™. hae} ( 4 SS ‘s i] ==" . = | PLANT HABITS AND HABITATS IN THE ARID PORTIONS OF SOUTH AUSTRALIA BY WILLIAM AUSTIN CANNON PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON QK ytd WasHINGTON, 1921 ‘a Digitized by the Internet Archive in 2010. with funding from University of Toronto http://www.archive.org/details/planthabitshabi00cann PLANT HABITS AND HABITATS IN THE ARID PORTIONS OF SOUTH AUSTRALIA BY WILLIAM AUSTIN CANNON PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON By WASHINGTON, 1921 CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 308 PRESS OF GIBSON BROTHERS, INC. WASHINGTON, D. C. CONTENTS. PAGE ttre et AP RESET aE CONS sofas frat cccta epee lars aims efoe ele otelip abe arel ale o's tele iace, aM neilala soe. 8 os ee eae a eatehe v PR RREUCRER UNG ERVIN soe chk tio te Soar hd ides SA teh cana etertae atti BAe” 6 -'o at ala) da alten i palntnc eal vuatalaly 1 Physical environment of the vegetation of Australia...............0--eseeeeeeeeee 5 General physiographical conditionS............-..seeeeeeee eee c cece te eeenees 5 eaturestotthe climate Ole AUIStrAltsen . ccctaaterctale) eavetel one vavs¥enareiaice o/ote) cine oneal alate) «Poise 8 Doreen ALS Se cone os eee reeks Ue he eek Mave Ot cca Nope 'e atBila ol wale aha link aka Chote aieinoes 8 BOTAN ie ao 0/5, diene Hens iene ev epteek ete ot iar perme ante le, alm oe br tees aah alata cae 14 HPC MeiLa pC TUEREGIOY. crasctia cove, wielele © ote sicnel ocala @ Ag «Rasa s oe ainzotal o Wo Ma, och anim. i nie syaiaceta 15 Light Ee airav bd duce rule Shas tol eleaa s otete: lai wlere(eneliel sseliaC duel wi siin\ ada eteyounhia (ieee seteiai dies (cr s/2\sgelahe 4): Seekers 16 BP EENPOEMBOLE ais la cicee oso.a%sisee- ofthe ciz o'oleie wa siete ae a pyhtry «Sadan 0S ai alepe rin oa dain eels 18 VVC. Aut bie sn oie ~ a's, a 5 Netora Wie lele Mia s ot0re arel mets os sre diaiana las tibialis eid tens ate ae « 20 CHA PEEranCni CMVINOUTHeNibe.c o4,.ce seo as in ee ers os cei agai ore ai cheiigeie slp Sansa val wien apace 20 Temperature, moisture, and aeration conditions of the soil...............+++++ 23 Aeration of. bne Soll: a7 sage a ie Sa ae ale ieee a aus ats ere aw 5 Sie ae afar an la piste ee igaye 24 ‘Demperature of the Soils i. .%\-iasiee «cia de wie Oni oars a pon eye nerds lols icac aie 25 Certain characteristics of the vegetation of dry regions.............-..0+e+eeeee ees 31 Physical environment of the vegetation of South Australia..............-..+20+-ee- 35 Physical eOgrApHy «si0.0:5 o)-10 s'ele 0% ILLUSTRATIONS. PLATES. . View looking north from O’Halloran’s Mount, Oodnadatta, showing lower plain with upper plain at the extreme right in background. . Lower plain near Oodnadatta, showing ‘‘gibbers” on the surface and t ical de- yP pression with species of Eremophila. . Eremophila freelingii in a shallow wash on the slope of upper plain near Oodnadatta. . Eremophila freelingii in a shallow wash on the edge of upper plain west of Neales - River, Oodnadatta. . Acacia cambadgei in a shallow wash on the slope connecting upper and lower plains west of Neales River, Oodnadatta. . Shoot-tips with leaves, Hremophila freelingit, from upper plain west of Neales River, Oodnadatta. . Shoot-tips with leaves, Hremophila latrobei, from a wash connecting upper and lower plains west of Neales River, Oodnadatta. . Acacia tetragonophylla, near west base of sandhills east of Oodnadatta. . Acacia linophylla on sandhills east of Oodnadatta. Short channel, Neales River, with Eucalyptus rostrata and Acacia stenophylla, small shrubs, on the banks, Oodnadatta. . Phyllodia of Acacia linophylla from sandhills near Oodnadatta. . Eremophila neglecta near base of sandhills east of Oodnadatta. . Neales River bottoms from the lower plain, Oodnadatta. . Shoot-tips and phyllodia of Acacia tetragonophylla, left, and A. cambadgei, right, from Neales River, Oodnadatta. . Leaves and phyllodia of Acacia stenophylla from Neales River, Oodnadatta. . Prominent development of horizontal roots in Acacia cambadgei, Neales River, Oodnadatta. . Vegetative reproduction in Acacia stenophylla from floodplain, Neales River, Oodna- datta. . Kochia sedifolia on low slope above Copley Plain on Yudnamutana road, Copley. . Zygophyllum fruticosum at edge of Copley Plain by Table Mountain. The trees in the background are Casuarina lepidophloia, Copley. _ Nitraria scheberi hillock colonies on Copley Plain. Table Mountain is in the back- ground at left, Copley. . Detail of edge of colony of Nitraria scheberi, showing horizontal prostrate branches by which the hillock colony is extended, Copley. . Shoot-tip of Eremophila freelingit from low hills on Mount Serles road, Copley. . Eremophila oppositifolia, showing leaves and flowers, from rounded low hills on Mount Serles road, Copley. . Pholidia scoparia, ‘‘broom,” from low hills on Mount Serles road east of Copley. . Cassia sturtii, constituting a mono-specific community on Mount Serles road, Copley. . Mono-specifie community of Eremophila freelingit, in low hills along Mount Serles road, Copley. . Mono-specific community of Pholidia scoparia in low hills on Mount Serles road, Copley. . Hakea leucoptera on southern slope of Table Mountain, Copley. . Casuarina lepidophloia, or “oak,” at south base of Table Mountain, Copley. Community of Zygophyllum fruticosum near Mount of Light, Copley. . Petalostylis labicheoides from south base of Table Mountain, Copley. Casuarina lepidophloia, Copley. . Petalostylis labicheoides at south base of Table Mountain, Copley. - Shoot habit of Hakea leucoptera, with fruit, from Table Mountain, Copley. . Melaleuca glomerata, the “white” tea-tree, in a small branch of Leigh’s Creek, Mount Serles road, Copley. . Melaleuca parviflora, the “black” tea-tree, near Myrtle Springs road, Copley. . Eucalyptus rostrata, the red gum, on Leigh’s Creek, Copley. . Eremophila alternifolia at side of small wash near Mount of Light, Copley. . Eremophila longifolia on edge of Copley Plain near Leigh’s Creek, Copley. Vv vI 15, 16, ii; 18, 19, 20, 21, 22, 23, 24, 25, 26, Breuash ILLUSTRATIONS. . Shoot-tip showing leaves and fruits of Melaleuca parviflora, or ‘“‘black”’ tea-tree, from Myrtle Springs road, Copley. . Tip of shoot of Hremophila alternifolia with flowers and leaves, Copley. . Leafy shoot of Acacia varians from a wash east of Copley. . Melaleuca glomerata, ‘white’ tea-tree, from Leigh’s Creek, Copley. . Eremophila longifolia, Copley. . Branch of Acacia tetragonophylla with short spinose phyllodia and inflorescence buds, Copley. . Acacia tetragonophylla in low hills on Mount Serles road, east of Copley. . Leafy shoot-tips with fruit of Fusanus spicatus, the ““quandang,” and F. acuminatus, the native ‘‘peach,’’ Mount Deception Range, Copley. . Myoporum platycarpum from low hills in Mount Serles road, Copley. . Shoot-tip with leaves and fruit of Loranthus exocarpi and leafy branch of host, Acacia sentis, Copley. . Loranthus exocarpi, at right, and Eremophila brownii, host, Copley. . Loranthus quandang, with oval leaves, and the narrow-leaved form of Acacia aneura, the ‘‘mulga,” its host. From Mount Searles road, east of Copley. . Loranthus linearifolius on Acacia tetragonophylla. The host is shown with character- istic spine-like phyllodia. Copley. . Loranthus exocarpi, with leaves and fruit and shoot-tip of its host, Myoporum platy- carpum, Copley. . Acacia aneura, the mulga, at Ooldea. . Eucalyptus oleosa by a wash at the eastern base of Mount Deception Range. The prominent stem base and enlarged crown of the taproot, both characteristics of the ‘‘mallee,” are shown. Copley. . Detail of branch of Acacia colletioides showing spine-like phyllodia Ooldea. . Narrow “leaf” form of Acacia aneura, the mulga, at Ooldea. Young fruits are shown on one of the branches. . Broad “leaf’’ form of Acacia aneura, the mulga, at Ooldea. . Eucalyptus pyriformis at Ooldea. Various species of Acacia and the mallee, Eucalyp-. tus incrassata var. dumosa, make up the surrounding woody vegetation. The floor is bare. . Eucalyptus leucorylon var. macrocarpa, middle ground, and E. incrassata var. dumosa, on the hillside beyond, near Ooldea. . Fruits of Eucalyptus pyriformis from Ooldea. The fruits are about 5 cm. in diameter. . Leptospermum levigatum var. minus, in flower, from the Ooldea Soak. . The shrubby Eucalyptus leucorylon var. macrocarpa in flower, from Station 408 near Ooldea. . Pholidia santalina from mallee community of low ridge west of Quorn. . Callistemon teretifolius, from ridge on Mount Arden road, Quorn. . Aphyllous Acacia continua from low hills on the Pichi Richi road, west of Quorn. Tip of branch of Acacia calamifolia, in fruit, showing the linear phyllodia. From open Casuarina forest on the Melrose road, east of Quorn. . Gravillea stenobotrya shoot showing leaves and fruits, from Station 408, near Ooldea. . Leaf habit of Eremophila rotundifolia, Tarcoola. . Tips of a branch of Acacia rigens, with phyllodia, Tarcoola. . A fruiting branch of Acacia tarculiensis showing characteristic phyllodia. From type habitat, Tarcoola. . Acacia rigens, the “myall,” with various halophytes, on plain north of Tarcoola. . Thicket of mallee, Eucalyptus oleosa, on sloping saltbush plain, foothills of the Flinders, east of Port Augusta, near Saltia. . “Beef wood,” Gravillea stenobotrya, on the crest of sandhill by Station 408, near Ooldea. . Forest of Eucalyptus rostrata on Saltia creek, east of Port Augusta. . Pine community, Callitris robusta, at Warren’s Gorge, near Quorn. . View in mallee scrub, about 2 miles north of Quorn. Eucalyptus odorata and E. oleosa in the background. Bunches of Triodia irritans in the foreground. 29,4 ILLUSTRATIONS. VII . Branches of Acacia sublanata, showing small and rigid phyllodia, Quorn. . Eutazxia empetrifolia, showing the small flowers and linear short leaves, Quorn. . Branches of Acacia pycnantha, the “golden wattle,’’ showing the character of the large phyllodia, Quorn. . View about 2 miles west of Quorn, taken from a grassy ridge and looking upon a ridge which is covered with mallee. In the intervening valley are a few specimens of Eucalyptus leucorylon var. pauperita. . Hakea leucoptera on the edge of the mallee scrub, about 2 miles north of Quorn. Small shoots which spring from superficial roots of the larger plants are in the foreground. . Western slope of ridge along Mount Arden road, Quorn, with Triodia irritans and Trichinium, dominant grasses. Dead fruiting stalks of Xanthorrhea semi- plana shown in the foreground; mallee, Eucalyptus sp., in the background. . Bossiea walkeri on summit of a sandhill by Station 408, near Ooldea. . Hakea multilineata on the crest of a sandhill by Station 408, near Ooldea, with Euca- lyptus incrassata var. dumosa, a mallee, in the flats below. Bunches of spinifex, Triodia irritans, are to be seen between the mallee. . Branch with withered flower-spike and leaves of Hakea multilineata, from Station 408, near Ooldea. . Melaleuca uncinata in fruit, from the sandhills by Station 408, near Ooldea. 30, a. A community of Acacia pycnantha, the golden wattle, by a streamway on the Mount Brown road, Quorn. . A large specimen of Eucalyptus leucorylon var. pauperita, by a wash on the Mount Arden road, Quorn. A comparison with the automobile will give an idea of its size. . Vegetative reproduction in Hakea leucoptera. A young shoot, removed from the soil, is shown taking its origin from a horizontal root. Quorn. . Exposure of roots of mallee, Hucalyptus sp., by a narrow wash, showing the abun- dance of superficial roots. Along the Mount Arden road, Quorn. . Root exposure of Eucalyptus leucoxylon var. pauperita by erosion of the bank of stream above Warren’s Gorge. The roots were washed out for a distance exceeding 16 meters. Quarn. . Scattered groups of Melaleuca parviflora, in the mallee scrub near Blanchtown. . Flood plain of the Murray River showing open forest of Eucalyptus rostrata partly submerged, Blanchtown. . View in mallee, Eucalyptus sp., scrub on Murray flats near Blanchtown. FIGURES. Physical divisions of Australia, after Gregory, 1916, to which has been added the 10-inch isohyet. The shaded areas have an altitude of 1,000 feet or more above the sea. Mean annual rainfall map of Australia, adapted from Hunt. . Duration of wet seasons, after Taylor, 1916. ‘The periods shown on the map include those months in which the average rainfall exceeds the geometric mean of the monthly rainfalls.” . Wettest months of the year, after Hunt’s meteorological map of Australia, 1916. . Mean rainfall of Australia for January, after Hunt. . Mean rainfall of Australia for April, after Hunt. - Mean rainfall of Australia for July, after Hunt. . Mean rainfall of Australia for October, after Hunt. Graphs, after Hunt, showing average monthly rainfall and mean monthly evaporation, in inches, for various places in Australia. . Mean humditiy of Australia for January, after Taylor, 1918. . Mean humidity of Australia for July, after Taylor, 1918. . Mean annual evaporation in Australia, after Hunt. . Average yearly temperature of Australia, after Hunt. . Mean temperature of Australia for January, after Hunt. . Mean temperature of Australia for July, after Hunt. VIII 10. DE 12. 13. 14. 15. 16. Wg 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. ILLUSTRATIONS. Chief physical divisions and geographical plan of South Australia, after Howchin and Gregory, with the 5-, 10-, and 15-inch isohyets. Graphs showing the annual (total) and ‘‘non-effective”’ rainfall for 1901-1906 at Oodnadatta (a), Copley (b), and Quorn (c), South Australia, based on records supplied by the Adelaide office of the Commonwealth Bureau of Meteorology. Acacia linophylla, transverse section of phyllode, semi-diagrammatic, X 72. The large proportion of mechanical tissue is indicated (sc), and the protected position of the chlorenchyma (ch). The relatively heavy covering of hairs is indicated by the stippling. Same. Detail of margin of phyllode to show the nature of the sclerenchyma and epidermal cells and the presence of glandular trichomes, < 700. Same. Detail of inner portion of chlorenchyma showing its relation to the fibro- vascular bundle at the left, x 700. Acacia continua, transverse section of chlorophyll-bearing stem, X 52.5. Acacia tetragonophylla, cross-section of phyllode, semi-diagrammatic, X 85. Casuarina stricta, transverse section, semi-diagrammatic, of chlorophyll-bearing stem, x 72. The chlorenchyma is shown partly protected by the heavy epidermis and partly by the furrows with the trichomes, of which the latter are not shown. The enlarged outer ends of the sclerenchyma also act in the same capacity. Eremophila alternifolia, detail of young stem with glandular trichome, X 52.5. Same. Transverse section of leaf showing old glandular trichome, heavy epidermis, and its covering of a resinous substance. Eremophila freelingii, semi-diagrammatic transverse section of leaf to show the size and frequency of internal glands (gl), X 52.5. Eremophila rotundifolia, longitudinal section, semi-diagrammatic, X 52.5, to show the relatively large internal glands and the very heavy covering of hairs (ér). Fusanus acuminatus, fragment of leaf showing chlorenchyma and a group of tracheids, X 350. Same. Cross-section of leaf to show the heavy epidermis consisting of two layers of cells, X 350. Gravillea stenobotrya, semi-diagrammatic transverse section of leaf. The various tissues are as indicated. Trichomes and stomata are confined to the ventral side, X 52.5. Same. Detail of leaf, dorsal side, in cross-section, to show the greatly elongated epidermal cells and well-marked palisades, X 350. Hakea leucoptera, leaf fragment, in transverse secticn, with very heavy epidermis and deeply sunken stoma and papillate processes in stomatal canal. The presence of sclerenchymatous fibers in the palisade chlorenchyma is shown. X 350. Hakea multilineata, semi-diagrammatic cross-section of leaf. The prominent devel- opment of mechanical tissue and dorsiventral nature of the leaf structure are indicated. X 52.5. Same. Fragment of leaf, cross-section, to show heavy epidermis, deeply sunken stoma, and pronounced palisade character of the chlorenchyma, X 350. Pittosporum phyllyreoides, fragment of dorsal side of leaf, transverse section, to show the 2- or 3-layered epidermis, X 350. Same, ventral side of leaf. The heavy outer epidermal wall, the single cell layer of the epidermis, and the superficially placed stoma are indicated. XX 350. Triodia irritans, transverse section of leaf, semi-diagrammatic, showing its infolded condition and the position and relative abundance of the main tissues, X 85. In figures 12 to 31 the tissues are designated as follows: ch, chlorenchyma; fz, conductive tissue; gl, internal gland; hd, hypoderm; sc, sclerenchyma; ep, epidermis; fv, fibro-vascular tissue. PLANT HABITS AND HABITATS IN THE ARID PORTIONS OF SOUTH AUSTRALIA. INTRODUCTION. Australia, especially South Australia, holds much of interest to the student of the vegetation of arid regions. Where rain is abundant plants compete with one another in a very real way for room in which © to live and for sunlight by which they gain energy for various life processes, but in regions of scarity rainfall, as in portions of South Australia, there is abundance of room and of light. Here the “struggle”’ is associated with the water relation; it is that of the individual plant with an arid environment, and not individual with individual. When viewed from this standpoint the island continent is seen to be the field of a vast botanical experiment in which may be observed the reaction of numerous species and innumerable individuals to a physical en- vironment, a leading characteristic of which is a relatively small water- supply. Moreover, owing to the great age of Australia, it is possible that nowhere else have plants been exposed to and influenced by an arid environment for a longer period of time. The physical background of the Australian plants is in a measure unique. The dry region is very extensive. Some idea of its size can be had by the statement that it has nearly as great an area as all Arabia, and as a matter of fact is larger than all other regions of the kind south of the equator. Living and developing under an environment of which the keynote is aridity, the flora of the continent as a whole bears a xerophytic stamp and appears to possess a degree of uniformity which constitutes one of its most marked characteristics. Wherever one goes in Australia, he encounters trees and shrubs with leathery, ever- green leaves. In the better-watered portions the trees are large and numerous and there is an extensive transpiration surface, but in portions less favored in this particular the trees are not large, a shrubby type of vegetation prevails, and the transpiration surface is also much restricted in area. Between the two extremes there are innumerable intermediate conditions in which the gradations are quantitative rather than qualitative. When studied in some detail, however, there may be found a bewildering variety of adjustments to the environment, often monotonous perhaps in outward appearance, and which have in a measure the force of belying the generalization just made. The physical and biological complexes which enter into our concep- tion of what constitutes an arid, more especially a desert, region are made up of many features. It is true that they center around the important factor of a small water-supply, but they all should be logically included in any definition of such regions with scanty rainfall. 