Zoogeography has been a fruitful preoccupation of New Zealand biologists throughout the history of New Zealand biology, and this is probably no accident. New Zealand's geographical isolation, and some striking peculiarities in its biota, both have made zoogeography a fascinating and rewarding pursuit. Furthermore, taxonomy has a very strong heritage in New Zealand, and taxonomy and zoogeography are closely allied and interdependent fields of study.

Zoogeography comes at two levels. In modern times, P. J. Darlington Jr., exemplifies the synthetic zoogeographer and there seems little doubt that his training and experience as a taxonomist (even though he is an entomologist and much of his zoogeography has dealt with vertebrates), has been fundamental to the success of his zoogeography. Most of us must confine our zoogeographical dabblings to dealing with the zoogeography of particular taxonomic groups in which we have a special knowledge and prime interest. And this is where the groundwork of zoogeography is largely done.

Zoogeography is one of the great synthetic sub-disciplines of biology and it requires an appreciation and knowledge of many other fields, both within and beyond the realms of biology — taxonomy, phylogeny, genetics, evolution, ecology, meteorology, geology, oceanography, and so on. The zoogeographer must have at his fingertips some knowledge of all these fields to facilitate interpretation of the distribution patterns he observes. Above all he must have an appreciation of the principles of logic and must examine the whole of the available evidence if he is to avoid falling into difficulties. It is so easy to argue circularly, or to accidentally ignore a part of the evidence. For instance, Ball and Fernando (1969) looked at the distribution of Dugesia, a freshwater triclad, and found that it occurs on all the southern continents, and New Zealand. Customarily, triclads are intolerant of sea water and are very poor at crossing even small ocean gaps, and the occurrence of a species on both sides of a body of water is taken to indicate that the two land areas where the triclad species is present were formerly connected by land. So Ball and Fernando concluded that the Southern Hemisphere distribution of Dugesia was a product of the former existence of Gondwanaland. However, they also pointed out that Dugesia occurs on the Crozet Islands, a group of very isolated islands that seem to be oceanic and relatively young. This suggests that Dugesia must sometimes cross ocean gaps. This being the case, we must consider the likelihood that the distribution of Dugesia page 89 in the Southern Hemisphere could be due to transoceanic dispersal. I'm not saying that continental drift and Gondwanaland did not have anything to do with the dispersal of Dugesia, but the evidence from the presence of Dugesia on the Crozet Islands forces us to consider that continental drift may not have. There is a basic need to consider all the implications of any zoogeographical hypothesis, and in the case of Dugesia, we need to examine how Dugesia got to the Crozet Islands, and consider the impact of this on the primary hypothesis, that the range of Dugesia is a result of the existence of Gondwanaland and subsequent fragmentation and dispersal of the continent.

We must be aware of the dangers of zoogeography as a science, but not be put off by these dangers. In fact, I believe that it is one of the responsibilities of good taxonomists to develop theories of evolution, relationship and zoogeography for the groups with which they work. No other worker is as capable of deciding where evolutionary relationships lie, and most zoogeography is an interpretation of phylogenetic relationships, dispersal, and geology.

Mayr (1969) subdivided taxonomy into ‘alpha,’ ‘beta,’ and ‘gamma’ taxonomy. Alpha taxonomy is the description of taxa, especially new ones, beta taxonomy is the study of the relationships of taxa, and gamma taxonomy is the study of intraspecific variation, evolutionary studies and zoogeography. Some taxonomists argue that a taxonomist should restrict his attention to alpha taxonomy because so many new species remain to be studied and described. They argue this way because their specialisation in some taxon allows them to make their greatest contribution to biology by describing and classifying unknown species. Mayr (1971) questions this attitude, noting that no one is better equipped to derive generalisations from taxonomic data than the specialist — ‘Indeed in 99 cases out of 100, if he does not do it, no one else will do it either… Let us remember that almost all of the major questions of evolutionary biology discussed during the past 50 years were either raised by taxonomists or were a product of taxonomic findings, (Mayr, 1971). Taxonomists do the groundwork necessary for all zoogeography, and it is their loss, as well as biology’s as a whole, if they do not take the extra trouble to determine what their taxonomic findings mean in a zoogeographic sense. It seems to me that zoogeography is a taxonomist's responsibility, and it is unfortunate that many a fine taxonomic treatment of some group or other, complete with elaborate distributional data has no discussion of the zoogeography of the group.

