i a ee, een anmets ihe Berohetes caren’ eae pean. vayotte: tae raat 2 ey a i Li) Ei 1 PROCEEDINGS of the LINNEAN SOCIETY NEW SOUTH WALES Marine Biological F Woods Hole Oceana eeoratory / C@anographie ioe Library PNI¢ institution ‘te JUL 08 1996 = The first section of this volume contains a collection of papers from students of Dr. PETER J. MYERSCOUGH which have been brought together to commemorate his retirement from the University of Sydney. The second section of this volume contains research reports submitted for publication in accordance with the publica- tion procedures of the Society, which can be obtained from the secretary. The Proceedings of the Linnean Society-of NSW are no longer published in parts and each issue represents a complete volume. VOLUME 115 A se s se & vrs 5) ae) is} << oS a fo) ie 3= = a. S i) = ”n a2 Sg is a. =) S w Ng /] near Darkes Forest, N.S.W. erscough surve / ter My Pe Proc. LINN. SOC. N.S.W., 115, 1995 SPECIAL SECTION OF THE PROCEEDINGS OF THE LINNEAN SOCIETY OF NEW SOUTH WALES ON PLANT ECOLOGY The Council of the Linnean Society of New South Wales decided to publish a special section ofits Proceedings on plant ecology and conservation biology to commemorate the retirement of Dr Peter J. Myerscough. Originally from Lancashire, Peter graduated from Oxford in 1963 and continued his research career in Scotland at the University of Edinburgh, publishing several papers on comparative plant biology and population dynamics. He moved to Australia in 1968 to take up a lectureship in Botany at the University of Sydney. Peter developed a major research interest in spatial pattern and environmental relationships of vegetation. His work on the Myall Lakes area in the mid north coast of New South Wales culminated in the publication of a major work jointly with Roger Carolin in 1986. He has also co-authored a diverse array of publications in the fields of plant population dynamics, estuarine ecology, freshwater wetland dynamics, germina- tion biology and fire ecology. One of Peter’s greatest contributions to the science of plant ecology has been through his teaching at undergraduate and postgraduate levels. This is perhaps best summed up by Peter himself, who said at the farewell dinner held in his honour, “I don’t think I’ve taught people very much in comparison to what they’ve taught me.” As well as Peter’s insuperable modesty, this illustrates his view of teaching as a two-way process between student and teacher. He offered exposure to a wide range of ecological ideas and studies without predisposition, but in a way that stimulated critical and original thinking in those who have been privileged to be his students. Peter continues to serve on the Council of the Linnean Society of New South Wales as he has since he was first elected in 1973, including a term as President in 1989. He has pur- sued his practical interest in the application of science to conservation through the Society’s Environment Committee and through his appointment to the Advisory Council for the National Parks and Wildlife Service of New South Wales between 1980 and 1988. Peter’s vision for the development of ecological knowledge and its application to conser- vation in Australia was outlined in his Presidential Address published in the Proceedings of 1990. This issue contains contributions from Peter’s students that address subjects in plant ecology. While the investigation of pattern and process is a theme common to all con- tributions, they span the full spectrum of experimental, comparative and descriptive approaches to ecology. The methods of investigation are equally diverse and include the study of experimental plant populations in the glasshouse, laboratory experiments, computer simulation, experimental manipulation of field populations, as well as the description and interpretation of vegetation patterns in nature. The breadth of topics and techniques covered by contributors reflects Peter’s own diverse interests and his far-reaching influence in Australian plant ecology. David Keith Council of the Linnean Society of New South Wales Proc. LINN. SOC. N.S.W., 115, 1995 Mero s ie agree Te _ nS ee Re he ee Patterns in Emergence of Acaciaand Grevillea Seedlings after Fire Tony D. AULD AND MARK TOZER (Communicated by D. KEITH) AULD, T.D.,and Tozer, M. Patternsin emergence of Acaciaand Grevillea seedlings after fire. Proc. Linn. Soc. N.S.W. 115: 5-15 (1995). Post-fire seedling emergence was found to be pulsed in Acacia suaveolens, a species which is known to have its seed dormancy broken by soil heating during the passage of a fire. This post-fire establishment declines to zero some 15 months after fire. Two species of Grevillea, G. buxifoa and G. speciosa, showed a similar, although slightly more varied response. Hence, both Acacia suaveolens and Grevillea spp. have germination cued to the passage ofa fire and are able to rapidly place seedlings into favourable sites for growth post- fire. However, seed dormancy in six Grevillea species was not significantly broken by heat- ing, although temperatures of 110-120°C were lethal to seeds. In these grevilleas, the heat produced by the fire is not the cue to break seed dormancy and it is expected that some other fire-related cue (charred wood, smoke) may be responsible for breaking seed dormancy. N.S.W. National Parks and Wildlife Service, P.O. 1967 Hurstville N.S.W. 2220; manuscript received 19 July 1994, accepted for publication 16 November 1994. KEYWORDS: germination, seed dormancy, seed heating, soil seedbanks INTRODUCTION Many species in fire-prone vegetation communities possess a soil seedbank. In the Sydney region of south-eastern Australia, some 89% of species in fire-prone vegetation have some form of soil seedbank. This may take the form of persistent (74%) or transient seedbanks (15%), the latter occurring where seeds have no dormancy and will germinate when sufficient moisture is available. The remaining species have either a canopy seed bank (10%) or no seedbank (1%) (Auld unpubl.). After a fire, conditions favourable for seedling growth and survival are enhanced via the release of nutrients in ash (Raison 1980), the increased availability of light and space and reduced competition levels. Consequently, the most favourable time for emergence of seedlings will be as soon as pos- sible after a fire. This situation will favour those species that are able to respond rapidly to the passage of a fire, i.e. those species whose germination is cued to fire. There is now widespread evidence that soil heating during the passage of a fire will break seed dorman- cy in many plant species with a soil seedbank. This has been demonstrated in several fire- prone communities in the world in several plant families, although particularly in legumes (Cushwa et al. 1968; Floyd, 1966,1976; Kruger, 1983; Keeley, 1987; Jeffery et al. 1988; Auld and O’Connell 1991; Bell et al. 1993). In the Sydney region, the first seedlings to emerge after a fire are legumes such as Acaciaspp., Dillwyniaspp. and Pultenaeaspp. and flushes of germination of such legumes are common after a fire (Auld, 1986). Other species which have soil seedbanks show no response to soil heating (Keeley, 1991; Auld, Keith & Bradstock unpubl.), although their seedlings may be numerous after fire. In such cases, the presence of post-fire seedlings may indicate the availability of non-dormant seed in the seedbank or the breaking of seed dormancy by other fire related cues such as charred wood (cf. Keeley et al. 1985; Keeley and Pizzorno 1986; Keeley, 1991), smoke (cf. de Lange and Boucher 1990; Brown, 1993), some other fire-related cue or the interaction of heat, charred wood and smoke. For the shrub species in sandstone communities around Sydney, there is generally a Proc. LINN. SOC. N.S.W., 115, 1995 6 T.D. AULD AND M. TOZER gradual decay of seeds from dormant to non dormant in the soil seedbank, independent of fire (Auld, 1986, Auld e¢ al. 1993). A similar pattern exists in other coastal plants, e.g. Acacia longifolia (Weiss, 1984), and some semi-arid and arid plants (Grice and Westoby 1987, Auld, 1995). Consequently, two simple alternative seedbank responses after fire are possible: 1) there is a pulse of germinants after fire as a direct response to dormancy breaking cues associated with the fire. This response would be expected in many legume species where the soil heating produced during a fire breaks seed dormancy. Germination levels would be expected to decline with time since fire. Such species would exploit the favourable post-fire environment for establishment; or 2) there is no pulse of germinants after a fire as the fire has no impact on seed dor- mancy. The level of post-fire emergence would reflect the rate of decay of seeds in the soil from a dormant to anon dormant state. Germination levels would be expected to be fairly constant through time, although there may be pulses related to seed fall ifa component of annual seed-crops are non dormant. Such species may still exploit the favourable post-fire environment for establishment if establishment is not successful at other times. This group of species would include those that show no response to soil heating or other fire cues or those species which can respond to soil heating but where soil heating during the fire has been insufficient to break seed dormancy (cf: Auld and O’Connell 1991). This study aimed to examine the emergence patterns of seedlings after the passage of fire. A legume species with a known heat response, Acacia suaveolens (Smith) Willd. (Auld 1986) was compared to Grevillea species which have a contrasting seed morphology and an unknown response to heat. METHODS Study area Sites were located in the Sydney Region (33°53’E, 151°13’S) of south-eastern Australia. Annual precipitation for Sydney is around 1300 mm, while the average monthly maximum/minimum temperatures are 26/18°C in summer and 16/8°C in winter. Effect of heat on seed dormancy in Grevillea The effect of heat on seed dormancy was examined in six Grevillea species. These were all shrubs comprising a component of the understorey (woodlands, forests) or dom- inants (heaths) of sclerophyll vegetation. Three species are common and widespread (G. buxifolia (Smith) R.Br., G. lineanifolia (Cav.) Druce and G. speciosa (Knight) McGillivray) while the remaining three species are rare plants (G. caleyi R.Br. 2ECi, G. longifolia R.Br. 2RC- and G. shivessii Blakely 2VCit, codes follow Briggs and Leigh 1991, ANZECC 1993). All these Grevillea spp. are killed by fire and rely on germination from a soil seedbank for establishment post-fire. Two species, G. caleyi and G. longifolia, have toothbrush flowers and seeds lacking an aril (Auld e¢ al. 1993), while the remainder all have ‘spider-flowers’ and seeds with an aril. Field collections of seeds for each study species were made during November, December and January of 1989, 1990 and 1991. For G. caleyi, G. longifolia and G. shiressii large developing fruits were bagged and seeds were collected after the fruits had dehisced. For the other species, most fruits were collected by hand when ripe, although some bagging was done for G. buxifolia. In the laboratory, intact seeds were stored in envelopes at room temperature. The effect of a range of temperatures at one duration of heating on seed germination were examined in the laboratory. Ten temperature levels were examined; ambient (control), 40, 50, 60, 70, 80, 90, 100, 110 and 120°C, for a single duration of exposure (10 mins). Thirty seeds were tested for each temperature treatment, Proc. LINN. SOC. N.S.W., 115, 1995 ACACIA AND GREVILLEA SEEDLINGS AFTER FIRE 7 although occasionally sufficient seeds were available to use additional seeds as replicates in the control. A small volume of air-dried soil from the field was enclosed in aluminium foil and preheated in an oven to the desired temperature. Once preheated, the soil was removed from the oven and seeds added. A thermometer placed in the soil was used to monitor the soil temperature. After exposure in the oven for the required time, seeds were extracted from the soil using a sieve (mesh 2 x 2 mm) and allowed to cool. Individual seeds were then placed on Whatman grade 2 filter paper moistened with distilled water in 9 cm diameter petri dishes. Germination was followed for eight weeks, after which all seeds which had not germinated were scarified and allowed to germinate in order to esti- mate the viability of the seed lot used. Any seeds which still did not germinate were tested for viability using the tetrazolium test (Lakon, 1949). Comparisons between temperature treatments were made via a two factorial model in GLIM (Aitkin et al. 1989) with a binomi- al error structure. Seedling emergence in the field Post-fire emergence of seedlings was recorded at four study sites in Garigal National Park in the northern suburbs of Sydney. Three of the sites were burnt in November 1992, while the fourth was burnt 2 months earlier. Vegetation at the sites was a mixture of heath and open woodland with a shrub understorey on Hawkesbury sandstone. At each site a quadrat was marked out and all emerging Acacia and Grevillea seedlings were tagged at 2-3 monthly intervals for up to 15 months post-fire. Quadrat size varied between sites and was governed by the abundance of emergents. Sampling involved repeated counts at the same quadrats over time so that each emergent seedling could be assigned to a particular time interval post-fire. It was assumed that correlations between counts from the same quadrats across time intervals were minimal. Emergent seedlings were Acacia suaveolens at all sites, Grevillea speciosa (3 sites) and G. buxifolia (2 sites). Patterns of emergence were examined for Acacia and pooled Grevillea spp. across time intervals using the number of seedlings emerged in a particular time interval as a fraction of the total number of seedlings emerged over all time periods. Data were analysed by a three factorial model GLIM using a binomial error structure. Appropriate post-hoc tests were made using a Z statistic (Zar, 1974). Arange of curves were fitted to proportional emergence across time to estimate the best fit to the data. RESULTS There was no clear evidence that heating broke seed dormancy in any of the six Grevillea species examined (Table 1, Fig. 1). There was considerable variation in the level of non-dormant seeds between species. In all species, temperatures of 120° C were lethal to seeds, while some seed death also occurred at 110° C. Significant temperature effects in species were related to this seed death at high temperatures. The one significant Site times Temperature interaction (G. linearifolia, Table 1) mainly reflected different seed mortali- ty levels at 100 and 110° C (Fig. 1c). For some species, there was a high level of seed dor- mancy at all treatments (G. caley, Fig. 1b), while for others, high levels of non dormancy were apparent in all treatments below the lethal temperatures (G. linearifolia, Fig. 1c, G. speciosa, Fig. 1f). Seedling emergence patterns in Acacia suaveolens and pooled Grevillea buxifolia and G. speciosa were similar, with an initial post-fire pulse declining to very small to no emer- gence some 15 months after the fire (Fig. 2). While one site had a significant increase in seedling emergence in Grevilleaat the second time interval compared to the first (Fig. 2b), the same overall pattern of a decline through time was apparent at all sites. This is reflect- ed in a significant three-way interaction in the GLIM analysis (Table 2). An exponential model was the best fit to the data in terms of decline in emergence through time (Fig. 2). Proc. LINN. SOC. N.S.W., 115, 1995 8 T.D. AULD AND M. TOZER These regressions accounted for 96% and 71% of the variation in emergence through time for Acacia suaveolens and pooled Grevillea species, respectively. Proportion of Seeds with Dormancy Broken O 50 100 150 Temperature CC) Fig. 1. Effect of heat on seed dormancy in Grevillea spp. Different lines represent different yearly samples and/or sites. Bars represent standard errors. The position of points on the x-axis have been adjusted slightly for clarity. a) G. buxifolia;b) G. caleyt,c) G. linearifolia, d) G. longifolia, e) G. shiressti; f) G. speciosa Proc. LINN. SOC. N.S.W., 115, 1995 ACACIA AND GREVILLEA SEEDLINGS AFTER FIRE 9 d) 0.8 Proportion of Seeds with Dormancy Broken Temperature CC) Fig. 1, cont'd. Effect of heat on seed dormancy in Grevillea spp. Different lines represent different yearly samples and/or sites. Bars represent standard errors. The position of points on the x-axis have been adjusted slightly for clarity. a) G. buxifolia; b) G. caleyi; c) G. linearifolia; d) G. longifolia; e) G. shiressii; f) G. speciosa Proc. LINN. SOC. N.S.W., 115, 1995 10 T.D. AULD AND M. TOZER (es D ONS g an > O & oO = O @) GSO 50 100 150 = (ep) @ TO o) J) i f) = O08 a Ne = -- x O PI &) = 06 1 eve amin O f \ o \ aL OA it \ \ OO \ \ 0.0 ow O 5O 100 150 Temperature (CC) Fig. 1, cont'd. Effect of heat on seed dormancy in Grevillea spp. Different lines represent different yearly samples and/or sites. Bars represent standard errors. The position of points on the x-axis have been adjusted slightly for clarity. a) G. buxifolia,b) G. caleyi,c) G. linearifolia; d) G. longifolia; e) G. shiressit;£) G. speciosa Proc. LINN. SOC. N.S.W., 115, 1995 Comparison of the effect of heat on seed dormancy in six Grevillea species using GLIM. Species G. buxifolia G. caleyi G. linearifolia G. longifolia G. shiressii G. speciosa 1, ns, non significant = 0.05>P>0.01 ** — 0.01>P>0.001 *eE P<0.001 GLIM comparisons of the proportion of seedlings emerging at various time intervals post-fire for Acacia suaveolens and Grevillea spp. Factor Species Site Time Species x Site Species x Time Site x Time Species x Site x Time 1, ns, non significant *** P<0.001, ACACIA AND GREVILLEA SEEDLINGS AFTER FIRE Factor Site Temperature Site x Temperature Site Temperature Site x Temperature Site Temperature Site x Temperature Site Temperature Site x Temperature Site Temperature Site x Temperature Site Temperature Site x Temperature Change df fo po DOO BD GO TABLE 1 Change df 1 9 9 oor oor oor TABLE 2 Change dev 0.04 0.05 3137 0.46 76.5 430.8 125.8 Change dev 0.01 26.48 9.16 0.42 11.26 10.92 0.99 304.9 83.3 42.66 44.64 8.17 10.13 19.47 3.13 9.54 182.9 10.9 Pp! 11 Proc. LINN. SOC. N.S.W., 115, 1995 12 T.D. AULD AND M. TOZER Proportion of Seedlings Emerged 0.0 0.5 1.0 1.5 Time (years) Fig. 2. Post-fire emergence of seedlings. Lines represent line of best fit (see text). Different symbols represent different sites. a) Acacia suaveolens; y = 8.117.e-"**™* b) Grevillea buxifolia and G. speciosa combined. y = 1.402.e-°”” Proc. LINN. SOC. N.S.W., 115, 1995 2.0 ACACIA AND GREVILLEA SEEDLINGS AFTER FIRE 13 DISCUSSION The Grevillea species examined contrast with Acacia spp. (see Auld and O’Connell 1991) in that they show no breaking of seed dormancy by heating. It is unlikely that longer durations of heating may have affected seed dormancy (Auld & O’Connell 1991) and such durations are not typical of soil heating in the field during the passage of a fire (Bradstock and Auld 1995). Consequently, while it would be expected that Acacia spp. would show a pulse of germinants post-fire, the expected response from Grevillea spp. is unclear. No pulse would be expected where there is no seed response to the passage of a fire. In such a situation seedling emergence is dependent upon the rate of decay of seeds from a dormant to anon dormant state in the soil seedbank. Alternatively, some cue asso- ciated with the passage ofa fire apart from heat may be responsible for breaking seed dor- mancy. Such a cue could be leachates from charred wood (cf Keeley et al. 1985) or chemicals derived from smoke (cf Brown, 1993). Should such a fire-related cue exist in Grevillea, then a pulsed response would be expected. The examination of post-fire emergence in the field confirmed the predicted pulse for Acacia suaveolens (Fig. 2a). The exponential regression (Fig. 2a) was able to explain some 95.6% of the variation in seedling emergence over time and is a good predictor of the likely emergence pattern in this species. Depending on the timing and magnitude of post-fire rainfall, the positioning of this curve may vary, however its shape should remain roughly the same. For Grevillea buxifolia and G. speciosa there was also a strongly pulsed pattern of post- fire emergence (Fig. 2b). In this case, the exponential regression accounted for some 71% of the variation in emergence over time. For Grevillea species there was more varia- tion between sites in the initial magnitude and timing of the pulse, however, the overall pattern was strikingly similar to A. swaveolens. In the case of Grevillea spp., the breaking of seed dormancy by heating of the soil during the passage of the fire cannot be invoked to explain the observed response. Clearly, there is a distinct post-fire pulse of germinants but the fire-related factor responsible for breaking seed dormancy remains unknown. It is possible that Grevillea spp. respond to fire cues such as smoke or leachates from charred wood and this remains an interesting area for future research. Auld and O’Connell (1991) predicted that the amount of the soil seedbank that is stimulated to germinate by the passage of a fire in species with dormancy broken by heat will vary depending on the amount of soil heating that occurs during a fire. This in turn is controlled by the amount of fine ground fuels that are burnt during a fire, a component of fire intensity (Bradstock and Auld 1995). In general, fires with high levels of fine fuel com- bustion will cause the greatest depletion of the soil seedbank through high post-fire ger- mination levels. Following from this, patterns of the depth of emergence of seedlings should be closely tuned to small-scale litter and fine fuel patterns. Predicted levels of seeds emerging from the soil seedbank in A. suaveolens range from 1.1 to 32.9%, depending on the amount of soil heating that occurs (Auld and O’Connell 1991). Additional seeds near the soil surface will be killed by lethal temperatures. Clearly under this scenario, seeds will be depleted from certain depths and remain available at greater depths, depending on the level of soil heating, i.e. there is a soil depth-related impact for species whose seed dor- mancy is broken by heat. It remains to be seen whether an equivalent soil depth-related response occurs in species whose dormancy is broken by a fire-related cue other than heat, although the seedbank will be depleted at the soil surface as seeds are killed by soil heating. The relative speed with which seedlings of particular species emerge in the post-fire environment will reflect their strategy of seed storage between fires and whether a persis- tent seedbank is maintained or not. Species that store seeds in the soil and have persistent seedbanks are likely to be the quickest to place seedlings in the favourable post-fire envi- ronment if they have their dormancy broken by the passage ofa fire. Species with a canopy seedbank must wait for seeds to be released from the woody cones, and this may take many Proc. LINN. SOC. N.S.W., 115, 1995 14 T.D. AULD AND M. TOZER weeks after the fire has passed. Species that rely on dispersal from outside the burnt area must wait for the season of fruit maturation in unburnt communities and this will not nec- essarily coincide with the favourable immediate post-fire period. Finally, there are those species that have transient soil seedbanks and that rely on a post-fire pulse of flowering to place seedlings in the post-fire environment. These species may take from a few weeks up to 2 years to flower after fire and therefore are the slowest to place seeds in the post-fire environment. For species with a soil seedbank, the pattern of post-fire emergence of seedlings should reflect moisture conditions, the influence of fire (heat, charred wood, smoke etc.) on breaking seed dormancy, and the breaking of seed dormancy by means other than the passage of the fire. While A. suaveolens and Grevillea spp. have different mechanisms for having seed dormancy broken in relation to a fire, this makes no difference to the timing of the post-fire pulse of emergent seedlings. This may be because both dormancy break- ing mechanisms operate at the time the fire passes. Alternatively, as seeds must wait until there is sufficient moisture in order to germinate, any differences between the timing of dormancy breaking by different fire-cues may be eliminated. ACKNOWLEDGEMENTS Thanks to Rachael Thomas, Maria Matthes, Karyn Maling and Andrew Denham who carried out the laboratory heating experiments and to Dave Brown, Andrew Marshall and numerous assistants who organised and carried out the experimental burns in the field. Ross Bradstock made helpful comments on the manuscript. References AITKIN, M., ANDERSON, D., FRANCIS, B. and HINDE, J., 1989. — Statistical Modelling in GLIM. Oxford University Press, New York. ANZECC, 1993. — Threatened Australian Flora. Australian Nature Conservation Agency. AULD, T. D., 1986. — Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: fire and the transition to seedlings. Australian Journal of Ecology 11: 373-385. AuLbD, T.D., 1995. — Soil seedbank patterns in four trees and shrubs from arid Australia. Journal of And Environments 29: 33-45. AULD, T.D., BRADSTOCK, R.A. and KEITH, D., 1993. — Fire as a threat to populations of rare plants. Australian National Parks & Wildlife Service Endangered Species Project No. 31. AULD, T.D. and O’CONNELL, M. A., 1991. — Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Australian Journal of Ecology 16: 53-70. BELL, D. T., PLUMMER, J. A. and TayLor, S. K., 1993. — Seed Germination Ecology in Southwestern Western Australia. The Botanical Review. 59: 24-73. BRADSTOCK, R.A. and AULD, T.D., 1995. — Soil temperatures during bushfires in relation to fire intensity: conse- quences for legume germination in south-eastern Australia. Journal of Applied Ecology 32: 76-84. Briccs, J. and LEIGH, J., 1991. — Rare or threatened Australian Plants. ANPWS special publication No 14. Brown, N. A. C., 1993. — Promotion of germination of Fynbos seeds by plant-derived smoke. New Phytologist. 123: 575-583. CusHwa, C. T., MARTIN, R. E. and MILLER, R. L., 1968. — The effects of fire on seed germination. Journal of Range Management 21: 250-254. DE LANGE, J.H. and Boucuer, C., 1990. — Autecological studies on Audouinia capitata (Bruniaceae). I. Plant- derived smoke as a seed germination cue. South African Journal of Botany 56: 700-703. FLoyp, A. G., 1966. — Effect of fire upon weed seeds in the wet sclerophyll forests of northern New South Wales. Australian Journal of Botany 14: 243-256. FLoyp, A. G., 1976. — Effect of burning on regeneration from seeds in wet sclerophyll forest. Australian Forestry 39: 210-220. Grice, A. C. and Wesropy, M., 1987. — Aspects of the dynamics of the seed-banks and seedling populations of Acacia victoriaeand Cassia spp. in arid western New South Wales. Australian Jowrnal of Ecology 12: 209-215. JEFFERY, D. J., HOLMEs, P. M. and REBELo, A. G., 1988. — Effects of dry heat on seed germination in selected indigenous and alien legume species in South Africa. South African Journal of Botany 54: 28-34. KEELEY, J. E., 1987. — Role of fire in seed germination of woody taxa in California chaparral. Ecology 68: 434-443. KEELEY, J.E., 1991. — Seed germination and life history syndromes in the Californian chaparral. The Botanical Review 57: 81-116. KEELEY, J. E., MORTON, B. A., PEpROsA, A. and TROTTER, P., 1985. — Role of allelopathy, heat and charred wood in the germination of chaparral herbs and suffrutescents. Journal of Ecology 73: 445-458. Proc. LINN. SOC. N.S.W., 115, 1995 ACACIA AND GREVILLEA SEEDLINGS AFTER FIRE 15 KEELEY, S. C. and PIzzorno, M., 1986. — Charred wood stimulated germination of two fire-following herbs of the California chaparral and the role of hemicellulose. American Journal of Botany 73: 1289-1297. KRUGER, K. J., 1983. — Plant community diversity and dynamics in relation to fire. In KRUGER, F.J., MITCHELL, D.T. and JARVIS, J.U.M. (eds.) Mediterranean — Type Ecosystems. The role of nutrients. pp. 446-472, Springer - Verlag, Heidelberg. LAKON, G., 1949. — The topographical tetrazolium method for determining the germinating capacity of seeds. Plant Physiology 24: 389-394. RAISON, R,J., 1980.— A review of the role of fire in nutrient cycling in Australian native forests, and of methodolo- gy for studying the fire-nutrient interaction. Australian Journal of Ecology 5: 15-21. WEIsS, P. W., 1984. — Seed characteristics and regeneration of some species in invaded coastal communities. Austrahan Journal of Ecology 9: 99-106. ZAR, J-H., 1974. — Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, N.J. Proc. LINN. SOC. N.S.W., 115, 1995 a plese Boas abe) j of Ss. . D a! (uate ‘ Ai en Mi : ae peers ak Ge. £ eal iis) hag itty ‘eth. poh ‘tier aed ff" Gee iL gee ss pe or —e conta my: in ou we: et cago ap ot vary Maia. eas ae chide Pe, Seb A 2 —y) ma eh, ota alee oh i ess ‘ e/a Oke 7 eat ve stil toy at ‘(uum clean Caton) ar, | ees 4 iT oot ® ee ae ee Lk arto fag gaa oe Os pains smote Ot lp 2 deh fl: » >’ AN aaa ae > ir acd oe aad ae Simple Models of Pattern and Process MICHAEL BEDWARD (Communicated by D. KEITH) BEDWARD, M. Simple models of pattern and process. Proc. Linn. Soc. N.S.W., 115: 17-23 (1995). Pattern and process dynamics in plant communities can be modelled using cellular automata. These are simulation models that are easy to build and modify, and that are espe- cially useful in visualizing the processes that are being modelled. To illustrate this approach, I describe a cellular automaton that simulates pattern and process in a simple bryophyte community growing on a steep rock face. NSW National Parks and Wildlife Service, P.O. Box 1967, Hurstville N.S.W. 2220; manuscript received 26 July 1994, accepted for publication 14 December 1994. KEY WORDS: Cellular automata, community population, lichen, moss. Come-and-go pervades everything of which we have knowledge, and though great things go more slowly, they too are built up of small ones and must fare as that which makes them. — Samuel Butler INTRODUCTION Pattern and process behaviour, a continual sequence of disturbance or senescence and regrowth (Watt, 1947), occurs in many types of vegetation but is particularly easy to observe in bryophyte communities that grow on steep surfaces. Ashton (1986) gives some interesting examples of pattern and process in bryophyte communities on rock faces and tree trunks. Heavy moss mats can slump from a surface creating new space that can then be colonized by algae, liverworts, lichens and eventually mosses. Moss mats may shade out and kill crustose and foliose lichens by overgrowing them. Alternatively, mosses can be parasitized and killed by lichens (McWhorter, 1921: Ashton, 1986) and the germination of moss spores can be inhibited by chemical compounds produced by lichens (Lawrey 1977). Such dynamics can produce a mosaic of different species and unoccupied space that is always shifting over time. Both the distribution and the area occupied by each species in the community can vary substantially. One way of viewing this shifting mosaic is long term observation. Another is to attempt to model the processes involved. Modelling can also help to test the logic of our current understanding of these processes and help us to develop new hypotheses. However, there are problems in trying to construct mathematical models of these types of communities. Many of the interactions between species are very localized and spatial con- figuration is important. For example, the pattern of growth of a moss mat depends on the arrangement of suitable space while the pattern of moss slumping from a steep surface is influenced by how connected the moss mats are. It is usually difficult to incorporate these sort of factors into mathematical treatments (Hastings, 1991; DeAngelis and Rose, 1992). An alternative to mathematical modelling is to use computer simulations. One very simple family of simulation models, cellular automata, have been used for a variety of eco- logical studies including fire patterns in vegetation (Green, 1983), the fire ecology of plant species (Green, 1985; Bradstock et al., in press), host-parasite dynamics in insects (Hassell et al., 1991) and evolutionary theory (e.g. Nowak and May, 1992). A cellular automaton is a collection of cells that interact in simple ways but where the whole collec- Proc. LINN. SOC. N.S.W., 115, 1995 18 SIMPLE MODELS OF PATTERN AND PROCESS tion can display very complex overall behaviour (Wolfram, 1984; Phipps 1992). Simulations using cellular automata proceed in discrete time steps. Each cell assumes one of a finite number of states, and its state at the next time step depends upon its current state, and the states of its neighbours. The neighbourhood of a cell can take many forms, e.g. the eight immediately adjacent cells or all cells within a specified radius, and in some models the size and shape of the neighbourhood can vary over time. The rules used to decide the state of each cell at each time step can be purely deterministic or include some element of chance (Phipps, 1992). In this paper, I describe a cellular automaton that simulates the dynamics of a bryophyte community growing on a steep rock face. Although simple, the model displays realistic pattern and process behaviour and could easily be extended or refined to study specific communities. I have implemented the model as a computer program, MOBI (Model of Bryophyte Interactions), that displays the shifting mosaic of species and vacant space on a computer screen. DESCRIPTION OF THE MODEL For simplicity, the model community presented here consists of only two species: a moss and a fruticose lichen that grows upon the moss and parasitizes it. Later, I will discuss some possibilities for including further species as well as environmental variables. The habitat, a steep rock face, is represented by a square grid where each cell can be either vacant, occupied by moss, or occupied by moss and lichen. At each time step during a sim- ulation, cells can change state according to rules that describe the growth and dispersal of the moss and the lichen, and the slumping of moss mats from the rock face. The size of the grid and the number of time steps in a simulation can vary. The temporal and spatial scale is flexible, but here I am assuming that each cell represents an area of about | cm* and each time step is about | year. Moss growth and dispersal At the beginning of a simulation, moss is assigned to a given number of randomly selected cells. The moss can then spread into adjacent cells to simulate the growth of moss clumps. The growth mechanism includes a stochastic element such that the probability of moss spreading into a vacant cell is: P moss(t+1) =MAX[1.0,C.f] where p moss(t+1) 18 the probability of the vacant cell being occupied at the next time step; Cis a constant; and f, is the fraction of neighbours of the cell that already have moss. The neighbourhood consists of the eight immediately adjacent cells. Figure 1 shows the pattern of growth that is produced using C = 2. Moss sporeling establishment was modelled by assuming that a rain of moss spores falls equally on all parts of the rock face. At each time step, moss sporelings could establish in suitable vacant cells with a specified probability that was constant throughout the simu- lation. Conditions on the suitability of vacant cells for sporeling establishment are explained further below. Moss slumping MOBI simulates the slumping of heavy moss clumps from the rock wall. To do this, each moss cell is assigned a weight. When the cell is first occupied by moss, or reoccupied after being vacant, a weight of | is assigned. The weight is then incremented by | at each time step until a preset maximum value is reached. The slumping of moss clumps is simu- lated at each time step by testing each moss cell to see if it will initiate a slump. The proba- Proc. LINN. SOC. N.S.W., 115, 1995 M. BEDWARD 19 bility of this for each moss cell is equal to the weight of the moss multiplied by a constant. Cells whose weights are below a specified minimum cannot start a slump. When a falling cell is found it becomes the focus for a slump area that spreads to contiguous moss cells using the same stochastic mechanism described above for moss growth. The growth of the slump area continues until no more moss cells can be reached, or until a specified maxi- mum area is attained, whichever happens first. Then all of the moss cells in the slump area Fig. 1. Simulated moss growth from ten randomly placed initial cells in a 100x100 grid after 15 time steps. are converted to vacant cells. A number of clumps of moss may slump from the rock face in a single time step. Newly vacant cells In some bryophyte communities, newly created space must be colonized by other species, such as algae, before moss spores can germinate and establish (e.g. Ashton, 1986). To simulate this, the number of time steps that must elapse before newly vacant cells are suitable for moss sporelings can be specified as a constant. This does not affect growth into these cells from adjacent moss clumps. Lichen dispersal and growth The lichen can only occupy cells that contain moss. Clumps of lichen can grow in the same stochastic manner described for moss growth. The initial lichen population is created by randomly assigning lichen to a specified number of moss cells either at the start of the simulation or at some later time. The dispersal of propagules, and establishment in moss cells, is modelled using a constant probability as for moss dispersal. Interaction of the lichen and moss The effect of the lichen on the moss is simply to prevent growth from infected moss cells. If all cells in a moss clump become infected the clump ceases to expand. Once in- fected, a moss cell can not free itself of lichen. The lichen does not affect the pattern of moss slumping. This is a very conservative and simplified interaction. Proc. LINN. SOC. N.S.W., 115, 1995 20 SIMPLE MODELS OF PATTERN AND PROCESS AN EXAMPLE APPLICATION I used MOBI to see how the rate and pattern of moss mats slumping from the rock face would affect the success of the lichen. I varied moss slumping by setting four different values for the maximum area of an individual slump. I also varied moss and lichen disper- sal to see what effect this would have on the behaviour of the model. Table 1 shows the design used while the complete list of MOBI variable values is given in the appendix. TABLE | Number of simulations for each combination of moss slump and dispersal variables Maximum moss slump area (cells) 100 250 500 1000 Dispersal none 20 20 20 20 moss only 20 20 20 20 moss and lichen 20 20 20 20 Figure 2 shows the average moss and lichen populations for each set of 20 replicate simulations. As the maximum size of a slumping moss mat increased from 100 cells (fig. 2a—c) to 500 cells (fig. 2g —-i) the proportion of moss infected by lichen decreased as indicated by the gap between the lines in each graph. The presence of moss and lichen dispersal made little difference to the results. Where the moss fell in very large clumps of up to 1000 cells (fig. 2j-1) dispersal had a greater effect. With neither moss or lichen dis- persal (fig. 2]), both species soon became extinct. Where the moss and lichen could both disperse (fig. 21), both persisted throughout the simulation period although only a small proportion of the moss was infected by lichen. 100 | 100 a | 100+mo b jae c 100 d = 250+mo | 250+mo+li 100 : 500 g ea 500+mo#li 2 Nee 100 % 1000 j 1000+mo k 1000+ 1 Ne Pe iL — ~ % Area occupied mo-+li SN nt ON 0 100 200 O 100. 200 O 100 200 Time Fig. 2. Average populations of moss and lichen for each set of 20 replicate simulations (see text). In each graph, the upper line is the moss population and the lower line is the lichen population. Codes within each graph indicate the simulation settings: e.g. 100 is max slump area of 100 cells with no dispersal, 100+mo is same slump area with moss dispersal, 100+mo+li is same slump area with moss and lichen dispersal. Proc. LINN. SOC. N.S.W., 115, 1995 M. BEDWARD 21 The mechanisms behind these results are easy to visualize by watching the moss and lichen mosaic as it unfolds in each simulation. Figure 3 shows the community in a single simulation for each level of moss slumping with no dispersal. In each simulation, 50 time steps have elapsed. Low levels of moss slumping led to small fragments of uninfected moss in a sea of lichen. When watching one of these simulations, the uninfected moss seemed to be constantly chased around the rock face by the lichen. Increasing the level of slump- ing led to the moss mats becoming more discrete which made it harder for the lichen to spread vegetatively. This also increased the chance of large patches of lichen being removed from the rock face. At the highest level of stumping the moss itself risked extinc- tion when there was no dispersal. € d Fig. 3. The state of the model in one example simulation for each of four moss slump areas: a. 100 cells; b. 250 cells; c. 500 cells; d. 1000 cells. Grey denotes uninfected moss; black denotes moss infected by lichen. DISCUSSION The results show that MOBI is capable of displaying complex and realistic pattern and process behaviour. The model could be refined and extended in many ways for specific applications to bryophytic communities. For example, the probabilities for growth, dispersal and moss slumping could be linked to a historical set of rainfall data. The interaction between the moss and the lichen could be tailored to data for particular species. A simulation could also include vascular plant species that establish in moss mats and accelerate slumping of the moss when they grow large. With these sort of elabora- tions, MOBI could be used as a tool for population viability analysis. Alternatively, it is possible to begin with a model such as MOBI that considers Proc. LINN. SOC. N.S.W., 115, 1995 22 SIMPLE MODELS OF PATTERN AND PROCESS biological processes explicitly, and then simplify the model to a more abstract and general form. Hassell et al. (1991) took this approach with their work on insect host-parasite popu- lations. By obtaining very similar results from models that were based on detailed mathe- matical formulations of species interactions, and alternative models that were purely qualitative cellular automata, they showed that it was the general pattern of species disper- sal rather than the fine detail of the species interactions that determined the behaviour of the system. This sort of approach seeks to identify common patterns that underlie the behaviour of many different kinds of biological systems (e.g. Green, 1993). The results of MOBI simulations could suggest useful hypotheses for other types of plant communities that display pattern and process behaviour. It is easy to design and build models that are based on cellular automata, such as MOBI, to explore pattern and process behaviour. Perhaps the most useful feature of these models is that you can watch the progress of each simulation and notice patterns that would not have been obvious from a mathematical treatment. This feature also makes cellular automata very useful teaching tools. ACKNOWLEDGEMENTS Thanks to Murray Ellis for help with the figures. References ASHTON, D.H. 1986. — Ecology of bryophytic communities in mature Eucalyptus regnans F. Muell. forest at Wallaby Creek, Victoria. Aust. J. Bot., 34: 107-29. BRADSTOCK, R.A. BEDWARD, M. and Scott, J.A. — Jn Press. Persistence of plant populations in landscapes subjected to recurrent fire: simulation of effects of varied frequency and size. In Proceedings of 2nd international conference on forest and fire research. Portugal. DEANGELIS, D.L. and Rosg, K-A. 1992. — Which individual based approach is most appropriate for a given problem? Jn D.L. DEANGELIS and L.J. GRoss (eds.) Individual based models and approaches in ecology: populations, communities and ecosystems, pp. 67-87. Routledge, Chapman & Hall. New York. CONNELL, J.H. 1978. — Diversity in tropical rain forests and coral reefs. Scence, 199: 1302-10. GREEN, D.G. 1983. — Shapes of simulated fires in discrete fuels. Ecol. Model., 20: 21-32. GREEN, D.G. 1985. — Simulated effects of fire, dispersal and spatial pattern on competition with forest mosaics. Vegetatio, 82: 139-153. GREEN, D.G. 1993. — Emergent behaviour in biological systems. In D.G. GREEN and T.J. BOSSOMAIER (eds.) Complex systems — from biology to computation, pp. 24-35. IOS Press. Amsterdam. HASSELL, M.P., Comins, H.N. and May, R.M. 1991. — Spatial structure and chaos in insect population dynamics. Nature, 353: 255-8. Hastincs, A. 1991. — Structured models of metapopulation dynamics. Biol. J. Linn. Soc., 42: 57-71. Lawrey, J.D. 1977. — Inhibition of moss spore germination by acetone extracts of terricolous Cladonia species. Bull. Torrey. Bot. Club. 104: 49-52. McWHOoRTER, F.P. 1921. — Destruction of mosses by lichens. Bot. Gazzette, 72: 321-5. Nowak, M.A. and May, R.M. 1992. — Evolutionary games and spatial chaos. Nature, 359: 826-9. Puipps, M.J. 1992. — From local to global: the lesson of cellular automata. Jn D.L. DEANGELIS and L.J. Gross (eds.) Individual based models and approaches in ecology: populations, communities and ecosystems. pp. 165- 87. Routledge, Chapman & Hall. New York. Watt, A.S. 1947. — Pattern and process in the plant community. J. Ecol. 35: 1-22. WOLFRAM, S. 1984. — Cellular automata as models of complexity. Nature, 311: 419-24. Proc. LINN. SOC. N.S.W., 115, 1995 M. BEDWARD APPENDIX 23 Variables that can be setin MOBI and values used for the simulations discussed in the text. Number of rows Number of cols Simulation period Initial moss population Initial lichen population Time to add initial lichen cells Lichen growth rate (relative to moss) Probability of moss sporeling establishment in a vacant cell Probability of lichen sporeling establishment in a moss cell Maximum moss weight Probability of any moss slippage at each time step Minimum weight for moss to start stumping Sina constant C (Probgiump = C.weight) Maximum area of slumping moss Moss with lichen can grow? Period that newly vacant cells are unsuitable for moss sporelings 100 100 200 time steps 100 100 4 1 0 or 0.001 (see text) 0 or 0.001 (see text) 10 1.0 5 0.01 100 to 1000 cells (see text) 2 no Proc. LINN. SOC. N.S.W., 115, 1995 eis aii MEN TS te its iY ridin eis Le supetaey Med etyelnce el. till dedi . = Real SU ee nak! 459 =i ace maactote: Weve es 0a. ered Ri: ; = hata ti wa ae. a re el, ean Ja five we i (Sau ST 129, Wagers he Meg Akal OG =i eS : ; THN, ii, Re > Sr Oth oad (einmrahiah sare: ae 6 Un ag Ss bei ey wash tigen ¥ X ; ‘Tige ma wt Ps oes ; ; ‘ ; er, 2 VR, nor Ayer iy p Ra Goa, 0 . open. ee ey i Eg ie Srtt, 5 Maas. WP. Gee Py? Bel a ee eee a ; ra’ Avex aA. LV Wey aad 4 4 Drees \ , Hs cagetpo.si tink \ ne: ot ar th sei duties jet oat gr? ude i be To ei oe a sina Shecee eis epee x Pe Trey Vea Var iabiteres ane Anh. . su el chal eps aber pas Baty 4 eos a Mien be or Sbiaer baceiactinal o apelsy @ Saki Yoel “ers — + Le wr ibe parts © soil: ap ayn ul. Dy "7 a a a 7, i Petes we wierd tire girs tty, Naveve- Sa faaiel = a a : aa é ay a we : ii 7 ; as TT a 7 7 ' —_ “) Wont : : : hs 7 a 7 Ure 7 © 1 7 ; a _ = : At af, 7 ii : cs is Demography of Woody Plants in Relation to Fire: Telopea speciosissima R.A. BRADSTOCK (Communicated by D. KEITH) BRADSTOCK, R.A. Demography of woody plants in relation to fire: Telopea speciosissima. Proc. Linn. Soc. N.S.W. 115: 25-33 (1995). Data on survival and reproduction within populations of Telopea speciosissima were collected over 4 years. These data were used in conjunction with an existing demographic model to predict population trends under regimes of frequent fire. The results indicate that recruitment will be absent when fire frequency is high (~ 5 year cycles) because young juvenile plants are not sufficiently developed to resprout. Under 10 year cycles of fire, recruitment may be sufficient to maintain stable populations if growth and maturation of juveniles is relatively rapid. Further studies of growth are needed to validate this prediction. NSW National Parks and Wildlife Service, P.O. Box 1967, Hurstville N.S.W. 2220; manuscript received 18 August 1994, accepted for publication 16 November 1994. KEY WORDS: Fire ecology, post-fire recruitment, Telopea. INTRODUCTION Research on the demography of woody plants in fire-prone, Australian vegetation has focussed on species which accumulate seeds in on-plant or soil storages (e.g. Auld 1987a; Bradstock 1990; Cowling et al. 1990). In some cases (Auld 1987a; Bradstock and O’Connell 1988; Bradstock 1990; Burgman and Lamont 1992) demographic information of this kind has been used in quantitative models to predict the population dynamics of species. Such methods have been used to draw conclusions about the persistence of popu- lations when exposed to a range of fire regimes. Among woody plant species that inhabit fire-prone environments, there are a variety of traits and syndromes of survival and reproduction (Kruger 1983). There is evi- dence that some longer-lived woody species may rely on rapid post-fire seed production rather than a long-lived seedbank in fire-prone habitats (e.g. Auld 1987b, 1990). Plants of the genus Telopea (waratahs) appear to possess reproductive characteristics keyed to fire. In Telopea speciosissima, a species with a distribution centred around the Sydney region of eastern Australia (Blomberry and Maloney 1992), a pronounced pulse of post-fire flower- ing has been described by Pyke (1983) and Whelan and Goldingay (1989). Seeds are apparently non-dormant, germinating readily upon wetting (Blomberry and Maloney 1992) and plants are also known to resprout following fire. In these respects the species exhibits life-history characteristics that are broadly similar to Angophora hispida (Auld 1987b, 1990), a common shrub in the Sydney area. The aim of this paper was to collect demographic data for T. spectosissima and to use the data in a population model to investigate how well populations may persist under regimes of frequent fire. Emphasis was placed on frequent fire in this study because of the fire-induced flowering response that seems typical of the genus. Past management prac- tices in some areas have been to burn waratah populations as often as possible to deliber- ately promote spectacular displays of post-fire flowering. This has resulted in the exposure of some populations to fires as often as every 4-5 years. Demographic studies of other woody species of plants (e.g. Bradstock 1990) predict that populations can decline under such regimes of frequent fire. Proc. LINN. SOC. N.S.W., 115, 1995 26 DEMOGRAPHY OF WARATAH IN RELATION TO FIRE METHODS The model developed by Bradstock (1990) for populations of serotinous resprouters can be used to estimate the number of fire tolerant juvenile plants (Rs) recruited per adult needed to maintain population density and structure. Rs = (1-Sfa) + (Sfa. (1-[Sua]})) (1) (Sj. (Suj)')" where: Sua and Suj are the annual rates of survival of adults and juveniles respec- tively (unburnt conditions) ; Sfa and Sfj are the survival rates of adults and juveniles during fires; iis the interval between fires and n is the number of cycles of fire experienced by fire- tolerant juveniles (i.e. a parameter that reflects rate of growth) before maturation. Equation 1) effectively indicates the demand for recruitment. The adequacy of supply of individuals (Rs*) and thus the likelihood that populations will remain stable in numbers can be estimated from Rs* =B.E.Ss. Suyj. Sufj (2) where; B is the viable seedbank (number of viable seeds per adult); E is the propor- tion emerging as seedlings; Ss is the proportion of seedlings surviving to 3 years of age; Suyj is the proportion of young juveniles (plants > 3 years old) surviving until they experi- ence their first fire; and Suf] is the proportion of young juveniles that survive their first fire, thus becoming fire-tolerant juveniles. Equations 1) and 2) constitute a population model which assumes that seedling establishment is concentrated into a single post-fire period or event. The object of this study was to collect sufficient data to estimate these population parameters, so that solutions to the model could be explored. In particular, population stability was estimated under two scenarios of growth (n = 1&2 alternatively, high and low respectively; see equation 1) above) when subjected to a constant fire frequency (i) of 10 years. Within each growth scenario alternative simulations were performed to explore the sensitivity of predictions to variations in survival. The growth values were chosen to repre- sent plausible rates of juvenile growth derived from subjective impressions during the study. Formal long-term estimates of growth in juveniles were prevented by a fire during the study (see below). The study was done within the Brisbane Water National Park about 80 km north of Sydney. Telopea speciosissima is present there in a number of small populations situated in low open forest on deep yellow earths derived from laterite (Benson and Fallding 1981). The species is absent from neighbouring forests and woodlands situated on more sandy soils. The methods used were broadly similar to those of previous demographic studies of woody plants in the Sydney region (Auld 1987ab; Auld et al. 1993; Bradstock and Myerscough 1988; Bradstock and O’Connell 1988). Observations were carried out over a six year period (1987-1993) within populations of differing fire history (time since last fire), to measure aspects of fruit production and rates of survival of adults, juveniles and young juveniles. Previous demographic studies of fire-prone resprouters (Bradstock and Myerscough 1988; Auld 1990) have identified the young juvenile life-stage (before plants are able to tolerate fires), as the most critical in the life-cycle. The minimum age of fire tolerance demarcates the effective upper limit to fire frequency above which no re- cruitment is possible. An experiment was performed to examine fire-survival in young juveniles. Measurements of survival and fruit production were carried out in areas last burnt in 1976, 1980 and 1986 (Fig. 1). Two separate samples of plants were tagged in each fire Proc. LINN. SOC. N.S.W., 115, 1995 R.A. BRADSTOCK 27 history class during early 1987. Within each site a random sample of 50-90 adjacent adult and juvenile plants was tagged. Height, number of stems, and position of lignotuber relative to the soil surface were measured on each individual. In the 1980a site large num- bers of young juveniles were present, having established since 1984. A sample of 100 plants was tagged and monitored. Details of height, lignotuber exposure and diameter were recorded. 