed in the manner related above will undergo a further division. More than five schizonts or 10 merozoites have not been observed in any one host cell. Merozoites resulting from the intermediate schizogony in the white cells are vermicular bodies, one extremity usually being rounded and the other more or

less pointed. Their dimensions range from 3.9μ to 7.1μ by 0.5μ to 1.0μ (av. for 50 examples, 5.5μ by 0.7μ). The cytoplasm stains light blue with Giemsa, somewhat more deeply so at the periphery, and the centrally (Fig. 12) or posteriorly (Figs. 11, 14) situated nucleus, which occupies the full width and from a third to half the length of the merozoite, is rich in chromatin and stains deep red with lighter red maculations. These merozoites are morphologically close to those of H. polypartita Neumann, but are relatively more slender than are those of the latter species, which measure 4μ by 1μ (in never-dried haematoxylin preparations). The course of merozoite formation broadly corresponds in the two species, in both of which the commonest number of merozoites formed in a single host cell is four. The maximum number so formed is 16 in H. polypartita as compared with 10 in H. bigemina. The merozoites are finally freed (Fig. 15), either as a result of their own active movements or of the breakdown of the host cells. Each of them now invades a cell containing haemoglobin, sometimes an erythrocyte (Fig. 20), but much more commonly an erythroblast (Fig. 16). Here they round up to a greater or lesser extent (Figs. 17, 18), usually becoming somewhat pyriform or else cucumber-shaped. Continued growth in size leads to the production of a reniform (Fig. 23), ovoid (Fig. 21) or irregularly rounded schizont. In rare instances two such schizonts are present in one red cell (Fig. 24), in consequence of the invasion of that cell by two merozoites. Schizonts within red cells range in length from 3.6μ to 9.2μ and in breadth from 2.2μ to 4.6μ (av. for 100 examples, 5.5μ by 3.0μ). The nucleus finally divides, the two daughter nuclei migrating to the ends (Fig. 25) or sides (Figs. 19, 24) of the schizont. Cytoplasmic cleavage may take place in the lateral or longitudinal (Figs. 26, 27) plane, and in the latter event may be diagonal (Fig. 28). If cleavage is lateral, the division products are rounded, like those illustrated by Laveran and Mesnil (1901) in their Fig. 5; if it is longitudinal, the young gametocytes are elongate-pyriform in shape (Figs. 29, 30). One red corpuscle which had been invaded by merozoites at two different times was seen. This corpuscle (Fig. 35) contained two developing gametocytes and a schizont as well. Two gametocytes are invariably formed from each schizont. Although Laveran and Mesnil once observed the commencement of a fourfold division, they never found more than two mature gametocytes within an individual red cell. The nucleus of immature gametocytes is fairly large and of very irregular outline. It is usually situated towards the broader end of the body as in the examples illustrated in Figs. 28–32 and in Fig. 6 of Laveran and Mesnil (1901). The gametocytes are able to move quite actively within the infected erythrocytes from an early stage in their development. They thus adopt a variety of positions with regard to one another and to the nucleus of the host cell. Sometimes they remain in close contact with one another on the same side of the nucleus of the host cell, either with their equivalent extremities corresponding (Figs. 34, 38) or in the tête-bêche position (Figs. 36, 37, 39, 40). They frequently take up positions on either side of the nucleus of the host cell, immature gametocytes often investing this structure very closely (Figs. 30, 31). With the approach of maturity the gametocytes become longer and relatively more slender, the anterior end being somewhat clubbed and rounded and the posterior one markedly attenuated. The organism is at this stage somewhat

crescentic in shape, but being very plastic it may (as figured by Laveran and Mesnil) be bent into the form of a U or have its fine posterior extremity sharply reflected (Fig. 35). The nucleus migrates into the posterior two-fifths of the body, but does not reach to within more than 2μ of the actual extremity. It becomes an elongated structure, ranging from 2.1μ to 4.4μ in length and from 1.0μ to 1.8μ in breadth. Occupying the full width of the body, it fills from a fifth to a quarter of the total length of the organism. Laveran and Mesnil failed to note any granular inclusions in the cytoplasm of the gametocytes of H. bigemina which they studied. In the New Zealand material, two or three small granules staining blackish red with Giemsa are usually present (Figs. 36, 37, 39), forming a compact group between the nucleus and the posterior extremity of the body. These granules closely resemble those described by Henry (1913c) from H. simondi, a species which Laveran and Mesnil (1901) described but from which they failed to record such granules. According to Henry, the granules of H. simondi are shed by the parasite to initiate a new developmental cycle in the vertebrate host. Similar granules are, however, also present in the intracellular merozoites of piscine haemogregarines (Laird, 1952) (see also Fig. 20). It is probable that they are metabolic products, perhaps associated in some way with the acts of entering and leaving host cells. They may also, like the pigment granules of haemosporidians, be derivatives of ingested haemoglobin. The bulk of the cytoplasm of that part of the body anterior to the nucleus stains a pale whitish blue with Giemsa. It assumes a somewhat darker colour peripherally, and may be more or less macular in appearance. That portion of the cytoplasm posterior to the nucleus stains dark blue. A small polar cap which stains rose-red may be present at the broader end of the body (Fig. 40), but the substance concerned is never as abundant as it is in fish haemogregarines of the rovignensis group (Laird, 1952). The overall length of mature gametocytes from all the New Zealand hosts ranged from 9.1μ to 14.9μ, and the overall breadth at the widest part of the body from 1.0μ to 2.2μ. The figures derived by Laveran and Mesnil (1901) from their material were 12μ by 1.5μ to 2μ. Average measurements were calculated for 100 mature gametocytes from Oliverichtus melobesia, Ericentrus rubrus and Tripterygion varium, the hosts from which the most abundant material was available. These measurements are detailed in Table IV, together with those of 50 normal erythrocytes from each of the hosts and the percentage increase in length estimated for the 50 parasitized red cells. Table IV. Normal Erythrocyte Percentage Length Gametocyte Av.length av.brdth. Increase of Lenght (μ) Breadth (μ) Host (μ) (μ) Infected Cell Range Av. Range Av. E. rubrus 10 68 79 3 12.3% 9 0-13 9 11.38 10–18 1.25 T. varium 12 36 7 99 5.5% 9.1-14.8 11.94 1.0-2.0 131 O. melobesia 13.33 9 01 3.7% 9.1-14.9 12.28 10-24 1.56 As Table IV indicates, there is a striking correlation between the size of the host erythrocytes and that of the mature gametocytes of H. bigemina, the average dimensions of the latter increasing in direct proportion to those of the red cells

of the host. The overall size range of the gametocytes is much the same throughout. Examples of less than 9μ in length are immature, the nucleus still being anteriorly or medianly situated. As regards the upper limit of the size range, only seven of the 100 gametocytes from T. varium and four of those from O. melobesia exceeded 13.9μ in length, the upper limit attained by those from E. rubrus. Both schizonts and gametocytes from all five New Zealand hosts were of identical morphology throughout. More often than not, the two gametocytes in an individual erythrocyte are dissimilar in size. A count of those in 50 consecutive parasitized erythrocytes chosen at random from among all five hosts revealed that in only six cases were the two gametocytes of exactly the same size. Erythrocytes containing gametocytes of H. bigemina become somewhat hypertrophied (Table IV), and the nucleus is frequently displaced (Figs. 24, 37, etc.). Hypertrophy may be very marked in doubly parasitized cells (Figs. 24, 35). The staining characteristics of the cytoplasm and nucleus are not affected. The maturation of the gametocyte accompanies that of the erythrocyte itself, as indicated in Table V (compiled from the same preparation from Ericentrus rubrus from which the merozoite counts in Table III were made). Table V. Schizont and Gametocyte Counts from 100 Successive Red Cells of E. rubrus Parasitized by H. bigemina. Number of Number of Gametocytes Type of Red Cell Schizonts Immature Mature Erythroblasts 31 2 0 Erythocytes 9 14 44 Temporary distortion of the host cell membrane may be brought about by the active movements of the gametocytes within it (Figs. 39, 40). These movements were observed in a fresh cover slip preparation of infected blood from E. rubrus, examined by dark ground illumination. Two mature gametocytes lying on either side of the nucleus of an erythrocyte were seen to straighten themselves out, then contract to the shape of a U. The straightening movement required some 15 seconds for its accomplishment, while contraction occupied only four or five seconds. When the gametocytes, moving synchronously, were both fully extended, the pressure of their extremities against the membrane of the host cell caused this cell to assume an almost rectangular shape, its sides being slightly convex and its ends markedly concave. On the contraction of the haemogregarines, however, the elasticity of the host cell membrane was such as to allow the erythrocyte to resume its normal shape immediately. Other cells were observed in which only one of the two gametocytes was active, and in others again the parasites moved alternately. After about ten minutes of intraerythrocytic activity the gametocytes commence to leave the host cells. Direct pressure against the membrane raises a nipple-like projection (Figs. 39, 40) which finally ruptures, allowing the haemogregarine to glide out into the plasma. Free gametocytes, also those ready to leave the host cells, are bluntly pointed anteriorly (Figs. 40, 43). Like those of other schizohaemogregarines the free gametocytes of H. bigemina move about quite slowly in a gliding manner, always with the broader end foremost. They are often a little narrower and rather

longer (Fig. 43) than intraerythrocytic forms, but otherwise closely resemble these. In addition to H. bigemina, a number of other species of schizohaemogregarines typically forming two gametocytes in each host cell have been described. Some of these “species” are morphologically indistinguishable from H. bigemina, and their gametocytes fall within the size range of those of the latter parasite (e.g., H. delagei Laveran and Mesnil of European skates). The description of H. gobii Brumpt and Lebailly (1904) is inadequate, and in no way differentiates this haemogregarine of the European Gobius minutus Gmelin from H. bigemina. It is likely, especially in view of the wide range of hosts now known to be parasitized by H. bigemina, that a re-examination of material from the four European species of flat-fishes from which Lebailly (1904, 1905) briefly described four different species of Haemogregarina, will disclose that he was actually dealing with one species only, H. platessae Lebailly, 1904, or perhaps even H. bigemina itself. Three other species of haemogregarines have been described from blennies. Kohl-Yakimoff and Yakimoff (1915) described H. londoni from the European Blennius trigloides. These authors found only intraerythrocytic schizonts and gametocytes. The schizonts of H. londoni fall within the size range of those of H. bigemina, but the gametocytes do not. These stages measure 9.94μ by 2.48μ to 2.84μ. They are thus broader than the mature gametocytes of H. bigemina, from which they further differ in that their extremities are of the same form and breadth. Sometimes they are surrounded by a cyst-like structure. Such a structure, which may perhaps be formed by the parasitized cell rather than by the parasites themselves, has not been reported for H. bigemina from any of its hosts. Kohl-Yakimoff and Yakimoff noted that some erythrocytes contained only one gametocyte. This could indicate either that the final schizogony does not always take place in H. londoni, or that two gametocytes had developed, one of them already having left the cell. If blood smears are taken from fishes parasitized by H. bigemina an appreciable time after the death of the host, or if cover slip preparations are examined under the microscope for 10 or 15 minutes before permanent smears are made, some of the host cells will be found to have lost one or both of the gametocytes developed within them. In so far as is yet known, H. londoni differs sufficiently from H. bigemina to justify its retention for the present as a valid species. H. salariasi was described from such limited material from Salarias periophthalmus Val. in Fiji (Laird, 1951) that full and satisfactory comparisons with previously described haemogregarines could not be made. It was hence accorded specific rank purely as a matter of convenience in reference. Only two mature gametocytes were found, these being present in one of the three parasitized red cells located. They were of very small size (7.7μ by 1.0μ and 8.3μ by 1.0μ), but otherwise closely resembled the equivalent stages of H. bigemina, with which H. salariasi may ultimately prove to be conspecific. Fantham (1930) described H. fragilis from Blennius cornutus (L.) in South Africa. He was dealing with stained material only, and the fact that his preparation contained many free forms indicates that it was probably made some little time after the death of the host. These free forms were of similar size to the free gametocytes of H. bigemina, measuring 11μ to 16μ (mostly about 13μ to 14μ) by about 1.5μ. Intracorpuscular forms measured about 9μ to 12μ

by 1.5μ, thus coming within the size range of the mature gametocytes of H. bigemina. Fantham made no mention of more than one parasite being present within a single host cell, and interpreted the haemogregarines as in the process of entering, not leaving, the erythrocytes. The nucleus of H. fragilis, unlike that of mature gametocytes of H. bigemina, is situated towards the broader end of the body. Specimen slides of H. bigemina from New Zealand hosts have been deposited in the collection of the Dominion Museum, Wellington (catalogue numbers Z16–17). Haemogregarina (Hepatozoon?) acanthoclini n.sp. (Pl. 8, Figs. 44–50). Six of 11 examples of A. quadridactylus collected at Tolaga Bay on February 14, 1951, were very scantily infected with this sporozoan. None of the remaining kelpfish studied were parasitized. Mature gametocytes were the only stages present, these occupying corpuscles which, despite a superficial resemblance to the smaller leucocytes (Fig. 44), are probably distorted and rather shrunken erythrocytes. The nucleus of the host cell is rather hypertrophied, indented or flattened, and displaced to such an extent that it is squeezed between the parasite and the cell membrane. It is sometimes pyknotic, its chromatin being grouped into a few very deeply staining aggregations. The parasite is reniform (Figs. 46, 47) or cylindrical in shape, the ends being rounded and of equal width. Only 11 were seen in all. These measure from 8.4μ to 9.4μ by 3.2μ to 4.0μ (av. 8.5μ by 3.5μ). The cytoplasm is dense and homogeneous, and stains a deep and uniform blue, quite distinct from the light blue colour assumed by that of other haemogregarines of fishes. It contains a few small, round vacuoles, and in some cases a dot of chromatic material which stains a bright red colour (Fig. 50). One (Fig. 50), two (Fig. 47) or more (Fig. 48) myonemes may be present. The nucleus is situated at or near the middle of the organism, and is usually round or ovoidal in shape. In one example it extends like a band almost across the body. The nucleus stains a uniform light pink colour. Together with the staining reaction of the cytoplasm and the presence of myonemes, this suggests that the parasite in question may belong to the genus Hepatozoon. Members of this genus have previously been described from reptiles, birds and mammals, but not from fishes. The nucleus ranges from 2.4μ to 3.3μ in its greatest diameter by 1.4μ to 2.6μ (av. 2.9μ by 2.1μ). In all six infections the parasite rate was less than one per 100,000 erythrocytes. The haematozoan from A. quadridactylus is quite distinct from any of the haemogregarines previously described from fishes. It is accordingly designated Haemogregarina (Hepatozoon?) acanthoclini n.sp., the generic designation being used in its broadest sense. A slide designated as the type of this species has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z18). Paratype slides are in the collection of the Department of Zoology, Victoria University College, and in my own collection.