1 2 PLANT HABITS AND HABITATS IN THE Owing, however, in part to the difficulty in evaluating the biological value of the accessory factors, such as light intensity, relative humidity of the air, rate of evaporation, and temperature of the air, and in part to the fact that such secondary factors may change in relative force with changes in the amount of the rainfall, it is difficult to express adequately what constitutes a “desert,” or even the degree of aridity. Nevertheless it is important to have some method of comparison. Thus, the extremes in amount of rainfall have been used (MacDougal, 1914:175), and the amount of evaporation for any given year has been compared with the precipitation for the same year, and, finally, com- parisons have been instituted between the moisture content of the soil and the rate of evaporation of the air (Shreve, 1915:92). The intensity of the aridity has also been expressed biologically in terms of the relative number of annuals in a region (Paulsen, 1912:159). It was not convenient in the present instance to use these methods, but it appeared necessary nevertheless (mainly for convenience in reference) to have some ready means of comparing one region studied with another, and in the end the device was resorted to of using the rainfall only. An arbitrary classification of regions based on the amount of rain was consequently adopted, which is as follows: A region having 5 inches, or less, of rain annually is a desert; one with a rainfall between 5 and 10 inches is arid; and a region in which the amount of rain is be- tween 10 and 15 inches is semi-arid. In all cases, therefore, in which reference is made in the text to regions so designated the appropriate rainfall will at once be understood. The study of plants in the field may be said to proceed mainly along three lines, which, although more or less intermingled, are fundamentally quite different. Thus, the leading emphasis can be placed on the plants as species, and their occurrence (local as well as general) can merely be catalogued; this is plant geography in a narrow sense. Or, the mutual relationships of plants can be investigated; this is one formal branch of plant ecology. Or, finally, the investigation can take into consideration mainly the relations of plants to the physical environ- ment in which they are placed; this third phase of the general subject is intimately related to experimental researches along lines suggested by field observations and is not to be dissociated from laboratory studies; this can be referred to as physiological plant ecology. It is the last type of ecological research which the writer has had especially in mind when making field studies, and though it has not been practicable to carry out direct experiments on subjects suggested by the observations, it has been of interest and profit to interpret the observations so far as possible in the ight of experimental results already accomplished on analogous lines and with analogous plants by various researchers. In addition to viewing the living plants from a physiological stand- point, another point of view has been of use, the comparative. In all instances the plants observed have been studied in the light of the ARID PORTIONS OF SOUTH AUSTRALIA. 3 writer’s previous experiences in the dry portions of North America, in southern Algeria, and in portions of Egypt, and these experiences have been of incalculable assistance in attempts at interpreting the various features of Australian plant life observed. It is impossible adequately to acknowledge the very many kind- nesses shown the writer while in Australia. A friendly and helpful spirit of assistance and cooperation was shown by a large number and on very many occasions; but especial acknowledgment must be given Professor and Mrs. T. G. B. Osborn, of the University of Adelaide, who helped greatly to make the visit pleasant as well as profitable. Dr. and Mrs. R. 8S. Rogers, of Adelaide, well known for their studies on the Orchidaceex, acted as guides on several botanical excursions into the Mount Lofty Ranges, and in other ways were helpful. J. M. Black esq., of Adelaide, an authority on South Australian plants, very kindly determined those plants which were collected by the writer. Among them Mr. Black found some new stations and a species of Kochia collected at Copley which was previously undescribed. Men- tion should also be made of the assistance of Alfred Cocks esq., of Adelaide, the former proprietor of “‘Wilgena”’ station, near which Tarcoola is situated, whose acquaintance with the ‘back blocks” of the state is very extensive. Thomas Gill esq., of Adelaide, was of as- sistance in procuring for the writer useful works on Australian explora- tion and in other ways; C. S. Owen-Smith esq., of Adelaide, was also helpful in various ways; and finally, not to mention others, G. A. Hobler esq., and Capt. E. W. Saunders, of the Commonwealth rail- ways, kindly placed conveniences at the writer’s disposal at Ooldea, and were of much assistance in other ways and at other times. Especial acknowledgment must be made of permission to use figures, tables, or data for the presentation of many characteristics of the physical environment of the South Australian plants. Figure 1 is from Australia, 1916, by Professor J. W. Gregory; figure 10 is an adaptation of a figure in the Geography of South Australia, 1909, by Howchin and Gregory. Those noted as being from Hunt are by H. A. Hunt esq., Commonwealth meteorologist, and have been taken from various publications of the Commonwealth Bureau of Meteorology. The figures from Taylor are by Dr. Griffith Taylor, physiographer, Commonwealth Meteorological Office, and arein part from publications of the bureau and in part from other publications as noted. Asa whole the figures were prepared to serve other than botanical ends and in most of them some changes have been made, inconsiderable in certain instances, to suit the needs of the present study. Figure 11 is based on data supplied by the Adelaide office of the Commonwealth Bureau of Meteorology. In the main the climatological data were supplied by the Commonwealth Bureau or were derived from its pub- lications, and in either case acknowledgment is made explicitly in the course of the study. 4 PLANT HABITS AND HABITATS IN THE The following species, determined by J. M. Black esq., were collected at Oodnadatta, Copley, Quorn, Port Augusta, Tarcoola, and Blanch- town, and in the regions contiguous to these places: Acacia aneura F. v. M. brachystachya Benth. calamifolia Sweet. cambadgei R. T. Bake. colletioides A. Cunn. continua Benth. hakeoides A. Cunn. iteaphylla F. v. M. kempeana F. v. M. linophylla W. V. Fitz. oswaldi F. v. M. pycnantha Benth. randelliana W. V. Fitz. rigens A. Cunn. salicina Lindl. sentis F. v. M. stenophylla A. Cunn. sublimata Benth. tarculiensis J. M. Black. varians Benth. Atriplex mumularrum Lindl. quinii F. v. M. spongiosum F. vy. M. vesicarium Hew. Bassia echinopsilla F. v. M. lanicuspis F. v. M. paradoxa F. v. M. Bossiva walkeri F. v. M. Bursaria spinosa Cavan. Calandrinia pusilla Lindl. Callistemon teretifolius F. v. M. Callitris robusta R. Br. Cassia artemioides Gaud. brownii R. Br. eremophila A. Cunn. sturtii R. Br. Cassinia aculeata (Lab.) R. Br. Casuarina lepidophloia F. v. M. stricta Ait. Cheilanthes tenuifolia Schwartz. Dodoneza attenuata A. Cunn. bursarifolia Behr et F. v. M. lobulata F. v. M. Enchylena tomentosa R. Br. Eremophila alternifolia R. Br. brownii F. v. M. freelingiui F. v. M. latrobei F. v. M. longifolia F. v. M. neglecta J. M. Black. oppositifolia R. Br. paisleyi F. v. M. rotundifolia F. v. M. Eucalyptus incrassata Lebill. var. dumosa Maiden. odorata F. v. M. oleosa F. v. M. rostrata F. v. M. Eutaxia empetrifolia Schlecht. Exocarpus aphylla R. Br. spartea R. Br. Fusanus acuminatus R. Br. spicatus R. Br. Geigera parviflora Lindl. Glycine clandestina Wendl. Gravillea stenobotrya F. v. M. Hakea leucoptera R. Br. multilineata Meiss. Helichrysum spiculatum D. C. Heterodendrum olxefolium Desf. Indigofera australis Wild. var. minor Benth. Jasminum lineare R. Br. Kochia cannoni J. M. Black (n. s.) decaptera F. v. M. eriantha F. v. M. planifolia F. v. M. pyrmidata Benth. sedifolia F. v. M. villosa Lindl. Leptospermum levigatum F. v. M. var. minus. Loranthus exocarpi Behr. linearifolius Hook. pendulus Sieb. quandang Lindl. Melaleuca glomerata F. v. M. parviflora Lindl. Menkea australis Lehm. Myoporum platycarpum R. Br. Nicotiana suaveolens Lehm. Nitraria schoeberi L. Olearia muelleri Benth. pimeleoides Benth. pannosa Hook. Pholidia santalina F. v. M. scoparia R. Br. Pimelea microcephala R. Br. Pittosporum phillyreoides D. C. Rhagodia parabolica R. Br. Salicornia tenuis Benth. (n. s.?). Sezvola collaris F. v. M. Senecio anethifolius A. Cunn. gregorii F. v. M. magnificus F. v. m. Sida corrugata Lindl. Solanum ellipticum R. Br. Templetonia aculeata Benth. egena Sweet. Trichinium incanum R. Br. spathulacum R. Br. Triodia irritans R. Br. Zygophyllum billardieri D. C. crenatum F. v. M. fruticulosum D. C. prismatothecum F. v. M. ARID PORTIONS OF SOUTH AUSTRALIA. . 5 PHYSICAL ENVIRONMENT OF THE VEGETATION OF AUSTRALIA. The vegetational environment of Australia, including the more arid portions, has a complex geographical background. The island continent is separated biologically, as well as physically, from other continents and has been so separated for an immense period of time. Long geological ages also have passed since a large portion of the surface was covered by the sea. The physiography is relatively monotonous, as might be expected from the fact that possibly the area may be regarded as a vast peneplain. The latitudinal situation also is of importance in influencing, really in shaping, the leading characteristic of its climate. Projecting as it does far into the interior, the state of South Australia shares in the general physical characteristics of the continent, but it also holds in certain regards a peculiar relation to the sister states. Its southern shores are washed by the cool seas, while the northern boundaries are parched and baked under a tropic sun. It has the most typically continental climate of all the states (Howchin and Gregory, 1909:17). Such are some of the features which have con- stituted and which now constitute the broad characters of the physical environment of the vegetation of Australia, including that of the central state, and under which by physiological reactions there has slowly developed the vegetation familiar to us at the present day. GENERAL PHYSIOGRAPHICAL CONDITIONS. The general physiographical conditions of Australia have many points of interest in connection with the present paper. Gregory (1916:25-27) states that: “ Australia as a whole is a great plateau land. It isa fragment of a large continent, the rest of which has been snapped off along great fractures. . The mountain system is not determined by any dominant lines of folding of the earth’s crust, like those which have formed the Alps and the Himalayas. Australia was folded at an early period in the earth’s history; and all its ancient fold-mountains have long since been worn down. The chief existing features in the relief of Australia are due to vertical earth-movements, by which some parts of the area have been raised to high plateaus and others have sagged downward into deep basins. The margins of the plateaus have been carved into valleys. . . . The eastern margin of the old plateau has been dissected by powerful streams into deep valleys, which are separated by steep-sided and flat-topped ridges; and in some districts river erosion has been so active that very little of the original surface has been left. eS Western Australia, on the other hand, owing to its smaller rainfall and feebler rivers, retains more of the old plateau surface. . . . The inner part of the plateau is a vast gently undulating country, with low rounded hills, except where some hard wind-etched boss of rock rises abruptly from the plains. Wide, shallow depressions run together like the converging branches of a river; and these valleys are divided by the irregularities of their floors into basins, which in wet seasons may contain lakes of little depth; but usually they are 6 PLANT HABITS AND HABITATS IN THE sheets of salt-incrusted clay, or damp mud and salt marsh. . . . Owing to the absence of recent fold mountains the relief of the continent depends on the weathering of the old plateau and the formation of highlands and low- lands by the uplift or subsidence of wide tracts of country.” The same writer divides the continent of Australia into the Eastern Highlands, the Great Plains, and the Western Plateau. The rela- tive extent and position of each division is indicated in figure 1. It Fig. 1.—Physical divisions of Australia, after Gregory, 1916, to which has been added the 10-inch isohyet. The shaded areas have an altitude of 1,000 feet or more above the sea. will be seen that the desert-arid regions lie in the western and cen- tral divisions and that possibly half of the area within the 10-inch isohyet has an altitude of 1,000 feet or more. On the other hand, a not inconsiderable proportion of the whole of the desert-arid regions is situated in the artesian basin of the Great Plains regions and is below the level of the sea. The Western Plateau is not level, but several mountain chains rise upon it; certain of these attain to considerable altitude, as, for example, Mount Woodrofe, of the Musgrave Range, which is over 5,000 feet above the sea, approximately 3,000 feet higher than the surrounding plain (Jack, 1915). ARID PORTIONS OF SOUTH AUSTRALIA. 7 To quote Jutson (1914:20), the interior of the great plateau is arid and has no permanent rivers. The drainage runs into shallow basins (with no outlet except at times of great flows of water), called “‘salt”’ or ‘“‘dry” lakes. There is, however, fresh water to be found where the catchment has been suitable, as, for example, the “soaks” like the well- known one at Ooldea, and basins in the rocks, as vividly described by Carnegie (1898:191). But, with the possible exception of the former, these have no significance so far as the vegetation is concerned. As the interior has little rain and no rivers, there is no water table. The picture that the great plateau presents as a whole is therefore a very arid one, both as regards the aerial and the subaerial plant environment. The great plateau is regarded by Jutson (1914:20-21) as being an old, uplifted surface, a vast peneplain, whose surface has been much destroyed, planed down, and often not recognizable. In the southern half of Western Australia the rocks of a very large area are probably pre-Cambrian, while in the southeastern corner they are Mesozoic or early Tertiary (Jutson, I. c.), which is the region known as the Nullarbor Plains. Where the great plateau of the western part of the continent joins the big central artesian basin, the rocks are also of the Mesozoic and more recent ages. As a whole, therefore, the great plateau is of very great geological antiquity. That portion of the desert-arid region which lies within the Great Plains varies in altitude from somewhat below the level of the sea to 1,000 feet or more above it. One of the characteristic physiographical features is the presence of steppes (Spencer and Gillen, 1912:5) or table-lands of Upper Cretaceous and Lower Cretaceous formations (Jack, 1915:13), which connect the Great Plains to the Western Plateau. On the east the plains gradually rise to the highlands of eastern Australia. The sandhills, which are especially to be found to the north and east of Lake Eyre, but occur to the northwest as well, are built of material derived by erosion from the desert sandstone of the steppes, and the “‘gibber” plains, or stony deserts, also trace their origin to these Cretaceous plateaus and are the residue remaining in place. A most striking feature of the Great Plains regions is the presence of several large lakes which in earlier geological times contained fresh water but are now saline wastes, usually carrying water only after heavy rains. Of these, Lake Eyre and Lake Eyre South are the largest, covering an area of 5,000 square miles when filled with water (Howchin and Gregory, 1909:100). Lake Eyre receives the discharge of several rivers of intermittent flow. At various places in the Lake Eyre basin natural artesian wells are found whose outlets are raised into small hills through the deposition of minerals held in solution. Also, numerous ‘bores,’ deep wells, have been sunk for economic purposes. The water that supplies the wells of whatever sort is derived from rains falling in the Eastern Highlands or is chiefly plutonic (Pittman, 1914). 8 PLANT HABITS AND HABITATS IN THE FEATURES OF THE CLIMATE OF AUSTRALIA. RAINFALL. There are four major rainfall regions in Australia, according to Taylor (19187: 10), namely, the summer-rain region of the north, the winter- rain region of the south, a region of uniform rains in the east, and a region of little rain in the center and middle west. The seasonal shifting of the climatic complex north-south, following the declination of the sun, operates to bring about the seasonal rains. Tropical storms, cyclones, reach inland in summer and cover the entire northern portion of the continent. They extend south barely as far as Oodnadatta. The winter storms, on the other hand, affect the southern portion and extend north only about as far as Farina, South Australia (Taylor, 1918: 10). These accompany the northern extension of the prevailing westerlies which in summer are far to the south of Australia. There is, therefore, a belt of territory running roughly east and west which is beyond the usual reach either of the winter or of the summer rains. This dry central region comprises, according to Taylor, 37 per cent of the area of the continent. TaBLE 1.—Rainfall at the capitals. A Capaure Percentage 5 SRO BE No. of of rainfall Capital. rainfall, nd? : : : SE ee ays rain. | in 6 wettest months. Ad elaIden st wasn eee 20.88 123 70 IBTishaners taku ae ieee oe 46.65 130 67.7 Elo bart yor terre Sse iat 23.39 144 55 Melbourne se nee we 2oLa2 133 49 Berths} {essere ass 32.91 115 86.9 Sydueyis fee 2 oa 24.49 159.5 59.9 The leading characteristics of the rainfall in the humid regions can be illustrated by that at the capitals of the different states. These are summarized in table 1. At Adelaide and at Perth rain occurs mostly in the winter season; at Brisbane it is mostly in summer; at Sydney it is mostly in late summer, autumn, and early winter, while at Melbourne and at Hobart it is fairly evenly distributed throughout the year. So far as the well-being of the vegetation as a whole is concerned the reliability of the rains (or the want of this) is of capital importance. And in a general way the reliability of the rains decreases with the decrease in the amount of the rainfall, which it will be seen only serves to intensify the effects of progressive aridity. Thus we find that in the dry interior of the continent there is a mean variation equal to 40 to 50 per cent from the normal (Taylor, 1918', fig. 4.) In some years very little rain falls at any season, while in others almost the entire yearly rain may fall within a few hours. While the rainfall throughout the interior is as a rule very unreliable, the portion of the dry mid-region ARID PORTIONS OF SOUTH AUSTRALIA. 9 which has the least reliable rainfall, according to the same author, is in Western Australia. At Roeburn, for example, the following extremes in precipitation have been recorded: In the year 1891 the rainfall was 0.13 inch only; in the year 1900, at the opposite extreme, the rainfall amounted to 42 inches. The region of greatest rain variability is roughly equal in extent to that which lies within the 10-inch isohyet, but it is situated somewhat farther north and hence is largely in the region of summer rains. Fig. 2.