Tuatara: Volume 12, Issue 2, July 1964
The Affinities and Derivation of the New Zealand Fresh-Water Fish Fauna
The Affinities and Derivation of the New Zealand Fresh-Water Fish Fauna
Like the New Zealand terrestrial vertebrate fauna as a whole, the fresh-water fish fauna of New Zealand is not large, as only six or seven families of fresh-water fish are native to New Zealand.
About 35 New Zealand species of fresh-water fish are at present recognised. These comprise 17 species of the family Galaxiidae in two genera (Neochanna, two species and Galaxias, 15 species); six species in the family Eleotridae in two genera (Gobiomorphus, four species and Philypnodon, two species); two species of Anguilla in the family Anguillidae; one species (which may be extinct) of Prototroctes (family Aplochitonidae); one species of Cheimarrichthys (family Cheimarrichthyidae); one lamprey Geotria, family Geotridae); one flatfish (Rhombosolea, family Pleuronectidae), which enters and lives freely in the sea and is a doubtful member of the fresh-water fauna.
Of these fishes all but one species of Galaxias (G. attenuatus), one eel (Anguilla australis) and the lamprey (Geotria australis) are endemic, although all the genera except Neochanna and Cheimarrichthys are shared with other regions. The family Cheimarrichthyidae, comprising only one species, is confined to New Zealand. About 92% of the species but only 11% of the genera in the fauna are thus endemic to New Zealand, indicating a relatively young fauna and/or incomplete faunal isolation from other regions. The high endemism at species level indicates that the fauna has been isolated in recent times but the generic similarities suggest that the fauna is not of great age.
The Nature of the Fauna
The New Zealand fresh-water fish fauna contains two elements:— 1. Temporary inhabitants of fresh-water, either anadromous or catadromous; 2. Permanent inhabitants of fresh-water. Most of the taxonomic groups in the fresh-water fauna have representatives in the first of these groups. Galaxias attenuatus breeds in estuarine conditions after maturing in the rivers, and the larval life and much of the juvenile life is spent in the sea. Prototroctes oxyrhynchus is thought by some authors to have behaved similarly (e.g. Arthur, 1884, p. 172). Gobiomorphus huttoni breeds in fresh-water but larval life is thought to be marine and this may apply to other New Zealand Eleotridae. The breeding place of Cheimarrichthys forsteri is undescribed and larval C. forsteri have not been found, although large adults full of eggs have been taken in autumn. Both species of Anguilla have marine breeding and larval life. Anadromous, partially fresh-water dwelling species include those species of Retropinna which are not entirely fresh-water dwelling (R. retropinna, R. osmeroides, R. anisodon) and which enter estuaries and lowland streams from the sea to breed; Geotria australis behaves similarly and Gobiomorphus basalis may be anadromous but is more usually resident in fresh-water. The permanent inhabitants of fresh-water show varying migratory patterns, where these are known.
The Broad Relationships of the Fauna
As outlined above, 92% of the species but only 11% of the genera of New Zealand's fresh-water fishes are endemic to New Zealand. Generic relationships are widespread, but mostly confined to the Southern Hemisphere.
Galaxias, using this name in a broad sense to include Paragalaxias and Saxilaga, etc., is distributed as follows:— Australia (about 27spp.), New Caledonia (1sp.), New Zealand (16spp. with one species on the Campbell and Auckland Islands), South America and the Falkland Islands (10spp.) and South Africa (2spp.). At the species level, each area comprises only endemic species except for the presence of Galaxias attenuatus in south-east Australia, New Zealand and South America (including the Falkland Islands).
The genus Retropinna is found only in the Australasian region, the six New Zealand and three Australian species being endemic to each area. Anguilla is a cosmopolitan genus with one species endemic to New Zealand and the other New Zealand species also present on New Caledonia and the south-east of Australia. The family Aplochitonidae is represented by one species of Lovettia from Tasmania, one species of Prototroctes from south-east Australia and Tasmania and another species from New Zealand, and the genus Aplochiton from South America. Geotria is a page 61 member of the family Geotridae, which comprises a southern group of lampreys. Geotria australis occurs in south-west and south-east Australia, Tasmania, New Zealand and Chile. Cheimarrichthys belongs to an endemic, monotypic family of uncertain relationships. Affinity has been suggested with the Indo-Pacific Parapercidae, but this relationship has not been studied and remains in doubt.