All figures drawn with the aid of an Abbé camera luida Laveran and Mesml 1901 (Illustrations prepared at a magnification of 2 2400X from Grema-starned heart blood smears) Figs. 1-15— Leu series from Fig. 16 — Ervthroblast of rubius Figs. 17-18 — Intr tropphozortes from E. rubrus Fig 19 — Intra sehizont from E. rubrus Fig 20 — Intr tiophozoite from Tripteruqion aum Figs. 21-22— Intr schizonts from T Fig 23 — Intraerthroetic zont from Ter on medium Fig 24 — Eythrote of ubrus doubly intecfed with two onts Figs 25-29 — Schizogony, from T. medium (25, 26, 28, 29) and E. rubrus (27) Figs 30-34 — Developing gametocytes, from T medium (30, 33) and T. medium (31, 32, 34) Fig. 35 — Divthrocyte of rubrus doubly intecfed at two different times and containing a schizont and two gametocytes resulting from an earlier schizogony Figs 36-40 — Intraelthrocytic gametocytes from T (36, 39), E. rubrus (37) and T. medium (38, 40) Fig. 41 — Free gametocte adhering to the greatly hypertiophied nucleus of a disintegrated erythiocyte of T. Figs 42-43 — Gametocytes free in the plasma of T. medium (42) and E. rubius (43).

Illustrations prepared at a magnification of 2.330X Figs. 44-30 — Haemogreana (Hepatozoon) acanthoclim nsp. drawn from Gremsa-stared heart blood smears of Acanthoclinus quadidactylus Fig 44 — Leucocyte of 1 quaddactylus Fig 45. — Erythrocyte of A quadrida tylus. Figs. 46-49. — Intraer gametocytes of H. acanthoclim. some (47-49) showing myonemes Fig 30 — Free gametovte. Figs 51-58 — Leptotheca subelegans n. sp. drawn from fresh bile preparations of Diplocrepis puniceus and Callogobius atratus Fig 51 — Elongate trophozorite from C. atatus Fig 52 — Sporulating trophozoite from D puniceus Figs 53-54 — Young spores still attached together in pars by a residium of cytoplasm derived from the trophozoite, from D. puniceus (53) and C. atratus (34) Fig 55 — Spore with one polafilament discharged by treatment with 5% phenol, from D. puniceus Fig 36. — Polar view of spore from C. atratus, showing lateral swellings Figs. 37-38 — Mature spores in front view, from D. puniceus (37) and C. atratus (58).

Illustrations prepared at a magnification of 2,530X Figs 59-61 — Darisia diplocrepis n. gen. n. sp. from urinary bladder of Diplocrepis puniceus. Fig 59 — Sporulating trophozorte from fresh bile preparation Fig 60 — Polar view immature spore, showing relatively short and stubby lateral appendages, and the crinkled condition of the polar filaments when discharged following treatment with 5% phenol Neutral red as intia vitam starn Fig 61 — Front view of mature spore. Neutral red as intra vitam stain. Figs 62-71 — dium atum Thélohan, 1892, from gall bladder of Diplocrepis puniceus, Olerichtus melobesta, Notoclinus fenestratus and Acanthoclinus quadridactylus Fig 62 — Rounded-up and vacuolated trophozoite from a cover slip preparation of fresh bile from O. melobesia Fig 63 — Sporulating trophozoite, from fresh bile of D. puniceus Fig 64 — Young spore, from fresh bile of D. puniceus Figs 65-66 — Young spores from D. puniceus (65) and O. mlobesia (66) Gremsa Fig 67 — Mature spore from fresh bile of D. puniceus Side view, showing polar filament discharged following treatment with 5% phenol Fig 68 — Side view of mature spore from D. puniceus Iron haematoxylin Fig 69 — Sutural view of mature spore from D. puniceus Gremsa. Figs. 70-71 — Mature spores from A quadridactylus (70) and O. melobesia (71), in sutural view, Iron haematoxylin.

Figs. 72-74 — Sphaeromura tripterygii n. sp. from gall bladder of Trpterygion varium and T. medium Fig 72 — Trophozorte, fixed in Schaudinn's fluid and stained with acetic-alum-carmine 16X. Figs 73-74. — Spores before (73) and after (74) autogamy Iron haematoxylin 2,530X. Fig 75 — Zschollella sp. from urinary bladder of Tripterygion raum Sporulating trophozoite, from co p preparation of fresh urine 2.530X Figs 76-77. — Myrosoma tripteryn n. sp. from subdermal connective tissue of Tripteryon arm Fig 76 — Young spores, connected together by a residuum of cytoplasm derived from the trophozoite. Drawn from fresh material at a magnification of 2,530X Fig 77 — Mataure spore, drawn from fresh material at a magnification of 2,530X. Figs 78-79. — Scyphidia (Gerda) acanthoclini n. sp, from gills of Acantholinus quadridactylus Fig 78. — Medium sized trohozoite, its scopula attached to the tip of a gill filament of the host. Whole mount, starned with on haematoxylin. 1,700X Fig 79 — Large trophozoite. semi-diagrammatic Cilia not indicated Whole mount, stained with upon heamatoxylin. 1,700X.

The scale-line at the right of each figure represents 10 microns at the same magnification Fig 80 — Scyphidia (Gerda) acantho n. sp. from gills of Acanthonus quaddactulus Large trophozoite drawn from whole mount stained with on haematoxylin For greater clarity, only on row of ha is indicated 1,700X Figs 81-83 — Caliperia longipes n. gen, n. sp. from gills of Oliveihtus melobesia and Eicentus rubus Fig 81 — Trophozoite, from te showing food inclusions O mlobesia 880X Fig 82 — Whole mount of trophozoite, starned with on haematoxylin Only the base of each posterior process is illustrated For greater clarity only one row of ha is illustrated 1,700X Fig 83 — Whole mount of trophozoite, stained with on haematoxylin. The entire animal is illustrated. Note siderophilous granules in the posterior processes 880X Fig 84 — Endosphaera engelma Entz. 1896. parsitizing Trichodina (Trichodina) multidentis n. sp. from gills of Triptery medium young trophozoite in whole mount of T. multidentis stained with on haematoxylin 2,530X Fig 85 — Trichodina (Trichodina) parabranchicola n. sp. from gills of various intertidal zone fishes Trophozoite in side view, indicating the disposition of the components of the skeletal complex Whole mount from gills of quadadactylus, on haematoxylin 880X

All figures drawn from gill smears starned with iron heamatoxylin. The scale-line at the right of each figure represents 10 microns at the same magnification Figs 86-87. 90-91. 93-97 — Trichodina (Trichodina) parabranchicola n. sp. Fig 86. — Adoral view, detached disc of small example having relatively long erlia, from Oliverichtus melobesia. 880X Fig. 87 — Adoral view, detached disc of large example from Acanthoclinus quad-ridactylus 880X Fig 90 — Three denticles in silhouette 2.530X Fig 91 — Macrounucleus. showing chromatin spherules and superposed micronucleus 1,700X Figs 93-96 — A series following binary fission, illustrating the macronuclear development accompanying growth of the organism semi-dragrammatic 880X Fig 97 — Silhouette of portion of the denticulate ring of a young, growing trichodinid, showing the interpoloation of a new denticle, 1. 700X. Figs. 88, 89, 92 — Trichodina (Trichodina) multidentis n. sp. Fig 88. — Adoral view, detached disc of large example from Tripterygion varrum 880X Fig 89 — Three denticles in silhouette. 2,330X. Fig. 92. — Cocconers sp, from a food vacuole of T. multidentis, n. sp.

Binary fission in Trichodina (Trichodina) parabranchicola. n. sp. and Trichodina (Trichodina) multidentis n. sp. Only units of the skeletal complex are illustrated, and the border membrane is omitted from figures 99-101 All figures drawn at a magnification of 2,530X from gill smears stained with iron heamatoxylin. Fig. 98 — Division of a very large example of T. parabranchicola, viewed from the aboral aspect. Macronuclear division is completed, cytoplasmic cleavage is taking place, and the denticulate rings of each of the daughter trichodinids are about to unite. The individual plates of the new rings are clearly distinguishable Fig 99 — Division product of T. parabranchicola (towards the lower limit of the size range) The denticles derived from the parent are only loosely inserted into one another but in this instance the plates of the new denticulate ring are not yet differentiated. Aboral view Fig 100 — Adoral view T. parabranchicola A later stage than that represented in Fig. 99 All but the cones of the denticles derived from the parent have here been absorbed, and these are no longer interconnected. The hooks of the new denticles have formed the cones are strongly developed and the characteristic number of units in the striated band is in process of being restored by the development of new stae between those derived from the parent Fig 101 — Aboral new o a detached skeletal complex of T. multidentis Absorption of the parent denticles is commencing while the cones of the cones of the new ring are becoming apparent.

Order Myxosporidia Bütschli Within this order are included those sporozoans having a resistant spore, the covering of which is in the form of a bivalve shell, from one to four polar filaments each coiled within a polar capsule, and one sporoplasm only. The vegetative stage is either coelozoic or histozoic. Upwards of 400 species have been described, a few from amphibians and reptiles, but the great majority from fishes. Classification The spore characters serve as a basis for the classification of the order. Tripathi (1948a) briefly summarized the history of the classification of the Myxosporidia, and himself proposed a new classification based entirely on the spore characters, earlier investigators having found themselves forced to rely to some extent at least on such additional criteria as the micro-habitat of the parasite or the ecology of the host. The classification put forward by Kudo (1920) and later revised by that author (1930, 1933) is the one most generally followed by modern protozoologists. Kudo separated his three sub-orders, Eurysporea, Sphaerosporea and Platysporea, on the basis of the relation of the sutural axis to the greatest diameter of the spore. This led to certain anomalies, aberrant species the generic affinities of which were not in question sometimes having the criteria of a sub-order other than the one to which their genus belonged. Thus Dunkerly (1925) indicated that although the sub-order Sphaerosporea Kudo, 1920, was defined to include those myxosporidians having spherical or subspherical spores, certain species of Chloromyxum (C. caudatum Thélohan, C. quadratum Thélohan), a genus of this sub-order, have spores which do not fulfil these requirements. Tripathi (1948a) also instanced C. histolyticum Perard in this regard, and pointed out a similar anomaly in Leptotheca—this genus belongs to Kudo's sub-order Eurysporea, but certain of its members, notably L. spherula Noble, closely approach the requirements of Sphaerosporea in their spore characters. Kudo's (1933) emended diagnosis of his sub-order Sphaerosporea still further accentuates the anomaly first mentioned by Dunkerly (1925), for it gives the shape of the spore as spherical only. Additional exceptions are to be found in the genera Sphaerospora and Sinuolinea (Sphaerosporea: Sphaerosporidae). As seen in front view, several species of Sphaerospora have spores distinctly broader than long. For example, the genotype, S. divergens Thélohan, may have more or less elongate spores (cf. Auerbach, 1912, Pl. 5, fig. 4; illustration reproduced by Kudo, 1920, Pl. IX, fig. 185); Davis (1917) stated that his S. polymorpha has spores which are “sometimes slightly compressed parallel to longitudinal plane”; and Kudo (1920) described the spore of his S. carassii as variable in shape, his Pl. IX, fig. 204 illustrating a decidedly ovoidal condition. As regards Sinuolinea, spores of S. arborescens are distinctly longer than broad (Davis, 1917), and those of S. bidens and S. cella have ovoidal central chambers (Jameson, 1931). Tripathi's (1948a) classification divides the Myxosporidia into two suborders—Unipolaria Tripathi, in which the spore is of variable shape and has from one to four polar capsules at or near the anterior end, and Bipolaria Tripathi in which the spore has two widely separated polar capsules, one at each extremity. The sub-order Unipolaria is divided into two super-families—Cera-

tomyxoidea Tripathi in which the sporoplasm lacks an iodinophilous vacuole, and Myxoboloidea Tripathi, in which an iodinophilous vacuole is present in the sporoplasm. There is but one family in the sub-order Bipolaria. The superfamily Ceratomyxoidea is subdivided into four families, and Myxoboloidea into two, on the basis of the number of polar capsules in the spore. Thus far, I favour Tripathi's classification over Kudo's. It not only does away with the anomalies involved in making the mere shape of the spore a subordinal criterion, but finally abolishes the use as family criteria of the microhabitat of the parasite and the ecology of the host—Kudo (1933) included two families in his sub-order Eurysporea, defining Ceratomyxidae Doflein, 1899, as “Typically coelozoic parasites of marine fish” and Wardiidae Kudo, 1933, as “Histozoic or coelozoic parasites of fresh-water fish.” Other considerations aside, it is obviously most undesirable to be unable to assign certain myxosporidians to such a major group as a family without full information on their microhabitat or the ecology of their hosts. Furthermore, Kudo's (1933) classification left certain genera of obviously close affinities in different sub-orders. Thus Sinuolinea Davis (Sphaerosporea: Sphaerosporidae) has been used (in part) as a repository for forms having very close affinities with Ceratomyxa Thélohan (Eurysporea: Ceratomyxidae). By Tripathi's classification, Ceratomyxa and Sinuolinea are considered to belong to the same family. Considering that Tripathi's classification of the Myxosporidia is mechanically superior to Kudo's and affords a more satisfactory framework within which to demonstrate the relationships and affinities existing within the order, I have adopted it in this paper. Sinuolinea Davis, 1917. Davis defined this genus (the recognised species of which are listed in Table VI) as follows:—“Spores approximately spherical; with or without lateral processes. Capsules rounded, not convergent when seen from above; capsular pores some distance apart, sometimes on nearly opposite sides of the spore. Sutural line forming a prominent ridge, which takes a sinuous course around the spore. Sutural plane usually distinctly twisted on its axis. Disporous and polysporous.” Table VI The Species of Sinuolinea Davis. Species Locality 1. S. dimorpha Davis, 1916 U.S.A., Atlantic Coast 2. S. capsularis Davis, 1917 U.S.A., Atlantic Coast 3. S. arborescens Davis, 1917 U.S.A., Atlantic Coast 4. S. opacita Davis, 1917 U.S.A., Atlantic Coast 5. S. brachiophora Davis, 1917 U.S.A., Atlantic Coast 6. S. gilsoni Debaisieux, 1925 Belgium 7. S. bidens Jameson, 1931 U.S.A., Pacific Coast 8. S. cella Jameson, 1931 U.S.A., Pacific Coast 9. S. murmanica Basikalowa, 1932 U.S.S.R. 10. S. cyclopterina Basikalowa, 1932 U.S.S.R. 11. S. rebae Tripathi, 1948 England In all cases the organ infected is the urinary bladder, although S. murmanica has also been recorded from the kidney of the host (Basikalowa, 1932).