—Mean annual rainfall map of Australia, adapted from Hunt. It can be seen that with the possible exceptions of the dry interior on the one hand and the humid regions on the other, the precipitation of the continent can be characterized as periodic. There are thus many days in the year when no rain falls and, as just suggested, these may occur in large degree consecutively. The actual number of rainless days may be surprisingly large, as the following will indicate; the figures are for the year 1912 only: In the humid regions the average number of days without rain for 26 stations is 306.2. In the semi-arid regions the average of 24 stations is 332 rainless days, the average for 16 stations in the arid regions is 328.4, and finally, the average number of rainless days in the desert, 6 stations, is 346.4. It is possible that the regular recurrence of rainless periods over most of Australia is a very important factor, although a very complex one, in giving the vegetation as a whole the xerophytic stamp it bears. 10 PLANT HABITS AND HABITATS IN THE , TaBLE 2.—Mean monthly and annual rainfall (in inches).* . | May.| June.|July. | Aug. |Sept. iMieans 3° 2.55 0.71 |0.71 |0.57 |0.18 |0.23 |0.64 |0.28 |0.11 |0.32 |0.34 |0.38 |0.38 | 4.85 Highest. 22.22: 5.18 [4.16 |4.32 |1.79 |1.54 |2.77 |1.68 |0.60 {2.21 |2.03 |1.41 [1.79 | 8.92 LOowest.2.3-.% 0.0 10-0. -|020=10:0:.410:0° .|0:0. 10.0-31020. 1020 (020 410-0 31020 1.54 Leigh’s Creek: : (Copley) Mean |0.67 |0.52 |0.80 |0.54 |1.06 {1.11 |0.50 |0.69 |0.75 |0.47 |0.53 |0.76 | 8.40 Highest....... 4.76 |3.05 |4.70 |3.78 |5.84 |4.72 |2.24 |3.02 |3.58 12.06 |2.35 |3.78 |15.6 Lowest........ 0.0 |0.0 {0.0 |0.0 |0.0 |0.0-|0.0 |0.0 {0.0 |0.0 |0.0 {0.0 1.95 Quorn: Mean eamcen x 0.64 |0.45 |0.61 |0.96 [1.58 |2.03 |1.57 |1.84 |1.30 |1.32 |0.87 0.65 |13.8 Highest....... 3.20 |2.53 |3.53 16.38 |6.86 |4.88 |6.15 |5.53 |4.00 |4.44 |5.36 |2.45 |25.7 Lowest... .¢ << 0.0 {0.0 |0.0 |0.0 |0.01 |0.23 |0.14 |0.06 0.0 |0.09 |0.0 {0.0 7.43 Adelaide: WMean'.,./.. tse 0.72 |0.63 |1.06 {1.85 {2.71 |3.10 |2.65 |2.50 {1.98 |1.72 |1.17 {0.96 |21.0 Highest....... 4.00. |2.67 |4.60 |6.78 |7.75 |8.58 |5.38 |6.24 |4.64 |3.83 |3.55 |3.98 |30.8 Lowést.. +. .i2. 0.0 |0.0 |0.0 /0.06 |0.20 |0.42 |0.36 |0.35 |0.45 |0.17 |0.04 |0.0 {11.3 Kalgoorlie: |S agers ae 0.43 {0.73 |0.90 |0.78 {1.31 |1.27 |0.91 |0.90 |0.52 |0.79 |0.58 {0.62 | 9.74 HMazhest..< = os 2.50 |4.68 15.02 |3.43 {3.12 |3.00 |2.08 |3.18 |3.29 |3.14 |2.76 |2.57 |16.4 Wowest: .... =. 0.0 {0.0 |0.0 |0.0 {0.0 |0.0 |0.22 |0.0 |0.0 |0.0 |0.0 {0.0 4.75 a ee Ln ee ‘Dec. to Apr/ Dec. ‘*The periods shown on the map include those months in which the average rainfall exceeds the geometric mean of monthly rainfalls.” Fig. 3a.—Duration of wet seasons, after Taylor, 1916. ARID PORTIONS OF SOUTH AUSTRALIA. ths Si ee ti Y SS 1k me. a January February March ay Muy H oy he "arn iN & # Ml Fre. 4a.—Mean rainfall of Australia for January, after Hunt. 11 12 PLANT HABITS AND HABITATS IN THE Fig. 46.—Mean rainfall of Australia for April, after Hunt. TABLE 3.—Average yearly evaporation and rainfall for representative Australian stations | Evaporation,| Rainfall, inches. inches. Evaporation,| Rainfall, inches. mache Station. Station. 66.13 33. Broken Hill 5.63 9.66 146.57 : Sydney .91 48.16 156.02 2 Melbourne........ .38 25.60 Kalgoorlie 87.74 ; Brisbane .95 47.05 Adelaide 54.28 a 3¥f 25.50 97.10 : Vi .67 9.94 ARID PORTIONS OF SOUTH AUSTRALIA. 13 i es Mh 1 t Tg a in. Te Wye n % Fig. 5a.—Mean rainfall of Australia for July, after Hunt. TaBLE 4.—Mean monthly and annual evaporation (in inches) .* i ff ff ff ff J MeRRERE oot. 122) (LOZ Sat O75) | OnOnh4.Selvore jmosO: | onOliliede te 192 2) |LOBGs ONO) Ob Highest mean. .|15.4 ,13.8 {12.6 | 8.4 | 5.5 1 4.1 | 4.4 | 6.4 | 8.5 |11.7 |12.8 |14.1 |108 MmowestimMeaneal orl | Oso Wao sO WosOlinond [mado Wicae [ic Model ors sec. fy scoot Se Adelaide: Means 2si.c45> ESHA eV ARTS | Meare oud fica eee pi | ete My feed (AT RS Le eee at LSet Rea hal | ov: br a cea le wy oad bans: § Highest mean../11.2 | 9.1 | 7.2 | 4.9 | 2.8 | 1.8 | 2.3 | 2.7 | 3.8 | 7.3 | 7.8 |10.4 | 60 Lowest mean..| 7.3 | 4.9 | 4.2 | 2.1] 1.2 8 on tek ie Ouse hi |t4osrinGsea i246 Eucla: WMigsitegs ccsee G28. |5G.0: voce. || Oelrocds | erouieesG: [roast ol A Moo. G: lone tise ham Highest mean..| 7.9 | 7.1 | 6.5 | 4.6 | 3.6 | 3.0 | 3.2 |} 4.1 |] 5.5 | 7.1 | 6.7 | 8.3 | 60 Mowestmcant pont Ove Weel alice eerie le: oh DoS) Qe iSO | boo GA 5S Coolgardie: Miesinte aeye.» SOO (Sion) 0 0 0 0 0 0. 0. 0 0 0 0 1 eoocoeo orooo°o RFOrnNnoo oonocjo SCO ONWSSOSD SCSOOROO fer) om CoOonRwoO wookoo COoNoOoO OWwNMO Zo ma ewWooce woorooo ooooco roonooe ooorcoco ooorco ocoooccoco oorwodo ooococo wNoorce POrFROCrO Ber OOF NoOCNnNocoO co © o TABLE 12.—Mean, highest, and lowest monthly rainfall at Oodnadatta, in inches, for a period of 26 years. ARID PORTIONS OF SOUTH AUSTRALIA. 49 the percentage of what may be called the non-effective rainfall in- creases with the decrease in the total average rainfall. At Quorn, for example, it is 18 per cent, at Copley it is 19 per cent, and at Oodnadatta it is 30 per cent for the years under consideration. Assuming that these figures represent the average percentage of the non-effective rain- fall, which may or may not be the case, and knowing the actual average rainfall for the three stations, the expected efficient rainfall can be easily calculated. For example, the average rainfall at Oodnadatta is 4.85 inches. This being the case, a correction for the percentage that does no good to the native vegetation would give an efficient average rainfall of 3.42 inches. The average rainfall for Copley is 8.70 inches, which would amount to 7.13 inches effective rainfall, and that at Quorn is 13.72 inches. At the latter station the effective amount would be 11.11 inches. The relation of the total, the effective and the non- effective, rainfall at Oodnadatta, Copley, and at Quorn is shown graphically in figure 11. In 1906 the non-effective rainfall at Oodna- datta was 43 per cent of the total recorded for the year. Thus the effective amount of precipitation for the year at Oodnadatta was 1.7 inches. This exceeds the percentage of non-effective rainfall of each of the other stations whose records were studied and indicates (together with the average percentage of non-effective rainfall for the stations given above) that the curve of aridity falls more rapidly, as one goes from a less to a more arid station, than the curve of the rainfall itself. 50 PLANT HABITS AND HABITATS IN THE VEGETATION AND PLANT HABITATS IN VICINITY OF OODNADATTA. PHYSIOGRAPHY. The country to the south and to the west of Lake Eyre and Lake Eyre South, as seen from the railway, has a monotonous topography, although not without a certain diversity. As one goes west from Maree (Hergott Springs) the land-surface on every side, except to the south, is fairly level. Some distance on the south lies the north- western spur of the Flinders Ranges, but in the other directions the general flat expanse is broken only occasionally by low hills or mounds. Maree lies on the southern edge of the great Australian artesian basin and in its vicinity, as elsewhere along the line of the railway, mounds mark the presence of springs. There are 50 or more of these springs in the southern and southwestern part of the basin. The water is charged heavily with mineral matter and upon evaporation leaves a calcareous deposit. ‘Thus the travertine mounds are formed. The water flows out of the summit of the mounds and extends for a greater or less distance from the base. The mound springs thus can carry vegetation and are often the only green spots, truly oases, on a parched and barren plain. But the leading characteristic of the country to the south of Lake Eyre South is its flatness. On every side it is a plain, and to the north it stretches as far as the eye can reach, a grayish-green plain that finally merges in a mirage, suggesting flat expanses of a lake. Shortly before reaching the station of Stuart’s Creek the railway crosses what is probably the southern edge of the bed of the lake of earlier times, running below the level of the sea, and here, on the rare occasions when filled with water, it may be seen from the railroad. Thus, in July 1918, there seemed to be water in the south arm of the lake. After leaving Stuart’s Creek the railway ascends gradually until at Oodnadatta it is about 400 feet above sea-level. The topography becomes more broken and presents several features of especial interest. Beyond William Creek some worn-down pre-Cambrian hills extend, and flat-topped hills may be seen to the west, which are probably outliers of the desert sandstone tableland (Upper Cretaceous). Occasionally sandhills were seen lying toward Lake Eyre. The railway line also runs among low hills and along salt flats, and finally, not far south of Oodna- datta, it crosses a wide and poorly defined river bottom, which at the time of my visit was dry. Thus we have the leading topographical elements of the Oodnadatta region, namely, the lower plain, itself somewhat diversified, on which the railroad runs, the flat-topped hills, the sand dunes, and the drainage channels. Oodnadatta, which is the present end of the railway running from Adelaide, and approximately 700 miles from that city, is situated near ARID PORTIONS OF SOUTH AUSTRALIA. 51 the western edge of the great Australian artesian basin. It is about 100 miles to the north of west of the northern end of Lake Eyre. The most important topographical feature of the immediate neighborhood of the town is the lower plain (Lower Cretaceous) on which it is situated. Outlying members of the higher plain, the desert sandstone tableland, can be seen to the westward (plate 14). These flat-topped hills are a striking feature of the topography of the region. The surface both of the lower and of the upper plain is usually covered with fairly small stones, the gibbers, which were left by the wind after all fine material had been removed (plate 1B). Although a superficial view over the lower plain gives the impression that it is quite flat, when seen more closely such is found not to be the case. It is slightly undulating and there are shallow depressions, apparently wind-scooped, here and there, where gibbers appear to be wanting, and which are but a few inches below the general level of the plain. Such, as will be mentioned below, are in fact miniature oases which support a characteristic vegetation in the midst of a plain otherwise quite barren. So far as I saw, the surface of the flat-topped hills, which are the re- mains of the upper plains, it is much like that of the lower plain, except that the effects of erosion are more marked. There are small depressions in this plain also, and it is covered with gibbers. Possibly, as will be remarked later, the upper plains constitute the most arid habitats of the great basin. The most prominent water-course in the vicinity of Oodnadatta is Neales River. At Oodnadatta the river-bottoms are defined by low banks, and are possibly 6 feet below the general level of the lower plain. But as the river goes southeast on its way to Lake Eyre, the course widens and is not so sharply delimited as it is near Oodnadatta. Neales River and the Macumba River constitute the only rivers leading into Lake Eyre from the western side of the basin. At a point west of Oodnadatta and about a mile distant, the bottoms of Neales River are nearly a mile across. The surface of the bottoms, or flood-plain, is level. There is no continuous and well-defined water-course here, but at various places along the bottoms are depressions, 100 meters or less in length, which contain water after rains. These depressions have the appearance of having been gouged out by water at the time of high floods, which rarely occur. It is doubtful whether the water of the rivers ever empty into Lake Eyre. Just north of Oodnadatta there are flats which may possibly drain into the Neales River; and several miles farther north, as viewed from the summit of O’Halloran’s Mount, there are extensive flats also, with narrow contributory chan- nels, which probably act as reservoirs for the reception and catchment of flood-waters. All of these are probably related to the drainage of the main river. Besides the plains and the drainage channels of whatever kind, in connection with the topographical features of the region one should 52 PLANT HABITS AND HABITATS IN THE also mention the sandhills and accompanying clay-pans situated 2 or 3 miles east and south of Oodnadatta. The sandhills lie on the lower plain and vary considerably in plan, extent, and height. The highest seen were estimated to have an altitude of about 15 meters above the plain and to be about 50 meters in diameter. They are round, oblong, or crescentic in plan. The windward slopes of some are gradual and the lee slopes steep. Ripple marks also are to be seen on most of the dunes. From such features it is concluded that not all of the sandhills are fixed. Adjoining the sandhills, or between them, are clay-pans which vary greatly in size, some being 1,200 meters, more or less, in diameter. The surface is hard and glistens in the sunlight. Although when seen the clay-pans were entirely dry, whenever rains occur they are covered with water which apparently escapes, mainly by evapora- tion, inasmuch as the surface of the pans is of fine texture, apparently silt, which would allow but relatively slow penetration. In one group of sandhills the clay-pans appeared to extend beneath the dune, although this was not surely determined. If true, it would follow that under such a condition there might be important water reservoirs in the sandhills and at no great depth. “Soaks” among the sandhills, as possibly at Ooldea, may have such an origin and struc- ture. Howchin (1909:103) says that sandy ridges and clay-pans form the southern plains of the central portion of Australia and occur largely on the eastern and northern sides of Lake Eyre, and points out that such an inland basin as that of Lake Eyre can not get rid of its worn-down material, the product of erosion, ‘‘such as occurs when the drainage of the country flows into the sea, whilst from a deficiency of moisture, vegetation is scarce and the soil but loosely held together. From this cause the soil and sand are constantly on the move.” From the foregoing sketch of some of the leading physiological features in the vicinity of Oodnadatta it will be clear that the character of the soil varies considerably. On the plains, for example, it appears to be fine sand with an admixture of pebbles of different sizes. The largest of these constitute the ‘‘gibbers,’”’ or desert pavement, which protect the fine soil beneath from being carried away by the wind and from rapid drying out. In the shallow hollows on the plains, as has already been alluded to, there are no “‘gibbers.’’ This also is true on the flood-plain of Neales River, where, however, the small stones are largely wanting and the soil is relatively sandy. On the plains the top- soil appears to be about 50 cm. deep, and on the flood-plain it is 2 meters, and probably much more, in depth. In places on the plains where erosion has occurred and the substratum is exposed, it is seen to be of a white color and often of considerable thickness. How generally this extended was not learned. From the appearance of the material it was assumed to be gypsum, although in certain places where the layer of material was thin, as where it covered ARID PORTIONS OF SOUTH AUSTRALIA. 53 stones, it may well be travertine or desert limestone. The nature of the sandhills has already been mentioned and need not be taken up further. As to the clay-pans of the sandhill neighborhood, it need only be said that in them, as elsewhere where the drainage is poor or lacking, there is an excessive accumulation of salts. In the clay-pans the flood- ing water is at first fresh, but as it evaporates it becomes highly charged, and upon drying leaves the salts in the fine silt or on its surface. In other situations during dry seasons salts are frequently exposed in considerable quantity, especially under conditions of poor surface- drainage. It seems probable, however, that where the salts are not present in the soil solution in sufficient quantity to form a visible crust upon the surface of the ground when the water has evaporated, as on the lower plain especially, yet they are relatively abundant. This characteristic of the soil of all arid regions is probably nowhere more marked than in the vicinity of Oodnadatta. _ So far as regards the water relations of the different physiographical areas mentioned, it would appear from inspection and from the character of the soil that they may be said to be as various as the areas themselves. At the one extreme should be placed the flood-plain of Neales River, which receives not only water directly from the rains, but also such as comes to it from the drainage of the river system of which it isa part. The gradient of the flood-plain is apparently slight, the soil is deep and relatively coarse-grained, and there is no.continuous and well-defined channel at the place visited. All of these features make for the maximum reception, storage, and retention of water. At the other extreme should probably be placed the flat-topped hills, representing the upper plain. Here the water is merely lost and not received by run-off, and the soil is moistened directly by the rains and in no other way. Percolation of the rains into the soil is indeed forwarded by the presence of stones of various sizes, but is relatively slow because of the fineness of the soil. Only in the slight depressions which are the centers of diminutive drainage systems is there appre- ciable accumulation of water after rains; otherwise, there are no proper soil reservoirs for water retention. However, owing to the closely fitted mosaic, the ‘‘gibbers,” the water is conserved very much better than would otherwise be the case. The sandhills and the lower plain occupy an intermediate position with respect to the water relation, although their relative position is not very clearly defined. So far as regards the quality of the soil solution and the retention and storage of the soil moisture, the sand- hills are to be considered less arid than the lower plain, but where the latter joins higher ground, as the upper plain, it may receive seepage water, and in this regard only the water relation of the lower plain can be said to be more favorable than that of the sandhills. Other than this the water relations of the lower plain are about the same as 54 PLANT HABITS AND HABITATS IN THE the upper one. In the vicinity of Oodnadatta, however, the soil solution of the lower plain probably carries a larger percentage of salts, which, as will be remarked below, makes for physiological dryness. The sandhills probably constitute excellent reservoirs for water, and through the rapid drying-out of the surface of the soil the capillary chain is broken and evaporation into the air above the ground is either wholly stopped or at least it is very decidedly retarded and reduced to a quantity that for practical purposes can be neglected. On the other hand, there are no indications of water erosion among the sand- hills, and such probably does not occur. Unlike soil of a fine physical character, as, for example, that of the plains, prompt penetration of the rains takes place even if the soil at the time is air-dry and none is lost by run-off. The dust mulch on the surface is probably 30 cm., more or less, in thickness, and the soil beneath, of homogenous sand, is moist to a depth dependent only on the amount of the rains. The only water lost from such depths is through the plant covering. The sandhills, therefore, have water relations which are considerably more favorable than would appear from a superficial examination. In the other areas, such as have poor surface drainage, the amount of water received and also in certain instances retained must be said to equal, if not to exceed, that of any of the other areas. In such areas the degree of aridity is related to the concentration of the soil solution. In the case of the clay-pans, for example, when first flooded, as previously mentioned, the water is fresh, or but slightly brackish, and this is indicated by the character of the plant and animal life to be found in them at the time. As the water escapes by evapora- tion, however, the salts become more and more concentrated until only such forms as are especially adjusted to withstand such dense solutions are able to survive. Carrying possibly not a small amount of moisture, such areas are nevertheless intensely arid, speaking in & physiological sense. From this sketch of the leading physiographic features of the vicinity of Oodnadatta, it can be seen that there are about four or five well-defined plant habitats. These are the upper and lower plains, the flood-plain of Neales River and its branches, the sandhills, and the clay-pans, with which might be included other areas having poor drainage. As will be shown later, these areas have for the most part characteristic vegetation, as might be expected from their marked differences, some of which have been outlined above. ARID PORTIONS OF SOUTH AUSTRALIA. 