Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 77, 1948-49
This text is also available in PDF
(5 MB) Opens in new window
– 469 –

A New Family of Living Ostracoda with Striking Resemblances to Some Palaeozoic Beyrichiidae

[Received, by Editor, January 8, 1948; issued separately, September, 1949.]

Introduction

The discovery of the two forms described in this paper makes it clear that the Ostracoda of the South Pacific are far from completely known. The Challenger Expedition made extensive marine collections which were described by G. S. Brady in the Reports, but very many forms from the New Zealand area remained unknown.

The two species here described were found in two dredgings from the collection of the late R. L. Mestayer, made in 1914 in the North Cape-Three Kings area; and the writer is indebted to Mr. A. W. B. Powell, of the Auckland Museum, for his generosity in making the material available.

It is unfortunate that the material contains only separated valves with no trace of the appendages. Both forms are very rare, only four valves of Manawa tryphena and one adult and one juvenile valve of Puncia novozealandica being known. There appears to be very little chance at present of obtaining more dredgings from that area, so the possibility of finding appendage material is slight. In spite of the unavoidable gaps in the knowledge of these animals the author has decided to place the information on record and erect a new family to include these two strange forms, so radically unlike other known living ostracods. Most remarkable is the striking resemblance that Puncia bears to some of the Palaeozoic Eurychilinas, and in this connection the author has to acknowledge his indebtedness to Dr. R. Bassler, of the United States National Museum, and, in particular, to Dr. Frank M. Swartz, of the Department of Earth Sciences, Pennsylvania State College, for affording him the benefit of their extensive knowledge of Palaeozoic Ostracoda.

Type specimens are lodged with the New Zealand Geological Survey collections.

Order Ostracoda Family Punciidae nov.

I propose this new family to include the genera Puncia nov. and Manawa nov., with Puncia as the type genus.

Family characters: Carapace elongate, semi-elliptical with a long straight hinge, with or without terminal teeth; free margins with a wide, double-walled, frill-like border, partitioned by radial septa into internal chambers and produced well below the true contact edges of the valves; muscle-scar pattern consisting of six more or less radially disposed scars.

– 470 –

Genus Puncia nov.
Genotype: Puncia novozealandica Hornibrook n.sp.

Carapace semi-elliptical, higher and slightly more inflated at posterior end, well-defined cardinal angles and long, straight hinge without teeth; two pointed dorso-lateral, somewhat backwardly directed, horn-like processes on each valve; free margins of valves provided with a wide, double-walled frill divided into about forty compartments by radial septa and produced well below the true contact margins of the valves; muscle-scar pattern consisting of six more or less radially disposed scars situated in a median sulcus. (The question of orientation is discussed below.)

Puncia novozealandica Hornibrook n.sp. (Plate 50, Figs. 1–6.)

Description: Carapace smooth and delicate, other features as for generic description.

Holotype: A right valve; length 0°56 mm., height 0°29 mm.

Locality: Mestayer Stn. 21, 75 faths., 6 ½ E. 5′ N. of North Cape, 8/9/1914.

Juvenile specimen from Mestayer Stn. 7, 98 faths., off Big King.

Genus Manawa nov.
Genotype: Manaua tryphena Hornibrook n.sp.

Carapace semi-elliptical, non-sulcate, higher at posterior end; hinge long and straight, produced in the right valve into two terminal, bracket-like processes which clamp around the bevelled ends of the left valve; free margins of valves provided with a wide, double-walled frill subdivided into about eleven chambers, produced well below the true contact margins of the valves and bearing a rim of clear shelly material; muscle-scar pattern consisting of six more or less radially disposed scars.

Manawa tryphena Hornibrook n.sp. (Plate 51, Figs. 1–8.)

Description: Carapace thick and ornamented with a mosaic pattern of reticular ridges with small pores between; other features as for generic description.

Holotype: A right valve; length 0°56 mm.; height, 0°26 mm.

Paratype: A left valve, length 0°50 mm., height 0°22 mm.

Locality: Mestayer Stn. 7, 98 faths., off Big King.

Discussion

The systematic location of Puncia and Manawa is puzzling, as they are so unlike any other known living Ostracoda. No ordinal placing has been attempted, as the sub-orders of living Ostracoda are based on appendages. These two forms evidently form a natural group and have essentially the same type of structure, except for the bracket-like teeth of Manawa, which at first suggest possible Cytherid affinities. This type of bracket-hinge appears, however, to be quite unique amongst the Ostracoda and quite different from anything seen in the Cytheridae, which have a tooth-and-socket arrangement comparable to the bivalve Mollusca. The “frilled” structure of the valves is paralleled only by the Palaeozoic Eurychilininae in which the frill is interpreted as a double-walled, septate structure by Swartz (in litt.),

Picture icon

Figs. 1–6–Puncia novozealandica n.gen., n.sp.
Fig. 1—light valve (holotype), × 135. Kg. 2–right valve, dorsal aspect (holotype), × 105. l fig. 3–right valve (holotype), × 95.
Fig. 4–juvenile, left valve, × 105. Fig. 5–right valve, internal lateral aspect (holotype), × 105. Fig. 0–juvenile, and view, × 100.
Fig. 7—Eurychilina reticulata, Ulrich. × 20. Ordovician. (After Bassler and Kellett, 1934.)
Figs. 8, 9–Chilobolbina dentifera Bonnemma. Ordovician. Left valves of male and female with brood pouch. × 15. (After Bassler and Kellett, 1934.)

Picture icon

Figs. 1–8–Manawa. tryphena n. gen., n.sp.
Fig. 1–right valve (holotype), × 90. Fig. 2–right valve, internal gen., n.sp.al aspect (holotype), × 80. Fig. 3–left valve (paratype), × 90. Fit.
Fig. 1–right valve (holotype), × 90. Fig. 2–right valve, internal valve, dorsal aspect (paratype), × 90. Gig. 6–right valve, dorsal aspect (holotype), × 90. Fig. 7–right valve, and view (holotype), × 92. Fig. (8–left valve, and view (paratype),-X 90.

– 471 –

who suggests that this internal structure should enhance the significance of the frill as a guide to genetic relationships and together with the associated dimorphic brood pouches should prove to be a major guide in the arrangement of the straight-backed family Beyrichiidae. It may be reasonably conjectured that the frill of Puncia and Manawa is of reproductive significance from analogy with the Eurychilinas, many of which develop an associated brood pouch. There is no evidence of any such brood pouch in the single valve of Puncia examined, but no conclusion can as yet be reached on this point without further material. Puncia, with its straight hinge, median sulcus, radially septate frill and semi-elliptical form with a pronounced backwards swing is a very close duplication of the Palaeozoic genus Chilobolbina and would almost certainly be placed in the subfamily Eurychilininae if it were found in Palaeozoic rocks. In spite of this remarkable resemblance the author is reluctant to assert that a representative of a group of Ostracoda, which apparently died out elsewhere in the Devonian, is still living in the South Pacific. Nevertheless, it does seem at least plausible that this case may be analogous to the discovery of living Crinoids, particularly in view of the fact that the Ostracod fauna of the South Pacific deeps is largely unknown. It may be, on the other hand, that this is a particularly striking example of parallel evolution.

The problem of the orientation of the valves is much the same as in the Beyrichiidae. There seems little reason to try and homologise the blunt or “plenate” end with the plenate end of other modern Ostracoda. The author has tended rather to look for analogy with those Palaeozoic forms with which Puncia seems to have its affinities and he believes that a very strong clue as to the anterior end of Puncia is to be found in the orientation of the dorso-lateral horn like prominences. In modern forms, where alar projections are present, they are backwardly directed, presumably to minimise the resistance of the water when the animals are moving forwards. This strongly suggests that the tendency for the paired prominences in Puncia to lean towards the plenate end is of direct and functional significance in the orientation of the carapace, and in the absence of further evidence the plenate end is here regarded as posterior in both forms. This is in agreement with the conclusions reached by Ulrich and Bassler (1908) from the location of brood pouches in Palaeozoic Ostracoda.

At present these two strange forms are known only from the Three-Kings-North-Cape area in fairly deep water and accompanied by a rich fauna of warm-water species. It seems likely to the author that future dredging operations in the South Pacific will bring to light other members of this family and perhaps specimens with the appendages intact, in which event the problem of the orientation of the Palaeozoic Beyrichiidae may be finally solved.

Literature

Bassler. R. S., and Kellett, Betty, 1934. Bibliographic Index of Paleozoic Ostracoda. Geological Society of America Special Papers, no. 1.

Ulrich, E. O., and Bassler. R. S., 1908. New American Paleozoic Ostracoda; Preliminary Revision of the Beyrichiidae, with Description of New Genera. U.S. Nat. Mus. Pr., vol. 35. p. 280.

– 472 –

The Systematic Arrangement of the New Zealand Galaxiidae
Part II. Specific Classification*.

[Read before the Science Congress, April 22, 1947; received by the Editor, March 4 1947: issued separately, September, 1949.]

Introduction

The present paper must be regarded as merely a preliminary review of the New Zealand species of Galaxias and Neochanna. It has not been possible to re-collect certain recorded species, and it is probable that new ones remain to be described, as much of the country has not yet been worked by collectors. Only those forms that have come under my own observation are described below, and except where otherwise stated the descriptions are based on the examination of ten or more adult specimens. Several apparently valid species, of which specimens are not available, are commented upon in the discussions, and unobserved species that appear to be merely nominal are also noted, but are not included in the synonymies. Mention is thus made of, and a reference given to, every specific name known to have been applied to New Zealand species. For greater convenience in reference these names are brought together in the following list, the order of arrangement being chronological and the present identification being given where determination is possible.