Seven of the species—S. dimorpha, S. capsularis, S. arborescens, S. gilsoni, S. murmanica, S. cyclopterina and S. rebae—have approximately spherical spores in which the suture line is markedly twisted on its axis. The other four—S. opacita, S. brachiophora, S. bidens and S. cella—are quite distinctive in that their spores have arm-like lateral extensions. The sutural line is only slightly curved in spores of S. brachiophora and S. cella, but is markedly twisted in those of S. opacita and S. bidens. With regard to S. brachiophora, Davis (1917) remarked that “Possibly this species should be made the type of a new genus, but the spore undoubtedly more closely resembles that of Sinuolinea than any other genus. In many respects this species is very similar to S. opacita, which occurs in the same host.” Jameson, noting these remarks, held that there are insufficient grounds for considering Davis's three species having spherical spores as distinct from Sphaerospora Thélohan. He claimed that “the only character that the spherical members of the genus Sinuolinea possess that is not present in some species of Sphaerospora is the sinuous suture line and that is hardly a sufficient ground for creating a new genus. Such a character is found in the genus Myxidium and is regarded as being of only specific worth.” Jameson (who failed to notice S. gilsoni) thus considered that S. dimorpha should revert to the genus Sphaerospora in which Davis (1916) had originally placed it, and further elected to transfer S. capsularis and S. arborescens to that genus. He retained Sinuolinea for S. opacita and S. brachiophora, and described his S. bidens and S. cella in this same genus. Sinuolinea, as emended by Jameson, contains those myxosporidians the spores of which have “first a clearly walled off, small spherical or oval central chamber in which the sporoplasm and pole capsules are located, and second a pair of prominent lateral appendages attached to this. The suture line is slightly curved but this is possibly not of great importance.” Jameson considered that the acceptance of his emended definition of Sinuolinea would necessitate the transfer to it of Ceratomyxa spinosa Davis, 1917, the only species of that genus in which the spore has a clearly demarcated central chamber. Several other species of Ceratomyxa have laterally extended shell valves, but the spore cavity is continuous with those of the extensions. Kudo (1933) reviewed the classification of the Myxosporidia, emending the definitions of certain of the genera and listing all the species known at that time. He neither accepted nor discussed Jameson's views on the status of the species of Sinuolinea, and retained Ceratomyxa spinosa in the genus to which it was allocated by Davis. Tripathi (1948a) followed Kudo in this regard, stating that “The spore of Ceratomyxa spinosa Davis, 1917 is not spherical, which is a necessary character of the members of the family Sphaerosporidae Davis, 1917…” Davis's definition of his Family Sphaerosporidae (1917, p. 219) reads as follows: “Spores pyramidal or approximately spherical; not distinctly longer than wide; with or without lateral processes.” As seen in front view, several members of the genus Sphaerospora itself are distinctly broader than long. For example, the genotype Sphaerospora divergens Thélohan may be more or less elongate (c.f. Auerbach, 1912; Pl. 5, fig. 4; reproduced by Kudo, 1920, Pl. IX, fig. 185); Davis (1917) himself stated that his Sphaerospora polymorpha is “sometimes slightly compressed parallel to longitudinal plane”; and Kudo (1920) described the form

of his Sphaerospora carassii as being “variable to some extent,” one of his figures (Pl. IX, fig. 204) illustrating a decidedly ovoid spore of this species. Similarly, Davis (1917) stated that the spore of his Sinuolinea arborescens is “rounded, slightly elongated along longitudinal axis” (his Pl. XXIII, fig. 110), giving the length of the spore as 15μ and the breadth as 12μ—this is hardly to be described as “not distinctly longer than wide.” Finally, the spores of both Sinuolinea bidens and Sinuolinea cella have ovoid central chambers, Jameson (1931) giving the dimensions of that of the former species as 8μ by 6.5μ to 11.5μ by 9μ and that of the latter species as 9μ by 8μ to 13μ by 10μ. It is thus considered that the definition of the major group to which the genera Sphaerospora and Sinuolinea belong should be emended so as to include species having spores oval in shape or with oval central capsules. In drawing a sharp distinction between those species of Sinuolinea having lateral extensions to the shell valves and those not having such extensions, Jameson (1931) contended that the species in the latter category only differ from the species of Sphaerospora in that their spores have a sinuous sutural line. He claimed that this is “hardly a sufficient ground for creating a new genus,” pointing out that the character of a sinuous suture line is regarded as being of only specific worth in the genus Myxidium. Jameson might also have instanced the genus Leptotheca in this connection for the spore of L. lobosa Davis (1917) has a sinuous suture line, whereas in most species of Leptotheca this line is more or less straight. As has already been indicated, Jameson proposed to retain only those species having a clearly walled-off spherical or oval central chamber, and a pair of prominent lateral appendages attached to this, in Sinuolinea. However, Jameson himself stated that in these species “The suture line is slightly curved but this is probably not of great importance.” The author agrees with Jameson in considering that Sinuolinea as defined originally by Davis (1917) [and more recently by Kudo (1933) and Tripathi (1948a)] embraces species having spores of two such distinct types that they should be accorded separate generic status. He does not agree with Jameson in retaining those species having spores with lateral extensions in Sinuolinea and transferring the others to Sphaerospora. Although there are isolated cases in other genera (c.f. Myxidium, Leptotheca) where some species have spores with sinuous sutural lines, these cases are the exception rather than the rule in the genera concerned. Seven species of Sinuolinea having spores lacking lateral extensions but possessing a very sinuous sutural line have now been described. It is contended that these species should be retained in Sinuolinea—which in any case would hardly be an appropriate generic name for the species having lateral extensions, Jameson himself having emphasized that the spores of these species have the sutural line only slightly curved. It is proposed, therefore, to establish a new genus, Davisia n. gen., to embrace those species of Sinuolinea Davis the spores of which have a clearly walled-off spherical or oval central chamber bearing a pair of lateral appendages. It is further proposed to follow Jameson in transferring Ceratomyxa spinosa to the company of these appendage-bearing myxosporidians previously considered to belong to Sinuolinea. Davisia n.gen. obviously has close affinities with Cera tomyxa Thélohan, 1892 (Kudo, 1933 emend.), the chief difference as regards spore characters being that the cavity of the central chamber is continuous with those of the appendages in the latter genus, but clearly walled-off from those of

the appendages in the former one. Elongate trophozoites having filopodia, so common in Ceratomyxa, have not been reported for any of the species here included in Davisia n.gen. Davisia diplocrepis n.gen., n.sp., (Pl. 9, figs. 59—61). Of 50 examples of Diplocrepis puniceus, ranging in length from 48 mm. to 90 mm. and collected at Point Jerningham, Lyall Bay and Island Bay, Wellington, in all months of the year (1950 and 1951), 44 had spores and trophozoites of D. diplocrepis n.sp. in the urinary bladder. There was no apparent pathological effect, apart from the fact that in extremely heavy infections the urinary bladder had a slightly milky appearance and its contents were very slightly viscous. Trophozoites. Irregularly rounded in shape, the size ranging from 4.1μ by 4.0μ to 32.3μ by 30.6μ. The commonest forms are approximately 20μ in diameter. There is a broad and clearly demarcated peripheral zone of hyaline ectoplasm (Fig. 59), and the rather granular endoplasm contains many greenish refractive spherules averaging 0.7μ in diameter. Disporous and polysporous, up to eight spores of various stages of development being present in a single trophozoite. Extremely slow non–progressional movements of broad-ectoplasmic lobopodia take place, and the trophozoites soon become rounded and motionless in fresh preparations under microscopic examination. Spores. Young spores (Fig. 60) have a more or less spherical central chamber. Their hollow lateral appendages are circular in cross-section. These appendages are shorter than those of mature spores, and at the point of attachment to the central chamber their bases are markedly inflated. The appendages, as seen in front view, may take up a variety of positions with regard to the central chamber. They may be straight and project at a slightly oblique downward angle, or be so recurved that they almost meet beneath the spore. The large shell valve nuclei, which are easily seen in fresh material, are located in the lateral appendages. Two rather smaller sporoplasm nuclei are present, lying close together one on either side of the rather curved sutural line. Small capsulogenous nuclei may be seen in contact with the round or ovoid polar capsules. Mature spores (Fig. 61) have an oval central chamber, the length of which (from the measurement of 50 fresh examples) is 9.0μ to 12.0μ (av. 10.7μ) and the breadth 12.1μ to 14 0μ (av. 13.0μ). The curved lateral appendages assume such an angle as seen in front view as to give the entire spore a crescent-shaped appearance (broken, of course, by the bulging contours of the central chamber). Both the shell valve nuclei and the capsulogenous nuclei persist to a late stage in development, and the sporoplasm, which almost or entirely fills the cavity of the central chamber, contains two nuclei positioned as in the young spore. The polar capsules, which are usually ovoidal, are somewhat flattened on the side adjoining the sutural line This line is clearly visible in living material, but there is no raised sutural ridge. The lateral appendages of the shell valves are somewhat narrowed proximally, expanding to an inflated base at the point of junction with the central chamber. They taper distally to a characteristic nipple-like projection. The appendages of an individual spore are usually of the same size, although they may be slightly unequal. They range in length from 10.0μ to 14.2μ (av 12.7μ) The greatest breadth of the appendages ranges from 2.8μ to 3.8μ (av. 3.2μ). The polar capsules of an individual spore are equal or subequal, ranging from 3.4μ to 3.8μ in length (av, 3.7μ) and from 3.2μ

to 3.5μ in breadth (av. 3.4μ). The tightly coiled polar filaments do not contact the periphery of the capsules, and have from 5 to 9 coils. Extrusion of the polar filaments is readily brought about with 5% phenol, these filaments being 40μ to 56μ in length (av. 47μ). The filaments of immature spores may be extruded by the same treatment, but these remain crinkled throughout their length (Fig. 60). Discussion. Davisia spinosa, D. opacita and D. brachiophora all occur in the urinary bladder of the same host, the halibut Paralichthys albiguttus (from Beaufort, North Carolina); while D. bidens and D. cella are both found in the urinary bladder of a toadfish, Porichthys notatus, in California. Davis (1917) stated that his (Ceratomyxa) = Davisia spinosa was rare, and recorded (Sinuolinea) = D. opacita and (Sinuolinea) = D. brachiophora on but one occasion each. In view of the considerable variation in size and shape exhibited by spores of D. diplocrepis n.sp. during their development, the possibility that Davis was actually dealing with, at most, two species instead of the three which he described, is not to be lost sight of. Taking Davis's descriptions as they stand, however, three distinct species are quite clearly characterized. Similarly, further research may well indicate that D. cella (Jameson) is a synonym of D. bidens (Jameson). Jameson found no trophozoites of the former species, spores of which are only differentiated from those of D. bidens by the somewhat greater size of the central chamber (9μ by 8μ to 13μ by 10μ, as compared with 8μ by 6.5μ to 11.5μ by 9μ), the greater length of the appendages (25μ to 35μ, as compared with 6μ to 10μ) and the less strongly curved sutural line. All these differences might well be correlated with increasing maturity, as in D. diplocrepis. All six species of Davisia are disporous, D. spinosa being monosporous as well, D. diplocrepis being disporous and polysporous, and the remaining four species being disporous only, in so far as they are known from very limited material. D. opacita and D. bidens have short lateral appendages, strongly recurved towards the postcapsular side of the spore. They share this character with certain of the immature spores of D. diplocrepis. Should a continuous series through such forms to arc-shaped spores with relatively longer appendages be found in some future investigation, D. brachiophora may prove to be the mature phase of D. opacita (and possibly spores of both these species may prove to be immature phases of D. spinosa), while D. cella may prove to be the mature phase of D. bidens. However, the former eventuality is not so likely as the latter, for spores of D. opacita differ from those of all other known members of the genus in having a prominent sutural ridge. Accepting the five earlier species of Davisia as they stand at present, the affinities of D. diplocrepis lie with D. spinosa, D. brachiophora and D. cella. The spores of these species have an oval central chamber, which measures about 13μ by 7μ in D. spinosa, from 9μ to 11μ by about 9μ in D. brachiophora and 13μ by 10μ to 9μ by 8μ in D. cella, as compared with 12.1μ to 14.0μ by 9.0μ to 12.0μ in D. diplocrepis. Although the central chambers are of roughly comparable size in all four species, the lateral appendages are longer in the earlier described species—–approx. 33.5μ in D. spinosa, 18μ to 22μ in D. brachiophora and 25μ to 35μ in D. cella as compared with 10.0μ to 14.2μ in D. diplocrepis. On the basis of these measurements D. spinosa and D. cella bear close resemblances to one another, as do D. brachiophora and D. diplocrepis. However, although Davis (1917) and Jameson (1929) failed to state the breadth of the lateral