55 CLIMATE. RAINFALL. The most arid part of South Australia, as well as of Australia as a whole, is situated in the Lake Eyre Basin, as has already been pointed out. The character of the rainfall for Oodnadatta may be taken as representative of other portions of this intensely dry region. Ap- parently the station with the least average rainfall is Kanowana, where it is 4.33 inches annually; at Oodnadatta, however, the average is 4.85 inches. The monthly average distribution of the rainfall at Oodnadatta is given in table 12, which shows that the seasonal rains are fairly equally divided, although averaging highest in the summer. Thus, at Oodnadatta the average is 1.80 inches in summer, 0.98 inch in autumn, 1.03 inches in winter, and 1.04 inches in spring. As table 12 suggests, the actual rainfall month by month for different years is very unequal in amount, so that a table of averages does not have the significance it otherwise might possess. The seasonal periodicity, so well marked over much of the state, is accordingly not so dependable at Oodnadatta and in the Lake Eyre Basin. However, as has just been indicated, the rains of summer, undependable as they may be and slight as they are, are nevertheless rather more in amount than those of any of the other seasons; and it is not impossible that the adjustment of certain of the native plants to their environment is sufficiently delicate so as to largely hinge on just such relatively small differences in the seasonal rainfall. So far as the character of the individual storms is concerned, relatively few of them are of so large amount as to be mainly lost by superficial run-off. On the other hand, a remarkably large percentage of the rainfall occurs in amounts too small to directly benefit plants. The proportion of the total rainfall which occurs in separate storms, amounting to 0.15 inch or less at Oodnadatta, lies between about 17 and 56 per cent of that for the year, as shown by the rainfall records of 1901-1906, inclusive (table 11). The average monthly effective rainfall for those years is as follows: January 0.06, February 0.64, March 0.03, April 0.32, May 0.24, June 0.29, July 0.34, August, 0.13, September 0.46, October 0.15, November 0.29, and December 0.20 inch. According to these records, therefore, the storms of summer, so far as the effective rainfall is concerned, are considerably less than those of winter. Inasmuch as the actual summer rainfall exceeds that of winter, it would be of much interest and importance to know, when a long series of years is taken into account, whether over such long period the effective rain- fall would hold the relation above shown to be the case during the years 1901-1906, inclusive. 56 PLANT HABITS AND HABITATS IN THE TEMPERATURE. As would be expected from the geographical relations, including the altitude of Oodnadatta and of the Lake Eyre Basin in general, extreme temperatures are the rule. The summers are very hot, the daily variation in temperature is relatively great, and the minimum tem- peratures of winter are lower than might be expected for the latitude. These features, of course, are to a large degree associated with a low rainfall and low relative humidity. The stabilizing influences of water and of water-vapor are largely wanting. Thus the air, and objects on the surface of the soil, including the vegetation, are quickly heated under the very favorable conditions of insolation obtaining, and conversely the radiant energy rapidly escapes with the passing of day. For these reasons there are, as table 7 indicates, many days during the year when the air has a temperature of over 90° F., and, on the other hand, many nights when an air temperature of less than 40° F. is recorded.’ In fact, in every month, except June and July, a temperature of 90° F., or above, is reached, and in 7 months of the year the night temperatures drop to below 40° F. The extremes thus far recorded at William Creek, about 100 miles from Oodnadatta, are 119° F. and 25.3° F. for a period covering 28 years. A summation of the mean monthly temperature for Adelaide, Port Augusta, Farina, and William Creek shows that the total amount of heat received increases at these stations in the order given: Adelaide, 747.8°; Port Augusta, 793.9°; Farina, 808°; William Creek, 822.6° F. It is not likely that the temperatures attained at Oodnadatta, or in the Lake Eyre Basin, are so high as to materially injure the perennial vegetation. On the other hand, a large amount of heat is required for the growth and development of such plants. Aside from this feature, it is probable that a leading effect of high temperatures on the vegeta- tion, especially perennials, is an indirect one, namely, the effect of such temperatures on the relative humidity of the air, and hence on the rate of evaporation. The capacity of the air to absorb water-vapor increases directly with the temperature. Hence, it is found that the drying power of the air is markedly greater where the temperatures are higher, other conditions being equal, than where they are less. The following will illustrate the point: The average mean tempera- ture for Adelaide is 63° F. (17.2° C.); for Port Augusta it is 66.2° F. (19° C.); for Farina it is 67.3° F. (19.6° C.); and for William Creek it is 68.6° F. (20.3° C.). The weight of aqueous vapor in a cubic meter of saturated air for these temperatures is as follows: 14.4, 16.1, 16.6, and 17.4 grams, from which it is seen that an amount capable of producing saturation under the mean-temperature conditions at Adelaide, and assuming the equality of other conditions, would not bring about satu- ration at William Creek, but only about 83 per cent of saturation. 1 See the records for William Creek, table 7. ARID PORTIONS OF SOUTH AUSTRALIA. 57 GENERAL FEATURES OF FLORA OF SOUTH AUSTRALIA. THE NORTHERN PORTION OF SOUTH AUSTRALIA. The flora of the northern portion of South Australia has a marked xerophytic stamp. As much, however, can be said for most of the perennials, if not all, in other portions of the state having a much larger rainfall. As Taylor remarks, possibly with another idea in mind, “St differs in degree according to the rainfall, but not in kind” (1918:89). This is possibly the most remarkable feature of the vegetation of this in- tensely dry region. That is to say, there are possibly no perennial forms peculiar to it as such—for example, as the different types of succulents developed in other arid and semi-arid regions, or possibly in regions which are desertic. There also are apparently no deciduous species. In other dry regions some of the species are deciduous. For example, the acacias of the Sahara, as well as those of the south- western United States, have foliage which falls away with the seasons; and in the arid southwestern part of the latter country a species (Fouquieria splendens) occurs which is deciduous with respect to the occurrence of the rains, losing and forming the foliage perhaps several times during the year, in a manner directly connected with the rains and the intervening dry spells. Therefore, there may be possible not a little variation in plants in extremely arid regions. Thus, although the perennial flora of the far north has apparently little or no striking peculiarities which set it apart from that of the regions of South Australia more highly favored with rainfall, there is still not a little diversity in it. Among the most interesting trees of the dry region are several species of Acacia which show important morphological, as well as physiological, differences. Certain of the hakeas exhibit remarkable differences as between species and marked adjustment to an arid environment. Of the shrubs, those of the genus Hremophila are of especial interest. Out of the 39 species, including those now put under Pholidia, Tate gives 17 species as occurring in the northern portion of the state. Of the shrubs, this is the desert plant par excellence. But the halophytes of whatever species constitute the most prominent element of the flora of this region. Tate gives 54 species of the Chenopodiacee and Amarantaces from the far north, most of which are highly salt-resistant. There are a few succulent annuals, or half shrubs. Among these are species of the Zygophyl- laces and Calandrinia balonnensis (Portulacez) ; the latter bears large, succulent leaves and is much sought after by animals. Stuart’s pea, Clianthus dampieri, is one of the most showy of the annuals. Of the grasses, among the most frequently met are the ‘‘bunch” forms, as Triodia irritans and Spinifex paradorus, which are widely distributed in the interior of the continent. In addition, according to 58 PLANT HABITS AND HABITATS IN THE Tate, there are 55 species of grasses, mainly annual, which are to be found also in northern South Australia. These for the most part are to be seen only after rains, possibly mainly those of summer. VEGETATION OF THE LAKE EYRE BASIN. As one enters the Lake Eyre region by the railway, the general im- pression is had, which becomes stronger upon further acquaintance with the region, that. the total amount of perennial vegetation has suddenly become less, despite the large halophytic population. More- over, the larger elements of the vegetation are more and more re- stricted to the water-courses and a relatively small amount is to be seen venturing away from them. The differences noted are possibly most marked with the tree flora. Clearly such a generalized statement is difficult, if not impossible, to demonstrate. One would, of course, expect a falling off in the amount of vegetation in a region having only 5 inches as opposed to one having nearly 9 inches—for example, that about Copley. But the point I wish to make is that the differ- ence is greater than the difference in the recorded rainfall would lead one to expect. To illustrate the idea crudely, we may suppose the amount of vegetation and the amount of rainfall to be represented in a figure by two curves. The curve representing the amount of vegeta- tion is assumed to decline parallel to that representing a decrease in the precipitation. However, with the decreased: rainfall there comes a rapid increase of other climatic factors making for aridity, and espe- cially there occurs, in the Lake Eyre region, a relatively large proportion of non-effective precipitation. We therefore get a diverging of the vegetation-rainfall curves which at the last is very sharp. The sudden falling of the vegetation curve I believe to be directly related to the large percentage of non-effective rainfall, although with it there naturally must be associated an intensification of other environmental physical factors working toward the same end. In other words, the effect is as if in place of contrasting regions having about 5 and 9 inches of rain, we are contrasting regions whose effective rainfall may be said to be about 3 and possibly 8 inches. VEGETATION AT OODNADATTA. A general idea of the amount of the perennial vegetation in the vicin- ity of Oodnadatta may be had by ascending O’Halloran’s Mount, which is about 4 miles to the north of the village (plate 14). This low hill is evidently a detached portion of the upper plain and it com- mands a very good view of the desert. As one surveys the surrounding country he is at first struck with the paucity or rather with the want of plant covering. In every direction the wide-spreading plains and flat-topped hills appear quite barren. The gibbers glisten like polished mirrors in the sun, but otherwise a monotonous reddish-brown color prevails. There are thus apparently no plants to enliven the scene. ARID PORTIONS OF SOUTH AUSTRALIA. 59 The effect is as if no living thing ever found or ever could find a lodg- ment and an abiding-place in the expanding waste. More careful examination, however, reveals the presence of plants. Across the lower plain, for example, there extend narrow ribbons of vegetation. These converging make gray-green bands and mark the drainage channels where only masses of vegetation are to be found. The in- consequential plant life of the plains as a whole does not appear on such a general view. Slight as is the present-day vegetation of the region, it perhaps is a matter for surprise that there is so much. Aside from the unfavorable physical environmental conditions which make for few plants, there are also biotic environmental elements which are relatively very destructive. Thus the region has supported and now supports an aboriginal population which depends wholly on natural products, animal and vegetable, for its living; and relatively recently the white inhabitants have drawn heavily on every useful native plant. Herbivorous animals of many kinds have also made destructive inroads on the vegetation. Just how all of these biotic factors have affected the flora as a whole will, of course, never be known, but that they are, and long have been, of great importance in many ways can not be seriously questioned. VEGETATION OF THE PLAINS. The vegetation of the plains was observed in four separate localities on the upper and on the lower plains. So far as these situations are concerned the perennial vegetation was found to be very meager indeed, but there were a few annuals which had sprung up following a small rain of the week preceding my visit. Both classes of vegetation, how- ever, were for the most part restricted in distribution to the small and slight depressions on the plains or to the diminutive drainage channels frequently associated with them. Taken as a whole, aside from such relatively favorable situations as above mentioned, the plains gave little suggestion of plant life. Among the annuals were found Bra- chycome ciliaris and Senecio gregorii, and among the perennials Bassia lanicuspis(?), which occurred very sparingly, and species of Eremo- phila. The Eremophilas were the most striking and interesting of the perennials. Of these, Hremophila freelingii (plate 2, A and B; plate 3B) appeared to be most numerous and was found on both plains. Plate 2B shows a typical example of the species in place on the upper plain. It is a much-branched shrub about a meter in height and bears dull- green leaves, narrowly oblong in form, mostly at the end of the branches. The species is fairly rare on the plains, but in and along the drainage channels the population is more abundant and the indi- viduals of a larger size than on the plains proper. In addition to the species referred to, HE. latrobei (plate 3c) has a distribution about like that of E. freelingit, but is possibly not so abundant. Ina wash at the 60 PLANT HABITS AND HABITATS IN THE south base of O’Halloran’s Mount may be found LE. neglecta (plate 5B). In the same situation a few examples of an undetermined half-shrub and several annuals, also undetermined, were seen. Although the species mentioned in the preceding paragraph are to be found very sparingly on both the upper and the lower plains, the former was especially in mind when writing the description; and, so far as one can judge, the vegetation of this plain, at least in the vicinity of Oodnadatta, has been little affected by the presence of man. No trees occur on it and the shrubs of the plain are apparently not found useful either by man or beast. On the lower plain, however, there are indications that formerly the number of individuals may have been greater than now. Dried remains of salt bushes are to be found here and there, and persons familiar with the region say that such forms were formerly fairly abundant on this plain. It seems prob- able, also, that after heavy rains, which occur at widely separated intervals, the face of the lower plain may be clothed with annuals, including grasses, and that they may be of such size and abundance as to completely hide the surface of the ground; and, although there are no trees on this plain, they may be found along water-courses con- necting it with the upper plain—for example, Acacia cambadgei occurs sparingly in such a drainage channel leading from the upper to the lower plain at a place about 4 miles to the west of Oodnadatta (plate 3a). Eremophila neglecta and FE. latrobei also occur in such situations. VEGETATION OF AND ABOUT THE SANDHILLS. Across the lower plain and about 3 miles to the east or southeast of Oodnadatta is an area where sandhills and clay-pans are the most characteristic features of the physiography. On the way to the hills the plain is found to sustain a sparse population of saltbushes of various kinds which were not studied particularly, and viewed from the plain the distant sand ridges appear to be fairly well covered with vegetation. Upon drawing near the hills a comparatively large number of small trees and shrubs are also to be seen. But the near view of the sandhills reveals the fact that the vegetation is very diffuse and composed almost exclusively of a single form. The clay-pans which lie between or by the sandhills were found to be devoid of vegetation, although when flooded such a form as the ‘‘nardoos,”’ Marsilia spp., may be found. On the rims of the clay-pans saltbushes occur, but not in abundance. The characteristic species of the sandhills is the sandhill mulga, Acacia linophylla. On the dunes the species has the habit of a large shrub, but occasionally a central stem isformed. It is about 5 meters in height. The habit of the species is shown in plate 4B. The leaves, or rather phyllodia, are relatively few in number (plate 5a), about 15 em. in length and about 2 mm. wide. They are fairly rigid and have an attenuated but not a spiny tip. At the time of my visit, ARID PORTIONS OF SOUTH AUSTRALIA. 