The Flora of Australia as a Measure of the Antiquity of the Angiosperms

KARL SUESSENGUTH 1

INTRODUCTION only in Australia and in New Guinea. These , IF ONE INTENDS to consider as difficult a mammals resemble reptiles more than any of problem as the early history of a large land the other mammals because they lay eggs, area, Australia would seem to be particularly have a cloaca, and still possess the number of suitable for such a study: its long geographic shoulder bones of primitive animals. As fos­ isolation and the great number of scientific sil evidence has shown, their ancestors ap­ investigations to which it has been subjected peared during the Triassic formation , to be­ make it a natural choice. come, in fact, the first of the mammals. They Australia's land connection with the island increased in number during the Jurassic pe­ groups of Malaysia (except for New Guinea) riod, but, to a great extent, they died out was ended in the Upper Cretaceous period. as early as the Eocene epoch. In Australia, According to physicists' calculations, based however, some of these primitive mammals on the rates of disintegration of radioactive have survived to this day, affording us illus­ elements, about 30-40 million years have tration of the concept of "endemism by passed since the Eocene epoch in early Ceno­ conservarion." zoic time. Inasmuch as the Upper Cretaceous The marsupials, too, are notably typical of period occurred before the Eocene epoch, it Australia. They are not limited to Australia, can be concluded that, on the whole, the Aus- for there are opossums in North and South .tralian flora and fauna have remained undis­ America, and Chironectides in South Ameri­ turbed for an extremely long time. Under ca; but the great majority of the marsupials these circumstances of isolation, ancient is found only in Australia. They have devel­ forms of plant and animal life have been oped there, it is interesting to observe, in a preserved, while, during the long periods manner analogous to the development of since the isolation began, new species of . placental mammals of the other continents, plants and animals have developed from notably the carnivores, rodents, insectivores, . them as well. It must not be forgotten, how­ . and ungulates. ever, that immigration also has occurred, in­ The survival of the Monotremata and of troducing new species into Australia's plant the Marsupialia can be attributed to the fact and animal life since the beginning of its that, before man's appearance on the 'Aus­ geographic isolation. tralian continent, no other placental mam­ The many investigations into the animal . mals existed in Australia to prey upon them. life of Australia have given unequivocal evi­ There were only mice, which sometimes dence of the continent's isolation. Today the wandered on driftwood from island to island, most primitive mammals-the Monotremata and bats; but these were not significant .ene­ (Ornithodelphia) and the species of Echidna, mies. (It may be possible, of course,that the Proechidna, and Ornithorhynchus-appear mice and the bats first arrived in Australia in times subsequent to the Upper Cretaceous 1 Botanische Staatssamrnlung, Munich 38, Ger­ many. Manuscript received November 20, 1948. period and, therefore, subsequent to the be-

[ 287] 288 PACIFIC SCIENCE, Vol. IV, October, 1950 ginning of Australia's geographic isolation.) tive, while others think that the Polycarpicae As far as the dingo is concerned, it is sup­ among the Choripetales are the most ancient. posed that aboriginal Austral inhabitants en­ Once, even certain of the Monocotyledones tering Australia from Malaysia were the (the Pand anales ) were considered for the first to bring this animal to the -southern distinction, although this claim, of course, continent. could not be proved by any significant argu­ It is also pertinent to our thesis to note ments. Inasmuch as this is hardly the place that zoologists believe that some of the opos­ for a discussion of the phylogenetic criteria sums (Caenolestes) migrated to North Amer­ by which plants are judged, the more perti­ ica from South America. Fossils of Marsu­ nent portions of my book, Neue Z iele der pialia have been found in Europe and in Botanik (1938), are suggested for reference. North and South America, an indication that If, in our search for the oldest angio­ they must have been distributed over vast sperms, and in our analysis of the Australian regions of the earth. The Marsupialia have flora, we hold the opinion that those families survived chiefly in Australia, and for this which are put at the head of the Monochla­ reason Australia today has the oldest and mydeae in the Engler and Prantl system of most primitive mammal types in the world. .classification are the ones which show the Now because, geologically speaking, mam­ most prim itive characteristics, we should be mals and angiosperms are of about the same supported in this assumption by the Casuari­ age, it is natural to ask if the oldest and most naceae found in Australia. In their original primitive of the flowering plants are also to distribution they extended from Sumatra to be found in Australia today. The answer to the Philippines, New Caledonia, and the Fiji this question would help us to determine the archipelago (Diels, 1926 ), and to Tahiti, the antiquity of the angiosperms. Austral Islands, and the Marquesas ( Brown, This is a question that is difficult to an­ 1935 ). As yet, however, there is no reliable swer,if only . because botanists are not in basis for the hypothesis that the Casuarina­ agreement on the most primitive species of ceae are more nearly related to the Gymno­ angiosperms. This one question asks other spermae than is any other family of the An­ questions : If the earliest species of angio­ giospermae. Neither can this supposition be sperms could be defined, would it be found proved for the Proteaceae and the Balanop­ that they exist preeminently or even exclu­ sidaceae, which are also placed at the begin­ sively in Australia? Or, if the earliest species ningof the Monochlamydeae in the Engler c annot be defined, is it possible to determine , and Prantl taxonomic system. In Australia from the Australian flora of today, which are more than half of all ofthe species of Mono­ the most primitive species of angiosperms chlamydeae are Proteaceae (a bout 600 .spe­ that have succeeded in persisting until this cies) , although the family has extended to time ? These are the problems to be investi- southern Africa, southern Asia, and South gated in this paper. . America (Vester, 1940) . The species of Proteaceae are almost exclusively ligneous WHAT ARE THE MOST PRIMITIVE FAMILIES plants, which would indicate that, phyloge­ OF THE FLOWERING PLANTS? netically, they are rather an old group, but in the absence of paleontological evidence As every botanist knows, the question of we cannot be certain that these species are primitiveness in flowering plants is a contro­ really older than many others we might con­ versial one. Some think that certain species sider, so we have to be content only with of the Monochlamydeaeare the most primi- supposing thatthey might be. Antiquity of the Angiosperms-SUESSENGUTH 289

On the other hand, if we hold the opinion, ing plants? Or may those various plant spe­ as many botanists do, that the Polycarpicae cies now found in Australia have migrated are the most primitive of the Angiospermae, to the continent since its separation because we should find that only two very small fami­ they-or rather, their seeds-could cross the lies of Polycarpicae are endemic in Australia: ocean gap while the animals were not able the Eupomatiaceae and the Himantandraceae, to do so? which are related .to the 'Eupomatiaceae, but To obtain a general view of the whole which possess neither calyx nor corolla flora of Australia, let us consider thecata­ (Diels, 1919: 126 et seq.). The other fami­ logue of F. von Mueller, his Census of Aus­ lies of Polycarpicae are not well represented tralian Plants (1 889). Because of the recent in Australia : the Magnoliaceae afford only advances in our knowledge, Mueller's list is 4 species, the Annonaceae 18, the Nymphaea­ neither complete nor infallible, yet it is not ceae 5, the Ranunculaceae 17, and the Myris­ likely that the proportions of the numbers ticaceae 1. In view of these facts, it cannot of species within large groups and of the . be claimed that the Polycarpicae are the most endemic species have changed significantly primitive of angiosperms in Australia and since that time. Therefore we may use the that a comparison with the preservation of Census without hesitation , all the more neces­ the earliest mammals could be made. sarily because there does not exist a later cat­ This brief appraisal is enough to show that alogue for the whole Australian territory the angiosperms which most botanists con­ (including Tasmania but not Ne w Zealand ). sider to be the oldest of flowering plants Mueller's catalogue lists 8,842 species, (Polycarpicae) do not exist exclusively or pre­ and, because it does not mention those spe­ eminently in Australia. But in the informa­ cies introduced in recent times (since about tion we have learned about the Casuarinaceae 1800), it is well fitted for our purpose. Of and' the' Proteaceae we may have found cer­ these 8,842 species, 7,734 (that is, 87.5 per tain clues which will be of value later when cent) are endemic in the larger sense of the we investigate their degree of primitiveness. word- that is to say, they are found in Aus- - tralia itself but may also extend to New W HICH SPECI ES OF ANG IOSPERMS Zealand and to parts of Polynesia as well. PERSISTING IN AUSTRALIA ARE The percentage of endemism is extremely THE MOST PRIMITIVE? high. Now we can set about answering the sec­ Table 1 may serve for comparison of the ond question, inquiring into the conclusions percentage of endemic plants found in Aus­ which can be drawn from a study of the his­ tralia with those found in other parts of the tory of the Australian flora. At first it may world. seem questionable in itself to compare the As a matter of fact, it is probable that early histories of flowering plants with those among the 8,842 species listed in Mueller's of mammals, particularly when it is realized Cens1JS of Australian Plants there may be a that the conditions governing their migra­ great many species which were introduced by tions were quite different. We assume that man , although this hypothesis cannot be those mammals living in Australia during substantiated in its details. If this is true," the Upper Cretaceous period have been iso­ however, the percentage of endenism in Aus­ lated since that time because of the conti­ tralia would be even higher than it is here nent 's isolation, and that only rarely have calculated. . they been joined by later immigrants. Can There is no doubt but that the longer a the same assumption be made for the flower- country has been isolated the more endemics 290 PACIFIC SCIENCE, Vol. IV, October, 1950