appendages in their material, a comparison of Davis's Pl. XXI, fig. 72 with Jameson's Pl. IV, fig. 13 indicates that these appendages are relatively much narrower in D. spinosa than in D. cella. The lateral appendages of the spores of D. diplocrepis are appreciably shorter than those of D. brachiophora, and the central chamber is more rounded in the latter species. Otherwise these two species have much in common. The polar capsules are of similar size (3.5μ diam. in the latter species and 3.7μ by 3.4μ in the former), and (from Davis's Pl. XXIII, fig. 113) the lateral appendages of D. brachiophora may end in a narrowed projection as in D. diplocrepis. The myxosporidian under discussion differs from the five previously described species now transferred to Davisia in that the lateral appendages of its spore are characterized by having an inflated base, which is particularly evident in the immature state. Although it is felt that further investigation of the earlier species may prove that other distinguishing features both among themselves and between themselves and the present species are not so significant as they seem at present while only brief and arbitrary descriptions based on limited material are available for comparison, it is considered best for purposes of reference to describe the New Zealand parasite as new. It is hence designated Davisia diplocrepis n.sp. Leptotheca subelegans n.sp. (Pl. 8, figs. 51–58). Trophozoites and mature spores were abundant in the bile of all three fishes found to be parasitized by this myxosporidian, 57 mm. and 72 mm. examples of Diplocrepis puniceus and a 45 mm. Callogobius atratus. Both the clingfishes were collected at Point Jerningham (3.9.50 and 26.2.51), while the goby was obtained at Island Bay, Wellington (14.7.51). The infected gall bladders were hypertrophied, the bile being distinctly sticky to the touch and of a light yellowish colour instead of the normal bright green. Trophozoites. Both rounded and clubs-shaped forms occur, the rounded forms ranging from about 10μ to 28μ in diameter. Elongate forms (Fig. 51) have a terminal mass of cytoplasm containing many greenish refractive spherules, and a long, attenuated extension which is occasionally (Fig. 51) branched laterally and terminally. Typical elongate trophozoites appear rather like African knob-kerries in shape. They range in length from 25μ to 85μ. An ectoplasmic zone is rarely apparent, and none of the examples seen have been observed to display the slightest locomotor activity. The species is disporous, sporulating trophozoites (Fig. 52) being of approximately ovoidal shape and averaging some 30μ by 25μ at their greatest diameters. The finely granular endoplasm contains numerous refractive spherules. These are particularly abundant peripherally, are of a greenish colour, and measure from 0.7μ to 1.0μ in diameter. Spores. In the mature state, these are elliptical (Fig. 57) to egg-shaped (Fig. 58) in front view. Those from Diplocrepis range from 16.2μ to 20.3μ in breadth (av. 17.3μ) and from 8.3μ to 10.7μ in height (av. 9.0μ); while those from Callogobius range from 19.6μ to 26.2μ in breadth (av. 22.6μ) and from 8.6μ to 12.5μ in height (av. 11.6μ) (50 examples of each measured). From the polar aspect they are seen to be distinctly compressed and more or less arched, and to have lateral swellings at the extremities of the concave face of the arch. Because of the peculiar shape of the spore, almost all of those present in cover

slip preparations are so oriented as to lie with the front surface uppermost. The thickness of the spore is substantially less than the height, ranging from 5μ to 6μ in seven examples measured—–these seven spores were the only ones of hundreds seen which were lying still enough in the liquid medium to be measured from the polar aspect (although many other spores rolling over and over in fresh preparations, in consequence of the flowing of the medium beneath the cover slip, were momentarily glimpsed from this aspect). The lateral limits of the shell valves are smoothly rounded, the valve walls being quite thin. One of the valves is often appreciably larger than the other (Fig. 54). The inequality may be so marked as to give the spore, as seen in front view, an egg–shaped appearance (Fig. 58). The sutural line is clearly visible by ordinary bright field illumination. This line traverses the spore transversely (Figs. 55, 56) or obliquely (Figs. 53, 54, 57), the valves uniting at a slight ridge which may be just sufficiently raised to break the even contour of the spore (Figs. 53, 55–58). In developing spores (Fig. 52) the sporoplasm, although large, does not fill the entire space available to it, but in mature examples it almost or quite fills the spore cavity (Figs. 55–58). The sporoplasm is homogeneous and binucleate. The polar capsules are spherical or subspherical, the side in contact with the sutural line often appearing somewhat flattened. Their diameter ranges from 2.1μ to 2.9μ in both hosts (av. 2.5μ). The polar filaments, which are easily extruded by means of 5% phenol, range from 64.5μ to 81.5μ in length. Mature spores frequently remain for a time associated in pairs (Figs. 53, 54), joined to one another by means of a residuum of protoplasm derived from the trophozoite. Discussion. The only member of this genus previously recorded as having spores displaying lateral swellings in polar view is Leptotheca elegans, described by Noble (1938) from a Californian kelpfish, Gibbonsia elegans elegans (Cooper). Noble later (1939, 1941) reported finding L. elegans in four additional Californian intertidal pool fishes belonging to the Families Cottidae, Clinidae and Gobiidae. Both rounded and club-shaped trophozoites, the former of equivalent size to those of the species under discussion, occur in L. elegans. Club-shaped trophozoites of the latter species differ from those of the New Zealand one in that they are not known to exceed 65μ in length. Noble (1938) observed sluggish movement in his trophozoites from Gibbonsia elegans elegans, but (1941) stated that those from Artedius lateralis (Girard) show no sign of movement (his measurements for the spore of L. elegans from the latter host, 22μ by 7μ, suggest that he may here have been dealing with a species of Ceratomyxa). The trophozoites of the New Zealand parasite differ from those of L. elegans from the type host in displaying no locomotor activity whatsoever. Both species have similarly shaped spores and polar capsules, and the size of L. elegans spores from Gibbonsia (17μ by 9μ) is equivalent to that of spores of the present species from Diplocrepis. As far as can be gathered from Noble's (1938) description, the only real point of difference between the two species lies in the character of the sutural line. This is indistinct in L. elegans spores, but in spores of the present species there is a distinct and slightly raised sutural ridge. Were it not for this latter point, I would have no hesitation in identifying the New Zealand Leptotheca as L. elegans. Disregarding the difference in zoogeographical dis-

tribution of these parasites, the intertidal pool habitat of the Californian host is very similar to that of the New Zealand ones. One of the hosts for L. elegans, Typhlogobius californiensis Steindachner, belongs to the same family (Gobiidae) as does the New Zealand Callogobius atratus. Furthermore, the habitat of T. californiensis is precisely the same as that of the New Zealand clingfish Diplocrepis puniceus, for according to Jordan and Evermann (1898) the former species has a ventral disc and lives attached to the underside of rocks in shallow water or surf. L. obovalis Fantham, 1930, parasitizes various South African fishes of the pelagic and intertidal zones, the hosts of this myxosporidian including a blenny and three species of kelpfishes (Fantham, 1919, 1930). This species, like the New Zealand one and L. elegans, is disporous. Its club-shaped trophozoites range from 30μ to 60μ in length, and its spores as observed in front view are ovoidal or somewhat arched. Spores from Blennius cornutus (L.) measure 12μ to 18μ by 6μ to 9μ when fixed and stained, and thus enter the size range of those of L. elegans and of the species under discussion. Fantham failed to describe the appearance of L. obovalis in polar view. It would be of considerable interest to ascertain from the study of fresh material of this species whether or not its spores have lateral extensions as do the Californian and New Zealand parasites compared herein. If it should prove that the shape of L. obovalis spores entails their being seen most frequently in front view, it might eventuate that all three species are very closely related to one another if not conspecific. It is considered inadvisable to follow Kudo (1933) in transferring L. obovalis to Ceratomyxa. The breadth of the spore of Fantham's species from the type host, Blennius cornutus (L.) is only equal to twice the height; and Kudo's emended generic diagnosis of Ceratomyxa requires the breadth of the spore to be more than twice the height. However, it is possible that some of the spores referred by Fantham to L. obovalis did not in fact belong to this species; for those from kelpfishes (1919), (Dentex) = Argyrozona argyrozona (Val.) (1919, 1930), and Lepidopus caudatus (Euphrasen) (1930) were all three times as broad as they were high. Fantham may thus have been dealing with a species of Ceratomyxa which he failed to recognise as such and which remains undescribed, as well as his undoubted Leptotheca. The same remarks also apply to the 22μ by 7μ spores which Noble (1941) found in the gall bladder of the North American sculpin Artedius lateralis (Girard) and identified as his L. elegans. Noble failed to state whether or not L. elegans induces any changes in the gall bladder or bile of its hosts. Fantham (1930) noted that the bile of Blennius cornutus parasitized by L. obovalis remained light green and clear, although that of his earlier hosts (1919) was slightly turbid and yellowish green or yellow in colour. As already stated, the New Zealand species produces marked alterations in the gall bladder and bile of its hosts. On the available data there is no alternative to considering the Leptotheca of D. puniceus and C. atratus as a distinct species, while recognizing its very close affinities with L. elegans and possibly with L. obovalis as well. To emphasize its relationships with the Californian parasite, it is hence described as Leptotheca subelegans n.sp.

Myxidium incurvatum Thélohan, 1892 (Pl. 9, figs. 62–71). Trophozoites and spores of this species were abundant in the gall bladder of three examples of Diplocrepis puniceus (50–69 mm.) from Point Jerningham (19.9.50) and of a 46 mm. example of Notoclinus fenestratus collected at Lyall Bay, Wellington, on 15th October, 1950. A 36 mm. Oliverichtus melobesia collected at Long Beach, Russell, on 6th February, 1951, was also heavily infected with M. incurvatum; as were 14 of 28 examples of the same clingfish (28 mm.–39 mm.) obtained at Princess Bay and Island Bay, Wellington, from July to September, 1951. Scanty spores of this species were observed in the bile of three kelpfishes (Acanthoclinus quadridactylus) from Point Jeningham (14.8.50) and Island Bay (9.9.51). Trophozoites. As seen in a drop of fresh bile these are irregularly rounded, and make slow, progressional movements by means of one or two broad and blunt lobopodia. The endoplasm is hyaline and finely granular, and contains numerous refractive spherules of a yellowish green colour. Larger and darker bodies, representing the initial stages of sporoblasts, can often be distinguished. The trophozoite is not surrounded by a clearly demarcated zone of ectoplasm, although the lobopodia present a less granular appearance than does the body proper. Trophozoites very soon become rounded and motionless in cover slip preparations, exhibiting from one to several large, clear vacuoles (Fig. 62). The majority of the examples seen ranged from 15μ to 20μ in diameter, although a few larger forms of up to 28.5μ were noticed. Monosporous (Fig. 63) and disporous trophozoites were observed. Young spores have round polar capsules and are straight in both front and side views, as Georgévitch (1916) indicated. Spores. These were studied in both fresh and Schaudinn/iron baematoxylin preparations, all measurements being made from fresh material. In front view they are spindle-shaped (Figs. 66, 68), in side view twisted into the shape of a thick figure S (Figs. 69–71). The shell valves are not striated, and their extremities are rounded. It is unusually difficult to make out the sutural line, which is sometimes distinguishable in side view running along the proximal sides of the polar capsules (Figs. 69–71). The measurements of 50 spores from N. fenestratus, and of the same number from D. puniceus and O. melobesia, are as follows:— N. fenestratus Length 11.5μ to 13.0μ (av. 12.3μ) Breadth 5.1μ to 6.2μ (av. 5.8μ) O. melobesia Length 8.1μ to 11.7μ (av. 10.4μ) Breadth 4.9μ to 6.9μ (av. 5.7μ) D. puniceus Length 8.5μ to 9.7μ (av. 8.8μ) Breadth 5.1μ to 5.9μ (av. 5.4μ) The polar capsules of the examples from N. fenestratus ranged from 3.7μ to 4.5μ (av. 4.0μ by 1.4μ to 2.2μ (av. 1.8μ), those of the spores from O. melobesia, 3.2μ to 4.4μ (av. 3.8μ) by 1.4μ to 2.2μ (av. 1.7μ), and those of the spores from D. puniceus, 34μ to 3.7μ (av. 3.5μ) by 1.2μ to 1.7μ (av. 1.5μ). Polar filaments, extruded by means of 5% phenol, ranged in length from 42.4μ to 73.4μ (av. 64.9μ). HC1 gave much less satisfactory extrusion, the filaments remaining partly coiled and measuring from 15μ to 30μ in length. As seen in side view, the axes of the polar capsules may be parallel to one another (Fig. 71), but more frequently are not (Fig. 70). Due to the curvature