61 July 11, the species was flowering and the plants were covered with aaots in immense numbers, which appeared to be seeking the nectar of the flowers. The small size of the trees and the large amount of dead wood attested to the very difficult conditions under which the species was living. So far as the roots could be seen, it was found that many were very near the surface, being often exposed through wind erosion. Although no long horizontal roots were found, it is known that in the mulga and under analogous conditions very long superficial roots are formed. Owing to this fact and to their being of a fairly uniform size, they are used by the aborigines in the making of spear handles. Although the mulga was the only woody perennial found in the sand- hills, a species of ‘‘spinifex’’ or bunch grass appears here and there, and is one of the most characteristic plants of the arid interior. On the plain and at a relatively short distance from the base of the sandhills some shrubs and trees are to be found. In places they are relatively abundant and also fairly large. The shrubs are wholly or mainly Eremophila neglecta (plate 5B), about 2 meters high and well clothed with leaves. The trees are of but a single species, “dead finish,” Acacia tetragonophylla, which is of wide occurrence in the drier portions of South Australia. The habit of the species is shown in plate 44 and plate 64. A leading and striking characteristic of the latter species lies in the small size of the phyllodia, which are borne in relatively large numbers, fairly well distributed on the smaller branches. The relative abundance of the species at the base of the sandhills is apparently attributable directly to the comparatively favorable water relations of the place. . The soil is fairly deep and covered with a sand mulch blown from the neighboring hills. There is a substratum of a white and not especially hard material which has the appearance of travertine limestone and indeed may be that. Whether this prevents the deeper sinking of water or absorbs it, to be yielded again to the plant-roots, was not de- termined. In case the substratum underlies the sandhills as well, which may be the case, a further reason for the relatively abundant vegetation at their base would be found. So far as the substratum is concerned, it was found to crop out along the path to the hills, and in such cases it formed a hard surface quite as in the “‘caliche,”’ or desert limestone, in the more arid parts of the United States, in southern Algeria, and other dry regions. Moreover, under such conditions it is fairly impervious to water. In the habitat at the base of the sandhills there were found a few specimens of the phanerogamous parasite Loranthus exocarpi on Eremophila neglecta. The paucity of this type of vegetation in the vicinity of Oodnadatta forms a striking contrast to the conditions obtaining at Copley, where the mistletoes are very numerous. 62 PLANT HABITS AND HABITATS IN THE Although there is thus seen to be a comparatively abundant peren- nial flora at.and near the sandhills, yet the largest number of woody plants about Oodnadatta, especially the largest number of trees, is to be found in water channels of whatever kind, either Neales River or washes at the bases of the flat-topped hills. As viewed from the vantage-point of the upper or even the lower plain (plate 5c), the Neales River bottoms has a considerable tree population, and par- ticularly a large number of shrubs. The present-day shrub and tree population of the flood-plain of the river can not be taken as indicative of what must have been the condition in earlier times. The inhabitants of the nearby town have actively removed all trees large enough to supply fuel. The shrubs, however, have probably very largely escaped such inroads. The same remark would also apply to the neighboring washes, where the water relations are suitable for some tree-growth. Of the species now to be found on the flood-plain of Neales River, the most conspicuous are Eucalyptus rostrata (plate 4c) and species of Acacia, among which are A. cambadgei, A. tetragonophylla, and A. stenophylla. There are also several species of saltbushes, several shrubs whose identity was not determined, and some annuals. Of the trees, Acacia cambadget is possibly the most numerous. This is the “stinking” acacia, or “gidya,’’ the name given it by the aborigines, which is said to refer to an edible (for aborigines!) larva which is to be found beneath the bark. The tree attains a height of 10 meters or more and has a compact habit of growth. Like the other species of Acacia growing in the far north, true leaves are not present on the mature plant, but their place is taken by phyllodia. In the case of the gidya the phyllodia are fairly large as well as abundant, so that a distinctly leafy effect is produced (plate 64) and the shade cast is dense—a rare occurrence in desert plants. A. tetragonophylla is also a small tree, but with a foliar habit quite different from the species last mentioned. Its phyllodia are needle- form, 8 to 28 mm. in length, and may occur in groups of a few each. They fall away fairly easily and the ground beneath the trees is usually thickly covered with them. It appears probable that the shedding of the phyllodia should-be considered a very effective means of re- ducing the evaporation surface and hence of cutting down excessive loss of water during especially severe drought. The foliar habit of A. stenophylla is unlike that of either of the species above mentioned. The species is a small tree or a large shrub at Oodnadatta and true leaves are formed on young shoots, although they are soon cast off, phyllodia taking their place (plate 68). The phyllodia are various as to form and size, but in general they are long, narrow, and even linear in the extreme form. In the portions of a plant which are sub- jected to the most intense illumination the phyllodia assume an up- right position, but where shaded they are pendant. ARID PORTIONS OF SOUTH AUSTRALIA. 63 The root-habits of certain species growing along the washes or by the separate basins on the flood-plain of Neales River were studied as far as possible. Observations of roots exposed by the washing away of the banks indicate that the general course of the roots of the trees was similar. Taking Eucalyptus rostrata and Acacia cambadget, especially, it was found that a prominent portion of the root-system is made up of large horizontal members which may extend for a long dis- tance from the central stem (plate 7a). Thus the superficial roots of Eucalyptus were seen to reach out 9 meters or more and to lie at a depth not exceeding 60 cm., although the distance beneath the surface of the ground was usually much less. In addition to horizontal roots, vertical ones were found, but none lying at angles between. In the case of several shrubs whose roots were observed, it was found that an analogous condition obtained—that is, there was evidence of direct reaction to the type of rainfall which is especially characteristic of the region. In Acacia stenophylla shoots arise from the superficial roots (plate 7B), with a resulting and characteristic dense thicket formation. Such vegetative reproduction was seen in several species in the other regions visited, especially at Quorn. 64 PLANT HABITS AND HABITATS IN THE © THE COPLEY ENVIRONMENT. PHYSIOGRAPHY. The topography and hence the plant habitats in the immediate neigh- borhood of Copley are extremely varied, owing mainly to the relation of the region to the Flinders Mountains. The main portion of the north end of the Flinders lies to the east, between Copley and Lake Fromme, but a lesser part is between the village and Lake Torrens on the west. Copley is thus in the bifurcation of the range, with out- lying hills and low mountains within one mile or more on either side. The plain at Copley constitutes little more than a valley, but it widens to the north, becomes lower in altitude, and soon constitutes a leading feature of the topography. Copley lies between the 500-foot and 1,000-foot contours and hence is near the upper level of the Cretaceous beds (Taylor, 1918:87). In the Flinders to the east are peaks which are among the highest of the entire ranges, including Mount Serle, the Freeling Heights (3,120 feet), and Mount Benbonyanthe (3,470 feet). The altitude of the division of the Flinders to the west of Copley is 2,000 feet or less. The observations upon which the present study is based were made within a 15-mile radius of Copley and were confined to an altitudinal range of approximately 1,000 feet. Within this small region, however, there is to be found a bewildering array of hills, valleys, slopes, washes, and flats which would be beyond the pur- poses of this study to describe accurately or in detail. The device will be resorted to, however, of attempting a characterization, so far as possible from the data at hand, of such of the habitats as appear to be of most interest in connection with this study. The Flinders on either side of Copley are for the most part of the Cambrian age, but there are also Mesozoic rocks. Thus the slate hills with vertical strata to the east of the town are possibly of the former, and the table mountain just southeast is of Mesozoic. The latter is part of the desert sandstone previously seen at Oodnadatta and which formerly extended over most of central and northern Australia (How- chin and Gregory, 1909:93). Thus the underlying rocks, and conse- quently the soil derived by disintegration from them, have a widely different physical character. Future ecological studies in the vicinity of Copley may well correlate the distribution of the vegetation with the nature of the soil and with the rocks from which it has been derived, as has been done by Osborne in the Mount Lofty Ranges near Adelaide (1914:114). Such observations as were made by me on soils will be given, to- gether with the running account of the most striking features of each of the habitats studied. It will suffice to say, as would be supposed from what is known of the local geology as above suggested, that the soils are very various. For example, on the plain about the village ARID PORTIONS OF SOUTH AUSTRALIA. 65 there is a large amount of sand, some of fine grain.and easily shifting with the wind. For this reason one finds fences covered with sand drift and sandy hillocks. There are also a few moving dunes of small size. In the low hills the soil appears to be all or part clay, or at least of very fine structure, asa sandy loam. ‘There is evidently very little humus. So far as the depth of the soil is concerned it can be said to vary greatly. On the plain near Copley it is4 metersormore. Among the hills, on the other hand, there are many areas where the rock is barely covered with soil, and of course there are many outcrops in which soil is entirely wanting. In any region, such, for example, as that about Copley, where a small rainfall is a prominent environmental factor of plants, the nature of the soil, including its depth, is of very great importance in shaping the character of the plant covering as well as its distribution. This fact is largely based on the circumstance that the soil constitutes the sole reservoir for the reception and storage of water. Owing to the broken topography the water-courses of the region are not only numerous but also in many regards extremely varied. In some the gradient is marked, in others it is relatively slight. The channels may be very well developed and hence considerably below the adjoining plain, or slope, as the case may be, or the converse may be true. The leading drainage channels of all sorts are contributory to Leigh’s Creek. This stream, dry much of the year, takes its origin a few miles to the east of Copley in the Flinders Ranges, and running northward it joins the Fromme River and ultimately discharges into Lake Eyre. To the south of Copley, and a little beyond the region studied, the streams run in the opposite direction and into or toward Lake Torrens. Although the drainage is generally very well defined, there are relatively small separated basins, mainly on the Copley plain, which have inadequate drainage or no surface outlet. In such depressions there is a large accumulation of salts. It will be seen from the above sketch that the leading physiographic divisions or units can be said to be the hills, the Copley Plain, and the washes. For convenience, the hills will be divided into (1) the Mount Deception Range, such of the Flinders Ranges as lie to the west of Copley, and (2) Table Mountain and Mount of Light, with the nearby hills, to the east. For the purpose of further distinguishing the locali- ties to be referred to below, they will be mentioned in reference to the roads which lead to them, as follows: North of town: Myrtle Springs road crosses the plain northwest of the village, goes through the low hills to the east of the Mount Deception Range, and finally through that range. Yudnamutana road goes north and northeast of town, passes along the western side of the Flinders Ranges, and at length penetrates them; this road also crosses the plain and runs over low hills, where it passes out of the Copley vicinity. 66 PLANT HABITS AND HABITATS IN THE East of town: The Mount Serle road crosses the plain for about a mile and then enters the undulating hill country. It passes to the north of the Mount of Light, and winds between lower hills until it also passes beyond the Copley district. South of town: The Beltana road goes in the midst of the Copley Plain for several miles and at length makes its tortuous way through the Flinders to Beltana and southern towns. CLIMATE. RAINFALL. The climate of Copley can be characterized as arid, with cool winters and hot summers. Owing to its position, about midway between the southern extension of the summer rains and the northern limit of those of winter, some precipitation is to be expected in every month of the year; for the same reason, the rainfall of any season is extremely variable. In table 13 are presented rainfall data for a period of 35 years, supplied by the Commonwealth Bureau of Meteorology. It will TaB.LE 13.—Average rainfall at Copley (Leigh’s Creek), South Australia, based on records for 35 years. Jan. | Feb. | Mar.| Apr. | May.|June.|July. | Aug. | Sept.| Oct. | Nov.| Dec. | Year. be seen that the extreme variation in yearly precipitation is from 1.54 to 15.64 inches. ‘The seasonal averages are as follows: Summer, 1.95; autumn, 2.4; winter, 2.3; and spring, 1.76 inches. Another feature of the rainfall at Copley does not appear in the table, but was seen in daily reports very kindly put at my disposal by Mr. Bromley, meteor- ologist at Adelaide. This relates to the maximum rainfall for one day. Records covering 6 years, 1901 to 1906, inclusive, were examined. The greatest rain for 24 hours during this period was on December 28, 1903, when 3 inches were reported. This, it will be noted, is about twice as much as occurred during the year of minimal rain as given in the table. ; The rainfall at Copley, however, as has been shown in another place, is not always in amount sufficient to be of direct use to plants. In fact, it was shown that for the years 1901 to 1906, inclusive, approximately 19 per cent of the rain occurred in showers of 0.15 inch or less. The amount of 0.15 inch rainfall was placed as the minimum after the following observations had been made at Copley, supplemented by others at Ooldea. On August 20, after a slow rain amounting to 0.21 inch, fairly fine soil containing some sand on the plain was found to have been moistened to a depth of 4 cm., and coarser soil also on ARID PORTIONS OF SOUTH AUSTRALIA. 67 the plain was wetted to a depth of 8to9cm. The observations were repeated a few days later with substantially the same results. In both instances the rains were followed by drying weather, so that a large percentage of the moisture taken into the soil must have been lost quickly by evaporation. It was assumed that had the rainfall been 25 per cent less, the penetration would also have been very much less, and the amount of water lost by evaporation would bave been rela- tively greater with respect to the amount recorded. Therefore, under such conditions there would be very little, if any, available for the use of plants, more especially for the use of perennials. However, as will be shown below, there is no doubt that, given favorable soil conditions, a rain amounting to 0.20 inch penetrates the soil suffi- ciently to moisten the horizon occupied by the roots of many annuals and also by a portion of the horizontal roots of certain perennials. TEMPERATURE. The temperature records for Copley were not examined by me, but the Commonwealth Bureau of Meteorology very kindly supplied the records for Farina, 40 miles farther north, which have been accepted as illustrating the conditions at Copley. A summary of these records for Farina has been given in table 7, which shows that the highest shade temperature reported for the station, for a period of 28 years, was 119° F. The absolute minimum was 25.3° F. The annual range may be considerable; for example, in 1912 the minimum reported was 27.3° F. and the maximum was 118° F., a total range of 90.7° F. During the same year the mean daily range was 18.6° F. and the maxi- mum range observed during the course of one day was 53° F. The mean yearly temperature at Farina is more than 10° F. below that at William Creek, as representing Oodnadatta, and it is nearly 4° higher than that at Adelaide. For other details, reference can be had to table 7. VEGETATION OF THE COPLEY REGION. The most striking features of the flora of the lowlands between Marree (Hergott Springs) and Copley are its abundance and the fact that it is practically all of one type, namely, halophytes. Viewed from some distance, the relatively slight individuality of the species is lost and the plains appear closely covered with a uniform growth. In and along washes and on the slopes of the hills, however, the vegetation is generally composed of sclerophyllous shrubs and trees which vary greatly in size, form, and occurrence. Even at this day, and in spite of the heavy demands of whatever nature made on the native vegeta- tion by the white inhabitants, there is a remarkable wealth of plants, not only as regards species but also as to number of individuals; and it is to be remembered that the rainfall of the region is by no means heavy. At Marree it is 6.08 inches and at Copley 8.40 inches, of 68 PLANT HABITS AND HABITATS IN THE which about 20 per cent occurs in amounts too small to benefit plants directly. The following glimpses of the vegetation, its leading character, and occurrence, taken at a few well-marked localities close to Copley, should supply in a broad way sufficient data to reconstruct its general features. For convenience the description of the vegetation at Copley will be grouped around the following physiographic units (habitats), which will be made the centers of ‘communities,’ but they are probably not of equal value, as will be at once apparent: (1) ‘‘Alkali’”’ plains, or lowlands, including slopes and benches where halophytic vegetation points to an excess of salts in the soil. (2) Low hills and the slopes of higher hills, or low mountains, as of the Mount Deception Ranges west of Copley and the lower portions of the Flinders Ranges to the east. This includes the subaerial delta fans(?) on either side of the Copley Plain. The leading features of the hill habitat can be said to be (a) possible variations in aspect; (6) a relatively good water relation through altitude and relation to higher hills or mountains; and (c) presence of rock outcrops with correspond- _ ing paucity of soil whose nature is determined by that of the rocks. (3) The Mount Deception Range immediately west of Copley. (4) The washes or streams in the hills and lowlands. VEGETATION OF THE ‘ALKALI’? PLAINS. The most important component of the vegetation of the lowlands, particularly of the Copley Plain, consists of halophytes in bewildering variety. These in large part are very similar in size and in general appearance and it needs fairly close study to distinguish many of them. They are 50 cm. more or less in height and usually of a grayish-green color; but some are small annuals and some occur in communities having a common ancestry, which are of considerable extent. Asa whole, the halophytic flora of the far north is of great economic importance, inas- much as it constitutes practically all of the forage of this vast region, being thus the basis of the pastoral industry.