TABLE 1 fore the sparse soil-cover left by the retreat­ R EPRESENTATION OF ENDEMIC PLANTS IN ing glaciers was removed by erosion, and AUSTRALIA AND IN OTHER PARTS OF THE WORLD because the short period during which allu­ vial soil has been collecting since glaciation APPROXIMATE P ERCE N TAGE 'OF has not been long enough for the develop­ E NDElIoUC REGI ON PLANTS REF ERE NC E ment of many new species. It is true, of Australia 87.5 calculated from course, that Scandin avia can hardly be com­ Mueller, 1889. pared with Australia, inasmuch as the pre­ New Zealand 73 calculated from vailing temperatures in Scandinavia are not Cheeseman, 1925 . Hawaii 90 according to St. at all favorable for the formation of new John, 1946 . species (Sterner, 1943 : 84) . Gal apagos 40 calculated from Even if we considered as not being en­ Stewart, 1911. Sokotra 33 according to Drude, demic to Australia those plants which are 1896. also found in New Zealand and in Polynesia, Balkans 26 calculated from Ha­ there remain, nevertheless, 7,501 species, or yek and Markgraf, 84.8 per cent, which are limited to conti­ 192 7-1933, and from Turrill, nental Australia and Tasmania. Obviously 1929. the reduction in number is a minor one. Iceland o calcu lated from Os­ An appreciation of the manner of the dis­ tenfeld and Gronrved, 1934. tribution of endemic species among the larger plant groups of the Australian flora For statements about sma ller islands, see O. D ru de, loco cit. will be gained from Table 2, which shows that among the Angiospermae, at least, the it shows; and we may assume that, other con­ ·endemic species are quite equally distributed ditions being equal, the percentage of its among the three groups into which the angio­ endemism would enable us to measure the sperms are divided, with 89.3 per cent for length of its period of isolation. Scandinavia, the Sympetalae, 90.2 per cent for the Dialy­ for example, has very few endemics (and peralae, and 92.9 per cent for the Monoch­ these are "weak" endemics in the systematic lamydeae. This high riumber of endemics is sense) because there was not enough time for not shared by the Monocotyledones, among it to be overgrown with flowering plants be- which only 79.7 per cent of the species are

TABLE 2 DISTRIBUTION OF ENDEMIC SPECIES AMONG THE LARGER PLANT GROUPS OF THE AUSTRALIAN FLORA

PERCENTAGE OF PERCENTAGE OF TOTAL NUMBER WHOLE FLORA NUMBER OF TOTAL NUMBER OF SPECIES REPRESENTED ENDEMIC SPECIES OF ENDEMIC PLANT GROUP IN GROUP BY GROUP IN GROUP SPECIES IN GROUP Pteridophytae 244 2.7 124 50.8 Gymnospermae 43 0.5 43 100.0 Monochlamydeae 1,130 12.8 1,050 92.9 D ialypetalae 3,64 1 41.2 3,286 90.2 Syrnpetalae 2,229 25.2 1,991 89.3 Monocotyledones 1,555 19.6 1,240 79.7 Totals 8,842 100.0 . 7,734 87.5 of the total number Antiquity of the Angiosperms-SUESSENGUTH 291 endemic. Relatively speaking, they show the to the Australian flora as it is now? Above smallest number of endemics among the all, did all of the many families of the Aus-' angiosperms, which is rather an interesting tralian flora exist then as they do now? These fact, inasmuch as in the flora of European questions are not easy to answer, and before countries a certain parallel can be found for investigating them it will be useful to give a these values (Schmidt, 1945). rather detailed account of the Australian flora The smaller percentage of endemics among itself. the Pteridophytae may be explained by the fact that the ferns are more readily dissem­ To begin with, it must be accepted that inated over greater distances by means of the evolution of the families of the flowering their spores. It may be that in this manner plants had begun in times earlier than those many species of ferns immigrated into Aus­ of the Upper Cretaceous period, for only a tralia, or emigrated from it, after its geo­ very few endemic families are found in Aus­ graphical isolation had begun. The same tralia, and these have only a very few species suppositionis valid for many species of the in them . These families are the Akaniaceae, Gramineae and for the Cyperaceae among Balanopsidaceae (which also appears in New the Monocotyledones. Guinea ), Brunoniaceae, Byblidaceae, Cepha­ When we consider the great number of Iotaceae, and Tremandraceae. These are the endemics present in the Australian flora, we only families that have developed endemically are tempted to jump to the important con­ in Australia since the Upper Cretaceous pe­ clusion that, in later times, only an inconsid­ riod, although they might possibly have been erable migration of plants took place into preserved in Australia from times even more Australia from abroad. If there had been any ancient than the Upper Cretaceous; since that considerable degree of migration, we should period there has not been time enough for be able to-find the species of plants now liv­ a further evolution of families. From this ing in Australia spread over other continents evidence we can conclude that it is very likely as' well, and especially over Malaysia. Ac­ that the primitive ancestral types of the other, tually, however, they are not so widely dis­ much larger, families of the Australian flora tributed, but are confined as endemics to existed during the Upper Cretaceous period. Australia. The evidence is such that we may If they had immigrated into Australia after safely conclude, therefore, that a pronounced Upper Cretaceous time-which is a possibil­ development of species took place on the ity we naturally have to take into considera­ Australian continent after the geographical tion-then they ought to be found in other isolation had begun. But, if few species have parts of the world as well. We shall learn migrated into Australia from abroad 'over later in detail how far this is true. But, in such a very long time (except in the cases any event, we must not assume that the six of the Pteridophytae, which show the fewest endemic Australian families also existed, at endemics) , is it not probable that all of the one time, in other parts of the world, only types which evolved into endemic species to die out later in those places, so that now were already in existence before the period they are native to Australia alone. We must of geographical isolation, that is, during the be cautious with this kind of conclusion, a Upper Cretaceous period? Did the numerous lesson which has been made obvious by the representatives of the characteristic families example of the mammals of New Zealand, of the Australian .flora already flourish in referred to by Diels (1897) in his work on those ancient times? Was the Australian flora the vegetative biology of New Zealand. We of those days similar, at least in its families, shall return to-this matter later in this paper. 292 PACIFIC SCIENCE, Vol. IV, October, 1950

According to A. Engler (1882), 425 of in column 1 of Table 3. These figures are the 1,393 genera of the Australian flora­ converted, in column 2, into percentage val­ that is, 30.5 per cent-c-are endemic. This is ues which can be compared with the figures a statement of great importance, for from it for the same plant groups in Australia we learn that the time interval between the (column 3 ). Upper Cretaceous period and the present ~n several of these groups-the Pterido­ time has been sufficient for the creation of phytae, the Gymnospermae, and the Mono­ a great number of new genera-almost a chlamydeae--the percentage values for Aus­ third of the genera found in Australia-or tralia do not differ much from those of central for the conservation in Australia of a part of Europe. In Australia the Monocotyledones them while in other continents they have be- and the Sympetalae appear somewhat less . come extinct. It has already been stated that frequently than they do in Europe; while the the same suppositions are valid for 90 per Dialypetales are found somewhat more fre­ cent of all the Australian species. In short, quently in Australia than in central Europe. the period from Upper Cretaceous time until In R. Mansfeld's catalogue of ferns and the present has been long enough to create, flowering plants ( 1940) , the figures given or, exclusively, to conserve, 30.5 per cent of for that part of central Europe included in the genera and at least 90 per cent of the Germany, Austria , Bohemia, and Moravia are species of the Australian flora. On the other presented in Table 4. hand , it has not been long enough a time to In northern Europe the Monocotyledones permit the creation, or the exclusive conser­ are even more plentiful. In England they .vation, of very many of the families of the form 25.3 per cent of the flora (Druce, Australian flora, particularly of the larger 1932 ) ; in Iceland and the Faroes, 30.8 per families. cent (Ostenfeld, 1934 ); in Greenland, 31.2 It is interesting, for the sake of compari­ per cent : (Osrenfeld, 1926); in Novaya sons not unimportant ro:the arrival at a con­ Zemlya, 33.3 per cent ( Eksram, 1897) ; and clusion, to see how the species of the larger in Spitzbergen, 31.2 per cent (Nathorst, groups are distributed in other parts ofthe 1883). In Portugal the Monocotyledones world. According to Hegi's Flora (1906­ form 20.3 per Cent of the flora (Palhinha, 1931 ), the larger groups of plants are repre­ 1939 ); in Italy , 18.6 per cent ( Buscalione sented in central Europe-e-Germany, Austria, and Muscatello, 1911-1913 ); in the Bal­ and Switzerland-by the numbers presented kans, 16.3 per cent (calculated from Hayek and Markgraf, 1927-1933 ); and in the ter­ TABLE 3 ritory of the Aegaean islands, 17.5 per cent REPRESENTATION OF ENDEMIC SPECIES AMONG (calculated from Rechinger, 1943). It be­ THE LARGER PLANT GROUPS OF CENTRAL comes apparent, then , that the number of. EUROPE AND AUSTRALIA species . of Monocotyledones is greater in COM· PERCENT ­ PARABLE northern Europe than it is in southern AGE OF F IGURES NUMBER "HIGH ER FOR Europe. PLANT GROU P OF SP EC IES * FLORA " AUSTRALIA With the Sympetalae quite the opposite Pteridophytae 74 2.3 2.8 representation is found: the northern coun­ Gymnosperrnae 11 0.3 0.5 tries have fewer of these, the southern coun­ Monochlamydeae 355 11.2 12.8 Dialypetalae . 1,043 32.9 4 1.2 tries have more : Spitzbergen has 13.5 per Symp etalae 1,042 32.8 25.2 cent; Novaya Zemlya, 16.5 percent; Green­ Monocotyledones 648 2004 17.6 land, 21.1 per cent; the Faroes, 25.5 per • According to Hegi (1906-1931) . cent; England, 25.4 per cent; Germany, 29.5 Antiquity of the Angiosperms-SUESSENGUTH 293