of the spore, the openings of these capsules are on opposite sides of the shell. The lightly granular sporoplasm, which does not completely fill the spore cavity, always contains two nuclei. These may be distinguished in fresh preparations under dark field illumination, as may the capsulogenous nuclei (Figs. 68–71) and one or both of the shell valve nuclei. In both fresh and Schaudinn/iron haematoxylin preparations, the undischarged polar filament is clearly visible, from five to eight closely wound coils being apparent. Discussion. M. incurvatum has been reported from about twenty different species of fish, from both marine and freshwater habitats. It is of cosmopolitan occurrence, having previously been recorded from various European waters and from the Atlantic and Pacific coasts of North America. Thélohan (1892, 1895) described this myxosporidian from six different North Sea hosts, including the blenny Blennius pholis L. The trophozoites from New Zealand fishes agree with his description in all essentials. Most authors have remarked on the presence of refractive granules in the endoplasm, although Georgévitch (1916) failed to observe these. The disporous condition is the typical one, although the monosporous one was recorded by Parisi (1912), Davis (1917) and Jameson (1929). Georgévitch and Dunkerly (1920, 1925) observed polysporous individuals. Spores of M. incurvatum are somewhat variable in size Their dimensions, according to various authors, are as follows:— Europe Thélohan (1895) 8μ to 9μ by 4μ to 5μ Parisi (1912) 10μ to 12μ by 5μ to 6μ North America Davis (1917) 8μ to 9μ by 5μ to 6μ (living material) Jameson (1929) 9.4μ to 10.9μ by 4.8μ to 6μ (preserved material) Noble (1941) Av. 11μ by 4.5μ Slight differences have also been noted in the shape and degree of curvature of the spore. These, together with the size differences, possibly indicate the occurrence of host-determined varieties or races. Thus, regarding the New Zealand material, the spores from N. fenestratus are somewhat longer and relatively rather narrower than those from O. melobesia, while those from D. puniceus are smaller still. Similarly, the measurements given by Noble (1941) suggest that his M. incurvatum spores from Dialarchus snyderi Greeley were longer and relatively narrower than those described by Jameson (1929) from another Californian marine fish, Sebastodes caurinus Richardson. However, Noble failed to state whether his measurements were made from fresh or fixed spores, although his measurements of other myxosporidians in earlier papers (1938, 1939) were based on living material. If he followed the same practice in regard to this species, the disparity could be attributed to shrinkage, for Jameson's figures were derived from the measurement of fixed and stained spores. There is a further difference between the parasites from these two Californian hosts which is not attributable to technique, Jameson's trophozoites being monosporous and Noble's being disporous. Although there are thus grounds for supposing that host-determined varieties or races of M. incurvatum do exist, there is little evidence to indicate the occurrence of geographical variation. The overall size range of the spores from all

known hosts in Europe, North America and New Zealand is much the same in each of these localities, although the upper limit of that range appears, from the available information, to be rather higher in the latter country than in the other areas. There is considerable variation in the figures published by earlier authors for the size of the pyriform polar capsules and the length of the polar filament in M. incurvatum. Noble's (1941) average of 3.8μ by 1.8μ for the dimensions of the capsule comes closest to the present ones (4.0μ by 1.8μ and 3.8μ by 1.7μ). The length of the extruded filament was given as 12μ by Thélohan (1892), who stated that extrusion was difficult to produce. The same author (1895) gave this length as 10μ to 15μ, while Parisi (1912) gave it as 28μ. Davis (1917) stated that when extrusion was brought about by means of HC1, the filament remained tightly coiled. This observation was confirmed in the New Zealand material, partially extruded and still coiled filaments measuring from 15μ to 30μ in length. Employment of Bond's (1938) technique utilizing 5% phenol gave much more satisfactory extrusion, the average length of the filament (64.9μ) being more than twice as great as the maximum derived from HC1-treated material. Earlier authors neglected to furnish details of the number of coils in the undischarged polar filament. The following figures are derived from their illustrations:— 7 Parisi (1912) 5 Davis (1917) 5–7 Dunkerly (1925) These figures compare favourably with those from my material (5–8). The spores of a number of other species of Myxidium are of broadly similar morphology to those of M. incurvatum. These species either differ from the latter one in their vegetative stages, or else have very much larger spores. The Myxidium of N. fenestratus and O. melobesia so closely resembles M. incurvatum as described by earlier investigators that it is considered, bearing in mind the wide geographical and host range of this species, that its description as new would be unwarranted. Sphaeromyxa tripterygii n.sp. (Pl. 10, figs. 72–74). This myxosporidian was first recorded from the gall bladder of a 72 mm. example of Tripterygion varium collected at St. Heliers, Auckland, on 12th January, 1951. Spores only were present. Spores and a trophozoite were found in the gall bladder of a 64 mm. example of the same host collected at Russell. North Auckland, on 7th February, 1951. Trophozoites alone were recorded from three examples of T. medium (67 mm.–74 mm.) inhabiting the same rock pool as this last blenny. Only one trophozoite was present in each of the infected gall bladders, while very few spores were found, six being present in one instance and five in the other. Trophozoites. Far exceeding the gall bladder itself in length and breadth, each trophozoite is folded upon itself into a hollow ball closely applied to the inner wall of the bladder, through which it is clearly visible in situ. Unfolded and flattened trophozoites measure up to 7 mm. by 5 mm. in the fresh condition. The example illustrated (Fig. 72) was fixed in Schaudinn's fluid and stained with Delafield's haematoxylin. It measures 5.47 mm. by 4.03 mm., and is thin, opaque and bluish white in colour.

Spores. The few available for study are in Schaudinn-fixed smears stained with Heidenhain's iron haematoxylin. They are arcuate, resembling boomerangs with truncated ends (Figs. 73, 74). The measurements of the 11 examples discovered are as follows:—Length along the median line, 17.2μ to 21.1μ (av. 18.9μ); length along the inner side of the arch, 16.4μ to 20.0μ (av. 17.9μ); length along the more strongly curved outer side of the arch, 18.6μ to 22.3μ (av. 20.2μ); greatest breadth 3.4μ to 3.8μ (av. 3.5μ). The pyriform polar capsules, located one at each end of the spore, measure from 46μ by 1.4μ to 5.1μ by 1.4μ (av. 48.μ by 1.4μ). A thick polar filament (not yet seen in the discharged state) is coiled lengthwise within the capsule. In the younger spores (Fig. 73) the rather granular sporoplasm contains two nuclei, and prominent polar capsule nucler are present at the inner ends of the capsules. These latter nuclei are not apparent in the older spores (Fig. 74), which also have only one sporoplasm nucleus, a synkaryon formed at autogamy. Discussion Since 1933, when Kudo listed 10 species of Sphaeromyxa, three further species have been described. The inclusion of the present species brings the total up to 14. All of these inhabit the gall bladder of marine fishes. Their distinguishing characters are summarized in Table VII. It is apparent from Table VII that the spores of Sphaeromyxa fall into two broad groups:— 1. The balbianii group Straight or slightly curved fusiform or ovoid spores having ovoid polar capsules (S. balbianii, S. gasterostei, S. longa, S. reinhardti, S. gibbonsia, S. ovula and S. lateralis). 2. The incuriata group Arcuate spores having pyriform polar capsules (S. incurvata, S. sabrazesi, S. hellandi, S. exneri, S. arcuata, S. curvula and S. tripterygii n. sp.) The known vegetative stages of the species of this genus, with the exception of S. reinhardti, are of very large size. Within the incurvata group, the trophozoites of S. exneri, S. arcuata and S. curvula have not yet been described, and measurements are not available for those of S. hellandi. The thin and leaf-like trophozoite of the species under discussion is comparable in size and form with that of S. incurvata, and is appreciably larger than that of S. sabrazesi, the diameter of which usually approximates 2 mm. (although Schröder, quoted by Kudo, 1920, recorded examples up to 5 mm. in diameter). The spores of S. incurvata and of S. exneri are very much larger than are those of the present species. They also differ from these in having actually and relatively larger polar capsules. The length of a single polar capsule expressed as a percentage of that of the spore itself approximates 40% in S. incurvata and S. exneri, but only 25% in S. tripterygii n. sp. S. sabrazesi, S. hellandi, S. arcuata and S. curvula have spores which, allowing for differences in methods of examination and fixation, are comparable in length with those of the species under discussion. S. arcuata may, however, be eliminated from further comparison. From Fantham's (1930) figures, the sporeb of this species are actually and relatively much broader, and have longer and relatively more slender polar capsules than those of S. tripteryii n.sp. The breadth of a polar capsule expressed as a percentage of its length is approximately 20% in the former species, but 30% in the latter.

Table VII. Summary of Data on the Species of Sphaeromyxa. Species. Trophozoite. Spore. Polar Capsules. Shape. size. Shape. Size. 1. S. balbianii Thélohan Up to 3–4 mm. in diameter Straight, fusiform, ends truncate 15μ–20μ by 5μ–6μ Ovoid 7μ by 4.7μ 2. S. incurvata Doflein Up to 5–7 mm. in diameter Arcuate, ends truncate 30μ–35μ (inner side of arch) by 8μ Pyriform 12μ–15μ by 4μ–5μ 3. S. sabrazesi Lav. & Mes. Up to 5 mm. in diameter Arcuate, ends truncate 22μ–28μ by 3μ–4.3μ Pyriform 8μ–10μ by 2μ–3μ 4. S. hellandi Auerbach Large rounded disc Arcuate, ends truncate 20.8μ–26μ by 5.4μ Pyriform 10μ–10.8μ in length 5. S. exneri Awerinzew — Arcuate, ends slightly tapering 75μ–80μ by 18μ–20μ Pyriform 30μ–35μ in length 6. S. gasterostei Georgévitch Large plasmodia Very slightly curved, fusiform, ends bluntly rounded 2 or 3 times larger than in S. balbianii Elongate–ovoid 2 or 3 times larger than in S. balbianii 7. S. longa Dunkerly — Slightly curved, fusiform, ends truncate 20μ by 5μ Ovoid — 8. S. reinhardti Jameson Very small Straight or slightly curved, fusiform, ends truncate 21.25μ–23.3μ by 3.75μ–5μ Ovoid — 9. S. arcuata Fantham — Arcuate, ends bluntly rounded 22μ–27μ by 6μ–8μ Pyriform 7μ–10μ by 1.5μ–2μ 10. S. curvula Fantham — Arcuate, ends bluntly rounded 19μ–22μ by 4μ–6μ Pyriform 7μ–9μ by 2μ–3μ 11. S. gibbonsia Noble Up to 2 mm. in diameter Slightly curved, elongate, ends rounded 27μ by 5.2μ Ovoid 10μ by 4μ 12. S. ovula Noble Up to 800μ in diameter Straight, elongate-oval, ends rounded 14μ by 4.3μ Ovoid 5μ by 3μ 13. S. lateralis Noble Up to 1.5 mm. in length Slightly curved, oval, ends rounded 26μ by 8μ Ovoid 8.6μ by 6.3μ 14. S. tripterygii n.sp. Up to 7 mm by 5 mm. Arcuate, ends truncate 17.2μ–21.1μ by 3.5μ Pyriform 4.6μ–5.1μ by 1.4μ

In S. sabrazesi, S. hellandi and S. curvula, the polar capsules, while being of much greater bulk than those of the New Zealand parasite, are of equivalent length-breadth ratio. The spores themselves are relatively broader in S. hellandi and S. curvula than in S. sabrazesi and S. tripterygii n.sp. The breadth of those of the first two species, expressed as a percentage of the length, is approximately 25%, while the equivalent figures for the last two species are 15% and 20%. Spores of S. tripterygii n.sp. are still further differentiated from those of these other three species by the fact that the length of a single polar capsule expressed as a percentage of that of the spore is only 25%, as compared with 35% for S. sabrazesi, 45% for S. hellandi and 40% for S. curvula. Longitudinal striations have been seen on the shell valves of spores of three species of the incurvata group:—S. sabrazesi, S. arcuata and S. curvula. Such striations are often evident only in living material under the most favourable lighting conditions, and the fact that they have not been reported from other members of this group is not necessarily indicative of their absence. As far as can be gathered from the figures published by earlier investigators, the outer side of the arch formed by spores of members of the incurvata group is usually more strongly curved than the inner one. Three species besides the present one—S. hellandi, S. exneri and S. arcuata—have the outer curvature of the arch sufficiently marked for the spore to appear humped or boomerang–like. It is concluded that the Sphaeromyxa of New Zealand blennies, having affinities with other members of the incurvata group but differing from them in detail as described above, is sufficiently distinctive to warrant description as a new species. The type slide of Sphaeromyxa tripterygii n.sp. has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z104). Zschokkella sp. (Pl. 10, fig. 75). A single sporulating trophozoite was found in the urinary bladder of a 50 mm. example of Tripterygion varium collected at Moa Point, Wellington, on 15th March, 1951. No other life history stages were present, and no further infections have been recorded. The following notes were made from a fresh cover slip preparation. The trophozoite (Fig. 75) which measures 18.4μ by 14.9μ, had rather granular cytoplasm containing 13 greenish refractive spherules averaging 0.7μ in diameter. An ectoplasmic zone could not be distinguished. Both of the young spores were reniform in shape, one of them being broader and more strongly curved than the other. Their greatest dimensions (length being measured along the median line) were 16.6μ by 5.4μ and 17.2μ by 7.4μ. The polar capsules, seen in anterior view, averaged 3.6μ in breadth. In each case the polar filament was wound in three coils Neither the sporoplasm nor the various nuclei of the spores could be made out. Discussion. The genus Zschokkella, in so far as at present known, comprises 14 species. Twelve of these have been described from the gall bladder, bile ducts, kidney or urinary bladder of marine and fresh water fishes, one from the gall bladder of a toad and one from that of a tortoise. In seven of the species the shell valves are striated, in the other seven they are smooth (Tripathi, 1948a). Striations could not be made out on the shell valves of the immature spores studied in the present instance. Other authors (e.g, Davis, 1917; Tripathi, 1948a) have drawn attention to the presence of large refractive granules in the trophozoites