20, — 15 1976 (a) 1976 (b) 10 5 0 Rens Qi josnireny 20, r 15} 1980 (a) + 1980 (b) NUMBER OF PLANTS Y Y y Gy Z % GY ] 7 Y Y WAAAY QQ G yg c) 100 200 300 400 0 100 200 300 400 HEIGHT (CM) Fig. 1. Fruiting and survival as a function of plant size in T. speciosissima populations of different fire history. Numerals indicate number of deaths during the study in each size class, while hatching indicates numbers of plants which set fruits. Proc. LINN. SOC. N.S.W., 115, 1995 28 DEMOGRAPHY OF A WARATAH IN RELATION TO FIRE The tagged plants were revisited annually, and survival was checked along with flow- ering and fruiting in adults. All sites were burnt by an extensive wildfire on the 23-24/12/90. Fire behaviour on the 23/12/90 was extreme, resulting in total scorch or consumption of the forest canopy. Based on fuel consumption patterns the highest fire intensity was experienced in the 1976 and 1980 sites. There, some of the stems of live adults (2-3 cm thick) were completely consumed. Little ash remained in these sites and anecdotal information on rate of spread combined with likely fuel quantities (see Conroy 1993) indicated a maximum fire intensity of about 20000 kW/m within these sites. Immediately after the fire, all sites were revisited and the tags checked. In most sites the tags had been placed on the branches of individuals. This posed a problem at some sites where the branches were consumed. Where possible the charred stumps of stems were located and re-tagged, with the assistance of pre-fire maps of locations of tagged plants. In some instances no plant remains could be found and tags were left where found. Some tags, particularly in one 1976 sample, were not found. Plants were re-sampled in subsequent years and survival monitored. Some unallocated tags were matched to plants following the commencement of resprouting. Care was taken to ensure that unallocated tags were re-matched to their original individuals. Post-fire survival was checked in late 1991 or 1992. An experiment aimed at investigating fire tolerance of young juveniles was performed in 1988-89 at the 1980a site. Forty individuals were selected and randomly allocated in equal numbers to alternative treatments: stem severed or; stem severed plus burning. The treatments were performed in February 1988 and the plants were revisited at six monthly intervals thereafter. Prior to treatment, height, number of leaves and stems, lignotuber position and diameter were measured. Burning was carried out with a propane torch following removal of surrounding litter according to the method described in Bradstock and Myerscough (1988). This treatment was found to approximate the level of heating that would occur in a low intensity fire (about 500 kW/m; Bradstock and Myerscough 1988). The effects of treatments on survival were contrasted with survival of untreated plants monitored at the same site over the same time. Survival frequencies of treated and untreated plants were compared using a G test of independence (Zar 1974). Seedling establishment rates were estimated in the 1986 sites by searching for seedlings in the area surrounding clusters of fecund adults, after seeds had been shed from freshly ripened fruits in 1989-90. The viable seed crop of these adults was estimated from fruit counts. Values for mean number of seeds per fruit were derived by harvesting forty fruits from the site in 1989 and counting the number of intact seeds in each fruit. Viability of seeds from this sample was tested in the laboratory. Four replicates of 25 seeds were placed on saturated filter paper within petri dishes for 8 weeks. Germinants were reg- ularly scored and removed and the viability of remaining seeds was estimated using the tetrazolium test according to methods described in Bradstock (1990). RESULTS Survival During the study, the plants that set fruit were of varied size (Fig. 1). Therefore for the purpose of describing survival, no attempt was made to distinguish between adults and juveniles. Few deaths were recorded among adults and juveniles before the 1990 fire (Fig. 1; Table 1). In contrast, survival was lower among young juveniles at the 1980a site over the same period (Table 1). Survival of the 1990 fire (Fig 1; Table 1) was highly varied within and between life-stages and between sites. Fire induced mortality was unrelated to size characteristics in adults and juveniles. The cutting/burning experiment performed on Proc. LINN. SOC. N.S.W., 115, 1995 R.A. BRADSTOCK 29 young juveniles indicated low survival when burnt and cut (Table 2). Analysis indicated that cutting and burning significantly affected survival (G = 35.05, d.f. = 2, p < 0.001). There was a significant difference between burning and cutting treatments (G = 5.88, d.f.=1, p<0.025). TABLE 1 Survival (Su-unburnt, Sf- burnt) of T. speciosissima plants from populations of varying fire history. See text for definition of survival rates and life-stages. Note removal of 29 young juvenile plants from the 1980a sample, post-fire, due to destruction of tags and disturbance of the site. a) adults and juveniles Site No. initially No. of deaths Su No. of fire-related Sf tagged pre-fire (1987-90) deaths (1991-92) 1980a 42 0 1.0 1 0.98 1980b 49 0 1.0 1 0.98 1986a 80 5 0.94 3 0.96 1986b 73 0 1.0 4 0.95 1976a 63 0 1.0 2 0.97 1976b 50 4 0.92 8 0.83 b) young juveniles Site No. of No. of pre-fire deaths (Suyj) No. of fire- Sfyj plants 1988 1989 1990 related deaths 1980a 97 8 (0.09) 8(0.10) 2(0.03) 39 0.22 TABLE 2 Survival of young juveniles subjected to cutting and burning treatments. Survival Treatments Cut and burnt Cut only Control Alive 5 13 73 Dead 15 8 8 Fruit production Fruits were mostly produced in the more recently burnt sites (Fig. 2). In these sites, the size of annual crops declined prior to the 1990 fire (Fig. 2). Fruiting was restricted toa relatively small number of plants except in the 1980a site but, as noted, was spread among plants of varied size (Fig. 1). The first post-fire fruit crop in the 1986 sites was produced in 1988 after which fruiting increased and declined. Seedling establishment and survival There were 12.6 + 0.7 intact seeds per fruit (mean + se). The viable proportion of intact seeds was 0.82 + 0.10. Estimates of seedling establishment in the 1986 sites varied considerably between different parent plants (Table 3). All seedlings tagged in 1990 were subsequently killed by the fire in these sites. Estimation of recruitment under different fire regimes Juvenile recruitment was estimated for the two growth scenarios (see above) and two survival scenarios based on results in Table 1 (high survival, Suj & Sua = 0.995, Sfj & Sfa = 0.97; low survival, Suj & Sua = 0.92, Sfj & Sfa = 0.80). Values of juvenile recruitment (Rs, derived from equation 1) varied over a hundredfold range according to the estimates of growth used (Table 4). Recruitment estimates overall, however, were less sensitive to variations in growth than survival. Proc. LINN. SOC. N.S.W., 115, 1995 30 DEMOGRAPHY OF A WARATAH IN RELATION TO FIRE 150 1976 (a) 1976 (b) 100 : 100 O < 50 50 FE > ” - e) iGo 1 1 =i F722. Poss ———— i 0 ti i u 6 150 1r i a ) -& Oo 1980 (a) 1980 (b) S 100 4+ {100 2 or a. : 2 6 || ; 50 5 th : 7 ar aS na S eel ba ee Ae Oe 7 E < z - 150 slits q 0 ol F 1986 (a) 1986 (b) ao 100 i} 100 LL. | ce) 50 F 50 \ o 5 Lum a 0 1987 1988 1989 1990 1987 1988 1989 1990 YEAR Fig. 2 Fecundity of T. speciosissima plants from populations of different fire history. Hatched bars indicate total fruit numbers per sample and solid bars indicate proportion of each sample that set fruits. Site 1980a not sampled in 1989-90. TABLE 3 Establishment of T. speciosissima seedlings three years after fire in groups of adjacent adults, as a proportion of the intact viable seedbank (E) Site No. of adults in group Total intact fruits No. of seedlings E 1986a 4 10 9 0.087 4 12 6 0.049 3 3 5 0.194 3 6 3 0.048 1986b 3 36 29 0.078 2 2 4 0.190 3 13 28 0.210 2 12 8 0.065 ] 12 ] 0.008 Proc. LINN. SOC. N.S.W., 115, 1995 R.A. BRADSTOCK 31 TABLE 4 Scenarios of juvenile recruitment (Rs) necessary for a stable population, as a function of different values of growth and survival parameters (see text for definitions and values) Growth Survival High Low High 0.091 5.41 Low 0.099 15.56 DISCUSSION Effects of fire regimes Based mainly on the survival of young juveniles in the burning/cutting experiment and the 1990 fire (Tables 1 and 2), it can be concluded that sustained high frequency fire (<10 year cycle) will cause populations to decline. At 8-10 years postfire <25% of young juveniles are fire tolerant: it could be expected that survival would be lower at an earlier age, because lignotubers would be less well developed. For example, the initial cohort of newly emerged seedlings in the 1986 sites were killed by fire in 1990. Therefore zero recruitment and a decline in density would be expected under a 4-5 year cycle of fire. The estimates of juvenile recruitment, derived from equation 1), can be used to give a more detailed indication of population trends under a 10 year fire cycle. In order to do this, some indication of the likely supply of young juveniles (Rs*) is needed based on seedbank, establishment and survival of young juveniles as specified in equation 2). Assuming half of emergent seedlings (Ss = 0.5, equation 2) ) survive to 3 years of age (the transition age to the young juvenile phase; Bradstock 1990), equation 2) can be solved using estimates of the other parameters derived from the study (i.e. Suyj = 0.93 per annum; Sfyj = 0.22; E= 0.1) for a cohort of seedlings emergent three years after fire. The seedling survival value of 0.5 is intermediate within the range found for other co-habiting proteaceous shrubs in the study area (Bradstock 1990; Bradstock and O’Connell 1988). These values give an estimate for supply of young juveniles of Rs* = 0.008.B (where B is the viable seedbank, equation 2)). The level of seedbank per adult (B) necessary to meet the recruitment scenarios in Table 4 would lie between the extremes of 1 or 1.1 (growth parameter n=1; high or low survival scenario, Table 4) to 61.6 or 176.8 (growth parameter n=2; high or low survival) intact fruits per adult, assuming that there is a mean of 11 viable, intact seeds per fruit (see above). It is apparent from data on post-fire fruiting in the 1986 sites that there would be adequate fruits and seeds to meet the high but not the low growth scenario (Fig. 2). Even taking into account fruit crops in the first and third year of post-fire flowering (not accounted for in the above calculations) in these sites, it is evident that there would be insufficient seedbank to maintain populations under the slower growth scenario. Fruiting of an order of magnitude greater than that measured would be necessary, an unlikely occurrence given that natural fruit production and seed set found in a study of a popula- tion south of Wollongong (Whelan and Goldingay 1989) were similar to that reported here. Either an increase or decrease in seedling survival would not substantially alter these conclusions. If, for example, seedling survival was substantially lower (Ss=0.10) about 5 times more seedbank would be required for populations to be maintained (low growth scenario): i.e. each adult would have to produce a total of about 5 fruits in total from post-fire flowering. This is about the level of fruiting sustained in the 1986 popula- tions after fire (Fig. 2). An increase in survival (Ss=0.90) would approximately halve the amount of seedbank required for population maintenance. This would be insufficient to Proc. LINN. SOC. N.S.W., 115, 1995 32 DEMOGRAPHY OF A WARATAH IN RELATION TO FIRE match the level of seedbank required under the low growth scenario. Population stability may be possible under a fixed fire 10 year cycle but a definite conclusion will require further work on juvenile growth and maturation rates. Data were not available to estimate the seedling survival parameter, mainly because the cohort of emergents tagged in 1990 (1986 sites) were eliminated by the 1990 fire. Implications for fire management Acquisition of growth data would provide a suitable basis for further development of the model used here to explore extinction risk as a function of fire regimes in the man- ner described by Burgman e¢ al. (1993). A version of the model predicting population extinction risk or viability would provide the most appropriate vehicle for exploring and comparing fire management scenarios. It would also provide a more appropriate means of exploring the outcome of stochastic variations in population parameters. In the absence of such a model, a number of general conclusions can be drawn con- cerning management of fire and the maintenance of Telopea populations in the study area. Without data on juvenile growth it could be concluded that a conservative manage- ment aim would be to avoid sustained periods (>5 cycles) of high frequency fire (interval between fire <10 years). Occasional longer intervals between fire will provide a window of opportunity for effective recruitment of juveniles (Bradstock 1990). In this respect the population response to fire and the fire management objectives for conservation of waratah populations are broadly similar to other woody species capable of resprouting in the Sydney area (Bradstock 1990, Auld e¢ al. 1993). The past management strategy of deliberate high frequency burning to promote displays of flowering would not be sustainable in the long-term. Coupled with the prob- lem of flower picking in some sites adjacent to roads (e.g. 1976 and 1986 sites) and the consequent reduction in seedbank and recruitment, such as management strategy could result in the depletion of populations. A more appropriate scenario for management is to allow the frequency of fire to vary in the longer term (Bradstock et al. 1995). Given that some of the populations monitored in this study (1986 sites) have experienced four fires in the period 1976-90, some thought needs to be given as to how a variable regime can be achieved in practice. ACKNOWLEDGEMENTS I wish to thank John Steer, Janet Cohn and Mark Tozer for their kind assistance with field work and Judith Scott for her help with preparation of the manuscript. References AULD, T.D. 1987a. — Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: survivorship throughout the life cycle, a synthesis. Aust. J. Ecol. 12: 139-51. AULD, T.D. 1987b. — Post-fire demography in the resprouting shrub Angophora hispida (Sm.) Blaxell: seed production, dispersal, seedling establishment and survival. Proc. Linn. Soc. N.S.W., 109: 259-69. AuLb, T.D. 1990. — The survival of juvenile plants of the resprouting shrub Angophora hispida (Myrtaceae) after a simulated low-density fire. Aust. J. Bot. 38: 255-60. AULD, T.D., BRADsTOCK, R. and KEITH, D. 1993. — Fire as a threat to populations of rare plants. Australian National Parks & Wildlife Service Endangered Species Program. Endangered species project No. 31. Unpublished Report, NSW National Parks & Wildlife Service. BENSON, J.S. and FALLDING, H. 1981. — Vegetation Survey of Brisbane Water National Park and environs. Cunninghamia, | (1): 79-113. BLOMBERY, A.M and MALONEY, B. 1992. — The Proteaceae of the Sydney Region. Kangaroo Press, Kenthurst, NSW. BRADSTOCK, R.A. 1990. — Demography of woody plants in relation to fire: Banksia serrata Lf. and Isopogon anemoni- folius (Salisb.) Knight. Aust. J. Ecol., 15: 117-32. BRADSTOCK, R.A.: MYERSCOUGH, P.J. 1988. — The survival and population response to frequent fires of two woody resprouters Banksia serrata and Isopogon anemonifolius. Aust. J. Bot., 36: 415-31. BRADSTOCK, R.A. and O’CONNELL, M.A. 1988. — Demography of woody plants in relation to fire: Banksia ericifolia Lf. and Petrophile pulchella (Schrad.) R. Br. Australian Journal of Ecology, 13: 505-18. Proc. LINN. SOC. N.S.W., 115, 1995 R.A. BRADSTOCK 33 BRADSTOCK, R.A., KEITH, D.A., and AULD, T.D., 1995. Fire and conservation: Imperatives and constraints on managing for diversity In BRADSTOCK, R.A., AULD, T.D., KEITH, D.A., KINGsFoRD, R., LUNNEy, D. and SILVERTSEN, D. (eds). Conserving Biodiversity: Threats and Solutions , Pp. 323-33. Surrey Beatty & Sons, Sydney. BuRGMAN, M.A. and LAMoNnr, B.B., 1992. — A stochastic model for the viability of Banksia cuneata populations; environmental, demographic and genetic effects. Journal of Applied Ecology, 29: 719-27. BURGMAN, M.A., FERSON, S. and ACKAKAYA, H. 1993. — Risk Assessment in Conservation Biology. Chapman & Hall. Conroy, B. 1993. — Fuel managementstrategies for the Sydney Region. In Ross, J. (ed.). The Burning Question: Fire Management in NSW. Pp. 73-83. Univ. New England, Armidale. CowLING, R.M., LAMONT, B.B. and ENRIGHT, N.J. 1990. — Fire and management of south-western Australian banksias. In SAUNDERS, D.A., HOPKINS, A.J.M. and How, R.A. (eds). Australian Ecosystems: 200 years of Utilization, Degradation and Reconstruction. Proceedings of the Ecological Society of Australia, 16. Surrey Beatty & Sons, Chipping Norton, Sydney. KRUGER, F.J. 1983. — Plant community diversity and dynamics in relation to fire. In KRUGER, F.J., MITCHELL, D.T. and JARVIS, J.U.M. (eds). Mediterranean-Type Ecosystems. The role of nutrients. Pp. 446-72. Springer-Verlag, Heidelberg. PyYKE, G.H. 1983. ee Relationship between time since the last fire and flowering in Telopea speciosissima and Lambertia formosa. Australian Journal of Botany, 31: 293-6. WHELAN, R.J. and GOLDINGAY, R.L. 1989. — Factors affecting the fruit set in Telopea speciosissima (Proteaceae): the importance of pollen limitation. J. Ecol. 77: 1123-34. ZaR, J.H. 1974. — Biostatistical Analysis. Prentice-Hall, USA. Proc. LINN. SOC. N.S.W., 115, 1995 neta beg oy ar a WE or ‘ i e ‘ eh java ne : i} } BN saute We Faye “ PRS eee (uae A eee Band as r ee fie nw ot naa Po :. ie yy ATTN Pah ina’ aslo Be OAL D it en ! ; d 4 i [ hast 1 ss, Me arty mag is en, Bui mn ne Lge ll tet The Population Dynamics of the Mangrove Shrub Aegiceras corniculatum (Myrsinaceae): Fecundity, Dispersal, Establishment and Population Structure PETER J. CLARKE (Communicated by D.T. KEITH) CLARKE, PETER, J. The population dynamics of the mangrove shrub Aegiceras corniculatum: fecundity, dispersal, establishment and population structure. Proc. Linn. Soc. N.S. W. 115: 35-44 (1995). The mangrove shrub Aegiceras corniculatum (L.) Blanco commonly occurs in estuaries north from Merrimbula in southern NSW. Studies of the population dynamics were undertaken at Jervis Bay as part of wider baseline studies of the marine environment of the Bay. Populations in Jervis Bay flowered regularly during spring and produced viviparous fruit by the following autumn. Predispersal mortality of fruits was very high (92%), but exclusion of herbivores reduced mortality to 53%. About 360 viable propagules were produced per plant each year, representing some 32% of above-ground productivity. Once dispersed propagules can remain bouyant in seawater for up to three months, but under brackish conditions sink within a week and do not refloat. During the dispersal phase, propagules landing on intertidal sediments had a low probability of establishing because of predators and tidal disturbance. Establishment, however, appeared to be in- trinsically slow because of the season in which propagules are dispersed. Shadehouse experiments also showed that propagules establish more rapidly in 10% and 50% seawater than in full seawater. Adult populations of Aegrceras corniculatum were conspicuously zoned in relation to the co-occurring mangrove Avicennia marina (Forsk.) Vierh. The former usually occurred at the higher edge of the mangrove zone in the marine zone of the estuary, whereas popu- lations in the riverine zone of the estuary dominated the lower edge of the mangrove zone. Establishment and recruitment appear to be episodic and highly patchy in space, although without an unambiguous measure of the age structure it is difficult to infer any population trends. Evidence from aerial photography suggests that the spatial extent of populations of A. corniculatum has remained static over the past 50 years, while that of Avicennia marina has spread. Based on broad regeneration syndromes, I predict that Avicennia marina would replace Aegiceras corniculatum under conditions of disturbance, but under long-term stable conditions the converse would apply. Peter J. Clarke, Botany Department, University of New England, Armidale, N.S.W. 2351, Australia; manuscript received 21 January 1994, accepted for publication 20 July 1994. KEYWORDS: Mangroves, Aegiceras corniculatum, Avicennia marina, co-existence, population dynamics. INTRODUCTION Aegiceras corniculatum (L.) Blanco, commonly referred to as the river mangrove, is a low tree or shrub that grows in the intertidal zone and has a widespread distribution throughout the shores of Australia and South-east Asia (Tomlinson, 1986). In New South Wales it is commonly found associated with, or adjacent to, Avicennia marina Forsk. Vierh, the grey mangrove, and it occurs in most estuaries open to the sea, northwards from its southern limit at Merrimbula (West et al., 1985). The floral phenology has been described for populations in Queensland (Duke et al., 1984; Hutchings and Saenger, 1987) and in NSW (Carey and Fraser, 1932; Clarke, 1994). In NSW flowering commencesas early as July and is completed by December while fruits are fully developed by the following April or May (Clarke, 1994). The fruit (capsule) contains a precociously developed embryo which often ruptures the seed coat whilst still attached to the parent. Once the propagules are on Proc. LINN. SOC. N.S.W., 115, 1995 36 POPULATION DYNAMICS OF AEGICERAS CORNICULATUM the ground the radicle penetrates the substrate and elongates and lifts the plumule. Subsequently the plumule extends through the remains of the fruit wall and the shoot emerges (Tomlinson, 1986). Whilst there are good descriptive accounts of the floral biology of Aegiceras (Tomlinson, 1986), there is little information about the fecundity of plants and factors limiting fecundity. Similarly, there are many accounts of the distributions in space, espe- cially zonations across the intertidal zone (e.g. Hutchings and Saenger, 1987; Clarke, 1993a), but studies of the population dynamics of seedlings and adults are rare (e.g. Osborne and Smith, 1990). Several studies have been undertaken to examine the growth of seedlings under glasshouse conditions, all of which show enhanced growth in dilutions of seawater (Clarke and Hannon, 1971; Ball and Farquhar, 1984). In mangrove species with precociously developed embryos the establishment and subsequent survival of seedling populations can be studied more easily than in those plants with a post-dispersal dormant phase. In these circumstances, where there is no buried seed bank, models about establishment and recruitment to adult populations can be tested by field experiments (e.g. Smith, 1987a; Clarke and Myerscough, 1993). The aims of this study were to: 1) determine the fecundity of Aegiceras corniculatum, henceforth referred to as Aegiceras, 2) examine factors limiting the production of viable fruits, 3) describe the dispersal properties of propagules, 4) determine what limits propagule establishment and seedling survival, 5) infer adult population dynamics from population structure and aerial photography, and 6) compare regeneration syndromes of Aegiceras with that of the co-occurring species Avicennia marina, henceforth referred to as Avicennia. METHODS Fecundity Numbers of floral buds, flowers and fruits of Aegiceras were followed at monthly intervals over a single reproductive season (Clarke, 1994). In total 741 buds were followed on eight trees randomly selected from widely spaced populations at Jervis Bay. The effect of predation by insects and other herbivores on the survival of fruits was assessed by bagging newly formed fruits (606) and recording the numbers of fruits that survived. The total number of mature fruits or propagules that shrubs produced was measured by counting the number of fruits caught by litter-traps which spanned the width of individual shrubs (see Clarke, 1994). Sixteen randomly selected shrubs were sampled from widely spaced populations at Jervis Bay. Litterfall was removed from each trap at monthly intervals for three years and the numbers of intact and herbivore-damaged fruits counted. Dispersal The dispersal properties of the propagules were examined in buoyancy experi- ments that examined the effects of salinity on the buoyancy properties of propagules. Ten propagules, of the same developmental stage, from four trees were placed in treatments of full seawater, 50% and 10% seawater. The number of propagules floating or sinking was recorded at regular intervals together with the presence of the pericarp and the viability of propagules. Field observations were also made with marked propagules at two locations. At each location 20 propagules were marked with a non-toxic pen and attempts were made at weekly intervals to recover marked propagules. Establishment and survvwal Patterns of establishment of propagules and survival of seedlings were examined in two inlets at Jervis Bay. In each inlet two widely separated plots were established within Proc. LINN. SOC. N.S.W., 115, 1995 P.J. CLARKE 37 existing strands of Aegiceras. In each plot four 50 x 50cm cages were randomly placed and fixed so that potential predators such as fish and crabs were excluded. Thirty mature propagules of Aegiceras were then placed in each cage and their establishment and fates followed by 12 months. To examine the effects of predators outside the cages 30 propag- ules were tethered on fishing line and placed outside the cages. A field experiment was also undertaken to examine the effect of sediment condi- tions on establishment. In each of two tidal inlets (Moona Moona Creek and Cararma Inlet, see Fig. 1) two plots were selected and within each area the surface sediment was either disturbed or left undisturbed. Four cages were randomly placed in each plot and five propagules of Aegiceras were placed in each cage. Finally, the establishment of propagules was examined in a shadehouse experiment where thirty propagules from three locations were placed on natural sediment water- logged with 10% seawater, 50% seawater, and 100% seawater. Callala Cararma Creek | Inlet Currambene Creek Jervis Bay = Moona Moona Creek Fig. 1. Location of study sites in Jervis Bay. Population patterns The density and height of Aegiceras were measured in 22, 5 x 5m plots randomly placed within four inlets at Jervis Bay (see Clarke, 1993a). Only plants exceeding the five leaf stage were measured in this way, otherwise they were treated as seedlings and sampled differently. If seedlings were present in plots then they were subsampled with a 0.5 x 0.5m quadrat. Where Aegiceras co-existed with Avicennia the frequency of occurrence at the waters edge or at the landward edge of mangrove stands was also recorded from 50 randomly placed transects in the upper and lower reaches of estuaries. A G-test of independence was used to test if stand locations (front or back) were independent of section in estuary (upper or lower). Proc. LINN. SOC. N.S.W., 115, 1995 38 POPULATION DYNAMICS OF AEGICERAS CORNICULATUM RESULTS Fecundity Flower buds initiate in May and the complete cycle from bud initiation to the abscission of fruits (propagules) takes about 14 months (for details see Clarke 1994). About 86% of floral buds flowered during the spring and summer months and of these about 62% formed young fruits (Fig. 2). Fruits matured over the summer months and by the time they were mature enough to produce propagules about 8% of the original flower bud population survived (Fig. 2). Bagging to exclude insects and other herbivores signifi- cantly increased survival of new fruits to 47% (F,7s = 38.6, P<0.001). 100 — Flower buds — Flowers 80 60 40 End of flowering surviving % Mature fruits 20 0 3 6 9 12 Months since bud initiation Fig. 2. Mean (s.e.) survival of Aegiceras flower buds, flowers, and fruits over three years and over all plants. Total number of buds followed = 740. The total number of viable fruit and the number of mature fruits attacked by herbi- vores did not differ significantly year to year (F513, = 0.98, P> 0.4), (Foisg= 0.58, P> 0.5). The mean number of viable fruits produced on an individual over three years was 356 (s.e. 44) and of these 58 showed signs of being affected by herbivores. Dispersal About 20% of propagules sank immediately when placed in treatments of different salinities. Of those that remained buoyant, propagules placed in 10% seawater sank sooner than those placed in 100% seawater (Fig. 3). After five days no propagules remained floating in tapwater while about half remained floating in 100% seawater (Fig. 3). Propagules placed in 100% seawater were also slower to lose their pericarps than those in 50% and in 10% tapwater. After a month all propagules had sunk and none showed any signs of decomposition. Few propagules marked and released in the field were recovered. Proc. LINN. SOC. N.S.W., 115, 1995 P.J. CLARKE 39 — io) —t—- 10% Seawater —l4— 50% Seawater (oe) 100% Seawater Number of propagules floating 0 24 48 UP 96 120 144 Hours Fig. 3. Mean (s.e.) number of Aegiceras propagules buoyant in treatments of 10%, 50% and 100% seawater. Establishment and survival Less than 10% of propagules tethered in the field survived the first month and none of these managed to establish (Fig. 4). Subsequently no propagules could be found attached to their tethers. In contrast those propagules placed in cages were able to germi- nate (split the pericarp), and establish (Fig. 4). Overall, very few propagules actually estab- lished and produced leaves (<1%), but those that did survived for up to two years when observations ceased. Of the 160 propagules used in the experiment to examine the effects of distur- bances only five managed to establish as seedlings and all of these occurred in disturbed plots. More seedlings established on sediments flooded with 10% and 50% seawater than with 100% seawater (Fig. 5). Propagules took up to three months to establish in all treat- ments but achieved their fastest establishment rates on the 10% seawater treatment (Fig. 5). In all treatments the first pair of leaves took up to six months to fully expand, thereafter when all seedlings were treated with 10% seawater for a further six months only a few seedlings developed further leaves. Population patterns The results from sampling adult populations have been reported elsewhere (see Clarke, 1993a). In summary, they show a highly skewed distribution with most plots con- taining only 1-4 plants. Adult shrubs had a remarkably normal height distribution around a mode of 60-80cm and 95% of these plants were multistemmed. Seedling densities were very high (>100m*), but were very localised as only five of the 22 plots where adults occurred also had seedlings. Proc. LINN. SOC. N.S.W., 115, 1995 40 POPULATION DYNAMICS OF AEGICERAS CORNICULATUM Number of germinated propagules 0 2 4 6 8 10 12 Months since propagules dispersed Fig. 4. Mean (s.e.) number of germinating Aegiceras propagules in caged O and uncaged [_] treatments in the field. Note that no propagules established in the uncaged treatment and that a total of four propagules estab- lished in the caged treatments. Both species of mangrove co-occurred in the 33 out of the 50 mangrove stands sam- pled by transects. When both species of mangrove co-occurred in stands, their position at the seaward edge of the stand or at the rear of the stand was not independent of where the stands were in the estuary (G= 7.8, P > 0.01). In other words, Aegiceras was more frequent at the landward edge of mixed stands in the lower estuary, whereas it was more frequent at the river edge of mixed stands in the upper estuary. DISCUSSION Reproduction and fecundity Shrubs of Aegiceras flower and produce fruits regularly even though the complete cycle from bud initiation to fruit abscission may overlap (Clarke, 1994). Some mortality (14%) of flower buds occurs prior to flowering and appears to result from insects burrow- ing into the base of the receptacle. Following a prolonged period of flowering, which pos- sibly promotes outcrossing, many young fruits are formed. It is not known how many of these contain embryos, but bagging to exclude herbivores indicates that most contain viable fruits. The dramatic increase in survival of fruits that were bagged (47%) compared with those left unbagged (8%) indicates that herbivores have a significant effect on the fecundity of plants. This contrasts with Avicennia where exclusion of herbivores using bags did not increase fruit survival, although fruit set was enhanced when they were treat- ed with insecticide (Clarke, 1992). Overall levels of fruit predation in Aegiceras are higher than that reported in Avicenniain NSW (Clarke, 1992), but are of a similar magnitude to those reported in tropical mangroves (Robertson et al., 1990). Observations of fruits collected from littertraps suggest that many fruits had been attacked by larvae that enter the base of the capsule near the calyx and consume the embryo leaving only the pericarp. Proc. LINN. SOC. N.S.W., 115, 1995 PJ. CLARKE 4] oS 10% seawater Bape are SOM, seawater ——_ 100% Seawater Number of seedlings standing Months Fig. 5. Mean (s.e.) numbers of Aegiceras propagules establishing (standing upright) to become seedlings in shadehouse experiments over a range of salinities. Overall, the numbers of viable propagules collected from mature shrubs of Aegiceras averaged c. 360 per year, thus about 4,500 flower buds per shrub are produced each year. This compares with some 61,000 flower buds and 2,000 fruits produced on a mature tree of Avicennia (Clarke, 1992). Nevertheless, the relative proportion of above-ground productivity devoted to reproduction in Aegiceras (c. 32%) is far higher than in Avicennia (c. 9%), which is remarkable for a perennial plant (see Harper, 1977 p. 660). Dispersal The dispersal phase in the life history of Aegiceras, like that of most mangroves, is rel- atively short when compared to many terrestrial shrubs because the propagules are viviparous. Factors influencing the distribution and fate of dispersed propagules include their buoyancy, period of obligate dispersal, and longevity for establishment (Rabinowitz, 1978). These factors together with abiotic (tides and currents) and biotic factors (predators and pathogens) influence not only the colonisation of new habitats but the rearrangement and replacement of populations within existing stands (Clarke, 1993b). Propagules of Aegiceras can remain bouyant in seawater for up to three weeks, but under brackish conditions sink within a week. This pattern is similar to that found in Avicennia, although propagules do not refloat under brackish conditions (Clarke and Myerscough, 1991). Proc. LINN. SOC. N.S.W., 115, 1995 42 POPULATION DYNAMICS OF AEGICERAS CORNICULATUM Establishment and survival Shadehouse experiments showed that the minimum time for propagule establish- ment and the transition to a seedling was at least two weeks and that propagules can take up to three months to establish, i.e. take root and lift the plumule from the ground. This establishment phase is much longer than that reported for Avicennia, which readily estab- lishes in the field over four weeks and up to 80% of propagules can establish when they are caged (Clarke and Myerscough, 1993). In contrast, establishment success of Aegiceras in the field was very low (<1%) and field observation of caged propagules suggested that propagules either failed to establish because the radicle did not develop sufficiently, or that small invertebrates consumed propagules. In the shadehouse there was a clear pref- erence for establishment of seedlings under low salinity conditions; nevertheless propag- ules took up to three months to establish and six months for the stem axis to emerge and leaves to develop. The difference between the more rapid and successful establishment of Avicenniaand the slower establishment of Aegiceras possibly relates to the timing of release: the former is released during summer, whereas the latter is released during autumn in south-east Australia. Under conditions of slow establishment, propagules of Aegiceras appear to be highly susceptible to herbivory as no propagules were recovered from their tethers. Similar results have been found in tropical mangrove forests in Queensland where, in the high intertidal zone, all propagules in Aegiceras were consumed within 14 days. However, in the low intertidal zone and in canopy gaps fewer propagules were consumed (Osborne and Smith, 1990). In the present study no such differential effects were tested, but caged and uncaged treatments were spread over a range of tidal positions, salinities, and canopy cover, and in all cases propagules appear to have been consumed. Population patterns and processes Adult populations of Aegiceras showed conspicuous patterns of distribution in rela- tion to Avicennia with which it commonly occurs. Populations of Aegiceras in the more saline parts of estuaries usually occur high on the shore, i.e. between Avicennia and the saltmarsh. However, those populations in the riverine parts of estuaries usually occur low on the shore, i.e. at the front of Avicennia. This pattern may be explained by the dispersal and establishment attributes of propagules. Propagules of Aegiceras dispersed into the seawater section of an estuary remain buoyant for longer and would tend to strand at the upper tidal limits i.e. the saltmarsh. Competition from the more vigorous and larger seedlings of Avicennia may also displace the zone in which Aegiceras can exist towards the saltmarsh. In the more brackish ends of estuaries, where tidal amplitude is reduced, propagules of Aegiceras sink and establish faster whereas those of Avicenniarefloat (Clarke and Myerscough, 1991) and are probably redistributed to the rear of the stand. Predators such as crabs may also influence these dis- tributional patterns, but because recruitment appears to be so episodic these models will be difficult to test in field experiments. Rare, but dense, occurrences of seedlings of Aegiceras together with homogeneous adult stands suggest patchy recruitment in space and time, although the average height distribution for populations in Jervis Bay seems to indicate steady recruitment and mortal- ity, assuming height is correlated with age. Population patterns on aerial photographs support this suggestion and show that the gross distribution of adults at Jervis Bay has not changed over 50 years, while those of Avicennia expanded both seaward and landward. Clearly a better understanding of the age structure of adult populations and the transition rates (recruitment) between age classes is required before detailed models of population dynamics can be developed. Finally, some general predictions can be made about how co-existing mangrove populations will interact based on regenerative attributes (sensu Myerscough, 1990). Both Avicennia and Aegiceras appear to reproduce regularly and at an early age relative to life- Proc. LINN. SOC. N.S.W., 115, 1995 P.J. CLARKE 43 span. Similarly, both species invest large amounts of resources into precociously devel- oped embryos, possibly at the expense of growth and maintenance at low latitudes (Clarke, 1994). Investment in large propagules in Avicenniaensures a high success rate for establishment, whereas the propagules of Aegiceras are intrinsically less likely to establish, possibly because of their smaller mass and the time of year that they are dispersed. Despite high levels of establishment, seedlings of Avicennia are unlikely to recruit unless they hap- pen upon a ‘regeneration niche’, i.e. an open disturbed patch (Clarke and Allaway, 1993). Seedlings of Aegiceras, on the other hand, do not appear to require disturbance for recruitment as evidenced by the presence of shrubs in the understorey of Avicennia stands, but are prone to extensive predation. From the limited understanding of these population attributes I suggest that Avicennia would respond rapidly to disturbance and inhibit recruitment of Aegiceras, as is suggested to occur in southern Queensland (Quinn and Beumer, 1984). Under more stable conditions, especially in the upper reaches of an estuary, populations of Aegiceras would establish and form dense stands in the understorey of Avicennia. If stable conditions persisted then establishment and recruitment of Avicennia seedlings would be inhibited and eventually the stands would be dominated by Aegiceras until the next disturbance. ACKNOWLEDGEMENTS David Keith, Frances Quinn, and John Williams kindly took time to preview this manuscript and made helpful comments. Lani Ratter, Gavin Tinning and Mark Fisher provided assistance in collecting the field data. Trevor Ward and Charles Jacoby of CSIRO Division of Fisheries encouraged the research and the Department of Defence funded the study as a part of the Jervis Bay Baseline Studies. References BALL, M.C. and FARQUHAR, G.D., 1984. — Photosynthetic and stomatal responses of two mangrove species. Aegiceras corniculatumand Avicennia marina, to long-term salinity and humidity conditions. Plant Physiol. 