List of Specific Names applied to New Zealand Species of Galaxias
Nominal Species Recorded by Present Identification
alepidotus Bloch and Schneider Bloch and Schneider 1801 alepidotus
fusciatuts Gray Gray 1842 fasciatus
altenuatus Jenyns Jenyns 1842 attenuatus
forsteri (alepidotus) Cuvier A Valenciennes Cuvier & Valenciennes 1846 alepidotus
brocohus Richardson Richardson 1848
reticulatus Richardson Richardson 1848
brevipinnis Gunther Gunther 1866
grandis Haast Haast 1872 alepidotus
olidus Gunther Hutton 1872
campbelli Sauvage Sauvage 1880 campbelli
lynx Hutton Hutton 1896 lynx
kokopu Clarke Clarke 1890 alepidotus
postvectis Clarke Clarke 1899
robinsonii Clarke Clarke 1899
abbreviatus? Clarke 1899
bollansi Hutton Hutton 1901 compbelli
buttoni Regan Regan 1905
burrowsius Phillipps Phillipps 1026 burrowsius

[Footnote] * Part 1. Generic and Subgeneric Classification. Trans, Roy. Soc. N.Z., vol. 75, 1945, pp. 124–137.

– 473 –
burrowsii Phillipps Phillipps 1927 burrowsius
paucispondylus Stokell Stokell 1936 paucinpondylus
castleae Whitley & Phillipps Whitley and Phillipps 1939 lynx
charlottae Whitley & Phillippa Whitley and Phillipps 1939
argenteus Gmelin Whitley and Phillipps 1989
koaro Phillipss Phillipps 1940 koaro
prognathus Stokell Stokell 1940 prognathus

The present investigation has revealed the occurrence of alepidotus and lynx in the North Island, from which they were not previously recorded, and it seems likely that further work will extend the known range of other species. A distributional list of the species dealt with in this paper is given below.

Species common to the North Island and South Island. Galaxias alepidotus fasciatus attenuatus lynx Neochanna apoda

Species known only from the North Island. Galaxias koaro Neochanna diversus

Species known only from the South Island Galaxias burrowsius paucispondylus prognathus vulgaris

Species known only from the Sub-Antarctic Islands. Galaxias campbelli Specific Characters and Methods of Description.

A good deal of the present confusion in Galaxiid taxonomy has resulted from the lack of a uniform system of description, and a loose observance by some systematists of certain basic rules of descriptive ichthyology. So important a dimension as the standard length is variously recorded as the total length, the length of the body and the length of the fish, the first and second of which are recognised terms with distinct meanings both differing from standard length, while the third could be construed most logically as over all length. In the present paper the term standard length is applied to the distance from the tip of the upper jaw to the hypural joint, which is the dimension used in proportionate measurements. The distance from the tip of the upper jaw to the extremity of the middle rays of the caudal fin is designated the total length, and is used in recording the maximum size observed. The length of the head is measured from the tip of the upper jaw to the edge of the opercular flap, as the limit of the bony operculum is not always determinable with certainty.

One of the principal disparities in methods of description occurs in records of the tail proportions. In the early descriptions of species.

– 474 –

of Galaxias the least depth of the tail is compared with the distance from the rear of the dorsal fin to the base of the caudal fin, but the majority of more recent workers substitute for the latter dimension the length of the caudal peduncle which, by reason of the extension of the anal fin base behind that of the dorsal in most species, is usually somewhat the shorter. The newer method has the advantage of dealing with a recognised section of the fish, and is used in similar measurements in many other families of fishes, but it may be urged in favour of the older method that in Galaxiidae the dorsal fin position possesses a particular significance while the location of the anal is quite normal. In the majority of the descriptions dealt with in this revision the measurement has been taken from the rear of the dorsal, and this method is retained in the present work to facilitate comparison.

The point of insertion of the ventral fins is usually recorded as equidistant from the tip of the snout or the base of the caudal fin and some not always easily definable point on the tail or the head. A more accurate method, and one that permits the convenient indication of variation, is to record the decimal fraction of the standard length at which the ventrals are inserted. Similarly the method of indicating the point of insertion of the dorsal fin by recording the number of times the distance from its origin to the base of the caudal is contained in a standard length is cumbersome compared with the direct record of the fraction of the standard length at which the dorsal is inserted.

The position of the anal fin in relation to that of the dorsal, which usually is rather vaguely expressed, can be recorded most satisfactorily by indicating the ray of the most anterior fin opposite which the other originates. In this it is desirable to count all rays however short. The method of dealing with only branched rays or developed rays, while convenient and satisfactory in many families of fishes, is unsuitable in Galaxiidae, as Neochanna apoda is highly inconstant in the proportion of simple to branched rays, and the Tasmanian species auguilliformis Scott (1935) is described as having all simple rays in the dorsal and anal fins.

The number of rays in the dorsal fin is not a character of much value, and is recorded in the present paper principally for the purpose of defining the position at which the anal originates. The number of anal rays is more important, and serves to distinguish attenuatus from all other species found in New Zealand. Enumeration of the pectoral and caudal rays does not seem to be worth while, but the number of ventral rays is an important character in defining a single species (burrowsius Phillipps 1926). The variability of this character has been discussed elsewhere (Stokell, 1945). Other examples of characters that provide distinctions in certain species, but are of no use in separating the majority, are the comparative length of the jaws and the degree of subdivision of the branched rays, of the anal fin. The protrusion of the lower jaw in prognathus Stokell (1940) is sufficient to distinguish this species from all others known in New Zealand. The degree of subdivision of the anal rays varies somewhat within species, and some intergradation occurs, but alepidotus Bloch and Schneider (1801) is distinctive in being the only living form in New Zealand in which the anal rays may be subdivided into 8–10.

– 475 –

Maximum size serves to separate extreme forms but is useless with regard to the majority. Most New Zealand species have a maximum total length of about six inches, but there are two (proynathius and paucispondylus) that never approach this size, and two (alepidotus and fasciatus) that considerably exceed it.

Certain characters tend to separate the New Zealand species into two groups, the most notable of these being the presence of canine teeth in the jaws. In all long-bodied species that have been observed canines are lacking, while the short-bodied and intermediate forms all show definite though not equal enlargement. It must be noted, however, that the few species available from other countries tend to fill the gap and suggest the existence of complete intergradation within the genus. The association of this character with the point of ventral insertion and the development of the pyloric cacca is discussed in an earlier paper (Stokell 1945), which also deals with the number of ventral rays, the point of ventral insertion, the point of anal origin, the lingual and entopterygoidal dentition, the number of vertebrae and the development of the gill-rakers. The number of gill-rakers merits little attention, as the one New Zealand species in which this character was supposed to provide the principal distinction (lynx) has been found to be quite normal.

The number of vertebrae is the most important of all specific characters, its particular value, apart, from the basic nature of all osteological characters, lying in the absence of correlation between it and the majority of the others. It might be expected that the number of vertebrae would be proportionate to the length of the body in relation to the length of the head, but the two characters have been found to be in no way associated. The number of vertebrae and the head in standard-length ratio of the species of Galaxias dealt with in this paper are given below. The number of vertebrae has been obtained by dissection and does not include the hypural.

Table 1—Number of Vertebrae and Head in Standard-length Ratio of New Zealand Species of Galaxias.
Species Number of Vertebrae Head in Standard-length Ratio
paucispondylus 51–53 5.1–5.7
vulgaris 52–55 4.2–4.8
burrowsius 52–55 5.4–6.1
koaro 54–56 3.8–4.7
prognathus 54–56 5.8–6.55
lynx 56–60 4.0–4.5
fasciatus 57–60 3.8–4.5
alepidotus 58–60 3.7–3.8
attenuatus 61–03 5.1–5.7
campbelli 60–64 4.45–4.9

It may be noted that the longest-bodied species (prognathus) has less vertebrae than the shortest bodied one (alepidotus); also that paucispondylus and attenuatus are identical in head in standard-length ratio, but the former has the minimum number of vertebrae and the latter comes within one of the maximum, while in koaro and prognathus, which have identical vertebral counts, the head in standard-length ratios differ greatly.

– 476 –

The size of the mouth is a character of some importance and can be recorded in a satisfactory manner by indicating the position at which the maxillary terminates in relation to the eye.

Certain of the proportionate measurements, such as the point of ventral fin insertion and the head in standard-length ratio, are variable with growth. In all species that have been investigated, the ventral insertion recedes with increase in size. The head in standard-length ratio changes greatly in some species while in others it appears to remain nearly constant. Increase in the size of the fish also causes the caudal fin to undergo a modification in form which, though of little consequence in extreme species, is sufficient to confuse the determination of intermediate ones. Species such as alepidotus and burrowsins have the caudal convex at all stages of growth, while in attienuatus and prognathus it is persistently concave, but in koaro Phillipps (1940) and campbelli Sauvage (1880) the concavity present in all small and moderately sized specimens becomes less pronounced as the fish grows and disappears entirely in very large individuals.

The depth of the body in relation to the standard length is much more variable than any of the characters discussed above, and is omitted from the present descriptions, as are all those characters still to be noted. In addition to the usual variations between individuals the relative depth of the body varies greatly with growth, the state of the reproductive organs, the amount of food in the stomach and the presence of internal parasites. It is usually associated with the length of the head, the longer-headed forms having the greater depth of body, and, quite apart from considerations of differential growth and the other disturbing influences noted, provides little distinction that, is not provided by the head in standard-length ratio.

The size of the eye is extremely variable, and is of little use in separating the New Zealand species except attenuatus, which is sufficiently distinguished by other characters. The majority of the other head measurements are also highly variable and to some extent subject to the personal influence of the recorder.

