Other formats

    TEI XML file   ePub eBook file  


    mail icontwitter iconBlogspot iconrss icon

Tuatara: Volume 17, Issue 1, May 1969

Observations on Growth and Behaviour of Galaxiidae in Aquariums

Observations on Growth and Behaviour of Galaxiidae in Aquariums


Although Recent Authors (e.g., Stokell, McDowall, and Woods) have published much work on the taxonomy and distribution of the Galaxiidae in New Zealand, there is, with the exception of Galaxias maculatus (Hefford, 1931a, 1931b, 1932, 1934a, 1934b; McDowall, 1968), little information on the behaviour and life histories of these fishes.

The Galaxiidae generally are secretive fishes, and their ability to exist in large numbers undetected by all but the most assiduous observers helps to explain why so little is known about them. This fact, together with recent warnings of some possible future extinctions (Stokell, 1955; McDowall, 1966a; Skrzynski, 1968), makes it very desirable that their life histories and behaviour should be fully understood in order that any necessary measures can be taken to preserve them.

This paper presents data on growth and behaviour of galaxiid fishes kept in aquariums, which are compared where possible with what has been found to occur in the natural environment. These data are not the result of any particular programme of investigation page 35 but rather the ‘by-products’ of other activities and are therefore far from complete; however, they provide information on artificial populations which may have relevance to studies in the natural environment.


As the chief aim of almost all the attempts to keep Galaxiidae was to observe the behaviour of the fish, conventional aquariums with angle-iron frames and glass on all sides were normally used. They ranged in size from 24in. × 12in. × 12in. to 72in. × 15in. × 18in. The back glass was usually blacked out. Aluminium covers incorporating electric lights were used to prevent the fish from jumping out of the tanks. The lights were used only to inspect the fish at night or to promote plant growth when this was required.

Water used in the tanks was usually uncontaminated stream water. Sometimes tap water from the Wellington City mains was used, but this was left to stand for several days before being stocked with fish. Once a tank was established the water was rarely changed.

If the fish became infected with a disease or parasite which proved difficult to eradicate by the various methods described in most aquarium handbooks, an extremely effective technique was to clean out the tank, fill it with 10 per cent formalin, and leave it for two days with the contained air lines, filters, etc. The formalin was then drained off, and the tank was refilled with fresh water and drained several times over a period of a week; the tank was then ready for re-use.

Depending on the species of fish kept, air pumps were used to operate filters, for aeration, and for water agitation for fish which normally inhabit fast waters.

Aquariums were maintained at room temperatures, but if sickness in a fish was thought to be due to too high temperatures, the fish was removed to a cooler situation.

The tank bottoms were covered with river gravel and with rocks so arranged that fish had the illusion of cover while being visible to the observer. This was accomplished by laying flattish rocks one on top of the other to form small caves running from front to back. Some tanks were well planted with water weeds, which gave additional cover, but this was not essential, as the galaxiids usually inhabit waters devoid of higher plant life.

Fish were measured (length to caudal fork) after being narcotised in a mild solution of MS 222.

The fish were fed with a wide variety of live foods augmented with some patent dried fish foods, minced liver, and fish. Whiteworms (Enchytraeus sp.), Daphnia, ostracods, and copepods were given to whitebait and small fish; earthworms, woodlice, crushed snails, grubs, caterpillars, flies, and moths — whatever was most easily available at the time — were given to the larger fish. The page 36 fish were given as much food at any one time as they would take eagerly, but frequency of feeding was not regular, being once, twice, or three times a week.

For a constant supply of live food, whiteworm cultures were most satisfactory, and these were fed on slightly moistened waste bread.

Whitebait — Adaptation and Behaviour

Young fish usually adapt to aquariums better than older fish, and whitebait — which Woods (1963) described as ‘the transparent, free-swimming and shoaling juveniles of at least five species of galaxias’— are abundant at a time when they are probably most adaptable to aquarium life.