The family Eleotridae is a large Indo-Pacific family of which the New Zealand species comprise the southernmost extension. The two genera present in New Zealand (Gobiomorphus, four species, Philypnodon, two species) are also present in south-east Australia where there is one species in each genus. The species in each area are endemic.
On the basis of existing distribution, it is apparent that the affinities of the New Zealand fresh-water fish fauna are southern, mostly with Australia but also with South America. Affinities with the South African fauna occur but are not marked.
The Derivation and Dispersal of the New Zealand Fauna
Fleming (1962, p. 152) dates the latest possible land connection between Australia and New Zealand as Cretaceous. Earliest fossil Isospondyli are Jurassic, so that it is possible that some groups of fresh-water fishes could have reached New Zealand by land routes. The only fossil traces of fresh-water fishes known in New Zealand are Pliocene (Stokell, 1945, p. 134), and fossil evidence is thus largely lacking. Postulated land connections between Australia and New Zealand are through the Lord Howe Ridge to the north-west of New Zealand connecting with either New Guinea or Queensland, but the New Zealand-Australian fresh-water fish relationships are most marked between south-east Australia and New Zealand. Use of a postulated land bridge for distribution of fish between Australia and New Zealand would mean that the fish groups involved must have had wider distribution to the north in Australia in Jurassic-Cretaceous times. As the Cretaceous was a period of marked cooling (Fleming, loc. cit.), it is possible that this was the case. Apart from the uncertainty of the existence and duration of such a land bridge, the establishment and use of fresh-water routes along the bridge also remains questionable. As suggested below, the use of such a bridge is not essential. The lack of affinities with the bulk of the South American fresh-water fish fauna indicates that the New Zealand fishes did not reach New Zealand from South America by land routes. New Zealand lacks all the primitive and primary fresh-water fish groups (e.g. the Ostariophysi) which abound in South America, and these groups are also mostly lacking from Australia. The general concensus of opinion (e.g. Simpson 1941, Stokell 1950, 1953, Myers 1953) gives no support for the older ideas of Gill (1893) and Oliver page 62 (1925) that the fresh-water fish were distributed to New Zealand by land routes. It seems fairly certain that the New Zealand fresh-water fishes must have arrived in New Zealand since its isolation from other land masses.
Geotria, with marine adult existence and its habit of attaching itself to fishes poses no problem of distribution. Similarly, the arrival of the two eel species in New Zealand is simply explained by transportation of their leptocephali in ocean currents and there is no problem of distribution mechanism. Cheimarrichthys forsteri seems to be a recent local marine derivative. The groups in which derivation and dispersal are less clearly indicated are the Retropinnidae, Aplochitonidae, Galaxiidae and also the Eleotridae.
The Eleotridae in New Zealand are representatives of a wide-ranging Indo-Pacific group of marine and estuarine species, which invade fresh-water in most regions. The closest relationships of the New Zealand eleotrids appear to be with Australia, as both genera present in New Zealand also occur in Australia. The presence of the two genera in both Australia and New Zealand implies double invasion or convergent evolution within the group in the two regions. Contemporary authorities on the Eleotridae in New Zealand (e.g. Stokell, 1959) consider the separation of the New Zealand Eleotridae into two genera unnatural, and that the New Zealand Eleotridae are more closely allied to each other than to the Australian species of each genus. Present opinion tends to favour the view that all the New Zealand freshwater Eleotridae should be included in one genus (Gobiomorphus) and that this group has probably arrived in New Zealand only once. That they did arrive in New Zealand and did not originate here and spread to the north is quite clear from the greater numbers and diversification of the Eleotridae in the tropics to the north-west of New Zealand. The New Zealand fauna is poor in eleotrid species and none are known any further south than New Zealand. The Australian species of Gobiomorphus (G. coxii) is similar to the New Zealand species and is probably closely related. No local marine species of Eleotridae indicate close generic relationship to Gobiomorphus.