Colour is of little use in separating the New Zealand species except attenuatus and alepidotus, the first of which has the belly and opereles much more silvery than in any other form, while the second is the only species in which light-coloured spots and short markings are imposed on a dark ground. The majority of the small upland species are mottled in shades of grey, green or brown, the difference between species both as regards colour and pattern being little more than occurs between individuals of a single species from different localities.

Several other characters such as the height and length of the vertical fins are omitted to simplify the descriptions. It is not suggested that these characters are entirely useless as specific distinctions, but it is maintained that those that have been substituted far outweigh them in value. In the absence of adequate material from other countries the present system is based principally on New Zealand species, and it is possible that it will not fully meet the requirements of overseas species when the family is considered as a whole. It seems more desirable for any unusual features of the species of other countries to be provided for in the descriptive systems employed there

– 477 –

than for the present system to be encumbered with minor specifications in anticipation of some of them being of value in respect of some overseas species. The ultimate necessity of showing all species in their correct, relation to each other demands the adoption of a universal system, but the preliminaries to such a work can be carried out most effectively in the countries in which the various species exist, and under descriptive systems according with local peculiarities.

A character that is not made use of in the present system but might be of assistance in separating the Australian species is the number and disposition of large open pores on the head. The usual arrangement on the dorsal surface of the head of New Zealand species is for one pore to be placed on the inner side of each anterior nostril. one on the inner side of each posterior nostril, two pairs in the interorbital space and one pore behind the upper part of each eye transversely in line with the posterior interorbital pair. In attenuatus the pore that is normally postorbital is above the posterior quarter of the eye and in advance of the posterior interorbital pair, while in fasciatus and alepidotus it is above the centre of the eye with the interorbital pair not far behind. The two last species have the second nostril pore in advance of the nostril, a similar arrangement occurring in burrowsius, while in prognathus it is confluent with the nostril. Neochanna apoda has the first and second nostril pores in advance of their respective nostrils but otherwise is normal.

In one Australian species that I have examined the pores are arranged as in the majority of the New Zealand species except that the second nostril pore is somewhat in advance of the nostril, but in another species there is an additional pore on each side so disposed between the rear interorbital pore and the postorbital pore that the three form an are with the concavity forward. Some New Zealand species have a curved row of about seven very small indistinct pores across the snout between the anterior nostrils.

The arrangement of large open pores on the side of the head provides no distinction between species, and is subject to greater variation than the dorsal arrangement. In all species examined there is usually one pore outside, and partly forward, of each anterior nostril, two pre-orbital, one sub-orbital and five or six around the edge of the pre-opercle, the first of which is below the sub-orbital pore and the last behind the centre of the eye.

On the ventral surface there are two large open pores on each side below the lower jaw, and in many species a curved row of small indefinite pores across the chin.

Systematic

The species described below have been divided into three groups, the long-bodied, the short-bodied, and the intermediate, but it must be understood that this division is merely for convenience in presenting the data and implies no suggestion of generic distinctness. This aspect of Galaxiid taxonomy has been dealt with in an earlier paper (Stokell 1945). The location of types is recorded when known, and when the information is second-hand its source is indicated. Apart from type localities the localities recorded are those from which the present material was obtained.

– 478 –

Key to the New Zealand Species

Galaxias
Head, contained more than 5 times in standard length.
Caudal fin concave.
Vertebrae 31–53 pauoispondylus
Vertebrae 54–56 prognathus
Vertebrae 61–63 attenuatus
Caudal fin convex. Vertebrae 52–33 burrowsius
Head contained less than 3 times in standard length.
Pyloric caeca indimentary or lacking.
Vertebrae 57–60 fusciatus
Pyloric caeca well developed. Entopterygoidal teeth weak and irregular.
Vertebrae 54–56 koaro
Entopterygoidal teeth well developed. Caudal fin concave.
Vertebrae 52–55 vulgaris
Vertebrae 56–60 lynx
Vertebrae 59–64 campbelli
Caudal fin convex. Vertebrae 59–60 alepidotus
Neochanna
Head contained more than 5 times in standard length. Teeth in jaws compressed apoda
Teeth in jaws conical diversus

Long-Bodied Group

In addition to having the head contained five or more times in the standard length, the members of this group are characterised by the shortness of the paired fins, the anterior ventral insertion and the absence of canine teeth. This association of characters is not maintained in robinsonii Clarke (1899), which appears to be a valid species but has not come under my observation. The description indicates a long-bodied fish somewhat resembling paucispondylus, but differing in the number of vertebrae (60), the possession of cainine teeth, the longer paired fins and the much greater total length attained. This is recorded as 8.2 in., which is nearly double the maximum observed in paucispondylus, and is a length that has been noted in only two species, i.e., fasciatus and alepidotus. Clarke recorded robinsonii from several Westland localities, but stated that it was rare. My own efforts to re-collect this fish in Frosty Creek and Lake Kanieri, two of the recorded localities, have been unsuccessful.

Galaxias attenuatus Jenyns

It is necessary to draw attention to the questionable distinctness of maculatus, Jenyns 1842), from attenuatus, Jenyns (loc. cit.), and to the nomenclature issues involved, as both names were first published in the same paper, maculatus having page precedence. It is to be noted that attenuatus agrees closely with maculatus in two of its most distinctive characters, namely, the high number of vertebrae and the high number of rays in the anal fin. Local specimens of attenuatus have ii-iv 13–16 rays in the anal fin and 61–63 vertebrae, while the figures given by Regan (1905) for maculatus are iv-v 11–14 anal rays and 62 vertebrae. The principal points of difference revealed by Regan's description are a somewhat longer head and pectoral fin in maculatus. Eigenmann (1928) records the number of ventral fin rays of maculatus as 8–9, but both Jenyns and Regan give the number

– 479 –

as 7. In New Zealand specimens of attenuatus the number ranges from 6–8 and is usually 7. Eigenmann made no record of attenuatus from Chile, from which locality it is recorded by Regan, who included in its range Patagonia, Tierra del Fuego and the Falkland Islands. It is scarcely conceivable that such a fish could have escaped the observation of so thorough a collector as Eigenmann, who obtained his specimens by the use of poison, and it is difficult to reconcile its absence from the west of South America with its plentiful occurrence on the east. The most probable explanation is that the form that Eigenmann recorded from Chile as maculatus is specifically identical with the form recorded by Regan as attenuatus, and that if any difference exists between this fish and the one known to New Zealand workers as attenuatus it is of not greater than sub-specific importance. A similarity of habit is indicated by Eigenmann's (page 14) account of the young of maculatus ascending the rivers in enormous numbers precisely as the young of attenuatus are known to do. It would be unsafe, however, to make any nomenclatural alteration without the examination of South American material, and in the absence of this the name attenuatus will be retained in the present paper.

Galaxias attenuatus Jenyns

Mesites attenuatus Jenyns (Zool., Voy. Beagle, Fish, 1842, p. 121). Galaxias attenuatus Cuvier and Valenciennes (Hist. Nat. Poiss., 23, 1846, p. 348).

Austrocobitis attenuatus Ogilby (Proc. Linn. Soc. N.S.W., 24, 1899, p. 158). Galaxias attenuatus Regan (Proc. Zool. Soc. London, vol. 2, 1905, pp. 365–384).

Austrocobitis attenuatus Whitley (Vic. Nat., 52, no. 3, 1935, p. 41). Galaxias (Galaxias) attenuatus Scott (Proc. Roy. Soc. Tasmania, 1935, p. 90).

B. 5–7 (usually 6); D. iv-v 8–10; A. ii-iv 13–16; V. 6–8 (usually 7); vertebrae 61–63.

Jaws about equal, without canines, entopterygoidal teeth strong, 5–9 on each bone, gill-rakers of moderate length, pyloric caeca short, rudimentary or absent. Maxillary reaching about to anterior of eye. Head 5°1–5°7 in standard length, dorsal inserted at 75–.80 of standard length, least depth of tail 40–.46 of the distance from rear of dorsal to base of caudal. Caudal fin with maximum concavity. Pectoral extending 36–.39 of the distance from its axil to the root of ventral, ventral inserted at .47–51 of the standard length, extending .33–.42 of the distance from its root to the origin of anal, anal originating below origin of dorsal, branched rays of anal usually bifid but occasionally subdivided into 3–4. Maximum total length observed 6 in.

Jenyns' types of the species were originally deposited in the Museum of the University of Cambridge, but were handed over to the British Museum in 1917. Type locality Bay of Islands, New Zealand.

Localities: The material examined during the present investigation was obtained from Papueru in the Auckland province, the Waiau lagoon, Southland, and many intermediate coastal localities.

Habits: In the adult state this species exists in estuaries, river mouths and swampy creeks at altitudes below 800 ft. The early life appears to be spent in the sea but the duration of this period is unknown. When the young fish are about 2 in. in total length they

– 480 –

enter the rivers in the spring and early summer, at which stage they form the whitebait of commerce. McKenzie (1904) records that spawning taken place in tidal water at spring tide, and that the eggs are left uncovered by water until the next spring tide, when hatching occurs. It appears, however, that such conditions are not indispensable to reproduction, as there are indications that at the mouth of the Selwyn River, which is not tidal in consequence of its discharge into a large brackish lagoon, attenuatus spawns in shallow water at the edge of the stream where the eggs remain covered by a constant depth of water.

Galaxias paucispondylus Stokell

Galaxias paucispondylus Stokell (Rec. Cant. Mus., vol. 4, no. 4, 1938, pp. 203–208).

