examples the broadest part of the body is at the central region. There is usually a short and clearly demarcated caudal spike (Fig. 2), which never exceeds 1.2μ in length, at the posterior extremity. This structure is always evident in life, but sometimes appears to be lacking in fixed and stained examples, particularly those towards the upper limit of the size range. The length of the oral pouch is about half that of the body in small examples, and about a third that of the body in large ones. A fine chromophile fibril borders the oral pouch anteriorly and laterally This fibril, although prominent in haematoxylin preparations, is relatively unresponsive to Giemsa. The nucleus has a large central endosome, which exhibits peripherally disposed chromatin granules. In the smaller examples the posterior flagellum is usually confined within the oral pouch, but in the larger ones it may protrude for a short distance. The latter flagellum is always markedly shorter than the anterior one, while the length of neither of them exceeds that of the body. The alveolar cytoplasm, which may contain small chromophilic granules and the spore bodies of ingested bacteria, stains light blue with Giemsa.

Kowalczyk (1938) found the average size of 400 random examples of R. phyllophagae from the Japanese beetle to be 7.8μ by 4.0μ. This is extremely close to the size of the retortamonad under discussion, which agrees morphologically with R. phyllophagae as described by Travis and Becker and by Kowalczyk. This flagellate of O. zealandica is thus identified as Retortamonas phyllophagae (Travis and Becker), the new host and locality not being regarded as sufficient grounds for its description as new.

Retortamonas pericopti n.sp. (Figs. 1, 3–7).

Thirty-nine of my 50 Pericoptus truncatus larvae proved to harbour the large retortamonad described hereunder.

This species has a slipper-shaped or fusiform body, which, in living examples, may measure up to 26μ in length and 9μ in breadth. Fifty random examples in haematoxylin preparations ranged from 10.9μ to 22.2μ in length and from 4.0μ to 7.9μ in breadth (av., 15.1μ by 5.9μ), while the same number of random examples from Giemsa-stained dry smears ranged from 11.1μ to 25.7μ in length and from 4.6μ to 8.9μ in breadth (av., 16.1μ, by 6.5μ) The length of the oral pouch seldom exceeds two-fifths of that of the body. The flagellate tapers posteriorly, and may exhibit spiral twisting (Figs. 3 and 6). There is usually a caudal spike which may reach 5μ in length but seldom exceeds 1μ in thickness at the base. This structure seldom appears to advantage in air-dried films, assuming a rather lighter shade of blue with Giemsa than does the body proper (Figs. 3–5), but is well marked in life and appears more or less hyaline in haematoxylin mounts. As in R. phyllophagae the anterior flagellum is the longer but is not, however, as long as the body. The posterior flagellum is rarely (Fig. 3) confined within the oral pouch, usually protruding for at least three or four microns beyond this structure. The average lengths of the anterior and posterior flagella are 14μ and 10μ respectively. The alveolar cytoplasm stains light blue with Giemsa, with darker blue maculations, especially towards the periphery. Numerous vacuoles may be present (Figs. 3, 5, 6), and food inclusions—particularly the spore bodies of bacteria (Figs. 3, 4, 6° 7)—are frequently demonstrable.

The nucleus, as seen in haematoxylin preparations, has a large endosome with peripheral aggregations of chromatin (Fig. 1). It is variously distorted, and the structural details are not apparent, in Giemsa-stained examples (Figs. 3–6). The basal granules giving origin to the flagella are located against the nuclear membrane adjacent to the inner end of the oral pouch.

This Retortamonas of P. truncatus differs morphologically from known species insofar as can be gathered from published data. It is of very much larger size than some of the other species from insects. Thus, while its dimensions range from c. 11μ to 26μ by c. 4μ to 9μ, R. agilis (Mackinnon) measures 4μ by 1.5μ to 11μ by

(Mackinnon, 1915), R. blallae (Bishop) ranges from 6μ to 9μ in length (Bishop, 1931) and R. phyllophagae (Travis and Becker) ranges from 4μ to 12μ in length and from 2μ to 8.5μ in breadth (Kowalczyk, 1938).

Species of comparable size are R. alexeieffi (Mackinnon) (7μ to 16μ by 5μ to 9μ, according to Mackinnon, 1915, and up to 22.5μ by 8.7μ according to Ludwig, 1946); R. gryllotalpae Grassi (7μ to 19.5μ in length, the breadth varying from one-fifth to one-half the length, according to Wenrich, 1932); R. caudacus Geiman (6μ to 22μ by 3μ to 6μ; Geiman, 1932); and R. wenrichi Stabler (12.2μ to 19μ by 3.8μ to 8.4μ; Stabler, 1944).