TABLE 4 REPRESENTATION OF ENDEMIC SPECIES AMONG THE LARGER PLANT GROUPS OF CENTRAL . EUROPE, THE A EGEAN ISLANDS, AND PORTUGAL

CENTRAL EUROPE* AEGEAN ISLANDSt PORTUGALt PLANT GROUP N UMBER OF PE RCENTAGEOF NUM BE R OF PE R C E ~ TAG E OF . NUMBER OF P ERCENTAGEOF . SP ECIES HIGHER F LORA SPECIES H I GH ER F LORA SPECIES HIGH ER FL ORA Pteridophyrae 73 2.3 4 1 1.2 51 1.9 Gymnospermae 12 0.4 18 0.5 12 0.4 Monochlamydeae 332 10.4 " 36 8 11.2 305 11.0 D ialypetalae 1,119 35.2 1,154 35.0 994 36.0 Syrnpetalae 998 3 1.4 1,138 34.6 843 30.5 Monocotyledones 645 20.3 574 17.5 557 20.2 Totals 3,179 3,29 3 I 2,762 *According to M ansfeld (1940 ). The unimportant differences between Hegi's figures and Mans­ feld's may be ascribed to differences in criteria for the recognition of species as well as to the fact that Hegi's Flora included Switzerland and the South Tyrol while Mansfeld's did not. t In his Flora Aegaea, Rechinger ( 1943) covers the territory of the Aegaean islands from Chalki­ dike in the north to Rhodes and Candia in th e south. tThe figures for Portugal are given by Ruy T elles Palhin ha in his Flora de Portugal ( 19 39). per cent; Portugal, 30.5 per cent; Switzer­ with respect to the numbers of their species, land, 30.5 per cent; France, 31.2 per cent; are these: Tyrol ( including South Tyrol) , 33.7 per FAMILY NUMBEROF SPECIES cent; the Aegaean islands, 34.6 per cent; Legurninosae . 1,065 Myrtaceae . 663 Italy, 35.7 per cent; the Balkan countries, Proteaceae . 59i 37.4 per cent. (These calculations are taken Compositae 539 from A. Schmidt, 1944. ) Cyperaceae . 38 0 Gramineae . 34 5 When the figures for the endemics of Aus­ Epacridaceae . 275 tralia are compared with the figures for those Orchidaceae . . 272 parts of Europe which are rich in endemic Euphorbiaceae . 226 Goodeniaceae . 220 plants , the contrasts are even more pro­ Rutaceae .. .. 190 nounced (Table 5). . These eleven families, with a total of 4,372 From these comparisons we learn that en­ species, include more than half of all the demic species of the Sympetalae are much Australian phanerogams, of which there are more numerous in southern Europe than they 8,555 species. It is worth noting how reo are in Austral ia, and that, at the least, the markably the Leguminosae, Myrraceae, Corn­ centers of development of the polyphyletic posirae, and Orchidaceae have developed. ,As Symperalae are not likely to have been lo­ we know, these families are by no means cated in Australia. If they had been, the per­ primitive. In this way the plants of the pres­ centages of representation would have been ent Australian flora give evidence that their reversed. ancestors (related systematically) must have been well-developed at the time of the Upper NUMBERS OF SPECIES IN FAMILIES Cretaceous period and even before. According to Mueller's figures ( 1889), The percentages of the endemic species in which are approximately correct even today, these 11 Australia n families are tabulated as the most important families in Australia, follows: 294 PACIFIC SCIENCE, Vol. IV, October, 1950

PERCENTAGE OF The Chenopodiaceae, with 111 species in ENDEMICSPECIES Australia; the Hallorhagidaceae, with 58 spe­ FAMILY IN AUSTRALIA Proteaceae 100 cies in Australia; and the Santalaceae, with Epacridaceae . 100 43 species in Australia, each has only two Goodeni aceae 99.9 Myrtacea e .. . 98.3 species which extend beyond the continent, ( including all Leptosper­ and the Amaranthaceae, with 100 Austral­ moideae-Chamaelaucieae, all ian species, has five species which extend their Leptospermoi d ea e : Leptospermeae- Calothamni­ range beyond Australia to N ew Zealand and nae, and all Leptospermoi­ Polynesia. Of the smaller families, many have deae-Backhousiinae ) species appearing in territories other than the Rutaceae .. 97.4 Orchidaceae . 94.8 Australian, and are therefore without value Cornpositae . 91.5 in evaluating endemism in Australia: Leguminosae ...... 90.6 ( including all Pap ilionaceae­ NUMBEROF SPECIES Genisteae-Bossiaeinae ) OUTSIDE Euphorbiaceae 88.0 FAMILY IN AUSTRALIA AUSTRALIA Cyperaceae ...... 70.3 Grarnineae . 69.3 N ymphaeaceae. 5 4 or 5 Guttiferae . . . 4 3 Of the larger families, 13 have only one Geraniaceae . . 8 5 species which has extended its range beyond Convolvulaceae 70 33 Lythraceae . . . 19 12 Australia to other countries; these families On agraceae .. 5 3 are listed here, together with the number of Rhizophoraceae 7 7 species in each family which are found only Cucurbitace ae . 28 13 H ydrocharitaceae 9 7 in Australia : Lemnaceae . .. . 6 6 NUMBEROF N ajadaceae } SPECIES Potamogetonaceae . 36 20 FAMILY IN AUSTRAL IA Aponogetonaceae Myoporaceae 76 Haemodoraceae 66 Many others of the smaller families might be Restionaceae 93 D illeniaceae . 95 added to this list, to support this contention. Saxifragaceae 35 Magnoliaceae 18 T remandraceae 17 THE PHYLOGENETIC AGE OF Annonaceae .. 16 SYST EMA TIC GROUPS Stackhousiaceae 13 Coniferae . . 29 Let us turn now to another question which Cycadaceae . . 14 is more easily answered: Do there exist, Casuarinaceae . 24 Pirtosporaceae . 40 among the families of Australian plants, any TABLE 5 COMPARISON OF PERCENTAGES OF ENDEMIC SPECIES IN THE HIGHER PLANT GROUPS OF AUSTRALIA WITH THOSE OF SOUTHERN EUROPE