* The following species of halophytes were observed on the Copley Plain or on slopes contiguous thereto: Atripler spongiosum, A. vest- cartum, A. quinii, Kochia pyrmidata, K. planifolia, K. cannoni, K. villosa, K. decaptera, K. eriantha, K. sedifolia, Enchylena tomentosa, Bassia lanucuspis, B. paradoxa, and Salicornia tenuis. In addition, “‘salt-loving”’ species of other families were found, among which were * In 1912 the Central, Lower North, and Upper North of South Australia supported the follow- ing number of live stock: cattle, 199,727; horses, 197,139; sheep, 2,674,856 (Handbook of South Australia, 1914, p. 148); there are also several thousand camels. Besides the domestic animals which derive their entire subsistence from the native flora, there should be mentioned the native and introduced wild animals which also subsist wholly on it. Of these, the most destructive are the rabbits. They occur during favorable periods in countless numbers and work great harm toa very wide class of vegetation, including that useful to sheep, cattle, horses, and camels. ARID PORTIONS OF SOUTH AUSTRALIA. 69 Nitraria schebert, Zygophyllum crenatum, Z. fruticulosum, and Z. prismatothecum.* The vegetation of the Copley Plain at present is very sparse, in places wanting, but in some places it is fairly abundant. Where absent, as on the higher portions, the dead stumps of shrubs show that in former times it also was covered by plant growth. But in a few small basins without surface outlet, where there is an excessive accumulation of salts, the edaphic conditions do not now and have not in recent times permitted the presence of plants. On the lower and also well-drained parts of the plain, on the other hand, species of Atriplex and of Kochia especially occur in good size and fair abundance. Plate 7c shows their leading characteristics. In order to learn the density of the population of saltbushes, a census was taken at the base of Table Mountain, about a mile south of Copley, at the upper limits of the plain where it joins the slope of the hill. Here were Atriplex and Kochia, with some specimens of a small and undetermined grass, a geranium, and a mustard. The area observed measured 10 meters square. In the first square 40 shrubs were counted, of which 36 were Atriplex. In the second square, somewhat nearer the hill, 42 shrubs were found, of which Zygophyllum fruticulosum was the dominant form, the balance being Salicornia and Kochia. Plate 84 shows the character of the plants and of the habitat. Somewhat farther (possibly 200 meters) on the plain and well away from the slope, Salicornia tenuis was dominant, with Zygophyllum present in fewer numbers than in the second square. At this place there were relatively large areas quite bare of halophytes other than Nitraria scheberi, which was growing on sandy hillocks. On the surrounding ground, where no perennials were found, there was a very dense covering of very small plants. These were mainly an annual species of Atriplex, which was largely in fruit. In such an area, | square meter, 1,200 individuals were enumerated. One of the characteristic features of the halophytic flora of the plain is the frequent occurrence of such forms on hillocks of various sizes, mostly small, which have been built up around them through wind action. Such hillocks were observed in connection with species of Zygophyllum, Atriplex, Kochia, and Nitraria. In the first three in- stances the result seems to be mainly incidental to the interference with the wind flotation of the sand, but in the last case there is a nice accommodation to the heaping sand on the part of the plant which merits some attention. Nitraria scheberi is a low shrub of rather diffuse habit. The branch- lets are often spinose and rigid. The habit and habitat of the plant occur characteristically on slopes, but also were found at the edge of the salt plain; hence their inclusion among the halophytes. It should also be said that certain species of Kochia especially occur where the soil does not appear to carry an excess of salts. In general this study does not undertake to distinguish between the species as to their toleration for salts. 70 PLANT HABITS AND HABITATS IN THE are shown in plate 8B. At the climax of its development a typical Nitraria hillock measures about 3 by 13 meters in horizontal plan. The height rarely exceeds 1.5 meters and is usually somewhat less. The origin and growth of a hillock, mound, or dune seems to be about as follows: In its earlier condition the seedling Nitraria has a well- marked main stem with branches fairly well raised from the surface of the ground. In this stage it does not interrupt the movement of the drifting soil. Early ‘in its development, however, the lower branches come to lie on the surface of the ground and thus provide an obstruction to free movement of the sand. The result is that such horizontally disposed branches become covered by soil; they develop roots and send up shoots which in growing maintain their position above the accumulating sand drift. Thus the plant adds to its size on every side; the central portion grows actively and keeps above the surface of the accumulating soil, and the characteristic hillock, ever increasing in diameter and height, results. The long diameter of the mounds seems to be usually at right angles to the direction of the prevailing wind, southeast and northwest, al- though there may be exceptions to this. The ‘‘runners”’ which extend the hillock colony, as well as characteristic features of such a colony, are shown in plate 8c. When the maximum size has been attained the hillock begins to break up in the following way: Portions of the plant which are about centrally located die out, for various reasons, and the soil about them begins to be removed, or possibly the converse is the initial step. However this may be, the soil placed in the midst of the dune is gradu- ally removed by wind action until the original level has been attained— a process of “‘base-leveling’’—and the two ends are completely sepa- rated. Thus two daughter mounds ultimately result in the natural course of disintegration of any Nitraria dune. From observations on the character of the flora which occurs on and around the base of a Nitraria mound, it seems probable that the accu- mulated soil of the hillock does not carry the large amount of salts found in the soil characteristic of the plain proper. Thus a greater proportion of nonhalophytic annual growth is found on the hillocks than on the lower contiguous ground. Root-HaBits oF PLANTS OF THE PLAINS. The roots of typical annuals and of a few perennials (halophytes) which were growing on the lowlands in the vicinity of Copley were examined. The depth of soil in all cases was greater than the depth attained by the roots, so that the maximum penetration took place. The roots of the perennials were for the most part studied along washes where they had been exposed, although in a few instances they were examined remote from drainage channels of whatever sort. The roots of annuals were seen on the open plain, but mostly where diminutive ARID PORTIONS OF SOUTH AUSTRALIA. ral hollows made possible relatively favorable water conditions, with results which can be briefly given. In one place the roots of several annuals, mainly of species I did not know, were examined. The shoots were 2 to 7 cm. long and for the most part either in flower or in fruit, and therefore fully matured specimens. A cruciferous species had a tap-root which varied in length from 4 to 8 em. Another species, with a rosette, had a prominent tap-root 7 to 9 cm. long. Of all the annuals seen at this place none had roots which penetrated over 12 cm. In another locality, about 2 miles east of Copley, but on a sloping plain, the roots of several annuals were dug up. Among these were several specimens of Zygophyllum crenatum. In specimen a the shoot consisted of 4 leaves and was 6 em. long, and the tap-root was found to be 8 cm. in length. It was little branched. In specimen 6 the shoot bore 6 short leaves and was 4 cm. in length. The tap-root was over 13.5 em. long. Specimen c had a tap-root over 8.5 cm. in length. Roots of Geranium pilosum, which is one of the most common annuals on the plain, and which was growing abundantly near the Zygophyllum, were also examined. In specimen a the small rosette had 7 leaves and flowers had been formed. The leading root was a tap-root, but this bore numerous fine laterals. It penetrated to a depth of more than 8.5 cm. In specimen 6 the shoot was somewhat larger and the tap- root was traced to a depth of over 11 cm. In another situation, where the plain was in part bare but where there were also slight and in- conspicuous hollows, there was a sparse growth of annuals. These were of various species, including an undetermined grass about 3 cm. high, a species of Geranium, a Zygophyllum, and some crucifers. As a rule the shoots of the annuals, which were fully matured, did not exceed 4 cm. in height. Although there was found to be a certain specific difference in the root-systems of these plants, it was learned that in no case did the penetration exceed 13 cm. and it was mostly less. In the case of the Zygophyllum, prominent laterals 8 cm. long, more or less, were found. As a result of the observations on the winter annuals of the Copley Plain, it can be said that, even under such relatively favorable local water relations, the roots do not penetrate the ground over 10 to 13 cm. and for the most part they lie closer to the surface than this. The roots of species of Atriplex and of Kochia, which were also situated on the Copley Plain, were examined in several specimens, and a certain parallelism in development was noted, similar to the development of the root-systems of most perennials, other than halophytes, which were seen. One of the features of the plain referred to is the presence of various dead perennials, mainly halophytes. Whatever may have been the cause of the death of the plants, it was noted that the root-system remained almost intact. The roots of 72 PLANT HABITS AND HABITATS IN THE such dead specimens had certain points of interest. It was found that soil no longer remained around the base of the halophyte when dead as it does around the living plant, but it is removed from the root-crown, exposing the origins of the superficialroots very completely. In every instance the dead root-crown is surrounded by a radiating circle of small roots which start away in a fairly horizontal direction. In one plant of undetermined species, where the superficial roots were exposed in the manner indicated, 23 were counted, all very close to the surface, but 3 or 4 were found more deeply placed. Living plants mostly were studied and the results, as above sug- gested, were very uniform. A Kochia which was growing on the plain at some distance from any wash was carefully removed from the ground. A tap-root which penetrated 39 cm. into the soil was seen to be fairly moist. All of the laterals, of which there were several, arose at a depth of about 10 cm. from the surface and were traced more than 60 cm. from their place of origin. A species of Atriplex, situated by the side of a small wash and not far from the Kochia just mentioned, was found to have a prominent tap-root and prominent laterals which ran in a fairly horizontal direction, and not far beneath the surface. A species of Salicornia, growing close by, had a root which penetrated to a depth of 15 em. and then, turning sharply, ran horizontally. Sev- eral large horizontal roots arose within 10 em. of the surface of the ground and many small laterals took their origin at about half that depth. So far as this species is concerned, therefore, the laterals constitute a very prominent feature of the root-system as a whole. That a recent rain of 0.21 inch penetrated to the roots of this species was evident from the fact that the branches and leaves were turgid. Of the root exposures, however, the best were found along a recent wash to the north of Table Mountain, where a small box canyon, about 100 meters in length by half that width, had been eroded. The sides were vertical and the walls about 1.2 meters high. The wash ran through a small plain on which were Atriplex and Kochia mainly, with Atriplex sp. dominant. The roots of several plants were partly exposed and were observed. Since the roots of all were about the same, a description of those of one will be sufficient. In this specimen the shoot was only in part living and it was apparent that a portion had been removed. Of the roots, it was found that 3 fairly large ones, representing tap-roots, went straight down over 1 meter. Several took their origin from these vertical roots just beneath the surface of the soil and ran in a horizontal direction, about 2 em. beneath the surface, to the base of a neighboring Atriplex, about 1.5 meter distant. The vertical roots gave off small and relatively unimportant branches at a depth of about 30 cm. The superficial horizontal roots bore numerous filamentous roots and such roots were seen on the most superficial of the other laterals. Such filamentous roots correspond to the ‘‘deciduous” ARID PORTIONS OF SOUTH AUSTRALIA. 73 rootlets which have been seen to occur on similarly placed superficial laterals of several species in the Tucson, Arizona, region. The roots of perennials growing on the Copley Plain, and described in the foregoing paragraph, are all either sharply vertical or as sharply horizontal in position. It is probable, from the large number of ob- servations, that this is the usual condition. However, an exception was found by a wash on the edge of the plain, where the soil is fairly coarse and the bank is about 2 meters high. Here an undetermined halophytic shrub of small stature had a root-system which was unlike that above described. In this instance there was a brush of roots, without the sharp differentiation into the vertical and horizontal members as elsewhere observed. This condition is probably attributa- ble to the fact that the soil at the place is relatively coarse, permitting a deeper penetration of the rain and better conditions of aeration than would more commonly be the case. VEGETATION OF THE Low HILus AND SLOPEs. The lower hills and slopes and the slopes of the higher hills are usually well covered with a perennial vegetation. Although some of it is of the halophytic type, it is largely composed of sclerophyllous shrubs, fairly uniform in appearance and generally of a relatively small size. The hill vegetation is so varied in species that it would require a much closer study than the present one to describe it at all accurately, as well as the aid of a large-scale contour map showing the leading geological features. Such, unfortunately, appears to be wanting. Certain leading characteristics, however, can be presented as a preliminary study. Possibly the most striking single feature of the hill perennial flora is the frequent segregation of species, so that a relatively large area may be populated by a single one to the exclusion of all others. This was noted again and again. Although there are many species of perennial habit on the hills, it is probably true that the most typical belong to two genera, Cassia and Eremophila. It will therefore probably convey the right impression if the vegetation of the hills was defined as the Cassia-Eremophila community. Following are some of the most conspicuous species of the low hills and slopes: Eremophila brownii. Eucalyptus oleosa. Atriplex vesicarium. E. freelingii. Cassia eremophila. Bassia lanicuspis. K. oppositifolia (plate 8c). C. sturtii. Hakea leucoptera. KE. latrobei. Myoporum platycarpum. Nicotiana suaveolens. Zygophyllum fruticulosum. Acacia sentis. Petalostylis labicheoides. Z. crenatum. A. aneura. Solanum illipticum. Z. prismatothecum. Cheilanthes tenuifolia. Menkea australis. Pholidia scoparia. Kochia planifolia. Pimelea microcephala. Fusanus acuminatus. K. sedifolia. Trichinium incanum. F. spicatus. Euchylena tomentosa. Casuarina lepidophloia. 74 PLANT HABITS AND HABITATS IN THE Mono-Speciric COMMUNITIES. Although the hill-slope habitat is here treated as a single unit, it is nevertheless far from uniform. However, as this habitat is well drained and hence the soil is well aerated, with relatively good water relations and various aspects, it may be proper to group the different slopes into one habitat. One of the features of the habitat is that any one element, for example the slope of a hill, is cf relatively large area, so that the environmental conditions of the particular area are correspond- ingly of relatively wide extent. This uniformity of environmental conditions extending over a relatively wide area is doubtless responsible for the frequent occurrence of a single species only, particularly of perennials, thoughout such an area. Thus it happens that mono- specific communities are common in the Copley region. Some of these will be mentioned here. The most conspicuous are those of various species of Eremophila. Of these, the ones along or accessible from the Mount Serles road may be considered typical and need only be de- scribed. To the east of the village the Copley Plain pushes into the hills, forming a bay-like area, with hills to the south, ending at the west in Table Mountain and lower hills on the north. The plain rises about 1.5 miles from the village and insensibly merges into gentle slopes which descend from the hills on the south. At the place in mind there is a considerable outcropping of slate or shale. With its northern aspect the habitat is relatively warm and is subject to the north winds, which at times are hot and dry. On this slope oceurs Eremophila freelingii, solely among perennials. This is a shrub about 1.5 meters in height, with an open habit of growth. The leaves are numerous and fairly crowded towards the ends of the branches (plate 9a). They are of leathery texture, about 3 cm. in length, and somewhat viscid. In the middle of July, when the area was first studied, they bore large numbers of lavender-colored flowers. No measurements were made nor computations of the populations per unit area. It can be seen from plate 10B, however, that the shrubs occur but sparsely. An- other area, in which F. freelingii occurs to the exclusion of other species of perennials, is by the Mount Serle road and about 5 miles east of Copley. Here are rounded slate hills also and the conditions are otherwise much as at the habitat just described (plate 9c). An additional mono-specific community observed was that of Pholidia (formerly Eremophila) scoparia, which was not far from the habitat last mentioned. It was different, however, in that the slope has a southern aspect and is somewhat less steep. Probably the water conditions are somewhat more favorable. Pholidia scoparia is rather strict in habit, “broom-like,”’ in fact, and has small leaves closely appressed to the branches, as shown in plate 9c. The shrubs have a canopy top and are 1.5 to 2 meters in height. Where they constitute ARID PORTIONS OF SOUTH AUSTRALIA. 