We have all at one time or another heard it stated that New Zealand has a unique fauna. This is true, as it is true of any other island or continent. But the statement often carries an implication that somehow the fauna is extraordinary and strange—quite different from what might be expected by a rational man.

Let us examine this connotation by resorting to a stratagem. Suppose that by some accident of history New Zealand was not discovered until last year and that a biologist is asked to comment on what vertebrates are likely to be discovered in this new unexplored land. As yet no expedition has landed, and he has nothing to guide his predictions other than his knowledge of animals in other countries. I am presumptuously casting myself in the role of this biologist. With what, I hope, is a minimum of unconscious cheating, I shall reconstruct what I guess would be his conclusions and the reasoning by which he would arrive at them.

From a knowledge of other faunas, several principles can be derived to help in predicting the nature of the New Zealand fauna:

(a) A high proportion of the forms occurring on islands, but seldom all of them, are derived from the fauna of the nearest land mass. If the island has been linked to the mainland in the past, the fauna will be more diverse than if the island is truly oceanic. [The rocks on the east coast of Australia argue for the existence of the Tasman Sea at least since the beginning of the Cretaceous (Glaessner, 1962) and so, from a strictly biological viewpoint, New Zealand can be treated as an oceanic island.]

(b) The probability of extinction on large islands is less than on small islands (Mayr, 1954). [New Zealand would therefore have a better chance of retaining an invading species than would for instance, a small volcanic island in a comparable position].

(c) The probability of a species reaching an island from a continent is intimately related to the distance between the two, and to the climate of the intervening region. [New Zealand is about 1,000 miles from Australia. This fact is probably the most important in deciding what might be here].

(d) ‘Dispersal of individual land animals over water is largely accidental … but in the course of time statistical probability comes into play and determines what sort of animals cross water most often and what islands they most often reach’ (Darlington, 1938). But for any process conforming to statistical laws the end result cannot be exactly predicted. Some effects will be very likely while others will be most unlikely, but one can never say ‘this is absolutely certain’ or ‘this is theoretically impossible’.

Fresh-water Fish

With considerable confidence I would predict the absence of strictly fresh-water fish. In most cases they are incapable of crossing salt-water gaps of more than a few miles. Where they do occur on islands it is usually possible to prove that at one time the island was connected to the mainland (Darlington, 1957).

My prediction is correct: there are no strictly fresh-water fish in New Zealand. Some species do spend their entire lives in fresh water, but are members of widespread families (Galaxiidae, Eleotridae and Retropinnidae) many species of which are tolerant of salt water (Stokell, 1955). The presence of the Galaxiidae is expected as this family is almost ubiquitous within the Southern Temperate Zone.

Amphibians

Frogs have reached the Seychelles from Africa (500 miles), but do not disperse widely in the temperate zone. The inference is obvious: there is little likelihood of frogs occurring in New page 51 Zealand. Unfortunately for theory, they are here. This is a major surprise, but the blow is softened a little by finding that these frogs (Leiopelma) are of an ancient stock that has no close relatives elsewhere. They may therefore have dispersed to New Zealand at some time in the distant past when the distribution of land differed from now.

Reptiles

One must first decide whether any reptiles are likely to be here at all. The distance from Australia would rule out fresh-water turtles and probably also land snakes. Snakes have crossed some wide gaps in the tropics (e.g. 600 miles to the Galapagos) but in temperate zones they appear to be incapable of dispersing far across water. There are none on the Azores or Madeira. 800 and 350 miles respectively from a continent, and even the Canary Islands (60 miles) have none. Hence the presence of snakes seems improbable, and conveniently none are here.

Lizards disperse further than do snakes. They have reached the Galapagos in the tropics but have not colonised Hawaii, 2,000 miles from a continent. Skinks and geckos have crossed to New Caledonia, presumably from Australia (800 miles), but outside the tropics they are less successful colonisers. They have reached Bermuda from North America (600 miles) but do not occur on the Azores. Thus New Zealand is very near their limit of dispersal, or beyond it. If New Zealand were smaller I would predict the absence of lizards, but as it has a fairly long coastline oriented across the prevailing wind, the probability of lizards reaching here is thereby increased.