PLANT GROUP AUSTRALIA ITALY THE BALKANS SARDINIA Pteridophytae 1.6 0.5 0.0 4.2 Gymnospermae 0.5 0.0 0.2 0.0 Monochlamydeae 13.6 2.0 11.8 O.U Dialypetalae 42.5 35.1 28.8 29.8 Syrripetalae 25.8 56.4 49.9 55.3 Monocotyledones 16.0 5.9 9.3 10.6 Antiquity of the Angiosperms-SUESSENGUTH 295 of the .larger systematic groups which are of all of the other tribes of Papilionatae are far very ancient age? The criteria for the recog­ less numerous. nition of phylogeneticall y old and new char­ In summary, as far as the Leguminosae acteristics have been treated at length in my are concerned, it can be said that the Mimo­ book (Suessenguth, 1938: 19 et seq.). soideae and the Papilionatae show the most Let us consider the Leguminosae first. pronounced development of primitive species Among the members of the subfamily Mimo- . in Australia. soideae the most important genus in Aus­ Among the tribes of the Labiatae, the Pro­ tralia is A cacia, which has more than 300 stanthereae are by far the most numerous ' in Australian species. The rest of the genera of Australia, having 89 species compared with the Mimosoideae in Australia number only 31 for all of the other tribes. If we study the about 23 species. subfamilies of the Labiatae, as they are con­ Is the genus A cacia, then, an old or a new sidered by Briquet (in Engler and Prantl, genus among the Mimosoideae? If we accept 1897 ), we come to the conclusion that the the general phylogenetic principle that free most primitive species must be those having stamens are more primitive than fused sta­ no gynobasic pistil-the members of the tribes mens, and if we agree that the group of spe­ Ajugoideae and Prosranrheroideae. When we cies with numerous stamens (now classified investigate these two groups we learn that in the tribes Ingeae and Acacieae) is older the Prosrantheroideae have ovules with endo­ than the group of species which have 10 or sperm, while the Ajugoideae do not show fewer than 10 stamens (now classified in the any endosperm in their seeds':"-evidence that tribes Eumimoseae, Adenanthereae, Piprade­ the Prostanth eroideae are the most primitive nieae, and Parkieae ), then we must conclude of the Labiatae. These primitive Prosranthe­ that the species of the tribe Acacieae, with roideae are confined exclusively to Australia. their free stamens, are more primitive than Among the Myrtaceae we think that the are the more or less synantherous species of subfamily having dry fruits-the Leptosper­ the tribes Ingeae and Parkieae. The Acacieae, moideae-s-is the most primitive. In Australia with the genus Acacia- in which the fila­ there are about 596 of these species with dry ments are free or only grown together to fruits, compared with only 41 species of form a short ring-are undoubtedly the most Myrtoideae which bear berries. In Australia primitive of the Mimosoideae, and apparently then , the older subfamily has about 14.5 the plants of the genus Acacia are the most times as. many species as does the younger primitive of all of the Acacieae. Therefore, one. The proportion of Leptospermoideae to Australia shows the greatest number of old~st Myrtoideae in the rest of the world is quite types among the Mimosoideae. different: there are 678 species of Leptosper­ moideae and 1,932 species of Myrtoideae, a The subfamily Papilionarae presents much ratio of 1:2.85. the same evidence. The most primitive tribes Among the Rutaceae the most primitive of the Papilionarae are those which have free species are placed in the subfamilies which stamens-the Sophoreae and the Podalyrieae. are inclined to apocarpy rather than in the The Sophoreae generally have pinnate leaves, subfamilies with united carpels (e.g., the while the Podalyrieae have simple or digi­ Flindersioideae, Sparhelioideae, Toddalioi­ tate, rarely pinnate , leaves. Because of their deae) or in those with bacciform fruits (e.g., simple leaves, the Podalyrieae may be con­ the Aurantioideae). The Rutoideae show a sidered the more primitive tribe. The Podaly­ tendency to apocarpy, and among their sub­ rieae number 350 species in Australia, while families several groups have developed: (a) 296 PACIFIC SCIENCE, Vol. IV, October, 1950 those with dorsivenrral flowers (the American flora, the Acrinidioideae being found in Cusparieae); (b ) those without endosperms ] apan , China, Manchuria, and the Himalayas, (the African Diosrneae); (c) those having her­ while the Saurauioideae are found in tropical baceous or suffruticose habit o( the Ruteae of Asia and America. All of the Australian spe­ the northern temperate zone); and (d) those cies of the Dilleniaceae belong to the more species with doubly digitate leaves (the Diet­ primitive subfamily, the Dillenioideae. One yolomeae of tropical South America ). The of the tribes of the Dillenioideae, the Aero­ remaining tribes .of the Rutoideae are the tremeae, is found outside of Australia, in Xanthoxyleae and the Boronieae. The Xan­ India and Ceylon; but this is a less prim itive thoxyleae have a tendency to produce uni­ tribe than is the Australian one, showing a sexual flowers, a characteristic which, for this number of derived characteristics, such as group, is regarded as a derived feature. The united carpels, a bushy habit, and pinnati­ most primitive types of the Rutaceae, then, partite leaves. The other tribes of Dillenioi­ are probably the Boronieae, and these types deae-the Tetracereae, Hibbertieae, and Dil­ are limited to Australia and N ew Caledonia. lenieae-which also have representatives in In Australia there are about 143 species of southern Asia and tropical America, are not the Boronieae and 26 species of the Xan­ well enough known at present for a decision thoxyleae, but only 9 species of the Flinder­ concerning the degree of their primitiveness sieae, 7 of the Auranrieae, and 3 of the Tod­ or evolution. dalieae. These figures, which could readily Among the Restionaceae, the Diplanthe­ be supplemented with more evidence, show reae have dithecic anthers and the Haplan­ clearly enough that the prim itive Rutaceae thereae have monothecic anth ers. Naturally, appear nowhere as plentifully as they do in those genera with dithecic anthers are 0 re­ Australia. garded as being the more primitive. They In considering the Loranthaceae, Engler appear only in southwestern Australia; the ( 1894) mentions the Loranthoideae as the genus Anarthria, with free anthers and a tri­ first state of development and the Viscoideae locular ovary, is the most primitive of them as the secondary one. The most primitive all. Among the Haplanthereae no differen­ of the Loranthoideae are non-parasitic trees tiation can be made upon phylogenetic char­ which have no berries-members of the acteristics, for here the Australian and the genera N uytsia and Gaiadendron. The species African species share some characteristics. of Nuytsia, with dry false fruits, are consid­ Nevertheless, the most primitive representa­ ered more primitive than the species of Gaia­ tives do not appear outside of Australia. dendron, with drupe-like false fruits. The The Centrolepidaceae show quite a similar species of Nuytsia are found only in western relationship: among them, too, the species Australia. Of the four species of Gaiaden­ with dithecic anthers are also the more dron, three are found in the Andes from Peru primitive ones. They are represented by the to Colombia, and one is found in eastern genus ]un cella in southern Australia and in Australia. Tasmania. Among the Dilleniaceae, two of the sub­ The Goodeniaceae, although not com­ families, the Actinidioideae and the Sauraui­ pletely endemic, 0 have most of their repre­ oideae, show a derived feature in the fusion sentation in Australia. The most primitive of their carpels and furth er development in genus in the family is Calogyne, which has that the Actinidioideae and most of the Sau­ bifid or trifid pistils. Two species of the rauioideae bear real berries. N either of these genus are found in Australia and the third in subfamilies is represented in ·the Australian south China. Antiquity of th e An giosperms-SUESSENGUTH 297

Of the seven tribes ofrhe family Protea­ the most pnrrunve are those which possess ceae, the Persoonieae is the most primitive, as no corona, as, for example, the Gymnolaima Engler has stated in his Natiirliche Pflanzen­ of Kiliman jaro, Africa, the Phyllanthera of familien ( II I/l : 127). The Persoonieae are Java, and the Pentam era of Sumatra. The found in Australia, Tasmania, New Caledo­ Asclepiadaceae are generally more highl y nia, and, to a lesser extent, in New Zealand. evolved than are the Apocynaceae, but they One species of Brebeiurn appears at the Cape do not originate in Australia. of Good Hope, but has developed farther Let us now go on to consider a rather com­ than its relatives in Australia, as is proved by plicated group, the Cyperaceae. Here the its floral axis which shows a cyathiform ex­ species of Scirpoideae, with their hermaph ro­ crescence at the base. In any event, the most ditic flowers, are more prim itive than are the primitive representatives of the Proteaceae Caricoideae, the flowers of which are rarely are almost completely limited to Australia. hermaphroditic. The nature of the axes in In the, Santalaceae, the members of the the inflorescences of the Caricoideae also tribe Antholobeae, with their superior ovary, proves to be a derived feature. Among the are considered primitive.. The genus Antho­ Scirpoideae there is a tribe, the Hypolyrreae, lobus is native to Australia. A closely related whose members have bracted flowers; and genus, Exocarpus, is found in Australia, Nor­ the ttansverse arrangement of these bracts folk Island , the Malaysian islands, India, (as occurs among the Hypolytrinae ) might Madagascar, and Hawaii, although most of be a more prim itive characteristic than is the its species are native to Australia. The possession of one or two median bracts (as genus Cbam pereia is found in Malacca and occurs among the Lipocarph inae ) . Among the Malaysian archipelago. It can be con­ the Hypolytrinae the genus Hypolytrum , chided, therefore, that the Santalaceae are of whose species show free bracts, is most prim­ Australian-Malaysian-e-rhar is to say, of post­ itive. A species of H ypolytrttm , H . latifolium Gondwanesian-s-origin. 1. C. Rich., is found in Queensland, but it is The most prim itive Apocynaceae are those also found in south Asia, Africa, America, 'in which the stamens are not tightly con­ and Polynesia. Two species of Lipocarpha nected with their stigma heads. These are also have an extensive range. The genus the Plumieroideae, especially a subgroup of Hypolytrum has its representatives in the them, the Pleiocarpeae, which have apocar­ tropical and subtropical ranges of both hemis­ pic ovaries, pistils split at the base, and more pheres. All of this evidence would seem to than two carpels. Among them are two show that the oldest living types of Cypera­ genera with the primitive arrangement of ceae-which is considered a rather "modern" alternate leaves: N otonerium Benth., an eri­ family- have their native habitat in the coid bush growing in southern Australia, and tropics, but by no means in Australia. Lepinia Decne., a tall tree found in Tahiti. In contrast to this, the oldest genus of the The most primitive species of the Apocyna­ Scirpoideae-Cyperinae, the genus Carpha R. ceae, therefore, are Australian-Pacific in their Br.-without disk, but with setaceous involu­ origin. cre, six setae, and a three-cleft pistil-is repre­ It might be noted in passing that this same sented by one species from Australia and conclusion cannot be drawn for the Asclepia­ New Zealand and by another in extra-tropical daceae, Here the Periploceae are the most Andean South America. primitive forms, judging by their tetrad pol­ In the large subfamily of the Caricoideae, len, the translators of which have no reti­ the Rhynchosporeae are the most primitive, nacula; and of these primitive Periploceae, inasmuch as, in most cases, they have three 298 PACIFIC SCIENCE, Vol. IV, October, 1950