74: 1-6. Carey, G., and Fraser, L., 1932. — The embryology and seedling development of Aegiceras majus Gaertn. Proc. Linn. Soc. N.S.W. 57: 341-360. CLARKE, N.J., and HANNON, N,J., 1971. — The mangrove swamp and salt marsh communities of the Sydney district. III. Plant growth in relation to salinity and waterlogging. J. Ecol. 58: 351-69. CLARKE, P. J., 1992. — Predispersal mortality and fecundity in the grey mangrove (Avicennia marina) in south- eastern Australia. Aust. J. Ecol. 17: 161-168. CLARKE, P.J., 1993a. — Mangrove, saltmarsh and peripheral vegetation of Jervis Bay. Cunninghamia 3 (1): 231-253. CLAREE, P.J., 1993b. — Dispersal of grey mangrove (Avicennia marina var. australasica) propagules in south-eastern Australia. Aquatic Bot. 45: 195-204. ‘CLARKE, P.J., 1994. — Baseline studies of temperate mangrove growth and reproduction; demographic and litter- fall measures of leafing and flowering. Aust. J. Bot. 42: 37-48. CLarKE, P.J., and ALLAWAy, W.G., 1993. — The regeneration niche of the grey mangrove (Avicennia marina): effects of salinity, light and sediment factors on establishment, growth and survival in the field. Oecologia 93: 548- 556. CLARKE, P.J., and Myerscoucu, P.J., 1991. — Buoyancy of Avicennia marina propagules in south-eastern Australia. Aust. J. Bot. 39: 77-83. CLARKE, P.J., and MYERSCOUGH, P.J., 1993. — The intertidal distribution of the grey mangrove (Avicennia marina) in south-eastern Australia; the effects of physical conditions, interspecific competition, and predation on establishment and survival: Aust. J. Ecol. 18: 307-315. Duke, N.C., BUNT, J.S., and WILLIAMS, W.T., 1984. — Observations on the floral and vegetative phenologies of north-eastern Australian mangroves. Aust. J. Bot. 32: 87-99. HarPER, J., 1977. Population Biology of Plants. Academic Press: London. HUTCHINGS, P., and SAENGER, P., 1987. — Ecology of Mangroves. University of Queensland Press: St Lucia. MYERSCOUGH, P.J., 1990. — Comparative plant ecology and the quest for understanding of Australian plants. Proc. Linn. Soc. N.S.W. 112 (4): 189-199. OsBoRNE, K., and SMITH., T.J., 1990 — Differential predation on mangrove propagules in open and closed canopy forest habitats. Vegetatio 89: 1-6. QUINN, R.H., and BEUMER, J.P., 1984. — Wallum Creek; a study of the regeneration of mangroves. In: COLEMAN, R.J., COVACEVICH, J., DAVIE, P. (eds) Focus on Stradbroke. Boolarong Press, Brisbane. RABINOWITZ, D., 1978. — Dispersal properties of mangrove propagules. Biotropica 10(1): 47-57. Proc. LINN. SOC. N.S.W., 115, 1995 44 POPULATION DYNAMICS OF AEGICERAS CORNICULATUM ROBERTSON, A.I., GIDDINS, R.L., and SMITH, T.J., 1990. — Seed predation by insects in tropical mangrove forests: extent and affects on seed viability and growth of seedlings. Oecologia 83: 213-219. SmiTH, T.J., 1987a. — Effects of seed predators and light level on the distribution of Avicennia marina (Forsk.) Vierh. in tropical, tidal forests. Est. Coast Shelf Sci. 25: 43-51. SMITH, T.J., 1987b. — Effects of light and intertidal position on seedling survival and growth in tropical tidal forests. J. Exp. Mar. Biol. Ecol. 110: 133-146. TOMLINSON, P.B., 1986. — The Botany of Mangroves. Cambridge University Press: Cambridge. WEST, R.J., THOROGOOD, C., WALFORD, T., and WILLIAMS, R., 1985. — An Estuarine Inventory for New South Wales, Australia. Fisheries Bulletin 2, Department of Agriculture, NSW. Proc. LINN. SOC. N.S.W., 115, 1995 Edaphics and Fire: An Interpretative Ecology of Lowland Forest Vegetation on Granite in Northeast Tasmania F. DUNCAN and M.J. BROWN (Communicated by D. KEITH) DUNCAN, F. and Brown, M.jJ. Edaphics and fire: an interpretative ecology of lowland forest vegetation on granite in northeast Tasmania. Proc. Linn. Soc. N.S.W. 115: 45-60 (1995) Forest, scrub and moorland vegetation near Old Chum Dam in northeast Tasmania is described. Analyses of the floristics and some environmental variables suggest that moisture availability, drainage and edaphic factors have a major influence on vegetation composition and structure. The vegetation itself, largely through differences in the flammability of its understorey, encourages fire frequencies and intensities which maintain its current heterogeneity. Trends observed in the study area are similar to those reported from comparable forested areas in Tasmania and on the southeastern Australian main- land. F. Duncan and M.J. Brown, Forestry Tasmania, 199 Macquamne Street, Hobart, Tasmania 7000; manuscript received 16 August 1994, accepted for publication 19 April 1995. KEYWORDS: vegetation, ordination, classification, edaphic, fire, Northeast Tasmania. INTRODUCTION The vegetation of a forested area near Old Chum Dam, in northeastern Tasmania, was surveyed in March 1990 as part of a wider study into the effects on the biota of forestry operations and forest conservation prescriptions. The vegetation is representative of that occupying much of the forested lowland country in the Northeast. There are few published descriptions of lowland vegetation in northeastern Tasmania, though the upland vegetation has been documented more thoroughly (e.g. Ellis, 1985; Davies and Davies, 1989). Remnant heaths, forests and woodlands occurring on the Great Northern Plain, to the north of the study area, are described (Kirkpatrick and Wells, 1987), and forest vegetation on Spurrs Rivulet, to the east of the study area, has been analysed (Peters, 1984). Stephens and Cane (1938) and Pinkard (1980) provide general descriptions of the vegetation of northeastern Tasmania, and its relationship with the environment. Statewide analyses of major vegetation types (e.g. Kirkpatrick, 1977; Jarman et al., 1984; Duncan and Brown, 1985; Kirkpatrick et al., 1988; Jarman et al., 1988; Pannell, 1992) include forest and non-forest communities found in the Northeast. Species nomenclature in this paper follows Buchanan et al. (1989). THE STUDY AREA The study area consists of undulating country in the upstream catchment of the Great Musselroe River, in the vicinity of Old Chum Dam (41°06'S 148°03'E). The general location is shown in Figure 3. Altitude varies from 100m to 250m above sea level. The area occurs within the humid warm climatic zone defined by Gentilli (1972). Average annual rainfall at Pioneer, some 9km to the west of Old Chum Dam, is 978mm and has a pro- nounced seasonality. January is the driest month (mean monthly rainfall of 45mm) and July the wettest (mean monthly rainfall of 112mm). Temperature records are available Proc. LINN. SOC. N.S.W., 115, 1995 46 LOWLAND FOREST IN NORTHEAST TASMANIA from St Helens, a coastal town 33km to the southeast. February is the hottest month (mean monthly maximum 22.9°C, minimum 11.9°C) and July the coldest (mean monthly maximum 14.2°C, minimum 2.5°C). Temperatures in the study area are likely to be more extreme than those on the coast. The ground rock is Devonian/Carboniferous granite, which crops out at several locations in the study area. The soils are gradational and comprise gravels and coarse- grained sands on ridges and upper slopes, with particle size becoming finer downslope, in drainage basins and along creeklines. Poorly drained flats, supporting moorland and scrub, also have relatively fine soils which are high in organic content. At a broad level, the vegetation in the general area varies in response to landform, and consequently moisture availability, drainage and fire history. Forests with rainforest understoreys are dominated by Acacia melanoxylon (and occasionally Eucalyptus obliqua), and are confined to humid creeklines and gullies. E. obliqua wet sclerophyll forests are common on shaded slopes. Dry sclerophyll forests dominated by E. obliqua or E. amyg- dalina are the most widespread vegetation types in the area. E. ovata woodlands occur on poorly drained sites, grading into shrublands and Gymnoschoenus sphaerocephalus (button- grass) moorlands as drainage becomes progressively more impeded. The study area is entirely contained in State forest, and some of its taller forests have been logged selectively in the past. Some disturbances, including construction of Old Chum Dam and systems of water races, are associated with tin mining prior to 1950. Vegetation in the general area is now being managed for a range of uses, including har- vesting of sawlogs and pulpwood. The study area includes a logging coupe (Gladstone 07), which was logged (using partial logging techniques) in late 1990. Ongoing studies in the coupe, and a comparable control area to the east of Old Chum Creek, will allow moni- toring of the impact of forest practices, including establishment of wildlife habitat strips. Some analyses of the fauna of the coupe and control area have been published (Taylor and Turner; 1992; Taylor et al., 1993; Cale, 1994; Walsh et al., 1994). METHODS Information on the vegetation and physical environment was collected from 66 sites of different dimensions. Fifty-three of the sites coincided with plots established in an area of about 350ha, in the course of the wildlife studies in the coupe and the control area. These plots were chosen to sample the range of forest types in the coupe, and comprised twenty-eight 50m x 50m plots, and ten plots of 10m radius centred on light traps used for insect collection. Two plant associations could be delineated in seventeen 50m x 50m plots, and the vegetation of both associations was sampled separately in these plots. An additional thirteen sites were sampled to include predominantly non-forest vegetation in the analysis, thereby providing a more complete picture of the vegetation of the general area. Eleven of the latter sites were located within the wildlife study area, and two were located some 5km to the north. The additional sites had a nominal 30m radius, except along creeklines, and were confined to relatively homogeneous vegetation. Floristic information obtained for all sites comprised lists of vascular plant species, and a Braun-Blanquet rating of their cover and abundance (Mueller-Dombois and Ellenberg, 1974). Height and cover of vegetation strata were recorded and qualitative cover ratings were given for bryophytes, ground litter, surface rock, bare ground, and logs on the ground. The number of stags (dead trees) standing on each site was also noted. Physical site information recorded included landform, aspect, slope, drainage and obvious past land use and fire history. The fire history of the area was interpreted in- directly, mainly from attributes of forest structure, distribution of eucalypt regeneration, characteristics of epicormic shoots on eucalypts, extent of charcoal and fire damage to trees (including ‘roll arounds’ on fire-scarred trees), and node counts on Banksia Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 47 marginata individuals (Brown and Podger, 1982; Podger et al., 1988). Sub-surface soil samples (5-15cm depth) were collected from most sites, for further evaluation of the relationship between vegetation and soil physical and chemical attributes. The floristic data were transferred to DECODA files (Minchin, 1991). The poly- thetic divisive program TWINSPAN (Hill, 1979) was used to classify sites on the basis of species cover/abundance data. Another classification based on species presence/ absence data gave similar results, but was marginally less informative. The sorted matrix of species and sites allows the composition of the vegetation group defined at each level of division to be perceived. The data were ordinated using hybrid multi-dimensional scaling (Minchin, 1987). This technique often gives a more realistic display of the trend in vegetation compositional change, and the relative disposition of sites, than other ordination techniques (Minchin, 1987). Soil samples from 31 sites were analysed for moisture loss, pH, organic content and texture (sand, silt and clay), using techniques described in Herbert et al. (1994). The sam- ples were chosen to include sites from the range of classifactory groups, with considera- tion also being given to the position of the sites on the ordination. These data were fitted to the ordination diagrams by vector fitting, using DECODA to explore the relationships between the trends in floristic variation and the edaphic factors. RESULTS AND DISCUSSION A total of 171 vascular species were recorded from the sites surveyed; they com- prised 28 pteridophytes, 48 monocotyledonous angiosperms and 94 dicotyledonous angiosperms. Only three species are Tasmanian endemics, the paucity of endemics according with trends described by Kirkpatrick and Brown (1984). No species of national or regional conservation significance were observed. Three exotic species were recorded. A full list of vascular species is available from the authors. The TWINSPAN analysis, using modified cover/abundance data, resulted in twelve interpretable plant groups (communities) being identified. The classification revealed a trend in floristic composition from tall Acacia melanoxylon and Eucalyptus obliqua forests with mesomorphic understoreys, to scrub and moorlands with diverse low shrub and ground strata of scleromorphic species. The TWINSPAN table showed that some species (e.g. Gleichenia microphylla, Xanthorrhoea australis, Atherosperma moschatum) have a high fidelity to particular classifactory groups, while others (e.g. Pteridium esculentum, E. obliqua, E. amygdalina, Gonocarpus teucrioides) occur in most of the groups delineated. Table 1 sum- marises the composition, structure, habitat and extent of each community. More detailed information on the communities is given in Appendix 1. Plotting of site scores on the primary and secondary axes of the ordination (Figure 1) suggested that several factors, some interrelated, were responsible for the distribution of native vegetation in the area. A distinct trend related to moisture availability, insolation and soil fertility is evident in the orientation of the environmental variables, when these are superimposed on the ordination (Figure 2). Most of the measured variables are aligned in much the same direc- tion, indicating a degree of autocorrelation. There is also a trend, running more or less orthogonally to the major alignment, which is related to site drainage. Acacia melanoxylon gully forest (Group A) occupies the most ‘favourable’ sites, such as well defined gullies and creekline corridors. These are relatively shaded, humid, pro- tected from fire, and have moderately high soil moisture contents, and higher propor- tions of silt and clay, than sites supporting other forest groups. The gully forests grade into tall E. obliquawet sclerophyll woodland and forest (Groups B and C) on sites which are less humid but are shaded and burnt infrequently (typically at intervals greater than 30 years). At the other extreme, heathy E. amygdalina dry sclerophyll forest (Group F), sedgey Proc. LINN. SOC. N.S.W., 115, 1995 LOWLAND FOREST IN NORTHEAST TASMANIA 48 Ha i peaidsapim ag we [e007] fer (sead Og — GT) Sous [q UeY) aseuIeIp popodunt JIOW JALY 1O I[HAIJ $S9] IIe YOTYM Says (sea 0¢ — GT) aseureip poepodunt yim saseyvos pure surseg (sxead (0g — 08) sadoys 19ddn papeys ‘sadoyjs 19Mo] pure a[ppru papeys Apseg (sieak (0G — 0€) sadojs 19MO] pure a[pprut papeys (sieak 001 — 08) souIp{aei9 Suoye pure sadojs 1aMo] papeys (seat QQT<) sv weons pue ‘sotyjn3 pure saurpyaoi9 Suoye s1OplI07) (Aouanbaayz astyq) yerqey peotdA 7, 1q dno18 0} septs :AdIOWsIOpuUy) pDuyppstun “7 Soot], (oy kydosrm pruayngy ‘wnjugnasa unipung “dds unuyr¢ ) sua} DIpesods pure (sepuvis Diuyvy ) sploUTUIeIS I9AO (DIDIIIIBIAA DIDIV “dds unusadsojdaT ‘vsosuonbs vanavjayy ) sqniys wMNIpaut asuaq sAd1O}SIOpuUl) puypopstup “yz ‘onbygo “yz SILI, (sepuvss piuyvy “dds vusadsoprdaT) spiouturei3 pure (wnjuagnosa wnipuajq) SUI I9A0 (wnundoos wnusadsoiqaT ‘vuuaqrunl vanuaynd ‘Syourutsa) DIDIV ‘DIDI}I9YLIO. DIIDIY ‘DIDAY DUDIAC) ) SQNAYS JOMOT O} UINIPIU Jsuap AJaIeV1OPOJ :‘AdIOISIOPUyL) puyvpstun “gq ‘onbygo “y ‘Sdor], (snywja pusadsoprqaT ) sploulureis pue (viqnp DIIIIN) “UNjUaNISa WNIpPUAIY) SULI} 19AO (DUUadiunl papuaynd ‘DIDyWYA. DIDI °3'd) squiys d1yYd1OUIOII[IS PUR (sUaISaLOgLD DUALT ‘nvjadn stuuapowmog °8°'3) sqniys s1yd10WOsauI WINIpPIUT asuaq TAdIOWSIOpUL)] (sypurwen xy) ‘onbygo “sq SIXT, (sipupid piuyDy ‘snaypja DusadsoprqaT ) sploulwess pure (wnzuanosa wnipuarg ‘Dignp DnIINT) “dds wnuyrg ) SUIOJ PUNOIS 19A0 (SyDAysnD DaYIDK) “DAvgLvg Vapo J, ‘DIr3ILDIJUD DIUOSYIICT) SUF payxun.n yim (aaoge se) sqnays styd1ourosauE [[e) 0) WINIPIW asuaq] *AdIO}SIOPUL) (sypumua “q) ‘onbygo “7 Sda1], “uOWIUIOD sayAydoArg pur SUIdy ondydidy * (sn2n7a pusadsoprdeT) sprourmes8 pure (wnuafipord wnyousiog “dds wnuyrag ) sustay puNOIS 19A0 DINIIDJUD DiUosyIGT pur (DyKydosiw DUDAC ‘ppDjaqD Stusapouog) sqn.iys s1yd1OWOsaUt [[e} OF WNIPaUT asuaq *AdIOWSIOpUy) (wnjoyssou Duwadsosayry ‘onbygo *s7) ‘Uopkxounjau “y ISdO1], IIN{INAYS Pue saioads INS AVIV YL) “DALD APNJS AY) U2 S9INGUIID 429Y] PUD SYNOAS I2)8240} [ VIaV I, pur[poom Aydouayos Arp puyppstup snadQqpany Aqqnids :7q purlpoom pAydo.aypos Arp puyppstun sniq{jpang — pnbyqo sniqqvany Aqqnios :[,q sa.105 |[Aydosayos durep vuyppatun snjdajypang — pnbygo snyqQquangy iq ISIIOF TAydo.ayzos 19M [eI pnbygo snydqvony 2) purypoom TAydo.aypos 19M [LI pnbygo snidqvangy *g JSF ANS UOMXOUD[IUL DIIDIW OW dnoiy Proc. LINN. SOC. N.S.W., 115, 1995 49 Bl fee ai F. DUNCAN AND M.J. BROWN Wi tha: peaidsapiy po es (sread g -¢) saseyeos pur sie pouresp Aj100g (saead G] - 8) sa8eyeos pur syepy poureip Aj100g (savak G{ —8) sadvyvos pur syepy poureip Aj100d jo sursirepy (steak 0% - OT) syey Apues (savak Qg<) sa3pu peoig (s1ead 03 - OT) souljaspu pure sodojs 19ddn pue a[pprur payejosuy (Aouanbayy a1nq) yeyqey peotdA y, (syedouf piUosiajDg ‘snurnu pusipoduy ‘xvuay snduvsojiqaT ‘snyoydasosanyds snuaoyssouuts) sploutureis asuap 19A0 (vsouzdnury susvdy ‘vipusvoue Dyaduardy ‘dds wnusadsoiqaT “dds vonaqnjayy) sqnays s1yd.10Ur10.12[9s MO] IsUap a}e1BPO| ‘ADIOWSIOpUy) (DID0.0 “*J) ‘SIOTJ, 1 dno8 0} rep1wIg :‘AdIO}SIOpUyy] DIDAO “iy SILI, (snurmu pusr~poquy ‘xnuay snduvo0jqaT ‘sipuvad piuypy ‘snyvydasosavyds snuaoyrsouuty ) sploururer3 asuap 19A0 ( (psowdnun) susvdy “dds vruosog “dds wnwsadsoidaT “dds nonapjay) sqniys s1ydrowo.1a]9s UNIpPaUt 0} MO] IsUIG *AdI1OSIOpU) (DjD0N0 “y) ‘Duyvpskup “SILI, (wnjquajnosa wnipuajg ) SUId} pure (sydouf Diuosiajog ‘unavoUos DusadsoprqaT “dds nruypy ) sprouturess asuap Ajaye1apout 19A0 (‘dds nny ‘ondiquy vazuny ‘sapionida SNJOY ‘SYDAISND Da0YLLOYIUDY “‘Dyours.vU vsyuog ‘dds vuwvnsvs07y) sqniys s1yd10WI0.19]9s WINTPaUT 0} MO] asUap AJA1e.19POUL PU as1aAIG ‘ADIOWSIOpUl) Duyvopstup “y ‘SdOr], sqiay pure (wnjuaqnosa wnipuajg) SUI} asreds 19A0 (DIDULSLDU pisyung ‘unundoos unusagsoiqaT ) sqn.iys I1yd10ur0.19[9s asredg *ADIOWSIOpUl) (sypurun. “y ‘vuyopstup 7) ‘stD10}]1] DULLDNSDIO|LY ‘SIL |, sqiay pure (wnjuanosa wnipiuayd ) SUIOJ ISUAP IDAO (SYDUIUIA DIDIY ‘DiU0JIUY DYDWOT ‘Dssacqur susvgy “univdoos wnusaqsoiqaT ) sqn.iys s1ydi0ur10.1a]Is asuap A[a}e19pOUr 0} asredg :ADIOWIOpuUy) pnbygo vy ‘vuyvpstup “y SdOL, aimjomns pure saiads snstiajoRrey) “Dav Kpngs ay} U2 $a]NgGLyID L1aY] PUD SdnoAsd 9145210} ,J p.jUOd | ATAV puryioow snqoydarosanyqs snUa0YyISOUUK I qn.ios /pur[poom MO] 1D)D0.0 sudqvong Ka8pas :[ qn.ids /pue[poom Duyopstup snidQqvong 5a8pas °] pueypoom [[Aydouays Arp puyppstup snadqvang Aueayy 7H JSIIOFJ Pasoyo $777.10992) DULLDNSDIONY 2S) 3sa.105 [[Aydo.ayos Ap pnbyqgo snjqQqvany —puyvpdtun snadQqvang Auyeayy “A Proc. LINN. SOC. N.S.W., 115, 1995 50 LOWLAND FOREST IN NORTHEAST TASMANIA woodland/scrub (Groups I and J) and moorland (Group K) occupy ‘unfavourable’ sites, which receive relatively high amounts of solar radiation, have higher fire frequencies (typically at intervals less than 20 years), and have sandy soils with low moisture contents (Group F) or have impeded drainage (Groups I, J and K). The influence of drainage on floristic composition can be seen by the position on the ordination (Figure 1) of scrubby woodlands (Groups E1 and E2), which have a dense understorey dominated by Melaleuca spp. and Leptospermum spp., and occupy poorly drained soaks and basins associated with minor drainage lines. The soil samples from sites with impeded drainage had higher moisture contents and organic carbon contents than those collected from well drained sites. Fig. 1. Ordination of the vegetation in the Old Chum Dam area. Vector 2 24] 0 1 2 3 Vector 1 A Acacia melanoxylon gully forest B Eucalyptus obliqua tall wet sclerophyll woodland C E. obliqua tall wet sclerophyll forest o x 0 + D E obliqua - E. amygdalina damp sclerophyll forest B E11 Scrubby E. obliqua - E. amygdalina dry sclerophy!l woodland 4 £2 Scrubby E. amygdalina dry sclerophyll woodland A F Heathy E. amygdalina - E. obliqua dry sclerophyll forest O G Allocasuarina littoralis closed forest =» H Heathy E. amygdalina dry sclerophyll woodland + I Sedgey E. amygdalina woodland/scrub o J Sedgey E. ovata low woodland/scrub % = =K Gymnoschoenus sphaerocephalus moorland Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 5] fig 2 here %Organic Cc | % Moisture “7Landform —~%Clay ‘sryophytes Vector 2 Vector 1 Fig. 2. Fit of vectors of some environmental variables associated with the ordinated sites. The role of fire in influencing the structure and composition of Tasmanian lowland forests, scrub and moorland is well documented (e.g. Jackson, 1968; Brown and Podger, 1982; Podger et al., 1988). It is described, for areas in northern Tasmania which are com- parable to the study area, by Brown and Buckney (1983) and Marsden-Smedley and Williams (1993). Their findings accord with local information and field observations on the fire history of the Old Chum Dam area. Characteristic fire-free intervals for each community are indicated in Table 1. The occurrence of Atherosperma moschatum as a secondary tree in Acacia melanoxylon gully forest, and the abundance and diversity of epiphytic species, indicates a period of over 100 years between fires for this vegetation type (Neyland, 1991). A melanoxylon itself germinates prolifically after wildfire or other major disturbance, but is also capable of gap-phase replacement following small-scale endogenous disturbances (Pannell, 1992). However, along period (over 200 years) without fire is likely to result in A. moschatumdom- inating such sites, and A. melanoxylon being represented by sporadic trees and abundant soil-stored seed. A fire-free interval of at least 30 years is surmised for EF. obliqua tall wet scle- rophyll forests and woodlands, with the intervals on more humid sites, characterised by the presence of trunked ferns, epiphytic ferns and young A. moschatum, probably approaching that of A. melanoxylon gully forest. Fires are more frequent in heathy dry sclerophyll forests and woodlands on better drained sites, but are generally less intense than those resulting in conflagrations in wet forest types. The relatively low densities and diversities of understorey shrubs on some dry sclerophyll sites are likely to have resulted from high fire frequencies, which prevent species from reaching reproductive age, exhaust their ability to regenerate from coppice or fail to stimulate germination of soil-stored seed (Purdie, 1977b; Gill, 1981). Such fire regimes favour vegetative increasers (Purdie, 1977a), such as Pteridiwm esculentum which Proc. LINN. SOC. N.S.W., 115, 1995 52 LOWLAND FOREST IN NORTHEAST TASMANIA regimes favour vegetative increasers (Purdie, 1977a), such as Pteridium esculentum which forms a dense lower stratum on many sites in the study area. Scrubby woodlands, occupy- ing drainage basins and soaks, appear to have escaped burning for at least 20 years, despite supporting dense understoreys of flammable myrtaceous shrubs. It is possible that moist soil and litter conditions may have inhibited the spread, into scrubby woodlands, of cool fires burning in adjacent heathy forests. Copses of Allocasuarina littoralis closed forest (Group G) occur locally on broad ridges which have escaped fire for at least 30 years. These stands are mainly surrounded by heathy forests with open understoreys, which tend to carry ground fires of low intensities rather than crown fires. The cool fires are unable to penetrate far into the A. littoralis forests, because of the relatively non-flammable foliage and litter of the dominants (Dickinson and Kirkpatrick, 1985), and the sparse nature of the understorey under the dense canopy. A similar situation has been described for Allocasuarina verticillata closed forest, which is associated with drought-susceptible, fire-shadow sites on dolerite in dry areas of the State (Harris and Kirkpatrick, 1991; Fensham, 1992). Flammable myrtaceous species, epacrids and graminoids dominate low woodlands, scrub and moorlands on sites with impeded drainage. In warmer areas of the State, such as the Northeast, rates of fuel accumulation of up to 3 tonnes/ha/year in Gymnoschoenus sphaerocephalus moorlands (Marsden-Smedley and Williams, 1993; Marsden-Smedley, 1994) encourage a fire regime which, coupled with seasonal waterlogging on these sites, maintains the dominance of G. sphaerocephalus and other graminoids, and reduces the abundance and regenerative potential of shrub species (Kirkpatrick and Wells, 1987). Fires are less frequent in low woodlands and scrub, and the abundance of shrub species, and the longer interval between fires, ensures their replacement from rootstock or seed following fire in this vegetation type. In nutrient-poor areas of western Tasmania, the rela- tive densities of woody species and graminoids are correlated with fire history, with fires in rapid succession resulting in ‘ecological drift’ from woodland or scrub towards moorland, extending this vegetation type far beyond its edaphic limits (Jackson, 1968; Brown and Podger, 1982). This situation is atypical in the Northeast, with moorlands being very local and strongly associated with poorly drained sites on relatively infertile soils derived from granite. The relationship between landform and distribution of plant communities near Old Chum Dam is shown in Figure 3a. The general responses of the vegetation to mois- ture availability, drainage and fire are indicated in Figure 3b. The gross trends observed in the Old Chum Dam area are consistent with those found or postulated for several other forested areas occurring at lower altitudes in north- ern and eastern Tasmania (e.g. Hogg and Kirkpatrick, 1974; Brown and Bayly-Stark, 1977; Kirkpatrick and Nunez, 1980; Harris and Kirkpatrick, 1982; Brown and Buckney, 1983; Duncan, 1983; Kirkpatrick and Wells, 1987). Similar trends have been reported for com- parable environments on the southeastern Australian mainland (e.g. Forbes et al., 1982; Keith and Sanders, 1990). Most of the hinterland forests of northeastern Tasmania have a long history of dis- turbance, and contain few plant species which are either endemic or have a priority for conservation. Consequently, they have not received the research or media focus afforded more charismatic forest environments in Tasmania. The detailed patterns of variation in the vegetation, and their implications for conservation of biodiversity, remain poorly known. Further research is warranted, as obligate habitats for localised and/or relictual biota (Mesibov, 1990; Taylor and Turner, 1992) may be present amongst the heteroge- neous native vegetation of this part of the State. Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 53 D B A C F D E1&2 Fa G bride ot ballet twas - --- ® transition as site susceptibility to drought increases — — ® transition as site susceptibility to waterlogging increases ——> effect of recent or frequent fire Fig. 3. Relationships between vegetation and environment in the Old Chum Dam area. (a) Representation of changes in the vegetation across the landscape (approx. vertical exaggeration 3:1), indicating structural differences between communities. Inset shows location of the study area. (b) Response of plant communities to changes in three major environmental variables. Plant groups: A Acacia melanoxylon gully forest; B Eucalyptus obliqua tall wet sclerophyll woodland; C E. obhqua tall wet sclerophyll forest; D FE. obliqua—E. amygdalina damp sclerophyll forest; El Scrubby EF. obliqua—E. amygdalina dry sclerophyll woodland; E2 Scrubby EF. amygdalina dry sclerophyll woodland; F Heathy E. amygdalina — E. obliqua dry sclerophyll forest; G Allocasuarina littoralis closed forest; H Heathy FE. amygdalina dry sclerophyll woodland; I Sedgey E. amygdalina woodland/scrub; J Sedgey E. ovata low woodland/scrub; K Gymnoschoenus sphaerocephalus moorland. Proc. LINN. SOC. N.S.W., 115, 1995 54 LOWLAND FOREST IN NORTHEAST TASMANIA ACKNOWLEDGEMENTS We thank Andrew Herbert, Sheryl Wolfe and Bill Brown for technical assistance, and Robert Taylor for initiating the wildlife studies in the Old Chum Dam area. References Brown, M.J. and Bayty-Stark, H.J., 1979. — Vegetation of Maria Island. Wildlife Division Technical Report 79/1. National Parks and Wildlife Service, Hobart. Brown, M.J. and Popcer, F.D., 1982. — Floristics and fire regimes of a vegetation sequence from sedgeland-heath to rainforest at Bathurst Harbour, Tasmania. Australian Journal of Botany 30: 659-676. Brown, M.J. and BUCKNEY, R.T., 1983. — Structural and floristic variation in the forest communities of the West Tamar, Tasmania. Papers and Proceedings of the Royal Society of Tasmania 117: 135-152. BUCHANAN, A.M., MCGEARY-BROWN, A. and ORCHARD, A.E., 1989. — A census of vascular plants in Tasmania. Tasmanian Herbarium Occasional Publication No. 2. Tasmanian Herbarium, Hobart. CALE, P., 1994. — Temporal changes in the foraging behaviour of insectivorous birds in a sclerophyll forest in Tasmania. Emu 94: 116-126. DAVIES, J.B. and Davies, M., 1989. — Plant communities of the Ben Lomond Plateau. Occasional Publication No. 1. Queen Victoria Museum and Art Gallery, Launceston. DICKINSON, K.J.M. and KirKPATRICK, J.B., 1985. — The flammability and energy content of some important plant species and fuel components in the forests of southeastern Tasmania. Journal of Biogeography 12: 121-134. DUNCAN, F., 1983. — Plant communities of the Douglas-Apsley region. Wildlife Division Technical Report 83/1. National Parks and Wildlife Service, Hobart. DUNCAN, F. and Brown, M,J., 1985. — Dry sclerophyll vegetation in Tasmania. Wildlife Division Technical Report 85/1. National Parks and Wildlife Service, Hobart. ELus, R.C., 1985. — The relationships between eucalypt forest, grassland and rainforest in a highland area in north-eastern Tasmania. Australian Journal of Ecology 10: 297-314. FENSHAM, R.J., 1992. — The management implications of fine fuel dynamics in bushlands surrounding Hobart, Tasmania. Journal of Environmental Management 36: 301-320. ForBES, S.J., WALSH, N.G. and GULLAN, P.K., 1982. — Vegetation of East Gippsland. Meulleria 5: 53-113. GENTILLI, J., 1972. — Australian Climatic Patterns. Nelson, Melbourne. Git, A.M., 1981. — Adaptive responses of Australian vascular plant species to fire. In: Gill, A.M., Groves, R.H. and Noble, I.R. (eds). Fire and the Australian Biota. Australian Academy of Science, Canberra. Harris, S. and KIRKPATRICK, J.B., 1982. — Vegetation of Schouten Island. Papers and Proceedings of the Royal Society of Tasmania 116: 117-141. Harris, S. and KirKPATRICK, J.G., 1991. — The phytosociology and synecology of Tasmanian vegetation with Callitris. In: BANKS, M.R., SMITH, S.J., ORCHARD, A.E. and KANTVILAS, G. (eds). Aspects of Tasmanian Botany —A Tribute to Winifred Curtis. Royal Society of Tasmania, Hobart. HERBERT, A., LAFFAN, M., GRANT, J. and HILL, R. 1994. — Laboratory procedures for soil analysis and preparation of plant materials. Soils Technical Report No. 2. Forestry Tasmania, Hobart. Hit, M.O., 1979. — Twinspan — A Fortran Program for Arranging Multivariate Data in an Ordered Two-Way Table by Classification of Individuals and Attributes. Ecology and Systematics Department, Cornell University, New York. Hocc, A.M. and KirKPATRICK, J.B., 1984. — The phytosociology and synecology of some southern Tasmanian eucalypt forests and woodlands. Journal of Biogeography 1: 227-245. JARMAN, S.J., BROWN, M.J. and KANTVILAS, G., 1984. Rainforests in Tasmania. National Parks and Wildlife Service, Hobart. JARMAN, S.J., KANTVILAS, G. and BRown, M,J., 1988. — Buttongrass Moorland in Tasmania. Research Report 2, Tasmanian Forest Research Council, Hobart. KEITH, D.A. and SANDERS, J.M., 1990. — Vegetation of the Eden Region, south-eastern Australia: species composi- tion, diversity and structure. Journal of Vegetation Science 1: 203-232. KIRKPATRICK, J.B., 1977. — The Disappearing Heath. Tasmanian Conservation Trust, Hobart. KIRKPATRICK, J.B. and BROWN, M,J., 1984. — Numerical analysis of higher plant endemism. Botanical Journal of the Linnean Society (London) 88: 165-183. KIRKPATRICK, J.B. and NUNEZ, M., 1980. — Vegetation-radiation relationships in mountainous terrain: eucalypt- dominated vegetation in the Risdon Hills, Tasmania. Journal of Biogeography 7: 197-208. KIRKPATRICK, J.B. and WELLS, J.M., 1987. — The vegetation of the Great Northern Plain. Papers and Proceedings of the Royal Society of Tasmania 121: 43-49. KIRKPATRICK, J.B., PEACOCK, R.J., CULLEN, P.J. and NEYLAND, M.G., 1988. — The Wet Eucalypt Forests of Tasmania. Tasmanian Conservation Trust, Hobart. MARSDEN-SMEDLEY, J., 1994. — Fuel Characteristics and Fire Behaviour in Tasmanian Buttongrass Moorlands. Parks and Wildlife Service, Hobart. MARSDEN-SMEDLEY, J. and WILLIAMS, K.J., 1993. — Floristics and Fire Management in West Tamar Buttongrass Moorlands. Report to the Tasmanian Forestry Commission, Hobart. MesiBov, R., 1990. — Velvet worms: A special case of invertebrate fauna conservation. Tasforests 2: 53-56. MINCHIN, P.R., 1987. — An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69: 89-108. MINCHIN, P.R., 1991. — DECODA, Database for Ecological Community Data. Version 2.4, A.N.U., Canberra. Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 55 MUELLER-DomBo!s, D. and ELLENBERG, H., 1974. — Aims and Methods of Vegetation Ecology. Wiley and Sons, New York. NEYLAND, M.G., 1991. — Relict Rainforest in Eastern Tasmania. Tasmanian NRCP Technical Report No. 6. Dept. of Parks, Wildlife and Heritage, Hobart and Dept. of the Arts, Sport, the Environment, Tourism and Territories, Canberra. PANNELL, J.R., 1992. — Swamp Forests of Tasmania. Forestry Commission, Hobart. PETERS, D.G., 1984. — Tasforhab. Jn: Myers, K., MARGULES, C.R. and Musto, I. (eds). Survey Methods for Nature Conservation. Vol 2. pp 47-66. CSIRO Division of Water and Land Resources, Canberra. PINKARD, G., 1980. Land Systems of Tasmania, Region 4. Dept. of Agriculture, Hobart. Popcer, F.D., BIRD, T. and BRown, M.J., 1988. Human activity, fire and change in the forest at Hogsback Plain, southern Tasmania. Jn: Frawley, KJ. and Semple, N.M. (eds). Australia’s Ever Changing Forests. Proceedings of the Ist. National Conference on Australian Forest History, Canberra. pp. 119-140. Defence Force Academy Special Publication 1, Canberra. PURDIE, R.W., 1977a. — Early stages in regeneration after burning in dry sclerophyll vegetation. I. Regeneration of the understorey by vegetative means. Australian Journal of Botany 25: 21-34. PuRDIE, R.W., 1977b. — Early stages in regeneration after burning in dry sclerophyll vegetation. II. Regeneration by seed germination. Australian Journal of Botany 25: 35-46. STEPHENS, C.G. and CANE, R.F., 1939. — The soils and general ecology of the north-east coastal region of Tasmania. Papers and Proceedings of the Royal Society of Tasmania 1938: 201-208. TAYLOR, R.J., DUDLEY, A. and CALE, P.G., 1993. — Reptiles and amphibians in sclerophyll forests surrounding Old Chum Dam in north-eastern Tasmania. Herpetofauna 23: 26-31. TAYLOR, R,J. and TURNER, E., 1992. — The spider and the fly. Australian Natural History 24: 12. WALSH, A.M., TAYLOR, R.J., CALE, P.G. and BRERETON, R.N., 1994. — Species composition and seasonal capture rates of terrestrial amphipods in lowland and highland forests in Tasmania. Papers and Proceedings of the Royal Society of Tasmania 128: 57-62. APPENDIX 1. Descriptions of plant communities, Old Chum Dam area GROUP A: Acacia melanoxylon gully forest Acacia melanoxylon gully forest is associated with gullies and creeklines, often form- ing a narrow corridor along these, but in some instances occupying well drained stream flats. Sites are relatively humid and protected from fire by topography and the mesomor- phic nature of the vegetation. The position on the catenary sequence explains the rela- tively high fine particle content of the soils, and consequently their greater water holding capacity and fertility compared with soils on adjacent slopes, which support drier and more flammable forest types. Surface rock cover is low or absent. The community is characterised by sparse Eucalyptus obliqua emergents, exceeding 30m, and a dense secondary tree layer (20-30m) dominated by A. melanoxylon. Atherosperma moschatum is present on more humid sites. Dicksonia antarctica, Olearia argo- phylla, Pomadernis apetala, Coprosma quadnfidaand Bursaria spinosa form a dense medium to tall shrub layer. The liane Parsonsia brownii connects the forest floor to the canopy. The ground layer is dominated by pteridophytes, particularly Blechnum nudum. Blechnum watts, Polystichum proliferum, Culcta dubia, Pteridium esculentum and the sword sedge Lepidosperma elatius are also present on most sites, the latter two fern species increasing in abundance as the community grades into wet sclerophyll forests or woodlands. Epiphytic ferns (Tmesipteris billardien, Polyphlebium venosum, Hymenophyllum spp., Rumohra adiantifor- mus, Ctenopteris heterophylla) are widespread, with abundance and diversity being greatest on sites most approaching rainforest (i.e. with A. moschatum present). The low light levels reaching the forest floor preclude the development of herbaceous species, though bryophytes are conspicuous on logs and on the ground. The group can be allocated to the A. melanoxylon gully forest community identified by Pannell (1992), and also has affinities with E. obliqua — Acacia dealbata — Olearia argo- phylla (OB 0110) wet sclerophyll forest (Kirkpatrick et al. 1988). Proc. LINN. SOC. N.S.W., 115, 1995 56 LOWLAND FOREST IN NORTHEAST TASMANIA GROUP B: Eucalyptus obliqua tall wet sclerophyll woodland Eucalyptus obliqua tall woodland with a wet sclerophyll understorey occurs adjacent to creeks and gullies, and is closely related to Acacia melanoxylon gully forest. The micro- climate appears to be slightly less humid, hence the paucity of Atherosperma moschatumand the absence of a rich epiphytic flora. The sites also tend to have more impeded drainage, resulting in higher soil moisture contents than Group A or Group C sites. As site dryness increases the community grades into E. obliqua tall wet sclerophyll forest (Group C), and as drainage becomes more impeded E. obliqua tall wet sclerophyll woodland grades into Group E]1 vegetation. The community occurs mainly as woodland, the upper stratum exceeding 30m. Eucalyptus obiqua dominates, with E. viminalis present as a minor species on some sites. These species also occur in the medium to tall shrub layer, which also includes Pomaderns apetala (better drained sites) and Melaleuca squarrosa and Acacia verticillata (sites where drainage is more impeded). Trunked ferns are prominent: Dicksonia antarcticaand Todea barbara were present on the five sites samples, and Cyathea australis on two sites. The ground layer is very dense, compared with that of Group A sites, and dominated by ferns. All of Blechnum nudum, Blechnum wattsu, Pteridium esculentum and Culcita dubia occur on most sites. Tall graminoids (Lepidosperma elatius, Gahnia grandis, Carex appressa, Baumea tetragona) are mainly associated with microhabitats having somewhat impeded drainage. Herbaceous species are sparse and low in diversity, and bryophytes are less common than in the Acacia melanoxylon gully forest. The community has a close affinity with E. obliqua — Melaleuca squarrosa — Monotoca glauca (OB 0111) wet sclerophyll forest (Kirkpatrick et al., 1988). GROUP C: Eucalyptus obliqua tall wet sclerophyll forest Eucalyptus obhiqua tall open forest with a wet sclerophyll understorey occurs on south facing middle to lower slopes. Soils are well drained. The community grades into Acacia melanoxylon gully forest or E. obliqua tall wet sclerophyll woodland as sites become more humid, but the boundary between the communities tends to be sharp, reflecting more fre- quent or recent fire in the EF. obliqua forest. The community forms a forest, the upper stratum exceeding 30m, dominated by E. obliqua, with E. viminalisa common subdominant or minor species. The small tree and tall shrub strata are very sparse, with eucalypt regeneration and Acacia melanoxylon the main components. The medium shrub layer (1—5m) is dense, resulting in the community hav- ing a distinctly layered appearance. The main species are Pomaderns apetala, Monotoca glau- ca, Acacia verticillata, Zieria arborescens and Coprosma quadrifida. Scleromorphic shrubs, notably Pultenaea juniperina and Lomatia tinctoria, are also present. A dense to very dense lower stratum is dominated by ferns (Culcita dubia and Pteridium esculentum) which often exceed one metre in height. Graminoids (Lepidosperma elatius, Dianella tasmanica) , grasses (Deyeuxia quadriseta) and forbs (Viola hederacea, Acianthus exsertus) are sporadic, but the shading of the forest floor by the fern fronds is responsible for herbaceous species, in par- ticular, being sparse and depauperate. On more humid sites the community is similar to E. obliqua — Acacia dealbata — Olearia argophylla (OB 011) wet sclerophyll forest, and on drier sites the community has strong affinities with EF. obliqua — Olearia lirata — Pultenaea juniperina (OB 010) wet sclero- phyll forest (Kirkpatrick et al., 1988). GROUP D: Eucalyptus obliqua — Eucalyptus amygdalina damp sclerophyll forest Eucalyptus obiqua —E. amygdalina open forest is a transition community between E. obliqua wet sclerophyll forest and E. amygdalina dry sclerophyll forest. The community is widespread in the study area, mainly occupying slopes with south to east aspects, and therefore not subject to severe insolation or summer drought. Soil moisture levels are Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 57 intermediate between those recorded on wet sclerophyll and dry sclerophyll sites. Soils are well drained, with some sites having a sparse surface rock cover. The community is dominated by E. obliqua and/or E. amygdalina. The dominance reflects the site dryness, and the two species are co-dominant on intermediate sites within the community’s range. Eucalyptus viminalis occurs commonly as a minor species. The canopy is relatively dense and between 20 and 30m in height, with some trees overtopping 30m. The small tree layer is comprised of eucalypt regeneration. The medium to tall shrub layer is very sparse, and mainly comprises Acacia terminalis, Acacia verticillata, Olearia lirata and eucalypt regeneration. Vegetation below one metre is moderately dense, with the relative abundance of scleromorphic shrubs (Pultenaea juniperina, Lomatia tinctoria, Epacris impressa, Leptospermum scoparium) and bracken (Pteridium esculentum) probably reflecting fire history. Culcita dubia is present on moister sites, while graminoids (Lepidosperma spp. Dianella tasmanica, Gahnia grandis) are present but sparse on many sites. Forbs and grasses contribute little biomass but are more abundant and diverse than in wetter forest communities. Species occurring on most sites include Lagenifera stipitata, Goodenia lanata and Chiloglottis reflexa. Bryophytes are virtually absent. The community has strong affinities with shrubby siliceous E. obliqua sclerophyll forest (Duncan and Brown, 1985) and E. obliqua — Olearia lirata — Pultenaea juniperina (OB 010) wet sclerophyll forest (Kirkpatrick et al. 1988). The name damp sclerophyll reflects the position of the community between these two vegetation types. GROUP E1: Scrubby Eucalyptus obiqua — Eucalyptus amygdalina woodland Dense myrtaceous scrub with emergent FE. obliqua and E. amygdalina is strongly associated with basins and soaks with impeded drainage. The community grades into wet sclerophyll woodland (Group B) or damp sclerophyll forest (Group D) as drainage improves. Soil moisture content is relatively high, reflecting the location of the sites on the catenary sequence. The sites sampled have not been burnt for at least 20 years, resulting in the development of the dense shrub understorey. Eucalyptus obiqua and E. amygdalina are co-dominant in most stands, with E. ovata being present on some sites. Trees are considerably sparser, lower in height and poorer in form than those in surrounding forests. A dense to very dense medium to tall shrub layer is dominated by Melaleuca squarrosa, which achieves greatest densities on the most poorly drained sites. Leptospermum scoparium and Acacia verticillata are also prominent in this layer. Acacia terminalis (towards the drier fringes of the community) and Leptospermum lanigerum are occasional. The ground layer is also dense and is dominated by sedges (Gahnia grandis, Lepidosperma elatius) and ferns (Blechnum nudum, Blechnum wattsii, Gleichenia microphylla, Pteridium esculentum). Grasses and forbs are extremely sparse under the dense canopy, though bryophytes are more common than in other communities dominated by E. obliqua. The community has affinities with E. obliqua — Melaleuca squarrosa — Monotoca glauca (OB 0111) wet sclerophyll forest (Kirkpatrick et al., 1988). GROUP E2: Scrubby Eucalyptus amygdalinawoodland Dense myrtaceous scrub with emergent E. amygdalina has strong floristic and struc- tural affinities with Group El. Both communities occur on poorly drained basins, and are characterised by high soil moisture contents. The main difference between the communi- ties is the absence of E. obliqua, and the lower height of the dense shrub stratum (2 —5m, compared with 5 — 10m), in the Group E2 sites. This suggests that Group E2 sites are marginally less fertile or more insolated than Group El] sites. The community grades into damp sclerophyll forest (Group D) or dry sclerophyll forest (Group F) as drainage improves. The emergent E. amygdalina are of spreading, woodland form, and are mainly less Proc. LINN. SOC. N.S.W., 115, 1995 58 LOWLAND FOREST IN NORTHEAST TASMANIA than 15 m in height. A dense medium (2 — 5m) shrub layer is dominated by Melaleuca squarrosa, with Leptospermum scopanum also present on the margins. The ground layer is also dense, and is dominated by ferns (Ptendium esculentum on the margins, Blechnumspp. on moister sites, and Gleichenia spp. forming tangles) and graminoids (Gahnia grandis, Baumea tetragona, Tetraria capillaris). Selaginella uliginosa grows on bare sites on the forest floor. The community has affinities with sedgey E. amygdalina woodland (Duncan and Brown, 1985). GROUP F: Heathy Eucalyptus amygdalina — Eucalyptus obliqua dry sclerophyll forest. Eucalyptus amygdalina — E. obliqua open forest with a heathy understorey is widespread in the study area, mainly occupying well-drained middle and upper slopes subject to moderate drought stress. The community is structurally and floristically similar to Group D and grades into this community as slopes become more humid. The commu- nity grades into heathy E. amygdalina forest (Group H) as sites become more insolated or drought-susceptible, and into scrubby E. obliqua— E. amygdalinawoodland (Group E1) or sedgey E. amygdalinawoodland (Group 1) as drainage becomes more impeded. The community is mainly dominated by E. amygdalina, with E. obliqua occurring as a co-dominant on moister sites and as a subdominant elsewhere. Eucalypt regeneration contributes to the 10 — 20m secondary tree layer cover. The medium shrub layer is very sparse, mainly comprising eucalypt regeneration and occasional Acacia terminalis. The low shrub/ground layer is moderately dense and is mainly dominated by bracken (Pteridium esculentum). Leptospermum scoparium is the main lower shrub species; Lomatia tinctora, Epacnis impressaand Amperea xiphoclada are also present but uncommon on most sites. The sparseness of understorey shrubs, and the relative density of bracken, suggests that this community is burnt frequently. Herbaceous species include Goodenia lanata (present on all sites) and Gonocarpus tetragynus, Eniochilus cucullatus, Chiloglottis reflecaand Xanthosia dis- secta (occasional). Bryophytes are extremely rare. The community has strong affinities with a facies of heathy E. amygdalina forest, described by Duncan and Brown (1985). GROUP G: Allocasuarina littoralis closed forest. Allocasuarina littoralis closed forest occurs locally on broad ridges or flats. Soils have a high sand content and appear to be well drained, though on some sites the community grades into woodland or scrub occupying sites with impeded drainage. Allocasuanna lit- toralis closed forest is more typically surrounded by heathy E. amygdalina — E. obliqua dry sclerophyll forest (Group F). The two communities have many species in common, but are structurally very different, reflecting the absence of fire in Allocasuanrina lhitoralis closed forest for some decades. The community is characterised by a very dense stratum (12-—18m) of Allocasuarina littoralis, overtopped by occasional eucalypts. The density of A. littoralis, coupled with the presence of a deep litter layer, inhibits the development of an understorey, except under gaps in the canopy and at the margins of the community. The main species recorded were Banksia marginata, Leptospermum scoparium, Lomatia tinctoria (shrubs) , Pteridium esculentum (bracken), Lepidosperma laterale, Diplarrena moraea, Gahnia grandis (graminoids) , Ehrharta distichophylla (grass) and Goodenia lanata (forb). The club moss Lycopodium deuterodensum is occasional. The community is not described in the literature but has some affinities with heathy E. amygdalina forest (Duncan and Brown, 19850. GROUP H: Heathy Eucalyptus amygdalina dry sclerophyll woodland Eucalyptus amygdalina woodland with a diverse, heathy understorey occurs locally in Proc. LINN. SOC. N.S.W., 115, 1995 F. DUNCAN AND M.J. BROWN 59 the general area, and more extensively to the north on the naturally vegetated remnants of the hinterland plains. The two sites sampled were about 5 km north of the main study area. Both were relatively flat, and characterized by very sandy soils, with low moisture and organic carbon contents. However, variations in microtopography may be responsible for the presence of many species typical of sites with impeded drainage. Some sites support- ing recently fired examples of this community were much less diverse