I would probably be non-committal about the possible presence of lizards, but would add a rider that if they are here at all they will be represented only by those groups in Australia that have powers of wide dispersal. The monitors and agamids can be dismissed because although they occur in Australia and are widespread they have not managed the crossing from Africa to Madagascar. Dismissed too are the pygopods which have not reached New Caledonia from Australia. Only skinks and geckos survive this trial by elimination.

In fact, skinks have reached here at least twice (two Australian genera are represented) and geckos at least once (three endemic genera). I am ignoring the records of Lepidodactylus lugubris and Gehyra oceanica (McCann, 1955) whose distributions in New Zealand suggest that they may have received an assisted passage.

I would not have suspected the presence of the tuatara (Sphenodon) in my most irresponsible imaginings. Like the New Zealand frogs it is completely unexpected.

Birds

Of all vertebrates, birds are by far the widest dispersers. Very few islands large enough to retain vegetation are completely lacking in land birds.

Following the principle that most of the fauna should be derived from Australia, I can list the families of land birds most likely to occur in New Zealand (Table 1). My rule of thumb is to assign a high probability to widespread families that breed in the south-east of Australia (Bassian province).

Of the 38 families listed, all but 11 are represented. Considering the chance processes with which we are dealing, this is a most gratifying proportion.

The following families are reported from New Zealand but are not on the list of high probabilities: Apterygidae, Acanthisittidae, Callaeidae, Turnagridae, Anhingidae, Glareolidae and Phaloropodidae. The first four are restricted to New Zealand while the last three do not breed here and have been reported only as stragglers.

Birds are zoogeographically the most predictable of all groups and I can therefore risk predicting down to the species level. There is a reasonable probability of an Australian species occurring in New Zealand if it is fairly wide ranging (showing that it has good powers of dispersal) and breeds in Tasmania (showing that it could probably tolerate the New Zealand climate). My criterion for the ability to disperse—that the species has penetrated into the Sunda Chain or the Western Pacific—is not a very satisfactory one as it tends to favour tropical species. It is forced on me by circumstance, for there is no convenient chain of temperate islands extending out from Australia.

By my count there are 43 species that qualify; they are listed in Table 2.

The score of 33 out of 43 is high and even I (in my capacity as the hypothetical biologist) am agreeably surprised. But my exuberance is dampened somewhat by finding in the list of wrong predictions one whose presence I would most confidently predict. This is the peregrine falcon which by its wide distribution and tolerance of a broad range of habitats earns a high place in the list of probabilities.

It should be noted that although the prediction about the scarlet robin was wrong, some of the robin's ancestors reached here at least twice in the past, giving rise to forms now specifically distinct from Petroica multicolor (Fleming, 1950).

On most islands of any size there is a stratified order of endemism ranging from endemic sub-species to endemic families, reflecting differing periods of time that the stocks have been isolated. I would correctly expect the same to occur here. Falla (1953) has shown that there is a continuous gradation of relationships, mainly with Australian birds, and Fleming (1962a) has tentatively dated the postulated invasions.

I would not have predicted the occurrence of the two endemic orders (kiwis and moas), even though the occurrence of flightless birds has a parallel in Madagascar. Loss of flight is not a common consequence of island colonisation except in rails, and there is no reason therefore to predict it. But the New Zealand bird fauna in this respect does not form an exception—only an uncommon case. Although failing to predict the presence of moas and kiwis, I would not be greatly surprised to find them here.

New Zealand and South America are separated by about 4,000 miles. This gap is a little wide even for birds, especially as they would fly into westerlies most of the way. Birds of South American derivation would hardly be expected here. Predictably, there are no land birds in New Zealand with unambiguous South American affinities.