Apart from the fact that taxonomists have much to say zoogeographically, it seems to me that the taxonomist can better understand and deal with his taxonomic problems if he looks at the zoogeography of the animals he studies. This is a dangerous area, as it is very easy to argue circularly, and very easy to be badly page 90 misled. The value of zoogeography is limited by the soundness of the taxonomy on which it is based. The zoogeography of New Zealand's butterflies, for instance, has been revolutionised as a result of far-reaching changes in the generic classification of these insects. Some butterflies thought to have South American relationships, are now believed to have European ones. There is a strong sense in which a zoogeographic study is a study not of a group of animals, but of our taxonomy of that group. And insofar as the taxonomy is in agreement with the animals, our zoogeography has meaning. So there is a heavy responsibility on the taxonomist to produce proper and sound taxonomies, whether or not he is going to indulge in zoogeographic speculations.

Let's take a look at New Zealand moas. On the basis of morphological evidence, 28 species have been described, almost all recent or subfossil, so essentially contemporaneous in New Zealand (Oliver, 1955). And remember that it seems that New Zealand was a very small land area in the Oligocene (Fleming, 1962). So, probably since then, as the land area expanded again, the moa fauna evolved to reach a total of 28 species. From an evolutionary-zoogeographic viewpoint, it is very hard to understand how this number of very large and rather specialised birds could have evolved and lived contemporaneously in the small area of New Zealand. What is apparently clear to the ornithologists on morphological evidence is rather incomprehensible from an evolutionary-zoogeographic viewpoint. The ornithologist must have the last word, but zoogeographers can, I think, have useful things to say about the taxonomies and may ask questions that can assist the taxonomist. I would very much like to see a comprehensive series of measurements taken on available moa bones, and have them submitted to detailed computer analysis to see if any of the material can be tied together as growth, sexual or geographic variations. It seems most unsatisfying that New Zealand has 28 moa species but less than 100 other land birds of diverse character and habits, and that New Zealand has more than threequarters of the entire ratite bird group.

Recently, I have had a look at the scope of the galaxiid fish genus Brachygalaxias. Several species have, at times, been included in this genus. The type species in South America has one unique character, it shares another peculiarity with two Australian species that are very similar to each other, and shares a further peculiarity with both Australian species and a further species in South Africa, (unpublished data). The question the taxonomist must answer is where to limit the genus. But the answer to the question has far-reaching zoogeographic implications. If the Australian and/or South African species are included in Brachygalaxias with the South American species, we are indicating strong belief in close phylogenetic relationships, and posing a zoogeographic question for which we page 91 must seek an answer. We must attempt to explain how species in Australia, South America and South Africa have a common ancestry which relates them more closely to each other than to any other galaxiids. It seems to me that a taxonomy with all the species in Brachygalaxias creates zoogeographic problems that the morphological evidence on relationship probably does not warrant. The alternative is to retain the South American type species in Brachygalaxias, and to regard the other species as fringe species of Galaxias, which, on the basis of certain morphological evidence, may be regarded as related to Brachygalaxias. By choosing this alternative, the taxonomist is not forcing the zoogeographer to seek explanations of relationships and distribution patterns that may be a taxonomic artefact. And yet, the question of relationship, and the intriguing zoogeographic problems remain open for consideration. The interplay of taxonomy and zoogeography in this way should produce taxonomies that have the greatest utility and retrieval value for users of the taxonomic treatment, for zoogeographic or other purposes.

While concordance between taxonomy, phylogeny, and zoogeography is nice — it's very satisfying to draw a cohesive and consistent picture — I think that we must beware of organising taxonomies to produce nice, perhaps preconceived zoogeographic pictures. This is a very real danger. Taxonomies must be based on taxonomic evidence and taxonomic principles. If the taxonomy results in a satisfying zoogeography, that is nice. If it doesn't, may be we can and should take another look at the taxonomy, but we should not and must not reorganise the taxonomy to tidy up the zoogeography.