anthers and an involucre. If we except the together to form a circular ring (except for genus Oreobolus-which is somewhat in a the slightly more developed Gomphogyninae special position because of its circum-Pacific and the Zanoninae, which have unilocular distribution and its single, terminal, one­ ovaries). The Fevillinae have trilocular ova­ flowered false spikelets (Suessenguth, 1942 ) ries, and, of course, are more highly evolved -·we note that genera with alternate or very than are the Fevilleae. All of these plants slightly distichous bracreal scales are more are native to tropical America, Brazil, and the primitive than are those with distichous West Indies. scales. Among these genera the most prim­ Among the 'Orchidaceae, the more prim­ itive are those which have three style branches itive species (the Diandrae-Apostasiinae) are and an involucre; of these genera three are not found in Australia. Species of the genus especially worthy of consideration here : N euwiedia, which have three fertile stamens, Lepidosperma, with nine-tenths of its appear in Malacca and the Malaysian archi­ species found in Australia, two in New pelago; those of the genus A postasia, which Zealand, and two in tropical east Asia have two fertile stamens, appear in the East T ricostularia, with five-sevenths of its Indies, the Malaysian archipelago, and trop­ species found in Australia, one in Bor- ical Australia. From this it is evident that neo,and one in Ceylon . the oldest types of the whole family belong Decalepis, with one species found at to the tropics and are found today in terri­ the Cape of Good Hope in South Africa. tories lying north of Australia. From this evidence we see that-again if we The distribution of the Piperaceae (as, in­ except the genus Oreobolus, which is of old­ deed, of many another smaller family), leads Pacific origin and which is difficult to classify us to expect a tropical origin for them. -most of the oldest types of the Caricoideae Of the Rubiaceae, the more primitive sub­ are to .be found in Australia, while the most family is that of the Cinchonoideae, whose primitive species of the 'whole family 'of the species have many seeds in each locule of Cyperaceae are found in the tropics: the ovary. Among the Cinchonoideae, the In order to complete the picture we should Cinchoninae are more primitive because of consider some of the families, the origin of their dry fruits. More highly developed which cannot be traced to Australia. groups, like the Condamineae and, to some Of the Anacardiaceae the most primitive extent, the Rondeletieae, have radiate flowers genus is Buchanania, native to tropical Asia, which are single or in panicles (but not in especially to the Malaysian territory, and to clusters), apterous seeds, whole or bipartite northern Australia. The most primitive spe­ stipules, and the habit of trees or shrubs. The cies of Buchanania have four to six free car­ Rondeletieae, however, show imbricated or pels, of which one is fertile. contorted vernarion of the corolla, and the The Compositae are impossible to trace contorted vernation, at least, is a derived to their origin, or to differentiate into their feature. Among the Condamineae the most most primitive groups, even if we exclude primitive species are those in which the se­ from consideration the tribes which are ob­ pals are of equal size and in which the petals viously derived, like the Liguliflorae and the are simply valvare and not reduplicatively Mutisieae. valvate. The family Cucurbitaceae is divided into The simply valvate species are placed in the Fevilleae and the Fevillinae. The Fevil­ the genera Condaminea (found in Andean leae are the more primitive, having five free South America ), Chimarrhis (found in An­ anthers and loculamenta which are not grown dean South America and in the Antilles ), Antiquity of the Angiosperms-SUESSENGUTH 299

Rustia (from Central America to Brazil ) , Most of these groups cannot be considered and T resanthera ( in Venezuela and in the primitive in the general phylogenetic sense- West Indies ). N one of these genera even so . as, for example, these nine of the 14 families : much as app ears in Australia. The redupli­ the Labiatae, Papil ionarae, Mimosoideae, catively valvate species of the Condamineae Restionaceae, Centrolepidaceae, Apocynaceae, are placed in the genera Portlandia (found Myrtaceae, Goodeniaceae, and the subfamily in the W est Indies and in Mexico ) , Isidorea Caricoideae of the family Cyperaceae. This (in Haiti and Cuba), Bikkia (from the Pacific would mean that the angiosperms which islands, New Caledonia, and Malacca ) , and . have developed in Australia since the Upper Morrierina (found in New Caledonia). Cretaceous period cannot be traced back to Of the Rondeletieae, the simpler species, the very earliest groups of angiosperms. These in which there is no contorted vernation of ancient groups must have developed in much the corolla and .in which the petals are not earlier times than the Upper Cretaceous. It evolved into showy organs, are placed in the is not likely that the nine families have genera R hachicallis ( found in the Antilles ), spread from Australia to other parts of the Bathysa (found in Brazil ) , and R ondeletia world after Australia's geographic isolation (found in the Antilles, Central America, and began and that the original primitive species the northern parts of South America ). have been conserved in Australia ever since All of this evidence proves that the Ru­ that time. On the contrary, it is much more biaceae did not originate in Australia, but probable that the primitive ancestral types primarily in the tropical regions of Central also existed in other parts of the world even and South America and in the West Indies, before the Upper Cretaceous period and that and only. in lesser part in the regions of they have died out there since that time, N ew Caledo nia and the Moluccas. just as most of the Marsupialia and the The most primitive species of the Valeri­ Monotremata have died out in parts of the anaceae rappear on the Asiatic mainland: world outside of Australia. Finally, it should Species of Nardostachys, with four stamens be remembered that it is also possible that and the clearly five-parted edge of the calyx, the Australian angiosperms of today might are found in the central Himalayas; species have had ancestors originating in other of Patrinia, with four. stamens, extend west­ continents before the beginning of Austra­ ward from Japan through central Asia to lia's geographic isolation. the Ural mountains and northward into It is likely, too, that many of the families Arctic territory. of the Australian plants have migrated into Now, in recapitulation, let us list all the Australia in times later than the Upper larger systematic groups of the angiosperms, Cretaceous period, especially those families the most primitive types of which are found found now in northern, tropical Queensland. in Australia: Labiatae, Mimosoideae, Papil­ All of this would mean that parent types ionatae, Myrtaceae (sub-family Leptosperrn­ of most of the derived families of Angie­ oideae ), Rutaceae, Santalaceae, Apocynaceae, spermae were already in existence before the Goodeniaceae, Proteaceae, Restionaceae, Cen­ Upper Cretaceous period, and that the de­ trolepidaceae, Loranthaceae; Dilleniaceae, velopment of the main branches of the Cyperaceae (subfamily Caricoideae ). This Angiospermae took place in even earlier summary and all of the evidence leading up times. Fossil discoveries lead us to suppose to it are of great importance for the proper that a strong and rapid development of evaluation of many of the problems and angiosperms has taken place since the Upper questions in the science of plant geography. Cretaceous period. Investigations of the 300 PACIFIC SCIENCE , Vol. IV, October, 1950

Australian flora do not confirm this impres­ in very recent times. The islands of New sion, ho~ever. Rather, they support the sup­ Zealand have never been connected with position that, in most of its essentials, the land areas inhabited by mammals, and until development of flowering plants goes back now no fossil relics of mammals have been .to even earlier times-to the period of Low­ found there; it is very unlikely, therefore, er Cretaceous formations, possibly even as that mammals did live in New Zealand at far back as the Jurassic period. Unfortunate­ one time but have died out there since. ly there are not man y fossil evidences of Now, if New Zealand has never been angiosperms preserved from Jurassic times, connected with land areas populated by and very few of these can be identified with mamm als, where did its flora come from? certainty. When the Cenozoic era began, And does this flora show still more prim itive the chief development of the angiosperms features than does that of Australia? must already have been finished. Particularly Diels ( 1897 ) has entered into a full dis­ primitive types might have been preserved cussion of these questions in his work on until then, of course, but there is no definite the Vegetationsbiologie von N euseeland. He fossil evidence as yet of this possibility." assumes that New Zealand has not been submerged since the middle of the Mesozoic COMPARISON OF AU ST RALI A WITH era. According to Hutton (cited by Diels, NEW ZEALAND 1897), New Zealand was connected with an Antarctic continent which existed during In this connection it might be significant the Lower Cretaceous period, toward the end to draw a parallel by investigating a land of the Mesozoic era. Diels thinks it is prob- ­ area near Australia and which has been able that, even during the later Triassic isolated from other continents for even a period in early Mesozoic time, the Austral longer time than Australia. Such a territory circumpolar lands were closely related to is New Zealand. No fossil mammals were each other, so that there was a geneticalcon­ found there, while, as we' know, primitive nection among the mountain floras of Tas­ mammals had entered Australia from south­ mania, southern Australia, the southernmost ern iAsia. In the event that some of these part of South America, and an Antarctic / mammals originated in Australia itself-a continent which probably was more tem­ ,f rather untenable supposition-they must perate in its climate in those early times have wandered out of Australia over land than it is now. This interrelationship of , ' bridges toward the north, eventually to reach floras would find its parallel in the Arctic, ", Europe and North America. In N ew Zea­ Alpine, and Altaic floras of the northern land, on the other hand; only a small rat hemisphere. has been found to represent the mammals, In his paper, Diels cites evidence to sup­ and this rat was probably imported by man port this supposition of the connection of the Antarctic and Austral land masses. In those 2Erdtman in 1948 published reports .in Grana Palynologica, that pollen had been found in the times the Antarctic continent must have been black lias formations of southern Sweden. Th e much larger than it is today, free from ice pollen .appears similar to that of Eucommia in its, northern parts, and certainly warmer species ( Eucommia is a genus in China, closely related to the Ulmaceae ) and it is not likely to during the Triassic period. In addition to have been deri ved from Gymnosperm ae. Inas­ Diers evidence, we can find ·further testimony much as the .black lias of Sweden is a Lower in comparative zoology and in plant geo­ Jurassic form ation, these pollen finds may offer graphy. Fossil relics of marsupial groups some evidence of the early developm ent of the angiosperms. now limited to Australia-species of the Antiquity .of the Angiosperms-SUESSENGUTH 301