than the sites sam- pled. The community is dominated by FE. amygdalina, of spreading form and less than 15m in height. Canopy cover varies from open woodland to forest, but is mainly less than 20 per cent. Eucalyptus amygdalina is also conspicuous in the tall shrub/small tree layer, as are Allocasuarina littoralis and Banksia marginata. The 1 — 5m shrub layer also contains these species, but is dominated by Leptospermum scoparium, Kunzea ambiguaand Xanthorrhoea aus- tralis. The low shrub layer and ground layers are particularly diverse, and include Boronia pilosa, Epacris impressa, Aotus ericoides, Allocasuarina monilifera and Hibbertia procumbens (shrubs and under-shrubs); Gahnia grandis, Gahnia radula, Lepidosperma concavum, Leptocarpus tenax, Patersonia fragilis and Hypolaena fastigiata (graminoids); and Pteridium esculentum, Selaginella uliginosa and Lindsaea linearis (pteridophytes). The community is included in heathy E. amygdalina forest (Duncan and Brown, 1985). The community is also analogous to Group 0110, described by Kirkpatrick and Wells (1987) for the Great Northern Plain. GROUP I: Sedgey Eucalyptus amygdalinawoodland/scrub. Sedgey E. amygdalina woodland/scrub is a transition community between heathy E. amygdalina woodland (Group H) and sedgey E. ovata low woodland/scrub (Group J). The community occurs on the margins of drainage basins and flats, on sites with impeded drainage, which are also insolated. Stunted E. amygdalina forms a distinct stratum between 5 and 12m in height. Eucalyptus ovata is occasional. A moderately dense 1 — 2m shrub layer is dominated by Melaleuca squarrosa, Melaleuca squamea and Leptospermum scoparium. Other shrubs typical of poorly drained sites (Boronia pilosa, Boronia parviflora, Epacris lanuginosa) are also present. The ground layer is moderately dense and is dominated by graminoids (Gymnoschoenus sphaerocephalus, Empodisma minus, Gahnia grandis, Leptocarpus tenax, Lepidosperma filiforme, Patersonia fragilis, Restio complanatus). Other species (e.g. Pteridium esculentum, Ehrharta dis- tichophylla, Stylidium graminifolium, Gonocarpus tetragynus) are found on better drained sites. Lindsaea linearis and Selaginella uliginosa are common ground cover species. The community has affinities with sedgey E. amygdalina woodland (Duncan and Brown, 1985) and is similar to Group 0110, described by Kirkpatrick and Wells (1987) for the Great Northern Plain. GROUP J: Sedgey Eucalyptus ovatalow woodland/scrub. Sedgey E. ovata low woodland/scrub occurs on poorly drained flats, at the margins of buttongrass moorland (Group Kk), or as a later successional stage of that community. Soils sampled had a relatively high moisture content. Sparse emergent, E. ovata overtop a dense 1 — 5m shrub layer, dominated by Melaleuca squarrosa, Melaleuca squamea, Leptospermum scoparium and Leptospermum lanigerum. Epacris lanuginosa and Sprengelia incarnata occurred on all sites sampled. The ground layer is also dense and is dominated by graminoids (Gymnoschoenus sphaero- cephalus, Gahnia grandis, Leptocarpus tenax, Patersonia fragilis, Restio complanatus). Pteridiwm esculentum (bracken) occurs on better drained sites. Gleichenia dicarpa forms locally dense thickets. Selaginella uliginosais the most conspicuous of the prostrate herbaceous species. The community has affinities with sedgey E. ovata woodland (Duncan and Brown, 1985), and can also be ascribed to the Common Wet Eastern Heathy facies of Eastern Proc. LINN. SOC. N.S.W., 115, 1995 60 LOWLAND FOREST IN NORTHEAST TASMANIA Moorland (Jarman et al., 1988). It can also be related to group 0100, described by Kirkpatrick and Wells (1987) for the Great Northern Plain. GROUP K: Gymnoschoenus sphaerocephalus moorland. Moorland dominated by buttongrass (Gymnoschoenus sphaerocephalus) occurs on broad flats (or plains), generally towards the centre of these landforms. Sites are poorly drained, and soils sampled had a high organic content, a relatively high moisture content, and a relatively low sand content. One of the moorlands has been burnt 2 to 3 years previously, but Kirkpatrick and Wells (1987) and Jarman et al. (1988) suggest that the distribution of moorlands of this type has an edaphic basis. The presence of several shrubs also common to sedgey E. ovatalow woodland/scrub (including E. ovata) suggests that the moorland could succeed to woodland/scrub in the absence of fire or other disturbance. Gymnoschoenus sphaerocephalus contributes most of the biomass to the community. Other graminoids are also common; they include Leptocarpus tenax, Xyris operculata, Baumea spp., Patersonia fragilis and Empodisma minus. Gleichenia dicarpa forms tangled patches. Shrubs to one metre occur throughout the community and include Melaleuca squarrosa, Melaleuca squamea, Leptospermum lanigerum, Leptospermum scoparium, Epacris lanuginosa and Comesperma retusum. The myrtaceous shrubs, and occasional individuals of E. ovata, are sporadic emergents above the sedgeland, to a height of 5m. Bare ground between clumps of sedges supports small graminoids and herbs, including Selaginella uliginosa, Gonocarpus micranthus, Drosera pygmaeaand Centrolepis fasciculata. The community is consistent with the structure and composition of the Lowland Eastern Sedgey facies of Eastern Moorland (Jarman et al., 1988). It would probably be included in Group 0100 from the Great Northern Plain (Kirkpatrick and Wells, 1987), though that group is fairly diverse, containing heath and scrub as well as sedgeland associations. Proc. LINN. SOC. N.S.W., 115, 1995 How Similar are Geographically Separated Stands of the Same Vegetation Formation? A Moorland Example from Tasmania and Mainland Australia. DAVID KEITH KeITH, D. A. How similar are geographically separated stands of the same vegetation formation? A moorland example from Tasmania and mainland Australia. Proc. Linn. Soc. N.S.W. 115: 61-75 (1995). Broad-scale, intuitively derived vegetation maps and classifications are used for a variety of purposes including evaluation of representativeness and determination of priori- ties in conservation planning. Such uses assume that stands of vegetation ascribed to partic- ular units of classification or map share characteristics (e.g. composition, structure) that differentiate them from stands ascribed to other units. To test this, moorland vegetation was compared at two widely separated locations that have been included within one vegeta- tion unit by several authors. Vegetation at the two locations had similar structure and simi- lar compositional gradients, though some statistically significant differences were demonstrated. There were major differences in floristic composition and richness, overall habitat characteristics and the responses of individual species to a soil gradient. For exam- ple, there were more species in common between moorland and woodland in the same area than between moorlands at different locations. The results highlight the limitations of using broad-scale, intuitively defined units of mapping and classification for conservation planning. Alternative approaches include regional partitioning of units or description of vegetation at finer scales, depending on the nature of heterogeneity within units. D. A. Keith, New South Wales National Parks and Wildlife Service, P.O. Box 1967, Hurstville, Australia 2220; manuscript received 24 April 1994, accepted 21 September 1994. KEYWORDS: Biogeography, classification, conservation planning, gradient response, map units, mire, realised niche. INTRODUCTION In recent decades vegetation has been classified and mapped over large parts of the world’s land surface (e.g. Kuchler, 1964; Sochava and Lukicheva, 1964; Carnahan, 1976; White, 1983). Australian examples that cover broad areas at small scales include the work of Beard and Webb (1974), Specht et al. (1974), Carnahan (1976), Beadle (1981), Kirkpatrick and Dickinson (1983), Baur (1988), Resource Assessment Commission (1992) and Pickard (1994). Units of classification and mapping are defined in terms of physiognomy, structure, floristics of the dominant stratum and/or environmental fea- tures, usually intuitively or according to some pre-determined framework. These maps and classifications are used extensively in resource economics, land- scape geography, comparative ecology and conservation planning. An important assump- tion that underpins such uses is that the units of classification and mapping delineate stands of vegetation which share features in common that distinguish them from stands ascribed to other units. In conservation planning, for example, a vegetation map is expected to identify areas with similar habitat characteristics, structure and species com- position, from which a representative sample may be selected for reservation (Austin and Margules, 1986). How reliably do small-scale maps and classifications fulfil this need? There is insufficient knowledge of the variability within units of classification and map- ping to understand the limitations of these widely used tools. Another aspect of variability within units of classification and mapping concerns Proc. LINN. SOC. N.S.W., 115, 1995 62 SIMILARITY OF SEPARATED MOORLAND VEGETATION genetic variation and the role of species within ecosystems. Genetic and ecosystem com- ponents of biodiversity are poorly understood, relative to species diversity, even though their significance for conservation is now widely accepted (WRI, [UCN and UNEP, 1992). Comparative studies of vegetation can elucidate patterns in these components of bio- diversity indirectly, as illustrated by Hutchinson’s (1959) notion of the realised niche. Niche differentiation between populations of the same species at separate locations may represent variability derived from genetic or ecosystem components of diversity. An understanding of such patterns will contribute to the assessment of maps and classifica- tions for conservation planning. Very few studies have examined properties of map units at local and regional scales (e.g. Burgman, 1988; Pressey and Bedward, 1991). Similarly, few studies have addressed the occurrence of species in relation to interactions between their responses to different environmental factors (e.g. Austin et al. 1983). Quantitative comparisons of vegetation units and their component species between areas separated by large distances are appar- ently non-existent. In this paper, the following characteristics of moorland vegetation were compared at two widely separated locations in south-eastern Australia: (1) overall floristic composition; (2) vegetation structure and species richness; (3) overall character- istics of the habitat (climate, landscape and soils); (4) variation in floristic composition in relation to local environmental gradients; and (5) gradient responses of species common to both locations. METHODS Study Areas Moorland is a very distinctive type of treeless vegetation found in waterlogged soils at low to moderate elevation from south-east Queensland to southern Tasmania. In the Bulli area, south of Sydney (34°14'S 150°54'E, Fig. 1) moorland occurs on a Hawkesbury Sandstone plateau, 300-400 m above sea level. Vegetation of the area was described by Davis (1941), Keith and Myerscough (1993) and Keith (1994a). In the Melaleuca area (also known as New Harbour district) in south-west Tasmania (43° 26'S 146°09'E, Fig. 1), moorland occurs on a peneplain and surrounding quartzite hills, from sea level to well over 400 m. Vegetation was described by Davis (1940) and Keith and Pellow (1989). A sim- ilar area of moorland was sampled at each location (3400ha at Bulli, 3100ha at Melaleuca). Similarities between these moorlands were first described by Davis (1940, 1941). Recognising a conspicuous dominant at both locations, Davis described each moorland as a Gymnoschoenus sphaerocephalus community. This view was subsequently taken up by Specht et al. (1974) who listed alliances dominated by G. sphaerocephalus for the central coast of N.S.W. and south-west Tasmania, and by Beadle (1981) who considered two inter- grading alliances, one dominated by G. sphaerocephalus and the other dominated by Calorophus minor (=Empodisma minus) and Leptocarpus tenax. Carnahan (1976) mapped the area around Melaleuca as gG3, a tussock grassland dominated by Cyperaceae, and it is likely that he would have classified the moorlands at Bulli similarly, but these cover areas too small to map at 1:6,000,000 scale. While each of these authors clearly recognized geographic variation within their respective units of classification, they emphasised similarities in structure, floristics and environment over the broad distributional range of the units. Describing the Bulli moor- lands Davis (1941) wrote, “The community is exactly similar to the extensive Button-grass Plains of Tasmania [Melaleuca]... The structure is identical, and the most prominent species (G. sphaerocephalus) is common to both. Many of the subsidiary species are com- mon to both expressions of the community, though the Tasmanian development is, as Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 63 + 145°E 150°E s fo wee 2 ) + 30°S 30°S+ tT 35°S 35°S+ Melbourne ZX ] 40°S () FE] Areas with mooreland vegetation , % pe SS vat 40°S+ 0 200 400 nl 5 Kilometres Melaleuca 145°E 150°E 155°E t + =f = Fig. 1. Map of south-eastern Australia showing location of Bulli and Melaleuca study areas and the distribution of moorland vegetation. Proc. LINN. SOC. N.S.W., 115, 1995 64 SIMILARITY OF SEPARATED MOORLAND VEGETATION would be expected, richer florisitically, than the present example [Bulli] being extra- limital.” Data collection Vascular plant taxa were scored as present or absent in contiguous 0.5 x 0.5 m quadrats along 30 m transects. The abundance of each taxon at each transect was expressed as the proportion of quadrats occupied. The location of the 60 transects recorded at Bulli was stratified according to 7 classes based on soil drainage and vegeta- tion structure (Keith and Myerscough, 1993). Due to time constraints, only 30 transects were recorded at Melaleuca, their location stratified by classes based on soil drainage and topography (Keith and Pellow, 1990). The height and cover of shrub and herbaceous strata were estimated at 5-metre intervals along each transect and mean values were calculated for each transect. Ten soil cores, 2 cm diameter x 7 cm depth, were sampled at regular intervals along transects and homogenised. Soils were analysed for pH, organic matter content (loss on ignition), exchangeable Na, K, Ca, Mg and Al, and total (acid soluble) P. Mean monthly minimum and maximum temperature data and mean monthly rain- fall data were obtained from the Australian Bureau of Meteorology for stations in each study area (Melaleuca and Maddens Plains). No temperature data were available within the Bulli study area, so data were obtained for a station at Lucas Heights (150 m eleva- tion), 22 km to the north. DATA ANALYSIS Overall floristic composition Similarity in overall floristic composition was examined by simple tabulation and ordination. Taxa were assigned to one of the following groups: represented only at Bulli; only at Melaleuca; or common to both localities. The proportion of moorland taxa at Bulli also represented in moorland at Melaleuca was compared with the proportion represent- ed in a nearby woodland community at Bulli (Keith 1994a) using the z test (Snedecor and Cochran 1963). An ordination was carried out on the transect data from both locations. An association matrix was calculated from the combined data matrix using the Kulzcynski coefficient (Faith et al., 1987). Configurations were fitted in 2, 3 and 4 dimensions using a global non-metric multidimensional scaling algorithm (Minchin 1990) according to the procedure described by Keith (1994a). Vegetation structure and species richness To provide a framework for comparison of vegetation structure and species rich- ness, separate classifications were performed on floristic data from each location. An unweighted pair-group arithmetic averaging (UPGMA) clustering procedure was applied to association matrices calculated using the Kulzcynski coefficient (Belbin, 1986). Floristic groups were defined in each dendrogram using the procedure described by Keith and Myerscough (1993). Each floristic group was characterised by its topographic position in the landscape: along drainage lines; on lower slopes and seepage zones; or on drier upper slopes. Floristic groups were defined as analogous between the two locations if they occupied the same topographic position. Mean height and cover of shrub and herbaceous strata and mean species richness for 15 m* were compared between analogous floristic groups using t tests (Snedecor and Cochran, 1963). Halhitat characteristics Climatic variables were compared between locations graphically. Landscapes were compared by examination of aerial photographs to determine the proportion of area and Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 65 types of landform occupied by moorland at each location. Soils were compared by tabulat- ing the total ranges of soil variables at each location. Environmental Gradients To examine the relationship between floristic composition and local gradients in soil properties, ordinations were derived for each of the two locations. Ordinations were performed on floristic data using the Kulzcynski coefficient and multidimensional scal- ing, as previously described. Vectors for each soil variable were fitted to floristic ordina- tions using a least squares method and a Monte-Carlo procedure to test the significance of correlations (Minchin 1990). Gradient Responses Gradient responses were compared between locations in a set of 13 species that were abundant at both locations. Five classes of soil phosphorus content, containing roughly equal numbers of samples at each location were defined: <90; 91-139; 140-185; 186-250; and >250 ppm. Differences in species’ response to the phosphorus gradient between the two localities were tested using an analysis of deviance by fitting logit-linear models, assuming a binomial error distribution (McCullagh and Nelder 1983). Models were of the form Lt = By + By.P + Bo.L + Bg (P.L), where pis the proportion of quadrats in a transect occupied by the species, Bo is the binomial error term, with the number of quadrats per transect as its denominator, P is the soil phosphorus class (1-5), L is the location (Bulli or Melaleuca), P.L is the interaction term and 8}, Bo and Bo are Co- efficients for P, Land P.L, respectively. Differences in gradient response were tested using a stepwise modelling procedure. First the null model was fitted, then the full model was added, then the interaction term was eliminated from the full model. The statistical significance of the change in deviance associated with elimination of the interaction term was assessed in relation to the chi-squared distribution (McCullagh and Nelder, 1983). RESULTS Overall floristic composition At the species level, only 12% of the combined moorland flora was common to both regions (Table 1). However, floristic differences diminish at higher taxonomic levels, with 36% of genera and 62% of families held in common. Shared taxa were spread evenly amongst growth forms, except that ferns are proportionately more represented. Taxa not represented in both floras fall into several categories: (i) local endemics (e.g. /sophysis tasmanica at Melaleuca, Pultenaea aristata at Bulli); (11) relatively widespread taxa whose range does not extend to one of the localities (e.g. Leptospermum scoparium at Melaleuca, Banksia robur at Bulli); (iii) taxa represented in both localities, but only in moorland habitats at one (e.g. Banksia marginata and Bauera rubioides in moorland at Melaleuca, but only in woodland at Bulli); and (iv) itinerant taxa from adjacent habitats (e.g. Anodopetalum biglandulosum common in adjacent rainforest at Melaleuca, Acacia terminalis common in adjacent woodland at Bulli). Of 167 vascular taxa present in the Bulli moorlands, 55 (33%) were recorded in a nearby woodland community (Sandstone Woodland of Keith 1994a), compared with 29 (17%) recorded in Melaleuca moorlands, a significantly different proportion (z=3.31, P<0.01). Stress values for ordinations in 2, 3 and 4 dimensions were 0.1238, 0.0883 and 0.0761, respectively. Inspection of scatter plots indicated that the configuration of points on the first two axes was similar in all ordinations. Therefore the results of the 2-dimen- sional ordination were presented. Separation of samples along the first ordination axis shows that major differences in florisitic composition were related to geographic separa- Proc. LINN. SOC. N.S.W., 115, 1995 66 SIMILARITY OF SEPARATED MOORLAND VEGETATION tion of the two locations (Fig. 2). However, in each of the two clusters there is a parallel arrangement of samples along the second axes, suggesting a common environmental trend in floristics. TABLE 1: Moorland floras of Bulli (NSW) and Melaleuca (Tas). Taxon Both Melaleuca Bulli Total Regions Only Only Species 29(12%) 74(31%) 138(57%) 241 Genus 46(36%) 27(21%) 54(43%) 127 Family 33(62%) 7(13%) 13(25%) 53 Vegetation structure and species richness The cluster analyses allowed recognition of 5 floristic groups at Bulli (after Keith and Myerscough 1993): Ti-tree Thicket (ITT); Cyperoid Heath (CH); Sedgeland (SL); Restioid Heath (RH); and Banksia Thicket (BT), and 4 groups at Melaleuca (nomencla- ture follows Jarman et al. 1988): Creek Copse (CC); Layered Blanket Moor (LB); Standard Peat (SP); and Alkaline Pan (AP). Fig. 3 shows floristic relationships among groups at each locality. TT and CC were considered analogous because both were thicket occurring along drainage lines. CH and LB were considered analogous because both were dense heath occurring in seepage zones. SL, RH and SP were considered analogous because all were open sedgeland-heath occurring on drier slopes. BT, a thicket on drier slopes at 3 - tC on a e e o 79 5 aoe Oo Qe e cue? 0,0 a ee eet ] t al ar @ 3 0 Ca oO L ey o e eat a N Es A Pipi Ayh te Gane r= ae) 4 e Oo aaa Oo OF o -1 L Se | ee | Ee Seer RS eh | AE eee oe | eee 7 SS | -1 0 1 2 3 4 Axis 1 Fig. 2. Ordination based on floristic composition of 60 samples from Bulli (open squares) and 30 samples from Melaleuca (closed circles). Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 67 S(15m?) Darkes Forest S(15m?) Melaleuca 24(2) Ti-tree Thicket 14(2) Creek Copse 37(2) | Cyperoid Heath 16(-) Alkaline Pan 40(3) Sedgeland 26(2) Layered Blanket Moor 55(2) — Restioid Heath 33(1) Standard Peat 40(2) | Banksia Thicket Fig. 3. Dendrograms showing relationships between floristic groups based on separate analyses of data from each of Bulli and Melaleuca. S, se and n give the mean species richness per 15 m’, standard error and number of samples for respective floristic groups. Bulli had no analogue at Melaleuca. AP, a heath community on flats subject to occasional tidal flooding at Melaleuca had no analogue at Bulli. Ail floristic groups at Bulli were significantly richer in species at 15 m* scale than their analogues at Melaleuca (Fig. 3, P<0.001). Thickets along drainage lines had a taller and denser shrub stratum at Melaleuca (CC) than at Bulli (TT), and their ground stratum was denser at Bulli, but not different in height (Fig 4). At Bulli, dense heath in seepage zones had slightly less shrub cover and a slightly taller ground stratum than at Melaleuca, but otherwise there were no differences (Fig. 5). The structure of open sedgeland-heath on drier slopes at both locations was also similar. Both strata were slightly taller at Bulli than at Melaleuca, while cover did not differ, except that ground cover was greater in Restioid Heath than in Standard Peat (Fig. 6). Hahitat Characteristics Mean monthly temperatures were 3-7°C greater at Lucas Heights (cf: Bulli) than at Melaleuca (Fig. 7). The greatest differences in temperatures occurred in summer. Mean annual rainfall is almost 700mm greater at Melaleuca than at Maddens Plains (Bulli) (Fig. 8). At Melaleuca, maximum rainfall occurs in winter months, while Maddens Plains expe- riences peak rainfall in summer. At Bulli, moorlands occupied approximately 15% of the landscape, the remainder being occupied by dry sclerophyll woodland and forest. Moorlands at Bulli were restricted to gently sloping headwater valleys, the interfluves and steeper gullies were invariably occupied by woodland and forest respectively. At Melaleuca, moorlands occupied approximately 85% of the landscape, the remainder being occupied by wet sclerophyll forest and rainforest. Moorland at Melaleuca was widespread on flats, steep slopes and summits, while forest was restricted to the most sheltered sites. Site 19 at Melaleuca (Alkaline Pan) was excluded from soil comparisons because tidal influence resulted in extreme values of some soil constituents, particularly exchangeable Na. Soil pH was similar at Melaleuca and Bulli, though soils were slightly less acidic at Bulli (Table 2). Moorland soil contained more organic matter at Melaleuca than at Bulli, though values overlap (Table 2). Levels of total soil phosphorus were similar at the two locations, but varied over a wider range at Melaleuca than at Bulli (Table 2). Exchangeable cations were, overall, more abundant in Melaleuca soils, however differ- ences varied between cations. Levels of exchangeable Ca and Mg were much higher at Melaleuca than Bulli, while exchangeable Na and K were slightly higher. Levels of Proc. LINN. SOC. N.S.W., 115, 1995 68 SIMILARITY OF SEPARATED MOORLAND VEGETATION exchangeable aluminium were much higher at Bulli than at Melaleuca (Table 2). TABLE 2: Variation in soil properties of moorland soils at Bulli (NSW) and Melaleuca (Tas). Exchangeable cations in milliequivalents per 100 g air-dried soil. Organic matter is % air-dry mass. Total phosphorus is ppm acid-soluble phosphorus in air-dried soil. Bulli Melaleuca Soil Property pH 3.4-4.1 3.1-3.8 Exchangeable Na 0.04-1.41 0.36 - 2.47 Exchangeable K 0.10-0.75 0.13- 1.24 Exchangeable Ca 0.10- 1.50 0.10-5.70 Exchangeable Mg 0.10 - 2.20 1.40 - 12.70 Exchangeable Al 0.04 - 6.00 0.04-0.90 Total Exch. Cations 0.38 - 12.46 2.44-21.5] Organic matter 0.2 - 43.6 10.2 - 69.1 Total P 60 - 290 20-350 TABLE 3: Correlations between floristic ordination vectors and soil properties at Bulli (NSW) and Melaleuca (Tas). Bulli Melaleuca Soil Property n R P n R P pH 18 3642 741 29 6153 .028* Exchangeable Na 18 8395 .002** 29 .7218 .002** Exchangeable K 18 .7894 .004** 29 .7048 .006** Exchangeable Ca 18 .6895 .048* 29 .7400 .004** Exchangeable Mg 18 8252 .004** 29 7391 .002** Exchangeable Al 18 9515 .000*** 29 .7615 .000*** Total cations 18 .9198 .000*** 29 -7528 .000*** Organic matter 18 .9140 .000*** 29 5811 .026* Total P 18 9133 .000*** 29 .6023 .026* Environmental Gradients For the Bulli floristic data, stress values for ordinations in 2, 3 and 4 dimensions were 0.1453, 0.1026 and 0.0821, respectively, while for the Melaleuca data they were 0.1229, 0.0845 and 0.0568, respectively. Correlations for vectors fitted in 4 dimensions are shown in Table 3, those for vectors fitted in 3 dimensions were similar. At both locations total exchangeable cations and exchangeable Al were highly correlated with floristic composi- tion and there were strong correlations with exchangeable Na, K and Mg. Correlations between floristic composition and exchangeable Ca and pH were stronger at Melaleuca than at Bulli. Total phosphorus and organic matter were much more highly correlated with floristic composition at Bulli than at Melaleuca. Gradient Responses There was a significant interaction between soil phosphorus and location in the dis- tributional models for 12 out of the 13 species examined (Table 4). Thus, with the excep- tion of Lepidosperma filforme, a species’ response to the soil phosphorus gradient varied with location. The full models accounted for between 27 and 85% of the total deviance in the null model (Table 4), suggesting that factors other than soil phosphorus and location Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 69 influence abundance, at least in some species. There were not sufficient data to examine other factors. TT < CC TT =CC ‘tl TT > CC HEIGHT (m) -_ N ry a a ® T T Tiga T ae} COVER % a > a Q N ocuUchPWCUCOWlCUCUCOmUCUCO T T T T 7) zs 3 Ce : conser Ground STRATUM Fig 4: Mean height (a) and cover (b) with standard errors for each of shrub and ground strata in thicket along drainage lines at Bulli (open bars) and Melaleuca (solid bars). CH = LB ] CH > LB HEIGHT (m) ® als CH < LB Shrub Ground STRATUM COVER % o bp A @ oo oOo 6 T T T Fig 5: Mean height (a) and cover (b) with standard errors for each of shrub and ground strata in dense heath in seepage zones at Bulli (open bars) and Melaleuca (solid bars). Proc. LINN. SOC. N.S.W., 115, 1995 70 SIMILARITY OF SEPARATED MOORLAND VEGETATION 0.9 - 0.8 | SL > SP RH > SP 0.7 - HEIGHT (m) 0.4 - 0.3 80 - 70 + 60 + COVER % a °o T a if] n ae) 40 + 30 + 20 Ground STRATUM Fig 6: Mean height (a) and cover (b) with standard errors for each of shrub and ground strata in open sedgeland- heath on drier slopes at Bulli (open and hatched bars) and Melaleuca (solid bars). 30 - —_ : IL = Yr SS oy NS He ae X Wa we y, Oo 20 Es N We ° Sie : y Ai @ oN, es NN Ye / = : a \ f al \ # eae 7 2 Z a ‘ / Zo D E Rea ise Rak ; . 2 ‘ Ue L % vA = N / 10 ase tl ‘ . SS A See % r Z | les | eo | Se: | | | @: ea ae ET ae l | l l Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH Fig 7: Mean monthly minimum and maximum temperatures for Lucas Heights near Bulli study area (unbroken line) and Melaleuca (broken line). Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 71 300 — 250 + 200 - 150 - 100 - 50 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Rainfall (mm) Month Fig 8: Mean monthly precipitation for Maddens Plains in Bulli study area (open bars, mean annual total 1546 mm) and Melaleuca (solid bars, mean annual total 2212 mm). DISCUSSION Similarities and differences Similarities between the geographically separated moorlands were limited. Of all the characteristics examined, there was greatest convergence in vegetation structure. Although there were a number of statistically significant differences in height and cover of various components, these differences were generally small in magnitude (Figs. 4-6). Some of these structural differences may relate to differences in fire history. Floristic simi- larities were limited, contrary to the remarks of Davis (1941), although some of the small number of shared species were visually conspicuous (e.g. Gymnoschoenus sphaerocephalus, Melaleuca squarrosa). The pattern of shared taxa reflects in part the intermittent isolation of Tasmania from the Australian mainland as a consequence of climatic fluctuation dur- ing the past million years (Barlow, 1981). Neither do the data support Davis’ (1941) con- tention that Melaleuca moorlands are floristically richer than those at Bulli. In fact, the reverse is true, even though Australian moorlands reach their greatest spatial develop- mentin Tasmania (Jarman et al., 1988). There were substantial differences in climate, landscape and some soil characteris- tics between moorlands at the two locations. Nonetheless, ordination of combined data suggests that there may be analogous floristic trends in relation to a local environmental gradient (Fig. 2). Correlations between floristic vectors and certain soil variables at both locations (Table 3) support this interpretation. The apparent commonality of gradients may reflect similarity in processes that influence the distribution and abundance of plant species at each locality. Dynamic soil gradients and recurring fires have both been impli- cated as driving forces in vegetation dynamics of moorlands and heathlands in Tasmania (Jackson, 1968; Brown and Podger, 1982; Bowman et al., 1986) and the central coast of New South Wales (Siddiqi et al. 1976a,b; Keith 1991; Keith 1994b, Keith and Bradstock 1994). Proc. LINN. SOC. N.S.W., 115, 1995 72 SIMILARITY OF SEPARATED MOORLAND VEGETATION While differences in gradient responses between species within regions have previ- ously been reported (e.g. Austin ef al., 1983), the possibility that differences may exist within species between regions has received little attention from community ecologists. Even though similar environmental gradients appear to regulate the composition of vege- tation at Bulli and Melaleuca, the response of any given species to these gradients may not be the same at different locations (Table 4). The statistical analyses are supported by the observation that some other species are represented exclusively in different habitats at each location (e.g. Banksia marginata, Bauera rubioides). There may be several explanations for such phenomena: genetic variability within species that affect their physiological range of tolerance (e.g. Hamerick 1983); physiological interactions within plants such that the level of one resource factor affects utilisation or tolerance of another (Tilman 1982); and ecological interactions between species such that a species’ local distribution and abundance depends on the presence or absence of its competitors and predators (Connell 1975). TABLE 4: Logit-linear models of species occurrence in relation to soil phosphorus (P) and location (L). Right-hand column indicates significance of interaction term (P.L). Change in Deviance P Model: Null +P+L+P.L -P.L Degrees of freedom: 51 9 5 Boronia parviflora 703.3 314.8 70.6 <0.001 Cassytha glabella 806.2 420.0 39.4 <0.001 Drosera binata 506.4 138.6 79.7 <0.001 Empodisma minus 1068.7 296.0 117.7 <0.001 Epacris obtusifolia 842.7 277.3 17.6 <0.01 Gymnoschoenus sphaerocephalus 1057.6 695.4 15.0 <0.01 Lepidosperma filiforme 663.8 561.1 7.9 ns Leptocarpus tenax 1157.0 615.6 100.1 <0.001 Lycopodium laterale 574.0 266.4 47.8 <0.001 Restio complanatus 1020.5 352.1 56.3 <0.001 Selaginella uliginosa 601.0 292.1 52.6 <0.001 Sprengelia incarnata 1334.2 728.9 116.3 <0.001 Xyris operculata 1092.8 339.2 44.6 <0.001 Implications for use of small-scale maps and classifications The high level of variability in moorland vegetation and its environment between, relative to within, the two locations examined in this study highlights the limitations of using intuitively defined, broad-scale classifications and maps in conservation assess- ments. The accuracy of such assessments depends on the extent to which representation of classification units in reserves reflects representation of species and their assemblages, since these are the primary objects of conservation goals. The strength of this relation- ship, in turn, depends on heterogeneity within classification units, which is inversely related to scale (Bedward et al., 1992). The intuitive method by which broad-scale units of classification and map are defined also reflects upon their reliability, although this effect is difficult to quantify because of its subjective nature. The results for geographically separated moorlands show that similarities in vegeta- tion structure and shared occurrences of conspicuous species do not necessarily reflect similarities in other features of vegetation, most notably overall floristic composition. The reliability of units of classification and mapping defined on this basis is therefore limited for certain uses. Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 73 Several studies in Australia have attempted to assess conservation needs and priori- ties using intuitively defined, broad-scale classifications (e.g. Specht et al., 1974; Benson, 1989; Resource Assessment Commission, 1992). The moorland example suggests that such assessments should be used cautiously. A high proportion of the total moorland in Australia could be reserved in the southern part of its range, but many moorland species and gradient patterns would not be represented unless reserves also sampled other parts of moorland distribution. Indeed, the results of the floristic analyses suggest that better representation might be achieved if Bulli moorland and woodland were lumped together and distinguished from Melaleuca moorland, than if the two moorlands were grouped within one unit and distinguished from Bulli woodland. Nonetheless, classifications and maps will remain principal tools for conservation planning. They offer an essential means of simplifying complex spatial patterns in biodi- versity and the conservation of species assemblages is a recognized goal in itself (WRI, IUCN and UNEP, 1992). It is the techniques of classification and mapping (intuitive cf quantitative methods) and the scale of application that require more attention than previ- ously received in conservation planning exercises. The example examined here is extreme because of the large distance between study sites, but it raises a broader question about the nature of heterogeneity in broad-scale classification units. What is the relation- ship between heterogeneity and distance between stands? The extent to which heterogeneity is predictable through spatial autocorrelation remains a crucial issue that requires resolution (Sokal and Oden, 1978). If distance rela- tionships account for much of the heterogeneity in broad-scale classification units, then conservation planning strategies would be more likely to achieve their goals if they incor- porated some form of regional partitioning to ensure that units were represented throughout their distributional range (e.g. Hickey and Brown, 1989; Brown and Hickey, 1990). If heterogeneity within broad-scale classification units is mostly independent of dis- tance, their usefulness in conservation planning may be very limited and efforts would be focussed more productively at finer scales in smaller regions. New statistical methods in spatial autocorrelation offer a means to achieve greater understanding of heterogeneity in ecosystems and its effect on widely used tools for conservation planning (Legendre and Fortin, 1989). ACKNOWLEDGEMENTS Belinda Pellow assisted with fieldwork at Bulli and Melaleuca. George Williams and Denny King shared their local bush knowledge of Bulliand Melaleuca moorlands, respec- tively. Denny King loaned us his dinghy to explore country surrounding Melaleuca Lagoon. The Sydney Water Board and Tasmanian Department of Parks, Wildlife and Heritage gave permission to work on lands under their control. Fieldwork in Tasmania was funded by the Wilderness Ecosystem Baseline Studies program. Michael Bedward and Bob Pressey gave helpful comments on a draft manuscript. References AUSTIN, M.P., CUNNINGHAM, R.B., and Goon, R.B., 1983. — Altitudinal distribution of several eucalypt species in relation to other environmental factors in southern New South Wales. Australian Journal of Ecology 8: 169- 180. AUSTIN, M. P., and MARGULES, C. R., 1986. — Assessing representativeness. Jn USHER, M. B., (ed.), Wildlife Conservation Evaluation, pp45-67. Chapman and Hall, London. BarLow, B. A., 1981. — The Australian flora: its origin and evolution. Flora of Australia 1: 25-75. Baur, G. N., 1988. — Forest types of New South Wales. Research Note No. 17. Forestry Commission of New South Wales, Sydney. BEADLE, N. C. W., 1981. — The vegetation of Australia. Cambridge University Press, Melbourne. BEARD, J.S.,and WEBB, M. J., 1974. — The vegetation survey of Western Australia: its aims, objects and methods. Great Sandy Desert. Part I of explanatory notes to Sheet 2. Vegetation survey of Western Australia, 1:1,000,000 vegetation Proc. LINN. SOC. N.S.W., 115, 1995 74 SIMILARITY OF SEPARATED MOORLAND VEGETATION series. University of Western Australia Press, Nedlands. BEDWARD, M., KEITH, D. A., and PRESSEY, R. L., 1992. — Homogeneity analysis: assessing the utility of classifications and maps of natural resources. Australian Journal of Ecology 17: 133-139. BELBIN, L., 1986. — PATN: Pattern analysis package. CSIRO, Canberra. BENSON, J. S., 1989. — Establishing priorities for the conservation of rare or threatened plants and plant associa- tions in New South Wales. Jn Hicks, M., and EIsER, P., (eds.) The Conservation of Threatened Species and their Habitats, pp17-82. Australian Committee for IUCN, Canberra. Bowman, D. M. J. S., MACLEAN, A. R. and CROWDEN, R. K., 1986. — Vegetation-soil relations in the lowlands of southwest Tasmania. Australian Journal of Ecology 11: 141-153. Brown, M. J., and HICKEY, J. E., 1990. — Tasmanian forest-genes or wilderness? Search 21: 86-87. Brown, M. J. and PopGcER, F. D., 1982. — Floristics and fire regimes of a vegetation sequence from sedgeland-heath to rainforest at Bathurst Harbour, Tasmania. Australian Journal of Botany 30: 659-676. BURGMAN, M. A., 1988. — Spatial analysis of vegetation patterns in southern Western Australia: implications for reserve design. Australian Journal of Ecology 13: 415-429. CARNAHAN, J. A., 1976. — Natural vegetation. Map (1:6,000,000) and commentary. Jn: PLUMB, T. W. (ed.), Atlas of Australian resources. Department of Natural Resources, Canberra. CONNELL, J. H., 1975. — Some mechanisms producing structure in natural communities: a model and evidence from field experiments. In: Copy, M. L., and DIAMOND, J. M. (eds.), Ecology and evolution of communities. Belknap Press, Harvard University Press, Cambridge. Davis, C., 1940. — Preliminary survey of the vegetation near New Harbour, south-west Tasmania. Papers and Proceedings of the Royal Society of Tasmania 40: 1-9. Davis, C., 1941. — Plant ecology of the Bulli district II: plant communities of the plateau and scarp. Proceedings of the Linnean Society of New South Wales 66: 1-19. FalTH, D. P., BELBIN L., and MINCHIN, P. R., 1987. — Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69: 57-66. HaMERICK, J. L., 1983. — The distribution of genetic variation within and between populations. Jn: SCHOENEWALD, C. M., CHAMBERS, S. M., MAcBRYDE, B., and THOMAS, W. L. (eds.), Genetics and conservation. Benjamin/Cummings, London. HIckey, J.E. and Brown, M. J., 1989. — Planning for regional biological conservation of Tasmania’s forest vegetation types. Proceedings of the International Foresters Association Conference, Leura. HUTCHINSON, G. E., 1959. — Homage to Santa Rosalia or why are there so many kinds of animals? American Naturalist 93: 145-149. JACKSON, W. D., 1968. — Fire, air, water and earth — an elemental ecology of Tasmania. Proceedings of the Ecological Society of Australia 3: 9-16. JARMAN, S. J., KANTVILAS, G., and BROwN, M. J., 1988. — Buttongrass moorland in Tasmania. Research Report No.2. Tasmanian Forest Research Council, Hobart. KeITH, D. A., 1991. — Coexistence and species diversity of upland swamp vegetation: The roles of an environmental gradient and recurring fires. PhD thesis, University of Sydney. KeiTH, D. A., 1994a. — Floristics, structure and diversity of natural vegetation in the O’ Hares Creek catchment, south of Sydney. Cunninghamia 3: 543-594. KerTu, D. A., 1994b. — Mosaics in Sydney heathland vegetation: the roles of fire, competition and soils. Calmscience Supplement 4: 199-206. KeITH, D. A., and BRADSTOCK, R. A., 1994. — Fire and competition in Australian heath: a conceptual model and field investigations. Journal of Vegetation Science 5: 347-354. KEITH, D. A.,and MyERscouGcuH, P. J., 1993. — Floristics and soil relations of upland swamp vegetation near Sydney. Australian Journal of Ecology 18: 325-344. KeITH, D. A., and PELLow, B. J., 1989. — Temperate moorland vegetation: a comparative study of Tasmania and the Australian mainland. Unpublished report to the Department of Parks, Wildlife and Heritage, Hobart. KIRKPATRICK, J. B. and DICKINSON, K. J. M., 1983. — Vegetation map of Tasmania, 1:500,000. Forestry Commission of Tasmania, Hobart. KUCHLER, A. W., 1964. — Potential natural vegetation of the Conterminous United States. Map (1:3, 168,000) and manual. Special Publication No. 36, American Geographic Society, New York. LEGENDRE, P., and ForTIN, M., 1989. — Spatial pattern and ecological analysis. Vegetatio 80: 107-138. McCuLLAGH, P., and NELDER, J. A., 1983. — Generalised linear models. Chapman and Hall, London. MiNCHIN, P. R., 1990. — DECODA. ANUTECH, Canberra. PICKARD, J., and Norris, E.H., 1994. — Vegetation of north-western New South Wales. Cunninghamia 3: 423-464 PRESSEY, R. L., and BEDWARD, M., 1991. — Mapping the environment at different scales: benefits and costs for nature conservation. Jn: MARGULES, C. R., and Austin, M. P. (eds.), Nature conservation: cost effective biological surveys and data analysis. CSIRO, Melbourne. RESOURCE ASSESSMENT COMMISSION, 1992. — A survey of Australia’s forest resource. Australian Government Publishing Service, Canberra. Sippigl, M Y., CAROLIN, R. C., and MyerscouGuy, P. J., 1976a. - Studies in the ecology of coastal heath in New South Wales III. Regrowth of vegetation after fire. Proceedings of the Linnean Society of New South Wales 101: 53-63. Sippigl, M Y., CAROLIN, R. C., and MyerscouGu, P. J., 1976b. — Studies in the ecology of coastal heath in New South Wales IV. Seed survival, germination, seedling establishment and early growth in Banksia serratifolia Salisb., B. asplenifolia Salish. and B. ericifolia L.f. in relation to fire: temperature and nutritional effects. Australian Journal of Ecology \: 175-183. SNEDECOR, G. W., and COCHRAN, W. G., 1963. — Statistical methods. lowa State University Press, Ames. SOCHAVA, V. B., and LUKICHEVA A, N., 1964 — Vegetation of the U. S. S. R. map (1:15,000,000) and text. In: Proc. LINN. SOC. N.S.W., 115, 1995 D. KEITH 75 GERASIMOV, I. P. et al. (eds.), Fizako-geograftscheski Atlas Mira, plates 240-1, pp. 280-3. Academy of Sciences of the U.S.S.R., Moscow. SoKAL, R. R., and ODEN, N. L., 1978. — Spatial autocorrelation in biology 2. Some biological implications and four applications of evolutionary and ecological interest. Biological Journal of the Linnean Society 10: 229-249. SPECHT, R. L., RoE, E. M, and BoUGHTON, V. H., 1974. — Conservation of major plant communities in Australia and Papua New Guinea. Australian Journal of Botany. Supplementary Series No. 7. TILMAN, D., 1982. — Resource competition and community structure. Princeton University Press, Princeton. WHITE, F., 1983. — The vegetation of Africa: a descriptive memoir to accompany the UNESCO/AESTAT NUSO vegetation map (1:5,000, 000) of Africa. UNESCO, Paris. WRI, [UCN and UNEP, 1992. Global biodiversity strategy. Guidelines for action to save, study and use earth’s biotic wealth sus- tainably and equitably. World Resources Institute, World Conservation Union and United Nations Environment Program. Proc. LINN. SOC. N.S.W., 115, 1995 cunniee fees ra oti Phe wist roar aly ES, aig neh wre seedy: ute 4 Dep mist af Why oF estonia Lae ta ih esha, X..