Zoogeography in our day is undergoing a profound revolution. The edifice of modern historical zoogeography is very clearly built upon the foundation of Darlington (1957) ‘Zoogeography — the geographical distribution of animals.’ This book deals only with land and freshwater vertebrates, and it follows as a basic premise, the view that the continents have always been roughly where they are now. And it is no discredit to Darlington that, largely since this book was written, geologists have become increasingly convinced that the continents have, in fact, not always been where they now are. Darlington, told by geologists, that the continental positions had not changed, had to rationalise the distribution of animals on this basis. Now, we are told that the continents have dispersed away from Gondwanaland, a large southern continent, and we will now have to establish a whole new set of rationalisations. Stable continent zoogeography, confined largely to vertebrates, left plenty of perplexing problems — characid fishes and ratite birds are two of the more horrendous zoogeographic problems for which neither Darlington nor anyone else has had any really satisfying answers. It is interesting that invertebrates, especially insects, had little wide-ranging attention from stable continent zoogeographers, and I suspect that the problems of characid fishes and ratite birds may have page 92 paled into insignificance compared with the problems that insects would have caused. It is interesting too that many of the really vexing problems of zoogeography, as formulated by stable-continent zoogeographers, are Southern Hemisphere problems, and that most of the attention being paid to continental drift is involed with the fragmentation and dispersion of the southern land masses on which these zoogeographic problems exist.

The increasing confidence of geologists and geophysicists that continental drift is a fact, means that drift warrants, or rather demands serious attention from zoogeographers. The impact of this on zoogeography is going to be substantial. I believe that zoogeography has very little indeed to say about the former positions of the land masses, and at this point, we are very much at the mercy of the geologists. And if they say that Gondwanaland split up some time during the Cretaceous, and that various fragments moved in certain directions, at given rates, at various times in the past, then we have little option but to accept this, and investigate the impact of these theories on zoogeography. Perhaps zoogeographers will help to refine the dating of continental fragmentation using paleontological evidence, but I believe that we can have little more to say than that.

Following his 1957 synthesis of animal distributions on the basis of stable continents Darlington (1965) has begun to restructure his zoogeography. This 1965 book is a reflection of Darlington's efforts at an approach to continental drift zoogeography, built upon the foundations of his earlier work, and the tensions between these two mutually exclusive approaches are evident.

The acceptance of continental drift has, as I see it, three impacts on existing zoogeographical knowledge:—

1.	It may help to explain the previously inexplicable.
2.	It may lead to the replacement of old explanations for patterns of distribution by new explanations.
3.	It may produce a whole new array of inexplicables.

Early attention is bound to be paid to the first of these, but I have little doubt that all three of the above will apply.

What is the impact of continental drift on New Zealand zoogeography? New Zealand is believed to have moved away from Antarctica about the Cretaceous, about 50-60 million years ago. It can be regarded as having been an isolated land area ever since, although the area of land is believed to have fluctuated broadly (Fleming, 1962). Thus the effect of continental drift, and the departure of the New Zealand land mass from Antarctica is related to: 1. The initial departure from Antarctica; and 2. The increasing distance between New Zealand and Antarctica.

There should be evident in the New Zealand fauna signs that New Zealand was suddenly cut off from other land areas. A whole series of taxa should show signs of having been contemporaneously isolated in New Zealand. However, because it was so long ago, the page 93 effects of this isolation have been masked by extinction, speciation, phyletic evolution, etc. But presumably the classic problems of New Zealand zoogeography (all ancient and apparently primitive forms) date back to this period. These include the tuatara, moas, kiwis, the native frog, parastacid crayfishes, hyridellid mussels, earthworms, peripatus, etc. Here is a series of taxa that are ancient and whose zoogeography has been a vexing problem but which tend to fall into place when we consider New Zealand's involvement with Gondwanaland in the Mesozoic. The New Zealand fauna is commonly looked at as a dispersal fauna, a fauna that has arrived here over the years by chance dispersal. Excluding the ratites, this is certainly true for the birds; it applies to the few mammals even though nothing seems to be known about the relationships of the short-tailed bat and it may be a Gondwanaland relict. But until we know something about the relationships of this bat, Gondwanaland is of no assistance in understanding how it got to New Zealand. Reptiles, apart from the tuatara, are clearly vagrants capable of having dispersed to New Zealand, and must have done so to explain close and apparently relatively recent Austral-Pacific relationships. Gondwanaland is probably not needed to get the tuatara to New Zealand either as it may be capable of dispersing across the sea. Even if we conclude that the presence of the tuatara in New Zealand is related to the former existence of Gondwanaland, we are only indicating that the unknown ancestors of the tuatara were somewhere in Gondwanaland, and we are little better off.