Abderitidae, and of the Sparassodontidae, ographical reasons-that the connection from which are related to the Dasyuridae-have Chile to east Australia could have been been discovered in Eocene deposits in Pata­ formed in a straight line. On the contrary, gonia; and Zittel ( 1895) concluded that "it this line passed farther south through an is an undeniable paleontologic fact that in Antarctic continent, which at that time was those times both regions were in mutual overgrown with plants. exchange or at least drew from the same According to Hutton and Wallace (cited sources." But only the western isle of the by Diels, 1897 ) , a Melanesian continent former Australian archipelago (the West connecting N ew Caledonia, Lord Howe Is­ Australia of today) participated in this ex­ land, Norfolk Island, and N ew Zealand, and change. The eastern islands, particularly New reaching as far north as the present north Zealand, did not-because they were not con­ Queensland, might well ·have existed in the nected either with western Australia or with Eocene epoch. There was no connection, Patagonia. The most primitive species of mar­ however, between this continent and western supials-species of M yrmecobius and Per­ Australia. In Miocene times west Australia agalea-are endemic to West Australia, and and east Australia were connected, but the it is a very significant fact that there is no west A~ stralian species never reached the fossil evidence to prove that marsupials exist­ tropics, and, therefore, did not get to Ne w ed in eastern Australia at any time before Zealand. the late Cenozoic era, that is to say, before From these few considerations we learn the central Australian sea had retreated (Zit­ that the situation in N ew Zealand is quite tel, 1895: · 294) . All this is evidence that different from that in Australia. Ne w Zea­ there must have been connecting land links land was closely related to the Antarctic between P,atagonia and western Australia. continent and to a Melanesian continent, but We find a very interesting parallel in the we cannot expect to find there the primitive distribution of two sections of .the genus species of the Australian continent. Th e dif­ Discaria of the family Rhamn aceae. The ferent character of the flora of Ne w Zealand section N oto pbaena (Miers) Suessenguth , in is proof of this expectation. In their Manual its present range, connects Chile and New of th e New Z ealand Flora, Cheeseman and Zealand. The section Eudiscaria .Stapf ap­ Oliver (1 925 ) list 1,59 1 species of vascular pears in the Argentine countries (that is, plants, with 1,415 phanerogams and 156 vas­ in the countries east of Chile ) , and -in Tas­ cular cryprogams, among all of which are mania, Victoria, and New South Wales. This 1,143 endemic species-e-Fz.S per cent-and distribution can be explained only by as­ 24 endemic genera. While Mueller's cata­ suming two land bridges leading through logue counts 592 species of Proteaceae in an Antarctic continent-one connecting Chile Australia, only two can be listed for New and New Zealand, in a strip slightly arched Zealand. The large Australian genera .of towards the south ; another, farther south Eucalyptus and Acacia are compl etely missing than the first, leading from eastern Pata­ in New Zealand. The floristic connection of gonia through the Antarctic contine nt to New Zealand with Australia is formed by Tasmania and southeast Australia. certain of the Myrtaceae (the genus M etro­ It is my opinion that all sketches of these sideros ) and by the family Epacridaceae. hypothetical land bridges which have been According to Grisebach (1872 : II, 633) published are not quite correct, for it is im­ these are the New Zealand families or groups possible-for phytogenerical as well as pale- which are represented by the most species: 302 PACIFIC SCIENCE, Vol. IV, October, 1950

PE RCE N TAGE OF VASCULAR PL AN TS sia and New Guinea apparently occurred N UM BE R OF REPRESENTED PLAN T .GROUP SPECIES BY GROU P earlier than did that of Australia. The great Compositae . . 22 1 14.1 number of endemics in New Zealand's flora, Ferns . 138 8.8 then , can be.attributed not to the conserva­ Cyperaceae . .. . 119 7.6 tion of primitive species, but rather to the Scrophulariaceae . 113 7.2 Gramineae .. . 113 7.2 formation of new ones. Um belliferae . . 62 3.9 Because of its temporary connection with Orchidaceae . . 57 3.6 Ranuncu1aceae . 50 3.2 Australia, Melanesia, and the Antarctic con­ Rubiaceae . . . . 47 3.0 tinent, New Zealand cannot give us any Epacridaceae . . . 3 1 2.0 assistance in solving the problem of the origin / --Ql1agraceae ..,,/., - ;31 2.0 of the larger and older Australian flora, so /' Leguminasae :-. . 26------..-<, 1.7 ] uncaceae . ... 25 1.6 important in any estimation of the age of Boraginaceae . . . 25 1.6 the angiosperms. The geologic and biologic The differences between the figures for records left on New Zealand are quite dif­ New Zealand and those for Australia are ferent from those of Australia, and they very striking: Leguminosae, ranking first in can not be traced very far back into geologic Australia, is not amon g even the first lOaf time. the families of New Zealand, and neither is Myrtaceae (in second place in Australia) , ORIGIN O FN EWT YP ES INAUST RALIA Proteaceae (in third place ), or Euphorbi­ How , then , can we explain the appearance aceae (in eighth place ). The Compositae, -of new types in Australia ? Perhaps the fol­ however, have achieved first place in the lowing supposition may be the simplest one. New Zealand flora (they are in fourth place Let us assume that during the Upper in Australia ) , and-the ferns being left out Cretaceous period, or possibly during the of .consideration-the Scrophulariaceae, the Middle Cretaceous period, but in any case a Umbelliferae, the Rubiaceae, and the Ranun­ short time before Australia's geographic iso­ culaceae have entered into the list of fam­ lation began, there existed the plant types ilies with the most species. The Labiatae, in A, B, C, D . .. in Malaysia as well as in contrast, are not to -be found at all in New Australia. Since that time, the Malaysian Zealand; they did not reach these islands types have evolved into types AI, Bt, CI, D l from Malaysia and Australia. , . .. that is, into new and different species or Of the 1,59 1 plant species found in N ew genera. The Australian types, however, have 2 2 2 Zealand, 428 species are not endemic. Among developed into types A2, B , C , D ••• into these, 366 are related to the Australian flora, different species or genera from both their and 108 are related to that of South America. -parent types, A, B, C, D and the col- l The genera with the greatest number of lateral types AI, B , CI, D' developing species in N ew Zealand are these : V ero­ in Malaysia. It is conceivable that the en­ nica, 84 species; Carex, 54; Celmisia; 43; demic plants in Australia have evolved in Coprosma, 40; Ranunculus, 38; Olearia, 35; this manner. Senecio, 30; Epilobium, 28; Poa, 25; My­ Of course it is also possible that this osotis, 23; and H ymenophyllum, 20. development of -endemics could have taken It cannot be said, however, that New place in later times, without it being neces­ Zealand 's flora is more primitive or has more sary for us to conclude that all species of primitive species than does the flora of Aus­ the parent series A, B, C, D.. . must have tralia,even though its isolation from Malay- been distributed throughout Malaysia and Antiquity of the Angiosperms-SUESSENGUTH 303