The family Retropinnidae is confined to the Australasian region, with three Australian species and six species in New Zealand. Distribution of the family is south-east Australia, Tasmania, New Zealand. Australian species are anadromous, breeding in fresh-water, and three New Zealand species have similar habits, but the other three species are entirely fresh-water dwelling.
The family Aplochitonidae is represented by two species in Australia (one species each of Lovettia and Prototroctes), a species of Prototroctes in New Zealand and the genus Aplochiton in South America and the Falkland Islands. The New Zealand species of Prototroctes is thought to be catadromous with marine larval page 63 life, although this is uncertain. Lovettia seali in Tasmania and Aplochiton marinus in South America both have marine stages in their life histories. Prototroctes in Australia is present in the south-east and in Tasmania.
Finally the Galaxiidae are present on all the southern land masses, and show greatest diversification in Australia (27 spp.) with reduction in the numbers of species towards the east (New Zealand 17 spp., South America and the Falkland Islands 10 spp., South Africa 2 spp.). The pattern of numerical distribution suggests origin of the group in the west, with eastward distribution in the west wind drift from Australia to New Zealand, South America, and South Africa. Except for the presence of Galaxias attenuatus in south-east Australia, New Zealand and South America, species in each area are distinct.
In the above discussion it is readily noticeable that when there are affinities between Australia and New Zealand, the distribution pattern is south-east Australia, sometimes Tasmania, and New Zealand. This pattern applies to the three species in common (Galaxias attenuatus, Anguilla australis and Geotria australis) and to the genera Gobiomorphus, Philypnodon, Prototroctes and Retropinna; in other words, almost all the groups which the author suggests have been derived from other than local marine species. There is thus a common pattern for these groups: viz. south-east Australia, perhaps Tasmania, and New Zealand, with the Geotridae and Aplochitonidae extending further east to South America, and the Galaxiidae present in both South America and South Africa. What is the centre of distribution and means of dispersal of the New Zealand fresh-water fish fauna?
In view of the improbability that the New Zealand fresh-water fish fauna used a land migration route, varying suggestions of derivation and dispersal must be examined. Allen (1956) presents the possibilities as follows: (1) Marine ancestry common to the groups in each area, the ancestors now no longer living; (2) Transoceanic migration of fresh-water forms which are euryhaline; (3) Parallel or convergent evolution of the groups in the different regions.
To hypothesise parallel evolution within the Galaxiidae, Aplochitonidae and Retropinnidae is unreasonable. To do so for one of these groups is less unreasonable, but this leaves the derivation and dispersal of the other two groups to be explained. If an explanation can be offered for the derivation of these other two groups, it is probable that the same reasoning would apply to the group for which parallel evolution is postulated. That three groups should evolve convergently in two, three or even four widely separated areas, all from unknown ancestry, is unlikely.
There is no evidence for or against a common marine ancestry for the groups under discussion, and no recent marine ancestors page 64 are known except for the Eleotridae. For all New Zealand groups including the Eleotridae, the closest relationships are with fresh-water species elsewhere. Marine ancestors giving rise to the Galaxiidae and Aplochitonidae would have had to spread through most of the southern oceans. That such an ancestor should disappear without trace or other apparent derivative is doubtful. The support for the hypothesis of recent marine ancestry for the New Zealand fresh-water fishes does not seem strong.
The third alternative is that euryhaline species have been distributed from their centres of origin by ocean currents or active migration. Both Myers (1938) and Darlington (1957) place all the families of fishes discussed in this paper in the peripheral fresh-water group (i.e. fishes with high salt tolerances) and most groups are found to have representatives which breed in the sea or have some stage of the life history which is marine. That such fresh-water fishes can be distributed across large ocean gaps is indicated by the presence of Galaxias attenuatus in Australia, New Zealand and South America, with little chance that this species was transported by other than ocean currents. Further support for transoceanic dispersal of fresh-water fish to New Zealand is the faunal relationship between south-east Australia (Tasmania) and New Zealand outlined above, and the presence of the warm east-Australian (Notonectian) sea current which impinges on much of the west coast of New Zealand. It is this current which is thought to carry the leptocephali of the New Zealand eel, Anguilla australis, each year from a more northern breeding site across the Tasman Sea. In view of the known ability of Gobiomorphus huttoni, Philypnodon hubbsi, Galaxias attenuatus, Retropinna spp., Anguilla spp., and Geotria spp. to tolerate sea water it is suggested that the fresh-water fish fauna of New Zealand has been derived from the north-west, mostly from Australia, by transportation in the east-Australian current. All the fishes concerned except the Eleotridae are free-swimming pelagic-type fishes and what is known of the life histories of the Eleotridae suggests that some of these have marine pelagic larvae.