B. 6–7 (usually 6); D. ii-iv 7–8; A. iv-v 7–9; V. 6–8 (usually 7); vertebrae 51–53.

Liower jaw somewhat the shorter, almost straight across the front, jaws without canines, entopterygoidal teeth rather weak, 2–5 on each bone, gill-rakers with maximum development, pyloric caeca short, rudimentary or absent. Maxillary reaching to or scarcely to anterior of eye. Head 5°1–5°7 times in standard length, dorsal inserted at 68–71 of the standard length, least depth of tail –37–48 of the distance from rear of dorsal to base of caudal. Caudal moderately concave with rounded lobes. Pectoral extending 4–5 of the distance from its axil to root of ventral, ventral inserted at .45–.49 of the standard length, extending .45–48 of the distance from its root to origin of anal, anal originating below 4th–6th dorsal ray, branched rays of anal bifid. Maximum total length observed 4°4 in.

Differs from attenuatus in having a lower number of vertebrae, a lower number of anal rays and a more anterior dorsal insertion.

Types: The holotype and ten paratypes are in the Canterbury Museum.

Type Locality: Acheron River, tributary of the Rakaia, Canterbury.

Localities: Acheron River, Harper River, Avoca River, Glenthorne Creek, Cass River.

Phillipps' (1940) reference to this fish as a mud fish is incorrect. Its habitat is rocky and shingly alpine streams at altitudes exceeding 1700 ft. It has not been found in lakes or tributaries of lakes.

Galaxias prognathus Stokell

Galaxias prognathus Stokell (Trans. Roy. Soc. N.Z., vol. 69, 1940. pp. 422–424).

B. 6–8; D. ii-iv 6–8; A. ii-v 7–10; V. 6–8 (usually 7); vertebrae 54–56.

Lower jaw conspicuously protruding, jaws without canines, entopterygoidal teeth weak, uniserial in some specimens, irregularly biserial or grouped in others, gill-rakers rudimentary as in Neochanna apoda, pyloric caeca lacking. Maxillary not extending to anterior of eye. Head 5°8–6°55 in standard length, dorsal inserted at 68°71 of the standard length, least depth of tail 33°39 of the distance from rear of dorsal to base of caudal. Caudal strongly concave with rounded lobes. Pectoral extending 33–43 of the distance

– 481 –

from its axil to root of ventral, ventral inserted .48–.50 of the standard length, extending: .32–.42 of the distance from its root to the origin of anal, anal originating below 4th–5th dorsal ray. branched rays of anal bifid. Maximum total length observed 2.9 in. Differs from attenuatus in the lower number of vertebrae, the lower number of anal rays and the more anterior dorsal insertion; from paucispondylus in the higher number of vertebrae; from both species in the shorter head and the conspicuous protrusion of the lower jaw.

Type: The holotype is in the Canterbury Museum.

Type locality: Wilberforce River, Canterbury.

Localities: The species has been taken from the Wilberforce River, its tributary the Harper, and its tributary the Avoca at altitudes exceeding 1500 ft. It has not been found elsewhere and appears to be rare, as only 18 specimens have been obtained as a result of extensive collecting at various seasons. A favourite habitat is where a side-stream rejoins the main-stream at such a gradient that the water percolates through the builders leaving their upper surfaces dry.

Galaxias burrowsius Phillipps:

Reasons for retaining this fish in the genus Galaxias have been given in an earlier paper (Stokell 1945), and further evidence in support of this course is now presented. The specimens examined during the preparation of the previous paper came from Tinwald, and differed from the original description in usually having teeth on the entopterygoids. They were also inconstant in the lingual dentition and the number of ventral fin rays. It was suggested that a form with constantly toothless entopterygoids might exist in the original locality (West. Oxford), where efforts to re-collect the fish had been unsuccessful, but this suggestion cannot now be entertained. Specimens obtained recently from a drain near Rangiora, which is in the same drainage system as the original locality, agree in dentition with the Tinwald fish. Of three specimens examined, one has toothless entopterygoids, one has one well developed tooth on each bone and the third has one bone toothless and the other carrying one tooth. It has been found that the number of entopterygoidal teeth differs in different species of Galaxias, and that the forms possessing the maximum (lynx, fasciatus) intergrade completely with burrowsius which merely represents the opposite extreme of a range of variation.

The disposition of teeth on the tongue is also variable in the Rangiora fish, two specimens having two rows as in typical Galaxias, while the third has two additional teeth on one side.

Galaxias burrowsius Phillipps

Galaxias burrowsius Phillipps (Trans. N.Z. Inst., vol. 56, 1926, p. 531).

Galaxius burrowsii Phillipps (Bibl. N.Z. Fishes, 1927).

Saxilaga (Liwagasa) burrowsius Scott (Proc. Roy. Soc. Tasmania, 1935, pp. 85–112).

Paragalaxias burrowsii Phillipps (Fishes of N.Z., vol. 1, 1940, p. 35).

B. 6–7; D. iii-iv 7–8 i; A. iii-v 7–8; V. 5–6 (usually 5); vertebrae 52–55.

Jaws about equal or the lower slightly protruding, without definite canines, entopterygoidal teeth usually present but weak and inconstant, 0–3 on each bone, lingual teeth may be disposed in two

– 482 –

rows, four rows or any intermediate arrangement. Gill-rakers short or rudimentary, pyloric caeca long. Maxillary not extending to anterior of eye. Head 5.4–6.1 in standard length, dorsal inserted at 71°72 of the standard length, least depth of tail .6–.9 of the distance from the rear of dorsal to base of caudal, peduncle flanges strongly developed, caudal fin strongly convex. Pectoral fin extending .25–.34 of the distance from its axil to the ventral, ventral inserted at .49–.51 of the standard length, extending .25–.30 of the distance from its root to the origin of anal, anal originating below 2nd–4th ray of dorsal, branched rays of anal bifid.

Last ray of dorsal and anal may be simple. Maximum total length observed 4.5 in.

Differs from the three preceding species in the lower number of ventral rays, the inconstant development of the entopterygoidal teeth, the convex caudal fin and the strong development of the caudal peduncle flanges.

Type: The original description makes no mention of a type, but if any type exists it should be in the Dominion Museum.

Localities: Specimens were obtained from a small creek on Mr. J. Merrin's farm which flows into the Ashburton River from the south side several miles below Tinwald, and from a drain on Mr. Judson's farm near Rangiora. The former locality is about 300 ft. above sea-level and the latter about 90 ft., but the late Mr. Burrows' farm at West Oxford, from which the species was first recorded, has an altitude of about 800 ft.

Habits: This species is able to survive a drought by burying itself in the mud in the same manner as Neochanna; but there is no reason to suppose that it adopts a mud-dwelling existence unless forced to do so. The Tinwald specimens were netted from mud-bottomed pools and from flowing water with a shingle bottom, in both of which habitats they were associated with the Eleotrid Philypnodon breviceps. No evidence of a mud-dwelling habit was observed in this locality, but at Rangiora specimens were dug up out of the damp earth and detritus at the bottom of a drain that had been dry for over a month. They immediately became active on being placed in water. The skeletons of several small eels which appeared to have died from lack of water were found in the drain.

Intermediate Group

A group of closely related species, intermediate in body length, possessing canine teeth, moderate to long paired fins and having a posterior ventral insertion presents the greatest difficulty in the classification of the New Zealand Galaxiidae. The forms composing this group are separated principally by the number of vertebrae, particulars of which are given in the following frequency table:

Species Locality No. of Specimens Number of Vertebrae
52 53 54 55 56 57 58 59 60 61 62 63 64
vulgaris Cant. streams 36 2 12 17 5
koaro L. Rotoaira 13 5 5 3
koaro L. Taupo 19 4 7 8
lynx L. Coleridge 28 3 5 15 3 2
campbelli Auckland Isls. 15 1 2 8 3 1
campbelli Campbell Isl. 14 1 3 4 4 1 1
– 483 –

Although there is overlap throughout, it will be noted that a different peak occurs in all species. Combining the locality groups the data may be presented graphically as in Text Figure 1.

The distinctness of vulgaris, lynx said campbelli is demonstrated by the wide separation and tolerably symmetrical form of the curves representing these species, while the difference in variation range and frequency in koaro, as compared with the others, forms a specific character of value. In allocating the present names considerable use has been made of data supplied by Dr. Ethelwynn Trewavas, of the British Museum, who very kindly made X-ray examinations of the types of several species for the purpose of determining the number of vertebrae and also provided other particulars, but the responsibility for the final identification rests entirely with myself.

Galaxias lynx Hutton

The original description of this species (Hutton 1896) is of little definitive value, but the identity of the present fish is established by the circumstance of exclusive occurrence in the type locality (Lake Coleridge) and by comparison with the type. Extensive collecting in Lake Coleridge and the streams and lakes within its natural catchment area has failed to produce more than one species of Galaxias, although three other species occur in nearby streams in other drainage systems. Two of these streams, the Harper and the Acheron, have been diverted into Lake Coleridge to augment the water supply for the hydroelectric power station, with the result that there now exist four species of Galaxias in water connected with the lake. There is as yet no evidence of a transference of the forms native to the Harper and Acheron to the natural feeders of Lake Coleridge, but at certain seasons there is a considerable incursion of lynx into the lower Harper.

