JOURNAL OF THE ROYAL SOCIETY OF WESTERN AUSTRALIA Volume 69 • Part 4 • 1987 ISSN 0035-922X THE ROYAL SOCIETY OF WESTERN AUSTRALIA PATRON Her Majesty the Queen VICE-PATRON His Excellency Professor Gordon Reid, Governor of Western Australia COUNCIL 1986-1987 President: J S P Beard, M A, B Sc, D Phil Vice-Presidents: J T Tippett, BSc, PhD J S Pate, Ph D, D Sc, FAA, FRS Past President: S J Hallarn, M A, FAHA Joint Hon. Secretaries: V Hobbs, BSc (Hons), PhD K W Dixon, B Sc (Hons), Ph D Hon. Treasurer: W A Cowling, BAgricSc (Hons), PhD Hon. Librarian: H E Balme, M A, Grad Dip Lib Stud Hon. Editor: B Dell, BSc (Hons). PhD Members: J Backhouse, M Sc A E Cockbain, B Sc, Ph D S J Curry, M A E R Hopkins, B Sc, Dip For, Ph D L E Koch, M Sc, Ph D J D Majer, B Sc, DIG, Cert Ed, Ph D K McNamara, B Sc (Hons), Ph D J Webb, B Sc, Ph D, Dip Ed Journal of the Royal Society of Western Australia, Vol. 69. Part 4. 1987. Aspects of variation in histology and cytology of the external nasal gland of Australian lizards H. Saint Girons and S. D. Bradshaw Laboratoire d'Evolution des tires Organises, University Pierre et Marie Curie, Paris 75006. and Department of Zoology, University of Western Australia, Perth 6009. Manuscript received 17 June 1986; accepted 16 September 1986. Abstract The histological and cytological structure of the external nasal gland was compared in 32 species of lizards representing the five families found in Australia. Considerable variation in the size of the gland was apparent, but size alone was not necessarily a reliable indicator of the gland’s ability to function as an cxtrarenal salt-secreting organ. The elaboration of an hyperosmotic saline solution is associated with the presence of salt-transporting cells possessing a characteristic striated appearance, due to repeated folding of the basal and lateral membranes. These “striated cells” are generally grouped together into homogeneous tubular segments (“striated segments”) which may occupy from 14% to 85% of the tubules in the gland depending upon the species. In the three large skinks studied, however. (Egernia kingii. Tiliqua rugosa and T. occipitalis) homogeneous “striated segments” do not occur, as salt-secreting and classical muco-serous cells intermingle throughout the length of the tubules and right up to the proximal end. Nasal salt-secreting glands are most highly developed in the Australian Varanidae, and occur to some extent in all the Scincidae studied here. Salt-secreting elements occur rarely and then only feebly differentiated in the Australian Gekkonidae and Agamidae and are completely absent in the only member of the Pygopodidae examined. Lialis burtonis. External nasal glands in Australian lizards appear to differ from those described in species from both the Old and New World in that, even when apparently capable of functioning as salt-secreting glands, they show no obvious correlation with either environmental aridity or mode of nutrition. Introduction Recent studies have shown the lacertilian external nasal gland to be markedly polymorphic (Gabe & Saint Girons 1971, 1976; Dunson 1976; Lemire 1983). Often of small size and composed uniquely of classic glandular cells, it may also be enlarged and incorporate salt- secreting cells which arc usually arranged in tubules having a characteristic striated appearance (“les segments sines” in the terminology of Gabe & Saint Girons, 1976). Athough the mechanism of secretion is still not understood (Lemire 1983), it is clear that these cells are responsible for the capacity of these so-called “salt glands” to elaborate an hyperosmotic saline solution which, when eliminated, represents an important avenue of electrolyte excretion for many species. The presence and state of development of these salt-secreting cells appear to vary according to both the taxonomic position of the animal as well as its ecological situation. From the literature it would appear that well- developed glands with striated segments are very common in the families Iguanidae and the Scincidae, frequent but more variable in their occurrence in the Varanidae, rare and very variable in their state of development in the Agamidae, absent or very' little developed in the Gekkonidae and completely absent in the Chamaeleonidae and Anguoidea (including the Helodermatidae). By contrast, in agamids of the genus Agama and in the skink Tiliqua rugosa (Saint Girons, Lemire &. Bradshaw 1977) the nasal gland is not composed of typical homogeneous striated segments but, instead, salt-secreting cells are interspersed with classic glandular cells in the secretory tubules and are not restricted to an intermediary zone as in other species. The ability of these glands to elaborate an hyperosmotic saline solution has been demonstrated unequivocally only in the case of the North African agamid lizard Uromastix acanthinurus. in about a dozen iguanids which are all herbivores and frequently desert- living or littoral species (see Lemire 1983 for references), and in three varanid species — one widely distributed in Australia (Green 1972), one littoral (Dunson 1974) and the other Saharan (Lemire 1983). In all the cases where the gland has been studied morphologically it is relatively large in size and packed with homogeneous striated segments representing from 65-95% of the total volume of the tubules. Little is known of the physiology of “salt glands” from species where the gland is only moderately developed or where the salt-secreting elements form onlv 25-60% of the epithelium and the interpretation of data from such species is difficult (see Gerzelli & Dc Piceis-Polver 1970 Braysher 1971. Saint Girons e/iiA 1977, Minnich 1979). It does seem clear however that the small nasal gland of the Saharan Agama species has no osmoregulatory role, despite the fact that salt-secreting cells are common throughout the gland, intermixed with classical glandular cells (Lemire 1983). From a simple morphological point of view it is apparent that, within each Old World family, striated segments are more 54692-1 117 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. common or more well-developed in species occupying arid regions and they are invariably highly developed in species which are primarily herbivorous. The present investigation forms part of a detailed study of the ecophysiology of Australian reptiles, particularly lizards inhabiting arid and semi-arid regions of the continent (Bradshaw 1981, 1986) and an opportunity was taken to extend our limited knowledge of the morphology of these “salt glands” by examining common species living in a variety of habitats and representing the 5 families occurring in Australia. MATERIALS AND METHODS A list of the 32 species studied is given in Table 1. Amongst these, Tifiqua rugosa and the 4 varanids have already been the subject of detailed study (see Saint Girons et al. 1977, 1981). In the case of other species, specimens were autopsied the same day or the day following capture using Nembutal (Abbott, sodium pentobarbitone) as anaesthetic. The entire head was fixed for a period of 24 hr in aqueous Bouin, decalcified in 5% trichloracetic acid, dehydrated and then blocked in paraffin. lO/xm serial sections were reconstituted in 6 series by mounting one section in every 10 or 20, depending upon the thickness of the head. These series were then stained successively with PAS-haematoxylin- picro indigocarmine, Gabe’s Single Trichrome and Azan for topographic studies and with Mowry’s PAS-alcian blue to detect mucins and with Danielli's tetrazoreaction for protids as described by Gabe ( 1 976). The volume of salt-secreting cells and striated segments, relative to either total cell volume or total secretory segments, and the relative size of the external nasal gland were estimated by eye following the method of Gabe & Saint Girons (1976). In addition, in those species where striated segments were quite distinct, photographs w^crc made of the gland at three different levels, enlarged, and from these were cut all secretory segments both striated and glandular. These were then weighed to give an estimate of the relative proportion of salt-secreting to classical glandular portions of the gland. It should be emphasised, however, that it is possible to give a rough approximation only of the relative proportion of salt-secrelirig and classical glandular cells when both are interspersed all along the secretory tubules. Table 1 List of species studied, habitat type and cytological and histological characteristics of the external nasal gland Species GEKKONIDAE Crenodactylus ocellatus Diplodaciylus sienodactylus. Gehyra variegata Heteronotia binoei Oedura lesueuri Rhynchoedura ornata Underwoodisaurus milii PYGOPODIDAE Lialis burtonis AGAMIDAE Ctenophorusclayt Ctenophorus isokpis Ctenophorus maculafus Cienophorus ornatus- Ctenophorus caudkinctHS.... Ctenoph\ but none of the Australian lizards is an obligate herbivore, like for example the North African agamid Uromastix acanthinurus. The three large skinks studied here, which are all partially herbivorous {Tiliqua rugosa, Tiliqua occipitalis and Egernia kingii) show no greater development of salt- secreting elements in their external nasal glands than do insectivorous species. Given the morphology of the external nasal glands of lizards falling in the third category, where salt-secreting cells are either absent or only marginally developed, any role of the gland in osmoregulation can be excluded. This includes all the species of the genus Ctenop/wrus studied, many of which live in particularly arid regions. This is consistent with what is knowm of the water and electrolyte physiology of these lizards which survive long periods of water deprivation through their ability to retain sodium ions in the body fluids at markedly elevated concentrations (Bradshaw & Shoemaker 1967, Bradshaw' 1981. 1986). Lizards of the genus Agarna living in the Sahara similarly tolerate hypernatraemia, rather than excrete sodium ions via an external nasal gland (Lemire 1983) and even Uromastix acanthinurus^ which possesses one of the most developed external nasal glands of any lizard, has a very' limited ability to excrete sodium ions as shown by the work of Bradshaw et a(. (1984) and experiences hypernatraemia in its natural habitat during periods of water deprivation (Lemire et al. 1982). Amongst the Ctenophoms species examined here, only caudkinctus and nuchalis show' any tendency towards development of salt-secreting cells in the nasal gland, as with Hetcronotia binoei amongst the geckos. The ecological and physiological significance of this is. however, obscure, particularly in the case of this gecko which has recently been shown to be parthenogenetic in parts of its range (Moritz 1983. 1984). In conclusion, it would appear that ecological and environmental correlates with nasal gland development, particularly aridity and mode of nutrition, are much less obvious in Australian lizards than in other parts of the world. In part this may stem from our poorer knowledge of the ecology of many of these species and it would be most useful for future ecophysiological studies if the osmoregulatory capacities of these glands could be defined. Acknow'ledgement:^~'X\y\% work was made possible through a travel grant to HSG from the Deparlement des Affaires Etrangeres. Paris, and was supported by Grant No. D 1 -83/ 1 5 1 22 to SDB from the Australian Research Grants Scheme and by funds from the University of W.A. References Bradshaw, S. D. (1981). — Ecophysiology of Australian desert lizards: studies on the genus Amphiboturus. IN Ecological Biogeography of Australia (ed. A. Kcasi), Vol. 2, pp. 1394-1434. Junk, Den Haage. Bradshaw, S. D. (1986). — Ecophyuology of Desert Reptiles, 350 pp. Academic Press, Sydney. Bradshaw, S. D. & Shoemaker. V, H. (1967). — Aspects of water and electrolyte changes in a field population Amphiholurus lizards. Comp. Biochem Physiol. 20: 855-865. Bradshaw. S. D.. Lemire, M.. Vemet. R. & Grenot, C. J. (1984). — Aldosterone and the control of secretion by the nasal gland of the North African desert lizard, Uromastix acanthinurus. Gen. Comp. Endocrinol. >4: 314-323. Bradshaw, S. D., Tom. J. A. & Bunn. S. E. (1984), — Corticosteroids and control of nasal salt gland function in the lizard Tiliqua rugosa. Gen. Comp. Endoermoi. 54: 308-313. Braysher. M. (1971). — The structure and function of the nasal salt gland from the .‘\usiralian sleepy lizard Trachydosaurus (formerly Tiliqua) rugostts: Family Scinddac. Physiol, /tool.. 44 : 129-136. Dunson. W. A. ( ] 974). — Salt gland secretion in a mangrove monitor lizard. Comp. Biochem. Phy.sioi. 47: 1245-1255. Dunson, W. A. (1976). — Salt gland in reptiles. IN Biology of the Reptilia (eds. C. Cans & W. R. Dawson), Vol. 5. pp. 413-445. Academic Press, N. Y. Gabc. M. (1976). — Histological Techniques. 1106 pp. Masson. Springer- Verlag, Paris. Gabe. M. & Saint Girons, H. (1971). — Polymorphisme des glandes nasales extemes des sauriens. C. R. Acad. Sci. Pari.i. 272; 1275-1278. Gabe. M. & Saint Girons. H. (1976). — Contribution a la morphologie comparcc des fosses nasales el de leurs annexes chez les Lepidosaunens. Mem. Mu.s. Nat. Hist. nal. Paris. A98; 1-87. Gcrzelli, G. & De Piceis-Polvcr, P. (1970). — The lateral nasal gland of Lacerta viridis under different experimental conditions. Monitore Zool. iral., 4 : 191-200. Green. B. (1972). — Water and electrolyte balance in the .sand goanna, Taranus gouldii. Unpublished Ph.D. thesis. University of Adelaide. Lemire. M. (1983). — Contribution b I’eiude dcs structures nasales des Sauriens. Structure et/onciion de la glande “a sels‘* des Iwards descrlicoles. These d‘Etat non publi6e de PUniversite Pierre et Marie Curie. Paris. Lemire, M., Grenot. C. J. & Vernel. R. (1982). — Water and electrolyte balance of free-living Saharan lizards, Uromastix acanthinurus (Agamidac). J. Comp. Physiol., 146: 81-93. Minnich. J. (1979). — Reptiles. IN Comparative Phvsiology of Osmoregulation in Antmals{cd. G. M. O. Maloiy), Vol. 1. pp. 391- 641. Academic Press. London. Moritz. C. (1983). — Parthenogenesis in the endemic .Australian lizard Heteronotia binoei (Gckkonidae). Science (H. Y.). 220: 735-737. Moritz, C. (1984). — The origin and evolution of parthenogenesis in Heteronotia htnaei (Gekkonidac). I. Chromosome banding studies. Chromosoma. 