The three last-named species differ sharply from that under discussion in the length of the caudal process relative to that of the body. According to Wenrich (1932) this process is from one-quarter to two-fifths of the length of the body proper in R. gryllotalpae, which further differs from the present species in that its flagella are longer than the body, the posterior flagellum being longer than the anterior one. R. caudacus has a long, needle-like caudal process, its oral pouch is half the length of the body, and the nucleus is located right at the anterior extremity (Geiman, 1932). R. wenrichi is unique in that the caudal process may vary from a short spike to a looped and twisted filament up to 58.5μ in length. This species differs further from the one from P. truncatus in that its flagella are of approximately equal length and are generally from one-and-a-half times to twice the length of the body proper. Grassé (1952) illustrated long-tailed examples of R. gryllotalpae discovered by him together with normal and tail-less forms in the type host in Europe (his Fig. 663), and was inclined to consider that R. wenrichi may be a synonym of the former species, although he did not commit himself in this matter. He also pointed out that R. caudatus is very close to R. gryllotalpae.

As far as can be gathered from the literature, the closest affinities of the New Zealand retortamonad he with R. alexeieffi (Mackinnon). This species was originally described from crane fly larvae (Dipera, Tipulidae; Tipula sp.) in England, and has since been reported from North American tipulid larvae (Geiman, 1933, Ludwig, 1946). Its size is comparable with that of the present species, according to the following figures from Ludwig (1946). As seen in haematoxylin preparations, it varies from 10μ to 22.5μ on length (as compared with 10.9μ to 22.2μ) by from 4.1μ to 8.7μ in breadth (as compared with 4.0μ to 7.9μ). The average size of R. alexeieffi in such preparations is very close indeed to that of the New Zealand species, being 13.7μ by 6.2μ (as compared with 15.1μ by 5.9μ)

Trophozoites of R. alexeieffi may either have a short caudal spike (Wenrich, 1932) or else the posterior end may be rounded, this last condition being the characteristic one according to Ludwig (1946). In the retortamonad from Pericoptus truncatus, on the other hand, a caudal spike is usually present. The length of the anterior flagellum as compared with that of the body appears to be much the same in both cases, but the posterior flagellum is relatively shorter in R. alexeieffi and its extremity seldom projects out beyond the oral pouch (Ludwig, 1946) The cytoplasm of R. alexeieffi frequently contains numerous ingested bacteria, that of the example illustrated by Ludwig in his Pl. II, Fig. 13 containing two inclusions very similar to the bacterial spore bodies so commonly ingested by the species under consideration. Finally, figures of R. alexeieffi published by various authors agree in showing the nucleus located at the very anterior tip of the body. This organelle is never so far forward in the New Zealand species.

Differing in detail from other species of its genus as outlined herein, the retortamonad of Pericoptus truncatus is accordingly designated Retortamonas percopti n.sp. Its closest known relative would appear to be R. alexeieffi (Mackinnon), from which it is to be distinguished by the relatively longer posterior flagellum, the relatively longer and more persistent caudal spike and the less anterior location of the nucleus. These differences might well prove to be more apparent than real following future com-

parisons of more detailed data, in which event it is not considered that the difference in hosts should be regarded as debarring the identification of the present flagellate with R. alexeieffi.

Monocercomonoides melolonthae (Grassi, 1879) (Figs. 12–16).

Numerous authors have recorded this species, from a variety of European and North American insects. There are also unconfirmed reports from amphibians (Swezy, 1916) and from lizards (Wood, 1935) in North America.

Monocercomonoides trophozoites were abundant in the hind gut contents of all the insects examined during this study. Spherical (Fig. 15), subspherical (Fig. 13), ovoid (Fig. 12) and pyriform (Figs. 14 and 16) examples were noted, ovoid and pyriform ones predominating. Living specimens, particularly pyriform ones, sometimes have a slight axostylar projection posteriorly. The pellicle of the posterior extremity appears to be somewhat sticky, for numerous individuals were observed to have bacteria or inorganic particles adhering there. None of the internal organella could be distinguished either by bright field- or by dark field-illumination. The organism moves rapidly through the medium, three of the flagella beating anteriorly and the fourth trailing behind. This trailing flagellum was seen to serve as an attachment organelle from time to time.

Giemsa-stained dry smears proved useful for the demonstration of the flagella, the axostyle and the two basal granules with their connecting rhizoplast. Although such preparations are unsuitable for the study of nuclear detail, and the flagellates concerned suffer a degree of hypertrophy and distortion, they may be employed to advantage in the rapid generic identification of these zoomastiginids. The particularly heavy staining of the flagella (see Plate) renders these conspicuous under quite low powered objectives, a useful feature for student demonstrations in teaching laboratories.