Derivation of the New Zealand fresh-water fish fauna by ocean current dispersal across sea gaps seems preferable to the derivation of at least four groups in widely separate areas, each group from an unknown and/or extinct marine ancestor. The numerical distribution of the Galaxiidae strongly supports dispersal from the west with origin in Australia, and in the lack of evidence of movement in the opposite direction, there is no reason to postulate the reverse movement.
The fresh-water fish fauna of New Zealand has been derived mostly from Australia by transoceanic dispersal of larval or adult fishes. Earlier arrivals were probably the more primitive Isospondyli — the Galaxiidae, Retropinnidae and Aplochitonidae — with the Eleotridae arriving more recently. A further emigration of Galaxiidae is suggested by the distribution pattern for Galaxias attenuatus. The distribution of G. attenuatus clearly shows that transoceanic dispersal can disperse fishes across the South Pacific Ocean to South America. This mechanism probably accounts for the fresh-water fish faunal relationships between Australasia and South America.
As Myers (1953) has said, there is nothing in the New Zealand fresh-water fish fauna to indicate land connections; the ‘key to relationships is marine wandering.
I wish to express my thanks to Mr. P. H. J. Castle, Zoology Dept., Victoria University of Wellington, and Dr. R. B. Pike and Mr. L. J. Paul, both of the Fisheries Laboratory, Marine Department, for their help in the preparation of this paper. I am grateful to Mr. K. R. Allen of the Fisheries Laboratory and Mr. J. M. Moreland, Dominion Museum, for reading the manuscript and making numerous valuable suggestions.
Allen, K. R., 1956. The Geography of New Zealand's Fresh-water Fish. N.Z. Sci. Rev., 14(3): 3-9.
Arthur, W., 1884. Notes on New Zealand Fishes. Trans. Proc. N.Z. Inst., 17: 160-172.
Darlington, P. J., 1957. Zoogeography, the Geographical Distribution of Animals. Wiley, New York, 675 pp.
Fleming, C. A., 1962. New Zealand Biogeography, A Palaentologist's Approach. Tuatara, 10(2): 53-108, 15 figs.
Gill, T., 1893. A Comparison of Antipodal Faunas. Mem. Nat. Acad. Sci., 6: 91-124.
Myers, G. S., 1938. Fresh-water Fish and East Indian Zoogeography. Smithsonian Rept. 1937: 339-364.page 67
Myers, G. S., 1953. The Palaeogeographic Significance of Fresh-water Fish Distribution in the Pacific. Proc. 7th. Pac. Sci. Congr., 4: 38-48.
Oliver, W. R. B., 1925. The Biogeographical Relationships of the New Zealand Region. J. Linn. Soc., Botany, 47: 99-140.
Simpson, G. G., 1941. Antarctica as a Faunal Migration Route. Proc. 6th. Pac. Sci. Congr., 2: 755-768.
Stokell, G., 1945. The Systematic Arrangement of the New Zealand Galaxiidae. Trans. roy. Soc. N.Z., 75(2): 124-137, 10 figs.
——, 1950. Fresh-water Fishes of the Auckland and Campbell Islands. Cape Exped. Ser. Bull., 9: 1-8, 1 fig.
——, 1953. The Distribution of the Family Galaxiidae. Proc. 7th. Pac. Sci. Congr., 4: 48-52.
——, 1959. Notes on Galaxiidae and Eleotridae with Descriptions of New Species. Trans. roy. Soc. N.Z., 87: 265-269.