The transparent juveniles of five species having marine whitebait (McDowall, 1964, 1965, 1966b, and in press) are the subjects of this section. They enter estuaries in mixed shoals on the flood tides during spring and can be obtained all round the New Zealand coasts. Although species composition varies from place to place (McDowall, 1965), and all except G. maculatus appear to be rare on the east coasts of both islands, the most likely places to obtain the different species can be predicted (McDowall, 1965; Woods, 1966). Galaxias brevipinnis tends toward snow-fed rivers, but G. fasciatus, G. postvectis, and G. argenteus prefer warmer bush-and swamp-fed waters. Galaxias maculatus will be found in all types of rivers except those running off steep hill faces directly into the sea.

Whitebait were safely transported in water in insulated (foam plastic) bins or in water sealed in plastic bags protected with sawdust or some other suitable material in a box or tin.

Released into a tank, whitebait will begin to feed in a few hours, especially if the tank has been ‘seeded’ with Daphnia in advance.

Successes in rearing batches of whitebait were variable, and the reasons for this are not understood. With G. maculatus, rearing was almost always 100 per cent successful, but with the other species there was sometimes failure. Galaxias fasciatus usually settled in well, provided that the depth of water was kept down to 10in. or less, but these fish were extremely slow growing in a tank, and in a mixed batch of fish were at an early disadvantage, being the smallest of the whitebait. They sometimes sickened and died off in numbers, even when kept separate from other species.

Although two separate small lots of G. postvectis all lived, less than 50 per cent of all whitebait introduced into tanks and known to have been this species survived. Mortalities were unexplained, deaths occurring at temperatures between 15.6 and 19.2°C, though more than 50 per cent of a smaller number of G. brevipinnis under the same conditions survived.

page 37

The fourth species of whitebait. G. argenteus, has not been handled in sufficient numbers to make comparisons, but of three whitebait known to have been G. argenteus, two were reared to the late immature stage and the third died under the same circumstances as the G. postvectis referred to above.

Galaxias brevipinnis adapted surprisingly easily to aquarium life, considering the fairly rapid waters which it normally inhabits. Woods (1966) stated this species died at temperatures between 17 and 20°C, but the author found it capable of withstanding 19°C for periods of a day or two, and up to 23°C once the fish had been established for a few weeks. At temperatures below 17°C this species usually attains the late immature stage without loss.

Post-Whitebait — Growth, Behaviour, Longevity

In nature, with the exception of G. maculatus, the whitebait shoals break up at some stage after entering fresh water, and the different species occupy their own distinctive habitats.

Galaxias maculatus remains in the lower, easily accessible reaches of the rivers or may find its way into coastal lagoons and swamps. Extensive rapids, low falls, and modern bridge culverts with a free fall at the outlet bar its upstream progress. Galaxias fasciatus and G. brevipinnis often inhabit streams with a flow of less than about 10 cusecs — the former in pools and the latter in the fast ripples — and both are capable of surmounting large falls while they are migrating upstream. Galaxias fasciatus is also found in swamps. Galaxias postvectis has been caught mostly in the pools of bush streams, but its adult habitat is not really known. Galaxias argenteus is found mostly in slow streams and deep swamps, but it has also been taken from some lakes.

Galaxias maculatus (Inanga)

This fish is probably the best known of all the native freshwater fishes, and this is due in part to the ease with which it can be reared in almost any kind of container. To the student it is important, because its growth to first maturity in an aquarium can be comparable with that of fish in the natural environment; it is therefore useful as a guide to the suitability of conditions in the aquarium.

Inanga have a naturally short life cycle (Burnet, 1965; McDowall, 1968). In a tank many of them died after about six months, when they were 1 + years old and fully mature. They then measured from about 55 mm to about 80 mm. However, death did not always occur at this time, for some fish were found to survive for a further twelve months, maturity either being delayed or gonad resorption taking place, a facet worth further study. These fish attained lengths of 110 mm or more. These age/lengths for 1 + and 2 + years fish are comparable with data on fish from the natural environment (Burnet, 1965; McDowall, 1968). The occasional fish which page 38 survived to an age of 3 + years in tanks did not attain the size of that age group in the natural environment 140 mm or more (Burnet, 1965).