Picture icon

Text Figure 1.
Number of Vertebrae

– 484 –

This species is plentiful in Lake Lilian, which drains into the Harper, and may have occurred naturally in this water before the diversion. Hutton's original description makes no mention of a type of lynx, but it appears that one was established, probably subsequent to publication. A jar in the Canterbury Museum containing one large specimen of about 5.7 in. in. total length, which is mounted on a glass support, and several young of about 2 in., which are lying in the bottom of the jar, is labelled “Holotype Galaxias lynx Hutton (Trans. N.Z. Inst., vol. 28, 1896). Lake Coleridge.” There appears to be no doubt that the large fish is the type of lynx. Particulars of this fish are: Number of vertebrae (as determined by X-ray examination) 56, head in standard-length ratio 4.5, dorsal fin inserted at .72 of the standard length, ventrals at 53 of same, pectoral extending .53 of the distance from its axil to the ventral, all of which specifications come within the range of variation of the Lake Coleridge species.

Regan's distinctive character of 12–14 gill-rakers on the lower limb of the anterior gill arch has been found to be incorrect. Hutton's type has 11, including one located in the angle of the arch, and no higher number was observed during the examination of 25 Coleridge specimens, the minimum in which was 9. There would thus appear to be some doubt regarding the identity of the specimens described by Regan as lynx, and it seems advisable to omit this reference from the present synonymy. It is to be noted that Hutton identified with lynx a form from Lake Wakatipu which he had earlier (1872) referred to olidus Gunther (1866). The most unusual feature of the Wakatipu fish is the colouration, which is described as pale yellow with minute scattered spots, an arrangement that has not been encountered in lynx from any locality but is indicated in Regan's figure and description. It would thus appear to be possible that the Wakatipu fish is not identical with lynx and that Regan's description is based, in part at least, on an extraneous species.

In 1939 Whitley and Phillipps described a fish from Lake Waikaremoana in the North Island under the name of castleae, the description being based on juvenile examples 48 mm. in length. Examination of specimens from this locality shows that the fish is lynx. The number of vertebrae ranges from 56 to 60, and in other characters the fish agrees with the Coleridge species.

A species having a head in standard-length ratio of about 5 was described by Regan (1905) under the name of huttoni. The locality given is Lake Rainiera, New Zealand, but this appears to be a misspelling of some local name, as no such water is known in this country. The difficulty in identifying any fish with huttoni is increased by certain points of disagreement between Regan's description and the figure presented by him, notably in the caudal peduncle proportions and the length of the ventral fins. A re-examination of one of the types has been made by Dr. Trewavas, who states that the largest is only 45 mm. in total length. The bones do not show in the X-ray negative forwarded, but a count of vertebrae was obtained later by staining. Particulars of the specimen examined are: Vertebrae 61, dorsal rays (all counted) 13, anal 16, pectoral 12, ventrals 7–7, dorsal insertion (computed from measurements supplied) at .76 of the standard length, ventrals at .47 of same, no enlargement of lateral

– 485 –

mandibular teeth. This number of vertebrae has been found in only two New Zealand species, attenuatus (61–63) and campbelli (59–64). The dorsal insertion of huttoni agrees with that of attenuatus, but is farther to the rear than in campbelli. The ventral insertion provides no distinction, as in this character campbelli exhibits considerable variation with growth. In a group of 9 juvenile specimens of campbelli, the largest of which has a total length of 59 mm., the ventral insertion averages 474, which comes within the range of attenuatus. The head in standard-length ratio of huttoni approaches agreement with that of campbelli, but is insufficient for attenuatus. A distinctive character excluding huttoni from attenuatus is the point of anal origin. In attenuatus the origin of the anal is opposite that of the dorsal, but in the X-ray photograph of huttoni it is about halfway along the dorsal base. In this feature huttoni agrees satisfactorily with campbelli. The number of anal rays of huttoni represents about the minimum in attenuatus and the maximum in campbelli. No importance attaches to the particulars of the mandibular dentition in huttoni, as similar-sized specimens of large-toothed forms such as lynx frequently show no lateral enlargement.

It is therefore clear that huttoni is not referable to attenuatus, and, although it agrees tolerably closely with campbelli, the suggestion of identity with this species is negatived by the circumstances of locality and habit. There are no named lakes in the Auckland Islands and Campbell Island, which are the only known localities of campbelli, and the species occurring there is stream-dwelling.

It seems likely that the locality of huttoni is in the South Island, as the late Professor Hutton, who apparently provided the material upon which the species is based, was at the Canterbury Museum during the relevant period. The only species known from the upland lakes of the South Island is lynx, the highest number of vertebrae observed in which is 60, but in view of the different method of enumeration is huttoni and the poorness of the specimens there seems a possibility that the two are identical. It has been suggested that Lake Kanieri in Westland might be the original of “Lake Rainiera,” but an investigation of this locality failed to reveal the presence of any Galaxiid, the only fishes obtained from the lake and associated streams being Gobiomorphus basalis, Salmo trutta and Perca fluviatilis. The examination was made in February in anticipation of the small fish shoaling at the mouths of tributaries in the manner of lynx. In this species the adults and the larval young are found in tributary streams, but small fish up to 50 mm. total length occur in the lake. These small fish represent the “whitebait” stage of lynx, and when running into tributaries of the Southern Lakes are actually referred to by local residents as “mountain whitebait” or “freshwater whitebait.” A similar “whitebait” stage occurs in koaro at Lake Taupo. The seven co-types of huttoni as shown in the X-ray negative obviously represent a sample of such “mountain whitebait,” the failure to obtain which from Lake Kanieri discredits the suggestion that huttoni came from there. This lake is only 430 ft. above the sea, and the temperature taken in shade near the surface was found to reach 70° F. The lowland character of the fauna is indicated by the absence of the Eleotrid Philypnodon breviceps.

– 486 –

It seems fitting to observe here that descriptive zoology is not advanced by the creation of species based on juvenile examples. In a family such as Galaxiidae, in which differential growth is prevalent, but is not subject to any universal law, no satisfactory estimation of the adult can be made from a consideration of juvenile characters. The natural features of the group render its classification sufficiently difficult without the introduction of artificial problems such as result from the creation of inadequately based species.

Galaxias lynx Hutton

Galaxias lynx Hutton (Trans. N.Z. Inst., 28, 1896, p. 317).

Galaxias castleae Whitely and Phillipps (Trans. Roy. Soc. N.Z.), 69, 1939, p. 228).

B. 6–9; D. iii–v 5–8; A. v–vi 9–10; V. 6–7 (usually 7); vertebrae 56–60.

Lower jaw the shorter, both jaws with strong lateral canines, entopterygoidal teeth rather weak, 7–9 on each bone, gill-rakers of moderate length, pyloric caeca strongly developed. Maxillary extending about to posterior margin of eye. Head 4.0–4.5 in standard length, dorsal fin inserted at 72–74 of the standard length, least depth of tail 50–65 of distance from rear of dorsal to base of caudal. Caudal fin moderately concave with rounded lobes. Pectoral extending 5–7 of the distance from its axil to the ventral, ventral inserted at .49–.65 of the standard length extending .6–.8 of the distance from its root to the anal, anal originating beneath 4th–6th dorsal ray, branched rays of anal sometimes divided into 4.

Maximum total length observed 5.9 in.

Type: Hutton's holotype is in Canterbury Museum.

Type locality: Lake Coleridge, Canterbury.

Localities: Lake Coleridge and its tributary streams and lakes, Lake Lilian, Harper River, Lake Pearson and the Craigieburn Creek, Lake Marymere, Lake Hawdon, tributary of Lake Brunner, Upper Ashburton River, Lake Hawea, Lake Wanaka, Lake Waikaremoana. This is the only South Island form known from alpine lakes and their tributaries. It may extend for a considerable distance down the outlet streams of lakes and may be associated there with purely river-dwelling species, but it has not been found in streams that are not connected with lakes.

Galaxias koaro Phillipps.

A species from the thermal lakes is closely related to lynx but has fewer vertebrae and a narrower range of variation (54–56). Phillipps (1940) described this fish from Lake Rotoaira under the name of koaro, and distinguished it from brevipinnis Gunther (1866), which he recorded as occurring in Lake Taupo, by its shorter snout and fewer branchiostegals (6). I have examined several groups from each of these waters and have found them to consist of a single species. The Taupo specimens possess 54–56 vertebrae and agree also in other characters with the Rotoaira fish. The number of branchiostegals is 7–9 in each. The specific name koaro applies therefore to the only Taupo form with which I am acquainted, as the low number of vertebrae disqualifies the fish for admission to brevipinnis, in which the number is recorded by Gunther as 65. The species brevipinnis appears to have been described from four specimens, and, as

– 487 –

far as I am aware, has never been re-collected and independently re-described, the descriptions published in this country being more or less composite ones in which up to seven species are involved. It appears also that Regains description of brevipinnis is based on material including more than one species, as he has considerably extended the original specification in respect of certain characters and has recorded the fish from localities (Dunedin, North Island) where other forms are plentiful, but from which no species agreeing with the specification of brevipinnis can be obtained. The prevailing uncertainty regarding the affinities of brevipinnis is illustrated by the widely differing forms that have been regarded as identifiable with this species. Clarke's (1899) species robinsonii, which is described as being short-headed, elongated and possessed of a long caudal peduncle, was referred to brevipinnis by Regan, whose synonymy includes grandis Haast (1872), which approaches or attains the opposite extreme in these characters. In addition to the high number of vertebrae already mentioned as a feature of brevipinnis this species is described by Gunther as having a pectoral fin extending half or rather less than half of the distance from its root to the ventral, a ventral terminating at a great distance from the vent, a rounded caudal, an obtuse snout, a head in standard-length ratio of 5, dentition similar to that of fasciatus, while the least depth of the tail is said to be two-thirds of the distance from the dorsal fin to the caudal fin. I have been unable to find such a fish in New Zealand waters or in museum collections, but have not had much experience of the locality from which the species appears to have come. The original material was presented to the British Museum by Captain Stokes, Master of the Acheron, which was engaged in marine survey work at Southland and Westland during 1847–51. The suggestion that brevipinnis came from near the sea within this territory is supported by the circumstance that Captain Stokes also presented specimens of fasciatus and alepidotus, both of which are lowland species occurring in the localities mentioned.