of species of Zschokkella, Tripathi stating that these disappear in fixed and stained preparations. The affinities of the myxosporidian under discussion cannot be accurately assessed until more material is available. Two of the previously described members of the genus, Z. acheilognathi Kudo, 1916, and Z. russelli Tripathi, 1948, need not be considered further in this connection, for both of these have very large polysporous trophozoites. The spores of three of the other species from fish, Z. hildae Auerbach (data from Kudo, 1920), Z. rovignensis Nemeczek, 1922 and Z. salvelini Fantham et al., 1939, are very much larger than those of the present species. Some of the remaining species are of comparable size. Before useful comparisons can be drawn between these and the New Zealand parasite, mature spores of the latter will have to be obtained, for the shape of young myxosporidian spores often differs markedly from that of mature examples of the same species. It is possible that the present species will prove to be systematically close to Z. ovata (Dunkerly, 1920) of English rocklings. Its immature spores closely resemble the mature ones of Dunkerly's species in both size and shape. Myxosoma tripterygii n.sp (Pl. 10, figs. 76 and 77). This species is described from the subdermal connective tissue of the caudal peduncle of a 102 mm. example of Tripterygion varium, collected at Lyall Bay, Wellington, on 1st September, 1950. Cysts were not evident, only a few disporous trophozoites and mature spores being seen. The following measurements and observations were all made from fresh material. Young Spores (Fig. 76). Ovoidal or pyriform, conjoined in pairs by a small residuum of protoplasm. Two sporoplasm nucler and the capsulogenous nuclei are clearly visible. The spores themselves measure from 11.1μ to 12.9μ (av. 11.6μ) by 7.9μ to 9.0μ (av. 8.3μ) (14 measured). Their polar capsules are also ovoidal or pyriform, their dimensions ranging from 4.8μ to 5.8μ (av. 5.4μ) by 3.7μ to 4.0μ (av. 3.8μ). Mature Spores (Fig. 77). These are circular in front view and apparently markedly flattened in side view, the position adopted by all of them in the liquid medium allowing of their being studied in front view only. Their diameter range from 11.7μ to 12.4μ (av. 12.1μ) (10 examples measured). The sutural edge is 0.6μ thick, and has from eight to ten triangular folds. The polar capsules are pyriform, and have a prominent projection anteriorly. They are relatively very large, measuring from 6.8μ to 7.8μ (av. 7.3μ) by 4.0μ to 4.6μ (av. 4.5μ). The polar filament, which is coiled five or six times, occupies the central portion of the capsule only, the sides remaining empty. The binucleate sporoplasm is rather small, and tapers anteriorly to a slender tongue which is pushed between the polar capsules and reaches a point about half way along them. Systematic Position. Mature spores of many of the species of Myxosoma are ovoid or pyriform. The spores of the following species, however, are circular or almost so in front view. M. cerebralis Hofer, 1903. Data from Kudo, 1920. M. encephalinum (Mulsow, 1911) Kudo, 1933. Data from Kudo, 1920. M. dermatobium (Ishii, 1916) Kudo, 1933. Data from Kudo, 1920.

M. catostomi Kudo, 1926. M. cndovasa Davis, 1947. Spores of M. cerebralis and M. endovasa do not exhibit folding at the sutural edge, while the diameters of those of M. encephalinum and M. dermatobium are very much smaller than in the present species (5μ to 5.5μ and 6.3μ to 7μ respectively, as compared with 11.7μ to 12.4μ in the New Zealand parasite). M. catostomi resembles the species under discussion in being disporous, and its spores are of much the same order of size (13μ to 15μ by 10μ to 11.5μ, according to Kudo, 1926, and 10μ to 12.7μ by 7.6μ to 10μ, according to Fantham et al., 1939). However, the spores of M. catostomi are rarely quite circular in front view, and have relatively much narrower and smaller polar capsules than do those of the present species (3.6μ to 5.5μ by 1.3μ to 2.7μ, as compared with 6.8μ to 7.8μ by 4.0μ to 4.6μ). The hosts of these five species are all fresh–water fishes of Europe, North America or Japan. The Myxosoma of Tripterygion varium has spores which differ from those of previously described members of the genus as detailed above. Thus, despite the paucity of material available for study, it is considered that it should be described as new. It is accordingly designated Myxosoma tripterygii n.sp., having the characters detailed herein. Class Ciliata Perty Order Peritricha Stein Classification. The broad classification of this order adopted here is that of Kahl (1935), who established the two sub–orders Sessilia and Mobilia. A majority of the Sessilia spend their adult existence attached to submerged objects. Kahl divided this sub–order into two tribes, the Aloricata and the Loricata, the former tribe including five families all the members of which lack a lorica—Astylozoonidae Kahl, Ophrydiidae Kent, Epistylidae Kent, Vorticellidae Fromental and Scyphidiidae Kahl. The latter family includes those sessile peritrichs which lack a lorica, an anterior neck and a stalk, and which are attached directly to submerged objects in the adult state. This family comprises the genera Scyphidia Dujardin, Paravorticella Kahl, Glossatella Butschli, and Ellobiophyra Chatton and Lwoff; to which is now added Caliperia n.gen. Those peritrichs included in the sub–order Mobilia are free–swimming in the adult state, and have a highly developed skeletal complex at the aboral extremity. There is but one family, Urceolariidae Stein. I have followed Fauré–Fremiet (1943) and Tripathi (1948) in recognizing only two genera, Trichodina Ehrenberg and Urceolaria Lamarck, in this family. In Urceolaria (to a sub–genus of which Fauré–Fremiet relegated Leiotrocha Fabre–Domergue) the denticulate ring of the skeletal complex is made up of obliquely arranged simple denticles lacking radial processes In Trichodina each unit of the denticulate ring bears a ray on its inner side and a hook on its outer one. Members of Cyclochaeta Jackson, relegated to the status of a sub–genus of Trichodina by Fauré–Fremiet, have cirri at the aboral extremity Cirri are lacking in the sub–genera (Trichodina) Fauré–Fremiet and Vauchomia (Mueller). In Vauchomia, which Tripathi (1948) reduced to subgegeneric status, the adoral spiral makes more than two complete circuits of the body before descending into the vestibule, whereas in (Trichodnia) it makes less than two such circuits.

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). 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). Table VIII. Denticular Counts for Trichodina parabranchicola n.sp. Number of Denticles 16 17 18 19 20 21 22 23 24 25 26 Number of individuals 1 5 7 12 21 13 40 28 12 9 2 Percentage of Individuals 0.7 3.3 4.7 8.0 14.0 8.7 26.7 18.7 8.0 6.0 1.3 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

Text-Fig. 1. Biometrical separation of Trichodina parabranchicola and T. multidentis n. spp. Data obtained from random samples of 150 individuals of each of these species from heavy infestations on Acanthoclinus quadridactylus and Tripterygion varium respectively. The stippled vertical columns represent T. parabranchicola, the hollow ones T. multidentis and the cross-hatched ones overlapping data. The black spot on each vertical column indicates the average diameter of the denticulate ring for each denticular number of each of the two species, while the white number on each spot indicates the number of trichodinida from which this average was derived.

Text-Fig 2 Illustrating the increase in height of the denticular hooks of Trichodina parabranchicola and T. multidentis n. spp, with increasing ring diameter. The data was obtained by random sampling of these species from heavy infestations on Acanthoclinus quadridactylus and Tripterygion varium respectively. Individuals of all sizes other than those very recent products of binary fission in which the hooks of the parent denticles had not yet been absorbed, were measured. This degree of selection was adopted in order to illustrate the rapid initial rate of development of the hooks in T. multidentis—no examples of this species showing no trace of the parent denticles and having hooks of less than 3 microns in height were seen. In T. parabranchicola, on the other hand, individuals in which the parent denticles have just been absorbed have the hooks of the new ring far less strongly developed.

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):— T. pediculus Ehremberg, 1838 (16–26) T. steini Claparède and Lachmann, 1858 (21–26) *T. baltica Quennerstedt, 1869 (21) *T. scorpaenae (Robin), 1879 (22) *T. labrorum Chatton, 1910 (21) *T. fariai Da Cunha and Pinto, 1928 (24–28) *T. clini Fantham, 1930 (24) T. myakkae Mueller, 1937 (17–24) T. vallata Davis, 1947 (18–21) T. symmetrica Davis, 1947 (21–28) T. tumefaciens Davis, 1947 (19–26) T. bulbosa Davis, 1947 (19–24) T. bursiformis Davis, 1947 (24–27) T. branchicola Tripathi, 1948 (20–26) 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).

Table IX. Denticular Counts for Trichodina multidentis n.sp. Number of Denticles 23 24 25 26 27 28 29 30 31 32 33 Number of Individuals 2 6 7 8 5 16 4 7 7 12 8 Percentage of Individuals 1.3 4.0 4.7 5.3 3.3 10.7 2.7 4.7 4.7 8.0 5.3 Number of Denticles 34 35 36 37 38 39 40 41 42 43 44 45 Number of Individuals 18 7 11 10 6 7 2 3 1 2 0 1 Percentage of Individuals 12.0 4.7 7.3 6.7 4.0 4.7 1.3 2.0 0.7 1.3 0 0.7 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:— 27 (14) 28 (14) (Fig. 101) 29 (14) 30 (15) 32 (16) 33 (16) 37 (18) 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 Fulton, 1923 (26–36) *T. blennii Fantham, 1930 (24–32) *T. mugilis Fantham, 1930 (32) *T. chelidonichthyos Fantham, 1930 (30) T. okajimae Ibara, 1931 (34–38) T. truttae Mueller, 1937 (28–31) *T. spheroidesi Padnos and Nigrelli, 1942 (21–31) *T. halli Padnos and Nigrelli, 1942 (26–34) T. vesicularum Fauré-Fremiet, 1943 (21–33) *T. tenuidens Fauré–Fremiet, 1943 (27–37) T. discoidea Davis, 1947 (18–30) T. platyformis Davis, 1947 (26–35) T. fultoni Davis, 1947 (25–30) T. californica Davis, 1947 (25–32) T. ranae da Cunha, 1950 (23–31) 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

species have not raised the upper limit of the denticular number above 25, and as suggested by MacLennan (1939) Wallengren's trichodinid with 49 denticles probably belonged to another species as yet undescribed. T. (Vauchomia) renicola (Mueller) has 56 denticles, and T. (Vauchomia) nephritica (Mueller) has from 36 to 40 (Mueller, 1938). As the trichodinid discussed herein must be regarded as new, it is designated Trichodina (Trichodina) multidentis n.sp., in reference to the high upper limit of its denticular number. A slide designated as the type of T. multidentis n.sp. has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z22), and paratypes are in my own collection and in that of the Department of Zoology, Victoria University College. Class Suctoria Claparède and Lachmann Endosphaera engelmanni Entz, 1896 (Pl. 11, fig. 84). This hyperparasite was encountered on one occasion only, in five examples of Trichodina multidentis n.sp. from the gills of a 90 mm. Tripterygion medium collected at Red Rocks (on Cook Strait, near Wellington) on December 26, 1950. Endoparasitic suctorians were originally mistaken for reproductive phases of the host by many protozoologists, the “embryo hypothesis” first being opposed by Balbiani (Lynch and Noble, 1931). The genus Endosphaera was established by Engelmann (1876), and the present species was proposed (without figures or description) by Entz (1896). Lynch and Noble (1931) were the first to publish an adequate generic and specific diagnosis of Endosphaera: “Small internal parasites belonging to the family Acinetidae; body spherical, without tentacles at any stage in the life-history; no test, cup, or stalk; reproduction by the formation of endogenous embryos; free–swimming embryo with three equatorial bands of cilia.” E. engelmanni parasitizes various peritrichous ciliates in both marine and freshwater habitats. Thus Lynch and Noble (1931) obtained their material from Opisthonecta henneguyi Fauré–Fremiet from a freshwater pond at San Francisco, California, while Padnos (1939) and Padnos and Nigrelli (1942) recorded this same suctorian from Trichodina spheroidesi, a peritrich described by the latter authors from a marine fish Spheroides maculatus (Bloch and Schneider), from the coasts of New York and New Jersey. According to Lynch and Noble (1931), the parasitic stage of E. engelmanni is spherical and ranges from 15μ to 41μ in diameter. It is attached to the pellicle of the host, in the cytoplasm of which it is embedded, by a short birth canal. The organism has a spherical macronucleus, one micronucleus and one contractile vacuole. Only the parasitic stage was represented in the material from Trichodina multidentis n.sp As none of the organisms contained embrayos, and they were all at the same stage of development as that illustrated in Fig. 84, it is probable that the infestation was a recent one. E. engelmanni was not seen in the living state, my material being limited to four parasitized trichodinids in one haematoxylin–stained gill smear, and one in a second such smear from the same fish. All the trichodinids concerned were young ones in which the macronucleus was still rounded. Each of them contained a single Endosphaera only. Altogether

69 trichodinids were present in the two smears, only 7% of these being parasitized by the suctorian. Morphology. The body outline is round or oval (Fig. 84), the length ranging from 17.0μ to 22.9μ and the breadth from 17.0μ to 18.7μ. These figures are further evidence that the infestation was a recently acquired one, for they are close to the lower limit of the diameter of E. engelmanni (given by Lynch and Noble as 15μ to 41μ). According to these authors, parasites lacking embryos are circular or very slightly oval in outline, their diameter averaging 21μ (range 15μ to 32μ). In the New Zealand material the diameter of the coarsely granular macronucleus ranges from 7μ to 10μ, as compared with 9μ to 17μ in young examples of E. engelmanni from Opisthonecta in California. The micronucleus is composed of a spherical mass of chromatin some 1·5μ in diameter, within a delicate membrane forming a vesicle of from 2μ to 3μ in diameter. There is a peripheral zone of hyaline ectoplasm, and the endoplasm, which stains only faintly, is of somewhat granular structure. A contractile vacuole of from 2.5μ to 4.0μ in diameter is located towards the periphery of the endoplasm (Fig. 84). The birth canal, which Lynch and Noble found to be apparent only in profile in favourably oriented examples, was not seen in the very limited material available for study. Gönnert (1935) described a second species of Endosphaera, E. multifiliis. This species forms more embryos per individual than does E. engelmanni. A comparison of my material with E. multifiliis, in regard to this feature, could not, of course, be made. However, the latter species is stated to be ectoparasitic (on German marine suctorians and vorticellids), and is thus distinct from the one under consideration. In so far as the New Zealand parasite is as yet known it does not differ in any major feature from E. engelmanni, and is accordingly identified with that species. The slide on which the example illustrated in Fig. 84 is mounted, has been deposited in the collection of the Dominion Museum, Wellington (catalogue number Z23). Discussion And Conclusions Discounting Zschokkella and Myxosoma tripterygii, only solitary infestations with each of these being noted, the only protozoans recorded from but a single host were Haemogregarina (Hepatozoon?) acanthoclini, Davisia diplocrepis and Scyphidia acanthoclini. Three species were each recorded from two hosts occupying the same environmental niche. Thus fishes belonging to two different families, the goby, Callogobius atratus, and the clingfish, Diplocrepis puniceus, both collected from the underside of stones, were parasitized by Leptotheca subelegans. The few examples of the blenny, Ericentrus rubrus, from which Caliperia longipes was collected, were found beneath stones in company with clingfishes (Oliverichtus melobesia) heavily infested with this ciliate. Sphaeromyxa tripterygii parasitizes two blennies, Tripterygion varium and T. medium, which very commonly occur together in intertidal pools. Trypanosoma tripterygium was found in the blood of three blennies, E. rubrus, T. varium and T. medium. The latter two fishes move freely about in intertidal pools, but E. rubrus is of cryptic habit.