Occasionally G. maculatus females in a tank were found by the author to be spent, but this was a rare occurrence and the spawn has not been seen. Presumably it was eaten by the occupants of the tank.

In a large tank (72in. × 15in. × 18in.) inanga formed a shoal, which included other species, and remained in mid-water except in temperatures below about 10°C, when they tended to become less active and sometimes lay on the bottom. Very large individuals, too, tended to leave a shoal and remain inactive under cover indefinitely. Temperatures as high as 24.5°C did not appear to inhibit normal activity.

The behaviour of G. maculatus toward all other species of fish was pacific unless the other fish was small enough to be eaten. Inanga fed readily on most animals small enough for them to eat, either dead or alive, and would also take patent dried fish foods. They fed from the surface, from mid-water, or from the bottom with impartiality, as has been found with wild fish (McDowall, 1968).

Galaxias maculatus was sometimes introduced to an aquarium as an adult, but it was very often infected with whitespot (Ichthyophthirius), and though this parasite may be dealt with in a number of ways, it was a great nuisance and highly infectious.

Galaxias fasciatus (Banded Kokopu)

The banded kokopu were much slower growing in an aquarium than any of the other whitebait species. Captured as whitebait at about 45 mm in length, they attained only 60-65 mm in a further twelve months. The only two specimens which survived to twenty-six months reached lengths of 103 mm and 104 mm, but one fish, which was 72 mm long when caught in April, increased by only 7 mm in the subsequent sixteen months. This was slower growth than the species attained in a stream near Wellington, where 1 + years fish ranged from about 70-100 mm, and 2 + years fish about 105-125 mm (Fisheries Research Division, unpublished data).

In the wild a few male G. fasciatus at lengths of only 62 and 63 mm were found to be ripe during April after what is assumed to have been only about six months in fresh water. Females have been known to shed spawn in a tank after eighteen months in captivity, having been well past the whitebait stage when captured (it is not known if the eggs were fertile, as they were eaten by adult fish before development could occur (R. Perrett, pers. comm.)). It seems this species does not normally mature until its second year, and that it has a life span of at least three, and possibly four or more, years, as specimens of 150 mm are not uncommon and some exceed 200 mm. As there are known lacustrine populations of this species (McDowall, 1966b.) tank breeding may be feasible.

page 39

In the pools of small streams banded kokopu tend to form shoals, and over a dozen fish may be seen which may span a large size range. In tanks these fish were generally solitary, though the species was not kept in numbers on its own and may well have shoaled under these circumstances. In a large, deep tank (72in. × 15in. × 18in.), with a mixed population of fish, a single small G. fasciatus remained near the surface of the water and was easily intimidated by other species. In smaller tanks G. fasciatus was better conditioned and lived peacefully with G. maculatus or G. postvectis and with Retropinna spp. and Gobiomorphus spp. of similar size range.

Under aquarium conditions the banded kokopu exhibited the same feeding habits as the inanga, but was more adept at catching large insects on or just above the surface of the water. It seized big live moths by the abdomen, dragged them down, and broke off the head and thorax of each by dashing the insect against a rock. It then slowly swallowed the abdomen, part of which protruded from the fish's mouth for some time afterward.

The author has had little success with the adaptation of adult Galaxias fasciatus to tank life. The fish have invariably broken out with a fungus (Saprolegnia) a few days after capture, even when no injury was apparent, though this has not been the experience of G. C. Kelly (pers. comm.), who has kept a number of captured adult G. fasciatus in covered bath tubs. However, rearing from the whitebait may offer the most satisfactory means of acquiring adult fish, and this would probably be most successful if the G. fasciatus could be separated from the other whitebait at an early date.

Galaxias postvectis (Short-jawed Kokopu)

The least known of the whitebait species, this kokopu has been taken by the author in large numbers as whitebait from the Buller River, and there may be substantial populations of it in the deeper pools of this river and its tributaries.

Once past the whitebait stage, this species appeared well suited to aquarium life. In tanks longer than 3ft. 6in., it attained lengths of 105 to 121 mm after twelve months' captivity, and one fish reached 135 mm after twenty-four months. No data are available on growth in the natural environment, but the observed growth was good, compared with that of the other kokopu.