Another name applied to a North Island species is abbreviatus by which Clarke (1899) referred to one of the species that he had been able to find in the North Island, the other being fasciatus. I am unable to find any account of such a species, and think it probable that abbreviatus is a misprint of attenuatus.

Galaxias koaro Phillipps

Galaxias huttoni Phillipps (N. Z. Journ. Sci. and Tech., 4, 1921, p. 119).

Galaxias brevipinnis Phillipps (Fishes of N. Z., vol. 1, 1940, p. 21).

Galaxias koaro Phillipps (Fishes of N.Z., vol. 1, 1940, p. 35).

B. 7–9; D. iv–v 6–7; A. iii–iv 8–10; V. 7; vertebrae 54–56.

Jaws about equal, with moderately enlarged lateral canines, entopterygoidal teeth variable in size and number, individuals may have one tooth on each bone, a row of seven on each bone or any intermediate arrangement, frequently a number of teeth poorly developed and occasionally one or two abnormally enlarged. Gill-rakers moderate, pyloric caeca strongly developed. Maxilliary extending about to middle of eye. Head 3°8–4°7 in standard length, dorsal inserted at .70–.77 of standard length, least depth of tail .5–.6 of the distance from rear of dorsal to base of caudal. Caudal fin moderately concave

– 488 –

with rounded lobes, becoming convex or almost so in old specimens. Pectoral extending .45–.60 of the distance from its axil to the ventral, ventral inserted at .52–.56 of the standard length, extending .51–.64 of the distance from its origin to origin of anal, anal originating below 5th–6th dorsal ray. Branched rays of anal usually bifid, but sometimes divided into 3–4. Maximum total length observed 5.8 inches.

Differs from lynx in having fewer vertebrae, and in the irregular number and development of the entopterygoidal teeth.

Type.: The original description makes no mention of a type, but if any type has been established it will presumably be in the Dominion Museum.

Localities: Lake Rotoaira, Lake Taupo, Lake Roto Pounamu and the Waikato River.

Galaxias compbelli Sauvage

The only species known from the Auckland Islands and Campbell Island closely resembles lynx, but is distinguished by a higher number of vertebrae (59–64). Four species have been described from these localities, the earliest being brocchus Richardson (1848) and reticulatus Richardson (loc. cit.), both of which are described as being short bodied (head in standard length 4) with rounded caudal fins. In the Auckland Islands fish the body becomes relatively shorter and the caudal fin loses its concavity with increase in size, and it is conceivable that specimens of the size mentioned by Richardson (8 ½ in. brocchus, 7 ½ in reticulatus) would approach agreement with his descriptions, but the attainment of such sizes seems extremely unlikely. The various collections that I have examined aggregate over 70 specimens, which is a much greater amount of material than appears to have been available to Richardson, but the largest individual observed measures only 5.9 in. Whether these discrepancies are owing to the vagaries of collecting or to some transposition of labels in the Erebus and Terror material cannot be determined here. A re-examination of Richardson's types is necessary and, pending this, it is advisable to exclude brocchus and reticulatus from the possible choice of names for the present fish.

In 1880 Sauvage described a specimen 60 m m. in length from Campbell Island under the name of campbelli, the description indicating a long-bodied fish (head in standard length 6.5). Apparently this feature induced Regan to refer the species to attenuatus, notwithstanding the excessive length of the pectoral fin, which Sauvage recorded as extending half of the distance to the ventral. Through the kindness of Dr. Leon Bertin, of the Museum National d'Histoire Naturelle, Paris, I have been able to examine one of the five co-types of campbelli, all of which are said to be small. This fish, which is 40.5 mm: in total length, proves to be a juvenile example of the common Auckland Islands and Campbell Island form. The vertebrae number 60, the anal rays 14 (all counted), the branchiostegals 7–7, the pectoral fin extends .58 of the distance from its axil to the ventral, the ventral is inserted at .49 of the standard length, it extends .62 of the distance to the anal, and the head is contained 5.4 in standard length, all of which specifications come within the range of the present form. The name campbelli is therefore the earliest that has been

– 489 –

definitely associated with the only species known from the Auckland Islands and Campbell Island, and is used in the present paper.

In 1901, Hutton based the species bollansi on a single specimen found in the throat of a merganser taken at the Auckland Islands. This fish is now in the British Museum, and has been subjected to X-ray examination by Dr. Trewavas, who kindly forwarded particulars of the specimen in addition to the negative. These indicate a fish 4.2 in. in total length with 62 vertebrae, 12 dorsal fin rays (all counted) and 13 anal rays. These figures are typical of the Auckland Islands fish, and there is also satisfactory agreement in the proportionate measurements. The head in standard-length ratio of the type is somewhat lower than has been observed in specimens from the Aucklands Islands and Campbell Island, but this appears to be explained by the undulations visible in the vertebral column which suggest that the body has been shortened as a result of compression in the bird's throat. The name bollansi therefore becomes a synonym of campbelli.

Regan (1905) referred the Auckland Islands fish to brevipinnis, and this course was followed by Waite (1909), who dealt with the material collected by the expedition organised by the Canterbury Philosophical Institute in 1907. In the original description of brevipinnis the pectoral fin is said to extend half or rather less than half of the distance from its root to the ventral, but in the Auckland Islands fish, the pectoral normally extends more than half of this distance. The examination of 25 specimens revealed only one in which the ratio was less (.45), but this fish had the abdomen abnormally distended with the result that the ventral fins had been forced downward and backward, thus unnaturally increasing the pectoral-ventral interval. The condition of this fish resembled that of individuals that have been found to be harbouring large cestodes in the body cavity, and as it was not dissected it was deemed advisable to exclude it from the material used for description.

The largest specimens of the Auckland Islands fish agree with the, description of brevipinnis in so far that the caudal fin is truncated, or almost so, but adult fish of 4.5 in. in total length have the caudal concave with rounded lobes. A lack of agreement is to be noted in the form of the snout, which is less obtuse than in most New Zealand species.

Gunther's statement that the ventral fin of brevipinnis “terminates at a great distance from the vent” is rather indefinite, but may be interpreted by analogy with fasciatus, which is described by the same author as having the ventral termination “a short distance from the vent.” Measurement of specimens of fasciatus from several localities has shown that in this species the distance from the tip of the ventral to the vent is .18– 33 of the distance from the origin of the ventral to the vent, while in the Auckland Islands fish the corresponding ratio has been found to be .28–.40. The distances represented by these values are too similar for Gunther to have made such a distinction between them as to have referred to one as short and the other as great. In the truly short-finned forms, such as attenuatus the corresponding value is about 6.

The Auckland Islands fish agrees completely with the description

– 490 –

of brevipinnis in dentition, and comes nearest to the latter species in the number of vertebrae, but it must be emphasised that no specimen with 65 vertebrae has been found and that 64 occurred only once in 29 specimens. The dominant number is 61. It is also to be noted that in the original description of brevipinnis the locality is recorded as New Zealand, no mention being made of outlying islands, and that, so far as is known, the Aucklands Islands fish does not occur on the mainland.

Galaxias campbelli Sauvage

Galaxias campbelli Sauvage (Bull. Soc. Philomatique, ser. 7, vol. 4, 1880, p. 229.

Galaxias bollansi Hutton (Trans. N.Z. Inst., 34, 1901).

Galaxias brevipinnis Regan (Pro. Zool. Soc. London, 2, 1905).

Galaxias brevipinnis Waite (Sub-Antarctic Islands of N.Z., 2, 1909, p. 586).

B. 7–9; D. iv–v 7–9; A. v–vi 8–9; V. 7; vertebrae 59–64.

Lower jaw somewhat the shorter, with strong lateral canines, upper with moderately enlarged canines, entopterygoidal teeth moderate, 6–7 on each bone, gill-rakers moderate, pyloric caeca well developed. Maxillary extending to middle of eye. Head 4.45–4.9 in standard length, dorsal fin inserted at .70–.75 of the standard length, least depth of tail .5–.6 of the distance from rear of dorsal to base of caudal. Caudal fin moderately concave with rounded lobes, becoming almost convex in old specimens. Pectoral extending .54–.68 of the distance from its axil to the origin of the ventral, ventral inserted at .48–.53 of the standard length, extending .60–.68 of the distance from its origin to origin of anal, anal originating below 5th–6th dorsal ray. Branched rays of anal bifid. Maximum total length observed 5.9 in.

Differs from lynx in the greater number of vertebrae; from koara in the same feature, and also in the constant and strongly developed entopterygoidal teeth.

Types: Four co-types of campbelli are in the Museum National d'Histoire Naturelle, Paris, and one in my own collection.

Type locality: Campbell Island.

Localities: Several streams on Auckland Islands and Campbell Island. Also recorded from water that is brackish at high tide, Chambers Inlet, Auckland Islands.

Galaxias vulgaris n.sp.