89: 151-1 62. Parsons, T. S. (1970). — The nose and Jacobsen’s organ. IN Biologv of the Repiilia (eds. C. Cans & T. S. Parsons). Vol. 2, pp. '99-191, Academic Press. London & New York. Saint Girons, H.. Lemire, M. & Bradshaw. S. D. (1977). — Structure de la glande nasale exierne de Tiliqua rugosa (Repiilia. Scincidae) et rapports avec sa fonction, Zoomorphologie, 88: 277-288. Saint Girons. H., Rice, G E. & Bradshaw, S. D. (1981). — Histologic compar^e et ullrasiructurc dc la glande nasale exierne de quelques Varanidae (Reptilia: Lacertilia). Ann. .Sci. Nat. ZooL. Paris, 13*^ Serie3: 15-21. 121 Journal of the Royal Society of Western Australia, Vol. 69, Part 4. 1987. Northern Sandplain Kwongan: regeneration following fire, juvenile period and flowering phenology by Paul G. van der Moezel, William A. Loneragan and David T. Bell Department of Botany. University of Western Australia, Nedlands, W.A. 6009. Received October 1986. Accepted March 1987. Abstract Fire is an integral factor in the ecology and management of the shrub lands of the Northern Sandplains in Western Australia. Documentation of fire effects on 192 species from a range of edaphic conditions revealed that 73% were capable of resprouling after fire. Both resprouling and reseeding species resumed flowering quickly following fire with 79% of the species flowering within two years. Particular species requiring longer juvenile periods, however, could have important management consideration due to their status as pollen species for apiculture or their conservation status. Flowering peaked in September in this study region and there were no major differences in phenology of more recently burned sites compared with mature shrublands. The interaction of the use of fire to protect human developments and the desired maintenance of areas of unburnt shrubland for honeybee pastures and biological species preservation is discussed. Introduction Fire is an integral factor in the shrub-dominated ecosystems throughout the world (Specht 1979). The plants of these communities possess numerous adaptations which enable them to regenerate after fire, such as sprouting from buds located in underground organs and firc-siimulaicd seed germination. Studies on the effect of fires on shrub-dominated heaths in Australia have shown that most of the species regenerate after fire by sprouting (Specht et al. 1958. Siddiqi et al. 1976, Russell and Parson 1978. Bell el al. 1984, Bell 1985). In these communities the species which lack the ability to resproul, i.c. obligate seeders, regenerate either by fire-stimulaicd germination of seed stored in the soil or by dispersal of seed held in woody fruits. The effect of fire on soil-stored seed is well documented (Stone and Juhren 1951. Went et al 1952. Cushwa et al. 1952, Christensen and Muller 1975). The depth and intensity of heat through the soil profile during a fire is an important factor in determining post-firc regeneration (Shea et al. 1979). .A very hot fire may kill underground organs (Hopkins 1979) and soil-stored seed, whereas a fire of low intensity may not stimulate some seed to germinate (McArthur and Cheney 1966). Some Australian plant species, eg. Xanihorrhoea spp. and a few species of the Orchidaceae actually depend on fire to stimulate flowering (Gill 1975). Most plants, however, require a certain time after a fire before reproduction begins. This period, termed the “juvenile period", is least important for plants regenerating after fire by sprouting but is an important characteristic of plants regenerating from seed. Time since last fire may also have an impact on the annual period of flowering. The Northern Sandplain shrublands have probably been subjected to periodic fires for at least the past 5 000 years (Churchill 1968). Under conditions prior to the settlement of the region by European man. the region probably received fires on a cycle of some 25 years (Bell 1985). Today fire frequencies are higher due to man- caused fires (Bell et al. 1984) and a conirolled-burning regime must be imposed on certain regions under management (Bell and Loneragan 1985). Some of the Northern Sandplain shrublands have been reserved in National Parks and Nature Reserves, but extensive areas have been cleared for agricultural, pastoral and mining land uses. Uncleared land serves a number of economic purposes including the tourist industry and the cut-flower and native seed collection trades. The Northern Sandplain native shrublands also serve as winter season “honeybee pastures" for cornmercial apiarists. The beekeeping industry uses the native shrublands and especially the pollen produced by winter-flowering species to maintain hives and to build up worker bee numbers for the honey production seasons in the south-western forest regions later in the year. Fire management in the Northern Sandplain must provide sufficient areas of prolific shrublands to serve the needs of commercial apiculture while protecting life and property in the adjacent wheat and pasture developments. Information on the impact of fires on the shrublands is of primary importance to the apiculture industry and aspects of species conservation. Conservation of community types and the flora and fauna of regions of this rich (Lamont et al. 1984) and highly endemic (Rye 1982) shrubland must also consider the impact of fire (Bell et al. 1984). The following study was designed to provide information on the influence of fire on the mode of regeneration, the length of the juvenile period and the flowering phenology for species of these Northern Sandplain shrublands. 54692-2 123 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Methods Twenty-six permanently marked 20 m x 10 m plots were established in Northern Sandplain shrublands near Badgingarra. Western Australia (30° 1 6’S, 1 1 5“26'E). The sites were representative of a range of topographic sites and ages-since-last-fire. Each site was initially categorized as lateritic upland or deep sand slope as these edaphic conditions have proved to produce the major florislic differences in the vegetation of this region (Bell and Loncragan 1985). The ages of more recently burned sues were determined by records of the Western Australian Bush Fire Board. Sites burned before 1 1 years ago could not be exactly documented and were grouped as > 1 1 years. Each site was visited monthly from March 1981 till December 1981 and a list of species in flower was compiled. Post-fire regeneration strategies were determined from recently burnt sites. Obligate seed regenerating species could be recognized because they initially have only a single erect stem. These seedlings were clearly differentiated from resprouting species which tend to be multi-stemmed. Geophytes were classed as sprouters since they regenerate after fire by producing new shoots from underground storage organs. Results and Discussion Resprouting after fire During the study 238 species were identified {Table 1, Appendix I). The range of sites allowed particular species to be assigned to edaphic preference categories. Among the 192 species identified in this way. approximately equal numbers were subjectively assigned to the generalist (or cdaphically indifferent), lateritic or sand specialist categories. The vegetation patterns in the Northern Sandplain shrublands have previously been shown to correspond strongly to the major differences in soil conditions, but fire was also shown to influence the floristic composition of stands in these shrublands (Bell and Loncragan 1985). Categorizing species of the Northern Sandplain study sites into mode of regeneration after fire revealed that 73% of the 192 species recorded in this way w'ere capable of resprouting after fire. This division of sprouters and obligate seeders is similar to shrubland sites in eastern Australia w'hcrc the reported percentages of sprouters includes 70% in South Australia (Specht el ai 1958), 73% in Victoria (Russell and Parsons 1978) and 80% for the coastal heaths of New South Wales (Siddiqi et al. 1976). The sclerophyllous shrub-dominated underslorcy of the jarrah forest of the Darling Range also contains a similar proportion of resprouting species (Christensen and Kimber 1975, Bell and Koch 1980). In a more limited study of predominantly deep sand sites in the Northern Sandplain, resprouter species represented 66% of the total (Bell c/ 1984). Table 1 Summary statistics for the Tire response survey m the Northern Sandplain shrublands near Badgingarra. Western Australia. Total Species Identified in studies Species Categorized for Total 238 Total Generalists Sand Laterite Edaphic Preference 197 72 61 64 Species Categorized for Total Sprouter Seeder Both Regeneration Strategy 192 126 51 15 Species Categorized for Total <2 yrs •>2<4 yrs >4 yrs Juvenile Period 108 87 16 5 Species in Phenology Study Total 149 The "sprouting" habit is considered an adaptation to recurring fire (Biswell 1974). Conversely, a long fire-free period was probably important in evolving the obligate seeding strategy (Keelcy and Zedler 1 978). However, it is uncertain w'hether the characters such as sprouting, woody fruits and hard seeds are adaptations specifically to fire or adaptations to other environmental factors, such as a low nutrient regime (Specht 1979), drought (Hnatiuk and Hopkins 1980) or insect damage (Morrow 1977). Whatever their origin, these adaptations ensure survival in the fire-prone regions of the Northern Sandplain. Many of the tall shrubs of the Badgingarra shrublands are obligate seeders, e.g. Hakea obliqua. Acienanthos cygnorum. Dryandra sessilis. This relationship of size with regeneration mode w-as also identified in Kings Park, Western .Australia (Baird 1977). The significance of this relationship, how'cvcr. is obscure. A number of the obligate seeders possessed the bradysporous habit (Specht 1979), i.e. seeds arc retained in woody fruits or cones until a fire opens the fruit. Examples of species with this habit in Western Australian heathlands are Hakea obliqua. Eremaea fimbriaia and Beaufortia elegans. Other obligate seeders such as Acacia piilchella and Kennedia prostrata possess "hard" seeds (Ewart 1908). These seeds remain viable and dormant for long periods in the soil until some event, usually fire, stimulates them to germinate. The effect of fire on these hard seeds is to crack the secdcoat making it permeable to moisture and oxygen (Beadle 1940; Floyd 1966, 1976), As hard seeds can remain viable in the soil for many years even after the parent plants have died, species richness and diversity often increases following managed fuel-reduction fires (Bell and Koch 1980). Flow ering phenology Throughout the 26 shrubland sites the maximum flowering period occurred in spring with a peak of 74 species recorded on September 29th (Fig. I, Apendix 1). Common winter flow'cring species were Leucopogon conosiephioides. Andersonia lehmanniana and Stylidium repens. In late winter and early spring Hibbertia crassifolia. H. hypericoides and Drosera hcterophylla were flowering abundantly in most sites. With the onset of spring many more species began their flowering period and no species dominated flow-ering throughout all sites. Calothamnus sanguineus was the only species which flowered throughout the sampling period. The rapid increase in flowering species towards spring was associated with an increase in flowering for species of the families Myrtaceae and Proteaceae and a decrease in the Epacridaceae (Fig. 2). The winter maxima for members of the Epacridaceae are important as members of this family are reported to be favoured apicultural species (Smith 1969). In comparing the flowering periods between sites varying in the lime-since-last-fire, only two species. flypocalymma xanthopetalum and Hovea stricta showed a different phenological pattern between the recently burnt and long unburnt sites. Hypocalymma xanthopetalum flow-ered in two year old sites as much as IW'O months before it flowered in any of the older sites. Hovea stricta flow'ered in two- and five-year-old sites, also two months before it flow^ered in the greater than eleven-year-old sites. In general, however, stand-age had little relationship to the period of availability for honeybee use once the juvenile period was completed. 124 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. 1981 Figure I. — Total number of species flowering in 26 heaihland study sites in the Northern Sandplain during the period April through December 1981. Figure 2. — Total number of Epacridaceae, Myrtaceae and Proteaceae species flowering in 26 heathland study sites in the Northern Sandplain during the period April-December 1981. The time required for plants to reach reproductive maturity after fire was recorded for 108 species (Appendix 1). Since all species were not present at every site, and sites of one and three years-since-last-fire were not available, the exact number of years to reach reproductive maturity could not be assigned to most species. Many species, therefore, were given values of <2, <4>2, etc. A value of <4>2 means that the species was seen flowering in a four-year-old site, was not seen flowering m a Iwq-year-oid site and was not recorded in a ihree-year-old site, so therefore it flowers within three or four years after a fire. Most of the species (79%) required only two years at the most to begin flowering after fire regeneration. Several species flowered profusely in the first few years following a fire, but then were only minimally reproductive in older sites. The most conspicuous of these plants was Vertkordia grandis which produced a mass of bright red flowers in the sites of less that two years old. The fire-stimulated success of species such as Vertkordia grandis, Stirlingia latifolia, Anigozanthos humilis, Pimelea sulphurea, and ' others may be a response to light as observed by Stone and Juhren (1951) or induced by substances produced during a fire. Gill and Ingwersen (1976) demonstrated that injection of ethylene, which is produced in large quantities during a fire, into Xanthorrhoea australis stimulated the species to flower as it would normally do after a fire. Even though many plants flower only after a fire, there was no difference in the number of species flowering in burnt and unburnt sites (Table 2). The study of several sites over a period of years would probably show some pattern in the number of species flowering after a fire but in this study, the variation between sites of the same age was too great to permit valid comparisons between sites of different ages. Table 2 