After the early deaths of some fish of this species (at the whitebait stage already referred to) the survivors thrived in temperatures which did not drop below 21°C for several weeks during summer and which exceeded 23°C at times. As whitebait and early juveniles they were mid-water swimmers, but after about two months they took up solitary positions under cover. At this stage there was some skirmishing over choice of refuge, but this soon ceased, and though the fish met when being fed, there was little aggressiveness between them, and even less towards other species.

page 40

Later, however, when the fish had been in fresh water for about twenty-one months, prolonged battles occurred near the surface of the water, with the combatants circling and biting at each other. It is not known whether these fights were due to the increasing size of the fish and their need for more territory or if they were connected with approaching sexual maturity, but one fish which died of disease at twenty-two months was found to be a ripe male. It was 126 mm long — only about half the size of some specimens which have been taken from the wild.

The short-jawed kokopu is by nature a bottom feeder, but it readily took food which was still falling through the water; it also learned to go to the surface for food if it saw other species doing so, though the under-shot lower jaw caused it to miss often. The fish gradually lost the habit of rising to the surface as they grew bigger. They were never observed to eat other fish or try to do so, even when small bullies (Gobiomorphus spp.) and whitebait were placed in the tank with them.

No attempt has been made by the author to adapt adult Galaxias postvectis to tank life, but G. C. Kelly (pers. comm.) has a specimen (still alive at the time of writing) which was an adult when captured four years ago.

Galaxias argenteus (Giant Kokopu)

It was probably G. argenteus which Graham (1956) kept for a few days in a salt water tank. Although he described it as G. fasciatus, the measurements given — 15in and 2½lb — are far greater than any other measurements recorded for this species.

Galaxias argenteus is the largest growing of the New Zealand Galaxiidae, but it may be mistaken for the banded kokopu during the early immature juvenile stage. Both McDowall (1966b) and Woods (1968) stated that the markings of juvenile G. argenteus are quite distinct from those of other species, but neither of them had examined very many juvenile G. argenteus, and both the young and adults of G. fasciatus may vary (see Woods, 1968, fig. 7). If there was any doubt about the identity of small live striped kokopu, their behaviour in a tank quickly indicated the presence of G. argenteus, for they were very pugnacious and harrassed G. fasciatus almost constantly.

There are very few records available of small G. argenteus from the wild — possibly because they have been mistaken for the banded kokopu — and there is no information on growth rate. Only a few specimens have been kept in tanks, and one reached 95 mm after fourteen months in fresh water. This appears to indicate a slow growth rate when it is considered that many fish of this species grow to 300 mm and that much larger specimens have been recorded. G. C. Kelly (pers. comm.) kept one of these fish, which was already at least 150 mm long when first caught, in a bath tub for over four page 41 years. Thus it would appear that the potential life span of G. argenteus is great and probably exceeds six years.

A large G. argenteus is a hungry feeder once it has settled down, but adults may take several weeks to do this, and then only come out from cover to feed at night. The steady reduction in numbers of small fish in the same tank as G. argenteus indicated that it was to some extent piscivorous, though its apparent slow movements seemed ill suited to the catching of agile species like G. maculatus and Retropinna retropinna, which were among the missing fish. It has been supposed that G. argenteus normally takes these fish from ambush, and G. C. Kelly (pers. comm.) states that it is capable of very rapid acceleration and deceleration. It is worth noting that in the only locality known to the author where the giant kokopu could be described as abundant — a flax swamp on the West Coast — other species, with the exception of a few eleotrids, are absent.

An apparent shunning of light, exhibited by the adult giant kokopu during the first weeks of captivity, is not always a feature of its behaviour in the natural environment. Although it may often be seen at night by torch light in places where it is not to be seen during the day, it may sometimes be at the surface of quiet, deep swamp pools during the day. At such times it readily takes a baited hook and can be caught far more easily in this way than by any other method.