Australia during the Upper Cretaceous period. these developments took place at almost the Many of them may have migrated to Aus­ same time (in the Upper Cretaceous period) tralia in later times, after the separation, and and if we agree that the possibility of sub-.. 2 may have evolved there into types A , B2, sequent immigrations into Australia must 2 2 C , D •• •• But if this is true of some plants, also be taken into account, then we would it is not likely to be true of the Legumi­ do well to remember that in their manner nosae: It is probable that the primitive of distribution angiosperms and mammals species of the Mimosoideae, the Papilion­ differ markedly in at least this major point: ateae, and of some of the other families listed Flowering plants are much more able to. above (p. 295) immigrated into Australia cross the sea-if only by means of driftwood a long time ago and have survived there un­ -than are mammals. This would seem to be changed, remaining generally identical with an assertion that could hardly be contested. their ancestors of the Upper Cretaceous And yet it is a strange fact that greater period, wherever these ancestors may have numbers of primitive plant types have not grown. been preserved. They became extinct, while For some of the other plant groups, it is the primitive types of animals- the Mar­ 2 2 2 2 possible that their species A , B , C , D .. . supialia and Monotremata-c--conrinued to live. may. have risen in different epochs. A\ B\ These animals link the mammals with the 1 2 2 2 C1, D •.. in Malaysia, and A , B , C , reptiles, but even at the present time no 2 D ••• in Australia, continued to live, while plants are known in Australia which link their common ancestors A, B, C, D... died the angiosperms with the gymnosperms. The out in both territories. Or, if we assume that botanical systematist will regret this fact, if 1 1 1 A = A, 'B = B, C - C, and so on, or if only because such proof of primitiveness we ta ke ~A2= A, B2 = B, C2 = C, and so would be a much more scientific, and there­ on, we might deal, then, with only two lines fore a more reliable, basis for the taxonomic of development instead of three, and only system. one of them need have changed-either My impression of the rise of Australian en­ the one in Australia, since the beginning of demics has been described with reference to it') isolation, or the line in Malaysia, since its relationship to Malaysia, both because the Australia's separation. In other words, the endemics of this area are more closely related local ancestors of the line A, B, C, D.. . to those of Australia and because of -Aus­ might have died out in one territory arid tralia's former connection with N ew Guinea might have been preserved in the other for (see Behrmann , 1937) . Perhaps these con­ a very long time. Yet this is not very prob­ clusions will seem quite natural to most able a chance inasmuch as most species of readers; nevertheless, I think it would be use­ living things-except for the mussels-gen­ ful to develop further 'conclusions based on erally have not been conserved unchanged certain concrete suppositions. over long periods of geologic time. As has been known for a long time, most In my opinion this line of approach is of the species of the plant families charac­ the most natural way of explaining the teristic of Australia grow in the southwestern problem. It does not relegate the appear­ maritime areas ( Hooker, 1860 ) . Fewer spe­ ance of all the endemic families of the Aus­ cies are found toward the north. According tralian angiosperms to the apocryphal dark- . to his catalogue, Hooker counted 3,600 spe­ ness of antiquity, and yet it does help us .to cies in the southwestern territory, known in understand the rise of the many endemics in his day as Swan River. and King George Australia. If we do not insist that all of Sound, but only 3,000 from the eastern area; 304 PACIFIC SCIENCE, Vol. IV, October, 1950

and only 2,200 from tropical Australia, where firms the conclusions derived from the study the endemic species are fewest in number. of mammals. Sometimes the correspondence Now what is the reason for this distribution? is curiously exact." Unfortunately, Ne wbigin Is it because Malaysia has exerted less of an does not give his proofs of these lapidary influence . upon these southwestern districts sentences, particularly for the plants. Let us, because they are so distant, and because they therefore, find our own proofs in some ex- . are separated from the interior-and there­ amples from the pertinent literature. fore from the northern shores and Malaysia­ A good many of the flowering plants, as, by vast deserts? Is this pronounced isolation for example, the families of the Papaveraceae . the reason why more endemic species have and the Geraniaceae, have migrated along developed and have been preserved in the the ridges of the Andes, from both North southern periphery of Australia than in its America and Central America, far into South other parts ? Or is there a more general rule, America (Vester, 1940 : 162, fig. 78 ). The as yet unexplained, that endemics are devel­ genus Ribes, which also migrated in this man­ oped more generously in southern lands? ner, has been cited for this fact by Newbigin If we compare South Africa with Australia, (1936 ). The same evidence of migration we can count an enormous number of en­ appears to be provided for some of the demics in Cape Colony; and if we compare Primulaceae, with Primula farinosa in the the most southern parts of South America­ Holarctic region and in Andean Patagonia Patagonia and Chile-s-with Australia, we can (Vester, 1940: 154, fig. 40) ; for the Beru­ find there, too, a great number of endemic laceae (ibid., 163, fig. 80) ; for the Empe­ species-l,200 of 1,600 species, according to traceae (ibid., 163, fig. 81); for the Oroban­ Grisebach's early evaluation in his Die Vege­ chaceae (ibid., 164, fig. 86 ); and for the tation der Erde (1872: II, 498 ) . ]uglandaceae (ibid., 176, fig. 153). It is not possible to indicate a preponder­ Fossil discoveries give further evidence ance of endemisms in the most southern part that, to a great extent, some of the plant of India , at least on the basis of the figures groups were forced out of northern areas into reported by Hooker and Thomson in their southern not only by diluvial glaciers, which, Introduction to the Flora Indica (1855 ). coming from the north, destroyed the Ceno­ Newer statistics concerning Indian endemisms zoic flora of central Europe and of central apparently are not yet published. N orth America, but also by hitherto - un­ In Europe, however, the majority of en­ known factors which were effective much demics is found in the southern areas, partic­ farther southward. The Magnoliaceae, for ularly in the Balkans and in Crete (Turrill, instance, at one time must have grown over 1929 ). In this connection, Newbigin (1936) vast parts of the Holarctic region, but today has made these statements about mammals: they are limited to South and Central Amer­ "It has been made abundantly clear that the ica, eastern and southeastern North America, great migratory movements have been from the West Indies, southwestern Asia, Malaysia, the wide land masses of the northern hemis­ eastern Australia, and New Zealand (Vester, phere towards the narrower and discontin­ 1940: 188, fig. 262 ). The ]uglandaceae, too, uous southern ones, and that extinction of have disappeared from large areas of north­ early stocks has been most marked in the western N orth America, from Europe (except Holarctic region, while the survival of mem­ for the Balkans ), and from central Asia (ex­ bers of these is especially characteristic of cept for the Caucasus ) , while in the more some of the southern lands. ... The past and southern regions they continue to survive. present distribution of the higher plants con- Further examples are to be found in the fam- Antiquity of the Angiosperms-SUESSENGUTH 305

ily Ebenaceae, according to Fernald (in '. and Antarctic regions, and it is not to be Vester, 1940: 174), and in the tribe Cyca­ denied that a large part of the "small belt­ deae and the genera Podocarpus and Arau­ like areas" of many families in the whole caria, according to Studt ( 1926). tropic: range may thus have been established It is doubtful whether any instances can in their present ranges (Vester, 1940 : 166 be found to prove a considerable migration et seq., figs. 93-11 3). Nevertheless, it seems of a plant flora in the opposite direction, that as if in Australia, the Andes, Africa, and is, from south to north. It is true, of course, Europe other factors had contributed to force that a'number of plants, reported by Suessen­ a great number of species of plants from the gurh (1942 ), have worked their way north­ north to the south, and in those areas in the ward from the South American Andes, reach­ southern hemisphere this displacement car­ ing as far north as Costa Rica and Mexico. ried the plants even farther south than the However, these migrations have taken place Tropic of Capricorn. only since the Miocene elevation of the Cor­ This phenomenon of displacement from dilleras in Central America, and they are north to south does not need the supposition rather insignificant compared with the major of some mystical power to explain it. In southward migrations. Africa, for example, a northward counter­ There is evidence, however, that north­ displacement of the ancient flora of the north­ eastern Africa has been reached from a north­ ern and middle part of the continent could .eastern direction by species of plants from not happen because it was blocked in that India and western Asia. direction by the broad Tethys sea of the early Although it might be expected that the Tertiary period (Eocene time, and so on) or Mediterranean floral elements might have by the deserts that are its relics. Australia, arrived in central Europe from the south, fol­ to give another example, in post-Tertiary lowingzhe retreating glaciers as they with­ time could not receive plants from any direc­ drew to the north, I do not think this argu­ tion but from the north, because it was ment is tenable, inasmuch as it is quite possi­ only there that Australia was· connected, if ble that representatives of the Mediterranean only temporarily, by land bridges with large flora might have found refuges in the climat­ masses of land, while in the south the Ant­ ically favorable parts of central Europe dur­ arctic continent at a later period was too cold ing the glacial advances. It is much more and too far distant to permit of plant migra- likely that the North American plants of . tion. In the Andes plant distribution is not . Cenozoic time were forced southward by the as easily explained. In the Balkans the lower­ glaciers, and then, after the glaciers had re­ ing of temperatures in the north by the treated, were permitted to return north, to glaciers may have played a part in the south­ recover vast territories of their former areas ward displacement of the plant life, so that of distribution. Nonetheless, these instances numerous types of plants died out in the of northward migration are abundantly sur­ north which continued to live in the south. passed by the notable removal to the south It would be of great interest to investigate of plants in Australia, the Andes, Patagonia; the degree of displacement in still other parts Cape Colony, and the Balkans, in all of which of the world. real displacements to the south have been Many other objections might be raised to demonstrated. During the cold periods of the dispute this claim of the southward displace­ glacial advances, all of the hydro-megatherms ment of plants. The major point of dispute and megatherms should have been concen­ is whether or not this southward displace­ trated towards the tropics from the Arctic ment of certain systematic groups-such as, 306 PACIFIC SCIENCE, Vol. IV, October, 1950