The frog is another problem, as are the crayfish, mussels and earthworms. These seem unlikely to have been able to cross oceans in the way the tuatara could have. Their very presence in New Zealand, has been enigmatic, but probably can be explained by New Zealand's involvement in Gondwanaland.

Much of the New Zealand freshwater fish fauna has relationships with southern areas (McDowall, 1964). The Southern Hemisphere has a characteristic southern temperate freshwater fish fauna that is very small. This comprises the galaxioid fishes (Sub-order Galaxioidei, Families Galaxiidae, Retropinnidae, Aplochitonidae, Proto-troctidae) and the southern lampreys (Family Geotridae). The galaxioid fishes are a southern temperature radiation comparable with the northen temperate salmonoid fishes (Sub-order Salmonoidei, Families Salmonidae, Osmeridae, Plecoglossidae, Salangidae); but the lampreys are merely a family in the south comparable with a family in the north (Family Petromyzonidae). These southern freshwater fishes appear as if they could have attained their present distribution in relation to the fragmentation of Gondwanaland, and there has been much interest in the past especially in the distribution of the Galaxiidae (Australia, New Caledonia, New Zealand, South America and South Africa). This distribution has been taken as evidence that there have been connections between these page 94 land areas. The distribution of the Galaxiidae is an apparently classic example of a Gondwanaland distribution, But if we look more closely at these fishes we find that there is a species (Galaxias maculatus) common to Australia, New Zealand and South America, without any evidence of even separate subspecies; or races. If the range of this species is related to Gondwanaland, how can we explain the conspecificity of these highly disjunct populations after about 50 million years of isolation? If we take another look, we also find that this species has a marine stage in its life cycle, so is easily capable of transoceanic dispersal. Boulenger pointed this out as long ago as 1902. but no one seemed to take any notice of him. Another galaxiid (Galaxias brevipinnis) is present in Tasmania, New Zealand and the Sub-Antarctic Islands, and it too has a marine stage in its life cycle. Species common to western and eastern Australia and/or Tasmania also have marine stages. If we look at the distribution of galaxiids in New Zealand, we find two types of distribution — widespread species and localised species — and we find that the widespread species have marine stages and the localised ones do not. All these facts indicate very clearly that there has been dispersal of galaxiid fishes through the sea (McDowall, 1970). The southern lamprey Geotria australis is found in western and eastern Australia, Tasmania, New Zealand, Chile and Argentina, and like the galaxiids, the lamprey has a marine life history stage; the conspecificity of populations in all these areas shows that it also has dispersed through the sea.

Caughley (1964) wrote a paper under the stimulating title ‘Does the New Zealand vertebrate fauna conform to zoogeographic principles?’ With the admitted tremendous advantage of hind-sight, Caughley asked us to go along with him as he constructed a theoretical New Zealand vertebrate fauna, on the basis of disperal from the faunas of likely contributing land areas. Caughley assumed that New Zealand is an oceanic island, ruling out land connections and continental drift, for his exercise. For what they are worth, his predictions were good, and his failures were heavily weighted towards the archaic elements in the New Zealand vertebrate fauna, i.e.: the frog, tuatara, moas and kiwis. These are the sections of the vertebrate fauna that are most likely to have had origins in Gondwanaland. As Caughley excluded land connections of any type, we would have expected him to have had difficulty predicting the presence in New Zealand of any flightless animals that don't disperse across the sea, and which must have reached New Zealand by land routes. Note that the group of animals that: Caughley had difficulty predicting for the New Zealand fauna is almost identical with the archaic element in the fauna Fleming, 1962).