Species in flower ai each site for 1 98 1 for the 26 Northern Sandplain kwongan study sites. Site Years Since Date 1981 Number Last Burn 29.4 28.5 10.6 7.7 5.8 1.9 29.9 20.1 1 19.12 Sand 2 >11 0 1 5 5 7 6 4 14 >11 — 0 1 2 4 5 b b b 16 >11 — 1 2 4 7 8 11 7 21 >1 1 2 2 3 10 8 9 7 7 11 2 0 0 5 8 9 4 5 12 11 — 0 1 2 4 9 9 10 7 1 10 0 1 2 3 5 9 5 3 17 7 0 1 2 1 3 7 11 12 6 19 6 0 3 4 6 7 7 14 11 8 23 5 2 3 4 5 5 10 10 7 8 3 4 0 3 3 4 7 9 14 10 4 5 2 0 0 0 3 5 9 14 8 4 25 2 2 2 1 5 7 7 11 12 8 9 2 2 2 1 3 4 9 11 4 6 Laierile 10 >11 — 2 2 b b b b b b 13 >11 — 3 3 2 6 6 9 4 4 15 >11 — 3 3 6 8 7 9 6 4 22 >1 1 2 1 I 5 10 9 20 9 6 8 11 0 I 2 5 9 13 b b b 18 7 0 0 I 8 12 7 10 5 1 20 6 0 1 3 3 8 4 10 5 6 24 5 1 3 5 9 7 13 8 4 4 4 0 1 2 5 9 6 11 4 4 6 2 0 1 2 4 8 8 1 1 8 4 26 2 0 0 0 7 10 8 16 9 4 11 2 — 1 — I 4 5 10 7 5 b - burnt — = unsampled There was no evidence that any species ceased flowering once a site reached maturity, i.e. greater than ten years-since-last-fire. The common belief that plants regenerating from seed have a longer Juvenile period than sprouting species was unfounded in these results. Species reproducing from seed such as Dryandra sessi/is, D. kippistiana, Petrophile media and Leucopogon siriaius, for example, flowered on seedlings which were only two years old. Others, for example. Hakea obliqua and Dryandra carlinioides, however, required four years before they flowered. In these species a fire interval of three years could be disastrous and may lead to their local extinction. In heath vegetation of South Australia, 125 Journal of the Royal Society of Western Australia. Vol. 69, Part 4, 1987. firing at intervals ofless than five years was found likely to eliminate Banksia ornata. Casuarina piisilla and Leptospennum myrsinoides\ species which take several years to reach reproductive maturity (Specht et ai 1958). In Victorian coastal heaths, Leptospermurn laevigatiim is killed by fire and requires four years before it flowers (Burrell 1968). Short fire intervals would be detrimental to the long term survival of such species. In the Northern Sandplain region near Badgingarra. the impact of fire on the availability of flowers important to the bee keeping industry must also be considered. Species of the Epacndaccae, Fabaceae, Mimosaceae and Aslcraceae are important pollen- producing species in the Northern Sandplain shrublands. During the winter months. Leiicopogon species dominated the available flowers at nearly ever\ site on both sand and laterite substrates which had a time-since- last-fire of four years or more. Since it only takes two years for seedlings of Leucopogon striafus to flow'er after a fire, it is unlikely that this species will be eliminated from an area by frequent burning. Another important pollen producer. Acacia pulchella, requires fire for establishment and flowering. Most plants of the healhland flower within four years after a fire. .Although the dominant Leucopogon (winter) and Proteaceae (spring) species present in sites greater than four years since last fire did not dominate in recently burnt areas, the overall number of individuals and species in flower and density of flow'cring appeared as great in a two year old site as in a ten year old site. Whether the bees can utilise the flora of a recently burnt site though, is as yet untested. Given that honeybees can fly up to 1 1 km from their hive in search of favourable plants (A. Kcsscll. pers. comm.), it would take a large fire to render an apiary site completely unusable. Species consermtion managcmeni Crown lands in the region must be managed to control the fuel build-up and a concomitant increase in the potential of uncontrollable fires starting within the shrub communiliies and spreading into the adjacent pasture and farmlands, thereby endangering human life and property (Bell and Loneragan. 1985). Other considerations, however, include the conservation of examples of this extremely rich flora and the maintenance of sufficient areas containing flowering species of importance to the honeybee. Conservation of native flora everywhere is of growing concern since the rate of extinction is increasing rapidly as a result of man's activities (Leigh and Boden 1979). It has been estimated that, in tropical rainforests alone, at least one species is disappearing every day (Myers 1 979). What this figure might be in the healhiands of Western Australia would be pure conjecture since many species here have vet to be described (Marchani and Keigheiy 1979). The southwestern corner of Western Australia is characterized by a high degree of endemism. Marchant (1973) estimates that 68% of a listed 3 600 angiosperms in the South-West Botanical Province are restricted to this province. Marchani and Keigherv' (1979). in high- lighting the lack of knowledge of the Western Australian flora, list over 2 000 species of vascular plants as being cither poorly known or possibly rare or restricted to a small geograVhic area. With further taxonomic revision of local genera, as much as 25-30% of the south-west flora may be classified as rare (Marchant and Keighery 1979). Twenty-two of the 238 species recorded in the study area can be classified as either rare, restricted or poorly known (Appendix 1). One species {FAicalyptus pendens) is classified as being rare and occurring in a restricted habitat. Only two small populations were observed in the laterilic sandplains of the southeastern section of the Badgingarra National Park. Seven species {Cassytha piibescens. Dampiera lindleyn Gastrolobium hidens, Hihbertia gkdwrrima, H. pilosa. Leucopogon crassifolia and Xanthorrhoea rejlexa) rank as poorly known, with only 2-5 specimens preserved in the Western Australian Herbarium. Six species {Blancoa canescens, Conospermum nervosum. Daviesia cpiphylla. Dryandra nana, D. tridentata and Hakea flabeilifoUaj are classified as restricted to areas of less than 100 km diameter. Of these rare or poorly known species, six regenerate from seed and arc therefore considered in most danger of elimination by fire, and, since a two year old site is probably incapable of carry ing a fire, they are unlikely to be eliminated by this means alone. More data arc needed on the flowering characteristics of Gastrolobium bidens, Leucopogon crassifolius and Conospennum nervosum to determine how many years are required after a fire before these species flower. The remaining species collected fr*om the study area which are not rare, restricted or poorly collected, all flower within five years after a fire. Most species were found to regenerate after fire b> sprouting and are therefore not in danger of elimination from fire. .Although the seed-regenerating species Hakea obliqua and Dryandra carliniodes require at least four years to flower following fire, neither are rare, nor restricted so are unlikely to be in danger of extinction from frequent fires. Bee pasture management The winter “hive-buildup" period in the Northern Sandplain healhiands is of major importance to the apicultural industry. During the early winter months, pollen collection from species of Leucopogon predominates. At least four years are required after a fire before' Leucopogon striafus. the most abundant of the winter flowering Leucopogon spp., returns to flowering in abundance comparable to unburm sites. During the first four years following fire, there are as many species flowering as in an unbumt site but since L. siriatus is not tlow'cring abundantly it is unknown whether recently burnt sites are capable of supporting an apiary site. Comparisons of the expected distances foraged by honeybees and the observations of fire scars visible in Landsat photographs of the northern sandplain indicate that fires of such magnitude to render an apiary site unusable have never been attained in years prior to 1984. Ob.servalions of flowering of old sites and especially important honeybee pollen species such as Leucopogon strialus indicates that no decline in flowering intensity appears with increasing site age. Therefore, there would appear to be no disadvantage to the apiarist in leaving a site unburned for many years and their claims for long-term fire protection may be based more on subjective visual assessment of apparent flowering intensity. Fire hazard reduction In monetary terms, fire is probably the cheapest management tool used in manipulating vegetation today. The introduction of prescribed burning in reducing the hazard of uncontrollable high-intensity fires has been widely used in forests, wilderness areas, nature reserves and national parks in the last decade or so in .Australia (Gill 1977). The development of large Journal of the Royal Society of Western Australia. Vol. 69, Part 4, 1987. areas of the northern Sandplain heaihlands for crop and agricultural uses since the 196()'s places these at some risk from wildfires. The disastrous Beekeepers Reserve fire of January 1984 burned over 1 17 000 ha (Burking and Kessell 1*984) and Northern Sandplain shrubland fires between Februar> and May 1985 destroyed another 63 400 ha (Davies 1985). The losses in economic terms to the beekeeping industry could reach more than 5 million dollars over the next eight years (Davies 1985). The impact on conservation, tourism and the cut-flower and native seed industries is difficult to estimate but could also be considerable. For these reasons, alone, a policy on fire management for the Northern Sandplain shrublands is essential. Before the establishment of farms and roads in the Badgingarra area, fires were a common occurrence on crown land during the late summer and autumn months. These fires appear to have been lit by lightning and usually burnt, uncontrolled, for one or two days before going out naturally (A. E. Eastwood, pers. comm.). Lightning is such a common occurrence during summer thunderstorms over this latentic country that as soon as an area is capable of carrying a fire the chances of it remaining unburni for any great length of time would seem to be low. The vegetation of sand and latcritic cornmunilies recovers to maximum cover in an average of eight years (Bell et al. 1984). More material capable of combustion, however, is available as the age of a kwongan stand increases up to at least 20 years. Therefore, the longer a “burnable” site is left unburnt, the more intense would be a fire and its potential destructive force. The policy of “let nature take its own course” has operated advantageously in most large national parks in the past and is still in practice in many at present (Gill 1977). However, in areas of multiple land use, controlled fires under chosen conditions are more desirable to uncontrolled fires which may prove difficult to confine. Controlled burning of the Northern Sandplain area would necessarily involve rotational burning to eventually produce a mosaic of differing ages sincc-Iast- fire. Areas of more than 1 1 years-since-last-fire are however, becoming increasingly scarce. Efforts should therefore be made to create some areas which are protected from fire for as long as possible to allow for further research. These areas should not border on farmland in case of a wildfire. Ideally they should be enclosed by a wide buffer strip which is burnt more frequently and a fire break so that the chances of the area being burnt arc reduced. Any manipulation of the environment by man should attempt to closely reproduce the events of nature as much as possible. Applying this to fire management policies, it would be desirable to have controlled fires at a frequency close to that under more natural conditions. Policies of very frequent burning, or complete fire suppression should be avoided. Results from this study suggest the natural fire frequency in the Badgingarra area could be as low as between 8 and 1 5 years. .A policy of controlled burning everv' 10 years in a mosaic pattern would: ( 1 ) be unlikely to cause any loss of species; (2) be unlikely to badly interfere with the apicultural industiy- unless the area burnt was ver>^ large; (3) would reduce the risk of uncontrolled wildfires: and (4) would closely simulate the actual fire pattern of the area under natural conditions. Acknowledgements — Funds from the Honey Research Committee of the Department of Primary Industries and the Western Australian Bush Fires Board are greatly appreciated. The position of Senior Lecturer in Plant Ecology of Dr Bell is supported by Alcoa of Australia Ltd. and Western Collieries Ltd. References Baird, A. M. (1977). — Regeneration after fire in King’s Park, Perth. Western Australia. J. RoySoc. ^Vest. Ausi., 60: 1-22. Beadle. N. C. W. (1940). — Soil temperatures during forest fires and their effect on the survival of vegetation. J. EcoL, 28: 1 80-92. Bell, D. T. (1985). — Aspects of response to fire in the Northern Sandplain Heathlands. In Fire Ecology and Management in IVestern Australian Ecosystems. J. K. Ford, (ed.). pp. 33-40. WAIT Environmental Studies Group Report No. 14. Bentley, Western Australia. Bell. D. T.. Hopkins. A. J. M. and Pate, J . S. (1984). — Fire in the Kwongan. In Kwongan-Plani Ltje of the Sandplain. Pate J. S. and Beard, J. S. (eds.) pp. 178-204. University of Western .Australian Press, Nedlands. Western Australia. Bell, D. T. and Koch. J. M. 11980). — Post-fire succession in the northern jarrah forest of Western Australia, Aust. J. Ecoi. 5: 9-14. Bell. D. T. and Loneragan. W. A. (1985) — The relationship of fire and soil type to floristic patterns within healhiand vegetation near Badgingarra. Western Australia. J, Rov. Soc. Aust.. 67: 98-109. Biswell, H. H. ( 1 974). Effects of fire on chaparral In Fire and Ecosystems. Kozlowski, T. T. and Ahlgren, C. E. (eds) pp. 321-364. Academic Press. N. Y. Burking. R. C. and Kessell. .A. C. (1984). — Damage report of the west- coastal wildfire and Us effects on the Western Australian Beekeeping Industiy. Western .Australian Department of Agriculture Report. Burrell. J. (1986). — The invasion of Victoria heathlands by Leptospermum laevigatum. Proc. Ecot. Soc. Aust., 3: 39. Christensen. N. L. and Muller, C. H. (1975).— Effects of fire on factors controlling plant growth in Adenostoma chaparral. Ecol. Monogr., 45: 29-55. Christensen, P. E. and Kimber, P. C. (1975). — Effects of prescribed burning on the flora and fauna of south-west Australian Forests. Proc. Ecol Soc. Aust.. 9: 85-106. Churchill, D. M. (1968), — The distnbuiion and prehistory of Eucalyptus diversicolor F. Muell.. E. morginata Donn ex Sm. and E. calophvlla R. Br. in relation to rainfall. Aust. J. Bot.. 16; 125-151. Cushwa. C. T., Marlin, R. E. and Miller. R. L. (1968).