75 the sole species, as at the habitat mentioned, they form one of the mcst striking communities of the whole region (plate 10c). A mono-specific community of quite a different: character was ob- served on the upper slopes of hills leading up to the eastern summits of the Mount Deception Range, about 4 miles to the west of Copley. When viewed from the plain, there appeared at this place a bright- green area, as of a grassy field. When it was reached, however, by the way of the lower hills, it was found to consist wholly of Zygophyllum crenatum, a low, spreading annual. There was no other species of any sort over the rather large area, possibly of several acres. Among other species, which occur in large numbers to the exclusion of others, is Cassia sturtit. This is a somewhat diffuse shrub about 1.5 meters high. It appears to have been much more abundant in former times in the vicinity of Copley, if one can judge from the large areas which now are covered with specimens no longer living, as on certain slopes of the Mount Deception Range. At present, how- ever, it is common enough. Frequently, as among the rounded hills along the Mount Serles road about 5 miles east of the village, it forms fairly large commuaities (plate 10a). It seems to be restricted to hills and slopes. Another example of a mono-specific community is that of Kochia sedifolia, which, unlike so many of the family, appears not to tolerate a large amount of salts in the soil, so that possibly it may not be con- sidered a strict halophyte. It occurs on slopes which are often, but not always, somewhat above the plain. The situation of the habitat to be described is along the Yudnamutana road, about 2 miles north- east of Copley, where an intermittent stream, making its way across a slope from hills farther east, has cut a fairly deep gorge. To the south of the edge of the latter the slope dips gently to the Copley Plain. The population is fairly sparse, as the habitat shows (plate 7c). The individuals are about 50 cm. high and of a very close, compact growth- habit. The species occurs to the exclusion of others over an area many acres in extent. Zygophyllum fruticulosum is another species which may occur solely in a habitat. This small shrub, of a lax growth-habit, grows on the lower slopes of the Mount of Light as well as on nearby slopes to the exclusion of other species of perennials. The habit and distribution of the species, as well as the character of the habitat, are fairly well shown in plate I1l1c. An additional species, which occurs with little or possibly no ad mixture of other species in certain areas, is Eucalyptus oleosa. This is one of the ‘‘mallees.”’ Although the mallee does not form a conspicuous member of the plant community at Copley, it is to be found scattered along the washes, and at a place about 4 miles west of the village and in the Mount Deception Range a mallee thicket has 76 PLANT HABITS AND HABITATS IN THE been formed. The species has a heavy root-crown and a thickened stem-base, from which numerous branches arise. The entire aspect of the species is shrub-like, with the ‘‘canopy”’ top so common among Australian shrubs and trees. ISOLATED SPECIES AND MIxED COMMUNITIES. Should the vegetation of the Copley region be studied from the standpoint of its distribution, it would be found that all gradations exist between the mono-specific communities above sketched and geographically isolated individuals. As a matter of fact, possibly most of the species occur mingled in this manner. Whether such small communities or isolated individuals represent recent introductions or survivals from populous mixed communities are matters of interest, but this study throws no light on the problem. It is evident that the controlling factor of the hill-slope flora is the single one of moisture and that it is apparently a fortuitous mingling of mostly unrelated elements. This is in sharp contrast to the flora of many species of the Copley plain, which is bound by edaphic ties. In other words, the conditions noted are the usual ones controlling the plant distribution in an arid or semi-arid region. It is not especially unusual to find a single species of woody perennials represented in any area by one specimen only. For example, one or two specimens only of Hakea leucoptera occur on the top of Table Mountain (plate 114), although there is a small colony of about 20 widely separated individuals on low slopes at the head of a draw at the east base of Mount Deception Range west of Copley. It was not seen apart from these two situations. At Copley the species has the habit of a small tree, about 5 meters high, with a pronounced canopy- shaped crown. The leaves are needle-like (plate 12p) and the whole appearance of the species is as one well adjusted to very arid condi- tions. It is one of the species which has probably suffered little by the advent of the white man and his varied activities. It is not eaten by his animals or used by him as a fuel. For these reasons the oc- currence of the species as to its distribution in the Copley region probably represents its reaction to the physical environment only. Such a remark might also be made regarding smaller woody peren- nials, as the Eremophilas and Cassias. Another woody perennial which occurs sparingly is Petalostylis labicheoides. ‘This species was found at the south and west base of Table Mountain, in a straggling group or two of few individuals. It was seen in no other place. The shrub has a very striking appearance, but imperfectly shown in plate 12, A and c. Several branches of ap- proximately equal length are fairly well clothed with compound leaves which carry 15 to 31 leaflets. Thus the leaf-area is large, although the leaflets are rather small. ARID PORTIONS OF SOUTH AUSTRALIA. 77 Among other species of the hill-slope habitat which occur rarely or scatteringly, mention need be made of two or three additional ones only. Thus, Cheilanthes tenuifolia was found only among rocks, south surfaces, on the northern slope of Table Mountain. Casuarina lepidophloia occurs in small numbers around the south and west base of Table Mountain (plates 11B and 128), and also in few numbers among the low hills on the Mount Serles road. In other places, as on the top of Table Mountain, the species occurs sparingly. This species forms a small tree in the vicinity of Copley and is of some economic use, making it probable that the number of individuals is very much less now than in earlier times. It is to be questioned, however, whether it ever occurred so abundantly at Copley as to constitute a pure com- munity, as at Quorn, for example. An additional species which occurs but sparingly at Copley is Pimelea microcephala. Two specimens of this small shrub were found at the eastern base of the Mount Deception Range and in the same habitat as Hakea leucoptera. It occurs in certain other places also, as to the south of Table Mountain. VEGETATION OF THE WASHES. There is no sharp distinction between the plants of the hill-slope habitat and that of either the plain or of other portions of lowlands, provided the drainage is good. Thus there are lines of stress between contiguous habitats where the species most characteristic of each are more or less mingled. Some instances of invasion may be found, but these appear to be rather uncommon. Inasmuch as the water rela- tions of the washes of whatever kind are particularly favorable, this fact is reflected in the vegetation growing in them. ‘Thus, species with large water requirements may be found in such a habitat, and as a general thing, although not without exceptions, such forms have a relatively large transpiring surface, may be of good size, and may occur in abundance. One of the most characteristic trees of the washes is the red gum, Eucalyptus rostrata. This becomes a fair-sized tree whose habit of growth is illustrated in plate 13c. In places, as on Leigh’s Creek near Copley, the red gum is confined to the floor of the wash. Here the bank is about 5 meters high and has all of the physical characteristics of the plain of which it isa part. Somewhat farther up the same wash, however, the transition from flood-plain to the adjacent higher ground is gradual and the plant communities characteristic of wash and plain are more or less mingled. Smaller contributing washes, whose depth may be little if any over a meter and whose width may be no greater, also carry the species. Such small washes are marked by a “‘single file” of the gums crossing the plain. The same wash entering the hills to the east of Copley is, with its smaller branches, likewise populated with the gums. It is apparent, therefore, that the species is very 78 PLANT HABITS AND HABITATS IN THE closely and directly dependent on a relatively good water-supply. It would be impossible to judge the earlier distribution and frequency of the species from its occurrence at present, owing to the past as well as the present demands of the white population for its use in various domestic purposes. However that may be, owing to the close depend- ence of the species on a good water-supply and to its not being very tolerant of an excess of salts in the soil solution, and, further, owing to the fact that the physiographic conditions are by and of themselves much restricted, it is hardly to be supposed that the species at any time had a distribution different from what it has now, even if the number of individuals may have been greater. Among the other woody perennials found in and along the washes are Melaleuca glomerata and M. parviflora. Although these species are confined to the washes, they are not generally distributed along them, as is the case of the red gum, but are segregated into isolated masses. Both species, for example, are to be found on the flood- plain of Leigh’s Creek at a place near the Myrtle Springs road and about 2 miles from Copley, and also in a small wash not far from the Yudnamutana road, about 3 miles north of the town. M. parviflora is a small tree (plates 138 and 15a) and is said to be tolerant of an excess of salts in the soil solution, and thus to be an ‘‘indicator”’ of brackish water. M. glomerata, on the other hand, is said to be an “‘indicator”’ of fresh water. However that may be, the two species are closely associated in the two habitats above referred to. M. glomerata is a large shrub (plates 13a and 15p) and under conditions such as occur along the Myrtle Springs road may form a dense, jungle-like growth with semi-prostrate stems, unique in this regard and different from most species of a semi-arid region. In both species the leaves are small, almost needle-like, but numerous, and hence the transpiration surface of any individual is large. Of other species belonging to the washes, there may be mentioned certain Eremophilas, some of which are wholly confined to drainage channels. Of these, the most pronounced, so far as the characteristic just mentioned is concerned, is EL. alternifolia (plate 144). Whenever in the Copley vicinity an Eremophila was found along a drainage channel, however small, as a slight depression on a slope, it was nearly always of this species. In spite of the fact that (as plate 15B shows) the leaves are fairly small, it seems necessary for the species to have much better water relations than any other of the genus occurring in this vicinity. Eremophila latrobet is a small tree; it was found in a small wash near the Yudnamutana road and nowhere else. re- mophila longifolia occurs in the same locality and elsewhere, but is not common. It also is a small tree. The habit of the species and of the leaf is shown in plates 148 and 16a. ARID PORTIONS OF SOUTH AUSTRALIA. 79 Of other species found only in or along washes, Acacia varians, ‘native willow,” becomes a fairly large tree with drooping willow-like habit (plate 15c). This species does not appear to be very common. A. sentis and A. tetragonophylla are confined to bottoms by drainage channels, or to such channels. The former is not abundant, but the latter forms small thickets which are nearly impenetrable, hence the local name ‘‘dead finish,’ not alone because of the abundance of the individuals, but also because of the short and sharply pointed phyllodia (plate 16, Bandc). Casuarina lepidophloia occurs along bottoms con- tiguous to washes not far from the one along which Acacia varians was found. Senecio magnificus occurs only where the water relations are relatively good, as in or along washes. Of other species confined to washes, or to their immediate vicinity, Heterodendrum olecefolium, Myoporum platycarpum (plate 178), and Jasminum lineare, all small trees, may be mentioned. Although the leaves of all are relatively large, those of the first named are the largest and, as will appear below, possibly rank next to those of Eucalyptus oleosa in size, indicative of an adjustment to relatively favorable water conditions. Parasitic PHANEROGAMS. The flowering parasites constitute a very conspicuous portion of flora of the Copley region. They represent two families, the Loran- thacee and the Santalacee. Of the former, Loranthus exocarpt, L. lineartfolius, and L. quandang, and of the latter family, Fusanus acumi- natus and F. spicatus were seen. The mistletoes are of very general occurrence, but the sandal-woods occur much more sparsely. Loranthus exocarpi is the most commonly met of the mistletoes. It is to be found on a relatively large number of hosts, among which were seen Hremophila brownii (plate 17D), H. longifolia, Fusanus acuminatus, Acacia sentis (plate 17c), Melaleuca glomerata, and Myoporum platy- carpum (plate 18c). Loranthus linearifolius was found on Acacia tetragonophylla (plate 188), and L. quandang was seen on Acacia aneura. Of these species, only L. exocarpi on Eremophila longifolia and L. quandang on Acacia aneura (plate 18a), the ‘‘mulga,’’ were seen to be in harmful abundance. In the latter instance, especially, the parasitic relation is said to terminate fatally for the host in every instance and within two years or more following infection. A striking peculiarity of the leaves of certain of the parasitic couples is the strong superficial resemblance which they hold. The sandal-woods are small evergreen trees and occur rather spar- ingly in the hill-slope habitat and apparently nowhere else. The fewness in the number of individuals is in part attributable to demands of various kinds which are made on them. They are a source of fuel and also are attacked by animals which devour the young branches and leaves. The fruits of one of them, the “‘native peach,” F’. acumi- natus, are used as food to a certain extent by dwellers in the region. 80 PLANT HABITS AND HABITATS IN THE Root-HABITs OF PLANTS OF THE WASHES. A few root exposures were found along different washes, where the erosion of the soil left them more or less in place. The leading results of the observations on them can be briefly given. Pholidia scoparia has a large development of roots which take their origin near the surface of the soil, and at less than 1 meter from the main root they attain to a depth of approximately 40 cm. and maintain this depth for an undetermined distance. There also is a well-marked tap-root. Essentially the same condition was seen in Hremophila freelingti, which was growing on the edge of a wash whose bank was over 3 meters high. In this case, however, numerous radiating roots were seen and no tap-root. The roots of unknown shrubs which were growing along a smaller wash had roots of the same general type as that given above for Pholidia. An undetermined species of Acacia was found to have two prominent roots which were superficially placed; the balance of the roots were not seen. The most deeply penetrating tap-toot found was of an unknown Atriplex, which went to a depth of about 2 meters. This species also had roots horizontally placed. By a wash leading down from the Mount Deception Range there is a small straggling grove of Eucalyptus oleosa where a few roots are ex- posed. As the bank was nota high one the exposure was not extensive. It showed, however, the swollen stem-base and enlarged root-crown, “mallee” characters, and the origin of the main roots (plate 198). It appears from these observations that the root-systems of peren- nials of the washes and hill-slopes, with the possible exception of that of Pholidia scoparia, consist of a deeply penetrating portion and a super- ficial portion which extend in a horizontal direction. Obligate deeply penetrating or obligate superficial root-systems probably do not occur. LEAF-FORM AND LEAF-SIZES. The leaves of all of the perennials observed at Copley are coriaceous, and although they vary considerably in size they have a certain monot- ony in outline. A few of the species have linear leaves, or phyllodia, and these may be greatly elongated or they may be short and almost spine-like. The widest as well as the longest leaves were those of Eucalyptus rostrata, which may be characterized as being narrow elongate. Thus, except for the juvenile leaves of this species, there appears to be no perennial in the region with leaves that are even ovate, not to mention the wider possible forms. A detailed account of the sizes and forms of leaves will be given ina separate section. Here it will be sufficient to present a summary show- ing the relative sizes of the leaves of several different species. The leaves of Eremophila alternifolia are given a value of 1; from this as the base, the relative size (area) of the leaves of several species is as fol- lows: Acacia aneura, 1.3; Eremophila brownti, 2; Myoporum platycar- ARID PORTIONS OF SOUTH AUSTRALIA. 