¥," et a Pe Spi roeviiaeal yepedetiora ifthe | uin terest eae d? dtd Stites; Aap hc moog = eT Priidatad Lian is he te Ve Comey Kae Pr abe Salas ih, Baws Mair’, a rT Shire. Pad Housia, M. eh, vo! pani: wectanels sane aaicaptical analy. Vegwonaier Ss hy, a a i] — ia mot 9° : Lo One i‘ ey a. oe { | yr) and included wider ranges of leaf cohorts (including different age classes, species, and heights) , higher levels of grazing were reported (Coley, 1983; Lowman, 1985; Lowman and Heatwole, 1992). Long-term measurements have also illustrated the high variability of herbivory, both temporally and spatially, within a stand (e.g. Coley, 1983; Lowman, 1985; Brown and Ewel, 1987; 1988). Proc. LINN. SOC. N.S.W., 115, 1995 M.D. LOWMAN 79 In this study, I compare aspects of the spatial and temporal heterogeneity of her- bivory between adjacent rain forest and dry schlerophyll tree crowns within New South Wales, Australia, (but the comparisons between Australia and other continents are still open to debate). I also emphasize the methodological challenges associated with studies of herbivory as a canopy process (because obviously the reliability of methods has an enor- mous impact on the accuracy of the results). Whereas biologists have successfully counted and measured the abundance of herbivorous molluscs on a two-dimensional system such as intertidal rocky shores (e.g., Underwood and Denley, 1984) the height and structural complexity of forest canopies make it more difficult to count and measure grazing impacts there. METHODS At least two representative stands of each of six different types of Australian forest were selected for field work. These were classified as wet forest: cool temperate, warm tem- perate, and subtropical rain forests; and dry forest: healthy sclerophyll forest, rural stands of eucalypts that typify the outback environment, and dieback stands of eucalypts that have recently come to dominate the agricultural landscapes throughout Australia (see Lowman, 1982; and Lowman and Heatwole, 1992 for further site descriptions). All of the forests were situated within 50 km of Armidale, New South Wales, at approximately 30° 20'S. At least five tree species from wet and from dry forests were selected for field measurements. Both common and rare species were selected in each forest type, since the overall aim was to examine herbivory at the community level. For obvious logistic reasons, it was not possible to measure the canopies of all tree species; but in some cases, adjacent trees to the study samples were also measured. Trees that were studied in greatest detail included: rain forests — Ceratopetalum apetalum D. Don (Cunoniaceae), Doryphora sassafras Endl. (Monimiaceae) , Dendrocnide excelsa (Wedd.) Chew (Urticaceae), Nothofagus moorei F. Muell. (Fagaceae), and Toona ciliata (F. Muell) Harms (Meliaceae); and dry forests — Eucalyptus blakelyi Maiden, E. viminalis Labill, FE. melliodora A. Cunn. ex Schauer, E. caliginosa Blakely and McKie, and E. nova-anglica Deane and Maiden. Leaf growth and herbivory was measured monthly for 5 years in wet forests (1979- 1984) and dry forests (1983-1988). Because of the longevity of the evergreen leaves of some species (e.g. D. sassafras leaves lived between 2 — 12+ yrs, Lowman 1992), more than five years of field measurements were made on some of the rain forest trees. Leaf cohorts were marked in the canopy representing different light regimes, heights, species, individ- ual crowns, and sites. In total, over 10,000 leaves were monitored over the duration of their lifespans. Isolated events in the life of a leaf were quantified, including date of emer- gence, length of survival, proportion of leaf-area losses to herbivores, date of senescence, and rate of decay. Only herbivory is reported here, although other information was neces- sary to calculate annual levels of grazing. The extent of replication of leaves within a crown was determined by pilot studies using leaf size to indicate environmentally different regions in the canopy (Lowman, 1985). For example, because C. apetalum leaves varied significantly in size with respect to light levels, canopy heights, individual trees and sites, leaves within each of these cate- gories were monitored. In contrast, D. excelsa had a homogeneous canopy, so all leaves within each tree were pooled as one population. In general, between 200-1000 leaves were measured to calculate herbivory for a species. Further information on the numbers of replicates and regions of crown sampled for both wet and dry forest are reported else- where (Lowman, 1992; Lowman and Heatwole, 1992, respectively). Proc. LINN. SOC. N.S.W., 115, 1995 ‘eIpesny ‘Saye YINOS MON Ul SINUNUIUIOD JsoIOF UTLIJUDIIFFIp NOYSnoiy) (Sassoy vare Fea] enuue Jo uonsodoid v se passaidxa) ATOAIqIaFY “7 “SUT VIIVYLSNV Ni SLSSAYOS NIVY AO LNAIGVED TVNOILVAATS NV DNO1V AYOAISYSH WOIdOYL YO TVOIdOHLENS SALVYeAdWal WHVM ALVYAdNAL 1009 Wa] ‘uy ‘OL TY e9 ‘U'd ‘sq | ‘OL ‘eg ‘UM eS ‘do e9 | ‘s'g e7 | 'S'9 ‘b'9 "UN ‘sO ‘ND “Uw'N ‘sq ‘WN ‘e'¢G HERBIVORY IN AUSTRALIAN FORESTS 80 es[eoxe epludo0Jpueg = ‘a"q eyeljio eUCO] = "9" esoAjeu sAyoejsoyel|y = ‘ug Wejjenw snueEjed = “wD winje]ol|ojt4y UoAPUSposABsy = "yy uinjejode winjeyedoyeseg = "e"9 tAesinw seissfjod = “Wd seajesses esoyd(iog = ‘s'q snpiu wntuedsy = ‘uy wintjojlagoe UopYyoAYyoeIg = “eg eyjosryeuutd sijjes0e19 = ‘d'CO eyerjio BUCO] ="I") euelHeny Palit = "UM uinjeyede winjeyedojyeseg = ‘e"D tweyBujuuns euUsoA}od = "9° seajesses e1oudAiog = 's'q sijesysne euoystAly = ‘e"7 BYOJELUES BWODI||ED = “S*O eBljojupenb ewsoidog ="b*9 Legals EIUNUIND = "S'°O BOUINIA BISUND = "AO seujesses esoydAiog ="s'd jesoow snbejoyjoN = “WN BosJOUBJUe B|UOSYIIGQ = "eG 115, 1995 Proc. LINN. SOC. N.S.W., 81 “eIpeMsny ‘Saye YINOS MAN UI SaNTUNUTUIOD 389.105 []AYdo.rg]Is Arp JUIIIFFIP MOYsno.sy} (sassoy vase Fea] fenuue Jo uonsodoid v se passaidxa) A1oarqaayy ‘Z “Fy S33uL SYNLSVd G3aLv10Sh WHR NV TIGCOOM sooiy Ayyeop - Ajayesapow ‘“payejos} $8011 41S payejOs| $901] esueg Ajeyesopoyy AHLIVAH Movaasid AHLIVSH e-uq SQ} VW O'S ‘pg Sy ‘aq | 1q°3 ‘AF "e-ug ‘e-uq qa "e-uq | AQ i | ‘wg -o'" “AQ ‘a3 ya S) yh nA Nag nn M.D. LOWMAN eoljbue-enou snydAjeong = e-ug epunquoy esoyudobuy = aes eSeunA Se a eo1|Bue-eaou “3 = e-ug nuewnoK *y = AQ eyenjays “= SA SIJBUIWIA "3 = AQ SIJPUIWIA “3 = AQ euee|dwiujep “3 = py e1opoljew “y= wy eyeipes “QJ =19 esouibyeo “3 = 95 eueisebpiig °F = 149 esoulbie9 "5 = 99 tAjayeiq snydAjeong = qa tAjaye1q snydAyeong = qy Ajaye1q snjdAjeong = qy Proc. LINN. SOC. N.S.W., 115, 1995 82 HERBIVORY IN AUSTRALIAN FORESTS Profile diagrams were constructed throughout several forest sites for each forest type, using a hypsometer and standard forestry techniques (see Lowman, 1982). Idealized forest diagrams were constructed from these measurements, and used here to map the herbivory within each forest community. RESULTS Averages of leaf surface area loss of all the leaves were calculated and mapped to illustrate the herbivory for each forest community (Figs. 1, 2). Herbivory in individual canopies ranges from negligible (e.g. < 3% for Toona ciliata) to over 300% of annual foliage production in dry sclerophyll trees where some eucalypts re-foliated three succes- sive times after defoliation (see also Lowman and Heatwole, 1992). Herbivory levels varied significantly both between species and between forest types, with the dieback stands exhibiting the highest grazing levels, some to such extremes that crown mortality was also observed. Herbivory in rain forests was quite different from neighbouring dry sclerophyll canopies. In rain forests, there were greater differences with vertical stratification from top to bottom of the canopy. For example, Ceratopetalum apetalum had 9.4% leaf area grazed in the upper canopy, as compared to 35% in the understory of the warm temperate rain forest, almost a 4fold difference. In contrast, eucalypt trees had more homogeneous herbivory throughout the crown of each individual; but more wide-ranging levels of grazing between species and sites. Some trees in dry sclerophyll woodlands lost as little as 8% leaf area per year (e.g. Eucalyptus blakelyi), whereas E. nova-anglicain rural pastures lost as much as 300% in a given year (i.e. scarab beetles ate the entire crown three times successively, followed by re-leafing). Although the dietary qualities of eucalypt foliage have been studied elsewhere (see Landsburg, 1990; Fox and Morrow, 1983), it is none- theless phenomenal that levels of grazing vary so enormously among neighbouring trees. In order of increasing levels of annual grazing, Australian temperate forests were ranked as follows: healthy dry sclerophyll woodlands (13%), subtropical rain forests (14%), warm temperate rain forests (22%), cool temperate rain forests (27%), healthy stands of sclerophyll trees in rural pastures (35%), and dieback sclerophyll trees in rural pastures (89%). DISCUSSION The measurement of herbivory in evergreen forest canopies may be more compli- cated than predicted before canopy access was a reality, because the cycles of leaf turnover are not always seasonally distinct (e.g., Lowman, 1992). The existence of many cohorts or leaf populations within one crown, requires a more complex sampling design to ascertain both annual defoliation and cumulative herbivory over a leaf’s life span. In Australian evergreen forests where leaf longevity was also extremely variable, the canopy was composed of a complex mosaic of different aged leaves, with different susceptibilities to herbivores. Leaf life spans ranged from as short as 4-6 months (e.g., Dendrocnide excelsa, Urticaceae) (Lowman, 1992) up to 25 years (e.g. Araucana sp., Aracaceae) (Molisch, 1928). The average age of an Australian subtropical rain forest canopy leaf ranged from 2—4 years (sun) to 4-12 years (shade) (Lowman, 1992). Over this ten year period, herbivory was measured using long-term monitoring techniques and repeated visits to measure leaves and their associated phenological changes (see Lowman, 1984b). This long-term sampling yielded grazing levels that were 2 — 3 times higher than those reported in short term studies of other evergreen forests (cf. Leigh and Smythe, 1978). It also revealed an enormous difference in grazing sus- Proc. LINN. SOC. N.S.W., 115, 1995 M.D. LOWMAN 83 ceptibility between different species and within different leaf cohorts on one tree crown. So what do these relatively high levels of insect grazing mean in terms of the dynam- ics of the forest canopy community? First, the variability in levels of grazing are higher than previously assumed, even in adjacent forests. And second, the tolerance of trees to levels of grazing appears much higher than previously thought, and exhibits a plasticity in susceptibility to defoliation that may be very important to subsequent management and regeneration of forest stands. The tolerance of the dry sclerophyll forest canopies to outbreaks is illustrative of their strong response to stress, probably a consequence of many thousands of years of adaptation to physical (as well as biological) limitations. Conversely, in the rain forest, the environmental ‘stresses’ may be more subtle within one crown, such as the changes in microclimate as one progresses from ground to upper canopy through the complex layers of foliage. Comparative studies of insects in these two habitats will provide further in- formation on their trophic structures, especially relative proportions of herbivores (Kitching et al., 1993 and unpublished data). The heterogeneity of defoliation is a consequence of a leaf’s environment and phenology, with different leaf cohorts exhibiting different susceptibilities to grazing (sensu Whittam, 1981). From these long-term studies of herbivory in forest canopies, Iam now able to isolate “hotspots” in the canopy, where grazing will be predictably higher (Fig. 3). These ‘hotspots’ represent foliage with greatest susceptibility to herbivores, such as new leaf flushes, colonizing species that are characterized by soft tissue, lower shade regions of the canopy where insects aggregate to feed in the absence of predators, and canopy regions that attract more insects due to the presence of flowers, epiphytes or vines (e.g., Lowman, 1992; Lowman, Moffett and Rinker, 1993; Lowman, unpublished). These regions are different between rain forest and dry sclerophyll canopies. In the dry forests, where the physical environment throughout the canopy is less stratified, grazing was more homogeneous throughout the canopy of an individual tree, but entire crowns of some species were grazing hotspots (e.g. E. nova-anglica). In contrast, the rain forest canopies exhibit less magnitude of inter-species variation, but obvious grazing preferences within individual crowns (e.g. young leaves in the mid-canopy). For example, Nothofagus moore: had approximately eight cohorts of leaves present within one tree crown at one point in time, each with varying levels of susceptibility to insect attack. Young leaves that emerged during spring (Oct. — Nov.) were the most preferred by common host-specific beetle larvae that emerged synchronously with flush- ing; whereas old leaves (> 1 yr) from summer flushes and from the previous year were highly resistant to grazing. In addition, herbivory varied significantly between branches and individual crowns, but not with light regime or height (Selman and Lowman, 1983). More large-scale comparisons between forest communities are needed to better understand the impact of herbivory as an ecological process. For example, the annual levels of defoliation in Australian tree species ranged from as low as 2-3% in subtropical rain forests to as high as 300% in nearby dry sclerophyll (Eucalyptus) stands (Lowman, 1992; Lowman and Heatwole, 1992). Does this imply that one forest is healthier than another? Are different mechanisms regulating insect defoliators and subsequent foliage responses between two forests? Are the trophic structures of herbivores and predators intrinsically different? The prospect of increased ecological comparisons between and within forests is an incentive to develop better protocols for field sampling of events such as grazing. The pro- cess of herbivory has important consequences in forest ecosystems, both economically in terms of pest management and ecologically in terms of maintenance of species diversity. For example, what species are appropriate to sample? Is there greater variation within or between forests? And how do we tackle these questions with statistical and biological accu- racy? And perhaps most importantly in the current urgency of forest conservation issues, can we apply such community level measurements to improve the management and Proc. LINN. SOC. N.S.W., 115, 1995 HERBIVORY IN AUSTRALIAN FORESTS 84 ‘saroads Suowre sanryenb a8eroj pure A8o0jouayd ut saduaiazIp ay) 0} anp ‘aur 19A0 Area [JIM .s10ds}0}7], soy Sty aq [[IM suUepuUNge JOasSuT J19Y4M pur ‘afqndaosns a10UL SIIeY} I8VI[OJ 0} pa}oe.ne oie saro0arqsoy a19yM ‘Adouvd ay} ut sjodsjoy, Jo uonKUasaidai MNeUIaYIS “¢ SLY AYHOAISHSH 8 = HLMOHD 3181DI1D4N HLMOUD 4V37 HDIH | AYOAISHSH MOT HLMOYD dVa71 ALVYsACOWN AYOAISYSH ALVYACOW HLMOYD AVATSLVYSCGOW Fw AYOAISY 3H HSIH HIMOUYS AVSTMO1 AYOAISYSH HOIH BS CGNAD31 ALIAILOV NVINNH ALISY3ZAIC ALISNAG ALIS sjue|d - S2ABB] - FOV STVNGIAIGNI Saloddads LHSOM LHSISH *SHOLDVSA SY31AW NI JONVLSIG = oF SESS Toa) Oe R>§6. For the non-mycorrhizal treatment, elimination of mycorrhizal inoculum from the soil mix was achieved by steam-heating of soil to 80°C for 30 minutes (Sylvia and Schenck 1984). Heated soil was let stand for two weeks after treat- ment to allow soil microflora to re-establish. For the mycorrhizal treatment, the soil mix was not heated, to allow natural infection of roots from inoculum present in the soil. Pots used in the experiment were sterilised in dilute bleach, rinsed in distilled water, and ele- vated off the bench on inverted pots treated similarly. Six additional replicate pots were sown at the lower density, with no added nutrients (F0-fertility level) , three on pre-heated and three on unheated soil. The full experimental design was two densities x three fertility levels x two mycor- rhizal treatments x four harvests x three replicates sown in randomised blocks containing factorial combinations of all treatments. Sowing occurred on 25 — 26 February 1991. The replicates for the fourth harvest became badly infected by a spray-resistant strain of Botrytis, so data are presented for only the first three harvests (weeks 6, 9 and 12) for the Proc. LINN. SOC. N.S.W., 115, 1995 92 MYCORRHIZAL STATUS AND SELF-THINNING F3-, F2- and F1-fertility levels. The pots from the F0-fertility level were unaffected by the Botrytis infection, and were harvested in week 15. Plants were sprayed as required with Rovral (May and Baker, West Footscray) and Fongarid (Bayer Australia, Sydney) against Fusarium and Botrytis, Kelthane (Hortico (Australia), North Laverton) against red spider; and Foliomat (Bayer) against insect lar- vae. (While Rovral (active ingredient Iprodione) can inhibit mycorrhizal infection in some cases, West et al. (1993) found no effect of this spray in a glasshouse experiment with Vulpia ciliata spp. ambigua. Mycorrhizal infection of basil roots still occurred in the experi- ment reported here, despite its use on five occasions over the 12 weeks). Sampling At harvest a circular quadrat was positioned in the centre of each pot (using PVC pipe, internal diameter either 6.2 cm or 10.3 cm; the smaller quadrat was used for early harvests when densities were high, and the larger quadrat at later harvests as densities fell). Plants with stems rooted in the quadrat were cut off at soil surface. A random sample of ten individuals was selected from the quadrat population and scored for height, leaf number and total area of the laminae using a Lambda Instruments Corporation model LI 3000 (Lincoln, Nebraska). Mean leaf area per plant and total leaf area per quadrat were calculated, and used to estimate Leaf Area Index (LAI) (total leaf area per quadrat/ quadrat area) and Leaf Area Ratio (LAR) (mean leaf area per plant/ mean plant dry weight). As a measure of size variability within populations, the Coefficient of Variation (CV) of plant height for each sample was calculated (standard deviation/mean). The pipe was pushed into the soil and used to extract a soil core. Root material was separated from the substrate by hand, after washing in a 2—mm sieve and subsequent flota- tion. Root length was determined on a root subsample using an Image Analysis system (Skye Instruments, United Kingdom); the relationship between length as given by Image Analysis and known length of cotton and root samples was found to be quadratic (7? = 0.9986), and so actual length of root samples was calculated by solving the equation. Shoot and root material (main root sample and root length sub-sample) were dried in a convective oven at 80°C for 24 h and weighed. Root length per plant was calculated from the total root length per sample (length:weight ratio of the subsample x the weight of the whole root sample) and density. Sampling for mycorrhizal association Soil samples for root extraction were taken from the border region adjacent to the root core sample, for all pots with pre-heated soil and selected pots with unheated soil. Roots were washed free from the soil, stained using the chlorazol black E method (Brundrett et al. 1984) and examined under the light microscope for the presence of vesicular-arbuscular mycorrhizae. Nutnent levels in growing medium The background level of nutrients in the unamended potting mix was examined using the procedure recommended for potting media by Warncke (1980). A500 mL sam- ple of the mix was saturated with distilled water, left to stand for 1.5 hours and filtered through a Buchner funnel under vacuum. Concentrations of selected nutrients in the fil- trate were determined by inductively coupled plasma-optical emission spectrometry (Zarcinas and Cartwright 1983). Results were (mg.L'; mean + S.E.): Ca = 231 + 19; Mg = 120 +8 and P = 2.4+0.3. These concentrations of Ca and Mg are rated optimal for growth by Warncke (1980); the concentration of P is rated low. Data Analysis Analysis of experiment Comparison of treatment and interaction effects on density, shoot and root biomass, LAI, LAR and CV of plant height was made by Analysis of Variance (ANOVA). Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 93 Before analysis the homogeneity of variances in the raw data was checked by Cochran’s test, and transformation used if necessary to achieve homogeneity. Missing values were replaced by cell means, and the degrees of freedom reduced accordingly. The full model (used to analyse biomass, LAI and LAR) included harvest, fertility level, density and myc- orrhizal treatment as fixed factors. Comparison of treatment effects was by planned (orthogonal) comparisons of main effects or main effects in interactions, if interactions were significant (Keppel 1982). Trend analysis was used to analyse fertility effects, as nutri- ents were added at levels that represented equal increases along a (logarithmic) scale (Day and Quinn 1989, Keppel 1982). Density was analysed for fertility level, density and mycorrhizal effects at first harvest, and for fertility and mycorrhizal effects on the Fl- and F2-fertility level populations from the lower density at second harvest. Interpretation of data on size variability from within thinning populations is diffi- cult, because the loss of plants by mortality affects the measure, and is concentrated in the smallest size-classes. Comparison of data on size variability was limited to the lower-sown density while these were still pre-thinning. Heteroscedasicity of variances precluded com- parison of all three fertility levels at first harvest: comparison of the F1l- and F2-fertility levels was made over the first and second harvests. Since there were multiple ANOVAs conducted for the experiment, a sequential Bonferroni correction to significance levels was used to protect against increased risk of Type I error (Rice 1989). Thinning lines — selection of data points Once self-thinning begins, a subset of data points from each experiment must be selected a posteriori to fiteqn. (1) (Mohler, Marks and Sprugel 1978; Westoby 1984; Weller 1987). Inclusion of pre-thinning data points will affect the position of the line, and argu- ments about whether populations have begun to thin or not have been common (Weller 1987, Lonsdale 1990). Pre-thinning populations accumulate biomass with no or little change in density (ie progress vertically up the biomass — density plot). All populations from the higher-sown density showed a substantial decline from sown density (> 15%) at first harvest, and subsequently accumulated biomass while being subject to severe mortal- ity. All populations from the higher-sown density were considered for inclusion in the cal- culation of thinning lines. Populations from the lower-sown density were also less than sown density at first harvest: however accumulation of biomass without substantial mortal- ity was evident in stands from the lower-sown density grown at the F2- and F 1-fertility lev- els, up to second harvest, and in one case third harvest. The variability in density of the pre-thinning populations represents the net effects of sowing, germination and establish- ment on density. The variability in density due to the above-mentioned factors was esti- mated by calculating the 95% confidence limits to a grand mean density for stands in each mycorrhizal treatment from the lower-sown density at first harvest (all fertility levels) plus second harvest (F2- and F1-fertility levels only) using log mean density as the variable. The 95% confidence limits were 9% of the grand mean for mycorrhizal populations and 6% for non-mycorrhizal populations. A decline in mean density of >10% from established density was required before populations from the lower-sown density were considered self-thinning. Once thinning has commenced, data points may still be excluded from line-fitting, if other factors affect either mortality or biomass sufficiently to move the point away from the self-thinning line. A density-independent component of mortality operated in non- mycorrhizal populations at first harvest (see Results). If this occurs without a compensa- tory increase in biomass, data points so affected will be laterally displaced from the thinning line to lower densities. There is evidence of this at the F2- and F1-fertility levels, and so the non-mycorrhizal populations from the higher-sown density at first harvest were excluded from line-fitting (Fig. 2 b,c,e,f). Some data points from third harvest at the F3- fertility level were excluded from line fitting because they showed strong declines in both Proc. LINN. SOC. N.S.W., 115, 1995 94 MYCORRHIZAL STATUS AND SELF-THINNING biomass and density from the previous harvest (Fig. 2(a), (d)); these data points were out- liers from the thinning lines (P< 0.005, see below). Fitting of thinning lines Thinning lines were fitted to selected data points on the log mean biomass (B ) — log mean density (N ) plot, for shoot and root biomass separately. (Results for total biomass closely followed those for shoot biomass, and are not presented). Since both variables were subject to variability, the functional relationship between them was described by the Major Axis of the data (fitted by Principal Components Analysis (PCA), Sokal and Rohlf 1981) following the convention adopted by earlier workers (Mohler, Marks and Sprugel 1978, Westoby 1984; Weller 1987). The r statistic for each line was used to report the strength of the relationship (Weller 1987). Limits to the slopes (Lj ,Lo) were calculated (Sokal and Rohlf 1981). It was difficult to detect whether the presence or absence of mycorrhizae affected self-thinning at each of the three fertility levels separately, because of the loss of one har- vest and the subsequent low number of data points. To make comparisons possible, data points were pooled across fertility treatments, where the thinning lines and data for these treatments were not significantly different in slope or intercept. For shoot biomass, data from the Fl- and F2-fertility levels were pooled, for comparison with the F3-level. Thinning lines for populations in the mycorrhizal and non-mycorrhizal treatments within each fertility level were calculated: (r was not significant for two of these data sets (Fig. 3(c)). However the data sets used for comparison of mycorrhizal effects were subsets of larger data sets with non-zero slopes, and so the calculated slopes (range -0.40 to -0.57) were taken as empirical descriptors of slope for the convenience of comparing treatment effects. No comparison of mycorrhizal effects was attempted for root biomass, because of the more complicated pattern of thinning (Fig. 2) and the low number of data points in some data sets. Root —shoot allometry Allometric relationships of the form log y= b+ mlog x, where m= slope and b= inter- cept, were used to investigate patterns of root — shoot allocation. Biomass allocation was examined via shoot mass - root mass allometry, and relative size of resource-acquiring organs via leaf area — root length allometry. In both cases, as the variables were subject to both variability and correlated errors (density was used to calculate each) Geometric Mean Regression (GMR) was used to describe the functional relationship between the two variables (Rayner 1987); limits (Ly ,L9) to the slope were calculated using the formula of Jolicoeur and Mosimann (1968) reported by Ricker (1984). Lines were fitted initially to data from the individual pots from all harvests in each density x fertility level x mycorrhizal status combination, to give a (maximum) possible 12 lines. Biomass — Canopy Volume relationships Since differences in the biomass contained in given canopy volume can account for some of the differences observed between thinning lines, log mean shoot biomass (B) was plotted against log mean canopy volume (V) (which was estimated as mean plant height (Lonsdale and Watkinson 1983)). Allometric relationships were used to investigate biomass packing in the same populations as were selected to fit thinning lines for shoot biomass. The Major Axis of the data was fitted to describe the functional relationship between Band V (errors were uncorrelated in the variables). Thinning lines were calculated for log V— Log Ndata for comparison with thinning lines calculated in terms of log B— log N, using the same set of data points and the same methods of line fitting. Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 95 Comparison of fitted lines For lines fitted by either PCA or GMR, heterogeneity of slopes was tested by the max- imum likelihood method proposed in Harvey and Mace (1982) (Rayner 1985). To do this, the data were rotated on axes so that the new origin was the bivariate mean of the pooled data, and the new X-axis was the (weighted) pooled slope. The test compares the rs from the rotated data sets: if the slopes are parallel, r approaches zero in all sets. To test for differences in elevation from the common pooled slope, an ANOVA of the residuals after rotation was used (Clutton-Brock and Albon 1980; Harvey et al. 1980). The residuals measure the distance of the data points from the common pooled slope along the minor axis. Rejection of the null hypothesis indicates significant differences in elevation of the data sets (along the minor axis) from the common slope, and that differ- ent functional relationships apply to the data sets being compared. This test is analogous to (but not exactly the same geometrically as) the test for differences in intercepts in Analysis of Covariance (ANCOVA). (Use of ANCOVA to compare elevations in data sets where both variables are subject to variation can lead to an increased risk of Type I error (Huitema 1980)). Lines not differing significantly in slope or intercept were pooled; the probability that differences between lines involved in pooling could have arisen by chance is reported to indicate the strength of differences between such lines. Suspected outliers from both thinning and allometric lines were tested by Grubb’s test (Sokal and Rohlf 1981), using residuals after rotation. RESULTS Mycorrhizal infection No VA mycorrhizae were detected in roots from plants grown on pre-heated soil, at any harvest (Table 2). Mycorrhizal infection was detected in roots of plants grown on unheated soil at 6 weeks, in some pots; widespread infection of roots in all pots with unheated soil was evident by 9 weeks (Table 2). TABLE 2 Results of sampling for vesicular-arbuscular mycorrhizae at each harvest (n = number of pots sampled). Harvest Soil treatment n Number of pots with VA mycorrhizae 6 weeks unheated 6 2 heated 18 0 9 weeks unheated 18 18 heated 18 0 12 weeks heated 18 0 Germination, establishment and pre-thinning At the higher-sown density, plant numbers were less than sown density at first har- vest (52 —83% of sown density in the mycorrhizal treatment and 40 — 57% in the non-myc- orrhizal treatment), and numbers declined between each subsequent harvest (Fig. 1(a), 3). At the lower-sown density, populations were also less than sown density at first harvest (78 — 96% of sown density in the mycorrhizal treatment and 76 — 79% in the non- mycorrhizal treatment, Figs. 1(b), 3). Substantial mortality (> 10% of established, see Proc. LINN. SOC. N.S.W., 115, 1995 96 MYCORRHIZAL STATUS AND SELF-THINNING Methods) by second harvest was only observed in the F3-fertility level stands at the lower- sown density; populations at the F1l- and F2- levels showed little change in density while increasing in biomass between first and second harvests (Figs. 1(b), 3(b,c)). Most of the stands from the lower-sown density treatment at the Fl-and F2-fertility levels did show sub- stantial mortality by third harvest. The populations sown at the lower-density at the F0-fer- tility level and harvested in week 15 showed self-thinning (Fig. 4(b,c) ). Established densities at first harvest were significantly lower for populations in the non-mycorrhizal treatment than those in the mycorrhizal treatment (P< .01). This reduc- tion was density-independent, being apparent at both sowing densities (Figs. 1, 3), and continued to be evident in the pre-thinning populations sown at the lower density at second harvest (P< 0.05). (a) HIGHER DENSITY (b) LOWER DENSITY otB Ss UR2aRee8 PROPORTION OF SOWN DENSITY (%) TIME (WEEKS) Fig. 1: Proportion of O. basilicum plants surviving at each harvest from the (a) higher density and (b) lower density populations in the mycorrhizal (closed symbols, solid lines) or non-mycorrhizal (open symbols, broken line) treatments grown at the F3- (@, ©), F2- (a, A) or F1- (@,<) fertility level. Biomass Shoot biomass was significantly affected by harvest, mycorrhizal treatment (non- mycorrhizal > mycorrhizal) and soil fertility level (Fig. 2(a)). The relationship between shoot biomass and (In) fertility level was best fitted by a quadratic relationship; the F2-fer- tility level yielded significantly more shoot biomass than the F1-, but the additional nutri- ents available in the F3-treatment did not increase biomass further (Fig. 2(a) ). Root biomass was significantly affected by harvest, and soil fertility level in interac- tion with mycorrhizal status (Fig. 2(b)). There was a linear increase in root biomass with soil fertility level in the non-mycorrhizal treatment, but no effect of soil fertility level on root biomass in the mycorrhizal treatment. In the populations grown at the F0-fertility level and harvested in week 15, mycor- rhizal stands had about 1.5 times the shoot biomass of non-mycorrhizal stands; root biomasses were similar in the two treatments (Fig. 4(b,c)). Self-thinning For shoot biomass the populations from the three fertility levels thinned along lines of similar slope (Fig. 3(a-c)); testing for heterogeneity of slopes showed differences were Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 97 non-significant (P> 0.5). The lines for the F1l- and F2-fertility treatments were close, and not significantly different in elevation from acommon slope (P>0.9), so a pooled line was calculated (Fig. 4(a)). The line for the F3-fertility treatment lay c. 0.10 — 0.12 log units below the pooled F1 + F2 line (Fig. 4(a) ); the elevation of these two lines from a common slope was significantly different (one-way ANOVA of residuals, P< 0.001). Pre-thinning populations from the lower-density treatment at the Fl- and F2-fertility levels passed beyond the thinning line for the F3-stands at harvests 2 and 3 (Fig. 3(a-c)). con (a) SHOOT BIOMASS (b) ROOT BIOMASS E 6.4 52 <) -mye n 54 < 63 = S 50 onl a 49 z 6.2 Ss 48 Zz S +#myc 6.1 47 = Fl F2 F3 Fil F2 F3 SOIL FERTILITY LEVEL (In added nutrients (g)) Fig. 2: Biomass (geometric means) plotted against fertility level for (a) shoot biomass and (b) root biomass. Treatment means for each fertility level shown in (a) were averaged over all harvests, densities and mycorrhizal treatments, and in (b) over all harvests and densities for populations in the mycorrhizal (+myc) and non-mycor- rhizal (-myc) treatments. Significant terms only in the ANOVAs of (In) shoot and (In) root biomass are shown as main effects (harvest (H), fertility level (F), mycorrhizal treatment (M)) or interactions (FxM). The relationship between biomass and fertility level as given by trend analysis is shown (in brackets after F or FxM) as not significant (N.S.), linear (1) or quadratic (q). Significance levels: * P < 0.05; ** P < 0.01; *** P < 0.001. Vertical bar gives the Critical Value for comparison between two means (= tg g5, x Standard Error of Comparison). The slopes of the self-thinning lines fitted to root biomass — density data from the F1-, F2-and F3-fertility treatments were significantly different (P< 0.01) (Fig. 2(d-f)), with slopes becoming more negative in the order F2 > F1 > F3. The relative position of the three lines on the root biomass — density plot was complex, with the line for the F3-fertility treatment lying under the F1- line; the F2- line cut across both of the other two lines (Fig. 4(b)). Mycorrhizal treatment had no effect on the elevation of thinning lines from a com- mon slope, either as a main effect or in interaction with fertility level (two-way ANOVA of residuals from four lines in Fig. 4(c)). Non-significance of the interaction (P > 0.25) means that the relative position of the thinning lines for shoot biomass was the same for both mycorrhizal and non-mycorrhizal treatments across the range of fertility levels analysed (Fig. 4(c)), with the F3-line lying under the pooled F1+F2-line irrespective of mycorrhizal status. Non-significance of the main effect (P> 0.25) means that mycorrhizal status did not significantly affect the position of thinning lines for shoot biomass within each fertility level. The single shoot biomass — density data point for populations in the mycorrhizal Proc. LINN. SOC. N.S.W., 115, 1995 MYCORRHIZAL STATUS AND SELF-THINNING 98 ‘10'0>d xx $G0'0 > d x ‘S[PA9] SOURITFIUBIG ‘sayNIsSUap SULMOS MOUS SIXe-K UO SMOLIL ‘(spoyjay_) 8uNIY-suT] WoIy papnypoxa o19M sjutod eyep pepoi19 ‘OZIS JUOJ Ag][euUSs ye UMOYsS are suoneindod Suruuryi-aid wo syutod weq *(w) azis ajdures pur (2) JUIIIYJZIO9 uoNr]a1109 ‘(67‘17) adojs 0} syn] wAIM ‘uMOYs aie (—) sau] Suruuryi-jfas 107 suonenby *(o‘@) AIsuap 19MO] 10 (4g) 19y83ry aye UMOS pur [aagq 41.195 (FD) -[ J 10 (3‘q) -zq ‘(p e) -€4 94} 18 UMOIS s}UaUNeIN (sjoquics uado) yeztyst103Aur -uou 10 (sjoquads pasoyd) yezrysIOIAUI oY) WO wNIyISHg ‘CE Jo SUONLNdoOd Jo (fF - p) sseUIOIG 100.1 pur (9-8) sseurorq 100Ys 103 sdrysuonrjer (AV) Asuap - (gq) sseuorg -¢ Buy (aeos Boy ‘,.wW) ALISNAG NVdAW OL oL Cy | 3 i") s 9 =U 9n6b6'0-=4 81'1-=°7 89°0-=/7 L=™ ««Z06°0-=4 0b8°0-=°7 oF'0-=!7 g=u 9516 0-~4 OL 1-=°7 160 T ae ee Tee N 30[ 06°0 - p0°9 = g 30] N 80] 09'0 - OL'b = g 30] N 30] LU - 6P'L = g 80] & SSVWOIA LOOU La :G) SSVNOId LOO 7a °(a) SSVAWOIG LOOU ca *(P) a < = iS} imo) z < a3) = 9=4 40L8'0-=4 86°:0-=°7 g¢:0-=!7 L= sLp8'0-=4 8L'0-=°7 67:0-=!7 6 N 30] $9°0 - 8b's = g 80] N 30] 16°0 - €6' = @ 30] =U aa8%6'0-=4 LL'0-=°7 Spo-=!7 N 80] 090 - 81'S = g Roy SSVWOIG LOOHS Ta °(9) SSVWOId LOOHS 74 :(9) SSVWOId LOOHS a :(®) 1995 115, Proc. LINN. SOC. N.S.W., E.C. MORRIS (a): SHOOT BIOMASS 99 FI+F2: log B = 5.12 - 0.56 log N, L, = -0.40, L, = -0.74, r = -0.877** n = 13 F3: see Fig. 3(a) MEAN BIOMASS (g.m72, log scale) (c):} SHOOT BIOMASS F3 +myc: log B = 5.10 - 0.57 log N L, = -0.32, L, = -0.89, r = -0.898 N.S. n=4 F3 -myc: log B = 5.27 - 0.62 log N L, = -0.44, L, = -0.85, r = -0.940*. n=5 F1+F2 +myc: log B = 4.94 - 0.52 log N L, = -0.41, Lz = -0.63, r = -0.957* n= 8 F1+F2 -myc: log B = 4.52 - 0.40 log N L, = -0.04, L, = -0.88, r = -0.669 NS. n=5 —— F1+F2 +my 3.0 10 2 B. » Qa Fi+F2 -my ese 25 10 3.5 4.0 45 10 10 10 MEAN DENSITY (m2, log scale) Fig. 4: Comparison of biomass (B) — density (N) relationships across fertility levels for (a) shoot biomass and (b) root biomass of O. basilicum grown the F3- (0), F2- (A), or F1- ( ) fertility levels, and (c) for shoot biomass across fertility and mycorrhizal treatments (closed symbols, solid lines = mycorrhizal (+myc) , open symbols, broken lines =non-mycorrhizal (-myc); F3- (¢,©), pooled F1+F2- (m, 4) ). Only data from thinning populations used to fit lines are shown at F3-, F2- and F1-fertility levels. Single data points for populations grown at the F0-fertility level are shown in (b) and (c) for the mycorrhizal (F0+myc, 9) and non-mycorrhizal (FO-myc, 4) treatments. Selfthinning lines are shown labelled with (a,b) fertility level or (c) fertility level and mycorrhizal treatment. Equations of self- thinning lines not given in Fig. 3 are shown, with limits to slope (1,/9), correlation coefficient (7) and sample size (n). Significance levels: N.S. not significant; * P< 0.05; ** P< 0.01. Proc. LINN. SOC. N.S.W., 115, 1995 100 MYCORRHIZAL STATUS AND SELF-THINNING treatment and grown at the F0-fertility level lay within the general region of data points from the other fertility levels, close to the thinning line for the F3-treatment (Fig. 4(c)); the corresponding data point for populations in the non-mycorrhizal treatment grown at the F0-fertility level lay about 0.15 log units below the thinning line for F3-fertility level populations (Fig. 4(c)). The root biomass - density data points for populations grown at the F0-fertility level lay within the general region of data points from the other fertility lev- els, with root biomass in the non-mycorrhizal treatment being slightly higher than root biomass in the mycorrhizal treatment (Fig. 4(b) ). Leaf Area Stands carried significantly more leaf area as soil fertility increased, the relationship between (In) nutrient level and (In) Leaf Area Index (LAI) being linear (Fig. 5(a)). Stands in the non-mycorrhizal treatment carried significantly more LAI than stands in the mycorrhizal treatment (Fig. 5(a)). Allocation of biomass to leaf area (as measured by the Leaf Area Ratio (LAR)) significantly increased with soil fertility, the relationship between (In) nutrient level and LAR being linear (Fig. 5(b)). (a) ra (b) ave 30 ja, 420 E 2 % a £ 25 es 415 ~ = — < < 3 Z 10 z 20 | aiadiely (il [Sadia | H*** = M* = Fes ees) = TIME (WEEKS) Fig. 5: (a) Mean Leaf Area Index (LAI) and (b) Mean Leaf Area Ratio (LAR) plotted against time for populations of O. basilicum. Symbols used are: (a) populations from mycorrhizal (closed symbols, solid lines) and non-mycor- rhizal (open symbols, broken lines) treatments grown at the F3- (@ , 0), F2- (a, A) or Fl- (@, 3) fertility level; (b) populations grown at the F3- (@) , F2- (a) or F1- (m) fertility level. Densities were pooled within means shown in (a), and densities and mycorrhizal treatments were pooled within means shown in (b). Significant terms only in the ANOVAs of (In) LAI and LAR (raw data) are shown as harvest (H), fertility level (F) , or mycorrhizal status (M). The relationship between (In) LAI or LAR and fertility level as given by trend analysis is shown (in brackets after F) as linear (1). Significance levels: * P < 0.05; ** P< 0.01; *** P< 0.001. Variability in size Comparison of the Coefficient of Variation (CV) of plant height in pre-thinning populations showed that size variability within populations increased significantly with fertility level (Fig. 6(a)). Once thinning commenced, size inequality within populations generally decreased. Root - shoot allometry There was no detectable effect of soil fertility level on the allocation of total biomass to shoot or root biomass (Fig. 7(a)). There was no evidence of differences in slope between shoot mass - root mass relationships (P>0.5), nor of differences in elevation from a common slope (P= 0.25). The overall treatment means of each of the three fertility Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 101 CV OF PLANT HEIGHT (%) TIME (WEEKS) Fig. 6: Coefficient of Variation (CV) of plant height from populations of O. basilicum in the mycorrhizal (closed symbols, solid lines) or non-mycorrhizal (open symbols, broken lines) treatment sown at the lower density and grown at the F3- (0 , @), F2- (a, A ) or Fl- (™@, 2) fertility level. Data from thinning populations are circled. Significant terms only in the ANOVA of CV of plant height (for the F2- and F1-fertility levels at weeks 6 and 9) are shown as fertility level (F). Significance levels: ** P< 0.01. levels for proportion of total biomass as root were 20 — 22%. A single line was fitted to data from all treatments (Fig. 7(a)). In the populations grown at the F0-fertility level, the pro- portion of total biomass as root was 24% in the mycorrhizal treatment and 35% in the non- mycorrhizal treatment (data not shown). (b) Fl: y = 0.46 + 0.81x L) = 0.62, Ly = 1.05, r = 0.830** n = 30 F2:y = 0.58 + 0.77xL, = 0.60, Ly = 0.96, r = 0.909** n = 25 F3: y = 0.54 + 0.94x L, = 0.75, Ly = 1.17, r = 0.814%" n = 34 (a) F1+F2+F3: y = 0.615 + 0.96x L, = 0.89, Ly = 1.03, r = 0.926** n = 105 2.0 10 MEAN SHOOT BIOMASS (mg.plant™, log scale) MEAN LEAF AREA (cm?.plant7!, log scale) 10 0 1.0 10 10 MEAN ROOT BIOMASS (mg.plant"!, log scale) MEAN ROOT LENGTH (m.plant~!, log scale) Fig. 7: Allometric relationships between (a) mean shoot biomass (y) and mean root biomass (x) per plant and (b) mean leaf area (y) and mean root length (x) per plant for populations of O. basilicum grown at the F3- (0), F2- (4) or F1- (4) fertility level. Lines and equations for allometric relationships are shown for (a) pooled F1+F2+F3 line, and (b) for each fertility treatment (F3- —— ; F2---- ;Fl- —--— ). Outliersin (b) are circled. Limits to slope (11,9), correlation coefficient (7) and sample size (n) are given. Significance levels: ** P< 0.01. Proc. LINN. SOC. N.S.W., 115, 1995 102 MYCORRHIZAL STATUS AND SELF-THINNING The leaf area carried per unit root length varied between the fertility treatments, with more leaf area per unit root length as soil fertility level rose. Slopes of the leaf area - root length relationships were homogeneous (P> 0.1); differences in elevation from a common slope of 0.85 were significant (P< 0.005). Trend analysis showed that there was a linear increase in the adjusted mean of the residuals for each fertility level, as fertility level increased. Shoot Biomass - canopy volume Populations grown at the F1l- and F2-fertility levels had more shoot biomass in given canopy volume than those grown at the F3-fertility level (Fig. 8(a)). The shoot biomass - canopy volume relationships for the three fertility levels were linear on a log - log plot, and were not significantly different in slope (P> 0.9). The lines for the F1- and F2-fertility levels were very close, and not significantly different in elevation from a common slope (P= 0.87); the pooled F1+F2 line lay significantly above the line for the F3-fertility level (P< 0.001) (Fig. 8(a)). Canopy volume - density Thinning lines calculated in terms of canopy volume were quite close together, and showed no significant differences in slope (P> 0.1) or in elevation from a common slope (P=0.19) (Fig. 8(b)). (a) F1+F2: log B = 3.53 + 0.89 log V L, = 0.69 Ly = 1.13, r = -0.915** n = 13 F3: log B = 3.34 + 0.82 log V = 0.58 Ly = 1.12, r = -0.8985** n= 9 (b) F1+F2+F3: log V = 2.15 - 0.71 log N L, = -0.62 Ly = -0.80, r = -0.958** n = 22 MEAN SHOOT BIOMASS (g.m"2, log scale) MEAN CANOPY VOLUME (m’, log scale) MEAN CANOPY VOLUME (m%, log scale) MEAN DENSITY (m7, log scale) Fig. 8: Relationships between (a) mean shoot biomass (B) and mean canopy volume (V) and (b) mean canopy volume (V) and mean density (N) for populations of O. basilicwm grown at the F3- (©), F2- (A) or Fl- (QC) fertility level. Lines and equations for (a) allometric relationships between Band Vof the pooled F1+F2- and F3-fertility levels , and (b) pooled V— Nself-thinning line for populations from all fertility levels are shown. Limits to slope (Lj ,L9), correlation coefficient (r) and sample size (m) are given. Significance levels: ** P< 0.01. DISCUSSION Mycorrhizal effects on self-thinning The mycorrhizal status of populations grown over a range of soil fertilities had no detectable effect on the biomass — density relationships of those populations as they self- thinned, in this experiment. The relative position of thinning lines for shoot biomass remained the same across the range of fertility levels used (F1 — F3) , for populations in both the mycorrhizal and non-mycorrhizal treatments (Fig. 4(c)). Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 103 Some important qualifications should be added to this conclusion. Mycorrhizal infection of plants grown on unheated soil was slow to develop, only being evident in all pots from that treatment at second harvest (Table 2). Any possible effects of mycorrhizae on self-thinning may have been restricted to the latter part of the experiment, and may not have had time to become apparent. However, if a strong effect of mycorrhizae on self- thinning had been present in the experiment, it should have become apparent by the end, as between one-third to one-half of the progression of populations along self-thin- ning lines occurred between the second and third harvests (Fig. 3). The range of soil fertilities achieved in the experiment must also be considered. The soil-based potting mix used in this experiment had good levels of some background nutri- ents; testing for levels of nutrients present in the mix gave values in the good — excellent range (Warncke 1980), before the addition of fertilizer. In addition, the growth of plants in the non-mycorrhizal treatment at the F1 — F3-fertility levels was better than that of plants in the mycorrhizal treatment, an effect that has sometimes appeared in other stud- ies (Fitter 1977, West et al. 1993). Whether this was because pre-heating of the soil has had some unknown side effects on growth, such as mobilising nutrients, or there was a net cost to the plant of maintaining the symbiont (Peng e¢ al. 1993) is unknown. Since the pres- ence of mycorrhizae did not confer an advantage in growth in this experiment, repetition of the experiment, at a soil fertility level where mycorrhizal infection did confer such an advantage, may well give a different result. The single data points from the F0-populations, if indicative of the self-thinning paths followed at this lower fertility level, support this view. Mycorrhizal infection con- ferred an advantage in (shoot) growth at the F0-fertility level, and populations from this treatment had greater biomass for roughly comparable thinning density than non-mycor- rhizal populations (Fig. 4(c)). If this represents a treatment effect (rather than random variation), it would mean that at fertility levels lower than those used here, mycorrhizal populations would thin along a higher biomass — density line than non-mycorrhizal populations. The conclusion that mycorrhizal status did not affect self-thinning at the range of soil fertilities used, followed from an analysis of data points selected by the investigator, and consequently some doubt could remain about its validity — other investigators might choose points differently. In particular, the position of the lines for the F2- and F1-fertility levels were determined by one data point at the high-density end, with no further data points for over half the length of the line (although pooling of the shoot biomass data for these two fertility levels overcame this problem somewhat). Exclusion of the first harvest data points for the non-mycorrhizal populations from the higher-density treatment, on the grounds of density-independent mortality, might not be justified (Fig. 3). It is not known what caused density-independent mortality at the establishment phase in this experiment. A density-independent reduction in established plant numbers below sow- ing density has been observed in a second experiment with O. basilicum, where a fungicidal soil drench was used at sowing (Morris, unpublished data). Germination and establish- ment of O. basilicum would appear to be sensitive to soil treatments such as pre-heating or drenching. Inclusion of the first harvest data points for the populations in the non-mycor- rhizal higher-density treatment at the Fl- and F2-fertility levels, while changing the param- eters of the lines, did not alter the conclusion. For shoot biomass, slopes of the three lines were still homogeneous; the lines for the Fl- and F2-fertility levels were still not signifi- cantly different in elevation, and when pooled gave a line of log B= 5.47- 0.64 log N, which was still significantly higher in elevation than the line for the F3-treatment. Alternatively, using the individual biomass - density value for each pot (rather than harvest means) for analysis gives a greater spread of data points along each line, and more power in the com- parison of fertility and mycorrhizal effects. The conclusion drawn from an analysis of indi- vidual pot data was still the same as that drawn from the analysis of harvest mean data reported here: mycorrhizal effects were not apparent, while fertility effects were. Proc. LINN. SOC. N.S.W., 115, 1995 104 MYCORRHIZAL STATUS AND SELF-THINNING Effects of soil fertility on self-thinning - the intensity of competition Previous work has shown that mortality proceeds the fastest in populations grown with the greatest supply of nutrients (Yoda et al. 1963, White and Harper 1970, Bazzaz and Harper 1976, Furnas 1981, Morris and Myerscough 1985, 1991), and this occurred in the experiment reported here. Thus, on a time basis, and using the extent of mortality as the measure, the intensity of competition was greatest in the populations grown at the highest level of soil fertility. On a biomass basis however, the intensity of competition can be judged by the biomass at which thinning commences, the biomass that thinning populations can subse- quently support for a given density, and the ground area (= N*) surviving plants require to support given biomass. In the experiment reported here, populations grown at the high- est-fertility level supported the least shoot biomass for a given thinning density, and the surviving plants from the highest-fertility level required the greatest ground area to sup- port given shoot biomass (Fig. 4(a) ). Thus, competition was most intense (on the basis of shoot biomass), in the populations grown at the highest-fertility level (as argued by Grime 1979). This contrasts with the conclusion drawn by previous workers from experiments of the same design, using a soil-based growing substrate. White and Harper (1970) and Bazzaz and Harper (1976) concluded that the intensity of competition (on a shoot- biomass basis) was the same in all stands over the range of soil fertility used, since all popu- lations thinned along a line of common slope and intercept. The result obtained in the experiment reported here also contrasts with that observed when a sand or perlite potting medium has been used, where competition has been the most intense in populations grown at the lowest levels of nutrient supply (on a shoot-biomass basis, Furnas 1981; ona shoot- and root-biomass basis, Morris and Myerscough 1985, 1991). Thus the effect of soil fertility on self-thinning seen in the experiment reported here has not been observed before. It is of both theoretical and practical interest to know which of a number of populations grown at different soil fertilities will commence thinning at the lowest biomass. A model can be suggested from the data gathered from this and earli- er experiments to explain the differing observations recounted above. The population that begins self-thinning at the lowest biomass presumably does so because total competition is the most intense in that stand. Total competition has above- ground and below-ground components. The rate at which either form of competition intensifies as biomass accumulates depends on the biomass allocated to and architecture of resource-acquiring organs, the physiological activity of those organs (Goldberg 1990), and relative size inequality within the populations. The intensification of shoot or root competition (or an interaction between them) to levels sufficient to induce mortality would thus determine whether stands grown on the most-fertile or the least-fertile sub- strate began thinning at the lowest biomass, or whether stands from all fertility levels thinned at a common biomass. While the intensity of shoot and root competition was not directly measured in this or earlier experiments, an estimation of their relative impor- tance can be made post-hoc by examining the allocation to and the dimensions of the organs of resource capture in the populations. In the experiment of Morris and Myerscough (1991) plants at the two lower levels of nutrient supply grew more root, and less leaf, which was deployed in smaller canopies, than those grown at the highest level of nutrient supply. Morris and Myerscough (1991) argued that below-ground factors were the major determinant of the beginning and course of self-thinning in both of the lower-nutrient supply treatments. In the experiment reported here, plants grown at all levels of soil fertility allocated the same amount of biomass to shoot or root growth (Fig. 7(a)). While conversion of biomass into leaves or roots did differ with soil fertility level, these differences were not correlated with the position of the self-thinning lines. Plants grown at the F1-fertility level required more root length to support given leaf area than those at the F2- or F3- (Fig. Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 105 7(b)). The extra root length required by plants at the F1-fertility level to support given leaf area can lead to intensified root competition, if the extra root is located close enough to neighbour’s roots for depletion zones to overlap, and the total supply of nutrients is insuf- ficient for the growth of all plants. So while the preconditions may have existed for the intensification of root competition as soil fertility level declined in this experiment, any such intensification (if present) was not sufficient to initiate self-thinning at the lowest biomass in the stands grown on the least-fertile substrate. The evidence available suggests that shoot competition became most intense in the populations grown at the F3-fertility level as shoot biomass accumulated. Shoot biomasses achieved at the F2- and F3-fertility levels were similar (Fig. 2(a) ); however, the deployment of this biomass in space (canopy architecture) differed markedly between the two fertility levels. While populations from all three fertility levels occupied the same canopy volume as they self-thinned (Fig. 8(b)), populations at the F3-fertility level required greater canopy volume to support given shoot biomass (Fig. 8(a) ). Thus a given shoot biomass was deployed in the greatest canopy volume in stands grown at the F3-fertility level; this would lead to interference with neighbours at a lower biomass in the F3-stands than in the F1- and F2- stands (Grime 1979). The separation of thinning lines between the F3- and pooled F1 + F2-fertility levels on the shoot biomass density plot (Fig. 4(a)) was accounted for by the separation between the lines for these same treatments on the shoot biomass - canopy volume plot (Fig. 8(a)). This effect has been reported for Helianthus annuus plants grown at different levels of shading (Lonsdale & Watkinson 1983). Shaded populations required greater canopy volume to achieve given shoot biomass, and also thinned along lines of lower intercept on a shoot biomass - density plot. The leaf area carried by populations increased linearly with fertility level, and this would contribute to an intensification of shoot competition, with suppressed plants being the most heavily shaded in populations growing on the most-fertile substrate. However this pattern developed more at second and third harvests (Fig. 6(a) ); LAIs in the F2- and F3-stands were similar at first harvest. LAR, which measures biomass allocation to leaf area, increased linearly with fertility level, and again, this would contribute to an intensifi- cation of shoot competition from the F2- to the F3-stands, which both had similar shoot biomass. But also again, actual LARs achieved in the F3-stands only became clearly differ- entiated from those at the F2-fertility level at second and third harvests (Fig. 5(b)). Size inequality within pre-thinning populations increased with fertility level also, which would mean that suppressed plants were relatively smaller than dominants as soil fertility increased. However, the greatest increase in size inequality in pre-thinning populations was from the F1- to the F2-stands: size inequaltiy in pre-thinning F3-stands was comparable to that in F2-stands (Fig. 7). I would propose that as plants grew in this experiment, root competition was not important in determining the relative position of thinning lines for shoot biomass. However, shoot competition per increment of biomass intensified most quickly in popu- lations grown at the F3-fertility level, and while the amount of leaf area carried in the F3- stands contributed to this, canopy volume occupied per unit shoot biomass was the factor most consistently associated with this intensificaton over the whole of the experiment. Amore general model can be suggested from the results of this experiment. If pop- ulations are grown over a range of soil fertilities, thinning will commence at the lowest biomass in the population in which total competition intensifies the most per increment of biomass. Whether this is the stands grown on the most-fertile substrate (this experi- ment), or the least-fertile substrate (Furnas 1981, Morris and Myerscough 1985, 1991) or all stands commence thinning at the same biomass (White and Harper 1970, Bazzaz and Harper 1976) would depend on the rate at which shoot and root competition intensify per increment of biomass. This model is based on a posteriori correlations, and would require testing by experiment before gaining acceptance. Proc. LINN. SOC. N.S.W., 115, 1995 106 MYCORRHIZAL STATUS AND SELF-THINNING ACKNOWLEDGEMENTS Many thanks to Peter McGee and Anne Ashford for advice on mycorrhizae, to Sally Durham, Nick Skelton and Juliet Thomas for expert technical assistance, and to the NSW Department of Agriculture at Rydalmere for access to their soil steam sterilizer. David King, Peter Myerscough and Mark Lonsdale commented on an earlier draft of this work; an anonymous referee suggested improvements to this draft. The work was supported by the Australian Research Council Small Grant Scheme. References ALLEN, E.B. and ALLEN, M.F. 1990. The Mediation of Competition by Mycorrhizae in Successional and Patchy Environments. Jn: Grace, J.B. and Tilman, D. (eds.) Perspectives on Plant Competition. pp 367-389. Academic Press, New York. Bazzaz, F.A. and HarpER, J.L. 1974. Relationship between plant weight and numbers in mixed populations of Sinapsis alba (L.) Rabenh. and Lepidium sativum L. Journal of Applied Ecology 13: 211-216. BRUNDRETT, M.C., PICHE, Y. and PETERSON, R.L. 1984. A new method for observing the morphology of vesicular- arbuscular mycorrhizae. Canadian Journal of Botany 62: 2128-2134. CHIARIELLO, N., HICKMAN, J.C. and Moonky, H.A. 1982. Endomycorrhizal role for interspecific transfer of phos- phorus in acommunity of annual plants. Science 217: 941-943. CLUTTON-BROCK, T.H., ALBON, S.D. and Harvey, P.H. 1980. Antlers, body size and breeding group size in the Cervidae. Nature 285: 565-567. Day, R.W. and QUINN, G. 1989. Comparisons of treatments after an Analysis of Variance in ecology. Ecological Monographs 59: 433-463. DUNN, C.P. and SHARRITZ, R.R. 1990. The relationship of light and plant geometry to self-thinning of an aquatic annual herb, Murdannia keisak (Commelinaceae). New Phytologist 115: 559-565. Fitter, A.H. 1977. Influence of mycorrhizal infection on competition for phosphorus and potassium by two grasses. New Phytologist 79: 119-125. FURNAS, R.E. 1981. A resource theory of self-thinning in plant populations. Ph.D. thesis, Cornell University, Ithaca. GOLDBERG, D. 1990. Components of Resource Competition in Plant Communities. In: GRACE, J.B. and TILMAN, D. (eds.) Perspectives on Plant Competition. pp 27-49. Academic Press, New York. GriME, J. P. 1979. Plant strategies and vegetation processes. Wiley, New York. HarVey, P.H., CLUTTON-BROCK, T.H. and Mace, G.M. 1980. Brain size and ecology in small mammals and primates. Proceedings of the National Academy of Science, United States of America’ 77: 4387-4389. Harvey, P.H. and Mace, G.M. 1982. Comparisons between taxa and adaptive trends: problems of methodology. Jn: King’s College Sociobiology Group (eds.) Current Problems in Sociobiology. pp. 343-361. Cambridge University Press. Heap, A.J. and NEwMaN, E.I. 1980. The influence of vesicular-arbuscular mycorrhizas on phosphorus transfer between plants. New Phytologist 85: 173-179. HuTcuincs, M.J. and Bupp, C.J.S. 1981. Plant self-thinning and leaf area dynamics in experimental and natural monocultures. Ozkos 36: 319-25. HuitemMa, B.E. 1980. The Analysis of Covariance and Alternatives. John Wiley & Sons, Brisbane. JAKOBSEN, I., ABBOTT, L.K. and Rosson, A.D. 1992. External hyphae of vesicular-arbuscular mycorrhizal fungi asso- ciated with Trifolium subterraneum L. 2. Hyphal transport of *P over defined distances. New Phytologist 120: 509-516. JOLICOEUR, P. and MOSIMANN, J.E. 1968. Intervalles de confiance pour la pente de l’axe majeur d’une distribution normale bidimensionelle. Biometrie-Praximetrie9: 121-140. Kays, S.C. and Harper, J.L. 1974. The regulation of plant and tiller density in a grass sward. Journal of Ecology 62: 97-105. KEPPEL, G. 1982. Design and Analysis. Prentice-Hall, New Jersey. Second edition. LonsDALE, W.M. 1990. The selfthinning rule: dead or alive? Ecology 71: 1373-88. LONSDALE, W.M. and WATKINSON, A.R. 1982. Light and self-thinning. New Phytologist 90: 431-445. . 1983. Plant geometry and self-thinning. Journal of Ecology’71: 285-297. McGEE, P.A. 1990. Survival and growth of seedlings of coachwood (Ceratopetalum apetalum) : effects of shade, mycorrhizas and a companion plant. Australian Journal of Botany 38: 583-592. MOHLER, C.L., Marks, P.L. and SpRUGEL, D.G. 1978. Stand structure and allometry of trees during self-thinning of pure stands. Journal of Ecology 66: 599-614. Morris, E.C. and MyerscouGu, P.J. 1985. Nutrient level effects on thinning and non-thinning crowding effects in even aged populations of subterranean clover. Australian Journal of Ecology 10: 469-479. . 1991. Self-thinning and competition intensity over a gradient of nutrient availability. Jowrnal of Ecology 79: 903-923. PENG, S., EISSENSTAT, D.M., GRAHAM, J.H., WILLIAMS, K. and HopcE N.C. 1993. Growth Depression in Mycorrhizal Citrus at High-Phosphorus Supply. Plant Physiology 101: 1063-1071. RAYNER, J.M.V. 1985. Linear relations in biomechanics: the statistics of scaling functions. Journal of Zoology, London, Series A 206: 415-439. READ, D.J. FRANCIS, R. and FINDLAY, R.D. 1985. Mycorrhizal mycelia and nutrient cycling in plant communities. Jn: Fitter, A.H. (ed): Ecological Interactions in Soil. pp193-217. Blackwell Scientific Publications, Melbourne. Rice, W.R. 1989. Analyzing tables of statistical test. Evolution 43: 223-225. Proc. LINN. SOC. N.S.W., 115, 1995 E.C. MORRIS 107 RICKER, W.E. 1984. Computation and uses of central trend lines. Canadian Journal of Zoology 62: 1897-1905. SHINOZAKI, K. and Kira, T. 1956. Intraspecific Competition among Higher Plants. VII. Logistic Theory of the C-D Effect. Jowrnal of the Institute of Polytechnics, Osaka City University. Series D. 7: 35-72. SOKAL, R.R and ROHLF, F.J. 1981. Biometry. W.H. Freeman and Company, New York. Second Edition. SyLvIA, D.M. and SCHENCK, N.C. 1984. Aerated-steam treatment to eliminate VA mycorrhizal fungi from soil. Soil Biology and Biochemistry 16: 675-76. WARNCEKE, D.D. 1980. Recommended Test Procedure For Greenhouse Growth Media. In: W.C. Dahnke (ed.). Recommended Chemical Soil Test Procedures for the North Central Region. North Dakota Agricultural Experimental Station, North Dakota State University, Fargo, North Dakota. Bulletin 499. WELLER, D.E. 1987. A reevaluation of the -3/2 power rule of plant self-thinning. Ecological Monographs 57:23-43. WEST, H.M., FITTER, A.H. and WATKINSON, A.R. 1993. The influence of three biocides on the fungal associates of Vulpia ciliata spp. ambigua under natural conditions. The Journal of Ecology 81: 345-350. Westopy, M. 1984. The self-thinning rule. Advances in Ecological Research 14: 167-225. Westoby, M. and Howell, J. 1981. Selfthinning : the effect of shading on glasshouse populations of silverbeet (Beta vulgans) . Journal of Ecology 69: 359-365. . 1982. Self-thinning in Trifoltum subterraneum populations transferred between full daylight and shade. Journal of Ecology 70: 615-621. WHITE, J. 1985. The thinning rule and its application to mixtures of plant populations. Jn: J. White (ed.) Studies on Plant Demography. pp. 291 -309. Academic Press, London. and Harper, J.L. 1970. Correlated changes in plant size and number in plant populations. Journal of Ecology 58: 467-485. Yona, K., Kira, T., OGAWA, H. and Hozumt, K. 1963. Self-thinning in overcrowded pure stands under cultivated and natural conditions. Journal of Biology, Osaka City University 14: 107-129. ZARCINAS, B.A. and CARTWRIGHT, B. 1983. Analysis of Soil and Plant Material by Inductively Coupled Plasma- Optical Emission Spectrometry. Division of Soils Technical Paper No. 45 , Commonwealth Scientific Industrial Research Organisation. Proc. LINN. SOC. N.S.W., 115, 1995 fe “opranttaL vestamnenise, PT MiMING bein Salone: HARE dno wnpinae fig \ boda jo cnbeoieriet A A Oni bas ADA apices gee ban me ao sit wae A? Bi Curd : ome i aarp aneee tiny Restle ; veel ta Spb ae : repo) Lh eve Saat . bey yay (aka ac yy t tyael ea eer =e ey 2 eee OP omer Geren, diac Ne, 4 3 r ; ay £ eae Lai ea a3 Ph BAe shi tA Ane ae einai 2 Dee Ay ; it 4. 5 E f uylkast newt pthc ahrriat teeth eer ane inward, Jescomal: Beer ti Le fe =) tk fie Laprrra, A et. ba ee ek + Lowers, WL iw YF peek Sat ubat. i 4 Aly i ‘ J * iy j NR aR, PA LOAN “aaa OA gbons is pa serwy tales ‘fa Tebapibicter i afeeedte Wy (ate PROC ae ar) 08 Th cor we ee oe i 5c 2 a ey a Miva Je. 4.4... hase +) 2) Swe i .?) nr } ured CGE of ire sdniyreetd ae BN. Toe Pee! oof Pccdnggey tid i . | hereets IC ated Wes ay 190% irre) YS Petdcu) Ceeilig eat noe hirer aya are pains Pee | 0 effire valent sc lhi'ea rer aaT 091, frortu J 4 es hex ¢Fae 4 wile VWs). 5 oie lig otk! COYTeSU ten care Yom 4 aah ial Gay et pecatlahele: dionadae wpe Pasi, S.. Bedinaer on, 0,04, Deana Sta Wied efed P ae Qe,. MAS, Corian Dy ead ian Canund mie Bil violin apg un We gph Pale, Paces bal) SWOT ; or Vaynae, £.MW, AS Late oe eeleperiy 0 bd wees Poe the igen! weting Jone Syrjalif Yotignn Vries, 116 ¢ fh, Et i viper A. E.se6 yee ny Lee | ant hues ‘Seell iy BO HALA va “ ‘tig vA plan i Olu : ire AN, Ge Dain Tile Grae te Sool ppp ed SArvitase Piles rion is acd, Lyne teat py Go ne Latico yee Meyda Ch TIS R. © <= ° Foot: Ape ee hee en tt Some Effects of Low-intensity Fires on Populations of Co-occurring Small Trees in the Sydney Region DAviD A. MORRISON (Communicated by D. KEITH) Morrison, D. A. Some effects of low-intensity fires on populations of co-occurring small trees in the Sydney Region. Proc. Linn. Soc. N.S.W. 115: 109-119 (1995). Nine species of small tree were studied one year after low-intensity prescribed fires in 1991 and 1992 at a site in the outer western region of the Sydney metropolitan area. All of the species except Hakea sericea proved to be fire-tolerant, with less than one-third of the plants killed by the fires. All of the fire-tolerant species had smaller stems killed by the fires, the size of surviving stems being related to their fire-tolerance characteristics for most of the species — Leptospermum trinervium and Persoonia linearis (with dormant epicormic and lignotuber buds) survived at relatively small stem sizes, with Casuarina torulosaand Jacksonia scoparia (with dormant buds at the stem base) surviving at larger sizes, and Acacia binervia and Casuanna littoralis (with no dormant buds) surviving only at the largest stem size. Both Acacia implexa (with root suckers) and Acacia parramattensis (with no dormant buds) were exceptions to this generalization. The size structure of the stem populations was significantly different in the burnt areas compared to an adjacent unburntarea for all six species for which there were data. All four of the species that are capable of producing new post-fire shoots at the stem base preferen- tially did so when the upper stem had been killed, and the number of shoots produced was usually unrelated to the size of the stem. Both of the species that are capable of producing new post-fire epicormic shoots almost invariably did so if the stem was alive post-fire, and for both species the number of shoots produced was related to the size of the stem. D. A. Morrison, Department of Applied Biology, University of Technology, Sydney, P.O. Box 123, Broadway, N.S.W., 2007; manuscript received 22 March 1994, accepted for publication 18 May 1994. INTRODUCTION Knowledge of the responses of plant species to fires is of intrinsic interest as well as being essential for the scientific management of plant communities (Gill and Bradstock, 1992). For example, two general types of regeneration strategy by plant populations after fire are usually recognized:— death of all adult plants during the fire followed by regener- ation solely from seeds (fire-sensitive species) ; and regeneration from protected dormant vegetative buds on adult plants that survive the fire (fire-tolerant species) (Gill, 1981). These two strategies result in dramatically different population dynamics, and vegetation management for species conservation may need to incorporate these differences into fire- management plans (Bradstock and Auld, 1987). There is an increasing amount of quantitative data on the post-fire behaviour of both fire-sensitive species (e.g. Bradstock and Myerscough, 1981; Auld, 1987; Bradstock and O’Connell, 1988; Lamont et al., 1991) and fire-tolerant species (e.g. Gill and Ingwersen, 1976; Lamont and Downes, 1979; Auld, 1986; Zammit and Westoby, 1987; Bradstock and Myerscough, 1988; Bradstock, 1990; Davies and Myerscough, 1991; Lamont and Runciman, 1993) in Australia. However, there have been few comparative studies of co- occurring species (e.g. Beadle, 1940; Hodgkinson and Griffin, 1982; Benson, 1985; Delfs et al., 1987; Clark, 1988; Cowling et al., 1990; Auld and O’Connell, 1991). The work reported here seeks to compare some of the population responses to low intensity prescribed fires of a range of co-occurring species of small tree from the Sydney region. In particular, the following questions were addressed: — Proc. LINN. SOC. N.S.W., 115, 1995 110 EFFECTS OF LOW-INTENSITY FIRES IN SYDNEY REGION 1) which of the species are fire-tolerant as adult plants, and what characteristics allow them to be so? 2) for those species with some degree of fire-tolerance, what is the minimum stem size necessary for post-fire survival? 3) for these same species, how does the post-fire death of plants affect the size/age structure of the population? 4) for those species that can regenerate post-fire shoots from dormant buds, what is the pattern of post-fire shoot production from these epicormic and/or stem basal buds? MATERIALS AND METHODS The work was carried out on the ‘Yarrawood’ property of the University of Technology, Sydney, at Yarramundi in the outer western region of the Sydney metropoli- tan area. The vegetation is an open-forest dominated by Eucalyptus punctata, E. fibrosa, E. eximiaand E. oblonga (vegetation type 10ar[iii] of Benson, 1992), occurring on both sand- stone and shale substrates. The vegetation is thus not uniform throughout the area, but varies with soil type (sand versus clay) and aspect (Benson, 1992). Most of the small tree species are disjunctly distributed on the property, depending on their habitat preferences. The majority of the vegetation has been subjected to prescribed fires since the late 1960s, although the western end of the property was last burnt by a high-intensity wildfire in 1968. About half of the property was burnt by a low-intensity prescribed fire in the autumn of 1991 and most of the rest was burnt by a low-intensity prescribed fire in the autumn of 1992. Both fires varied spatially in intensity, with scorch height varying from 2 — 4m. Study samples were taken from these two areas | year after each of these fires and also from the unburnt western area in April of 1992 and 1993. All of the samples from the two burnt areas were combined for data analysis. Plants were sampled by locating as many individuals of the small tree species as possible in each of the three study areas. Small tree species were defined as those species with adult plants with stems usually above 2m tall on the ‘Yarrawood’ property. For each individual plant the following characteristics were recorded:— whether the stem was dead or alive (i.e. whether it had clear evidence of green shoots, either surviving pre-fire shoots or new post-fire shoots); stem circumference at 1m height; number of post-fire aerial (epicormic) shoots; number of post-fire shoots at the stem base (either from a lignotuber or from the bottom 30cm of the trunk, depending on the species). Individual plants were measured only if their stem was greater than Im tall or they had clear evidence of post-fire shoot regeneration. This sampling programme assumed that stem death was always a response to the most recent fire (in the burnt areas), that no plant with a stem greater than Im tall was completely consumed by the most recent fire, and that stem growth since the fire was randomized across all samples and produced variability that was no larger than the variability due to measurement precision. The stem circumferences of plants in the different sample areas were compared using log-likelihood ratio contingency tests on the frequency histograms for each species (Wilkinson, 1989). The median stem circumference necessary for post-fire survival (i.e. the stem size at which there is a probability of 0.5 of the stem surviving the fire) was esti- mated using the trimmed Spearman-Karber method for each species (Hamilton ef al., 1977). The number of shoots produced by plants with live and dead stems were compared using log-likelihood ratio contingency tests on the frequency histograms for each species (Wilkinson, 1989). The relationships between stem circumference and the number of post-fire shoots for each species were assessed using Spearman rank-order correlation coefficients (Minitab Inc., 1991) Adult plant density was also recorded in the area burnt by the prescribed fires (one Proc. LINN. Soc. N.S.W., 115, 1995 D.A. MORRISON 111 or both fires, depending on the species) and in the unburnt area. The number of individ- ual live plants greater than 2m tall of each small tree species was recorded in each of six replicate 15m x 15m quadrats in each of these two areas. Plant abundance was compared between the two areas using Kruskal-Wallis tests for each species (Minitab Inc., 1991). RESULTS Nine species of small tree were common enough on the ‘Yarrawood’ property to be studied (Table 1). Other small tree species recorded, for which less than 50 individuals were located, include:— Acacia longifolia Andrews (Willd.), Acacia trinervata Sieber ex DC., Banksia serrata L.f., Exocarpos cupressiformis Labill., Persoonia levis (Cav.) Domin, and Xylomelum pyriforme (Gaertner) Knight. Of these nine species, only H. sericea had a significantly different abundance of adults between the burnt and unburnt areas (Table 2), suggesting that all of the other species have some degree of tolerance to low-intensity fires as adults. For the other species, up to one-third of all of the plants located post-fire had been killed by the fires (Table 2); death of the plants could not be determined for the two species with root suck- ers (A. mplexaand J. scoparia). TABLE | Species of single-stemmed small tree studied, and their frre-regeneration characteristics Maximum Stem Stem basal Root Species Family height (m) aerial buds buds buds Acacia binervia (H. L. Wendl.) Mimosaceae 10 — — — J. F Macbr. Acacia implexa Benth. Mimosaceae 6 —_— — suckers Acacia parramattensis Tind. Mimosaceae 7 — — _— Casuarina littoralis Salisb. Casuarinaceae 8 — — — Casuarina torulosa Aiton Casuarinaceae 8 — stem base — Hakea sericea Schrader Proteaceae 3 = — = Jacksonia scoparia R. Br. Fabaceae 3 — stem base suckers Leptospermum trinervium (Smith) Myrtaceae 4 epicormic lignotuber _— J. Thompson Persoonia linearis Andrews Proteaceae 4 epicormic lignotuber — All of the eight species with adult fire-tolerance had relatively smaller stems killed by the fires (Fig. 1, Table 3) and the minimum stem size necessary for post-fire survival varied widely between these species (Table 4). Furthermore, the size structure of the stem populations was significantly different in the area subject to prescribed fires compared to the unburnt area for all six species for which there were data (Fig. 1, Table 3). All four of the species that are capable of producing new post-fire shoots at the stem base preferentially did so when the upper stem had been killed (Fig. 2, Table 3), and the number of shoots produced was usually unrelated to the size of the stem (only for C. torulosa was the relationship between stem circumference and number of basal shoots statistically significant) (Fig. 3). Both of the species that are capable of producing new post-fire epicormic shoots almost invariably did so if the stem was alive post-fire (Fig. 4), and for both species the number of shoots produced was statistically significantly related to the size of the stem (Fig. 5). Proc. LINN. Soc. N.S.W., 115, 1995 112 TABLE 2 EFFECTS OF LOW-INTENSITY FIRES IN SYDNEY REGION Density of live adult plants of the small tree species in the area subject to the prescribed fires in either 1991 or 1992 and in the unburnt area, and the number of plants apparently killed by the fires. Density (plants/225m°) * Species Unburnt area Acacia binervia 0.00 (0.00) Acacia implexa 0.00 (0.00) Acacia parramattensis 3.83 (2.46) Casuarina littoralis 1.17 (0.75) Casuarina torulosa 9.00 (5.59) Hakea sericea 30.33 (7.49) Jacksonia scoparia 6.17 (2.68) Leptospermum trinervium 24.00 (6.83) Persoonia linearis 10.17 (2.46) * Mean (standard error). Burnt area 2.50 (2.50) 4.00 (4.00) 2.33 (1.50) 3.83 (2.59) 11.67 (3.52) 0.00 (0.00) 1.67 (1.28) 10.83 (4.85) 4.83 (2.06) + Species has suckers, and so death of a plant was not determinable. TABLE 3 Kruskal-Wallis test H P 1.00 0.318 1.00 0.318 0.15 0.703 0.15 0.703 1.09 0.297 9:47 0.002 2.94 0.087 1.89 0.169 2.58 0.108 % of plants killed by the low-intensity fires (n) 95.9 (212) —+ 17.4 (138) 33.6 (321) 8.2 (170) 100.0 (56) —+ 5.0 (261) 0.0 (153) Results of the log-likelihood ratio contingency tests for the comparison of the frequency histograms of stem circumference and Species Acacia binervia* Acacia implexa* Acacia parramattensis Casuarina littoralis Casuarina torulosa Hakea serncea+ Jacksonia scoparia Leptospermum trinervium Persoonia linearis number of basal shoots for the small tree species. Stem circumference of alive versus dead stems in the burnt area G P G P 117.01 <0.001 — — 46.27 <0.001 — — 49.62 <0.001 15.92 0.007 179.84 <0.001 46.62 <0.001 121.25 <0.001 26.15 <0.001 43.38 <0.001 53.52 <0.001 129.21 <0.001 17.78 0.001 70.64 <0.001 19.36 0.001 * No stems were found in the unburnt area. + No live stems were found in the burnt area. TABLE 4 Median stem size that the species of small tree must reach before the stem is capable of surviving a low-intensity fire Species Stem circumference (cm) * Acacia binervia (22.2) 24.9 (28.0) Acacia implexa (6.9) 7.7 (8.5) Acacia parramattensis (8.0) 8.6 (9.3) Casuarina littoralis (15.2 ) 16.7 (18.5) Casuarina torulosa (10.6) 11.9 (13.4) Jacksonia scoparia (12.5) 13.7 (15.0) Hakea sericea Leptospermum trinervium (5.7) 6.3 (6.9) Persoonia linearis (5.7) 6.4 (7.2) * median (95% confidence limits). Proc. LINN. Soc. N.S.W., 115, 1995 Stem circumference of alive stems in the burnt versus unburnt areas Number of post-fire basal shoots of alive versus dead stems G P 105.18 <0.001 50.18 <0.001 52.33 <0.001 121.77 <0.001 Number of plants Number of plants Number of plants Number of plants OSNS) 1015) A. binervia n=218 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 Circumference midpoint (cm) A. parramattensis n=198 20 25 30 35 40 45 50 55 60 Circumference midpoint (cm) C. torulosa n=239 15 20 25 30 35 40 45 50 55 60 Circumference midpoint (cm) J. scoparia n=253 10 15 20 25 30 35 40 45 50 55 60 Circumference midpoint (cm) (continued next page) D.A. MORRISON Number of plants Number of plants Number of plants Number of plants 113 A. implexa n=272 OR 45) 0/15 20625: 30n35m40"45750255.60 Circumference midpoint (cm) C. littoralis n=349 1 0 5 10 15 20 30 35 40 45 50 55 60 Circumference midpoint (cm) H. sericea n=116 0 5 10 15 20 25 30 35 40 45 50 55 60 Circumference midpoint (cm) | L. trinervium n=342 O50 Circumference midpoint (cm) Proc. LINN. Soc. N.S.W., 115, 1995 114 EFFECTS OF LOW-INTENSITY FIRES IN SYDNEY REGION Number of plants Ww oO 10 15 20 25 30 35 40 45 50 55 60 P. linearis n=228 Circumference midpoint (cm) Fig 1. Frequency histograms of stem circumference at 1m height for nine single-stemmed small tree species. Alive stems in the burnt areas (open bars); dead stems in the burnt areas (filled bars); alive stems in the unburnt area (hatched bars); n: number of stems sampled. 100 Number of plants Number of plants fo) ° | C. torulosa n=171 7 I # I Ta eT TTT O 2.4 8 8 OO ae Wé es 6) 2 7A Number of basal shoots L. trinervium n=158 Diatirritirritbissibiriabiriitisiitisiitisiitiriid ae | ° ny b a & 10 12 14 16 18 20 22 Number of basal shoots Number of plants Number of plants 100 : J. scoparia 20 n=204 OR A OCB TO dW 1 We 18 2o 22 Number of basal shoots 100 fs ‘ P. linearis 90 n=176 OA ee By WO a 4b WG ales 4) 72 Number of basal shoots Fig. 2. Frequency histograms of the number of post-fire basal (lignotuber or base of stem) shoots per stem for four small tree species. Alive stems (open bars); dead stems (filled bars); n: number of stems sampled. PROC . LINN. Soc. N.S.W., 115, 1995 D.A. MORRISON 115 25 254 C. torulosa J. scoparia 2 r=0.46 * a) ] r=-0.16 ns ° ° 204 ° ° | = i= 4 ” ” 4 o = 154 ‘ e 0 0 ] 5 pe} fo} J 5 5) °° ) 10 4 e 4 4 e o o | E ° 2 = ] e eeeese Ss =] 5 5 e e eo e z z '; e é e e a e e e e e 4 e ee ee e ha Sa a OR a A 0) 5 10 15 20 25 Circumference (cm) Circumference (cm) 2575 " : 2575 i| L. trinervium | P. linearis 2 ] . r=0.09 ns 2 ] 6 r=-0.05 ns Oo 2074 oOo 20 e fo} fo} Sou : = << — e oS 1G 7 byeLe) e oS 4 e ° e & — —_— e ° 10 ° 10 ° e = Q — O i e e e aS 4 2 e ee e e e 37 ¢@ e 5 a4 0 ee § $4 : e Zz | e ee = e ee O +-0—?#- #0 0000-00. 1 7] 77 17] () 5 10 15 20 25 Circumference (cm) Circumference (cm) Fig. 3. Relationship between the number of post-fire basal (lignotuber or base of stem) shoots per stem and stem circumference at 1 m height for four small tree species. r: Spearman rank correlation coefficient; P<0.001, ns: not significant. 60 60 L. trinervium P. linearis 50 n=188 50 n=153 Number of plants Number of plants O 4 8 12 16 20 24 28 32 36 40 44 48 52 56 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 Number of aerial shoots Number of aerial shoots Fig. 4. Frequency histograms of the number of post-fire aerial (epicormic) shoots per stem for two small tree species. Alive stems (open bars); dead stems (filled bars); n: number of stems sampled. Proc. LINN. Soc. N.S.W., 115, 1995 116 EFFECTS OF LOW-INTENSITY FIRES IN SYDNEY REGION 60 60 ‘ L. trinervium ° P. linearis 2 r=0.73 * cy r=0.42 * 6 (50 5 «50 8 ° ° ° cm fa “ “ o I = as o co) 0 0 i) a) _ _ o o a a = = Ss rm) za za Circumference (cm) Circumference (cm) Fig. 5. Relationship between the number of post-fire aerial (epicormic) shoots per stem and stem circumference at 1m height for two small tree species. r: Spearman rank correlation coefficient; * P<0.001. DISCUSSION The nine co-occurring species of small trees studied displayed a wide range of responses to the low-intensity fires encountered at ‘Yarrawood’. Only H. sericea appears to be incapable of surviving low-intensity fires as adult plants (i.e. is fire-sensitive), thus relying entirely on regeneration of new individuals from the canopy-stored seedbank for continuation of the populations. This strategy does not appear to be successful at ‘Yarrawood’, as no seedlings of this species were observed to become established after either of the two prescribed fires and there were no surviving adults. Local extinction of a fire-sensitive species will occur if an inter-fire interval (the time between successive fires) is shorter than the time taken for the plants to reach first reproduction (the primary juvenile period) (e.g. Bradstock and O’Connell, 1988; Pannell and Myerscough, 1993) and this may well be the case for this species at ‘Yarrawood’. Most of the species showing adult tolerance of the low-intensity fires do not show an absolute ability to survive the fires, with up to 34% of the adult plants apparently being killed by the fires. Only for P. linearis were no plants located without post-fire shoots, although this may simply be a sampling error since it is clearly better to label live plants before the fire and then check for post-fire death (my estimates of the number of deaths are thus minimum estimates). Death of the plants could not be determined for the two species with root suckers (A. implexa and J. scoparia), and the physiology and ecology of root-suckering in relation to fire is a neglected area of research (Ashton, 1981). However, all of these eight species can be classified as fire-tolerant for low-intensity fires using the ‘general rule’ (less than one-third of plants killed by the fire) of Gill and Bradstock (1992). It is quite common for species to display variable survival rates even within the one fire (e.g. Beadle, 1940; Hodgkinson & Griffin 1982; Clark, 1988), as dis- played by the small tree species, presumably as a result of variation in both fire intensity and size-related fire resistance of the plants (Hodgkinson & Griffin, 1982; Clark, 1988). Many of the smaller stems of the individuals were killed for all of the fire-tolerant species. Individual plants must thus attain a certain minimum stem size before they are fire-tolerant (i.e. so that the temperature of the meristem tissue is not raised to lethal levels), and this size is presumably related to several growth characteristics that may Proc. LINN. Soc. N.S.W., 115, 1995 D.A. MORRISON 117 protect the living tissue of the plant from the heat of the fire (Gill, 1981). First, as sec- ondary growth progresses bark is formed on the outer surface of the trunk, which may provide a protective insulating layer (the periderm protecting the cambium) that becomes increasingly effective as the stem ages. Secondly, as the stem height increases an increasingly larger amount of the foliage may be held above the scorch height of the flames (i.e. the plant canopy may not be subject to 100% scorch). Thirdly, there may be protected dormant vegetative buds, either on the aerial parts of the stem itself or at its base, and the degree to which these buds are protected may increase as the stem ages. Two of the eight species (L. trinervium and P. linearis) have stems with both aerial and basal (lignotuber) protected dormant vegetative buds as well as quite specialized insulating flakey bark, and stems of these species can survive the low-intensity fires at quite small sizes (6 — 7cm circumference at 1m height); two of the species (C. torulosa and J. scoparia) have stems with only basal (non-lignotuber) protected dormant vegetative buds, and stems of these species must be larger before they can survive the low-intensity fires (11 — 14cm circumference); and two of the species (A. binerviaand C. littoralis) do not have any protected dormant vegetative buds, and stems of these species must be relatively large before they can survive the low-intensity fires (16 — 25cm circumference). Thus there is an apparently logical sequence, whereby species with fewer fire-protection mech- anisms require stems of larger size before they are capable of surviving fires. The two exceptions to this sequence are A. implexaand A. parramattensis, the stems of both of which appear to be able to survive fires at very small sizes (7 — 9cm circumference) without protected vegetative buds. However, both of these species grow rapidly (as do many other acacias), thus lifting the canopy above flame scorch height on stems that have quite small circumferences, and it may also be worth investigating the characteristics of the bark (e.g. thermal diffusivity, thickness, flammability; Gill, 1981) of these species. No growth data exist for any of these species (cf Pannell and Myerscough, 1993), but if they did then it would be possible to also calculate the length of time necessary for each of these species to become fire-tolerant (assuming that stem size is related to age). This time is clearly important for vegetation management purposes — if an inter-fire interval is shorter than the time required for a species to become fire-tolerant then new individuals will not be recruited to the population and local extinction will result (e.g. Bradstock and Myerscough, 1988). It is important to note that the response of the species to high-intensity fires may be quite different to that observed for the low-intensity fires at ‘Yarrawood’. The minimum fire-tolerant size of the stems for high-intensity fires would presumably be much larger for each species, as the heat influx to the stem and the scorch height will both be increased. It is likely that adult stems of the four species without protected buds (A. binervia, A. implexa, A. parramattensis and C. littoralis) may not survive high-intensity fires at all (i.e. the species are fire-sensitive) and nor may the two species with protected buds only at the stem base (C. torulosaand J. scoparia), although plants of A. implexaand J. scoparia may survive due to their root suckers. This may explain why Benson (1981) lists A. parramattensis and C lt- toralis as fire-sensitive species rather than as fire-tolerant. However, Fox (1988) also con- siders C. torulosa to be fire-tolerant. The size-structure of the populations in the burnt and unburntareas is markedly dif- ferent for all of the fire-tolerant species. If the pre-fire structure was similar at some time in the recent past then these differences must be the result of differences in the effect of the subsequent fire regimes (i.e. intensity, frequency, season) on the post-fire re-establish- ment of the populations in the two areas. For most of the species there are relatively more smaller stems in the unburnt area, as would be expected if it is the increased fire frequen- cy in the burnt area that is causing the structural differences. Consequently, it may be reasonable to conclude that the local fire regime has had a significant impact on the rela- tive abundances of these species, and will continue to do so for as long as the current fire management practices continue. If new individuals of these small tree species are not Proc. LINN. Soc. N.S.W., 115, 1995 118 EFFECTS OF LOW-INTENSITY FIRES IN SYDNEY REGION allowed to be recruited to the populations at some time in the future then the populations will eventually become senescent. None of the largest stems were killed by either of the prescribed fires, suggesting that the populations have not yet reached this senescent stage. ° All four of the species with protected buds at the stem base (C. littoralis, J. scoparia, L. trinervium and P. linearis) do not usually produce new shoots unless the upper part of the stem has been killed, irrespective of whether these shoots are from lignotuber buds or not, and the number of shoots produced is usually unrelated to the size of the stem. Therefore, these basal shoots may be viewed as a back-up mechanism that is only employed by the plants when the protection of the stem itself from the heat of the fire fails. Both of the species with protected epicormic buds almost invariably produce post- fire shoots if the stem is still alive, irrespective of whether part of the pre-fire canopy is still alive or not, and the number of shoots produced is directly related to the size of the stem. Therefore, these aerial shoots may be viewed as part of an active post-fire regeneration strategy by the stem rather than as a passive survival of the fire (as in A. parramattensis, A. binervia and C. littoralis), as the canopy is actively replaced or augmented depending on whether it was destroyed by the fire or not. It is clear from the data presented here that there can be no simple classification of plant responses to fires that adequately covers the potential range of post-fire behaviour (of. Gill, 1981; Gill & Bradstock, 1992). Most of the species studied at ‘Yarrawood’ showed considerable spatial variability in their response to the low intensity fires, and several of the species may show considerably different responses when subjected to high-intensity fires. Furthermore, at least three of these species would fit into more than one of the categories defined by Gill (1981), as they have several recovery mechanisms. It is unlikely that any simple sub-division of these categories could be devised to incorporate variable responses, and it is therefore necessary to consider the type of fire being studied before species are assigned to particular categories (cf. Gill & Bradstock, 1992). ACKNOWLEDGEMENTS Thanks to Geoff Cary and more than 100 second-year students of UTS for assistance with the data collection; to Rod Buckney, Geoff Cary and David Keith for helpful discussions; and to Rob Chatterton for many kindnesses during my visits to ‘Yarrawood’. References ASHTON, D.H., 1981. — Fire in tall open-forests (wet sclerophyll forests). Jn: GILL, A.M., GROVES, R.H., and NOBLE, I.R., (eds) Fire and the Austrahan Biota pp. 339-366. Canberra: Australian Academy of Science. AULD, T.D., 1986. — Post-fire demography in the resprouting shrub Angophora hispida (Sm.) Blaxell: Flowering, seed production, dispersal, seedling establishment and survival. Proc. Linn. Soc. N.S.W. 109: 259-269. AULD, T.D., 1987. — Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: Survivorship throughout the life cycle, a synthesis. Aust. J. Ecol. 12: 139-151. AULD, T.D., and O’CONNELL, M.A., 1991. — Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Aust. J. Ecol. 16: 53-70. BEADLE, N.C.W., 1940. — Soil temperatures during forest fires and their effect on the survival of vegetation. J. Ecol. 28: 180-192. 3ENSON, D.H., 1981. — Vegetation of the Agnes Banks sand deposit, Richmond, New South Wales. Cunninghamia 1: 35-57. BENSON, D.H., 1985. — Maturation periods for fire-sensitive shrub species in Hawkesbury sandstone vegetation. Cunninghamia 1: 339-349. BENSON, D.H., 1992. — The natural vegetation of the Penrith 1:100 000 map sheet. Cunninghamia 2: 541-596. BRADSTOCK, R.A., 1990. — Demography of woody plants in relation to fire: Banksia serrata and Isopogon anemoni- folius. Aust. J. Ecol. 15: 117-132. BRADSTOCK, R.A., and AULD, T.D., 1987. — Effects of fire on plants: Case studies in the Proteaceae and Fabaceae in the Sydney region and the implications for fire management in conservation reserves. Jn: Conroy, B., (ed.) Proc. LINN. Soc. N.S.W., 115, 1995 D.A. MORRISON 119 Bushfire Management in Natural Areas pp. 91-119. Sydney: National Parks & Wildlife Service of N.S.W. BRADSTOCK, R.A., and MyERSCOUGH, P.J., 1981.— Fire effect on seed release and emergence of seedlings of Banksia ericifolia L.f. Aust. J. Bot. 36: 414-431. BRADSTOCK, R.A., and MYERSCOUGH, P.J., 1988. — The survival and population response to frequent fires of two woody resprouters Banksia serrata and Isopogon anemonifolius. Aust. ]. Bot. 36: 415-431. BRADSTOCK, R.A., and O’CONNELL, M.A., 1988. — Demography of woody plants in relation to fire: Banksia ericifolia L.f. and Petrophile pulchella (Schrad.) R. Br. Aust. J. Ecol. 13: 505-518. CLARK, S.S., 1988. — Effects of hazard-reduction burning on populations of understorey plant species on Hawkesbury sandstone. Aust. J. Ecol. 13: 473-484. CowLING, R.M., LAMONT, B.B., and ENRIGHT, N.J., 1990. — Fire and management of south-western Australian banksias. Proc. Ecol. Soc. Aust. 16: 177-183. Davies, S.J., and MyerscouGu, P.J., 1991. — Post-fire demography of the wet mallee Eucalyptus leuhmanniana F. Muell. (Myrtaceae). Aust. J. Bot. 39: 459-466. DELFs, J.C., PATE, J.S., and BELL, D.T., 1987. — Northern sandplain kwongan: community biomass and selected species response to fire. J. Roy. Soc. W.A., 69: 133-138. Fox, M.D., 1988. — Understorey changes following fire at Myall Lakes, New South Wales. Cunninghamia 2: 85-95. GILL, A.M., 1981. — Adaptive responses of Australian vascular plant species to fire. In: GILL, A.M., GROVES, R.H., and NoBLE, LR., (eds) Fire and the Australian Biota pp. 243-271. Canberra: Australian Academy of Science. GILL, A.M., and BRADSTOCK, R.A., 1992. — A national register for the fire responses of plant species. Cunninghamia 2: 653-660. GILL, A.M., and INGWERSEN, F., 1976. — Growth of Xanthorrhoea australis R. Br. in relation to fire. J. App. Ecol. 13: 195-203. HAMILTON, M.A., Russo, R.C., and THURSTON, R.V., 1977. — Trimmed Spearman-Karber method for estimating median lethal concentrations in toxicity bioassays. Environ. Sci. Technol. 