— The effects of fire on seed germination. J. Range Manage.. 21: 250-254. Davies. J. (1985). — The impact of fire on the apiary industry. In Fire Ecology and Management in Western Australia Ecosystems. J. R. Ford, (ed.) pp. 209-213. WAIT Environmental Studies Group Report No. 14. Bentley, Western Australia. Ewart, A. J. (1908). — On the longevity of seeds. Proc. Roy. Soc. Viet 2V 1 - 210 . Floyd. A. G. (1966). — Effect of fire upon weed seeds m the wet sclerophyll forests of northern New South Wales. Aust. J. Bot.. 14; 213-256. Floyd, A. G. (1976). — Effect of burning on regeneration from seeds in wet sclerophyll forest. Aust. For. 39: 210-220. Gill, A. M. (1975). — Fire and the .Australian flora: .A review. Aust. For., 38- 4-25. Gill, A. M. (1977). — Management of fire-prone vegetation for plant species conservation in Australia. Search. 8: 20-26. Gill, A. M. and Ingwersen, 1. (1976). — Growth of Xanthnrrhoea australis R Br, in relation to fire. J. .ippl. Ecol.. 13: 1 95-203. Hnatiuk, R. J. and Hopkins, A. J. M. (1980). — Western Australian species- rich Kwongan (sderophyllous shrubland) affected by drought Au.st. J. Bot., 28; 573-585. Hopkins. A. J. M. ( i 979). — Ecological aspects of the lignotuber. Abstract of paper presented at ihe Symposium on the Biology of Native Australian Plants. Perth. W. A. 6-10 August. 1979. Kccley, J. E. and Zcdler, P H. (1978). — Reproduction of chaparral shrubs after fire; A comparison of sprouting and seeding strategies. Amer •Midi. Nat.. 99: \42-\6\. Lamonl. B. B.. Hopkins. A. J. M. and Hnatiuk. R. J. (1984).— The flora- composition. diversity and origins. In Kwongan-Planl Life of the Sandplams. Pate, J. S. and Beard. J. S.(eds.) pp. 27-50, University of Western Australia Press, Nedlands. Western .Australia. Leigh. J. and Boden, R. (1979). — Australian flora in the endangered species convention-ClTES. Australian National Parks and Wildlife Service Special Publication No. 3, pp, 93. McArthur. A. G. and Chene)/. N. P. (1966). — The characterization of fires in relation to ecological studies, iust. For. Res.. 2: 36-45. Marchant, N. G. (1973). — Species diversity in the south-western flora. J. Roy. Soc. West. Aust., 56: 23-30. Marchant. N. G. and Keighery. G. J. (1979).— Poorly collected and presumably rare vascular plants of Western Australia. Kings Park Res. Notes No. 5. Morrow, P. A. ( 1 977). — Host specificity of insects in a community of three co-dominant Eucalyptus species. Aust. J. Ecol. 2: 89-106. Myers. N. (1979 ). — The Sinking Ark. Pergamon. Oxford. England. 127 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Russell, R. P. and Parsons, R, F. (1978). — Effects of time since fire on heath floristics at Wilson’s Promontory, Southern Australia. Aust. J. EcoL, 26; 53-61. Rye, B. L. (1982). — Geographically restricted plants of southern Western Australia. Dept. Fish Wild!. West. Aust. Rept. No. 49. Shea, S. R., McCormick, J. and Portlock, C. C. (1979). — The effect of fires on the regeneration of leguminous species in the northern jarrah (Eucalvpius marginaia Sm.) forest of Western Australia. Aust. J. EcoL. 4: 195-205. Siddiqi, M. Y., Carolin, R. C. and Myerscough, P. J. (1976). — Studies in the ecology of coastal heath in New South Wales. III. Regrowth of vegetation afterfire. Proc. Linn. Soc. N. S. W.. 101: 53-63. Specht, R. L. (1979). — Heathlands and related shrublands of the world. In Ecosystems of the World. Vol. 9A. Heathlands and Related Shrublands R. L. Specht. (Ed.) pp. 1-18, Chapter 1. Elsevier; Amsterdam. Smith. F. G. ( 1 969). — Honey Plants of Western Australia. West. Aust. Dept. Agric. Bull,. 3618. Specht. R. L, Rayson. P. and Jackman, M. E. (1958). — Dark Island heath (Ninety-Mile Plain, South Australia), VI. Pyric Succession: changes in composition, coverage, dry weight, and mineral nutrient status. Aust. J. Bot., 6: 59-88. Stone, E. C. and Juhren, G. (1951). — The effect of fire on the germination of the seed of Rhus ovata Wats. .Amer. J. Bot.. 38: 368-372. Went, F. W., Juhren, G. and Juhren, M. C. (1952). — Fire and biotic factors affecting germination. 33: 351-364, Appendix 1. Vascular plant species of the northern sandplain honey bee pasture region. Annotated information where known includes edaphic preference, most preferred fire response mode, the juvenile period between fire and flower production and the dates of 1981 where flowering was recorded. Annotation codes. 'Flowering mainly restricted to period 1 or 2 years following fire; ^Flowering restricted to period 2-4 years after fire; ’Flowers earlier in season in 2-4 year old sites; ■‘Species that are apparently rare and have a restricted geographic distribution (after Marchant and Keighery 1979); ^Species poorly known (after Marchant and Keighery 1979); ^Distribution restricted to an area with 100 km diameter (after Marchant and Keighery 1 979): ’Distribution restricted to an area with 1 60 km diameter (after Marchant and Keighery 1979). Species GYMNOSPERMAE Cupressaceae Actinostrobus acuminatus Parlat ANGIOSPERMAE-MONOCOTYLEDONAE Cyperaceae Caususdtocea R. Br Mesomelaena stygia (R. Br.) Nees Mesomelaena teiragona (R. Br.) Benth Schoenus eurvifolius {B.. Br.) Benth Haemodoraccac Anigozanthos humiUs Lindl Blancoa canescens (Lindl.) Bail! Conostylis androstemma Lindl Conoaiytis aurea Lindl Conostylis filifolia F. Muell Conostylis teretifolia J. W. Green Haemodorum paniculatum Lindl Macropidiafuliginosa Drum Phlehocarya ciliata R. Br Iridaceae Patersonia ocddentalis R. Br Liliaceae Burchardia umbellala R. Br Johnsonia puhescens Lindl Laxmannia grandiRora Lindl La.xmannia sp. aft. sessiUflora Thvsanotus gtaucus Endl Thysanotus muhiflorus R. Br Orchidaccae Diuris longifotiQ R. Br Glossodia 6ru«t>n«5(Endl.) A. S. George... Prasophyllum parviflorum Lindl Pterostviis nana R. Br Thelymitra sp. aff. variegata Lindl Poaceae Neurachne alopecuroides R. Br Reslionaceae Alexgeorgia arenicola Carlquist Anartkna tae\'is R. Br Ecdeiocolea monosiachya F. Muell Hypotaena exsuka R. Br Lyginea harbata R. Br. Xanihorrhoeaccae Caleciasia cvanea R. Br Dasypogon hromeliifoUus R. Br Kingia australis R. Br..., Xanthorrhoea reflexa Herbert.’ ANGIOSPERMAE-DICOTYLEDONAE Apiaceae Xanthosia huegelii (Benth.) Steud Edaphic Preference Fire Response Juvenile Period Flowering Period 456789 9 11 29 28 10 7 5 1 29 20 Generalist Sprout X Generalist Sprout Generalist Sprout Generalist Sprout Generalist Sprout Sand Sprout <2 X X Sand Sprout <2’ X X X X X Laterite Sprout X Generalist Sprout <2 X X Sprout <2 X X Generalist Sprout Sprout Laterite Sprout 'Sprout Generalist Sprout <2 X Generalist Sprout <2 X Sand Seed <2 XXX Generalist Seed X X Seed Sand Seed X Sand Seed ^2 X Laterite Sprout 1 X Laterite Sprout X Laterite Sprout X X Sprout Laterite Sprout Sprout Generalist Sprout Generalist Sprout Generalist Sprout Sand Sprout Generalist Sprout Generalist Sprout <2 X X X X X Sand Sprout <2' X X Laterite Sprout <2' Laterite Sprout <2' Sand 128 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Species Asteraceae Angianthus tomentosus Wendle Arcintheca calendula (L.) Leogns Helipterum cotufa (Benth.) D. C Podotheca gnaphalioidesiGrah.) F. Muell Podotheca pvgmaea A. Gray Vrsinia amhemoides(L.) Poir Caesalpintaceae Labichea punctata Bemh Casuarinaceae Allocasuahna humilis Otto & Dietr Allocasuarina mkrostachya Miq Chloanthaceac Lachnostachys verbascifoUa F. Muell Pityrodia bahlingii (Lchm.) Benth Dilleniaceae Hibbertia acerosa{K. Br.) Benth Hibhertia aurea Steud Hibbertia crassifoiia (Turcz.) Benth Hibhertia glaberrima F. Muell Hibbertia huegilii {End\..) F. Muell Hibbertia hypericoides (DC.) Benth Hibbertia ptiosa Steud Hibbertia racemosa (Endl.) Gilg Droseraceae Drosera drummondii Lehm Drosera eryihrorrhiza Lindl Drosera heterophylla Lindl Drosera macrantha Endl Drosera menziesii R. Br Drosera paleacea DC Drosera pallida Lindl. Epacridaceac Andersonia heterophylla Send Andersoma lehmanniana Sond Astrohma microdonm (F Muell.) Benth Astrolnma pallidum R. Br. Astrohma serratifolium (DC.) Druce Astruloma stomarrhena Send Astrohma xerophyllum (D(7.) Sond Conostephium pendulum Benth Leucopogon striatus Leucopogon conostephioides DC Leucopogon crassijhrus F. Muell. ^ Leucopogon cryptanthus Benth Lysinema ciliatum R. Br Euphorbiaceac Monotaxis grandiflora Endl. Goodeniaceae Dampiera juncea Benth Dampiera lindleyi De Vnesse^ Dampiera spicigera Benth Dampiera slenostachya E. Priizel. Lechenaulfia hihba Lindl. Lechenaultia flonbunda Benth Lechenauluaformosa R. Br. Scaevola canescens Benth Scaevola gfandulifera DC Scaevota paludosa R. Br...„ Velleia trinervis Labill Verreauxia vilhsa E. Pritzel Haloragaceae Glisschrocaryon aureum var. aureum (Lindl.) Orch. Lauraceae Cassylha pubescens R. Br.* Lamiaceae Hemiandra pungens R. Br Lobeliaceae Lobelia gibbosa Labill. Appendix 1 — continued Flowering Period Edaphic Fire Juvenile Preference Response Period 4 5 6 7 8 9 9 11 12 29 28 10 7 5 I 29 20 19 Generalist Seed <2 Seed I Sand Seed Sprout I Seed 1 Seed 1 Laterite Sprout <2 X Generalist Sprout X X X Laterite Sprout Sand Sprout Sand Sprout <2 X Sand Sprout X X Laterite Sprout X Generalist Sprout <2 X X X Sand Seed & Sprout <2 X X X Generalist Sprout <2 X X X Generalist Sprout <2 X X X X Laterite Sprout X X X Sand Sprout Laterite Seed Sprout X Generalist Sprout <2 X X Sand Sprout Sprout <2 X Seed Sand Seed & Sprout <4 X X X Generalist Sprout <4 X X X Laterite seed <4 X X X X X X Laterite Sprout <2 X X X Generalist Sprout <2 X Generalist Sprout <2 X X X Sand Seed <4 X Sand Sprout <5 X Generalist Sprout & <2 X X X X X X X Seed Generalist Seed <5 X X Sand Generalist <5 X X X Sand Seed s4 X X X Generalist Laterite Sprout Sand Sprout X X Laterite Sprout <2 X Seed Sand Sand Sprout X X Sand Sprout X X Sand Seed Sand Seed 2 X X Sand Seed X Sand Sprout X X 129 Journal of the Royal Society of Western Australia, Vol. 69. Part 4, 1987. Species Loganiacea Logania spermacocea F. Muell Loranlhaceac Nuylsiajloribunda (Labill.) R. Br Malvaceae Plaglanihus monoica Ewart Mimosaceae Acacia auroniiens Lindl Acacia cedroides Bcnth Acaaa lasiocaroa Bemh Acaaa puichella R. Br Acaaa spinosissmia Benth Acacia stenopiera Benth Acacia terctifolia Benth Myrtaccae Baeckea camphorosmae Endl Baeckea cnspijhra F. Muell Baeckea grandiflora F. Muell Beauforiia hraaeosa Deils Beaufonia elegans Schau Beauforiia eriocephala W. V. Fiizg Beaufonia squarrosa Schau Calothamnus sanguineus Labill Calothaninus torulosus Schau Cahtrix hrachyphylla Turez Cahlrix Bavesccns A. Cunn Cafyrri.x muricaia F. Muell Conolhamnus irinervis Lindl Dar^'inia speciosa (Meissn.) Benth Eremaea beauforlioidcs Benth Eremaea fimhriala Lindl Eucah-ptiis macrocarpa Hook Eucalyptus pendens Brooker’ Eucalyptus todtmna F. Muell Hypocatymma xanihopetalum F. Muell. Leplospermum spinescens Endl Melaleuca depressa Diels Melaleuca scabra R. Br. Melaleuca irichophylla Lindl Pileanthus filifolius Meissn Veriicordia densiflora Lindl Verticordiagrandiftora Endl Veriicordia grandis Drumm.’ Veriicordia spicala F. Mtiell. Veriicordia ovalifolia Meissn Papilionaceae Davicsia aphyllaiF. Muell.) Benth Daviesia Japhnoides Meissn Daviesia divaricaia Uenlh Daviesia epiphylla Meissn.* Daviesia juncea Sm Daviesia nudiflora Meissn Daviesia pectinata Lindl Daviesia pedunculula Benth Daviesia prnisii Meissn Daviesia quadrilatera Benth Daviesia striata Turez Gastrohbium Indens Meissn.^ Gasirolobium ilic(fblium Meissn Gastrolobtum obovaluni Benth Gastrohbium oxylohioides Benth Gasirolobium spinosum Benth Gastrohbium knightianum Lindl Hovea siricia Meissn. L Isotropis cuneifoUa (Sm.) Domin Jacksoma Jloribunda Meissn Jacksonia restioides Hueg Jacksoma sternbergiana R. Br Kennedia prostrata R. Br Oxvhbium capitatum Benth Sphaerolohium macranthum Meissn Phytolaccaceae Gynystemon ramuhsus Desf Tersonia hrevipes Moq Pittosporaceae BiUardiera bicolor (Putterl.) E. M. Bennett Appendix 1 — continued Flowering Period Edaphic Fire Juvenile Preference Response Period 4 5 6 7 8 9 9 1 1 12 29 28 10 7 5 1 29 20 19 Generalist X X Sand Sprout Sprout & Seed Laterite <2 X X Laterite Seed Generalist Seed <5>2 X X Sand Sprout <5>2 XXX Laterite X Generalist X Laiertite Laterite Sprout <4>2 XXX Seed Sand Seed <2 X X Laterite Sprout & <2 X X Seed Seed <2 X X Generalist Sprout <4 XXXXXXXXX Laterite Sprout :<5>2 X Sand Sprout <2 Sand X X Sand Generalist Sprout ^2 X Generalist Sprout Generalist .Sprout XXX Generalist Seed Sand Sprout Laterite Sprout Sand Sprout Generalist Sprout ^2 XXX Generalist Sprout X Generalist Sprout X X Generalist Sprout XXX Laterite Sprout <4>2 X X Generalist <4 X X Generalist X Generalist Seed <4 X Sand Sprout <2 XXX XX X X Laterite Sprout -2 X Laterite Sprout X X X X Sand Sprout <2 X Laterite Sprout Laterite Sprout X Generalist Sprout <2 X X Generalist Sprout <2 X X Generalist Sprout <2 X Generalist Sprout X X Sand Seed ^2 XXX Laterite Sprout X Laterite Seed X X Laterite Laterite Sprout <5 X Laterite Sprout Laterite Seed X X Laterite Sprout Generalist Sprout <2 3 X X X Generalist Sprout <22 X X Sand Seed & <2 X X Sprout Generalist Laterite Seed Laterite Seed Laterite Laterite X Sand Seed <2 X Sand Seed <2 X X X X Laterite 130 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Species Polygalaceae Comesperma cafyniega Labill Proteaceac Adenanthos eygnorum Dies Banksia aiienuata Meissn Banksia candoHeana Meissn.’ Banksia menziesii R. Br Banksia prostrata R. Br Banksia sphaerocarpa R. Br Banksia .sp. aff. sphaerocarpa R. Br Conospermum acerosutn Lindl Conospermum incurvum Lindl Conospermum nervosum Meissn.® Conospermum stoechadis End! Conospermum triplinervium R. Br Dryandra hipinnalifida R. Br Dryandra cartinoides Meissn Dryandra kippistiana Meissn Dryandra nana Meissn . '' Dryandra mvea R. Br. Dryandra sessUis (R. Br.) Druce Dryandra shuttftrAvrthiana Meissn Dryandra (rideniata Meissn. ® GreviUea pUuhfera (Lindl.) C. A. Gardn. GreviUea shultleworthiana Meissn GreviUea synapheae R. Hakea auncu/ata Meissn. Hakea conchifoha Hook Hakea corymiwsa R. Br Hakea cosiata Meissn Hakea JlabeilifoHa Meissn, ® Hakea incrassata R Br Hakea (issocarpha R. Br Hakea obUqua R. Br Hakea prostrated. Br Hakea rusctfolia Labill Hakea sulcata var. scoparia R. Br Hakea undufata R. Br. Isopogon asper R. Br Isopogon linearis Meissn. ’ Isopogon teretifoUus R. Br. Lambvriia mullijlora Lindl Persoonia dUlwynioides Meissn Peirophde inconspicua Meissn Petrophile linearis R. Br Petrophde macrosiachya R. Br Petrophile media R. Br.... Petrophile serruriae R. Br. Petrophde striata R. Br Stirlingia latifolia (R. Br.) Steud Stirlingia simplex Lindl Strangea cyanchtcarpa F. Muell.’ Synaphaed petiolaris R. Br. Synaphaea polymorpha R. Br Rhamnaceae Cryptandra arbutijlora Fenzl. Crypiandra pungens Steud. Spyndium irideniatum (Steud.) Benth Spyndium sp. aff. iridenialum (Steud.) Benth. Trymalium ledijolium Fenzl Rutaceae Boronia ramosa (Lindl.) Benth Eriostemon spicaius A. Rich Stackhousiaceac Stackhousia brunonsis Benth Stackhousia puhescens A. Rich Sterculiaceae Commersonia pulchella Turez Lasiopetalum dnimmondii Benth.’ Lasiopeialum sp. Thomasia grandijlora Lindl Appendix 1 — continued Flowering Period Edaphic Fire Juvenile Preference Response Period 4 5 6 7 8 9 9 29 28 10 7 5 1 29 Generalist Sprout <2 X Sand Seed Sand Sprout & <2 Seed Sand Sprout <4>2 X X X X X Lalerile Sprout <2 X X X X X X Laierite Sprout <2 Generalist Seed Sc Sprout Generalist Seed & Sprout -2 X X X Sand Sprout X Sand Sprout <5 Laterite Seed Sand Sprout <2 X Sand Sprout <2 X X X Laterite Sprout & Seed Generalist Seed :<4>2 X Laterite Seed <2 X X Generalist Sprout <2 X Generalist Sprout X Laterite Seed ^2 X Generalist Sprout Sand Sprout <2 X Generalist Sprout X X X Sand Sprout X Laterite Seed <2 X X X Latcntc Sprout <2 X X X Laterite Sprout <2 X X X X Generalist Sprout X Sand Seed <6 X Generalist Sprout Laterite Sprout <2 X Generalist Sprout X X X X Sand Seed <4 X Sand Sprout X Sand Laterite Seed X Laterite <2 X Sprout <4>2 X Generalist Sprout <2 X X Laierite Sprout <4 Generalist Sprout Sc X Seed Sand Sprout <2 X Generalist Sprout <4 X Generalist Seed <2 X X Generalist Sprout & X Seed Sand Sprout Sprout <2 X Sand Sprout Laterite Sprout X X Generalist Sprout Sc <2 X X Seed Laierite X X X Generalist Seed ^2 X X X Laterite Laterite Generalist Sprout <2 X Laterite Sprout & <2 X Seed Sand Sand Generalist Sprout ^2 Laterite Sprout <2 X X X 11 12 131 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Species Stylideaceae Stylidium adpressum Bcnth Stvlidium lepiophyHum DC Stvlidium piliferum ssp. minor (Mildbr.)Carlq Stylidium tepens R. Br. Thymeleaceac Pimelea angusd/olia R. Br Pimeleafloribunda Mcissn Pimelea imhricata R. Br. Pimelea suaveolens (Endl.) Meissn. Pimelea sulphurea Meissn Tremandraceae Tetratheca confertifolia Steetz Violaceae Hybanthus calycinus (Sieud.) F. Muell Hybanthus Jloribundus (Walp.) F. Muell Appendix 1 — continued Edaphic Preference Fire Response Juvenile Period Flowering Period 456789 9 11 12 29 28 10 7 5 1 29 20 19 Seed Seed Laterite Seed <22 X Generalist Seed <2 XXX Generalist Sprout Seed Laterite Seed <22 Sprout & X Seed Laterite XXX Laterite <2 X X 132 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Northern Sandplain Kwongan: community biomass and selected species response to fire by Jeanette C. Delfs, John S. Pate, and David T. Bell Department of Botany, University of Western Australia, Nedlands, WA 6009. Received October 1986. Accepted March 1987. Abstract The mixed, taxonomically diverse shrublands of the Northern Sandplains near Badgingarra, Western Australia recover rapidly following fire and by seven years, above ground biomass has reached a maximum of about 16-18 t ha"'. Such rapid build up of biomass is typical of fire-prone communities of kwongan dominated by long-lived autoregenerating species. Sampling from a series of deep sand sites of known fire history provided material for case studies of biomass recovery and development of polycarpic fire ephemeral (Tersonia brevipes Moq. in DC), four obligatory reseeding species (Leucopogon conostephioides DC.. Petrophile media R. Br., Beaufortia edegans Schauer. and Hakea obliqua R. Br.), a species normally exhibiting both seed- and resprouting- regeneration {Jacksonia floribunda Endl.), and two long-lived resprouting species never recorded at the sites as currently regenerating from seed {Hibbertia hypericoides CDC.) Benth. and Hypocalymma xanihopetahan F. Muell.). Very great differences were observed between representatives of each fire response category in growth rates and shapes of developing shoot canopies of the species. Community biomass was directly correlated with foliage projective cover allowing an easily obtained value to estimate fire fuels. The implications of the data are considered in relation to the prediction of fuel loads and design of controlled burning regimes for the region. Introduction Upland regions in the Irwin District of the kwongan of the South-West Botanical province (Beard 1980) are dominated by shrublands of highly uniform visual appearance but of great floristic richness and diversity (Lamonl et ai. 1984. Bell and Loneragan 1985). The mediterranean weather pattern of the region, by providing cool wet winters and long dr>' summers, promotes a rapid accumulation of above ground biomass between successive fires. These conditions, compounded with the tendency for certain plant species to be highly flammable in living or dead state (Pompe and Vines 1966), present serious summer fire control problems for land managers in the region, especially where fire-prone natural plant communities are intermixed with farmland committed to arable crops or pasture. Fire management in the Northern Sandplains is aimed at fulfilling two major criteria. First and foremost farmers and pasiuralists must be protected from wildfires emanating from adjacent native vegetation. Secondly, sufficient areas of native plant communities must be maintained for conservation purposes, while still serving the needs of a commercial apiculture industry, the wildflower seed and cut-flower trades, and tourism (Bell et ai 1984). These areas of native vegetation must in turn be protected from fires originating from roadsides or burning operations in adjacent farmland. Techniques for estimating fuel loads in the shrublands of the Northern Sandplain have recently been developed (Schneider and Bell 1985), and certain general characteristics of the response to fire by vegetation in the region have been considered in relation to apiculture (van der Moezel et ai 1987). The vegetation generally increases to a near maximum value of 70% foliage projective cover within 7-10 years (Bell et ai 1984), rates of recovery being more rapid in plant communities of the laieritic peneplain surfaces than on adjacent quartzitic sands. The floristic composition of these two edaphically-distinct substrata is also quite different, so differing recovery profiles might relate as much to the species present as to contrasting soil types (Bell and Loneragan 1985). The first objective of this study was to document community foliage canopy cover and biomass changes occurring after fire in the vegetation of a scries of deep sand sites, using a combination of projective cover density measurements and direct assessments of dead and living plant material harvested from randomly selected quadrats within a series of sites of known fire history. A second objective was to examine in detail patterns of biomass recovery in a number of common species typifying the major classes of fire response (see Bell et al. 1984) shown by flora of the study area. The data obtained are discussed in relation to the development of rational and effective fire management policies for the region. 133 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Methods The study area was centred around Badgingarra (30°23’S, 115"30’E), approximately 200 km north of Perth, Western Australia. Individual sites extended from the Badgingarra National Park (30"20'S, 115“25’E) northwards to the region of Juricn Road (30°14’S. 1 1 5M6’E). They were selected on the basis of records of the Western Australian Bush Fires Board to represent stands burned 5 and 9 months, and 2. 3. 4, 5, 6, 7, 8. 11, 12 and 17 years prior to sampling in March 1982. The study sites lav over deep quartzite, nutrient-poor sands, analytical data for which have been recently published (Pate et al. 1985). Records of community above-ground biomass were determined by collecting all living and dead plant material from ten randomly selected 1 m^ quadrat samples at each site. The samples from each quadrat were weighed individually in the field using spring balances, and weighed subsamples of bulked material from each site taken back to the laborator\' for oven drying (65'C), to enable field biomass fresh weight data to be converted to dry matter. Each quadrat was assayed for foliage projection cover betore its biomass was collected, so that relationships between biomass and cover and between biomass and age since last fire could be determined by linear regression analysis. (A) 60 - 2 £ 40 - > o o ■H. 2 4 6 8 10 12 14 16 (B) < 12 O II 2 4 6 6 10 12 14 16 AGE (YEARS) A detailed analysis of develoing canopy structure was made for 8 common shrubby species (Table 1) which collectively included all major categories of response to fire represented in the community. Twenty individuals of a species were collected from all sites at which that species was present. The above ground parts of each individual plant w-cre air dried in intact state and then partitioned horizontally into a scries of 10 cm segments (except for the large species Jlakea obliqua with 20 cm increments). Each stratum of the shoot was then measured for foliage diameter and dry weight. Combining data for each sample of 20 plants, shrub structure profiles were then constructed depicting mean canopy shape and weight distribution for each species for the range of ages since last burn (Gibberto et al. 1977). Results and discussion Increase in community biomass following fire Canopy cover and above-ground biomass in the study area increase rapidly in the first 7 years following fire, thereafter tending to remain at levels of approximately 16-18 t ha'* (Fig. 1). This period of rapid increase in biomass correlates with the main flush of regeneration of woody shrubs establishing from fire-resistant underground root stocks. A study ot 152 species from sandy habitats in the Badgingarra-Jurien region has indicated that 66% resprouted in such manner loilowing fire, and that at least a similar proportion of biomass of a site was likelv to consist of these resprouler species (Bel! et al. 1984'). Other studies on sandy sites frpm the region, and of the floras of neighbouring latcrilic sites have shown even higher proportions of resproutmg species (Bell and Loneragan 1985, van dcr Moezel et al. 1987). The plateau of biomass al approximately 16-18 t ha'* places the Northern Sandplain kwongan at the low' end of values for above-ground biomass recorded f 9 r mediterranean-climate shrubland ecosystems in California. France and the eastern Australian states (see Gray and Schlesinger 1981, Bell et al. 1984). Biomass achievements of Northern Sandplain shrublands are, Figure 1. — Foliage projective cover percentage (A) and total above ground biomass (litter plus above-ground plant material) (B) of deep sand shrubland sites in the Northern Sandplain region between Badgingarra and Jurien. Western Australia. however, more than double those reported for the open matorral of Chile (Mooney et al. 1977), and lie within the range of 11-26 t ha'* reported for mature shrub- dominated communities in southwest Cape Province, South Africa (Kruger 1977). Elsewhere in the kwongan of Western Australia, biomass values for closed shrub communities on calcareous sands at Two Peoples Bay near Albany were reported to have reached near- maximum biomass al 16 t ha** after 9 years regrowth (Bell et al. 1984); a site on deep sands south of Eneabba carried 7 t ha'* after 9 years (Hopkins and Hnatiuk 1 98 1 ), and a mature shrub-dominated stand of unknown age on depauperate grey sands at Tulanning Nature Reserve showed 13 l ha'* (Brown and Hopkins 1983). As mentioned above, relatively early achievement of a plateau in biomass with age in kwongan is probably related primarily to the large (Fig. IB) proportional contribution of sprouters to the ecosystem, but it might equally refiect a limited overall carrying capacity in the dry , nutrient-poor sites typical of this class of vegetation. Cover to biomass relationships Analyses of data from all study sites showed that foliage projective cover percentages were directly related to biomass according to the following regression equation (biomass (t ha'*) - 2.99 -t- 0.25 cover (%), df = 1 15, r = 0.85, p < O.OI). An essentially similar pattern of biomass recovery following fire is reported for South African fynbos communities (Kruger 1977), although the perennial herbaceous component of these communities is greater than kwongan. Other published data on cover density have provided estimates of vegetation recovery following fire in a number of mediterranean-climate vegetation types. For coastal heaths at Dark Island, South Australia, cover percentages increase steadily to a maximum of approximately 70% after 10 years since burning, and 134 Journal of the Royal Society of Western Australia, Vol. 69. Part 4. 1987. remain fairly constant for the next 1 5 years {Specht ct al. 1958). Cover values following fires in the chaparral of southern California, however, show two peak periods, the first after 2-5 years coincide with dominance by annuals, herbaceous perennials and short-lived subshrubs, the second peak, after 8-17 years, with the resurgence of larger woody shrubs typical of the climax mature vegetation of the region (Horton and Knaebcl 1955. Keeley and Kccley 1981). According to Spccht ei al. (1985) cover values for fire-prone vegetation possessing a herbaceous phase in its pyric succession arc generally not well correlated with biomass, due to the much greater weight to cover ratio of later stage samples (Specht c/ a/. 