Like the banded kokopu, G. argenteus would take its food from any zone of the water, and worms, crushed snails, raw meat, and moths were all acceptable.

Adult giant kokopu adapted to tank life fairly well and did not seem to be as prone to disease as G. fasciatus.

Galaxias brevipinnis (Taiwharu)

This fish is an inhabitant of ripples and cascades, where it dwells under stones, and the fact that it settled fairly well to aquarium life was a paradox which it shared with the torrent fish (Cheimarrichthys fosteri).

Captive taiwharu made good growth for the first year, and lengths in excess of 90 mm have been recorded in this time, with one fish reaching 106 mm after fourteen months in captivity. These measurements are comparable with data on a small population in a Wellington stream where the same year class ranged from about 87 to 103 mm and from 78 to 102 mm in November of two successive years (Fisheries Research Division, unpublished data).

As opposed to the satisfactory growth rate for fish captured as juveniles, a fish which was 80 mm long when first captured in September 1965 (1 + years) had increased by only about 15 mm during the subsequent three years—a fact which highlights the desirability of obtaining fish while they are juveniles.

A specimen of G. brevipinnis has been kept in captivity for six years by G. C. Kelly (pers. comm.), and three marked fish with a page 42 range of 156-190 mm recaptured in a Wellington stream represented 50 per cent of fish fin-clipped exactly two years before, which all ranged from 133 to 190 mm when first handled (Fisheries Research Division, unpublished data). The above data seem to indicate that the taiwharu normally has a life span in fresh water of at least four years and possibly much more.

In behaviour the captive taiwharu was an inquisitive fish, constantly active, as was noted by Woods (1963), and after capture as whitebait it shoaled freely with such other fish as G. maculatus and R. retropinna for many months — behaviour which has certainly not been observed in the natural environment, where, as an adult, it would be unlikely to overlap the habitats of these fish. If a much larger fish of another species was placed in the same tank with it, G. brevipinnis reverted to an undercover existence and was not seen for days at a time

Galaxias brevipinnis is naturally a bottom feeder, but like G. postvectis it learned to take food from the surface of the water and became more adept at it than did the latter species. Owing to its small mouth it needed to be supplied with reasonably small animals as food. These were supplemented with dried food, which the fish learned to take.

Other Galaxias

Galaxias divergens (Dwarf Galaxias)

Of the Galaxiidae dealt with in this paper the dwarf galaxias at first proved the most difficult to keep alive and healthy. Several early attempts to keep this fish were entirely unsuccessful, the longest survival being three months.

Current work, however, indicates that this species may have only a two-year life cycle (C. L. Hopkins, pers. comm.), and this may account for some of the early adult mortalities, though not for similar losses of fry.

Recently, success has been gained by keeping these fish in a cool room, in shallow water not more than 2 or 3in. deep and well agitated by aerators. Gradual increase in the depth of water after a time apparently caused the fish to leave their cover and swim about in mid-water, where they were unbalanced and distressed.

Four specimens have now been held in tanks for ten months, three being young adults when first caught and the fourth a fry. At the beginning of August spawn was laid by one of the three adults, but the eggs failed to develop and it is thought that no male fish is present (C. L. Hopkins pers. comm.).

Free-swimming fry of G. divergens captured in the field may be reared on very fine dried foods, but once the fish have taken to the bottom, live foods are necessary, and white worms have proved satisfactory. Currently, free-swimming fry are being fed on newly hatched brine shrimps.

page 43

Galaxias vulgaris (Common river Galaxias)

This species does not occur in the North Island, and the author has had only six live specimens. One died shortly after arrival in Wellington, and a second was eaten by a G. argenteus during an experiment, but the other four survived for twenty months before dying of a severe infection of whitespot.

The average increase in length of the four latter fish was 14 mm over the twenty months (means of 72 and 86 mm at capture and at death respectively), but it is not known how this would compare with the growth of fish in the natural environment.