11: 714-719 [correction 12: 417, 1978]. HODGKINSON, K.C., and GRIFFIN, G.F., 1982. — Adaptation of shrub species to fires in the arid zone. Jn: BARKER, W.R., and GREENSLADE, P.J.M., (eds) Evolution of the Flora and Fauna of Arid Australia pp. 145-152. Adelaide: Peacock Publications. LAMONT, B.B., CONNELL, S.W., and BERGL, S.M., 1991. — Seedbank and population dynamics of Banksia cuneata: the role of time, fire, and moisture. Bot. Gaz. 152: 114122. LAMONT, B.B., and Downes, S., 1979. — The longevity, flowering and fire history of the grasstrees Xanthorrhoea preisstiand Kingia australis. J. App. Ecol. 16: 893-899. LAMONT, B.B., and RUNCIMAN, H.V., 1993. — Fire may stimulate flowering, branching, seed production and seedling establishment in two kangaroo paws (Haemodoraceae). J. App. Ecol. 30: 256-264. MINITAB INC., 1991. — Minitab Reference Manual Release 8. State College: Minitab Inc. PANNELL, J.R., and MYERSCOUGH, P.J., 1993. — Canopy-stored seed banks of Allocasuarina distyla and A. nana in relation to time since fire. Aust. J. Bot. 41: 1-9. WILKINSON, L., 1989. — SYSTAT: The System for Statistics. Evanston: Systat Inc. ZAMMIT, C., and Wesrosy, M., 1987. — Population structure and reproductive status of two Banksia shrubs at various times after fire. Vegetatio 70: 11-20. Proc. LINN. Soc. N.S.W., 115, 1995 r ee ERSrOn ata nae Si x BORA igen es AE SAR by, Scere oy ae thie Me Lee i ave a oa ee fee few eM ALT Se : ‘ Ae we Si, iar iia # psa Whe Naat ‘feor.’ ae vlscaciant eae URES DAIS he ibe ere ne sit oan Or m Vaot Fea fora Big net es He eRe F sibeatth adh iving?, dn adult erAeeS 8) CHR Getta eget brig a tiei sus wares a meat s Tinie CA erent! ©. iW) Hire Lan y BLU TEASED APA sek eee Eye td TS for amstall j rt - ra 7 4 hay fi an +, ar re CRE py " ~r 4 Pa she } é +5 i ‘i b, :} ry ft hel yal clivelasaice a i>) ees Pei Fon (iva ty EATS free Toe GS13. LG REY A A i its > 4 Ary) Fets7s Ca a i sesie | wae ive ist, fey CHL AM. Gia RD 5 7 a i ro aa? Ae fiat ihain (ALATA ep Riv, (tty lve tele? Sirety itil Accalensy aif fev . he » : 1e Trays } ir Tae Sivas ‘ DE ET i Propet aves (Rats ) Aloe: a @ esi ‘ 2 eel vy clitalitulaveer 1 ahert erred for Lerner ela, Be Oe i ae i he eee Va LO VEY ‘fag g 7 Xs Meath, Tos — " . Teeciz tiny ens eee desire tern ie he f <= f A ° , > G 4 yow Gs i i oA Doar pte Sf em Ciba pideay ep ih. fh Ute Aguch Peal ONE > tlhe SAY . aw ee , inf yn ; ee ‘iiwhe 0 : _ Bisa 217k. 1) ~ abies bit Paguc pee Eotirpernwd Few Noiithe Walt, Fa Be A A Nett) owls Gy hi a silt hey wis bd Vhaweetnary aaiiedetinna nif ‘Ur @°¢fA iw rep Agd : ; ' ’ _— fiws ru fiss ies = Fie yy ehee ir at agar denne Bee Pagans HUG HNT vane, Here ecsempegiowi 2d: S48 torr &2, | Ar pint ven 2 yw @ Tie La nee tT tara arid fo wend Jt) (an 6S ASL WLGAT ER. ALA» and AAD, ¥ ra 40K7 A Fives ‘oie Tee OF jaVaue the ane tien ihe Mrinbereanar mau FE ete are wy ete ide axw) de feel Lowa tor fine Tannaeeriewil ina peer Ry eR ae Cabin : i ea et), G8 SE hr RE Factors Affecting Seedling Recruitment of Allocasuarina distylaand A. nanaat Burnt and Unburnt Sites JOHN R. PANNELL (Communicated by D. KEITH) PANNELL, J.R. Factors affecting seedling recruitment of Allocasuarina distyla and A. nana at burnt and unburnt sites. Proc. Linn. Soc. N.S.W. 115: 121-133 (1995). Allocasuarina distyla (Vent.) L. Johnson and A. nana (Sieb. ex Spreng.) L. Johnson are strongly serotinous shrubs, typically releasing their seed (as samaras) en masse after fire. In long-unburnt stands, however, there is a light seed rain from the canopy comprising some viable seed that will germinate after rain. This study examined several factors thought to be responsible for the notable absence of seedling recruits in unburnt vegetation. A sizable proportion of seeds falling from the canopy in dry weather is harvested by ants (seed half life approximately 3.5 days). Heating the soil surface had no effect on seedling establishment in either species. Emergence of both species was enhanced by increased experimental irrigation, suggesting that water deficits may be important in limiting estab- lishment in the field. A. distyla established better in soil covered by litter, conceivably as a result of enhanced soil moisture levels. Recruitment of A. nanaseedlings was reduced by lit- ter on the soil surface, a result probably related to smaller seed size and failure to emerge from the litter. The enclosure of small mammals improved seedling recruitment at sites unburnt for 25 years, although increased number and sizes of recruits at experimental sites in burntrelative to unburnt vegetation are likely to have been mainly a result of both better light conditions and reduced competition. J-R. Pannell, Department of Plant Sciences, University of Oxford, OXI3RB, England; manuscript received 7 December, 1993, accepted for publication 18 May, 1994. KEY WORDS: Seedling establishment, fire, seed predation, Allocasuarina, serotiny, seed rain. INTRODUCTION It has long been observed that the seedling stage of the plant life cycle often exhibits the highest mortality (Miles, 1972; Harper and White, 1974; Harper, 1977). Thus patterns of plant distribution may be largely determined early in the life cycle by factors which affect seed germination and seedling survival. This paper presents the results of field and glasshouse experiments which focus on several such factors for Allocasuarina distyla (Vent.) L. Johnson and A. nana (Sieb. ex Spreng.) L. Johnson, two common and widespread plant species in the Sydney region. A. distyla and A. nana are serotinous, holding their seed as samaras (henceforth referred to as seeds) for extended periods in woody ‘cones’ (infructescences) in their canopies. Seed is typically released en masse following the death of the branch on which it is held, usually as a result of fire (Pannell and Myerscough, 1993). However, a small amount of seed is also dispersed in a light seed rain from older cones on live branches and some of the seeds may be viable (Pannell and Myerscough, 1993). Despite this seed rain there is a notable absence of any seedlings in unburnt vegetation. Similar observations have also been made for other species in fire-prone environments (e.g. Christensen, 1971; Ashton, 1979; O’Dowd and Gill, 1984; Andersen and Ashton, 1985; Andersen, 1988; Keeley, 1992; Williams and Johnson, 1992). The ways in which fire alters the biological and edaphic environment have been well documented (O’ Dowd and Gill, 1984). For example: (1) fire removes much of the biolog- Proc. LINN. SOC. N.S.W., 115, 1995 122 SEEDLING RECRUITMENT OF ALLOCASUARINA ical opposition from a site so that there is more physical space for new growth (Evans, 1976); (2) considerable quantities of mineral and organic nutrients are added to the soil as ash (Pryor, 1963; Purdie, 1977) or as a result of direct heating of the soil (Humphreys and Craig, 1981; Bradstock, 1991); (3) the reduction in canopy cover allows more light into the understorey (McPherson and Muller, 1969; O’Dowd and Gill, 1984); (4) the removal of litter from the ground or the increase in litter cover due to scorched canopy lit- ter fall can interact with evaporation from the soil (O’Dowd and Gill, 1984). Litter can interfere physically with establishing seedlings or alter the soil surface temperature as a result of changes in soil insolation levels (O’ Dowd and Gill, 1984); (5) the burning of the canopy and of litter can halt the release of growth-inhibiting allelopathic toxins into the soil (del Moral and Muller, 1970); (6) a bushfire can interfere with the activities of grani- vores or herbivores (Leigh and Holgate, 1979). Several studies suggest that the en masse release of seed from serotinous plants after fire leads to a satiation of seed predators and an increased likelihood of survival for individual seeds (e.g. Ashton, 1979; Wellington and Noble, 1985; Andersen and Ashton, 1985; Andersen, 1988; Lamont et al., 1991); (7) the various factors altered by fire can themselves interact with the time or season of burning. Bradstock and O’Connell (1988), for example, suggested that late summer/autumn fires favour establishment; (8) different species may respond to the effects of fire in contrast- ing ways if their requirements for germination and establishment differ (Bradstock and Myerscough, 1981; Bradstock, 1991). In this study the influence of the following factors on the emergence and establish- ment of Allocasuarina distyla and A. nana seedlings was investigated: (1) the presence or absence of litter on the soil surface; (2) the effects of heat on the soil and the burning of litter; (3) the effect of germination and growth in soil from different provenances; (4) the effect of differences in soil moisture; (5) the effects of seedlings growing in stands at dif- ferent times since fire; (6) stochastic effects attributable to random differences between sites; (7) the effect of seed harvesting by ants; (8) the effect of browsing by vertebrates. Factors 1, 2, 3 and 4 were investigated in a glasshouse experiment. Factors 5, 6, 7 and 8 were investigated in a second experiment under field conditions. Rates of seed harvesting by ants were further examined by observing the removal of seed from seed caches in unburnt vegetation. Rates of seed dispersal from the canopy were estimated by counting seed caught in traps beneath the canopy in along unburnt coastal population of A. distyla. METHODS The seed rain Seven circular seed traps were located randomly along a 30 m transect beneath the canopy within each of three dense populations of A. distyla, last burnt 25 years previously, on the Lambert Peninsula, Ku-ring-gai Chase National Park (NP). They were approxi- mately 0.2m* in area and were made of finely woven nylon sewn onto a stiff wire frame. Their legs were coated thickly in sticky ‘tanglefoot’ to deter ants. The seed traps were visit- ed at fortnightly intervals from May until August, 1990, and any captured seeds were col- lected and tested for germinability in the laboratory. At each site the density of cones and seed per area were estimated by counting the number of cones above a | metre-wide strip along the transect, multiplying by the average number of seeds per cone, and dividing by 30 to give an estimate in seeds/m’. Seed harvesting by ants Four separate sites were chosen in tall heath stands on the Lambert Peninsula, Ku-ring-gai Chase NP, last burnt 25 years previously. The sites were all similar, being locat- ed on flat ground and dominated by Allocasuarina distyla. At each site eight closed petri dishes with three entry holes cut into their sides (approximately 1 cm’) were located ran- Proc. LINN. SOC. N.S.W., 115, 1995 J.R. PANNELL 123 domly along two independently laid 30 m transects. Ten A. distyla seeds were placed in four of the dishes, while ten of A. nanawere placed i in the other four. The transects at each site were laid randomly up to 50 m apart, running perpendicular to a baseline transect. The petri dishes were recovered after twenty-four hours and the number of seeds removed was recorded. The entire procedure was replicated four times, each time at inde- pendent (different) sites. The results were analysed by a 4-factor ANOVA with species of seed treated as a fixed factor, time as a random factor, sites as a random factor nested in time and transects as a random factor nested in the times x sites interaction. During the course of the study petri dishes were found disturbed on a number of occasions, probably by birds or small mammals (Andersen and Ashton, 1985), so the data were analysed in two different ways: (1) using all the data, ignoring any possible effects of the disturbed petri dishes; (2) using an unbalanced data set with all values from disturbed petri dishes excluded from the analysis. In both analyses data were numbers of seeds removed in twenty-four hours from a depot of ten seeds. Variances were homogeneous, with degrees of freedom defined by the largest cell (Cochran’s statistic = 0.0048, first analysis; = 0.032, second analysis; k = 64, v= 3; P>0.05). Glasshouse experiment Soil and litter were sampled from three sites, one in Royal NP, one in Ku-ring-gai Chase NP and one in Bouddi NP in the Sydney region. At each site sixteen samples of soil and twenty-four of litter were chosen randomly from a 30 x 30 m grid. Soil was removed from an area of 15 x 15 cm’ toa depth of approximately 10 cm and placed carefully into a one-litre plastic container with bottom drainage holes. Litter covering about the same area was bagged separately. In the laboratory all the samples were placed in ovens at 60°C for about thirty-six hours to dry. The following treatments were then applied to the soil samples from each site: (1) the top 2 cm of soil plus 40 g of litter were heated with the intense flame of a bunsen burner from above for about five minutes (fire does not heat soil substantially below about 1.5 cm in depth; Humphreys and Craig, 1981); (2) half the burnt and half the unburnt samples were covered with 40 g of litter (about 2 cm litter depth); (3) half the samples were given 25 mL water twice weekly; the other half were given 50 mL twice weekly. These quantities corresponded to an even annual precipitation of approximately 115 mm and 230 mm respectively, well below the mean annual rainfall for the sites where the two species occur. Fifty seeds of A. distyla were scattered on the soil surface (beneath any litter) over one half of each container, while fifty seeds of A. nana were scattered over the other half. (The containers were partitioned in two by fly-screen gauze dividers). The experimental treatments were arranged in a fully orthogonal design with species and soil provenance. The numbers of surviving seedlings at the end of four months were analysed by a 5- factor ANOVA with species, fire, litter and water treatments treated as fixed factors, and soil provenance treated as a random factor. Variances were homogeneous (Cochran’s statistic = 1.11;k=48,v=1;P>0.05). The above-ground dry weights of harvested seedlings were analysed similarly. Four replicate seedlings were sampled randomly for each treat- ment combination, except that only two were available in one of the cells. Variances could be stabilized by a square-root transformation of the data with one outlier removed (Cochran’s statistic = 0.101; k= 48, v=3; P>0.05). Field experiment Two replicate sites were chosen within the Sydney region for each of three different ages (times since fire) of vegetation: recently burnt sites (time since fire less than one year) were located in Bouddi NP and Brisbane Water NP north of Sydney; sites burnt nine years previously were in Royal NP and on the Lambert Peninsula in Ku-ring-gai NP and sites burnt twenty-five years previously were located in Royal NP and Bouddi NP. Site local- ities and floristics are summarised briefly in Table 1. Proc. LINN. SOC. N.S.W., 115, 1995 124 SEEDLING RECRUITMENT OF ALLOCASUARINA TABLE | Locations and prominent species occurring at sites selected for the field experiment. Grid references are those of the Central Mapping Authority of NSW. Species nomenclature follows Beadle et al. (1982). Time since fire Site location 1. Species occurrence Site location 2. Species occurrence <1 year Brisbane Water NP _—Allocasuarina distyla Bouddi NP Allocasuarina distyla GR 414875 Banksia ericifolia GR 473883 Banksia oblongifolia B. oblongifolia Hakea dactyloides Isopogon anemonifolius H. gibbosa Darwinia fascicularis Isopogon anemonifolius Angophora hispida I. anethifolius Acacia oxycedrus Lambertia formosa A. suaveolens Angophora hispida Leucopogon microphyllus Pultenea elliptica Epacris microphylla Lasiopetalum ferrugineum 9 years Ku-ring-gai NP Allocasuarina distyla Royal NP Allocasuarina distyla GP 408827 Banksia ericifolia GR 157205 Banksia ericifolia B. oblongifolia B. oblongifolia Petrophile pulchella Persoonza levis Isopogon anemonifolius Petrophile pulchella Hakea teretifolio Hakea teretifolia H. dactyloides Darwinia fascicularis H. gibbosa Epacris microphylla Persoonia lanceolata Leucopogon microphyllus Acacia oxycedrus Dillwynia retorta A. suaveolens Olax stricta A. ulicifolia Caustis pentandra 25 years Bouddi NP Allocasuarina distyla Royal NP Allocasuarina distyla GR 470882 Banksia ericifolia GR 203291 Banksia enicifolia B. oblongifolia Banksia marginata Isopogon anethifolius Grevillea sericea Hakea teretifolia Isopogon anemonifolius H. gibbosa Leucopogon microphylla Woollsia pungens Acacia ulicifolia Acacia myrtifolia Lasiopetalum ferrugineum Random co-ordinates were used to locate twenty points within a 15 x 15 m grid at each of the six sites described above. At each point, eighty seeds of A. distyla and eighty seeds of A. nanawere sown over an area of about 15 cm in diameter. The two species were partitioned from one another by fly-screen gauze dividers for ease of seedling identifica- tion. The following exclosure treatments were applied to the sown seeds: (1) an ant exclo- sure; (2) a mammal exclosure; (3) ant and mammal exclosures together; (4) no exclosures; (5) a sham (control) exclosure. Ants were excluded from plots by surround- ing them with a cylindrical galvanized iron ring (15 cm diameter, 8 cm deep) which was sunk 3 cm into the ground. The outside of the cylinder was thickly coated in ‘tanglefoot’, a water-resistant, sticky substance, to deter ants. The control was set up without tanglefoot. Small mammals were excluded from plots by exposure cages made out of PVC mesh (15 x 20 x 10 cm; pore size about | cm’). The control was left open on two sides. In ensuring the efficacy of the exclosure cages it was necessary to interfere with the existing conditions of the treatment plots (such as litter cover) to some extent and care was taken to disturb the uncaged plots comparably. In order to control for the possibility of natural A. distyla emergents four additional plots were randomly located at each site and disturbed in a similar manner, except that no seeds were sown. No seedlings emerged at any of these plots over the course of the study. The experimental treatments were orthogonal with species and time since fire, except that the control was arranged asym- metrically in the design (see results). The experiment was commenced in mid April, 1990, and seedlings were finally harvested at the end of August, 1990. J.R. PANNELL 125 The numbers of emergents after (1) six weeks and (2) four months were analysed by a 5-factor ANOVA with all factors treated as fixed except sites (random and nested in time since fire). Variances in the first analysis were stabilised by In (x + 1) transform (Cochran’s statistic = 0.076; k = 60, v = 3; P > 0.05); those in the second were established by a In(1nx +1) +1 data transform (Cochran’s statistic = 0.065; k = 60, v= 3; P >0.05). There were insufficient numbers of seedlings after four months to allow their dry weights and heights to be analysed in a balanced design of the full experiment. Five seedlings of each species per site were therefore randomly selected from plots at which both ants and mammals had been excluded. (The dry weights and heights of seedlings of A. distyla were found not to depend substantially on the exclosure cages). The above- ground dry weights and the heights of these seedlings were analysed by fully balanced 3-factor ANOVAs with species and time since fire treated as fixed factors and site (nested within time since fire) treated as a random factor. A In (x + 1) data transform stabilised the variances in both analyses (Cochran’s statistic = 0.168, dry weights; = 0.223, heights; k=12,v=4;P>0.05). RESULTS AND DISCUSSION The seed rain The rate at which seed was captured in the seed traps was very low, and there were no easily discernible trends from site to site. For each month from May to August the amount of seed collected was pooled across all sites and ranged from a minimum of 0.25 seeds/ week/m? in June to amaximum of 2.3 seeds/week/m? in August. Averaged across all three sites, there were approximately 46 cones/m’, or about 3200 seeds/m’, in the canopy above the traps. Thus the seed rain per week, expressed in terms of seeds per canopy- stored seed, was very roughly between 8/100 000 (June) and 7/10 000 (August). Over the four months of the study a total of six cones were caught in the traps. This corresponds to a cone-capture rate of 0.11 cones/week/m’. Alternatively, 2.4 cones fell per week for every 1000 cones in the canopy. Of the individual seeds caught in the seed traps, only about 35% were viable. Very few seeds were recovered from the six cones caught in the traps. Those which were released from between the bracteoles were largely decayed and inviable, and only four seeds germinated from the 370 bracteole pairs counted on the cones. This result confirms the finding of Pannell and Myerscough (1993) that cones are abscised from the canopy when they old and after most of the seed has already died. Spatial and temporal variation in ant harvesting Seed removal differed significantly between transects within sites (P <0.025), but not between sites (P >0.20). The seeds of the two species were removed at different relative rates between times (species x time interaction: P <0.05). Marked small-scale spatial varia- tion in rates of seed removal is possibly associated with the territorial foraging patterns of ants and could have important consequences for spatial patterns of seedling recruitment following fire (Andersen, 1988). The lack of significant variation between sites is not sur- prising; Mossop (1989) found no significant inter-site differences in ant foraging pat- terns, even between rainforest and dry sclerophyll communities. There were no main effects of time on seed removal rates. Other studies have shown that seed removal rates can be seasonally related to ambient temperature (Cremer, 1965; Christensen, 1971; O’Dowd and Gill, 1984; Andersen and Ashton, 1985; Mossop, 1989) and to stochastic events such as rainfall (Andersen and Ashton, 1985). In the present study, the fact that weather conditions were uniform over the course of observations would account for the relatively low variability in seed harvesting rates over the course of the study. PrRoc.LINN.SOC.N.S.W., 115, 1995 126 SEEDLING RECRUITMENT OF ALLOCASUARINA TABLE 2 ANOVA of numbers of seedlings after 6 weeks and 4 months in the field. There were three random sites within each of the three Jixed fire histories (see text). Treatments were fixed: ‘Ants’ and ‘Mammals’ are the effects of an ant and mammal exclosure, respectively; ‘Cont. us. ET’ is the effect of the sham exclosure control relative to the experimental treatments. n = 4 replicates Six weeks Four months Source df. F-ratio P Fratio P Fire: F 2 5.95 6.04 Site: P (F) 3 6.56 <0.001 4.17 <0.01 Species: S 1 75.0 <0.01 101 <0.01 (Treatment): T (4) Cont. vs. ET: C 1 0.706 1.56 (Among ET): ET (3) Mammals: M 1 5.88 24.2 <0.05 Ants: A 1 11.2 <0.05 11.3 <0.05 AxM 1 0.016 0.20 FxS 2 4.16 5.23 FxT (8) FxC 2 0.106 .0145 FxXET (6) FxM 2 0.167 0.235 FxA 2 1.48 0.684 FxAxM 2 1.20 6.66 P (F) xS 3 0.73 0.660 P (F) xT (12) P (F) xC 3 13.6 <0.001 12.4 <0.001 (F) xET (9) P (F) xM 3 3.99 <0.01 2.54 3 3.67 <0.05 3.163 <0.05 P (F) xAxM 3 1.18 0.442 SxT (4) SxC 1 17.6 <0.05 8.84 SxET (3) SxM 1 1.58 1.10 SxA 1 4.17 <0.05 9.95 SxAxM 1 0.377 0.167 FxSxT (8) FxSxC 2 6.35 0.964 FxSxET (6) FxSxM 2 0.202 0.752 FxSxA 2 0.759 5.19 FxSxAxM 2 4.95 8.23 P (F) xSxT (12) P (F) xSxC 3 0.326 0.565 P (F) xSxET (9) P (F) xSxM 3 0.741 0.823 P. (F) xSxA 3 0.208 0.143 3 0.419 0.265 Residual 180 Total 239 The estimated half-life of a seed under the conditions of the study was 3.56 days, sim- ilar to that found by Wellington and Noble (1985) for Eucalyptus incrassata seeds and Pierce and Cowling (1991) for seeds of South African fynbos species. It gives an indication Proc. LINN. SOC. N.S.W., 115, 1995 J.R. PANNELL 127 of the order of magnitude which can be expected, but this estimate should be treated with caution since removal rates were almost certainly not constant. The estimated removal rates differed depending on whether all the seeds in the depots were considered (0.177 per day) or whether each depot was treated as a single seed (0.353 per day), with depots spaced by a mean of 3.75m. Other workers have found similar differences (e.g. Andersen, 1988). Hughes and Westoby (1990) suggested that they may reflect the relative propor- tions of individual and group foragers in the ant community, since group foragers would tend to remove more seeds from fewer depots. This explanation would suggest that, in the present study, group foraging predominated. It is likely that A. distylaand A. nana seed removed by ants is consumed in the ants’ nests. The alternative possibility that seed removal is a dispersal event is unlikely. The seeds of many legumes are typified by the presence of a specialised food body (elaio- some), a feature which is assumed to be an adaptive reward for seed dispersal agents (Hughes and Westoby, 1990, 1992). However, seeds which provide no food reward for potential dispersal agents (like those of A. distylaand A. nana) are probably removed to be eaten (Pierce and Cowling, 1991). Ashton (1979) found this to be the case for Eucalyptus regnans. Such seed is lost to the potential soil-stored seed bank. Glasshouse experiment It has been suggested that the addition of mineral and organic nutrients to the soil resulting from a bush fire may substantially improve the seedling establishment (e.g. Pryor, 1963; Christensen and Muller, 1975; Purdie, 1977; Humphreys and Craig, 1981). This was not the case for A. distylaand A. nana under glasshouse conditions (F-ratio <1; P> >0.25). In both species more individuals established under the high water treatment rela- tive to the low water treatment (P <0.05), a result which was related to differences in emer- gence and not to seedling mortality. In the field, levels of soil moisture strongly affect seedling survivorship (e.g. Miles, 1972; O’Dowd and Gill, 1984; Bradstock, 1985). However, this effect is quite different from the moisture effect on emergence found here where ungerminated seeds are dormant and may still germinate under a more favourable moisture regime. Litter present on the soil may reduce seedling establishment (Williams and Johnson 1992), with seedlings dying before their radicles reach the soil (Purdie, 1977). This appears to have been the case for germinants of A. nana rather than for A. distyla (com- pare Figures | and 2), a difference which can be explained by the fact that A. nana seeds are the smaller and have fewer reserves for initial growth. O’ Dowd and Gill (1984) noted the potential increase in seedling mortality with increased litter fall following leaf scorch during fire. They also suggested that post-fire seedling mortality could be partly due to water deficits as a result of increased soil insolation through the burnt canopy. The obvi- ous corollary of this is that litter cover may have a positive effect on seedling recruitment due to the conservation of soil moisture under it (Manders and Richardson, 1992; Keeley, 1992). In the glasshouse litter seemed to have acted in this way on the early recruitment of A. distyla seedlings, enhancing recruitment and survivorship of seedlings in each of the soil moisture levels (Figure 1). This suggests that seedling establishment may vary between species with fires of different intensities; increased establishment could be expected for relatively large-seeded plants which can emerge through deep fallen scorched litter following low intensity burns, while the converse could be expected for rel- atively small-seeded plants which do not have the reserves to emerge from the litter cover. Field experiment Treatment effects at the end of the four months were already apparent after six weeks (compare analyses in Table 2; see Figures 3 and 4). There was also a reasonably con- sistent decline in seedling numbers in both species between six weeks and four months Proc. LINN. SOC. N.S.W., 115, 1995 128 SEEDLING RECRUITMENT OF ALLOCASUARINA after sowing, a trend more pronounced at sites burnt twenty-five years previously (Figures 3 and 4). Seedling recruitment differed significantly between sites. Although the signifi- cant site effect depended on its interaction with the treatment effects (see below), a rea- sonably consistent trend across all of the treatments was the considerable difference in numbers of recruits between the recently burnt sites and other sites, especially those burnt nine years previously. A. distyla responded more to site differences in seedling height and dry weight than A. nana (Figure 5; S x P(F) interaction: heights, P <0.001; weights, P <0.05). A distyla seedlings were taller and weighed more at both the recently burnt sites and at sites burnt twenty-five years previously (Figure 5). A multitude of factors may affect seedling establishment success at the recently burnt sites (Christensen and Muller, 1975; O’Dowd and Gill, 1984; Bradstock, 1991; Williams and Johnson, 1992). Increased insolation was probably important (Miles, 1972), although it is likely that direct competition for biological space at the scale of the seedlings (i.e. close to the ground) was a more important factor at the sites burnt only about a decade ago than at those burnt two and a half decades ago, since the latter had an almost clear understorey. The greater above-ground size of A. distyla seedlings grown at the recently burnt sites indicates that they might survive a hot, dry spell in summer better than those which established at unburnt sites. This possibility is foreshadowed by the experi- mental results. The decline in the number of seedlings between six weeks and four months after sowing was greatest at sites last burnt twenty-five years previously, and it is likely that this difference between sites would have become more pronounced during hot, dry weather in summer (Trabaud and Campant, 1991). Since larger seedlings are more likely to survive conditions in summer, early autumn emergence and subsequent initial establishment during winter can be expected to enhance long-term survivorship. Seedlings which emerge at sites most conducive to fast growth (i.e. burnt sites) are likely to establish best. The corollary of this is that seedlings will establish most successfully at sites burnt in late summer (with en masse seed release in early autumn). A similar conclusion was reached by Bradstock (1985) for seedlings of four Proteaceae species. Bradstock (1991) concluded that seedling predation may be important in limiting seedling establishment in proteaceous species at unburnt sites. In the present study, the way in which the exclosure of ants and small mammals influenced seedling establishment was complicated by the significant interactions of the sham exclosure control with species in the first six weeks (Table 2, P <0.05) and its interaction with sites after four months (Table 2, P <0.001). The latter interaction indicates that there was a significant physical effect of the exclosure cages used in the experiment at some sites. For example, there were particularly substantial differences in recruits at recently burnt sites between the null treatment (no exclosures) and sham exclosures (Figure 3a and b). At these sites (to a less- er extent also at sites last burnt nine years previously) the effects of ants and mammals on seedling numbers were confronted by the physical effects of the exclosures. However, in sites last burnt twenty-five years prior to the experiment, the sham exclosures had no sig- nificant effect on seedling recruitment (compare Figures 3a and b, and Figures 4a and 4b). In the stands of older vegetation, then, it is reasonable to conclude that the activity of ants and mammals was responsible for reducing the numbers of seedling recruits, espe- cially during the first six weeks (note the significant [P <0.05] P(F) xA and P(F) x M terms in Table 4). Although the experimental design does not allow statistical discrimination between the individual effects of the two exclosures used, the data suggest that the ant exclosure was chiefly responsible for the confounding effects. This is not surprising as it consider- ably modified the micro-environment of the experimental plots, most particularly the drainage patterns; seeds surrounded by an ant exclosure would not have been washed away in the heavy rains which fell in the weeks immediately following the experiment’s induction, whereas many seeds unprotected in this way were probably lost in the run-off. This would largely account for the fact that the physical effect of the exclosures was promi- Proc. LINN. SOC. N.S.W., 115, 1995 J.R. PANNELL 129 (a). Low water treatment Z Soil 1 Soil 2 HB Soil3 30 (a). Low water treatment. ZZ Soil 1. EX Soil 2. 2074 HE Soil3. a D eS 3 pte 8 £ = > ° ° 5 (b). High water treatment 7) a2 = E ° (b). High water treatment. 2 8 E 3 z No litter eniteover Litter No litter Litter Soil cover. I 2 Fig. 1. Seedling numbers of A. distyla for the glass Fig. 2. Seedling numbers of A. nana for the house experiment. Low water treatment was 25mL glasshouse experiment. See text for details of experi- twice weekly; high water treatment was 50 mL twice mental treatments. Error bars are one SE; n = 4. weekly. Soils were taken from different sites (see text for details). Error bars are one SE; n = 4. nent at the recently burnt sites where run-off from the bare soil was probably fastest. The effect of ant activity on seedling numbers was most likely due to their removal of seed, as was foreshadowed by the observation of seed-harvesting by ants reported above. This effect might have been much greater had the period subsequent to seed sowing been warmer and drier (very heavy falls of rain occurred in April). It is now reasonably well established that foraging activity of ants is seasonally related to temperature and that ants are inactive during wet spells (Hughes and Westoby, 1990). This would suggest that seed dispersal may be most effective in cooler months when ant foraging activity is low. CONCLUSIONS In dense, old (25 years since fire), stands of A. distyla, the seed rain ranged between 0.25 and 2.3 seeds/week/m’, a figure which probably underestimates the rate typical of summer months. Only about 35% of this seed was viable so that, on average, about one seed capable of germinating fell per square metre each week. Field observations indicate that none of these seeds ever establish as successful seedlings. Some of the possible rea- sons for this lack of recruitment in the inter-fire period were investigated in this study. During the period of the study, between 1.3 and 2.4 seeds were removed by ants each week from artificial seed depots separated by a mean distance of about 4 m. It is thus conceivable that a large proportion of the seeds which fall from the canopy in unburnt stands of Allocasuarina are harvested and eaten by ants. This hypothesis was supported at some sites by improved seedling establishment within ant exclosures. Viable seeds which escape being eaten by ants (e.g. those which fall prior to a wet period) will germinate under the canopy after sufficient rain. Such germinations are likely to occur occasionally Proc. LINN. SOC. N.S.W., 115, 1995 130 SEEDLING RECRUITMENT OF ALLOCASUARINA Six weeks £3 Four months (a) Null treatment (no exclosures) (b) Sham exclosures Number of seedlings (e) Ants and mammais excluded O(a) O(b) 9(a) 9(b) 25(a) 25(b) Site 3 Fig. 3. Seedling numbers of A. distyla after 6 weeks and 4 months from 80 seeds sown. See Table 1. for site localities and species lists; see text for details of experimental treatments. 0(a) and 0(b) were sites recently burnt; 9(a) and 9(b) were sites burnt 9 years previously; 25(a) and 25(b) were sites burnt 25 years previously. Error bars are one SE; n=4. Proc. LINN. SOC. N.S.W., 115, 1995 (a) Null treatmen; Six weeks | Four months (no exclosures, (b) Sham exclosures (c) Ants excluded Number of seedlings (d) Mammals excluded O(a) 0(b) 9(a) 9(b) 25(a) 25(b) Site 4 Fig. 4. Seedling numbers of A. nanaafter 6 weeks and 4 months from 80 seeds sown. See Table 1. for site localities and species lists; see text for details of experimental treatments. 0(a) and 0(b) were sites recently burnt; 9(a) and 9(b) were sites burnt 9 years previously; 25(a) and 25(b) were sites burnt 25 years previously. Error bars are one SE; n = 4. 131 J.R. PANNELL & on O 2 < N A. nana 60 oO oO vt N (ww) jyBjey Bujjpees 20 J A. distyla 9) _— fo) uw) (Bw) yyBjem Buy|pees Site z seek gz | 1e0A Cz z seek 6 @ juing | quing gs harvested after 4 months by, Fig. 5. Above-ground dry weights (a) and heights (b) of A. distylaand A. nanaseedlin in the field. See Table 1. for site localities and species lists. Error bars are one SE; n Proc. LINN. SOC. N.S.W., 115, 1995 132 SEEDLING RECRUITMENT OF ALLOCASUARINA in nature, but in contrast to seedlings establishing at recently burnt sites recruitment to adults fails in unburnt stands. The results of this study indicate that whereas a proportion of Allocasuarina seedlings may be browsed by small mammals, the chief factors limiting successful establishment under a canopy are likely to be poor light quality and competi- tion for resources. ACKNOWLEDGEMENTS I would like to thank Peter Myerscough for his supervision and helpful criticism and the National Parks and Wildlife Service of NSW for permission to work on its lands. Tony Underwood assisted with the analysis of the field experiment. References ANDERSON, A.N., 1988. — Immediate and longer term effects of fire on seed predation by ants in sclerophyllous vegetation in southeastern Australia. Australian Journal of Ecology 13: 285-93. ANDERSEN, A.N., and ASHTON, D.H., 1985. — Rates of seed removal by ants at heath and woodland sites in south- eastern Australia. Australian Journal of Ecology 10: 381:90. ASHTON, D.H., 1979. — Seed harvesting by ants in forests of Eucalyptus regnans in central Victoria. Australian Journal of Ecology 4: 265-77. BEADLE, N.C.W., EVANS, O.D., CAROLIN, R.C., and TINDALE, M.D., 1982. — Flora of the Sydney Region, 3rd ed. A.H. and A.W. Reeds, Sydney. BRADSTOCK, R.A., 1991. — The role of fire in establishment of seedlings of serotinous species from the Sydney Region. Australian Journal of Botany 39: 347-56. BRADSTOCK, R.A., and MYERSCOUGH, P.J., 1981. — Fire effect on seed release and the emergence of seedlings in Banksia ericifolia. Australian Journal of Botany 36: 414-431. BRADSTOCK, R.A., and O’CONNELL, M.A., 1988. — Demography of woody plants in relation to fire: Banksia ericifolia and Petrophile pulchella, Australian Journal of Ecology 14: 505-18. BRADSTOCK, R.A., 1985. — Plant population dynamics under varying fire regimes. Ph.D. Thesis, School of Biological Sciences, University of Sydney. CHRISTENSEN, N.L., and MULLER, C.H., 1975. — Effects of fire on factors controlling plant growth in Adenostoma chapparal. Ecological Monographs 45: 29-55. CHRISTENSEN, P.E., 1971. — Stimulation in seedfall in karri. Australian Forestry 35: 180-90. CREMER, K.W., 1965. — Effects of fire on seed shed from Eucalyptus regnans. Australian Forestry 29: 251-62. DEL Moral, R., and MULLER, C.H., 1970. — The allelopathic effects of Eucalyptus camaldulensis. American Midland Naturalist 83: 254-82. Evans, G.C., 1976. — A sack of uncut diamonds: the study of ecosystems and the future resources of mankind. Journal of Ecology 64: 1-9. Harper, J.L., and WHITE, J., 1974. — The demography of plants. Annual Review of Ecology and Systematics. 5: 419-63. HUGHES, L., and Westosy, M., 1992. — Fate of seeds adapted for dispersal by ants in Australian sclerophyll vegeta- tion. Ecology 73: 1285-99. HUuGueEs, L., and Wesropy, M., 1990. — Removal rates of seeds adapted for dispersal by ants. Ecology 71: 138-148. HUMPHREYS, F.R., and CRale, F.G., 1981. — Effects of fire on soil chemical, structural and hydrological properties. In: GiLL, A.M., GROVES, R.H., and NoBLE, IR., (eds.). Fire and the Australian Biota pp 177-200. Australian Academy of Science, Canberra. KEELEY, J.E., 1992. — Demographic structure of California chaparral in the long-term absence of fire. Journal of Vegetation Science 3: 79-90. LAMONT, B.B., LE-MAITRE, D.C., COWLING, R.M., and ENRIGHT, N.J., 1991. — Canopy seed storage in woody plants. Botanical Review 57: 277-317. LeiGH, J.H., and Hocate, M.D., 1979. — The responses of the understorey of forests and woodlands of the Southern Tablelands to grazing and burning. Australian Journal of Ecology 4: 25-45. MANDERS, P.T., and RICHARDSON, D.M., 1992. — Colonization of Cape fynbos communities by forest species. Forest Ecology and Management 48: 277-93. McPHERSON, J.K., and MULLER, C.H., 1969. — Allelopathic effects of Adenostroma fasciculatum, ‘Chamise’, in the California Chapparal. Ecological Monographs 39: 177-198. MILES, J., 1972. — Experimental establishment of seedlings on a southern English heath. Journal of Ecology 60: 225- 34. Mossop, M.K., 1989. — Comparison of seed removal by ants in vegetation on fertile and infertile soils. Australian Journal of Ecology 14: 367-74. O’Dowp, D.J., and GILL, A.M., 1984. — Predator satiation and site alteration: mass reproduction of alpine ash Eucalyptus delegatensis following fire in southeastern Australia. Ecology 65: 1052-66. PANNELL, J.R. and MyerscouGu, P.J., 1993. — Canopy-stored seed banks of Allocasuarina distyla and A. nana in relation to time since fire. Australian Journal of Botany 41: 1-9. Proc. LINN. SOC. N.S.W., 115, 1995 J.R. PANNELL 133 PIERCE, S.M., and COwLING, R.M., 1991. — Dynamics of soil-stored seed banks of six shrubs in fire-prone dune fynbos. Journal of Ecology 79: 731-47. Pryor, L.D., 1963. — Ash bed growth responses as a key to plantation establishment on poor sites. Australian Forestry 27: 48-51. PuRDIE, R.W., 1977. — Early stages of regeneration after burning in dry sclerophyll vegetation. II. Regeneration by seed germination. Australian Journal of Botany 25: 35-46. SPECHT, R.L., 1981. — Responses to fires of heathlands and related shrublands. Jn: GILL, A.M., GROVES, R.H., and Noble, I.R., (eds.) Fire and the Australian Biota pp. 395-416. Australian Academy of Sciences, Canberra. TRABAUD, L., and CAMPANT, C., 1991. — Problems in the natural post-fire regeneration of the Salzmann Pine (Pinus nigra ssp. salzmannii (Dunal) Franco. Biological Conservation 58: 329-44. WELLINGTON, A.B., and NoBLE, I.R., 1985. — Seed dynamics and factors limiting recruitment of the mallee Eucalyptus incrassata in semi-arid, southeastern Australia. Journal of Ecology'’73: 657-66. WILLIAMS, C.E., and JOHNSON, W.C., 1992. — Factors affecting recruitment of Pinus pungens in the southern Appalachian Mountains. Canadian Journal of Forest Research 22: 878-87. Proc. LINN. SOC. N.S.W., 115, 1995 a tie, oN ote CSTs saldyeod toe dtsveg pasties ali nth xeuathes > hime Sh iaipenoceegeton ii .038 pee hy inks he ce 8 0 EOE Hes y as Lfetoines | ] Va * , iint Bt a tiene Py’ pike atite 24 tax Peters EN, 7h, Ae AD Sayre ot Le ins ; aun. ure, j 474. = The & het ah Wiz LM. 1208. ~— Fam ot ser dai ee Brake . 298 Pe >¢495 3 * Se Peiirweem FAS, ar 16. mR, Ma. ¢ tie) os et a M Cached, CB aes i ) aig ot Seweie, Crater. uk i He ae Unie £4 7 LOS. — fag ygregiys structure of Californie chapaiyal iat thie! haem wanes Ss ant oe iM ‘ . Lane? AL. LAM. O41, Sree. 8.58, ayes saab ‘YAM feta : Gt emeed Brees s** oe Le ee ee, De ee “iesdialit tnt aoheomonrt itunes Senay Coble ty wg Palen ened Lcnithi Awnratiie ts fied yf Myo d W45. Mcgee (FY, Aare ikaw, TSM _ HONE, —- Caborinaticres Fu emergent By r * | canta la pL Morigen int 4 | ea i Pen'y aide. yee == Miclopatin vOioant “angler : Caveman shyreed beta Momngray ; -_ Wine 4 12 Shstliieiiemaeaiaaaiiniaei saneiptenenir Nees, Mz, JAdk — Fimacimm lal ne punt pet af Caiotgy Ma vines i a —_ 1h}... ay Teed, RAE. % ibe ae Paci, pul Aye ‘ wens 0 Vion aan “tis was me eon Groupings of Tidal River Systems in Northern Australia Based on Mangrove Species A.G. WELLS (Communicated by D. KEITH) WELLS, A.G. Groupings of tidal river systems in northern Australia based on mangrove species. Proc. Linn. Soc. N.S.W. 115: 135-148 (1995). Eighty two tidal systems encompassing the entire coastline of the Northern Territory and the major portion of the Kimberley region of Western Australia were sam- pled for the presence of mangrove species. Four geomorphologically distinct tidal river systems were identified. Through clas- sificatory analyses tidal systems were grouped on presence/absence of mangrove species. Climatic variations, particularly increasing seasonal aridity in many regions across Northern Australia appear to strongly influence mangrove species diversity and resultant tidal system groupings through the survey area. A.G, Wells, LOA Rosemead Road, Hornsby, NSW 2077; manuscript received 20 September 1994, accepted for publication 19 April 1995. KEYWORDS: estuaries, mangrove ecology, northern Australian mangroves INTRODUCTION The use of a classificatory approach has a long tradition in many areas of natural sci- ence. Methods of classification suitable for examining plant/environment relationships from local to regional scales range from the phytosociological, developed by Braun- Blanquet (1932) to an ever-expanding array of techniques including Association Analysis and other numerical clustering approaches of handling large data matrices (see Mueller- Dombois and Ellenberg 1974). The results of classification using any of these techniques are evaluated in terms of the clarity of the groups produced and the ease with which rational hypotheses can be for- mulated concerning these groups and their environmental relationships. The applica- tions of these approaches to ecology have been reviewed by Williams (1971), Clifford and Williams (1973), Frenkel and Harrison (1974), Clifford and Stephenson (1975) and Sandland and Young (1979). In this paper numerical classificatory techniques are applied to mangrove plant communities through much of the Australian Monsoon region to explore floristic group- ings of the tidal systems from an environmental viewpoint. The potential of these approaches for an understanding of mangrove plant distri- bution is related to the fact that some species appear to be useful indicators of specific habitats. Schimper (1903) was among the earliest to emphasize that many mangrove species are useful biological indicators within intertidal areas reflecting changes in local microtopography, edaphic and climatic conditions. Other workers including Fosberg (1961, 1975), Macnae (1966, 1968), Chapman (1970, 1975, 1976), Zahran (1975) and Cintron et al. (1978) have stressed that increasing climatic aridity results in a reduction of floristic diversity of mangroves in both tropical and sub-tropical regions. METHODS Field work was conducted throughout the period 1975-1979 in 82 tidal systems com- prising 110 tidal waterways across northern Australia. This included 527 km of waterways in the Kimberley and Joseph Bonaparte Gulf regions of Western Australia and a further Proc. LINN. SOC. N.S.W., 115, 1995 136 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA 3998 km of waterways through the Northern Territory - from the Western Australian bor- der to the Queensland border. Fringing riverside mangrove communities were assessed on both banks for species composition and cover abundance at 2.5 km intervals. At each site associations were assayed for 20m deep quadrats stretching 100m along the river. Herbarium specimens were inspected at the Northern Territory Herbarium, Australian Institute of Marine Science, Townsville, James Cook University, Townsville, Department of Forestry Sarawak, Malaysia (Sarawak specimens deposited with Northern Territory Herbarium), Phuket Marine Biological Station, Thailand, University of Papua New Guinea, and Mangrove Research Centre, Forest Research Institute, Philippines for species verifications. Specimens have been deposited with the Northern Territory Herbarium and John Ray Herbarium, University of Sydney. Holotype specimens of new mangrove species resulting from these surveys of Avicennia integra N.C. Duke and Sonneratia spp. also described by Duke (1987, 1988, 1994) are located in the Herbarium of the Northern Territory, Darwin. New combinations in the genus Avicennia have been reported by Everett (1994). Groupings of tidal systems and mangrove species were obtained using the Multbet non-combinatorial information statistic program within the Taxon package (Dale e¢ al., 1980). In this study 24 species recorded in the survey area (Table 1) are included in a matrix analysis of 24 species x 82 sites. TABLE 1 Mangrove species from the survey area used in classificatory analyses. ACANTHACEAE Acanthus ilicifolius L. AVICENNIACEAE Avicennia marina subsp. eucalyptifolia (Valeton) J. Everett Avicennia integra N.C. Duke BOMBACACEAE Camptostemon schultzii Mas. COMBRETACEAE Lumnitzera littorea (Jack.) Voigt Lumnitzera racemosa Willd. EUPHORBIACEAE Excoecania agallochaL. MELIACEAE Xylocarpus australasicus Ridl. Syn X. mekongensis Pierre Xylocarpus granatum King MYRSINACEAE Aegiceras corniculatum (L.) Blanco MYRTACEAE Osbornia octodonta¥. Muell. PLUMBAGINACEAE Aegialitis annulata R. Br. RHIZOPHORACEAE Bruguiera exanistata Ding Hou RUBIACEAE SONNERATIACEAE Bruguiera gymnorhiza (L..) Lamk. Bruguiera parviflora (Roxb.) W & A ex Griff. Bruguiera sexangula (Lour.) Pior. Ceriops decandra (Griff.) Ding Hou Cenops tagal (Perr.) C.B. Rob. var. australis C.T. White Ceriops tagal (Perr.) C.B. Rob. Rhizophora aficulata Blume Rhizophora stylosa Griff. Scyphiphora hydrophyllacea Gaertn. Sonneratia alba}... Sm. Sonneratia lanceolata Blume; Duke and Jackes Proc. LINN. SOC. N.S.W., 115, 1995 A.G. WELLS 137 Study Setting The survey area lies between 11°S and 16°S across the northern coastline of Australia. Climate is strongly conditioned by the seasonal shifting of prevailing winds and marked changes in air-mass properties. Two distinct seasons can be identified - the ‘wet’ season with dominant winds from the north-west to west, occurring from November- March and the ‘dry’ season with prevailing south-easterly winds from May-September. April and October are transitional months (Specht 1985; Southern 1966; Gentilli 1971). Associated with these seasonal changes are significant variations in air tempera- tures, relative humidity, evaporation and precipitation (Bureau of Meteorology 1975). Variation in mean annual rainfall for the survey area and temperature- rainfall diagrams for selected sites are shown in Figures | and 2 respectively. aN RAINFALL (mm) SS . = 128°E 132° 136° ARAFURA MB 500-1750 TIMOR & 8 1,250-1,500 [Pe] 1,000-1,250 oe {"] 750-1,000 fF 500- 750 Carpentaria (52 g, “% Fig. 1. Variation in mean annual rainfall for the study area (all years of record to 1975, Bureau of Meteorology. Within the greater portion of the survey area, from the Kimberley region of Western Australia to Gove in the Northern Territory, semi-diurnal tidal patterns occur. Within the Gulf of Carpentaria diurnal tidal patterns are normal and lunar differences, particularly for spring and neap tides are not nearly so pronounced and may even be opposite in effect to those experienced in regions of the survey area experiencing semi-diurnal tides. Tidal bores occur on some river systems across the northern coastline during spring high tides. Spring tidal ranges vary from up to 11 metres in the Kimberleys, 8 metres in the Darwin region, 3 metres in the Gove region and 2.5 metres along the south-western por- tion of the Gulf of Carpentaria (Australian National Tide Tables). Tidal Systems Four geomorphically distinct estuarine systems were surveyed. These were: (1) those where drainage was structurally controlled, as in the northwest Kimberley region of Western Australia, where entire river courses follow geological jointing planes (e.g., Glenelg, Prince Regent, Roe, Hunter and Mitchell rivers); (2) those waterways of varying lengths which enter embayments after meandering across small or extensive alluvial floodplains; (3) those waterways entering into harbours or ports (e.g. Port Keats, Port Paterson, Bynoe and Darwin Harbours); or (4) short coastal inlets which do not drain any substantial upstream catchment areas (e.g. Mini Miniand Iwalg Creeks, Ilamaryi River). The term tidal system used throughout this paper refers to groups of individual waterways possessing a common sea entrance. The distribution of the tidal systems exam- ined in this study is shown in Figure 3. Individual waterways drain varying catchment areas Proc. LINN. SOC. N.S.W., 115, 1995 138 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA although systems in the Alligator Rivers area may partially interconnect during some wet seasons (Williams 1979). 