1985). The rapid regrowih of the Northern Sandplain shrublands means that stands achieve a capacity to support a fire very soon after the previous fire. Indeed, instances of fires burning through regions carrying only a three-year-old fuel load have been reported for the Beekeepers Reserve north of Jurien, albeit only under conditions of exceptionally intense late summer temperatures (>40“C) involving low humidities (<15%). high winds (>40 km hr'^) and dry fuel conditions (Burking and Kessell 1984). It is apparent, however, that on average, stands of age 2-6 years will have considerable less biomass than older counterparts and would be accordingly less prone to wildfires. Prescribed burning on a rotation of five to seven years, or even less, would therefore appear to be an eminently sensible means of reducing fuel loads to less dangerous proportions. However, this advantage must be weighed against potential problems in conservation of individual species, especially rare or restricted fire sensitive species which normally take a number of years after germination before commencing to flower and set seed (sec Hopper and Muir 1984). This w'ill constitute a particularly serious problem where the species in question retain their seed load in the plant canopy rather than in the soil (Table 1). Table 1 Characlerisiiics of species selected for detailed study on recovery after fire. Species Family Regeneration Mode Seed Store Leucopogon conostephioides.. . Epacridaceae Obligate seed Soil Hakea obtiqua Proieaceae regenerator Obligate seed Plant Beaufortia elegans Myrtaceae Obligate seed Plant Tersonia brevipes Phylolaceae regenerator Obligate seed Soil Petrophile media Proieaceae regenerator Obligate seed Plant Hibbertia hvpericoides Dilleniaceae regenerator Resprouter Soil Jacksonia floribunda Papilionaceae Both a Soil Hypocalymma xanthopetalum Myrtaceae resprouter and reseeder Resprouier Soil Growth and developing canofyy characteristics of selected species (a) Polycarpic fire ephemerals Monocarpic and polycarpic fire ephemerals arc relatively sparse in the Northern Sandplain in terms of number of species and relative biomass (Pate el al. 1 985) in comparison with the highly prolific post-fire herb flora of Californian chaparral (Muller el al. 1 968, Keeley and Keeley 1981). The species Tersonia brevipes is a typical short-lived polycarpic perennial of kwongan in showing fire-obligate germination, extremely fast early growth rates, early maturity, high reproductive effort in proportion to vegetative biomass, and a relatively short life span (see Pate et al. 1985). Fast early growth of these successional species is generally held to promote an immediate conservation of nutrients following disturbance such as fire, and thus minimize leaching losses of nutrients in such circumstances (Marks and Bormann 1972, Likens el al. 1978, Foster et al. 1980, Nilsen and Schlesinger 1981). Initial growth in Tersonia brevipes (Fig. 2a) was predominantly in a vertical direction though formation of a short lived leafy shoot, but a semi-woody creeping habit is then quickly attained through subsequent development of a number of basal axillary shoots. By four years the radiating stems of plants of the species may encompass an area up to 240 cm in diameter with all biomass restricted essentially to within 10 cm of ground level. By 4 years mean plant weight had reached 220 g, but by 5 years virtually all plants had senesced and died within the area (sec also Pate et al. 1985). 0 4 0-75 2 3 5 TERSONIA BREVIPES LEUCOPOGQN CONOSTEPHIOIDES Figure 2. — Mean shrub dimensions of the short-lived fire ephemeral, Tersonia brevipes. following fire in the Northern Sandplain shrublands (A). Histogram of dry weight for 10 cm increments of radial distribution from the root system. Mean shrub dimensions of obligatory reseeding species, Leucopogon conostephioides, following fire (B). Data includes height, 10 cm increment diameter and dry weight distribution and total dry weight. (b) Obligatory re-sceding species Patterns of regrowih following fire were essentially the same in the two seeder species Leucopogon conostephioides (Fig. 2b), and Beaufortia elegans (Fig. 3a). Over the first two years each species grew' mostly in a vertical fashion, but thereafter increasingly in diameter as well as in height. In Leucopogon conostephioides. mean plant above ground dr>' weight increased from 0.20 g plant'* at two years to 0.36 g after 3 years, 2.7 g after 5 years and 9.6 g plant * after nine years growth (Fig. 2b). Mean heights of the 2-. 3-, 5- and 12-year plants were II. 1 7, 23 and 44 cm, respectively. As plants aged, biomass tended to be distributed disproportionately toward the upper part of the stem, giving a decidedly ‘Top heavy” plant. 135 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Figure 3. — Mean shrub dimensions of obligatory reseeding species, Beaufortia elegaris (A), and Petrophilc media (B) following fire in the Northern Sandplain shrublands. Beaufortia elegans seedlings were generally larger than those of comparably-aged Leucopogon conostephioides. Plants of B. elegans growing in a 17-year old stand were nearly 100 cm tall and weighed more than 75 g (Fig. 3a). In these old plants the stratum between 60 and 70 cm above soil level contained the greatest amount of biomass. Petrophile media, another obligatory reseeding species, was unfortunately encountered only at sites within the age range 3-17 years. When mature, this species showed similar '*top heavy” biomass distribution profiles to those of Beaufortia elegans. although above ground parts of B. media plants were generally more than twice as large and heavy as B. elegans (Fig. 3b). The largest obligatory reseeding species encountered in the deep sand communities of the study region was Hakea obliqua (Fig. 4). In areas estimated to be 17 years old since the last fire, plants had a mean height of 2.7 m and a mean above ground dry weight of 1.44 kg. The distribution of this biomass with height was more uniform in this species compared with the previous three examples, as readily apparent from the generally spindly profile of the species in the field. Table 2 Linear regression equations and slalisiics for the relationship of mean plant height and age for four obligatory reseeding species of the Northern Sandplain shrublands. Regression Equation d.f v p Age (yr) = — 1.32 -f 0.41 Leucopogon conostephioides height (cm) 50.99 < 0.01 Age (yr) = — 2.89 + 0.21 Beaufortia elegans height (cm) 80.94 < O.OI Age (yr) = — 4.28+0.29 Petrophile media height (cm) 40.96 < 0.0 1 Age (yr) = — 2.90 + 0.07 Hakea obliqua height (cm) 90.99 < 0.01 Highly significant linear regressions between height and age were found for each of these four obligatory seed-regenerating species (Table 2). This relationship has already been suggested as a useful means for predicting age of sites for when fire records are not available (Bell 1985). We would now further suggest that, using such regression equations, and data on mean heights of a range of seeder species, one would have a simple method for estimating stand age at a site for which fire records were not available. Following this, using the age-biomass data of Fig. 1.. predictions could be made of fuel loads in the region, and thus determine whether or not a prescribed burn were both feasible and desirable. i6Gr DRY WEIGHT ( g ) 60 0 80 r I > DIAMETER (cm) 20 0 20 HAKEA OBLIQUA Figure 4. — Mean shrub dimensions of the obligatory reseeding species Hakea obliqua following fire in the Northern Sandplain shrublands. (c) Resprouting species also regenerating freely from seed Species which possess the ability both to resprout, and to establish abundant seedlings following fire may be considered to have distinct advantages over species exhibiting only one or other of these regeneration strategies. For instance, Keeley (1 977) noted that the most abundant chaparral species in California. Adenostomafasciculalum. reproduces following fire both by resprouting and from germinating seed, as do a number of successful species of the Californian coastal sage vegetation (Malanson and O’Leary 1 982). In the Northern Sandplain of Western Australia, Jacksonia florihunda. constitutes a common species possessing the above mentioned abilities (Bell et al. 1 984). By being able to distinguish between unscarred seedlings established following the last fire and fire- scarred resprouting individuals which had clearly survived at least one fire at the site, it was possible to compare growth patterns and morphologies of virgin seedlings and previously established survivors. Seedlings of Jacksonia florihunda were then found to produce above-ground biomass al ver>' slow rates, yielding after eight years, heights of approximately 40 cm and above ground dry weights of only 2.7 g dry weight (Fig. 5a). By comparison resprouling individuals in the some 8-year study site averaged nearly 80 cm in height and carried above-ground biomass averaging of 79 g dry weight (Fig. 5b). 1 36 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. (A) 10 • 0. ♦ (j 0 04 075 3 (B) 04 075 2 3 120r JACKSONIA FLORIBUNDA JACKSONIA FLORIBUNDA ( REGENEHATION FRO« SEED) (REGENERATION FROM RESPROUTINGl Figure 5. — Mean shrub dimensions of Jacksonia Jloribunda seedlings (A) resprouling individuals (B) following fire. There are inherent difficulties in interpretation of data on biomass of surviving plants of a resprouter species across a sequence of sites because the mean age of survivors ot one population may be very different from that of another site. This is especially so if earlier fires or other environmental events have given very different patterns of recruitment at the sites in question. The present data accumulated iov Jacksonia floribunda typify this problem; c.g. the population of plants in the region burned two years prior to sampling had a mean total di 7 weight of 1 26 g plant"’ compared with only 79 g"' in a neighbouring site known to have had an 8 year interval since the last burn (Fig. 5b). (d) Long-lived resprouting species, regenerating extremely rarely from seed Rapid regeneration of resprouter species was also demonstrated for two common non-clonal, resprouting species, Hibbcnia hvpericoides (Fig. 6a) and Hypocalymma xanthopetalum (Fig. 6b). These two species are exceptionally common members of the deep sand communities of the Northern Sandplain (Bell and Loneragan 1 985), but, in the authors experience, have never been observed to be regenerating successfully from seed (Belj cf al. 1984). As with other long lived sprouters, each exhibits highly heterogenous population structures in terms of number, mass and length of regenerating shoots per plant, tap root diameter, and inflorescence number and fruit reproduction. Unfortunately, the real age of resprouting individuals cannot be assessed with certainty, as growth rings in tap roots are not readily apparent, especially where root stocks have become split or partly destroyed by termites. In any event there is no proof that any growth rings which are present have been produced on a strictly annual basis. HIBBERTIA HYPERICOIOES HYPOCALYMMA XANTHOPETALUM Figure 6. — Mean shrub dimensions of the long-lived autoregenerating species. Hibberiia hypericoides (A) and Hypocalymma xanthopetalum (Bl following fire. These two long-lived resprouting species showed essentially similar habit to resprouting Jacksonia floribunda. and, regardless of size and age, mean diameters of their above ground stems were consistently greater than that of the earlier-mentioned seed regenerating species. In contrast to seeders, resprouters gained dry matter most rapidly over the first three years after a fire (Fig. 6), and with age, showed no tendency for the biomass of their shoots to become concentrated especially towards the top of their shoots. GENERAL CONCLUSIONS Species of the Northern Sandplain that regenerate by resprouting appear to have inherent advantages over obligate seeder species in possessing a deeply penetrating massive root stock, from w'hich nutrients can be mobilized to give quick recovery of above ground biomass after fire. The multiple shoots generated from these root stocks give the regenerating shrub a highly characteristic shape, with wide basal diameter and biomass initially concentrated mainly at the base of the plant. Regrowth of resprouters is very rapid in the first 2-3 years after fire followed by a slow and gradual increase over at least the next 1 4 years. The seed regenerators studied in the Northern Sandplain shrublands have the same general habit as reported for other fire-sensitive Western Australian species (Baird 1977). Typically a single main stem is established and persists, the root system is typically shallow and of fibrous character (Dodd et al. 1 984), and biomass is eventually located mostiv in the upper reaches of the plant. As shown by the silhouettes of shoot shape and mass distribution described in this paper those of seeders contrast markedly with those of root- grown resprouters. Moreover, since the seeders tend to establish in spaces between the regenerating crowns of the resprouters, and with time may even overtop the sprouters, both arc able to coexist successfully for many years of a post fire interval. Indeed, a properly balanced mix of seeders and sprouters, with essentially complementary' shoot canopies and rooting morphologies, is likely to maximize utilization of existing ground cover and resources of water and nutrients. Within this framework also, fast growing 137 Journal of the Royal Society of Western Australia, Vol. 69, Part 4. 1987. cphemerals such as Tersonia brevipes occupy a critical role early in a pyric succession by progressively recovering nutrients released from fire into plant biomass (see Pate c/ a/. 1985). Were it possible to extend information on canopy shape and weight distribution to all major species of some chosen aged community, and to combine this with measurements of density of these species, it would be possible to construct computer-simulated graphical representations of typical biomass structure, and thus assist in predicting how fuel loads arc distributed over time and space. Such information would be particularly valuable to a belter understanding of fire management of the community. This study has looked in detail at the responses to fire of only a small sample of species from a highly diverse fiora. For the meantime, faced with a paucity of data on regeneration strategics, ecologists must follow a conservative path when using fire as a management tool in these regions (Bell ei ai 1984. Hopper and Muir 1984). On the one hand, there is the danger that loo- frequeni fires might result in the loss of those obligatory seed-regenerating species which take an unusually long time to achieve first reproduction (van der Moezei ei al. 1987). Important apicultural species might well fall within such a category. On the other hand, long periods of fire prevention in kwongan generally lead to the build up of dangerous levels of fuel, and thus increase the possiblility of large scale wild-fires sweeping through shrubland and intervening paslurcland. Until the Norther Sandplan cscosystcm is much better understood, fire management policies should consist of planned mosaics of strategically reduced fuel zones, enclosing less frequently burnt regions, in which already indentificd, endangered seeder species might be able to survive. Designation of such a policy within National Parks, bee pastures, and specific recreation areas would further ensure sensible planned long-term maintenance of the species and their unique parent communities in the interests of all parlies concerned. Acknowfedgcmems.This, study was carried with funding assistance from the Western .Australian Bushfires Board, and the Honey Research Board of the Australian Department of Primary Industries. Dr W. .A. Lonera^n provided valuable assistance in the field and provided constructive criticism to early drafts of the manuscript. Joint funding of the Senior Lectureship of Dr D. T. Bell in the Department of Botany by Alcoa of .Australia Ltd and Western Collieries Ltd is gratefully acknowledged. References Baird, A. M. (1977). — Regeneration after fire m King's Park. Perth, Western Australia. / Roy. Soc. West Ausi . 