Mammals

The presence of large mammals can be virtually discounted. The farthest they have dispersed across water is only 200 miles (the Malagasy hippopotomus). Similarly, the presence of small mammals is rather unlikely as they have not colonised New Caledonia from Australia, a much less arduous task than reaching New Zealand. The marsupials do not appear to have great powers of dispersal for they have not reached Asia from Australia, although phalangers have reached the Celebes.

Of the flightless small mammals, the only possible (though not highly probable) contenders are the rats. One might expect that the Australian water-rats (Hydromyinae) would be well suited to dispersal, but their failure to enter the Sunda Chain rules against this contention. The genus Rattus, however, is widespread in the Pacific and travelled along the Sunda Chain in the late Tertiary or Pleistocene. If there are any flightless land mammals in New Zealand they should belong to this genus. But Rattus has not dispersed here; the three species in New Zealand have been introduced by man.

Bats disperse widely and have crossed water gaps as great as 2,000 miles (America to Hawaii). They are therefore likely to be an element in the New Zealand fauna. There are but two genera here, both of which are widespread in New Zealand (Dwyer, 1962). One is a species that occurs also in Australia while the other is the sole representative of a presumably autochthonous family. The problem is not that bats are here but that there are so few; by analogy with New Caledonia I would have expected about twice as many genera.

Any seal breeding on the New Zealand coast should be a sea lion (Otariidae). The walruses (Odobenidae), northern seals (Phocinae) and monk seals (Monachinae) occur only in the page 55 Northern Hemisphere, while the elephant seals (Cystophorinae) and Antarctic seals (Lobodontinae) are largely restricted—with the exception of the Californian elephant seal—to polar or sub-polar waters. These groups need not be considered further.

Of the sea lions, Arctocephalus is almost certain to occur as various species of this genus breed at the margins of most southern temperate lands (Davies, 1958). The genus is represented here by Arctocephalus forsteri. I would also expect one of the larger sea lions to be present. Neophoca is the most likely as it occurs on the coast of Australia, and Eumetopias is another, though lesser, possibility. But the coast of New Zealand is devoid of breeding sea lions of these genera, although Neophoca hookeri breeds at the sub-antarctic islands to the south. Why this species has not extended farther north is a mystery. No other species of the genera of larger sea lions (Otaria, Zalophus, Neophoca, and Eumetopias) are restricted to sub-polar regions.

Conclusions

How have the predictions stood up against the facts? I can claim success with fish, poor scoring with the amphibians, and reasonably accurate predictions for reptiles, birds and mammals. A few occurrences are quite unexpected, but these are mostly of members of Fleming's (1962b) endemic or archaic grouping—forms that have been here for a long time and have no close relatives in other areas. Most countries have their quota of such forms, and I might have predicted that there would be a residual group defying individual prediction.

In all, I have been fairly successful. There have been surprises but probably no more than would be produced by a similar analysis of another fauna. There would be no way of predicting, for instance, the occurrence of that remarkable bird, the kagu, in New Caledonia, or iguanas in Madagascar and Fiji. The New Zealand vertebrate fauna cannot therefore be considered a special case for it conforms to the same zoogeographic principles that operate elsewhere.

Addendum

In this investigation of the vertebrate fauna, the main working hypothesis has been that most forms were derived from Australia. This is a fair assumption but should not be extended to imply that all animals in New Zealand share this history. It is a mistake to assume that forms necessarily arose elsewhere and migrated to the particular region being considered, but this assumption is implicit in much of the zoogeographic literature. For instance, a page 56 group of snails common to Australia and New Zealand is often called the ‘New Zealand element’ in Australia and the ‘Australian element’ here. Someone must be wrong.

Conceptual clichés are common in zoogeography and often lead to inaccurate conclusions. The three unwarranted assumptions most often encountered, usually in heavily disguised form, are as follows:

Acknowledgments

I am grateful to Mrs. M. Grange, Dr. R. A. Falla, Mr. D. H. Brathwaite and Mr. M. J. Wraight for commenting on earlier drafts of this paper.