The most abundant and widely distributed of the upland species occurring in Canterbury has never been described, and is commonly regarded as being referable to brevipinnis, from which it is sharply differentiated by the number of vertebrae (52–55). Superficially the Canterbury fish agrees best with olidus, which Gunther (1866) based on a single specimen of questionable locality. An X-ray examination of the type shows that the species is based on a deformed specimen with a number of vertebrae fused so that a dependable count is not possible. Dr. Trewavas states that in the jar with the type there is a headless specimen in which, however, the vertebral column is intact, and which has 57 vertebrae. I am unable to determine from an examination of the X-ray negatives if these two specimens are specifically identical or distinct, but the position of the Canterbury fish is in no way affected by this circumstance. If the two are regarded

– 491 –

as identical, and the headless specimen is accepted as practically a co-type of olidus, the number of vertebrae is too high for the admission of the present form. On the other hand, if they are regarded as distinct, the sole representative of the species olidus is a deformed specimen of uncertain locality with which it is impossible to identify any fish. The Canterbury fish is therefore described as a new species under the name of vulgaris.

Galaxias vulgaris n.sp.

B. 6–8; D. iii–v 6–7; A. iv–v 8–9; V. 6–8 (usually 7); vertebrae 52–55.

Lower jaw slightly the shorter, both with moderately enlarged lateral canines, entopterygoidal teeth strong, 6–7 on each bone, gillrakers long, pyloric caeca well developed, maxillary extending to or beyond middle of eye. Head 4.2–4.8 in standard length, dorsal fin inserted at .71–.474 of the standard length, least depth of tail .5–.6 of the distance from rear of dorsal to base of caudal. Caudal fin slightly to moderately concave with rounded lobes. Pectoral extending .4–.5 of the distance from its axil to the origin of ventral, ventral inserted at .504–.54 of the standard length, extending .47–.60 of the distance from its origin to origin of anal, anal origin below 4th-7th dorsal ray. Branched rays of anal bifid.

Maximum total length observed 5.4 in.

Differs from lynx and campbelli in having fewer vertebrae and shorter pectoral fins; from koaro in having somewhat fewer vertebrae and a wider range of variation, the constant occurrence of strongly developed entopterygoidal teeth, and the exclusive stream-dwelling habit.

Types: Holotype in Canterbury Museum. Paratype in Dominion Museum.

Type locality: Rubicon River, Springfield, Canterbury, altitude 1,500 ft.

The species is named on account of its abundant occurrence in the upland streams of Canterbury.

Localities: Numerous streams at altitudes exceeding 800 ft. within the basins of the Hinds, Rakaia, Selwyn, Waimakariri, Ashley, Hurunui and Waiau rivers. Not found in lakes or tributaries of lakes.

In 1924, considerable numbers of this species were collected from the Rubicon River and liberated in Lake Rubicon, which previously contained no indigenous fishes. This lake, which is only about 10 acres in extent, has been ponded at the head of a short valley by the Rubicon River shingle fan. It has no outlet and is fed by soakage from several very small streams which seldom make direct connection with the lake. During the following year specimens were seen occasionally, but no evidence of reproduction was noted. A very rainy season followed, and during a period of connected water the Galaxias abandoned the lake and established themselves in the largest tributary, which previously was fishless, where they reproduced and have persisted ever since.

Short-Bodied Group.
Galaxias fasciatus Gray.

Whitley and Phillipps (1939) state that the correct name of this species is G. angenteus Gmelin, and give a list of synonymous names,

– 492 –

one of which (G. grandis Clarke) I am unable to find any account of, while another (Esox alepidotus Bloch and Schneider) is that of a distinct species. Gmelin's account is not available to me, and owing to the unfortunate circumstance of Whitley and Phillipps having regarded fasciatus and alepidotus as identical, there is nothing to indicate which, if either, of these names argenteus is prior to. This name is singularly inappropriate for either of the above-mentioned species, and seems unlikely to have been applied to any New Zealand Galaxias except attenuatus, the silver belly and opercles of which may have suggested it.

A Westland species described by Clarke (1899) under the name of postvectis was referred to fasciatus by Regan. It may be observed that Clarke pointed out several structural differences between his fish and fasciatus, which latter species was well known to him and was re-described in his paper. Two short-bodied species, brocchus and reticulatus, were described by Richardson (1848) from the Auckland Islands, but were identified with fasciatus by Gunther (1866) and later by Regan, both of whom had the advantage of access to the types. So far as reticulatus is concerned this identification is scarcely satisfactory, as the tail in this species is considerably longer than in any specimen of fasciatus in the present collection. As already noted under the heading of campbelli, neither broechus nor reticulatus agrees with the only form known from the Auckland Islands.

In 1939 Whitley and Phillipps described a fish from Queen Charlotte Sound under the name of charlottae. The description indicates a large short-bodied fish closely resembling fasciatus, but differing in having a somewhat more anterior ventral fin insertion and in the absence of canine teeth. The colouration, which is described as uniform, provides no distinction, and the slightly concave caudal fin figured in charlottae is a feature of fasciatus. A jar in the Dominion Museum labelled “Galaxias sp.” with the recorded locality and collector agreeing with those of charlottae, contains two specimens, one of which is almost uniform in colour and may be the type, but in the present state of disorganisation resulting from the occupation of the Museum by the Air Force it is not possible to ascertain if this is so. These specimen's have canine teeth and appear from an external examination to be fasciatus.

Galaxias fasciatus Gray

Galaxias fasciatus Gray (Zool. Miscel., 1842, p. 73).

Galaxias fasciatus Regan (Pro. Zool. Soc. London, 1905, pp. 363–384.)

Galaxias argenteus Whitely and Phillipps (Trans. Roy. Soc. N. Z., 69, 1939.

B. 8–9; D. iv–v, 7–9,; A. iii–v 10–12; V. 7; vertebrae 57–60.

Jaws about equal, lower with conspicuous lateral and anterior canines, upper with 2–3 canines on each side. Entopterygoidal teeth usually enlarged anteriorly, 6–10 on each bone. Gill-rakers long, pyloric caeca rudimentary or absent. Maxillary extending to posterior third of eye. Head 3.8–4 5 in standard length, dorsal inserted at .74–.78 of standard length, least depth of tail 9–1.02 of the distance from rear of dorsal to base of caudal, caudal caudal slightly concave or rounded. Pectoral extending .52–63 of the distance from its axil to root of ventral, ventral inserted at .52–53 of the standard length, extending .65–.68 of the distance from its root of origin of anal, anal originating

– 493 –

below 1st–3rd dorsal ray, branched rays of anal divided into 3–4. Maximum total length observed 9.25 in.

Types: Gray's types of the species are recorded by Regan as being in the British Museum.

Localities: The present material was obtained from Westland. Banks Peninsula, Nelson, Wairarapa, the Chickens Islands, and Stewart Island.

Galaxias alepidotus Bloch and Schneider

This species was noted by Forster in 1777, which appears to be the firs record of a Galaxiid from any country. It does not seem to have received the attention from collectors and systematists that appears to be its due as the first recorded species, and this seems the more remarkable in view of its conspicuous size and plentiful occurrence in lowland waters up to the close of last century. It is still taken occasionally from creeks in the neighbourhood of Lake Ellesmere, from which locality Haast (1872) described it as a new species under the name of grandis. Although Haast's species, as described, differs somewhat from the present form, and no type or specimen agreeing in length with the one described by him can be found in the Canterbury Museum, there is no doubt that they are identical. Apart from attenuatus the species described below is the only one known to have inhabited the locality concerned. The length of Haast's specimen, which is given as 19 in., three lines is a further indication, of the identity of this fish, as the only form known to attain this length is the one represented by the present Fig. 10. The original of this figure is 17 in. in total length, and Clarke records a maximum length of 23 in. and a weight of 6 lb. in Westland specimens. Haast's fish is further identified with the form shown in Fig. 10 by its colour, which is described as “brownish black above, yellowish brown beneath, with yellowish spots and short streaks.” The species alepidotus is one of the few in which coloration provides a useful distinguishing character, the pale-yellow spots, streaks and crescents on a dark ground being in sharp contrast to the arrangement in other spotted species such as truttaceus of Tasmania, in which dark spots are imposed on a light ground. Haast also recorded the species grandis from “Lake Hall, the outlet of which falls into the Paringa River” in Westland, from which locality it was obtained by Clarke who, in 1899, described it under the name of kokopu, his description, which fortunately includes the number of vertebrae, agreeing with the form shown in Fig. 10 of the present paper.

Galaxias alepidotus Bloch and Schneider

Esox alepidotus Bloch and Schneider (Syst. Ichth., 1801).

Galaxias alepidotus Cuvier (Regne. Anim., 1817).

Galaxias forsteri (alepidotus) Cuvier and Valenciennes (Hist. Nat. Poiss., 23, 1846, p. 351).

Galaxias grandis Haast (Trans. N.Z. Inst., 5, 1872, p. 278).

Galaxias kokopu Clarke (Trans. N.Z. Inst., 31, 1899, p. 84, pl. 4).

Galaxias alepidotus Regan (Proc. Zool. Soc. London, 2, 1905, pp. 363–384) Galaxias argenteus Whitley and Phillipps (Trans. Roy. Soc. N.Z., 69, 1939, p. 230).

Galaxias fasciatus Phillipps (Fishes of N.Z., 1, 1940, p. 21).

B. 8–8; D. iii–v 8; A. iv–v 10–12; V. 7; vertebrae 58–60.

– 494 –

Lower jaw somewhat the shorter, with well-developed lateral and anterior canines, upper with progressive enlargement of lateral teeth followed by abrupt reduction. Entopterygoidal teeth strong, 7–9 on each bone, gill-rakers well developed, pyloric caeca long. Maxillary extending to posterior third of eye. Head 3.7–3 8 in standard length, dorsal inserted at .74–.75 of the standard length, least depth of tail equal to or more than distance from, rear of dorsal to base of caudal, caudal rounded. Pectoral extending .53–.57 of the distance from its axil to root of ventral, ventral inserted at .53–.54 of the standard length, extending .52–.68 of the distance from its root to the origin of the anal, anal originating below 2nd–4th dorsal ray, branched rays of anal subdivided into 8–10 in large specimens.