Myxidium incurvatum was recorded from four hosts, all of cryptic habit, the clingfishes, Diplocrepis puniceus and Oliverichtus melobesia, the blenny, Notoclinus fenestratus, and the kelpfish, Acanthoclinus quadridactylus. Haemogregarina bigemina was recorded from five hosts, one clingfish and four blennies. Three of these (O. melobesia, E. rubrus and N. fenestratus) are of cryptic habit, while two (T. varium and T. medium) customarily browse among coralline seaweeds in the unsheltered parts of pools. All the protozoans considered thus far are endoparasites, with the exception of Scyphidia acanthoclini and Caliperia longipes. These latter species are ectocommensal, drawing their nourishment not from the host but from microorganisms in the water flowing through the gills of the host. It is of interest that they appear restricted to fishes spending the greater part of their time feeding among algae and under stones, for most of the known hosts of members of the family Seyphidiidae are molluscs, animals with which the fishes concerned enter into more intimate association than do those blennies having no adaptations to a cryptic existence. Trichodina multidentis was recorded from seven hosts, and T. parabranchicola from eight. These species, like the scyphidians referred to above, are ectocommensal, but unlike the latter organisms their attachment to the host is of a temporary though recurrent nature. Their host specificity is thus understandably less marked than is that of any of the other protozoans discussed. Each of these trichodinids, however, exhibits a general preference for a group of hosts. T. parabranchicola is, in nature, the dominant species (and frequently the only trichodinid present) on the clingfishes and kelpfishes; while T. multidentis is always the dominant species on the non–cryptic blennies (T. varium and T. medium) and is usually the dominant one on the cryptic E. rubrus as well. Both these trichodinids readily move from host to host when infested fishes from the two main niches, Diplocrepis and Tripterygion for example, are confined together in a small aquarium. The question of the rigidity of host specificity is one that can, of course, only be satisfactorily answered by cross–infection experiments. Such experiments, in so far as the haematozoa are concerned, must await the discovery of invertebrate hosts. Infection with coelozoic Myxosporidia is usually brought about by the ingestion of spores, and cross–infection experiments with these parasites would be somewhat easier to conduct. Scyphidians infest new hosts by means of free swimming telotrochs, while trichodinids move freely from host to host in the adult state. At least three of the protozoans recorded herein belong to species occurring in other parts of the world. These are Haemogregarina bigemina Laveran and Mesnil, Myxidium incurvatum Thélohan and Endosphaera engelmanni Entz. H. bigemina parasitizes two European blennies of similar habit to Tripterygion, Blennius pholis L. (Laveran and Mesnil, 1901, 1902; Neumann, 1909; Henry, 1913) and Blennius montagui Fleming (Neumann, 1909). It is also found in the European Blennius gattorugine Bloch (Laveran and Mesnil, 1901, 1902), a fish more often dwelling in deep water than in the intertidal zone; and in two viviparous blennies, fishes relatively rarely found in rock–pools, the European Zoarces viviparus L. (Bentham, 1917) and Zoarces anguillaris Peck, from which Fantham et al. (1942) recorded it on the Canadian Atlantic coast. Species systematically

close to H. bigemina have been described from blennies of the intertidal zone in Italy (H. londoni Kohl–Yakimoff and Yakimoff, 1915), South Africa (H. fragilis Fantham, 1930) and Fiji (H. salariasi Laird, 1951). Numerous other species of Haemogregarina have been described from European intertidal zone fishes of the genera Gobius, Cottus and Scorpaena (Brumpt and Lebailly, 1904; Neumann, 1909; Kohl–Yakimoff and Yakimoff, 1915). Some of these have been described as new largely on the basis of their occurrence in a new host, and there is little doubt but that considerable synonymy exists. From the scanty information supplied by Brumpt and Lebailly (1904), for example, it is quite impossible to differentiate their H. gobii from H. bigemina. The host of H. gobii, Gobius niger L., is found in intertidal pools in association with Blennius pholis and B. montagui. Cross infections with H. bigemina—now known from hosts of three different families, the Blenniidae, Zoarcidae and Gobiesocidae—might thus take place between these blennies and the goby, G. niger, in nature. Myxidium incurvatum has been reported from three blennies, Blennius pholis, L. (Thélohan, 1892, 1895; Dunkerly, 1920), Blennius ocellaris L. (Dunkerly, 1920) and Blennius montagui Fleming (Dunkerly, 1925), a sculpin (Fam. Cottidae), Cottus scorpius L. (Awerinzew, 1911), and a rockling (Fam. Gadidae), (Motella tricirrata) = Gaidopsaurus tricirratus (Bloch) (Thélohan, 1892), in European intertidal pools. This species has also been recorded by Noble (1941) from a tide pool sculpin in California, (Dialarchus) = Oligocottus snyderi Greeley. A number of coastal fishes which occasionally become trapped in tide pools are also hosts for M. incurvatum. These include European pipe–fishes (Fam. Sygnathidae) (Thélohan, 1892, 1895; Parisi, 1912; Georgévitch, 1916) and the scorpaenid, Scorpaena scrofa L. (Thélohan, 1895), also three Californian scorpaenids (Jameson, 1929, 1931). Littoral but not intertidal zone hosts for M. incurvatum include European turbots and sand dabs (Dunkerly, 1920) and a killfish, Fundulus majalis (Walbaum) (Fam. Poeciliidae), from the Atlantic coast of North America (Davis, 1917); while pelagic ones include Callionymus lyra L. (Fam. Callionymidae) and Trachinus draco L. (Fam. Trachinidae) in Europe (Thélohan, 1892, 1895; Dunkerly, 1920), and the mullet, Mugil cephalus L., from the Atlantic coast of North America (Davis, 1917). The previous record of Endosphaera engelmanni from a marine fish trichodinid (Padnos, 1939) concerns Trichodina spheroidesi Padnos and Nigrelli, 1942, which infests Spheroides maculatus (Bloch and Schneider) on the Atlantic coast of North America. S. maculatus, a swellfish (Fam. Tetraodontidae), occurs on sandy shores but does not remain in the intertidal zone. In the cases of these three cosmopolitan protozoans, then, the evidence indicates a decided lack of host specificity. That these parasites, or at all events H. bigemina and M. incurvatum, should infest intertidal pool fishes in Europe and at the antipodes, might be taken as indicating that the ancestral stocks of the modern hosts were in remote times already infested with hereditarily stable protozoans; and that any modifications which have since taken place in these parasites have been of a physiological nature only and unassociated with morphological change. As both H. bigemina and M. incurvatum have hosts in the littoral and pelagic zones as well as in the intertidal one, however, there is also a possibility that these species are in reality continuous in their distribution, parasitizing oceanic hosts from which they have not, as yet, been reported.

Several of the protozoans described as new herein are systematically close to species already known from intertidal zone fishes in other parts of the world. Thus species of Sphaeromyxa belonging to the incurvata group have been described from such fishes in Europe—Doflein (1898) described S. incurvata from the gall bladder of Blennius ocellaris L., while Georgévitch (1916) recorded S. sabrazesi Laveran and Mesnil from the rockling (Motella tricirrata) = Gaidopsaurus tricirratus (Bloch). Leptotheca subelegans has close affinities with L. elegans Noble, a gobiid host (Typhlogobius californiensis Steindachner) of which occupies precisely the same habitat as does Diplocrepis puniceus, a clingfish host of the New Zealand parasite. L. subelegans also has much in common with L. obovalis, a South African species the hosts of which include blennies and kelpfishes (Fantham, 1919, 1930). In this connection, it is of interest that Fantham's other hosts for L. obovalis included the littoral and pelagic frost fish, Lepidopus caudatus (Euphrasen), a species also occurring in New Zealand seas. Both the New Zealand trichodinids have affinities with species from intertidal zone fishes of Europe, T. parabranchicola with T. branchicola, described by Tripathi (1948) from English and Mediterranean rocklings, and from English sculpins, sticklebacks and blennies, and T. multidentis with T. tenuidens, described by Fauré–Fremiet (1943) from sticklebacks in France. All the previously described myxosporidians referable to Davisia n.gen. occur in North American seas. Both the North American hosts for myxosporidians of this genus are coastal fishes, one of them, Porichthys notatus Girard (Fam. Batrachoididae) being of rather similar habits to the New Zealand host of D. diplocrepis, Diplocrepis puniceus, occurring beneath rocks in shoreward areas. The most distinctive element among the New Zealand protozoans recorded herein, is Caliperia longipes n.gen., n.sp. Other species not having close affinities with ones already described from intertidal zone hosts elsewhere are Trypanosoma tripterygium, Haemogregarina acanthoclini, Myxosoma tripterygii and Scyphidia acanthoclini Representatives of all the protozoan genera recorded herein, with the exception of Caliperia n.gen., have been recorded from intertidal zone fishes in other parts of the world. With the same exception, representatives of all of them but Scyphidia have been recorded from pelagic fishes as well. Most members of the Family Scyphidiidae occur on freshwater hosts or on marine invertebrates of the intertidal zone. The same remarks apply also to the Family Vorticellidae, of which the only representatives known from marine fishes are Vorticella striata Dujardin and Zoothamnium duplicatum Kahl. Precht (1935) recorded both these species at Kiel, the former from the snout of the stickleback, Gasterosteus aculeatus L., and the latter from the pectoral fins of a sculpin, Cottus scorpius L. Beyond these vorticellids, the only protozoan genera recorded from intertidal zone fishes in other parts of the world but not so far from New Zealand belong to the Myxosporidia—Sphaerospora is known from blennies in Europe (Théolohan, 1895) and from Californian toadfishes (Jameson, 1931), Ceratomyxa from blennies, gobies, sculpins and rocklings in Europe (Théelohan, 1892; Parisi, 1912; Dunkerly, 1920) and from various blennies, kelpfishes and sculpins in North America (Mavor, 1915, 1916; Ellis, 1930; Jameson, 1929; Noble, 1941), Chloromyxum from blennies (Parisi, 1912) and sculpins (Dunkerly, 1920) in Europe, and Trilospora from blennies and gobies in North America (Noble, 1941).

The present studies suggest that the environmental niche occupied by fishes is an important factor in determining the composition of their protozoan fauna. This is particularly evident with regard to Trichodina parabranchicola and T. multidentis, ciliates having a wide host–range, the dominance of which in nature appears to be largely determined by the habitat of the host. The occurrence of the myxosporidians recorded herein also supports this hypothesis, suggesting that at least in the case of those species having a wide host range, Myxidium incurvatum for example, the hosts in any one locality are determined by the particular feeding ground in which spores accumulate. Where fishes occupying distinctive niches within the one general habitat are parasitized by the same haematozoan—as in the instances of Ericentrus, Oliverichtus and Tripterygion parasitized by Haemogregarina bigemina—the explanation may perhaps lie in the greater catholicity with regard to habitat displayed by the invertebrate hosts. Literature Cited Auebbach, M., 1912. Studien über die Myxosporidien der norwegischen Seefische und ihre Verbreitung. Zool. Jb. Syst., 34, 1—50. Awerinzew, S., 1911. Studien über parasitischen Protozoen. VII. Ueber die Sporenbildung bei Myxidium sp. aus der Gallenblase von Cottus scorpius. Arch. Protistenk., 23, 199—204. Ariake, B., 1929. Five new species of Trichodina. Annot. Zool. Japon., 12, (1), 285—288. Basikalowa, A., 1932. Wissenschaftliches Zentralinstitut für Fischereiwirtschaft, Moscow. p. 136 (Original not seen; summary of Sinuolinea descriptions published by Tripathi, 1948, p. 115). Bentham, 1917. Quoted by Wenyon, C. M., 1926. Protozoology. Baillière, Tindall and Cox, London. Vol. II, p. 1403. Bond, F. F., 1938. Cnidosporidia from Fundulus heteroclitus Lin. Trans. Amer. micr. Soc., 57, (2), 107—122. Boyer, C. S., 1927. Synopsis of North American Diatomaceae, Part II Proc. Acad. Nat. Sci. Philadelphia, 79, Suppl., 229—583. Brettschneider, L. H., 1935. Der Feinbau von Trichodina entzii sp.n. Réc. des Travaux dédiés au 25me Anniv. Scient. du Prof. Pavlovsky (All Union Inst. Exp. Med.), Moscow. Brumpt, E., 1913. Précis de Parasitologie (2nd Ed.). Mason et Cie, Paris. xxviii + 1 — 1011. — and Lebailly, C., 1904. Description de quelques nouvelles espèces de trypanosomes et d'hémogrégarines parasites des teléostéens marins. C.R. Acad. Sci., Paris, 139, 613—615. Chatton, E., 1910. Protozoaires parasites des branchies des labres; Amoeba mucicola Chatton, Trichodina labrorum n.sp. Arch. Zool. exp. gén, 5, 239—266. — and Lwoff, A., 1923. Un cas remarquable d'adaptation: Ellobiophrya donacis n.g., n.sp., péritriche inquilin des branchies de Donax vittatus (Lamellibranche). C.R Soc Biol., Paris, 88, 749—752. — — 1929. Contribution a l'etude de l'adaptation. Ellobiophrya donacis Ch. et Lw. péritriche vivant sur les branchies de l'Acéphale Donax vittatus da Costa. Bull. biol., 63, 321—349. Claparede, E., and Lachmann, J., 1858. Etudes sur les infusoires et les rhizopodes. Vaney, Geneva. Cunha, A. X. da, 1950. Trichodina ranae n.sp., un urcéolahe parasite de la vessie urinaire de la grenouille. Mem. Est. Mus. Zool. Univ. Coimbra, No. 202, 1—4. Davis, H. S., 1916. The Structure and Development of a Myxosporidian Parasite of the Squeteague, Cynoscion regalis. J. Morph., 27, 333—377. — 1917. The Myxosporidia of the Beaufort Region. A Systematic and Biologic Study. Bull. Bur. Fish., U.S., 35, 201—243. — 1947. Studies on the Protozoan Parasites of Fresh–water Fishes. U.S. Dept. Interior, Fish. Bull. 41, 1—29. Diller, W. F., 1928. Binary Fission and Endomixis in the Trichodina from Tadpoles (Protozoa, Ciliata). J. Morph., 46, (2), 521—561.