81 pum, 5; Fusanus acuminatus, 7; Fusanus spicatus, 13; Acacia varians, 15; Eucalyptus rostrata, 20 to 26. As this is a portion only of the species observed, no safe conclusion can be drawn from the relation above them. It is of interest to note the relatively large leaf-area of Eucalyptus rostrata and Acacia varians, which are strictly confined to the habitat of the washes. But the complete exposition of leaf-sizes of such a habitat as the wash, however, would show great variation. Thus Eremophila alternifolia is confined, and quite strictly so, to this habitat, and it has the smallest leaves measured. It should be said, in this connection, that a great reduction in the transpiring surface, even to a condition of total aphylly, is not uncommon in species with large water-requirement. The opposite condition, however, namely, the occurrence of large-leaved species in dry habitats, does not take place. Therefore, the general rule can probably be said to hold, that there is a direct relation between leaf-size (or transpiration surface) and water-relation, at least in an arid region such as that about Copley. VEGETATION OF SOUTHWESTERN SOUTH AUSTRALIA. In the brief discussion to follow, on the plants of the southwestern portion of the state and on their environment, the vegetation of only three stations along the line of the trans-Australian railway will be considered, namely, Port Augusta at the head of Spencer’s Gulf, Ooldea not far from the western border, and Tarcoola about halfway between the two. At Port Augusta the rainfall is about 10 inches an- nually and at the other stations it is somewhat less. Port Augusta is at sea-level, but Ooldea and Tarcoola each have an altitude of about 500 feet. In spite of the low topography and the small rainfall the region is much more varied and interesting than at first might be sup- posed. At the extreme west and extreme east are plains, and in the mid-region low hills characterize the topography. Except for the outcrops of pre-Cambrian rocks, of relatively small area, the region is of the Cenozoic age; Recent to Pleistocene between Port Augusta and Ooldea; and Miocene and Eocene west of Ooldea. With a single exception, therefore, this portion of South Australia is geologically much more recent than the far north. In the plains regions halo- phytes are the dominating plants, and in the hilly mid-region there is a large variety of species which occur in larger numbers than might be supposed, judging from the rainfall alone. They all, however, reflect the severity of the environmental conditions under which they and their ancestors have developed for an immense period of time. VEGETATION AND THE ENVIRONMENT AT OOLDEA. PHYSIOGRAPHY. Ooldea is a station, really a construction camp at the time of my visit, on the new railway crossing the continent, and lies about one- third of the distance between Port Augusta on the east and Perth on 82 PLANT HABITS AND HABITATS IN THE the west. It is 427 miles from Port Augusta and about 80 miles, in a direct line, from the Great Australian Bight. The main features of physiographical interest are the sandhills to the east and the Nullarbor Plain on the west. It lies thus at the division between the two. The Nullarbor Plain is a formation of very great interest and unique- ness. Its east-west extent is about 450 miles and its north-south extent about 200 miles. The plain is a sea-floor, of the Mesozoic age, underlain by Paleozoic rocks (Jutson, 1914:71). Rocks of the same age outcrop to the west of the plain in Western Australia, and passing beneath the plain appear again about 100 miles to the east and again at intervals farther toward Port Augusta, as at Tarcoola. The Nullarbor Plain is composed of limestone of a thickness approxi- mating 500 feet. The surface of the plain is nearly level, as would appear from the fact that the transcontinental railway runs for 330 miles, nearly across it, without a curve. The topography, however, is not absolutely flat, but is very gently rolling and dips (less than 1 foot to the mile) to the east. The limestone is covered by about a foot of red soil, although fragments of the underlying rock are scattered about on the surface. The most striking feature of the plain, however, aside from its level character and great extent, is the presence here and there of slight depressions, known as “dongas,” which vary in size from a few to hundreds of acres. There are, in addition, as would be expected in the limestone country, various fissures or openings of various shapes and sizes which communicate with subterranean cavities. Such open- ings are locally known as “blow holes” from the air-movements asso- ciated with them. The “blow holes” often appear to be situated in depressions of small extent. From the structure of the plain it will be seen that such water as falls on it must quickly disappear from the surface. The soil is so thin that it constitutes a very inadequate water-reservoir. In the dongas, however, there is a greater accumula- tion of soil and hence greater possibilities in the way of water-storage, and in such situations the vegetation is relatively abundant and wholly unlike that of the surrounding plain. The sandhills region east of the Nullarbor Plain has an east-west extent of 100 miles, more or less, and a larger extent in the north- south direction. As seen from Ooldea and from the railways trav- ersing it, the region has a certain monotony. The hills are of nearly equal size and run from north to south, or approximately so. A hardpan resembling travertine, or desert limestone, is often present, lying about 1 meter beneath the sandy surface. In the fairly deep hollows between the ridges, however, the sand appears to be deeper than on the ridges themselves. A short distance to the north of Ooldea, where an impervious substratum underlies the hollows, water collects. This may be either brackish or “sweet.’’ Such a place is ARID PORTIONS OF SOUTH AUSTRALIA. 83 the well-known Ooldea ‘‘soak,” which has played an important part in the activities of men of all sorts who have been in these parts. Here, in a sandy depression, is to be found a fairly good supply of potable water which appears to be the only water of its kind for a very great distance. Explorers have relied on it for their supply, the aborgines come to it frcm long distancss in the course of their wander- ings, and it has been an important factor in the construction of the railway, providing good water for camps. The origin of the sand on the Nullarbor Plain and of the sandhills appears not to be surely known. So far as the sand on the plain in the vicinity of the sandhills is concerned it may have been moved there by wind action from the east, as is popularly supposed. Another hypothesis is to the effect that the Nullarbor Plain was laid down not far from shore and hence that it had a large admixture of coarse material, which was set free upon the elevation and the subsequent erosion of the plain. So far as the origin of the sand in the sandhill region prcper is concerned, Howchin (1909:67) suggests that it may have been derived from the ‘‘waste of the granitic rocks, which form the bedrock of the country.’”’ The same author says also that the calcareous material, which was derived from “breaking down and solution of Kainozoic limestone, has cemented the sand, in places, into calcareous sandrock, or has laid down from solution a crust of calcareous limestone.”’ Possibly the calcareous core of the sandhills, above noted, had some such origin. The sandy soil in the Ooldea vicinity is fairly coarse and permits the rapid absorption of water. If, for example, a bucket of water is emptied quickly on level ground it spreads only slightly beyond the place where it is poured and sinks from sight almost immediately. This being the case, it would be expected that water derived from the rains would also sink quickly and without appreciable run-off. Sucha condition was observed at the time of a storm of September 14. On this day there fell, between 1 and 4 p. m., 0.21 inch of rain. At 4> 15™ p. m. the soil was moist to a depth of 3.5em. No rain occurred during the night, but on the following morning the soil was found to be moist to a depth of 9.5 em. Apparently no run-off had taken place. From these observations, and from the absence of all indications of washing from storms, it is concluded that such water as falls on the sandy soil at Ooldea is absorbed at once, and further, from the nature of the soil itself, that it is well conserved. These conditions must be taken into account when possible reasons for the relatively heavy vegetation of the sandhill region are being sought. CLIMATE, Ooldea is within the 10-inch isohyet. It is in the region of winter rains, but apparently more or less precipitation is to be expected in summer as well. It will be of great interest, in view of the fairly abun- 84 PLANT HABITS AND HABITATS IN THE dant vegetation, to establish not only the amount of rainfall but its distribution through the year. Very little is accurately known as to the rainfall at Ooldea, for the reason that permanent settlement began only with the building of the transcontinental railway, which was opened late in the year 1917. There are, however, records for one year at a railway camp a few miles to the east and for nearly one year at Ooldea itself. In addition, rainfall reports are available for Fowler’s Bay to the south and at Tarcoola to the east, covering many years. Nothing is known regarding the rainfall on the Nullarbor Plain except from inference. The precipitation for the years 1917 and 1918 at 396-mile Siding and at Ooldea, supplied by the Bureau of Meteorology of the Commonwealth, is given in table 14. TaBLe 14.—Monthly rainfall at 396-mile Siding and at Ooldea, South Australia, November 1916 to September 1918, inclusive. Jan. | Feb. | Mar.} Apr. | May.|June.| July. | Aug. | Sept.} Oct. | Nov.| Dec. fe | | J Sf ff ff LDU sre rrr area ere eters eal | evetie te tdcore, otlet cldreccik ean es ciray alls Genterarcel reer etell tote seeped epee ae 0.19} 0.32 LOD 3s Sas a ae 2.5 | 1.15] 1.18] 1.24) 0.51] 1.49} 1.48} 0.22} 0.76) 1.69] 1.16]..... Ooldea: 1 Tee A Gee iy FSI atid WAT oe | AA | (a eRa ye| VERaN KA MP eye al be teal rg aul ie A Niel fal 0.73 LOUS scejeis eee O20 HOL0A2OF O36 60263)1 06) (O16 IE 19 | Os2012 eal seems No record where ..... is given. Beyond the general facts that the summers at Ooldea are very hot and the winters are cool, little is known of the temperature. Although 80 miles, or more, intervenes between Ooldea and the sea, it is prob- able that here, as at Tarcoola, where the distance is even greater, the southern winds are cool in summer and cold in winter, and that the converse is true of the winds from the opposite direction. Such winds, moreover, are undoubtedly of very great importance as factors modify- ing not alone the temperature but also the relative humidity of the air, and thus they very directly affect the vegetation. This had already been mentioned in the opening section and need not be further dwelt on in this place. HABITATS. During my short stay at Ooldea, besides exploring on foot its immediate neighborhood, I visited, through the kind assistance of Mr. Edwards, engineer in charge of construction, such localities as were most accessible from camp by motor car (railway), carriage, or saddle-horse. Thus the localities seen included the eastern side of the Nullarbor Plain, the transition between the plain and the sandhills, and the sandhills including the (1) Ooldea Soak, (2) the “‘oak’’ forest 6 miles south, (8) the Condensers and Station 408, together with the region traversed between Ooldea and the last two, and (4) the vicinity of Ooldea itself. It will be seen that there are at least three easily ARID PORTIONS OF SOUTH AUSTRALIA. 85 distinguishable habitats, and probably more—for example, the hol- lows between the sandy ridges and the dongas on the plain. Study of the plants of these habitats would undoubtedly show that each has a very characteristic vegetation to be associated with the differences which mark each habitat. VEGETATION OF THE NULLARBOR PUAIN. The eastern side of the Nullarbor Plain, as already mentioned, has a very slightly undulating surface. When one is well out on the plain he can see for many miles on all sides. No topographical feature marks one direction as opposed to another. At first there is a sense of barrenness, accentuated by the low relief, which is second only to that to be felt in the region about Oodnadatta; a more careful exami- nation of the plain does not fully justify this impression. A sparse gray-green plant-covering, monotonous in form as well as in color, is to be seen on every side. This is composed of halophytes of various species, among which are chiefly Atriplex vesicarium and Kochia sedifolia. There is also an ephemeral flora composed largely of grasses which spring up after rains; but at the time of my visit (September), only dried remains of such were to be found. These remarks on the flora of the plain apply to the plain proper, or rather to the highest ground, which composes by far the greatest percentage of its surface. Here and there, in looking about, one sees widely separated masses of green of small extent, really oases, which appear quite like small islands in a dull-gray sea. These are the dongas with their char- acteristic vegetation. Such dongas as were visited between Watson and Ooldea had a few low trees and shrubs and a fairly considerable dead, herbaceous flora; but the plants, the woody forms especially, appeared to consist of but a small number of species, although so few dongas were seen that no dogmatic statement in this regard would be warranted. Certain of these depressions near the Fowler’s Bay road had Acacia tetragonophylla, A. aneura, and a ‘weeping sandal-wood,”’ possibly Pittosporum phillyreoides. The kind as well as the abundance of plants in the dongas illustrate in a very striking manner the very great importance of the topography and of the substratum as factors in the vegetation of an arid region. A similar condition is to be found in southern Algeria (Cannon, 1913:31). There, for example, are depressions (dayas) which are the centers of small drainage systems in which there is an accumulation of soil and which have subterranean drainage, so that an excess of salts in the soil solution appears not to occur. Owing in part to drainage into the basins after storms, and in part to the great amount of soil in them, constituting important reservoirs, the dyas and dongas alike have relatively good water relations and their fairly abundant flora follows as a matter of course. 86 PLANT HABITS AND HABITATS IN THE VEGETATION ABOUT COLDEA. At and about Ooldea, and in fact on all of the sandhills of this region, so far as my observation extended, there is a surprising wealth of woody vegetation. Views across country reveal an undulating surface well covered with low, spreading trees and shrubs. The trees are so numerous that often the branches intermingle, completely hiding the ground beneath. In the immediate vicinity of Ooldea the leading trees and shrubs are species of Acacia, among which the following were seen: Acacia aneura, A. brachystachya, A. colletioides, A. kempeana, A. linophylla, A.oswaldti, A.randelliana, A.salicina,and A. tetragonophylla. The following notes on the local distribution, especially of these species, were made: Acacia aneura, the ‘“‘mulga,” forms a small, fairly dense tree (plate 19a) and occurs in two very distinct forms, that with narrow and that with broad phyliodia (plate 20, B and c). It is fairly abundant at Ooldea. Acacia brachystachya is also common about Ooldea, forming a large and attractive shrub. The phyllodia are linear, long, and, like most of the foliage of the sandhill community, of a gray-green color. Acacia colletioides is common about 6 miles to the south. It is a rather diffuse shrub and bears short spinescent phyllo- dia (plate 20a). Acacia kempeana, a shrub of compact habit of growth, has a relatively heavy covering of fairly large phyllodia and is common at Ooldea. Acacia linophylla, also common about Ooldea, forms a small tree, possibly 3 to 5 meters in height. In foliage-habit it is quite the same as the same species previously seen at Oodnadatta. Acacia oswaldii, a small tree, was seen about 6 miles south of Ooldea. Acacia randelliana is a shrub, 3 or more meters high, and common about Ooldea. Acacia salicinia, a small tree with rather prominent phyllodia, is frequently seen in the neighborhood. Acacia tetragono- phylla, which forms a small tree, was seen only on the edge of the plain and appears not to occur commonly in the sandhills. With the exceptions noted, the acacias appear to be very generally distributed on the sandhills about Ooldea, but how far they are con- fined to the ridges was not noted. Among other woody species seen in the same habitat were the “‘bullock bush,”’ Heterodendrum oleefolium, a small tree well covered with small leaves. Dodonea attenuata and Oleria muellert also were fairly common on the sandhills. Among other species of interest was the ‘‘oak,” Casuarina lepido- phloia, in some respects the most important species of the region. It occurs in the hollows between the ridges to the south of Ooldea as well as on the ridges themselves. The ‘‘oak” makes open forests and is said to extend as such to the bight on the south. It is of great im- portance at present, as the wood is largely used as fuel in distilling water at the “condensers,” for uses of the railway in the region. The demand is so heavy for wood that fairly young trees are cut, with serious portent to the species in this portion of the state. ARID PORTIONS OF SOUTH AUSTRALIA. 87 With other woody species, fairly common, may be included Euca- lyptus incrassata var. dumosa. This has the mallee habit of growth and is also utilized as fuel along with the “oak.”’ In distribution this species frequently, if not usually, occurs to the exclusion of other woody forms, but a spinifex, Triodia irritans, is often associated with it. In case, however, other woody species occur with the mallee, the grass seems to be wanting. As to the habit of the mallee, it has open, canopy-like shoots, of many stems, and thus is easily recognizable from a distance. In addition to this, the most abundant of the genus about Ooldea, there are two others of the genus. Of these, Hucalyptus pyriformis occurs very sparingly between Ooldea and Ooldea Soak, 3 miles to the north (plate 21a). The species is of characteristic mallee form, and bears large flowers and fruits (plate 22a). Accord- ing to Maiden, it is one of the handsomest species of the genus, ‘‘be- cause of the large size and showiness of the flowers and the large size and ornamental character or at least grotesqueness of the fruits.” The fruits are about 5 cm. in diameter and, occurring in groups, they give a striking appearance to the open shoot. The third species of Eucalyptus seen among the sandhills was a dwarf’ form, LE. leucoxylon var. macrocarpa (plates 218 and 22c). This was seen at Station 408, where it forms a small thicket. It is from 1 to 2 meters high, appar- ently somewhat smaller than is usual for the variety. Several woody species were seen only here and there and appeared not to be very common, or at least not to be of very general distribution in the region. Among these may be included Eremophila alternifolia, which is fairly abundant near Ooldea and also occurs on flats between the sandy ridges (Black, 1917). The same author also says that Eremophila paisleyi and Pholidia scoparia are to be found about Ooldea. All of these species of Eremophila were seen at Copley, where £. alternifolia is confined to the banks of washes or to other situations where the water conditions are relatively good. Fusanus acuminatus, the ‘‘native peach,” is not uncommon in the sandhills near the camp. Many other species occur scatteringly, several of considerable interest. Of these, Leptospermum levigatum var. minus was seen at Ooldea Soak only (plate 228). As has been mentioned already, Ooldea Soak is a hollow among the sandhills about 3 miles to the north of the camp. There is more than one depression at the place, but all appear much alike. The higher ground carries a very good population of small trees, acacias and mallee, but in the depressions there is very little growth. It is here that in certain hollows, but not in all, the Leptospermum is to be found. The species is confined to such depressions as have potable water and avoids such as have water that is brackish. It is an interesting fact that, although the water-table of the portion of the soak where the Leptospermum grows is little if any over 1 meter deep, and the water 88 PLANT HABITS AND HABITATS IN THE itself is perfectly suited to the needs of the sandhills species growing round about, yet they appear never to have encroached on the soak more than they do at present. Of other isolated species, one of the most curious is the flat-stemmed and leafless Bossiwa walker (plate 294). This low shrub forms a small, thicket-like growth. It was seen in several fairly widely separated places, among which was a sandy ridge immediately north of Sta- tion 408. At this place, and not far from the Bassiea, there occur a few specimens of a yet more curious shrub, Hakea multilineata. This is a tall shrub with very long coriaceous and strap-like leaves of up- right habit (plate 29,B andc). The large and stiff spikes of flowers, “bottle brushes,”’ add to the wierdness of the plant. In the same neighborhood as the two foregoing were Acacia salicina, Callitris verrucosa, Eucalyptus leucoxylon var. macrocarpa, Melaleuca uncinata (plate 29D), and Gravillea stenobotrya. Of these, Callitris, the ‘pine,’ is a small tree at the only place, Station 408, where it was seen. Gravillea stenobotrya, one of the ‘‘beef woods,” is a species of con- siderable interest (plates 24a and 25c). It becomes a fairly large shrub, 3 meters or more in height, and has a very open habit of growth. The leaves are long and narrow and do not appear to be very abundant. It is thus apparent that the transpiring surface is relatively small. The specimens of the beef wood especially examined were situated on the summit of a sandy ridge at the condensers.