60: 1-22. Beard, J. S. (1980). — A new phytogeographic map of Western Australia. .‘iusi. Herb. Res. Notes, 3; 37-58. Bell. D. T (1985). — Aspects of response to fire m the Northern Sandplam hcaihlands. In Fire Ecology ami Management in Western Australian Ecosy.'Hems, Ford. J. (cd.) pp. 33-40. West. Ausl Inst. Tech. Environ. Studies Grp Repl. No. 14. Bentley, Western Australia. Bell, D. 1 Hopkins. A. J M. and Pate, J. S. (1984). — Fire in the kwongan. 1 n KviX}ngan Plant Lije of the Sandpiatn. Pale. J . S. and Beard. J. S. (cds) PP- 178-204. University of Western Australia Press. Nedlands. Western Australia. Bell, D. T. and Loneragan. W, A. ( 1985). — The relationship of fire and soil type to flonstic patterns within healhland vegetation near Badgingarra, Western Australia. J. Rov- Soc. West. .Aust.. 67; 98- 109. Biswell. H. H. ( 1 974). — Effects of fire in chaparral. In Fire and Ecosystems. Kozlowski, T T and .Ahlgren, C. E. (eds). pp. 321-364. Academic Press. New York. Brown, J. M. and Hopkins. A. J. M. (1983). — The kwongan (sclerophyllous shrublands) of Tuianning Nature Reserve. Western Australia. .■iusl.J-Ecol..S:(}2-li. Burking, R. C. and Kessell, A. C. (1984). — Damage report of the west- coastal wildfire and its effects on the Western .Australian Beekeeping Industry. Western Australian Department of Agriculture Internal Report. Como. Western Australia. Dodd. J.. Heddic. E. M., Pale. J. S. and Dixon. K. W. (1984). — Rooting patterns of sandplain plants. In Kwongan-Plani Life of the Sandplam. Pate. J. .S. and Beard, J. S. (eds) pp. 146-177. University of Western Australia Press. Nedlands Western Australia. Foster, M. M.. Vitousek, D. M. and Randolph, P. .A. (1980). — The effects of ragweed {Ambrosia artemtsifolia L.) on nutrient cycling in a Isl- yearold field. .4mer. Midi. Nat.. 103: 106-1 13. Gibberto. .)., Mooney. H. A. and Kummerow, J. (1977). — Shrub structural analysis. In Chile-Calijornia Mediterranean Shrub Attas-A Comparative Analysis. Thrower. N. J. W. and Bradbury. D. E. (eds) pp. 144-147 Dowden. Hutchinson & Ross. Inc., Stroudsburg. Pennsylvania. Gray. J. T. and Schlesingcr. W. H. (1981). — Nutrient cycling in mediterranean type ecosystems. In Resource Vse by Chaparral and Matorral. Miller, P. C. (ed.) pp. 259-285. Springer-Vcrlag. New York. Hopkins. A. J. M. and Hnatiuk, R. J. (1981). — An ecological survey of the kwongan south of Eneabba, Western Australia. Aust. Wildl. Bull.,9: 1-33. Hopper, S. D. and Muir, B. G. (1984). — Conservation of the kwongan. In Kwongan-Platu Life of the Sandplain. Pate, J. S. and Bear. J. S. (eds) pp. 253-2^6. Llniversily of Western .Australia Press, Nedlands. Western Australia. Horton. J. S. and Kraebel. C, J (1955). — Development of vegetation after fire in the chamise chaparral of southern California. Ecology. 36; 244-262. Keeley. J. E. (1977). — Fire-dependent reproductive strategies in Arctostaphyhs and Ceanothus. In Symposium on the Evnironmenial Con.sequence.% of Fire and Fuel Management in Mediterranean Ecosystems. Mooney, H. .A. and Conrad, C. E. (eds) pp. 391-396. USD.A Forest Service Gen. Tech. Report WO-3, Washington, D.C. Keeley, J. E. and Keeley, S. E. (1981). — Post-fire regeneration of southern Californian chaparral. Amer. J. Boi.. 68: 524-530. Keeley. J. E. and Zcdler. P. II. (1978). — Reproduction of chaparral shrubs after fire: a comparison of sprouting and seeding strategies. Am. Midi Nat.. 99; 142-161. Krugcr. F. J. (1977). — A preliminary account of aerial plant biomass in fynbos communities of the Mediterranean-type climate zone of the Cape Province. Bothalia. 12: 301-307. Likens. G. E.. Bormann. F. H.. Pierce. R. S. and Reemers. W. .A. (1978). — Recovery of a deforested ecosystem. Science, 199: 492- 496. Lamont, B. B.. Hopkins, A. J. M. and Hnatiuk, R. J. (1984). — The flora- composition. diversity and origins. In Kwongan-Plant Life of the Sandplain. Pate. j. S. and Beard. J. S. (eds) pp. 27-50. University of Western Australia Press. Nedlands, Western Australia. Malanson. G. P. and O’Leary. J. F. (1982). — Posi-flre regeneration strategies of Californian coastal sage shrubs. Oecotogia. 53: 355- 358. Marks. P. L. and Bormann, F. H. (1972). — Revegelation following forest cutting; Mechanisms for return to steady-stale nutrient cycling, Science. 176; 914-91 5. Mooney, H. A.. Kummerow. J , Johnson. .A. W.. Parsons. D. J.. Keeley, S., Hoffman. .A.. Hays. R. I.. Giliberio, J. and Chu. C. (1977). — The producers-their resources and adaptive responses. In Comergent Evolution in Chile and Cdlifornia Mediterranean Climate Ecosystems. Mooney. H A. (ed) pp. 85-143. Dowden, Hutchinson and Ross Inc, Stroudsburg. Pennsylvania. Muller. C. H.. Hanawall. R. B. and McPherson. J. K. (1968).— Allelopaihic control of herb growth in the fire cvcle of California chaparral. Bull. Torrey Bat. Club. 95; 225-23 1 . Naveh. Z. (1974). — Effects of fire in the Mediterranean region. In Fire and Ecosystems. Kozlowski. T. T and Ahlgren, C E. (eds) pp. 401-434. Academic Press. New V’ork. Nilsen. E. T. and Schlesingcr. W. H. (1981). — Phenology, productivity and nutrient accumulation in the post-fire chaparral shrub Lotus scoparius. Oecotogia. 50: 217-224. Pate. J. S., Casson, N. E.. Rulto. J. and Kuo, J. (1985). — Biology of fire cphemerals of ihc sandplains of the kwongan of south-western Australia. .-Iwit y Plant Physio!.. 12: 641-655. Pompe. A. and Vines, R. G. (1966). — The influence of moisture on the combustion of leaves, .-iust. For . 30: 231-242, Schneider. B. H. and Bell. D. T. (1985). — A simple, effective technique for rapid measurement of fuels in low shrub communities. .Aust. For. Res.. 15; 79-84. Specht. R. L.. Rayson. P. and Jackman, M. E. (1958). — Dark Island Heath (Ninety-Mile Plain. South .Australia). VI. Pyric succession: change in composition, coverage, drv weight, and mineral nutrient status. Ausl. J. Bm.. 5; 59-88. van der Moezei, P. G., Loneragan. W. .A. and Bell. D. T. ( 1 987). — Northern Sandplain Kwongan; Regeneration following fire, juvenile period and flowering phenology. /. Rov. 5’tJc- West. .Aust. 69: 123-132. Wells, P. V. (1969). — The relation between mode of reproduction and extent of speciaiion in woody genera of the California chaparral. Evolution, 23: 246-267. 138 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Northern Sandplain Kwongan: effect of fire on Hakea obliqua and Beaufortia elegans population structure by David T. Bell, Paul G. van der Moezcl, Jeanette C. Delfs and William A. Loneragan Department of Botany, University of Western Australia. Nedlands. W.A. 6009. Received October 1986, Accepted March 1987. Abstract Individuals of obligatory reseeding species are killed outright by fires and species persistence in the Northern Sandplain shrubland ecosystem requires re-establishment by seed. Hakea obliqua and Beaufortia elegans share the same basic fire response syndrome — fire sensitivity of mature individuals, lack of seed dormancy and seed storage on the plant, but their adaptations relating to seedling establishment differ. Hakea obliqua has few but relatively large seeds and early seedling growth is rapid. Beaufortia elegans seedlings grow very slowly by comparison but continued existence in the shrubland ecosystems is ensured by the massive numbers of seed w'hich are dispersed following fire. The massive, synchronous production of small seed apparently satiates seed harvesting predators and sufficient numbers of remaining seeds find conditions of the post-fire habitat favourable for re-establishmcnt of the species. Season of the burn had considerable impact on the re-cstablishmenl of these obligatory reseeding species. Seedling regeneration was most effective following autumn burns and least effective following spring fires. The implications for management in these Northern Sandplain shrublands are that ecologically unfavourable seasons (in this case, winter and spring) should be excluded from prescribed burning regimes if the objective of management is to maintain all components of the pre- fire ecosystem. Introduction Fires in the Northern Sandplain shrublands play a major role in floristic structure of the vegetation (Bell and Loneragan 1985). Nearly one-third of the species in this region are killed outright by fire and rely on reseeding to maintain their position in these communities (Bell r'/ 1984, Bell 1985. van der Moezel et al. 1987). Fire also plays an integral role in the reproductive biology of many plants in this environment by inducing synchronous flowering and seed production events (Gill 1981), causing seed release and dispersal (Cremer 1965a) and stimulating the germination of soil- stored seed (Purdic 1 977). The local post-fire persistence of obligate reseeding species is dependent on the events of seed dispersal, seed germination and seedling establishment. Seed store in these species can include both soil-and plant-borne propagules (Vlahos and Bell 1986). Fire stimulates the release of seed from bradysporous species (Cremer 1965b, Cowling and Lamoni 1985) and the germination of soil-borne seed (Floyd 1976). Establishment is then dependent on the allocation of seed-stored nutrients and energy for early growth, acquisition of resources for subsequent growth, escape from insect and mammal predation, and survival in competition with other species of the habitat. Seed contents and metabolic rates can effect establishment success. Northern sandplain shrubs present a range of seed sizes (Pate and Dell 1984) and early growth rates can vary' enormously. Population densities following fire can also be strongly influenced by seed predators. .Ants have been reported to collect up to 80% of the seeds shed following fire in stands of Eucalyptus delegatensis (Grose 1960. Cramer 1 966). The massive, synchronized release of seed results in satiation of predators and subsequent seed escape (O'Dowd and Gill 1984). This paper highlights the influence of fire on population densities in two obligatory reseeding species of the Northern Sandplain shrublands, Hakea obliqua R.Br. and Beaufortia elegans Schau. The species Hakea obliqua (Proteaceae) is an erect shrub growing up to four metres tall in the deep sand shrublands of the region surrounding Badgingarra (Beard 1979). It has sharply pointed, terete leaves, 5-8 cm long and 3-4 mm thick. The flowers are white and are grouped in sessile axillary clusters along most of the length of the branches. The fruit is a woody structure measuring approximately 4x3x2 cm and covered by numerous corky outgrowths. 139 Journal of the Royal Society of Western Australia, Vol. 69. Part 4. 1987. Two hemispherical seeds about 1 cm long and 0.5 cm thick are contained within each fruit capsule. The seeds have membranous wings and corky outgrowths on the convex side which is embedded in the fruit. Flow'cring generally occurs in early spring and begins when the plants are four years old (van der Moezel el al. 1987). Fruits start to accumulate on the plant from this age. remaining closed until opened by the effects of fire. The seeds are not dispersed immediately after fruit dehiscense but are generally held for up to tw'o weeks by an attachment of the tip of the seed wing to the fruit. The phanerocotvlar seedlings (cotyledons exposed from the testa and borne at ground level upon germination) show moderate rates of early seedling growth (Delfs. unpublised data). By mid-October of the first growing season mean seedling weights average about Ig planr^ Bv two vears plant weight has more than trebled and until* at least 17 years both biomass and height increase linearly with age (Delfs et al. 1987). Plants of 17 years since last burn have mean biomass and height of 1 440 g and 330 cm, respectively. Beaufort ia clegans (Myrtaceae) is much srnaller at maturity in comparison, rarely exceeding 1 m in height (Delfs et al. 1987). The species has a dense crown of small (0.5 x 0.2 cm) leaves and pink clusters of flowers borne terminally. Flowering occurs in November and December (van der Moezel et al. 1987). Fruits are generally clustered together numbering 5-10 and measure approximately 0.5 x 0.5 cm. Seeds arc numerous but small (0.45 g) and are enclosed in the fruit until burnt. In contrast to Hakea ohliqua. the seedlings of Beaufortia elegans are phaneroepigeal (exposed and elevated above the soil surface). The cotyledons are green and foliar (as are Hakea obliqua). but the hypocotyl is not nearly as elongated in Beaufortia elegans. The juvenile period in Beaufortia elegans is short despite its obligatory reseeding habit. Flowering occurs in less than two years following establishment (van der Moezel et al. 1987). Early seedling growth is verv slow with seedlings of at least 15 weeks of age weighing less than 0.2 g or l/40th of the weight of comparable aged Hakea obliqua. By two years mean plant weights average approximately 0.35 g and reach approximately 75 g after 1 7 years (vs 1 440 g for 1 7 yr old Hakea obliqua) (Delfs et al. 1 987). Methods Population densities for Hakea obliqua and Beaufortia elegans were determined from a range of sites of known age since last burn in the vicinity of the junction of the Brand Highway and the Jurien Road (30M4’S, 115°16’E), approximately 20 km north of Badgingarra. Western Australia. Establishment of seedlings was determined in sites burned within the year prior to the winter seed germination period. Five adjoining 10x10 m stands burned the previous autumn and spring were sampled to establish the stand seed load of Hakea obliqua prior to burning. The number of seedlings established by .August and the number of seedlings still surviving after tw^o months growth were recorded. For Beaufortia elegans the number of seeds per plant was established from counts made on plants collected m five continuous one metre square quadrats. Seedling establishment density and two-month mortality figures were determined in two Im^quadrats. 