Fixation with osmic vapour, followed by methyl alcohol and Giemsa resulted in the appearance of large, black-staining inclusions in the cytoplasm (Fig. 12). These—probably volutin granules—were never evident in flagellates fixed by any other method. The cytoplasm is alveolar and more or less granular, and usually contains a few vacuoles of varying size (Figs. 12–15). Ingested bacteria are occasionally evident. In haematoxylin-stained specimens the nucleus is seen to have a large central endosome, and in many instances a halo of chromatin granules is apparent to one side of this structure between it and the nuclear membrane (Figs. 14 and 15). This is a generic feature according to Kowalczyk (1938). Giemsa-stained examples, on the other hand, have the endosome hypertrophied and the nuclear membrane distorted. The anterior margin of the nucleus of such forms is usually somewhat concave, the curvature of the concavity often coinciding with that of the rhizoplast (Fig. 16). Because of this concavity the antero-lateral limits of the nucleus have the appearance of horn-like projections. This is without doubt an artefact consequent upon dry fixation, but it is nevertheless a useful aid to rapid recognition of the flagellate in Giemsa films.

The trailing flagellum, which as Ludwig (1946) indicated is usually longer than the three anterior ones, originates from the same basal granule as does the axostyle. All four flagella of 20 consecutive Giemsa-stained trophozoites from Pericoptus truncatus were measured, with the following results: Trailing flagellum, 15. 0μ to 22. 5μ (av., 18. 2μ); anterior flagella, 13.1μ to 17.8μ (av., 14.8μ), 13. 3μ to 17. 4μ (av., 15. 1μ) and 12.7μ to 17. 0μ (av., 14. 6μ). The anterior flagella of course often become reflected in smears, and are so drawn in the accompanying illustrations for space economy.

Three different series each of 50 consecutive trophozoites from the selfsame P. truncatus larva were measured. Living examples (made sluggish by lengthy exposure to the heat and light of the microscope lamp) were found to range from 5. 0μ to 11. 4μ in length (av., 7. 8μ) and from 3. 8μ to 10. 3μ in breadth (av., 6. 1μ).

Giemsa-stained ones in a dry film measured from 5.3μ to 10 7μ (av., 8.4μ) by from 3.6μ to 10.5μ (av., 7.1μ) Trophozoites in never-dried haematoxylin preparations, though, ranged from 4.3μ to 7.9μ in length (av., 5.8μ) and from 3.2μ to 5.7μ in breadth (av., 4.4μ). These differences according to the technique employed are in accord with those noted by Minchin (1909) for trypanosomes in air dried and never-dried blood films.

No significant morphological or size differences could be ascribed to the difference in hosts. Thus 50 consecutive trophozoites in haematoxylin smears from Odontria zealandica ranged from 4.1μ to 7.9μ (av., 5.5μ) by from 3.0μ to 6.1μ (av., 4.2μ), while the same number in similar preparations from Hemideina thoracica ranged from 4.2μ to 8.1μ (av., 5.6μ) by from 2.9μ to 6.4μ (av., 4.4μ).

The Monocercomonoides discussed herein is morphologically indistinguishable from M. melolonthae (Grassi) as described by numerous authors. It would appear to be of rather smaller average size than usual, the figures of some other investigators, derived from wet-fixed preparations, being:

6μ by 3μ to 9μ by 5μ (Macknnon, 1912; Tipula sp., England)

5μ to 12μ by 4μ to 10μ, av. (250 random examples), 6.6μ by 5.5μ (Kowalczyk, 1938; Popillia japonica Newm., U.S.A.)

6.2μ to 12 5μ by 4.3μ to 10μ, av (150 random examples), 10μ by 7.5μ (Ludwig, 1946; Tipula abdominalis, U.S A.).

However, the size range of the present flagellate substantially overlaps the above ranges, which in any case differ quite considerably from one another; and the average dimensions of haematoxylin-stained examples are only of the order of one micron below those given by Kowalczyk (1938)

Monocercomonoides globus Cleveland et al., occurring in North American wood roaches, is a distinctive species having a ribbon-shaped axostyle (Cleveland et al., 1934), while M. panesthiae, described by Kidder (1937) from a Philippine Islands wood roach, has a hyaline axostyle becoming free posteriorly and all four of its flagella appear to originate from a single basal granule. The latter species should thus be transferred to the genus Monocercomonas Grassi. Monocercomonoides cetoniae (Jollos) and M. ligrodis Travis, both described from beetle larvae, are synonyms of M. melolonthae, according to Grassé (1952)

The common orthopteran Monocercomonoides, M. orthopterorum (Parisi), is of markedly smaller size than M. melolonthae, ranging from 3.7μ to 7.0μ by from 2.1μ to 3.8μ (av., 6.2μ by 3.1μ) (Ludwig, 1946) This flagellate, which on grounds of size alone bears comparison with the New Zealand one, differs from the latter in its consistently more elongate shape The organism under consideration, differing from other species of Monocercomonoides from insect hosts as indicated herein, resembles M. melolonthae (Grassi) most closely and lacks well marked characteristics which would justify its description as new It is accordingly regarded as referable to the latter species.