Maximum total length observed 17 in.

Differs from fasciatus in the possession of strongly developed pyloric caeca, the anal rays being finely sub-divided and the head being longer; also, in the much greater size attained.

I am unaware of the existence of a type.

The present description is based on the dissection of five specimens—two from creeks flowing into Lake Ellesmere, one from Southland, one from Wairarapa, and one recorded from the Hutt District, Wellington. Specimens in museum collections have also been examined externally.

Neochanna Gunther.

Neochanna apoda Gunther

Neochanna apoda Gunther (Ann. Mag. Nat. Hist., 20, 1867, p. 305).

Neochanna apoda Regan (Proc. Zool. Soc. London, 1905).

B. 6–7; D. x–xiii 4–7 i = 18–19: A. xiv–xix 0–3 i = 18–19; vertebrae 55–57.

Jaws about equal, profile of snout only slightly convex. Majority of the teeth in both jaws compressed, with convex cutting edges, a few at rear may be conical or almost so, uniserial, no enlargement of lateral teeth, lingual teeth conical, hooked, in two rows, entopterygoids toothless. Gill-rakers rudimentary, or moderately developed, pyloric caeca well developed, 1–2. Maxillary extending about to posterior of eye. Head 4.9–5.6 in standard length, dorsal fin inserted at .72–.74 of same, extremely fleshy, tail with strongly developed dorsal and ventral ridges continuous with dorsal, caudal, and anal fins, its least depth, including ridges, 1.4–2.0 times the distance from rear of dorsal fin to hypural joint, caudal fin strongly convex. Pectoral extending .21–.24 of the distance from its axil to the origin of the anal, anal originating below second-fourth ray of dorsal, extremely fleshy, branched rays of anal bifid. Maximum total length observed 6.8 in.

The present description is based on the dissection of five specimens and the external examination of a number of others.

Gunther's type is recorded as being in the British Museum. Type locality: Westland. Localities! (South Island) Greymouth. (North Island) Palmerston North, Rongareu, Wairarapa. Habits: This fish is normally free-living, but during droughts it is frequently found alive in the mud of streams and swamps from which the water has been absent for months. Male specimens taken

Picture icon

Fig. 1.—Galaxias attenuatus.
Fig. 2.—Galaxias paucispondylus.

Picture icon

Fig. 3.—Galaxias proguathus.
Fig. 4.—Galaxias burrowsius.

Picture icon

Fig. 5.—Galaxias lynx.
Fig. 6.—Galaxias koaro.

Picture icon

Fig. 7.—Galaxias campbelli.
Fig. 8.—Galaxias vulgaris.

Picture icon

Fig. 9.—Galaxias fascialus
Fig. 10.—Galaxias alepidotus.

Picture icon

Fig, 11.—Neochanna apoda.
Fig. 12.—Neochanna diversus.

– 495 –

in October had the milt almost fully developed and appeared to be within a week or two of spawning.

Neochanna diversus n.sp.

In 1945 I noted the existence of a second species of Neochanna, but did not feel justified in naming it on the single specimen available. Since then I have observed a specimen of the same species in the reference collection of the Dominion Museum and have received six others per Canterbury Museum from Mr. C. W. Devonshire, Kaitaia, North Auckland. The present description is based on the examination of these eight specimens, six of which have been dissected.

Neochanna diversus n.sp.

B. 6–7; D. iii-iv 7–10; A. iii-vi 9–11; vertebrae 56–58.

Lower jaw somewhat the longer, profile of snout strongly convex, teeth in both jaws conical as in Galaxias, uniserial, without canines, lingual teeth conical, hooked, in two rows, entopterygoids toothless. Gill-rakers developed as in typical Galaxias, pyloric caeca 1–2, varying from well developed to extremely rudimentary. Maxillary extending about to anterior of eye. Head 5.4–5.9 in standard length, dorsal fin inserted at .70–.75 of the standard length, its base enclosed in a fleshy sheath, exposed part of fin extremely fragile. Tail with strongly developed ridges continuous with dorsal, caudal and anal fins, its least depth including ridges .70–.83 of the distance from rear of dorsal fin to hypural joint, caudal fin convex, asymmetrical, its upper half having the more acute profile. Pectoral fin extending .17–.23 of the distance from its axil to the origin of anal, anal origin about opposite that of dorsal, may be slightly anterior or slightly posterior, anal base enclosed in a fleshy sheath similar to that of dorsal and continuous with caudal ridge, exposed part of fin extremely fragile, branched rays of anal bifid.

In colour, uniform blackish grey on back and sides, somewhat lighter below. Maximum total length observed 4.7 in.

Differs from apoda in the conical teeth in the jaws, the lower number of rays and the higher proportion of branched rays in the dorsal and anal fins, the shorter mouth and greater convexity of the profile of the snout.

The species is named on account of its disagreement with several of the original generic characters.

Type: Holotype in Canterbury Museum.

Type locality: Holotype from Kaitaia, North Auckland.

Localities: In addition to the type locality the species has been obtained from Waihopo and Mangawai, North Auckland.

Habits: Mr. R. N. Hastie, of Mangawai, obtained the Mangawai specimen from the mud of a creek during summer, but Mr. C. W. Devonshire, who collected the Kaitaia material, states that three specimens taken on July 30 were free-swimming. Two of these specimens were dissected and proved to be fully developed males, one being just ripe for spawning, while in the other about half the milt had been discharged. The spawning season of diversus would thus appear to be about three months earlier than that of apoda.

– 496 –

Acknowledgments

I wish to express my thanks to Dr. E. Trewavas, of the British Museum, for the examination of types in that institution; to Dr. Leon Bertin, of the Museum National d'Histoire Naturelle, Paris, for a co-type of campbelli; to Mr. A. W. B. Powell, of the Auckland Museum, for the loan of a collection from the thermal lakes; to Dr. R. A. Falla, Director of the Canterbury Museum, for access to the collection in that institution; to Mr. A. C. O'Connor, of Wellington, for a collection from Wairarapa; to Mr. D. Cairns, of the Department of Scientific and Industrial Research (per Canterbury Museum), for specimens of lynx from Lakes Wanaka and Hawea; to Mr. J. K. Shaffrey, of Inchbonnie, for specimens of lynx from that locality; to Mr. A. Kean, Conservator of Fish and Game, for specimens of lynx from Lake Waikaremoana; to Mr. Heenan, Under-Secretary of the Department of Internal Affairs, for specimens of lynx from Lake Waikaremoana; to Messrs. Judson and Leech for assistance in collecting specimens of burrowsius at Rangiora; to Mr. W. J. Phillipps, Acting Director of the Dominion Museum, for access to the collection in that institution; and to Mr. C. W. Devonshire (per Canterbury Museum) for specimens of Neochana diversus from Kaitaia, North Auckland.

References.

Bloch and Schneider, 1801. Systema Ichthyologiae.

Glarke, F. C., 1899. Notes on New Zealand Galaxiidae, more especially those of the Western Slopes; with description of New Species, etc. Trans. N.Z. Inst., 31, pp. 78–91.

Eigenmann, C. H., 1928. The Freshwater Fishes of Chile. Nat. Acad. Sci., 22, 2nd mem., pp. 78–51.

Foster, J. R., 1777. Voy. World Resolution, 1, p. 159.

Gray, J. E., 1842. Zool. Misc., p. 73.

Gunther, A., 1866. Cat. Fish. British Museum, vol. 6.

Haast, J. von, 1872. Notes on some Undescribed Fishes of New Zealand. Trans. N.Z. Inst., vol. 5, p. 270.

Hutton, F. W., 1872. Contributions to the Ichthyology of New Zealand. Trans. N.Z. Inst., vol. 5, p. 270.

—— 1896. Notes on some New Zealand Fishes with Description of a New Species. Trans. N.Z. Inst., vol. 28, p. 317.

—— 1901. On a Marine Galaxias from the Auckland Islands. Trans. N.Z. Inst., vol. 34, p. 198.

Jenyns, L., 1842. Zoology, Voyage of the Beagle. Fish, pp. 119–121.

McKenzie, D. H., 1904. N.Z. Illustrated Magazine, vol. 10, pp. 122–126.

Phillipps, W. J., 1926. New or Rare Fishes of New Zealand. Trans. N.Z. Inst., vol. 50, p. 531, pl. 88.

—— 1940. The Fishes of New Zealand, vol. 1.

Regan, C. T., 1905. A Revision of the Fishes of the-Family Galaxiidae. Pro. Zool. Soc. London, vol. 2, pp. 363–384.

Richardson, J., 1848. Zoology of the Voyage of the Erebus and Terror.

Sauvage, —, 1880. Bull. Soc. Philomatique, ser. 7, vol. 4, p. 229.

Scott, E. O. G., 1935. Observations on Fishes of the Family Galaxiidae, Part 1. Proc. Roy. Soc. Tasmania, pp. 85–112.

Stokell, G., 1940. A New Species of Galaxias. Trans. Roy. Soc. N. Z., vol. 69, pp. 422–424.

—— 1945. The Systematic Arrangement of the New Zealand Galaxiidae, Part. 1. Trans. Roy. Soc. N. Z., vol. 75, pp. 124–137.

Waite, E. R., 1909. The Sub-Antarctic Islands of New Zealand, vol. 2, pp. 585–588.

Whdtley, G. P., and Phillipps, W. J., 1939. Descriptive Notes on Some New Zealand Fishes. Trans. Roy. Soc. N. Z., vol. 69, pp. 229–231.