14 Hakea obliqua • 12 10 8 • • ^ 6 1 t • • ieeds Plant o ro > • • • • fm • CO *1 d z 1 1 I • ' ' ■ ■ 0 5 10 15 1200 , Beaufortia elegans ^ 900 • • 600 • • 300 • 0 m 0 5 10 15 Age Since Last Burn (yrs.) Figure 1. — Plant borne seed store in Hakea obliqua and Beaufortia elegans in Northern Sandplain shrublands of known age since last burn. Results Seed store before fire Hakea obliqua and Beaufortia elegans retain seed in protective fruits until released following death of the supportive tissue. Estimates of the seed store build-up following fire from the site of known age since last burn indicated that the large seeded species. Hakea obliqua produced far fewer seeds per plant than the small seeded species Beaufortia elegans (Fig. 1 ). Only the occasional plant of < 5 years held fruit in these two obligatory reseeding species. By about II years following fire the accumulated number of seeds per plant in Hakea obliqua averaged approximately 6 while the number of seeds stored in Beaufortia elegans plants of comparable age was more than 800. Seedling recruitment following fire Estimates of seed density prior to the burns and seedling density following fire indicate that Hakea obliqua re-establishes a higher proportion of available seed compared with Beaufortia elegans (Table 1). Hakea obliqua might be expected to establish approximately one seedling from seven plant-stored seed while Beaufortia elegans might be expected to establish only one in twelve. Seed freed from fruits by hand have high rates of germination (98% in Hakea obliqua. 96% in 140 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. Table 1 Population recruitment in Hakea obliqua and Beaufortia elegans from recently burned sites in the Northern Sandplain shniblands. Species Burn site Age before fire Adult plants pre-fire ha Seeds pre fire ha Seedlings post-fire ha Percentage of seed establishment Percentage of pre-fire population Hakea obliqua Autumn >15 700 23 520 3 800 16 543 Hakea obliqua Spring 11 3 640 6 400 860 13 24 Beaufortia elegans Autumn >15 28 000 29 600 000 2 354 000 8 8 407 Beaufortia elegans) (J.C. Delfs, unpubL). Seed germination occurs within a week of imbibition and without artificial treatments (J.C. Delfs, unpubl.). Apparently large numbers of seed are predated between the time of the fire and the first significant rainfalls of late autumn and winter. The very much larger populations of Beaufortia elegans resulted from the much greater pre-fire density. In Hakea obliqua the recruitment of new seedlings was very much higher in the region burned in autumn only a few months prior to the winter rains. In the region burned the previous spring, however, the Hakea obliqua population was only one-fourth the density of the pre-fire condition. In the interval between initial counts and density measurements after two months, little mortality had occurred. In two separate sites within the autumn burn, mortality was less than 10% in the first two months (Table 2). Table 2 Seedling mortality of Beaufortia elegans in early months following establishment. Site no Adults pre-fire Seedlings Survival per m^ post-fire after 2 months 1 8 1 491 97% 2 10 863 90% Discussion Species comparisons Retention of seed on the plant until affected by fire is a strategy adopted by many species present in Australian plant communities (Gill 1981). Delayed dehiscence (bradyspory) is common in species from the families Proteaceac. Myriaceae and Casuarinaceae (Gardner 1957. Spcchi et ai 1958). .A fire is usually required for seed release but dehiscence may also occur when the woody fruit is dehydrated (Gill and Groves 1981). Species which retain seed until firing, such as Hakea obliqua and Beaufortia elegans, can exploit the open, well-lit, nutrient-rich, pyrogenic seed bed. Seeds accumulate on the plants during the inter fire period, then following post fire release and subsequent establishment ensure the survival of the species for the period until the next fire as both these species do not resprout following fire. The comparatively large seed of Hakea obliqua germinates quickly following the first rain and growth is relatively rapid during the first growing season following fire. The proportion of seedlings surviving the first summer drought might be expected to be high in this species compared to the populations of the small seeded, slow growing Beaufortia elegans. In California chaparral, moisture conditions in burned areas arc less favourable than in unburned control areas (Christensen and Muller 1975). Also, mineral nutrient changes which accompany the fire have little or no effect on post-fire germination responses but subsequent growth and survival on burned areas is thought to be enhanced by better nutrition and reduced grazing pressure. Summer conditions in the Northern Sandplains can be very dry. hot and windy; conditions most likely to be ver>' detrimental to first-year seedlings. Seedlings which develop deep roots rapidly might be expected to have an advantage in preventing summer season dessication. Beaufortia elegans appears to have opted for the production of very large numbers of seed of small size. This massive, synchronous reproductive event has possibly evolved in relation to seed predation. Ants predate or bury large numbers of seed (Briese and Macauley 1981). The large numbers of seed released into the habitat following fire results in satiation of predators and subsequent seed escape (O'Dowd and Gill 1984). Summer mortality might be expected to be comparatively greater due to the much smaller first- growing season size of Beaufortia elegans (Delfs et ai 1987). .Although only 8% of the seed store of an area resulted in seedling establishment, less than 1% of the resulting germules would ultimately need to survive to replace the parent population. Satiation of seed predators has been shown to be operative in the ultimate seed escape and germination of Eucalyptus delegatensis in the A.C.T. (O'Dowd and Gill 1984) and E. incrassata in north-western Victoria (Wellington and Noble 1985b). Successful post-fire population maintenance in Beaufortia elegans seems to rely on predator satiation and ultimately a safe site (sensu Harper 1977) where nutrients, light and moisture conditions are favourable to supplement the characteristically slow growth rate of this species. Both Hakea obliqua and Beaufortia elegans share the same basic fire response syndrome-fire sensitivity in the adult, lack of prominant seed dormancy and seed storage on the plant. This adaptive strategy is advantageous when the fire frequency occurs at intervals longer than the primary juvenile period but shorter than the life span of the plant. Beaufortia elegans is capable of fiowering in the second growing season while Hakea obliqua usually requires at least three full growing seasons before flowering in the fourth spring (van der Moczel et al. 1987). Plants in the site known to be at least 17 years since last burn were all vigorous indicating that the life span of both these species is probably considerably longer than 17 years. The natural frequency of burning in the Northern Sandplain shrublands is suspected to be of the order of 25 years (Bell 1985). Under the natural frequency of burning a significant buildup of plant- borne seed reserves would occur. Maintenance of population numbers would then depend on habitat conditions, especially in the first year following the fire. In addition to the size of the seed reserves before episodic fires, the population dynamics of these species will also be highly dependent on rates of seed loss following dispersal and before germination and early 141 Journal of the Royal Society of Western Australia, Vol. 69, Part 4, 1987. seedling mortality. Neither Beaufortia elegans or ffakea ohliqua seed appeared to germinate in the region of the spring burn before the following winter. The seeds, generally dispersed in the first month following a fire, were, therefore, susceptible to seed predation for approximately 8 months prior to the first winter rainfalls. A considerable fire season difference in the proportion of Hakea ohliqua seeds released which actually lead to established seedlings the following winter was observed. Seed predators are suspected of causing the reduced recruitment following the spring burn. Seed harvesting ants were the main cause of seed loss between dispersal and germination in Eucalyptus incrassala, a Victorian mallee with fire-induced seed fall (Wellington and Noble 1985a, 1985b). Seed predation was also the major cause of limited recruitment of South African bradysporous species (Bond 1984). Seediittg mortality Early mortality was minimal in Beaufortia elegans despite highly dense seedling stands. In Cape Province, South .Africa, post-emergence seedling predation tends to be minimal in burned areas in contrast to areas of more mature shrubland where seedling predation can be very heavy (Bond 1984). This generalization also appears to apply to the Northern Sandplain. Seedling loss in shrubland habitats can be considerable. Sever^ studies have documented high seedling losses in the first year or tw'o after fire (Horton and Knaebel 1955, Hanes 1971. Wellington and Nobel 1985a). The losses have most often been attributed to competition and drought stress (Schultz et al. 1955, Hanes 1977. Christensen and Muller 1975). Density independent mortality of seedlings of Ilakea ohliqua and Beaufortia elegans would tend to be high during the first drought season. Density dependent mortality might be expected later in the life cycle of these species. Senescence in these species due to factors related merely to age has not been observed in these species due to the lack of long unburned regions in the study area. Season of burn Season of burn appears to have a considerable impact on the continuing success of obligatory reseeding species in the Northern Sandplain. Seedling regeneration is most effective following autumn burns; least elTective following spring burns. The seasonal differences are due: 1) to the length of time for predation of dormant seeds before the winter germination period and 2) mortality of seedlings due to the competitive advantage provided to the rapidly-resprouting species by the longer interval between the spring burn and germination compared to the autumn burn to winter germination interval. Winter and spring burns are clearly unfavourable for the maintenance of obligate reseeders in the shrublands of this region. The implications for fire management of the Northern Sandplain and the problems that arise are fairly clear. If the objective of management is to maintain it// the species present, it is imperative that the prescribed burning seasons should be strictly defined to exclude ecologically unfavourable seasons. Prescribed burning events can’ihen be defined by the opportunity of suitable weather conditions in the ecologically favourable season. Aspects of the buildup of fire fuels arc currently being studied (eg. Schneider and Bell 1985), however, fire behaviour studies in the Northern Sandplain shrublands should be a priority for future research. Until further knowledge of the reproductive strategies of shrubland species is documented a conservative path of fire management should be followed (Hopper and Muir 1984). For the Northern Sandplain, late summer and autumn appear to be the ecologically most favourable season for prescribed burns. Fires during these seasons, however, have the potential to be uncontrollable and fire managers must be provided with guidelines for the safe limits under w'hich prescribed burns may be conducted during these seasons. Acknowledgemenis. Funds for Ihis research were contributed by the W.A. BushHres Board, the Honey Research Board of the Australian Department of Primaiy Industry and the Department of Botany. The position of Senior Lecturer in Plant Ecology of Dr Bell is supported by grants from Alcoa of Australia Ltd and Western Collieries Ltd. References Beard, J. S. (1979). — The Vegetation of the Moora and Hill River Areas, Western Australia. Vegmap Publications, Perth. Bell. D. T. (1985). — Aspects of response to fire in the Northern Sandplain heathlands. In; Fire Ecology and Management in Western Australian Ecosystems, Ford J. (cd.) pp. 33-40. West. Aust. Inst. Tech, Environ.' .Studies Grp. Rep. No 14.. Bentley, Western Australia. Bell. D. T.. Hopkins. A. J. M. and Pate, J. S. (1984). — Fire in the kwongan. In Kwongan-Flant Life of the Sandplain. Pate J. S. and Beard J. S. (eds) pp. 178-204. University of Western Australia Press. Nedlands. Western .Australia. Bell, D. T. and Loneragan, W. A. (1985). — The relationship of fire and soil type to flonstic patterns within heathland vegetation near Badgingarra. Western Australia. J Rov. Soc. Wes/. Aust., 67: 98- 109. Bond. W. .A. (1984). — Fire survival of Cape Proteaceae-influence of fire season and seed predators. Vegetatio. 56; 65-74. Briese. D. T. and Macaulay. B. J. (1981). — Food collection within an ant community in semi-arid Australia, with special reference to seed harvesters. .-Imj/. J. Ecoi, 6: 1-19. Christensen, N. L. and Muller. C. H. (1975).— Relative importance of factors controlling germination and seedling survival in AdenostomachtipzrmX. Am. .Midi Nat., 93: 71-78. Cramer. K, W. (1966). — Treatment of Eucalyptus regnans seed to reduce losses to insects after sowing. Aust. For.. 30: 162-174. Cremer, K. W. (1965a).— EITecis of fire on seed shed from Eucalyptus regnans. Aust. For., 29; 251-262. Cremer, K. W. (1965b). — How eucalypl fuits release their seed. Aust. J. Bot.. 13: 1M6. Cowling. R. M. and Lamoni. B. B. (1985).— Seed release m Banksia: the role of wet-dry cycles. Aust. J. Ecoi, 10: 1 69- 171. Delfs, J. C., Pale. J.S., and Bell, D. T. (1987). — Northern Sandplain kwongan: community biomass and selected species response to fire. J. Roy. S