Polymastix melolonthae (Grassi, 1879) (Figs. 8–11)

P. melolonthae, which is known from the larvae of numerous genera of beetles and from tipulids in Europe and the U.S.A, was present—usually in large numbers—in the hind gut contents of all 10 Odontria zealandica larvae and of 48 of the 50 Pericoptus truncatus larvae examined.

This fairly large, fusiform flagellate moves through the medium with a distinctive corkscrew motion. After prolonged exposure to the heat and light of the microscope lamp it tends to attach itself to the cover slip by the posterior tip of the body, hanging head downwards with all four flagella beating anteriorly. Perhaps its most distinctive feature is the presence of numerous, more or less longitudinally aligned ectocommensal bacteria, which are clearly visible by dark field illumination.

The following description is based on haematoxylin-stained material except where otherwise stated. Although examples having the posterior extremity rounded are not

uncommon (Fig 8), the body usually comes to a point posteriorly (Figs. 9 and 10). The hinder part of the body is sometimes bifurcated, as described by Kowalczyk (1938) and figured by Grassé (1952; his Fig. 625-B). A spherical to subspherical nucleus, averaging about 2μ in diameter and having a prominent central endosome of from 1.0μ to 1.6μ in diameter, is located towards the anterior extremity. Two closely adjacent basal granules connected by means of a rhizoplast are situated just in front of the nucleus. The granules appear more widely separated in air dried smears due to distortion, and the rhizoplast is then often clearly demonstrable with Giemsa (Fig. 9). An anteriorly thickened and dark-staining axostyle originates at one of the basal granules and curves about the nucleus keeping close to the membrane (Fig. 8). The axostyle could never be traced further posteriorly than in the example illustrated in Fig. 10, and the fine terminal filament shown in Grassé's Fig. 625-G was never made out The cytoplasm is alveolar and frequently contains vacuoles and food inclusions (Figs. 8, 9), but details are difficult to distinguish because of the ectocommensal bacteria attached to the pellicle. Osmic-fixed trophozoites in Giemsa smears often exhibit dark-staining spherules similar to those noticed in such preparations of Monocercomonoides melolonthae (Fig. 9).

Numbers of ectocommensal bacteria referable to the genus Fusiformis Holling (Actinomycetales: Mycobacteriaceae) are always attached to the firm pellicle as illustrated in Figs. 8–10. These have the form of elongate rods, pointed at the ends and ranging from about 2μ to 6μ in length. From two to eight siderophilous granules are distinguishable in suitably destained examples. The ectocommensals were observed to desert their host on the death of the latter, and Grassé (1952) noted that this is also the case at encystment. Morphologically identical bacteria are common in the free-living state in the gut contents of the host larvae (Fig. 11). The longitudinally attached ectocommensals of Polymastix melolonthae were referred to Fusiformis melolonthae Grassé by Grassé (1952), who gave their length as from 2μ to 7μ. It is possible that this name may prove to be a synonym of F. termitidis Hoelling, the genotype. This species, which occurs in the intestinal tract of termites, was described by Bergey (1934) as from 3μ to 5μ in length, the length increasing with the age of the organism, and as having from two to eight darkly staining granules suggestive of nuclei.

A pair of flagella originates from each of the basal granules. All four flagella are often directed anteriorly, although one, which is appreciably longer than the others, may trail behind. The length of the one with the tendency to trail was found to range from 20. 3μ to 35. 1μ, and that of the other three from 13. 1μ to 22. 2μ.

Fifty wet-fixed trophozoites stained with haematoxylin were measured in a preparation from Pericoptus truncatus and the same number in a slide from Odontria zealandica, no selection being exercised in either case. Those from Pericoptus proved to range from 7. 9μ to 16. 2μ by from 4.1μ to 7. 9μ (av., 13.1μ by 5. 3μ), while those from Odontria ranged from 7. 9μ to 19. 6μ by from 2. 9μ to 6. 9μ (av., 1. 9μ by 5. 3μ).

These measurements exceed those of Polymastix phyllophaiae Travis and Becker, the trophozoites of this species ranging from 4. 75μ to 12. 3μ in length and from 3. 8μ to 5. 4μ in breadth (Travis and Becker, 1931), the upper limit of their size range thus coinciding with the average for the flagellate under discussion. P. wenrichi Geiman, on the other hand, an entozoan of North American crane fly larvae, attains a much larger size than does the New Zealand species, ranging from 12μ to 35μ in length and from 8μ to 22μ in breadth (Geiman. 1932). Grassé (1952) has recently described P. hystrix from termites (Neotermes aburiensis); but this species is remarkable in having a vestiture of very long bacteria attached to the pellicle by one end and is unique in its genus in possessing a parabasal apparatus—Grassé himself felt qualms at including this flagellate in Polymastix.