for example, the Magnoliaceae, which are its isolation from Malaysia, were not very said to be analagous in this respect to many likely-or at least were not very plentiful­ animal groups-can be attributed only to the the enormous degree of endemism which lowering of temperatures in the northern re­ Australia now exhibits being evidence against gions during the time of glaciation . any considerable change in later times. It cannot be established with certainty CONCLUSION whether or not the plants of the primitive These investigations have offered statistical genera of the 10 major Australian families evidence that the phylogenetically older types are not only endemic by preservation but are of about 10 large taxonomic groups of the also plants which have originated in Aus­ higher plants are found, either exclusively or tralia and which have existed there since their in their great ma jority, in Australia. By anal­ beginning to become the ancestors from yzing the floras of the lands near Australia which the families have spread throughout today, and by drawing analogies from the the world. In the majority of cases I do not floras of the southern parts of all of the other think it likely that these plants have been continents, it can be concluded that the an­ disseminated from an Australian center in­ cestors of the Australian plants must have asmuch as paleontological evidence concern­ existed in Australia during the times of the ing animals shows that many anima ls which Upper Cretaceous period. This can be the at one time were widely distributed have only conclusion because it would be impossi­ been preserved alive in Australia while they ble for the Australian flora to form one vast have become extinct in other regions. Obvi­ atavism, if only because atavisms are .rare, ously, what has happened to animals could when they are encountered at all, and usually also have happ ened to plants. play no part in the formation of species. . We can conclude, however, from the in­ It should be stressed that in order to reach direct evidence presented by the Australian this conclusion no contrived assumptions flora that the development of primitive fam­ were made regarding the primitiveness of the ilies of the angiosperms must have taken · characteristics of the Australian plants : Con­ place during the Middle and Lower Cretace­ cepts and judgments of primitiveness were ous period or, possibly even earlier, during based entirely upon the well-established cri­ the Jurassic period. But this conclusion, while teria of the .older taxonomic systems ( the it is supported by the endemic nature 'of the Naturliche P/lanzenfamilien of Engler and .Australian flora-which, of course, was iso­ Prand, for example) and upon the general lated when the connection of Australia with discussions of phylogenetically important Malaysia was ended in the Upper Cretaceous characteristics given by Wettstein in his hand­ period-has yet to be confirm ed by the dis­ book of systematic botany (1 93 5) and . by covery of fossil evidences of angiosperms in ·myself (Suessenguth, 1938 ) , without de­ formations of those Middle Mesozoic times. pendence upon rules or criteria established particularly for the Australian flora. The spe­ REFERENCES' cial questions of the phylogene tic age of BEHRMANN, W. 1937. [N o tid e supplied.} ·Australian families put in this paper, and the Frankfurter Geog. Hefte, [abrg. 11: 19-20. answers proposed to them, have not been presented before,' to my knowledge. BRIQUET, 1. 1897. Labiatae. I n: Engler and The data obtained in these investigations Prantl's N atiirliche Pfianzenfamilien. IV would suggest that natural immigrations of Teil, Abt. 3a : 183- 375. Verlag W. Engel­ plants .into Australia,a fter the beginning of mann, Leipzig. Antiquity of th e Angiosperms-SUESSENGUTH 307

1910. Prodrom e de la flore Corse. ERDTMANN, G. 1948. [N o tide supplied.} Tome 1, .lvi + 656 pp. Georg et Cie., Grana Palynologica 1: 265-271. Geneve, Bale, Lyon. GRISEBACH, A 1872. Die V egetation del' BROWN, F. B. H. 1935. Flora of southeast­ Erde. Vol. 1: viii + 633 pp.; vol. 2: x ern Polynesia, Ill. Bernice P. Bishop Mus., + 635 pp., 1 map. R Engelmann, Leipzig. Bul. 130: 15. HAYEK, A , and F. MARKGRAF. 1927- 193 3. BUSCALIONE, 1., and G. MUSCATELLO. Prodromus florae peninsulae balcanicae. 1911-1913. Endemismi ed esodemisrni Feddes Repertor., Beihefte XXX, 1 (1927): nella flora Italiana. Malpighia 24 ( 1911) : 1-1193; 2 (1 931 ): 1-1152; 3 (1 93 3): 46 5-496; 25 (1912 ) : 85- 100, 157-172, 1-472. 283-346,453-500; 26 (1 913): 65-72, HEGI, G. 1906-1931. Flora vo n Mitteleu­ 145-192, 345-380. ropa. Band I-XIII. clviii + 7740 pp. I. CHEESEMAN, T. F. 1925. Manual of th e F. Lehmann, Miinchen. New Z ealand Flora. Ed. 2. xli + 1163 HOOKER, I. D. 1860. Flora of Australia, its pp. New Zealand Board of Science, W el­ origin, affinities, and distribution. Botany lington. of the Antarctic Voyage Ill, Flora of T as­ DIELS, 1. 189 7. Vegetationsbiologie von mania 1: cxxviii + 42 2 pp. , 200 tab . col. Neuseeland. Englers Bot. [abrb. 22 : 202- Reeve Brothers, London. 300. (See especially pp. 291-299. ) and TH. THOMSON. 1855 . Intro­ 1919. Ueber die Gattung Himan­ duction to the flora I ndica. Vol. 1: xvi + tandra. Englers Bot. Jahrb. 55: 126-134. 285 pp., 1 map. W . Pamplin, London. --- 1926. Verbreitung der Casuarina­ HUTTON, F. W. Geographical relation of ceae.Pflanzenareale, l st row, Ist book, the N ew Zealand fauna. New Z eal. Inst., 6th map. Fischer, Jena. Trans. and Proc.,5: 237-256.(Cited by Diels, 1897.) DRUCE, G. C. 1932. Comitalflora of the British Isles. xxxii + 40 7 pp. T. Bunde MANSFELD, R. 1940 . Verzeichnis der Farn­ and Co., Arbroath. und BliitenpRanzen des Deutschen Reichs. DRUDE, OSKAR. 1890. Handbu ch del' Pflan­ D eut. Bot. Gesell. Bel'. 58: 323 pp. zengeograpbie. xii + 582 pp., 4 maps, 3 M UELLER, F. YON. 1889. Second systematic figs. ( See especially p. 132 et seq.) Verlag census of A ustralian plants. Part 1: 244 I. Engelhorn, Stuttgart. pp. McCarson, Bird and Co., Melbourne. EKSTAM, O. 1897. Neue Beitrage zur Kennt­ NATHORST, A G. 188 3. Studien iiber die nis der Gefasspflanzen Novaja Semlja's, Flora Spitzbergens. Englers Bot. [abrb. 4: Englers Bot. Jahrb. 22 : 184-201. 432-448. ENGLER, A 1882. Entwicklungsgeschichte NEWBIGIN, M. 1936. Plant and animal . del' Pflanzenwelt seit dem T ertii r. Vol. 2: geography. xv + 298 pp., 39 figs. Me7 vi + 386 pp. , 1 map.(See especially p. thuen and Co., Ltd., London. 34.) W. Engelmann, Leipzig. OSTENFELD, C. H. 1926. Flora of Green­ 1894. Proteaceae. In: Engler and land. Dansk e Vidensk. Selsk. Biol.' M ed­ Pranrl's N aturliche Pflanzenfamilien. II del. VI , 3: 71 ·pp. Copenhagen. Teil, 1 H alfte: 127. W. Engelmann, Leip­ --- and I. GRONTYED. 1934. Flora of zig. Iceland and th e Faroes. xxvi + 195 pp., 308 PACIFIC SCIENCE, Vol. IV, October, 1950

2 rpaps. Levin and Munksgaard, Copen­ der Koniferen. Hamburg Inst. f. AUg. Bot. hagen. Mitt. 6 (2) : 167-308. PALI NHA, Ruv TELLES. 1939. Flora de Por­ SUESSENGUTH, K. 1938. N eue Ziele der tugal. Ed. 2a. 933 pp. Bertrand, Ltd., Botanik. 160 pp. 1. F. Lehmann , Miinchen­ Lisboa. Berlin. RECHINGER, R. K. 1943. Flora Aegaea. ---1942. Neue Pflanzen aus Costa Rica. Akad. der Wiss. Wien, Math.-Nat. «i; Englers Bot. Jahrb. 72 (2) : 270-302, Denkschr. 105 ( 1) : xx + 924 pp., 25 3 tab. tab., 3 maps. Springer, Wien. TURRILL, W. B. 1929. The plant-life of the Balkan peninsula. xxii + 490 pp., 1 map. ST. JOHN, H. 1946. Endemism in the Ha­ Clarendon Press, Oxford. waiian flora. Calif. Acad. Sci. Proc. IV, 25 (1 6) : 377- 380. VESTER, H. 1940. Areale und Arealtypen der Angiospermen-Familien. Bot. Arch. STERNER, R. 1943. Endemismer i den nor­ 4 1: 1-194,303 figs. diska floran. Sartryk ur Fauna och Flora. W ALLACE, A. R. 1880. Island life. London. Uppsala. WETTSTEIN, R. v. 1935. Handbuch der STEWART, ALBAN. 1911. A botanical sur­ systemat. Botanik. 4 Aufl., II Band: 1149 vey of the Galapagos Islands. Calif. A cad. pp., 709 figs. F. Deuticke, Leipzig-Wien. Sci. Proc, 1: 7-288. ZITTEL, K. A. v. 1895. Grttndziige der Pa­ STUDT, W. 1926. Die heutige Verbreitung laeontologie. Miinchen und Leipzig.