Thus, to answer my earlier question, the impact of continental drift on the zoogeography of New Zealand, from the point of view of vertebrates, is relatively slight in terms of the entire vertebrate page 95 fauna, although it does help to solve some old and vexing problems. The picture may be a different one for invertebrates. The impact is likely to be much more profound, but at present I don't think that the taxonomies are sufficiently well established to permit much progress in this area. Brundin (1965) has claimed to have proved trans-Antarctic relationships for the chironomid midges, but Darlington (1970) has been very critical of Brundin's methods of analysis. It seems probable that ‘trans-Antarctic ‘relationships exist. But I wonder whether they are anything like those proposed by Brundin.

One of the interesting impacts on zoogeography of continental drift is on the origins of bipolar groups. It has long been recognised that there are pairs of groups of organisms with one of the pair in the north-temperate — sub-Arctic and the other in the south-temperate — sub-Antarctic. Gill (1893), and more recently Hubbs (1953) have examined this subject, the problem being to explain how groups apparently intolerant of the high temperatures of the tropics, could have dispersed across the tropics. Hubbs suggested that some of the fishes could have crossed the tropics by descending into deeper, cooler water, and others may have done so when the tropics were cooler, as during the Pleistocene glaciations. But for many groups, neither suggestion is helpful.

For some old groups, e.g. the galaxioid (southern) — salmonoid (northern) pair, or amongst the plants the Nothofagus (southern) — Fagus (northern) pair, the former existence of a formerly united continental mass may assist the explanation of how and why there are northern temperate groups with obvious, closely related radiations in the southern temperate.

The field of New Zealand zoogeography is wide open, and as we see the taxonomy of New Zealand invertebrates develop, I think we will see a great flowering of New Zealand zoogeography. If this flowering had taken place 20-25 years ago when the continents were regarded as having been stable, and when we knew little about the taxonomy of our invertebrate fauna, I think that we would have had zoogeographical chaos. Now, the time seems ripe for this zoogeography really to flourish.

Literature Cited

Ball, I. C., and Fernando, C. H. 1969: Freshwater triclads (Platyhelminthes, Turbellaria) and continental drift. Nature, Lond. 221: 1143-1144.

Boulenger, G. A. 1902: The explanation of a remarkable case of geographical distribution among fishes. Nature, Lond. 67: 84.

Brundin, L. 1965: On the real nature of trans-Antarctic relationships. Evolution 19: 496-505.

Caughley, G. C. 1964: Does the New Zealand vertebrate fauna conform to zoogeographical principles? Tuatara 12 (1): 49-56.

Darlington, P. J., Jr. 1957: Zoogeography, the geographical distribution of animals. Wiley, New York 675 pp.

——, 1965: The biogeography of the southern end of the world. Harvard University Press, Cambridge. 236 pp.

——, 1970: A practical criticism of Hennig-Brundin ‘Phylogenetic Systematics,’ and Antarctic biogeography. Syst. Zool. 19 (1): 1-18.

Fleming, C. A. 1962: New Zealand biogeography — a palaeontologist's approach. Tuatara 10 (2): 53-108.

Gill, T. 1893: A comparison of antipodal faunas. Mem. Natn. Acad. Sci. 6: 91-124.

Hubbs, C. L. 1953: Antitropical distribution of fishes and other organisms. Proc. Seventh Pacif. Sci. Congr. 3: 324-329.

McDowall, R. M. 1964: The affinities and derivation of the New Zealand fresh-water fish fauna. Tuatara 12 (2): 56-67.

——, 1970: The galaxiid fishes of New Zealand. Bull. Mus. Comp. Zool. Harvard 139 (7): 341-432.

Mayr, E. 1969: Principles of systematic zoology. McGraw-Hill, New York.

——, 1971: Methods and strategies in taxonomic research. Syst. Zool. 20 (4): 426-433.

Oliver, W. R. B. 1955: New Zealand Birds. Reed. Wellington. 661 pp Revd. Ed.