The behaviour of this species in tanks varied greatly according to circumstances, but it was generally more active than was expected for a fish which normally dwells under cover. In a large tank with 16in. of water depth and with a variety of shoaling and bottom-dwelling species G. vulgaris would spend most of the time actively swimming in mid-water with the shoaling fish. This was very surprising, since Benzie (1961) found that adult G. vulgaris exhibited no intraspecific social behaviour.

In smaller tanks with less than 12in. of water G. vulgaris was found to be aggressive toward such mid-water species as smelt (Retropinna retropinna), but shy of bottom dwellers like Gobiomorphus huttoni. It spent much time swimming around just off the bottom, with its nose down on the gravel. If the tank contained no other species, it would rest on the bottom, either under or near to cover, but it became very active when food was offered.

Although mainly bottom feeders, captive Galaxias vulgaris would take food from the surface, as this species has been observed to do in the quieter waters of its natural habitat (Benzie, 1961), Both live insects and dried foods were taken readily by captive fish.


Neochanna apoda (Brown Mudfish)

Davidson (1951) gave an account of the keeping of the brown mudfish in aquariums, in which she met with varied success. Contrary to what is usual with most other fish, the adults seem to adapt to tank life far better than fry and juveniles. Young fish captured by the author have usually died within two months, having grown an average of 6 mm in the interim, and no visible cause of death has been noted.

Growth of adult fish appeared to be satisfactory, but it has not been possible so far to compare it with that of wild stock, which may vary considerably in any case according to whether or not drought conditions force the fish to aestivate.

Tank breeding of these fish is possible, for the author currently has three juveniles which were first observed as fry on July 27, page 44 1968, three days after six adult fish had been removed. These adults had been the sole occupants of the aquarium since their capture on March 9, and no material had been added to the tank subsequently. Weed growth in the tank had been allowed to become very dense, and despite frequent inspections of the tank, no spawning activity had been recognised, and indeed none of the fish had appeared to be gravid. Unusual behaviour had been noted in mid June, when all six fish had taken to resting on the weed growth very near the surface of the water and remaining perfectly still even when the tank cover was removed. How long the juveniles survive, in view of previous experiences with young fish noted above, remains to be seen, and observations are continuing.

Neochanna apoda has been observed to spent the daylight hours resting on the bottom of the tank, under weeds, stones, or whatever cover was available, with several fish sharing the same cover and lying packed close together in physical contact. At night they were seen to emerge and move slowly around the tank or lie motionless among the weeds close to the surface of the water as described above. If the tank was aerated, they would sometimes swim actively up and down the stream of bubbles emitted by the air line.

Although up to seven mudfish have been kept in one 30in. × 12in × 12in. tank, little aggressiveness has been noted among them, but M. M. Davidson (pers. comm.) reported an instance of cannibalism among fish which were being transported after capture.

On several occasions when a single N. apoda was placed with other species it became extremely aggressive (Eldon, 1968) and entered into battle with species of Galaxiidae, Retropinnidae, Eleotridae, and small eels, with which it was confronted. However, possibly because of its rather small mouth, it seldom inflicted any real damage. Paradoxically, when a single eleotrid was placed in a tank containing several mudfish, to enable a study to be made of the reaction of both species, it was completely ignored by the mudfish.

Whiteworms were a convenient food for mudfish and were taken readily, but the fish would also take earthworms and some insects provided that they were not too large.

Neochanna diversus (Black Mudfish)

The author has no personal experience of this species, but W. Skrzynski (pers. comm.) kept four specimens in a tank for three and a half years, and they grew from an estimated 50 mm at capture to an average length of 90 mm in that time.

These fish did not breed, though both sexes were found to have been present when the fish were examined after their deaths; death may have been premature, as all four fish died shortly after they had to be moved to another tank.

There was no aggressiveness among the black mudfish, which were kept in 2in. of water in a 24in. × 12in. tank.

page 45

Although supplementary food in the form of whiteworms was given to the N. diversus from time to time, the tank was largely selfsupporting, the fish apparently feeding on a population of snails (Potamopyrgus sp.) which maintained itself on the plant and algae growth.