Doflein, F., 1898 Studien zur Naturgeschichte der Protozoen. III. Ueber Myxosporidien. Zool. Jb., Anat., 11, 281—350. Dunkerly, J. S., 1920. Fish Myxosporidia from Plymouth. Parasitol., 12, (4), 328—333. — 1925. The Development and Relationships of the Myxosporidia. Q.J.M.S., 69, No. 274, N.S., 185—216. Ellis, M. F., 1930. Investigations on the Protozoan Fish Parasites of the St. Andrew's Region. Trans. Nova Scot. Inst. Sci., 17, (4), 268—275. Englemann, T. W., 1876. Ueber Entwickelung und Fortpflanzung von Infusorien. Morph. Jb., 1, 573—634. Entz, G., Sr., 1896 Fauna regni hungariae. Animalium hungariae hucusque cognitorum enumeratio systematica. Keg. Soc. Sci. Nat. Hungarica, 6, 1—29. Fabre–Domergue, P., 1888. Etudes sur l'organisation des Urceolaires. J. Anat., Paris, 24, 214—250. Fantham, H. B., 1919. Some Parasitic Protozoa Found in South Africa, II. S. Afr. J. Sc., 16, 185—191. — 1930. Some Parasitic Protozoa Found in South Africa, XIII. Ibid., 27, 376—390. — Porter, A., and Richardson, L. R., 1939. Some Myxosporidia Found in Certain Fresh–water Fishes in Quebec Province, Canada. Parasitol., 31, (1), 1—77. — — 1942. Some Haematozoa Observed in Vertebrates in Eastern Canada. Ibid., 32, (2), 199—266. Faure–Fremiet, E., 1905. Le structur de l'appareil Fixateur chez les Vorticellidae. Arch. Protistenk., 6, 207—266. — 1943.Etude biometrique de quelques Trichodincs. Bull. soc. zool. Fr., 68, 158—169. — and Mugard, H., 1946. Une trichodine parasite endovésicale chez Rana esculenta. Ibid., 71, 36—38. Fulton, J. F., 1923. Trichodina pediculus and a New Closely Related Species. Proc. Bost. Soc. Nat. Hist., 37, (1), 1—29. Georgevitch, J., 1916. Note sur les myxosporidies recueillies à Roscoff. Bull. soc. zool. Fr., 41, 86—95. — 1916a. Note sur les myxosporidies des poissons de la baie de Villefranche et de Monaco. Bull. l'inst. ocean., No. 322. — 1917.Esquisses protistologiques. Bull. soc. zool. Fr., 42, 99—107. Gonnert, R., 1935. Uber Systematik, Morphologie, Entwicklungsgeschichte und Parasiten einiger Dendrosomidae nebst Beschreibung zweier neuer Suktorien. Arch. Protistenk., 86, 113—154. Griffin, L. T., 1933. Studies in New Zealand Fishes. Trans. N.Z. Inst., 63, 330—333. Hendey, N. I., 1937.The Plankton Diatoms of the Southern Seas. Discovery Repts., 16, 151—364. Henry, H., 1910. On the Haemoprotozoa of British Sea-fish.Jour. Path. Bact., 14, 463—465. — 1912. Haemogregarina anarrhichadis from Anarrhichas lupus, the Catfish. Parasitol., 5, (3), 190—196. — 1913. A Summary of the Blood Parasites of British Sea–fish. Jour. Path. Bact., 18, 218—223. — 1913a. A New Haemosporidian from Scomber scomber, the common mackerel. Ibid., 18, 228—231. — 1913b. A Haemogregarine and a Leucocytozoon from Gadus aeglefinus. Ibid., 232—239. — 1913c. The Granule Shedding of Haemogregaina simondi. Ibid., 18, 240—249. Hirshfield, H., 1949. The Morphology of Urceolaria karyolobia sp. nov., Trichodina tegula sp. nov, and Scyphidia ubiquita sp. nov., three new ciliates from Southern California limpets and turbans. J. Morph., 85, (1), 1—33. Hutton, F. W., 1872. Fishes of New Zealand. Catalogue, with Diagnoses of the Species. Colonial Mus. and Geol. Surv. Dept., N. Z., xvi + 1—93. Ibara, Y., 1931 Trichodina okajimae n.sp., a New Species of Trichodina from Salamanders, J.Elisha Mitchell Sei. Soc., 46, 214—217. Jameson, A. P., 1929. Myxosporidia from Californian Fishes. J. Parasit., 16, (2), 59—68. — 1931. Notes on Californian Myxosporidia. Ibid., 18 (2), 59—68. Jones, E. W. K., 1940. The Occurrence of a Marine Leech, Abranchus blenni, n.sp., resembling A. sexoculatus (Malm.) in North Wales. J. Mar. Biol. Assoc., U.K., 24, (2), 533—541. Jordan, D. S., and Evermann, B. W., 1898. Fishes of Middle and North America. U.S. Nat. Mus. Bull., 47, (3), xxiv + 2183—3136.

Kahl, A., 1935. Urtiere oder Protozoa. I: Wimpertiere oder Caliata (Infusoria). 4. Peritricha und Chonotricha. In Dahl: Dic Tierwelt Deutschlands. 30 Teil, 651—886. Kohl–Yakimoff, N., and Yakimoff, W. L., 1915. Hamogregarinen der Seefische. Centralbl. Bakt., 76, (I Abt.), 135—146. Kudo, R. R., 1916. Contributions to the Study of Parasitic Protozoa. III, Notes on Some Myxosporidia Found in Some Fresh–water Fishes of Japan, with the Description of Three New Species. J. Parasit., 3, 3—9. — 1920. Studies on Myxosporidia. A synopsis of Genera and Species of Myxosporidia. Ill. Biol. Monographs, 5, (3 & 4), 1—265. — 1926. On Myxosoma catostomi Kudo 1923, a Myxosporidian of the Sucker, Catostomus commersonii. Arch. Protistenk., 56, 90—114. — 1930. Myxosporidia. In “Problems and Methods of Research in Protozoology,” edited by R. Hegner and J. Andrews. Macmillan Co., New York., pp. 303—324. — 1933. A Taxonomic Consideration of Myxosporidia. Trans. Amer.micr. Soc., 52, (3), 195—216. Labbe, A., 1899. Sporozoa. Das Tierreich. 5 Lief., 1—180. Laird, M., 1951. A Contribution to the Study of Fijian Haematozoa. Zool.Pubs. Viet. Unn.Coll., No. 10, 1—15. — 1951a. Studies on the Trypanosomes of New Zealand Fish. Proc. Zool. Soc. Lond., 121, (2), 285—309. — 1952. New Haemogregarines from New Zealand Marine Fishes.Trans. Roy. Soc. N. Z., 79, 589—600. Laveran, A., and Mesnil, F., 1901. Deux espèces nouvelles d'hémogrégarines des poissons. C.R.Acad. Sci, Paris, 133, 572—577. — 1902. Sur les hématozoaires des poissons marins.Ibid., 135. 567. Lebailly, C., 1904. Sur quelques hémoflagellés des téléostéens marms C. R. Acad.Sci. Paris, 139, 576—577. — 1905. Sur des hématozoaires nouveaux parasites de la barbue. C. R. Soc.Biol., 59, 304. Lynch, J. E. and Noble, A. E., 1931. Notes on the Genus Eudosphaera Engelmann and on its Occasional Host Opisthonecta henneguyi Fauré–Fremiet. Univ. Calif. Pubs.Zool., 36, (5), 97—114. MacLennan, R. F., 1939. The Morphology and Locomotor Activities of Cyclochaeta domcrguer Wallengren (Protozoa). J. Morph, 65, (2), 241—255. Mavor, J. W., 1916. On the Life History of Ceratomyxa acadiensis, a New Species of Myxosporidia from the Eastern Coast of Canada. Proc. Amer. Acad. Arts and Sci., 51. 551—574. Minchin, E. A., 1907. On a Haemogregarine from the Blood of a Himalayan Lizard (Agama tuberculata). Proc. Zool. Soc. Lond., 1907, 1098—1104. Mueller, J. F., 1937. Some Species of Trichodina (Ciliata) from Fresh Water Fishes Trans. Amer. Micr. Soc., 56, (2), 177—184. —, 1938. A New Species of Trichodina (Ciliata) from the Urinary Tract of the Muskalonge, with a Repartition of the Genus J.Parasit, 24, (3), 251—258. Nemeczec, A., 1922. Ueber Zschokkella rovignensis spec. nov. Arch. Protistenk., 45, 390—400. Neumann, R. O., 1908. Ueber protozoische Parasiten im Blut von Meeresfischen Munch Med. Woch., 55, 1460—1461. — 1909. Studien über protozoische Parasiten im Blut von Meeresfischen. Zeitschi. Hyg. u. Infekt., 64, 1—112. Noble, E. R., 1938 Two New Myxosporidia from Tide Pool Fishes of California J. Parasit, 24, (5), 441—444. Noble, E. R., 1939. Myxosporidia from Tide Pool Fishes of California.Ibid., 25, (4), 359—364. Noble, E. R., 1941. On Distribution Relationships Between California Tide Pool Fishes and Their Myxosporidian (Protozoan) Parasites. Ibid., 27, (5), 409—415. Noble, G. A., 1940. Trichodina urechi n. sp, an entozoic Ciliate from the eehiuroid Worm, Urechis caupo. Ibid., 26, 387—405.

Oliver, W. R. B., 1923. Marine Littoral Plant and Animal Communities in New Zealand. Trans. N.Z. Inst., 54, 496—545. Padnos, M, 1939. Endosphaera engelmanni endoparasitic in Trichodina Infecting the Puffer, Spheroides maculatus (Linnaeus). Anat. Rec. 75, (4, Suppl.), 156. Padnos, M, and Nigrelli, R. F., 1942. Trichodina spheroidesi and Trichodina halli spp. nov. Parasitic on the Gills and Skin of Marine Fishes, with Special Reference to the Life–history of T. spheroidesi. Zoologica, N.Y., 27, 65—72. Parisi, B, 1912. Primo contributo alla distribusione geografica dei missosporidi in Italia. Atti soc. ital sc. nat., 50, 283—290. Phillipps, W. J., 1927. Bibliography of New Zealand Fishes. N.Z. Maine Dept., Fish. Bull. No. 1, 1—68. Precht, H., 1935. Epizoen der Kieler Bucht. Noia Acta Leopoldina, 3, (N.F.), (15), 405—474. Quennerstedt, A., 1869. Bidiag til Sveriges Infusorie–fauna Hft. III, Acta Univers. Lundensis, 1869. Reichenow, E., 1910. Haemogregarina stepanowi. Die Entwicklungsgeschichte einer Hamogregarine. Arch. Protistenk., 251—350. Reichenow, E., 1927—29. p. 925 in Reichenow's edn. of Doflein's Lehrbuch der Protozoenkunde, Berlin, vi + v + vii + 1—1262. Robin, C., 1879. Mémoire sur la structure et la reproduction de quelques infusoires tentaculés, sucenis et flagellés. J Anat. Phys., 15, 529. Saville–Kent, W., 1880—82. A Manual of the Infusoriaa. London. Vol. I, x + 1—472. Vol. II, 473—913. Vol. III (51) Plates. Shortt, H. E., 1923. Data and incomplete reference given by Gatenby and Beams, 1950, 11th. Edn. of Holies Lees's Microtomist's Vade–Mecum. Churchill, London, xiv + 1—753 (p. 152, Entry 310). Thelohan, P., 1892. Observation sur les myxosporidies et éssai de classification de ces organismes. Bull. soc. philom., 4, 165—178. Thelohan, P., 1895. Recherches sur les myxosporidies. Bull. sc Fr et Rely., 26, 100—394. Thompson, Sally, Kirkegaard, D., and Jahn, T. L., 1947. Scyphidia ameiuri, n.sp., a Puritrichous ciliate from the Gills of the Bullhead, Ameiurus melas melas. Trans. Amer. micr. Soc., 66, (4), 315—317. Thompson, P. E., and Huff, C. G., 1944. A Saurian Malaria Parasite, Plasmodium mexicanum n. sp., with both elongatum- and gallinaceum–types of exoerythrocytic stages, J Inf Dis, 74, 48—67. Tripathi, Y. R., 1948. A New Species of Ciliate, Trichodina branchicola, from Some Fishes at Plymouth. J. Mar. Biol. Assoc. U.K., 27, (2), 440—450. Tripathi, Y. R., 1948a. Some New Myxosporidia from Plymouth with a Proposed New Classification of the Order. Parasitol., 39, (1 & 2), 110—118. Wallengren, H., 1897. Bidrag till Kannedomen om Fam Urccolarina Stein. Lunds. Unit,. Arsskrift, 33, (8), 1—48. Yonge, C. M., 1949. The Sea Shore. Collins, London. xvi + 1—311.