Many species of Galaxias will live for a considerable time in a suitable aquarium and there is no reason why spawning behaviour of some of the species should not be eventually observed in captivity, especially if some form of water circulation can be maintained. In the event of successful breeding techniques being evolved it may be possible to save seriously threatened species, such as G. burrowsius, from extinction (Skrzynski, 1968).

If tanks are to be used to study the growth and behaviour of a single species of Galaxiidae, additional tanks for study of interspecific behaviour are an advantage. The main fault with the author's arrangements has been too much mixing of species, partly to observe reactions and partly because of shortage of space. This situation frequently causes trouble when one species dominates the tank or falls sick and affects the others. The only exception to this rule of segregation would be the keeping of one or two G. maculatus with the study fish to act as controls.

Except for G. argenteus and possibly G. postvectis only shallow tanks of 12in. or less are required, but the surface area should be as large as practicable.

Dried foods and raw meat may be used to supplement the fish feed, but live foods should always form the major part of the diet.


Details are given of the behaviour and growth of nine species of Galaxiidae kept in captivity, and these are compared with data collected in the field.


I thank those people whose personal communications helped to complete this paper, particularly Mr. G. C. Kelly of the Soil Bureau, D.S.I.R. The help and interest of staff of the Fisheries Research Division are gratefully acknowledged.

Literature Cited

Benzie, V., 1961. A comparison of the life history and variation in two species of Galaxias, G. attenuatus and G. vulgaris. (M.Sc. thesis lodged in University of Canterbury library. Christchurch, New Zealand.)

Burnet, A. M. R., 1965. Observations on the spawning migrations of Galaxias attenuatus (Jenyns). N.Z. Jl Sci. 8: 79-87.

Davidson, M. M., 1951. The New Zealand mudfish (Neochanna apoda). The N.Z. Aquarium Bulletin 12: 33.

Eldon, G. A., 1968. Notes on the presence of the brown mudfish (Neochanna apoda Günther) on the West Coast of the South Island of New Zealand. N.Z, Jl mar. Freshwat. Res. 2: 37-48.

Graham, D. H., 1956. ‘A Treasury of New Zealand Fishes.’ 2nd ed. Reed, Wellington. 424 pp.

Hefford, A. E., 1931a. Whitebait. Rep. Fish. N.Z. 1930: 15-6, 21-3.

——, 1931b. Whitebait. Rep. Fish. N.Z. 1931: 14-7.

——, 1932. Whitebait investigations. Rep. Fish. N.Z. 1932: 14-5.

——, 1934a. Whitebait. Rep. Fish. N.Z. 1933: 13-4.

——, 1934b. Whitebait. Rep. Fish. N.Z. 1934: 14-5.

McDowall. R. M., 1964. A consideration of the question ‘what are whitebait?’ Tuatara 12: 134-46.

——, 1965. The composition of the New Zealand whitebait catch, 1964. N.Z. Jl Sci. 8: 285-300.

——, 1966a. A guide to the identification of New Zealand freshwater fishes. Tuatara 14: 89-104.

——, 1966b. Further observations on Galaxias whitebait and their relation to the distribution of the Galaxiidae. Tuatara 14: 12-8.

——, 1968. Galaxias maculatus (Jenyns), the New Zealand whitebait. Fish. Bull. N.Z. (n.s.) 2. 84 pp.

——, (in press). The galaxiid fishes of New Zealand. Bull. Mus. comp. Zool. Harv.

Skrzynski. W., 1968. The Canterbury mudfish, Galaxias burrowsius Phillipps, a vanishing species. N.Z. Jl. mar. Freshwat. Res. 2: 688-97.

Stokell, G., 1955. ‘Fresh Water Fishes of New Zealand’. Simpson and Williams, Christchurch. 145 pp.

Woods, C. S., 1963. ‘Native and Introduced Freshwater Fishes.’ Reed, Wellington. 64 pp.

——, 1966. Species composition of whitebait (Galaxias spp.) Rec. Canterbury Mus. 8: 177-9.

——, 1968. Growth characteristics, pigmentation, and the identification of whitebait (Galaxias spp., Salmonoidea). N.Z. Jl mar. Freshwat. Res. 2: 162-82.