4. MANINGRIDA 3. DARWIN | DALY RIVER 2. SNAKE BAY N i) is) iS 8 (ww) TIV4NIVE NV3W thi RAINFALL 27s EVAPORATION 8 6. KURI BAY 8 a 8 MEAN TEMP °C 7 KALUMBURU [ r 8. WYNOHAM | 2 ROPER RIVER Fig. 2. Temperature-rainfall diagrams for ten selected sites in the survey area. Solid black area represents the ‘wet’ season. Data by courtesy of the Bureau of Meteorology, years of record to 1975. RESULTS A classification of the data according to MULTBET analysis is shown in Figure 4. Here coherent groups of river systems have been arbitrarily truncated at the 12 group level. These in turn may be profitably lumped into three categories (A,B,C) as shown in Figure 4. Grouping river systems rather than individual waterways as sites eliminates confu- sion due to between waterway variations in species composition found in any particular system. Category A systems represents floristically the least diverse sites (between 414 species). Such systems are seen in this analysis to occur only in the Gulf of Carpentaria and Joseph Bonaparte Gulf, in more seasonally arid portions of the survey area. Systems inter- mediate in their level of floristic diversity (Category B with between 11-16 species) repre- sent systems throughout the Kimberley region of Western Australia and around Joseph Bonaparte Gulf and seawater systems across the northern coastline of the Northern Jerritory with only the Limmen Bight system from within the Gulf of Carpentaria. Category C systems with between 14-21 species are seen to occur only in the least season- ally arid areas across the northern coastline of the Northern Territory (11-12°S lat.) Proc. LINN. SOC. N.S.W., 115, 1995 LYAATVS 28 TAVd, 18 ‘NOLMAN, 08 sNIBLSNIB, 62 NOSNIGOHU 82 SMO1134 LV4 ZZ sOATINVS, 92 SSAGAWIHOUV, SZ AONVW 1d, #2 «HOG, €Z NVAYW4AM 22 AWOVYVS, tL ‘GWO1NOD, 02 YNHLYVOIW 69 LHDIa NAWWIT 89 SNMOL 29 IdNYVAVN 99 YAdOu S9 VdVMIA ¢9 ASOW €9 AVLINAW 29 139 A.G. WELLS LYVH 19 Y3ay1VM 09 SNOLV100% 6S AVG ATUATAW 8S ONNNIHIGONNYNE ZS NHOfP Y3L3d 9S OLVO SS NUNWOHYOD bS Vd IVEOD/YVNINIVHVE €S GOODSVH 2S VONNYVMYVG IS AYAddI1S 0S VIVYNM 6Y IOMNYNMVYEVM 8F IOMYVIVA/NVHONINONE Zp LIVHLS 1130V9 9b LIVHLS NOSNIHOLNH S¢ YSAIY NA 100M tb AQA19 €p LANN3G 2y VILIGHVO LP AY¥OLIYYal NYIHLYON VYNIVAY Vu5voird Ov vunavrd 6E VONVGVGNVON 8€ TISQVO/HLATE ZE NOSNIMWOLMOO0dH3SAIT 9E AIMYVOIVEONNN SE SINVIVW bE fIODNYNM €& YARS0VWOOD Ze VIUdvV LE ONIM OF IAUVAV TI 62 SIVMI 82 INIW INIA 22 VTTIANJSYNW 92 YOLVOITIV LSV3 SZ YOLVDITIV HLNOS ¢2 YOLVOITIV LSAM €2 NVWQOTIM 22 AGIV1s QV 12 OONVONIL 02 NVONOT 6L HLV8/NOLSNHOFP 81 OOONVNVHGNV ZL O1V4sNd 9b YNOSHVH NIMHYVO SL YNOSHYVH JONAS FL NOSHA1LVd LYOd EL AIVd 2 STAOW tL SLV3-y LYOd OL JOIUNVANZLIS 6 VIHOLOIA 8 quo Z YAGNSYYVM LYOd 9 TIBHOLIW S YALNNH ¥ JOU € 4LN39534 SONIYd 2 Y31VM 3DYOsD | VITVELSAY NY3SLSIM \\) @ 419 hy ow/0d0u0g ydasor Fig. 3. Tidal systems of the survey area. Proc. LINN. SOC. N.S.W., 115, 1995 140 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA GROUP 12 ___WILDMAN, WEST ALLIGATOR,SOUTH ALLIGATOR EAST ALLIGATOR,MURGENELLA, DJIGAGILA. 44 ___HABGOOD,GOROMURU,CATO,PETER JOHN. 4 ADELAIDE, BLYTH/CADELL, DARBITLA,GLYDE HUTCHINSON STRAIT,CADELL STRAIT, Cc BUCKINGHAM, KALARWOI. 3 ___GOOMADEER, MAJARIE,NGANDADAUDA WURUWURUWOI,SLIPPERY,BURUNGBIRINUNG 6 ___LIVERPOOL/TOMKINSON, DARWARUNGA, BARALMINAR/GOBALPA. 5 ___ANDRANANGOO, JOHNSON/BATH, DONGAU TINGANOO, KING, NUNGBALGARRIE,BENNETT WOOLEN, MELVILLE BAY. 10 ___GEORGE WATER,BYNOE HARBOUR,DARWIN B HARBOUR, BUFFALO,MINI MINI,ARRLA, WURUGOIJ, DJABURA, KURALA. PRINCE REGENT, ROE,HUNTER, MITCHELL 9 —___PORT WARRENDER, ORD, FITZMAURICE PORT KEATS,PORT PATTERSON,|IWALG,ILAMARY! LIMMEN BIGHT. 8 ____HART, MUNTAK,ROSE, A 7 —___DALY,KOOLATONG, WALKER,ROPER,McARTHUR WEARYAN, ROBINSON. 2 —__‘COULOMB’,’BOHR’.’PLANCK’,FAT FELLOWS, ‘EINSTEIN’,/(NEWTON’,’PAULI’. 1 VICTORIA,MOYLE, YIWAPA,NAYARNPI, TOWNS ‘FARADAY’, ARCHIMEDES’, ’GALILEO’,CALVERT. wee | iter fe tee \i 160 120 80 40 Fig. 4. Dendrogram of site groupings obtained from MULTBET analysis of the eighty two tidal systems in the survey area. /\i is change in value of the information statistic of Dale et al. (1980)). Figure 5 presents the results of a classification of species groups throughout the 82 sites surveyed. 7 species groups have been lumped into three higher-order categories A, B, C, as shown on the dendrogram. The species composition of each of these three groups is shown in Table 2. Category A species were recorded infrequently within river systems across the north- ern coastline of the Northern Territory. They are largely absent from the more arid portions of the survey area. Category B species also occur, in most instances, in systems throughout less seasonally arid portions of the survey area. These species are considerably more common in occurrence than species repre- sented in Category A (Wells 1984). Category C mangrove species are ubiquitous at least in some portion of most tidal systems in the survey area. A two-way table of groupings of tidal systems based on presence/absence of man- grove species is shown in Figure 6. Here it is seen that a group of ubiquitous species (Species-Group 5), comprising Avicennia marina subsp. eucalyptifolia, Excoecaria agallocha, Aegiceras corniculatum, Aegialitis annulata, Osbornia octodonta, Ceriops tagalvar. australis and Proc. LINN. SOC. N.S.W., 115, 1995 A.G. WELLS 141 SPECIES GROUP 6 ___Lumnitzera racemosa; Xylocarpus australasicus Cc Bruguiera exaristata. Avicennia marina subsp. eucalyptifolia Excoecaria agallocha,Aegiceras corniculatum, ~ Osbornia octodonta, Aegialites annulata, Ceriops tagal var australis, Rhizophora stylosa 5 2 ——Bruguiera gymnorhiza . 1 ___Acanthus ilicifolius, Ceriops decandra 7 ___Camptostemon schultzii, Bruguiera parviflora, Sonneratia alba. Lumnitzera littorea, Ceriops tagal, A 4 —Rhizophora apiculata, Scyphiphora hydrophyllacea. 3 Avicennia integra, Xyiocarpus granatum, Bruguiera sexangula, Sonneratia lanceolata. —_—.-$>>—->-—— Li 240 120 Fig. 5. Dendrogram of species groupings obtained from MULTBET analysis of mangrove species recorded for the eighty two tidal systems in the survey area ( iis change in value of the information statistic of Dale et al. (1980)). TABLE 2 Composition of species groups from numerical analysis based on site distributions. Group A Group B Group C Avicennia integra Acanthus ihcifolius Aegiceras corniculatum Bruguiera sexangula Brugutera gymnorhiza Aegialitis annulata Ceriops tagal Bruguiera parviflora Avicennia marina subsp eucalyptifolia Lumuitzera littorea Camptostemon schultzi Bruguiera exaristata Rhizophora apiculata Ceriops decandra Ceriops tagalvar. australis Scyphiphora hydrophyllacea Sonneratia alba Excoecaria agallocha Sonneratia caseolaris Lummnitzera racemosa Xylocarpus granatum Osbornia octodonta Rhizophora stylosa Xylocarpus australasicus Rhizophora stylosa occurs at most sites. The grouping of tidal waterways on the basis of pres- ence/absence of particular species or groups of species presents problems, not least of which is the increased importance of the ‘rarer’ species. Many species often emphasize specific habitats, assuming disproportionate importance and are ultimately responsible for many of the groupings formed in the analyses. The ubiquitous species, however, often contribute little to most groupings. Considerable care is thus required in interpreting some groupings. Proc. LINN. SOC. N.S.W., 115, 1995 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA 142 eqTe eTze1euUOS eroTyrared erarninig TrzaTAyos uouwazsozdued 2 dnos9 eReASsTIeEXa eTarnénrg snotseTeizsne sndzreo0TAx esouaoerl elazzrTuuntT 9 dnol9 PSOTAAS eIOoYdozrTYyYy stTerysne *zrea Tebez sdorra eReTNuUe STIITeETbay eRUOpO}RDO0 ePTUIOgSO un}ZeTNITUIOD seraotTbay eyooTTebe erreoas0oxq @®oo @oo @e@eo OOO @ @e2eee@@ @@00 @ ®@ eyoyndAyeona ‘dsqns eulew eiuuecAy @ G dnol9 PaoeTrAydozpAy eroydrydhos ezernorde ezroydozryy Tebe2 sdorzra) Pe8IORZIT ePrazzruwntT bp dnol9 eyejoeouey &TePLaUuUOS PTnbuexes erarnbénig um}euer6 sndzresoTAx eibaju) BTUUSDTAY ¢ dno19 ezTyzouwAS ezartnbnig Z dno9 eIpuesap sdorra) SNTTOFJTOTTT snyzueoy | dnosg MANGROVE TIDAL SEEGIES SYSTEM Group | 8. VICTORIA 11. MOYLE 64. YIWAPA OO0O00 1e) CcCO000 (ome) OO OO (ome) O0O00 Oo0OO00 O0O0O00 90000 OO0O0°0 O000 OO0O00 Oo000 oO O O00 O00 loo me) @ee@e@ee0 0 @®@e2eee ®@ @eeeee e @@eeeee ®e @®000 @@ ®@ @©eeee@2@ ®@ @©eeeee ®@ O 9 66. NAYARNFI 67. TOWNS OOO @oo OOO QO O [e) oO "FARADAY ' TANG OocoO OOO O0OCO OO0O00 OO (oe) "ARCHIMEDES ' "GALILEO' CALVERT Uexo 76. 82. O00 O0O00 OQ0o0o O0OO00 (oe e) a00 O00 ox Ie (oem Ome (oe) OO000 OOO0O0O O (ooo) Group 2 [om ewe) 000 @0 ® @0oe OO @ OO ® O0O00 O0O00 o0o0°0o OO000 O0OO0°0O ce) [oeme) OO (ome) Cc Oo ome) OO "COULOMB ' "BOHR' 70. O oO U2 74. 77. Ue 80. 8l. [oO me) (oe e) COO OO00Q (oe) O0OO00 "PLANCK' OO0O00 OOOO O000 O oO FAT FELLOWS "EINSTEIN' "NEWTON ' "PAULL! Group 3 32 GOOMADEER 34. MAJARIE OO ®@ OO ® @@@0 6 @ @ O000 O0O00 OOO00 oO Oo O000 O000 @e@eoee@ ee @0oe ome O0o@00 OO0O00 O0O00 @ O ce) 0 @ ®@ @ OoO0O00 O0O00 O@00 OOO0O°0O 38. NGANDADAUDA 48. WARAWURUWOL 50. SLIPPERY Sie oO O0OO00 e) 0o@o00 @®eee eee O0OO00 @ e@ @ @ @ @ e @ @ @ @ e @ @ @ ® @ @ O @ @ @ @ @ @ e @ Oo e) oO oO ie) Oo O e) e) O Oo (@) @) ) @ Oo @ @ @ @ @ z 3 & ra) ae 2 On Sty fQ os Oz N O000 @eeeee e @ @®ooo @ooo @®ooodo O000 O000 BLYTH/CADELL DARBITLA Sle OOO0O0 OoO0O00 41. 43. GLYDE @ooo O000 45. HUTCHINSON STRAIT 46. CADELL STRAIT e @ e e @ (e) @ @ e @ e@ e @ @ @ e@ e@ @ @ e e 6 @ @ @ ® Oo @ (@) @ Oo @) (@) @) Oo e@ Oo e) Oo @ (@) @) @ e@ oO e @ e@ H g ina 3 2 eee SINS ae. 3 4 O86 BO x — , ~ © > ~—wT @®0o0@ee ® OO0°0O © om ) 18. JOHNSTON/BATH 19, DONGAU 0@e@ OO @ @ O0O00 O0OO00 OOOO0O O0OO0O0O O0O0O00 OOO0O°0 OoO@e@o°0d 0@e@e0 0O0@@ 20. TINANGOO 30. KING 0O0@0 NUNGBALGARRIE BENNET 35. 42. O0O0@0 @oee 44. WOOLEN RIVER e@0@ee 58. MELVILLE BAY Group 6 36. LIVERPOOL/TOMKINSON ® ® 5. ® @®ooo @eeoe ® O000 @e0e® @@eoe# DARWARUNGA ® O000 BARALMINAR/GOBALPA @ I3\6 Fig. 6. Continued on next page. 115, 1995 Proc. LINN. SOC. N.S.W., 143 A.G. WELLS ege eijesauUOS eloyiased esainbnug HZyNYIS uowa}sajdweD @ 2 dnoi9 ejejsuexa esaininig @ snaisejesisne sndieo0jAx @ esOWwaoes BJaZJIUWN] @ Q dnol9 esojAjs BsOYdoZ/YyYy sijesisne “eA /25e) sdoag pyejnuue sijelbay B}]UOPO}IO B/UJOgSO wnjejnaiujoo sesaoibay eyooyjebe e1seD909XF pyoudxjeona “dsqns Bue BILU@OIAY@® G dno19 zaoeR|AydospAy esoydiydAIS ejyejnoide esoydoziyy jebe) sdoiad BasO}}I} BLAZHJUWNT p dnoJj9 BIPjOSOUe] BI]eEJQUUOS einduexas esainbnig wnjeues6 sndiecojAX eiBejul BLUUaDIAY ¢ dnoi9 eziyiouwAb evaininig Z dnol9 espuedap sdolaD SMOp/dI1 SnujUeDY (Oe) @0o @®@o0e0 @ @®@eeee@ @ @@e@eeee @ @®@eeeee0 O000 OO000 oO ® @ ® @0o @o @o @0o @o [oexe) (ome) DALY 12. 59. 60. 65. OO O00 OORT e) (oe me) @ ee @@ @0 ®@ @®oo e @ @ @ ® @ @ ® @e @ “@ e @ @ ® (2) @ ® @ @ @ @ @@ @ ® @®@e0e @e@e@e ® ®@ @ 6 @ @e@ @ O @ @ @@e @ee @@ ®@ @ee @ee0e Oo @ @ @ee O ®@ @®@ee @ @e0e @@ee @ee e@ee0e @®@ee @ @ee Proc. LINN. SOC. N.S.W., 115, 1995 @©@eee000 @®@eee eee @@eoeee @®@0ee80 0 @®@08@@ ®@ ®@ @@eoeodood ®@ @®eeeeesd @©@eee80 0 @eee8 @ @ @©@ee@0e80 80 @@0e8620@@ ® @©@ee808@@ ®@ @@08 6 @ ®@ ®0@0@@2®@ ®@ @@eee@e0e@ @©ee6e€6@0@ 0 @@e0eee@ @ @©e@e@ee0e00 @eeeeee @@e@eeee ®@ @©@e@ee0ee8 @ Ce et et @@eeeee @ @©@eee8@ ®@ @eee08e @ @ @@eee28 @ @e@@8e@e8e@ @@e@08@ ®@ @@eoee@e @®@ee0e028 ®@ @@eeoe ®@ @eeo0e0 0 @eeeee0@ @©e@@e@e8 0 @eee@802 @ @©e@eeeee e@ @@eeeoeee@ @ OO000 000 ®@ 0000 O000 OO000 O000 000 ®6 000 ®@ @®00 ® O000 O000 O0O00 O000 O000 O000 O000 O000 Oo000 O000 OO0O00 O000 O0O00 000 ®@ 000 ®@ OO0O00 Oo000 O000 O00 ®@ OO000 O0O00 O0o@@o oe00 O@e@eo0 00 @ ®@ O000 O000 OO000 O000 OO0O00 O0O0O000 O0O00 O000 O000 9000 O000 O000 O000 O000 (ome e me) O@00 O000 OoO000 Oo00°0 O0O0°0 O0O00 O000 O000 O000 (oom ene) O000 O000 OO000 O000 O0O00 @ooo O0OO00 O000 Oo000 O0O000 O0O00 000 ®@ 000 ® o0o0O@@ OO @ ®@ ®@®o00e @®o0o0oe @®o00e @o00oe @®o0oe @®ooo O 1@) oO Oo @ @ oO O oO oO oO O O oO fe) ce) oO O oO oO @ @ oO oO oO oO @ @ @ @ oO oO ® oO [@) Oo oO (ome) [oee) (ome) OO oe) oe) [oeze) OO (ome) oe) (ome) oxo) (ooze) loee) [oee) (ooo) oO ®@ O ®@ (ome) ox ) 0 @ om ) 0 ®@ O ®@ @@ @ ® @o Ct ) ee OO @o @@ SOUTH ALLIGATOR EAST ALLIGATOR DJABURA KURALA Group || HABGOOD GOROMURU CATO PETER JOHN Group I2 WILDMAN WEST ALLIGATOR MURGENELLA DJIGAGILA PRINCE REGENT ROE PORT WARRENDER ORD DARWIN HARBOUR BYNOE HARBOUR BUFFALO PORT PATERSCN IWALG GEORGE WATER MINI MINI WURUGOIJS LIMMEN BIGHT FITZMAURICE Group 10 KOOLATONG WALKER McARTHUR WEARYAN MITCHELL PORT KEATS ILAMARYI ROBINSON ROPER Group 8 HART MUNTAK ROSE Group 9 HUNTER 3. 4. 5. 6. 7. 9. 10. 13. 28. 29. 68. 1. 14. 15. 52. 54. 55. 56. 22. 23. 24. 25. 26. 40. 2. Fig. 6. Two-way table of site /species groups for the eighty two tidal systems in the survey area. Group numbers refer to multbet analysis, given for tidal systems in Figure 4 and for species in Figure 5. 69. 72. 78. 6l. 62. 63. 16. 27. 31. 33. 39. 49. 144 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA DISCUSSION There is a gradual decline in mangrove species richness southwards on both the east and west coasts of Australia (Saenger et al., 1977; Semeniuk et al., 1978; Love, 1981; Bunt, Williams and Duke, 1982; Wells, 1983; Duke, 1992; Semeniuk, 1993; Adam, 1994). At sites across the northern coastline of the Northern Territory (latitudes 11-13°S), 27 mangrove species have been recorded (Wells 1983). However, between latitudes 13 — 16°S, only 10- 14 mangrove species are recorded in the area described in this paper. Such a reduction in species richness within only a few degrees of latitude is unusual and is not considered to result from latitudinal sifting, as on the east Australian coastline within an identical latitu- dinal range there is no decline in species richness (cf. Dowling and McDonald 1982). Smith and Duke (1987) have devoted considerable attention to physical determi- nants of inter-estuary variation in mangrove species richness for northern Australia and utilizing data from the surveys reported here and their own data for the east Australian coast have concluded that estuary length, size of the surrounding catchment, rainfall vari- ation and frequency of tropical cyclones have significant effects on species richness down the east Australian coastline, but not for mangrove forests throughout the Gulf of Carpentaria, and the northern coastlines of the Northern Territory and Western Australia. Reasons for species diversity are undoubtedly complex; however, their claim that estuaries with larger tidal amplitudes have fewer species than estuaries with smaller tidal ranges is not borne out by a simple inspection of Fig. 6 which shows many groupings of tidal systems (cf Groups 1, 7, 9, 10) — with similar species diversity — to include estuar- ies with both large and small tidal amplitudes [Australian National Tide Tables 1976+, Messel et al. (1979-82)]. In other cases many of the most floristically diverse estuaries cf Andranangoo, Goromuru, Habgood, Peter John, and Cato occupy extremely small catchments. Consideration of the Site - Groups derived from the 82 tidal systems presents a tighter picture of floristic similarity and variation between latitudes and different climatic environments throughout the survey area. In grouping together all tidal waterways enter- ing into a particular trunk stream, floristic variation between waterways in a particular system are profitably eliminated. Site-Group 1 (Fig. 4) includes the Victoria, Moyle, Towns and Calvert systems, as well as several short coastal systems along the south-western shores of the Gulf of Carpentaria. As discussed by Wells (1984), all of these systems lie within extremely season- ally arid regions of the survey area and this is evidenced by a marked reduction in species diversity. Site-Group 2 includes several more coastal systems along the south-western shores of the Gulf of Carpentaria. From an inspection of Figure 6 it is seen that this group is floris- tically most similar to Site-Group 1. As these systems also lie in seasonally arid regions of the survey area (Fig. 1 and 2), there is considerable merit in fusing Site-Groups | and 2 as systems of low species diversity. Site-Groups 3, 5 and 6 are themselves floristically most similar to each other (Fig. 4 and Fig. 6), and includes systems occurring across the northern coastline of the Northern Territory only (Fig. 3). Sites within these groupings possess between 15-21 species, although it is apparent that irregular occurrences of less commonly recorded species (in Species-Groups 2, 3 and 4 of Fig. 6) do not in any way distract from the overall pattern and groupings. Floristically, Site-Groups 3, 45 and 6 have on average at least twice the number of mangrove species as Site-Groups | and 2. Site - Groups 7 and 8 (Fig. 4) are also closely related floristically to Site-Groups 1 and 2, and although considerably less diverse in species than Site-Groups 3, 4, 5 and 6 have sep- arated principally on irregular occurrence of Acanthus ilicifolius, Bruguiera gymnorhiza, Camptostemon schultz, Bruguiera parviflora, Lumnitzera littorea and Scyphiphora hydrophyl- lacea. Increasing levels of seasonal aridity in these regions of the survey area maybe largely responsible for irregular occurrences of the latter mentioned species. Proc. LINN. SOC. N.S.W., 115, 1995 A.G. WELLS 145 Site-Groups 9 and 10, also quite similar, separate from each other principally on occurrences of Ceriops decandra, Bruguiera gymnorhizaand Scyphiphora hydrophyllacea— pre- sent only in Group 10. Increasing levels of seasonal aridity in these regions of the survey area are considered to be largely responsible for irregular occurrences of the latter men- tioned species. These groupings include all systems from the Kimberley, in Western Australia, and some sites from Joseph Bonaparte Gulf and the northern coastline of the Northern Territory (Fig. 6.). Site-Groups 11 and 12 represent systems, with the exception of Djigagila, which occur either within Arnhem Bay or Van Diemen Gulf on the northern coastline of the Northern Territory where the regionality of these two groups is dramatically shown. Floristically these systems are most similar to each other and have recorded between 14-20 mangrove species. The occurrence of Sonneratia lanceolata is considered a major distin- guishing species for both these groups from other groupings in this analysis and the addi- tional presence at all sites in Group 12 only of Avicennia integra has resulted in the splitting of Group 11 from Group 12. The groupings shown in Fig. 4 and Fig. 6 have been made principally on sporadic occurrences of many of the less frequently observed mangrove species in the survey area. Wells (1984) has shown that many mangrove species are disadvantaged in their abil- ity to colonize all regions in monsoonal north Australia due to the presence of unfavourable currents during the wet season, the period when most fruiting occurs. Nevertheless, during cyclonic storms surface current directions are often reversed for considerable periods (Aust. Pilot Vol. 5,). Therefore floating fruits and hypocotyls of those mangrove species with restricted discontinuous distributions across the northern coastline of the Northern Territory could eventually reach all waterways of the survey area. That all species recorded in this study do not occur somewhere within all tidal sys- tems is of considerable interest. The three major site categories (A, B, C) for the 82 tidal systems appear to reflect local climatic variations across northern Australia. Sites occurring in category A (Fig. 4) show the lowest level of floristic diversity. They occur only within the Gulf of Carpentaria and Joseph Bonaparte Gulf. However, although the degree of seasonal aridity has been shown by Wells (1984) to be greater in these areas than elsewhere, such aridity is not the only factor responsible for much of the low diversity of mangrove species(Smith and Duke, 1987). The Daly River drains a large catchment area (51,800 km’) and remains fresh above 40 km throughout most years (Messel et al. 1979-82). The resultant lack of extensive periods of brackish water inundation on this system has resulted in only 10 species being recorded. In fact, the Daly River System, although occurring in what has been shown in Fig. 1 as the wettest region, has amongst the lowest diversity of mangrove species for any site in the survey area. Rhizophora stylosa, a species which is fairly ubiquitous at most other sites, is only represented here by an occasional shrub, while absence of Cenops tagal var. australis is most likely related to the nearly perennial freshwater inunda- tion of sites. Macnae (1966), in particular, has pointed out that Cenops tagalvar. australis is often absent or infrequently observed on what are considered ‘continually wet’ portions of coastlines, although in the Daly River region there are distinct ‘wet’ and ‘dry’ seasons. Tidal systems occurring in Category B (Fig. 4) are distributed almost entirely along the coasts of the Arafura and Timor Seas and are intermediate in floristic diversity between Categories A and C. Low species richness for systems within the Kimberley region of Western Australia (up to 14 species) and from Joseph Bonaparte Gulf (up to 12 species) appears to be largely a response to dry season aridity experienced in these regions. However, lack of suitable sites for colonisation by mangrove species due to rapidly rising land gradients, macro-tidal fluctuations (up to 11 metres) in many tidal waterways of the Kimberleys and unfavourable surface sea currents during fruiting periods may also con- tribute to low species richness (Wells, 1984). Proc. LINN. SOC. N.S.W., 115, 1995 146 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA Other river systems in Category B, such as Port Paterson, Bynoe and Darwin Harbours, Buffalo, Mini Mini, Iwalg, Arrla, Wurugoij and Djabura Creeks; Ilamaryi and Kurala Rivers, although occurring across the northern coastline of the Northern Territory —a less seasonally arid area— are also quite poor in mangrove species. All these tidal systems experience seawater salinities for most of the year and in some cases even become quite hypersaline by the end of the dry season (Messel et al. 1979 — 82). Such saline conditions could limit establishment of some mangrove species. The Limmen Bight System is included within Category B because of the presence of occasional shrubs of Camptostemon schultz. This species has not been recorded from any other system sur- veyed down the east coast of Arnhem Land in the Gulf of Carpentaria. Reasons for the absence of Camptostemon schultzii in other systems in the Gulf of Carpentaria are unknown although it is considered that its absence here and the extreme rarity of the species at sys- tems around Joseph Bonaparte Gulf may be related to a slightly higher level of seasonal aridity and more frequent occurrences of desiccating south-east trade winds and lower night temperatures during the dry season. Camptostemon schultziiis extremely abundant in mangrove swamps throughout the Kimberley region of Western Australia, which although experiencing considerable seasonal aridity, is largely protected from the south- east trade winds (blowing across from the interior of the Australian continent) by the high cliffs (300 metres) abutting mangrove swamps (Wells, 1981). Such rock outcrops may also act as microclimatic heat sinks raising dry season night temperatures and might be partial- ly responsible for some restricted distributions of mangrove species in the Kimberley (cf Xylocarpus granatum). Category C systems are floristically most diverse with 14-21 species. Such systems occur only across the northern coastline of the Northern Territory (11-12°S latitude) and are in regions receiving high annual precipitation with most also receiving at least a mini- mal amount of precipitation throughout ‘dry’ season months. For Category C sites, mean annual relative humidities are generally higher and mean annual evaporation losses much lower (Bureau of Meteorology, 1975) than for sites occurring throughout the Gulf of Carpentaria, Joseph Bonaparte Gulf and Kimberley regions. It is considered that the species/site groupings given in Fig. 6, in particular, are greatly influenced by climatic vari- ation within the survey area. Many other factors have been noted as influencing floristic diversity of mangrove species within tidal waterways across northern Australia (Wells, 1984; Smith and Duke, 1987; Semeniuk, 1993). However, only when biogeographic, microclimatic and edaphic data for individual sites within these mangrove communities are available will a better understanding of known variations in species distributions in this survey area be attainable. ACKNOWLEDGEMENTS I thank in particular Professor Harry Messel C.B.E., former Head of the School of Physics, University of Sydney, for assistance both in the field and in providing the consid- erable logistical and financial support (principally through the Science Foundation for Physics) that enabled all surveys along the northern coastline of Western Australia and the Northern Territory to be undertaken. The field assistance of all members associated with the University of Sydney-Northern Territory Conservation Commission Joint Crocodile Research Project, and in particular Fred Duncan is gratefully acknowledged. Dr W.T. Williams, Australian Institute of Marine Science, Townsville provided computational assistance and Drs Peter Myerscough and Bill Allaway of the School of Biological Sciences, University of Sydney provided critical appraisal of the manuscript. Thanks to Ms Kim Mawhinnew who typed the manuscript. Proc. LINN. SOC. N.S.W., 115, 1995 A.G. WELLS 147 References ApaM, P. 1994. Saltmarsh and mangrove. Pp. 395-425 in: “Australian vegetation” (R.H. Groves, ed.). Cambridge University Press. AUSTRALIAN NATIONAL TIDE Tables. 1976+. Australian Hydrographic Office Publication. Australian Government Publishing Service, Canberra. AUSTRALIAN PILOT. 1972. North, north-west and west coasts of Australia from the west entrance of Endeavour Strait to Cape Leeuwin, volume 5 page 21. Hydrographic Dept., Ministry of Defence, U.K. BRAUN-BLANQUET, J. 1932. Plant Sociology. Transl. and ed. by G.D. Fuller and H.S. Control. McGraw Hill Book Co,. New York. BUNT, J.S. and WILLIAMS, W.T. 1980. Studies in the analysis of data from Australian tidal forests. 1. Vegetational sequences and their graphic representation. Australian Journal of Ecology 5: 385-90 BUNT, J.S., WILLIAMS, W.T., and Duke, N.C. 1982. Mangrove distributions in north-east Australia. Journal of Biogeography 9: 111-20. BUREAU OF METEOROLOGY. 1975. Climatic averages, South Australia and Northern Territory Dept. of Science and Consumer Affairs, Australian Government Publishing Service, Canberra. CHAPMAN, V.J. 1970. Mangrove phytosociology. Tropical Ecology 11: 1-19. CHAPMAN V.J. 1975. Mangrove biogeograhy. Pp. 1-22 in: “Proceedings of the International Symposium of Biology and Management of Mangroves, Volume 1”, (G.E. Walsh, S.C. Snedaker and H.J. Teas, eds), Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida USA. CHAPMAN, V.J. 1970. Mangrove vegetation. J. Cramer, Vaduz, 447pp CINTRON, G., LuGo, A.E., POOL, D.J. and Morris G. 1978. Mangroves of arid environments in Puerto Rico and adjacent islands. Biotropica 10: 110-21. CLIFFORD, H.T. and STEPHENSON, W. 1975. An introduction to numerical classification. Academic Press, New York, 229 pp. CLIFFORD, HT and WILLIAMS, W.T. 1973. Classificatory dendrograms and their interpretations. Australian Journal of Botany 21: 151-162. DALE, J., HAIN, D., LANCE, G., MILNE, F., Ross, D., THOMAS, M. and WILLIAMS, W. 1980. Taxon users manual. D.R.C. Manual No. 6, Commonwealth Scientific and Industrial Research Organization. Division of Computing Research, Canberra. Dow Linc, R.M. and McDonaLp, TJ. 1982. mangrove communities of Queensland. Pp. 79-93 in “Mangrove eco- systems in Australia. Structure, function and management”, (B.F. Clough, ed.), Australian National University Press. Duke, N.C. 1988. An endemic mangrove species Avicennia integra sp.nov. (Avicenniaceae), northern Australia. Australian systematic Botany 1: 177-80. Duke, N.C. 1992. Mangrove floristics and biogeography. Jn “Tropical Mangrove Ecosystems” (A.I. Robertson and D.M. Alongi, eds.). American Geophysical Union, Washington D.C. Duke, N.C. 1994. A Mangrove hybrid Sonneratia xurama (Sonneratiaceae) from Northern Australia and southern New Guinea. Australian systematic Botany 7: 521-526. Duke, N.C. and JAckEs, B.R. 1987. A Systematic revision of the mangrove genus Sonneratia (Sonneratiaceae) in Australasia. Blumea 32: 277-302. EVERETT. J. 1994. New combinations in the genus Avicennia (Avicenniaceae). Telopea 5: 583-794. FosBERG, F.R. 1961. Vegetation-free zone on dry mangrove coasts. United States Geological Survey: Professional paper 424 (d): 216-18. FosBERG, F.R. 1975. Phytogeography of Micronesian mangroves. Pp. 23-42 in “Proceedings of the International Symposium on Biology and Management of Mangroves (G.E. Walsh, S.C. Snedaker and H.J. Teas, eds.). Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida. FRENKEL, R.E. and HARRISON, C.M. 1974. An assessment of the usefulness of phytosociological and numerical clas- sificatory methods for the community biogeographer. Journal of Biogeography 1: 27-56 GENTILLI, J. 1971. Climates of Australia. World Survey of Climatology 13: 4-7. Love. L.D. 1981. Mangrove swamps and salt marshes. Pp. 317-34 in: “Australian Vegetation” (R.H. Groves, ed.). Cambridge University Press. Macnag, W. 1966. Mangroves in eastern and southern Australia. Australian Journal of Botany 14: 67-104. Macnag, W. 1968. A general account of the fauna and flora of mangrove swamps and forests in the Indo-West Pacific region. Pp. 73-270 in “Advances in Marine Biology, Volume 6” (Russell, F.S., ed.). MESSEL, H., VORLICEK, G.C., WELLS, A.G. and GREEN, W.J. 1979-82. River systems in the Northern Territory of Australia and their crocodile populations. Series of seventeen monographs, Pergamon Press, Sydney. MUELLER-DOMBOIS and ELLENBERG. 1974. Aims and Methods of Vegetation Ecology. Wiley, New York. SAENGER, P., SPECHT, M.M., SPECHT, R.L. and CHAPMAN, V.J. 1977. Mangal and coastal salt-marsh communities in Australia. Pp. 293-345 in: “Ecosystems of the world” (V.J. Chapman, ed.). Volume 1, Wet coastal eco- systems. Elsevier, Amsterdam. SANDLAND, R.L. and YOUNG, P.C. 1979. Probabilistic tests and stopping rules associated with hierarchical classifica- tion techniques. Australian Journal of Ecology 4: 399-406. SCHIMPER, A.F.W. 1903. Plant geography upon a physiological basis. Translated by W.R. Fisher. Clarendon Press, Oxford. SEMENIUK, V., 1993. The mangrove systems of Western Australia. 1993 Presidential Address. Journal of the Royal Society of Western Australia 79: 99-122. SEMENIUK, V., KENNEALLY, K.F., and WILSON, P.G. 1978. Mangroves of Western Australia. Western Australian Naturalists’ Club. Handbook No. 11, Perth. 89 pp. Proc. LINN. SOC. N.S.W., 115, 1995 148 TIDAL RIVER SYSTEMS IN NORTHERN AUSTRALIA SMITH, T.J. and DUKE, N.C. 1987. Physical determinants of inter estuary variation in mangrove species richness around the tropical coastline of Australia. Journal of Bioseography 14: 9-19. SOUTHERN, R.L. 1966. A review of weather disturbances controlling the distribution of rainfall in the Darwin-Katherine region, Northern Territory. Bureau of meteorology of Australia working paper, 65: 3203. SPECHT, R.L. 1958. The climate, geology, soils and plant ecology of the northern portion of Arnhem Land. Pp. 333- 414 in: “Records of the American Australian Scientific expedition to Ahem Land. 3. Botany and plant ecology” (R.L. Specht and C.P. Mountford, eds.) . Melbourne University Press, Melbourne. WELLS, A.G. 1981. A survey of riverside mangrove vegetation fringing tidal river systems of Kimberley, Western Australia. Pp. 94-121 in: “Biological survey of Mitchell Plateau and Admiralty Gulf, Kimberley Western Australia. Part 3”. Records of the Western Australian Museum, supplement. WELLS, A.G. 1983. Distribution of mangrove species in Australia. Pp. 57-76 in: “Tasks for vegetation science” (H.J. Teas, ed.). Dr. W. Junk Publishers, The Hague. WELLS, A.G. 1984. Mangrove vegetation in northern Australia, PhD University of Sydney. WILLIAMS, A.R. 1979. Vegetation and stream pattern as indicators of water movement on the Magela floodplain, Northern Territory. Australian Journal of Ecology 4: 239-47. WILLIAMS, W.T. 1971. Principles of clustering. Annual Review of Ecology and Systematics 2: 303-26. ZAHRAN, M.A. 1975. Biogeography of mangrove vegetation along the Red Sea coasts. Pp. 43-51 in “ Proceedings of the International Symposium on Biology and management of Mangroves” (E.G. Walsh, S.C. Snedaker and H.J. Teas, eds.). Institute of Food and Agricultural Sciences, University of Florida, Gainsville, Florida. Proc. LINN. SOC. N.S.W., 115, 1995 PROCEEDINGS of the LINNEAN SOCIETY NEW SOUTH WALES The following papers are research reports dealing with natural history in all its branches. The Society invites authors wishing to publish in these proceedings to contact the secretary for details of procedures and format. VOLUME 115 © Linnean Society of New South Wales Proc. LINN. SOC. N.S.W., 115, 1995 PRESIDENTIAL ADDRESS 1993 Mangrove Macroalgae: A Review of Australian Studies ROBERT J. KING [Delivered 24 March 1993] KING, R.J. Mangrove macroalgae: a review of Australian studies. Proc. Linn. Soc. N.S.W. 115: 151-161 (1995). Algae associated with mangroves have received particular attention in phycology since Dr. Erika Post recognised the distinctive nature of the macroalgal community and began a series of papers (1936-1968) on their systematics and distribution. This address reviews Post’s studies on the taxonomy and systematics of the algae of the Bostrychia- Caloglossa association especially as they relate to Australia, and then reviews the contribu- tion of Australian studies on the taxonomy, ecology, ecophysiology, distribution and biogeography of mangrove macroalgae. Robert J. King, School of Biological Science University of New South Wales, Sydney NSW 2052. INTRODUCTION The recognition of mangrove macroalgae as a distinct field of study within phy- cology dates from Post (1936a) and her major paper on the systematics and biogeography of what she termed the Bostrychia-Caloglossa association. Her use of the term association was general, in the sense of an assemblage, rather than in the rigorous sense it is now used in ecology. Since that time there has been a wide recognition of the distinctive nature of these algal communities with numerous papers by Post and others on floristics. Partly this reflects general trends in algal ecological studies since the 1970s which have placed increasing emphasis on algae of estuaries and soft bottom shores rather than those of rocky shores, on production biology, and studies of tropical ecosystems (King 1990a). In this review, a number of different types of studies will be considered: taxonomy and sys- tematics, floristics and biogeography, local distribution, productivity and biomass, and ecophysiology especially in relation to salinity. Given the pivotal role played by Post in her series of papers (Post 1936a et seq.) itis appropriate to consider her studies first. THE CONTRIBUTION OF ERIKA POST Erika Post’s entire scientific career was concerned with the study of the macroalgae associated with mangrove and saltmarsh communities, and she published some 34 papers on the subject over the period 1936-1968. All of her papers were single author publica- tions and with one exception written in German. A full list of her publications is given in the bibliography. Post’s first and major contribution (Post 1936a) presented the results of the algal collections of the 1929 German Scientific Expedition in the Sunda region and was based on collections of Prof. W. Troll. Recognising the need for a reassessment of the taxa involved she undertook a complete revision of three major taxa Bostrychia, Caloglossa and Catenella. Later papers extended this approach but with an emphasis on distributional studies. Caloglossa later received further detailed attention (Post 1943) and two new species of Bostrychia were also described (Post 1939a, 1941). A major review of work was Proc. LINN. SOC. N.S.W., 115, 1995 152 MANGROVE MACROALGAE published in 1963. Post lived in Kiel (Schleswig-Holstein, Germany) but had no formal attachment with either the Marine Sciences Institute (Institut fur Meereskunde) Kiel or the Botanical Institute of the Christian Albrechts Universitat in Kiel. Despite this limitation Post almost single handedly promoted a global view of the study of the taxonomy, morphology and geographic distribution of mangrove algae. It was she who coined the terms ‘ Bostrychietum’ and ‘Bostrychia — Caloglossa Assoziation’. In retrospect it is easy perhaps to criticise certain aspects of Post’s taxonomic works. She appears to have given little attention to typification of species and failed to adhere strictly to the rules of taxonomic nomenclature. In addi- tion she was unable to fully appreciate variation within taxa. However, as King and Puttock (1989) have pointed out for Bostrychia, Post was consistent in her use of names throughout her papers and hence it is comparatively easy to correlate her species names with more recent taxonomy. Post’s major contributions were concerned with the systematics and distribution of mangrove algae, although she did report ecological observations albeit based on herbari- um labels and correspondence from collectors. Unfortunately many of her later papers are little more than extensive lists of species and localities based on such sources. The fact that data are often presented in the form of a compilation rather than a synthesis and in German means that these valuable data remain relatively under-utilised. Some of the later papers, however, do address particular biological issues such as the ability of typical man- grove algae to withstand desiccation (Post 1963c) and to grow ata range of salinities in the natural environment (Post 1963d). From the beginning Post was in contact with phycologists and collectors around the world and maintained a prodigious correspondence. In her first paper she thanked 95 individuals for their help in providing literature and material. That list reads like a ‘Who's Who’ of phycology at the time. Included are Prof. A.H.S. Lucas (Girrahween), Prof. E. Nicholes (Crawley) and Prof. F.J. Rae (Melbourne) from Australia, and Dr. L.M. Cromwell (Auckland) and Dr. R.M. Laing (Christchurch) from New Zealand. Post never visited Australia or New Zealand, nor would it appear ever made any field collections in mangroves. Nonetheless on the basis of her extensive correspondence she wrote a num- ber of papers about algae in specific areas: in the Australasian region, New Zealand (Post, 1936b, 1966b) and Australia (Post 1964c). The last of these papers, somewhat oddly enti- tled ‘Bostrychietum aus dem National Park von Melbourne’, is based on collections made by Sophie C. Ducker at Wilsons Promontory, Victoria (with comparisons made with earli- er collections of among others A.B. Cribb, F. von Mueller, H.T. Tisdall and H.B.S. Womersley) and a collection by Dr. J.A. Carnahan from the Minnamurra River (with com- parisons made to Valerie May’s collections in the Georges River estuary). Post continued her correspondence throughout the 1960s collecting both litera- ture and specimens. An example of the former is the following request to R.H. Anderson, Royal Botanic Gardens and National Herbarium, Sydney and dated 21st July 1961: ‘Dear Mr. Anderson, I try now already a very long time to get some papers of AHS Lucas of which I only would be interested in the verbally text (& figures?) for all what is said about Bostrychia (= Amphibia) , Caloglossa (= Delesseria leprieurii) and Catenella (& Murrayella) and I would be very grateful to you for the corre- sponding photocopies which of course I will pay. The papers in questions are: 1) Lucas and Perrin F.: The seaweeds of South Australia Part II. Red Seaweeds. Handbook of the flora &fauna of South Australia, Adelaide, p. 111-458, 202 figures (Are they also mentioned in the ‘Introduction’ in Part I, 1936?) 2) Lucas A.H.S.: Marine Algae in ‘British Association for the Advance- ment of Science) Handbook for N.S. Wales’, p.459-463. Proc. LINN. SOC. N.S.W., 115, 1995 R.J. KING 153 I have no idea, if there are some more recent publications in which my genera are mentioned. In the hope of your kind help believe me. Yours faithfully, EE osts Her requests for algal material were quite specific and with details such that a non- phycologist might respond appropriately. In a letter (17 May 1965) to Dr. Knowles Mair [Director of the Royal Botanic Gardens, Sydney 1964-1970] to thank him for recent corre- spondence she noted: ‘Of course I would also be interested in N.S.W. (dried material of the small blackish-purple mosslike mangrove algae (a) feathery, b) leafy, c) rosary like) which grow (mostly) in thick clusters on dead stumps, tree bases and the upper portions of the pneumatophores, uncovered between tides, if possible from any locality except Hawkesbury River, Kirribilli Point, Ryde, Abbotsford, Gladesville, Cook’s River, Georges river and Minnamurra River estuary. Please leave the tufts just as they are (unsorted). Believe me, Yours faithfully, E. Post?’ Predictably it was Valerie May [Jones] (King and Briggs 1988) who answered both of these requests noting when posting the collected materials (22/9/65) that: ‘Many of our mangrove areas are being destroyed, bridges, embank- ments, parking lots replace former collecting areas and even those collections I have been able to make are often from ecologically damaged areas, suffering from much silting up. Some areas, such as Georges River Bridge, where I collected material for you some years ago, have no signs of mangroves left now, only stone embankments. In other areas such as at Roseville Chase, Sydney, the mangroves still remain after bridge building but they now stand in a firm substrate and there is no trace of their characteristic algal flora, — oyster shells replacing it. Several areas from which I have collected this time are so silty — due to man’s activity — that I expect the algal flora to disappear relatively soon’. Whether or not that algal flora has gone, what has disappeared are the extensive col- lections which formed the basis of Post’s work. As already noted Post had no formal rela- tionship with any scientific institutions. Following her death in 1980 her apartment was cleared and only late in that process were staff of the Marine Sciences Institute, Kiel involved. Her reprints and papers on mangrove algae of the southern hemisphere are now in my collection. Included in her possessions was a small compound microscope of unknown manufacture about which Mr. R. Oldfield (pers comm.) has provided the fol- lowing information. “This compound drum microscope was probably made in Britain between 1840 and 1890 and could equally as well have been sold in a toy shop as by an opti- cian. Cost: 16s Od retail/9s 6d wholesale. It had three single screw on lenses which could be combined”. Whether or not this instrument was Erika Post’s research microscope, her lack of facilities was real, and her major contribution in mangrove algal studies is there- fore especially noteworthy. FLORISTICS Post (1936a et seg.) drew attention to the characteristic algal flora associated with mangroves, the Bostrychia — Caloglossa— association consisting of Bostrychia (now including Stictosiphonia — see King and Puttock 1989), Caloglossa, Catenella and Murrayella. Murrayella and Catenella are by no means as widely distributed or abundant as the other Proc. LINN. SOC. N.S.W., 115, 1995 154 MANGROVE MACROALGAE genera which almost invariably occur with mangroves and are also commonly associated with saltmarsh vegetation. The red algae associated with mangroves in Australia are listed with their distributions in King and Puttock (1994a). Along with these there is a number of other algae including green algae such as Rhizoclonium, Enteromorpha and Percursana and phaeophytes’ examples of which would be Dictyotopsis, Colpomenia and Hydroclathrus which are often found in mangrove areas and are widely distributed. With the exception of Dictyotopsis (Allender 1978) these genera are not restricted to mangrove habitats. Post (1936a) referred to such species as facultative members of the association, and algae which are frequent but not abundant in mangrove algal communities should be consid- ered separately from algae which also grow in adjacent communities on rocky shores or contiguous subtidal seagrass communities and simply continue to grow in the sheltered mangrove after they have been washed in. Examples of such species in eastern Australia include Sargassum spp., Codium affin. fragile, and Microdictyon umbilicatum (Velley) Zanardini. In unusual circumstances there may be a local element that is a genuine com- ponent of the mangrove algal flora. An example is provided in northern New Guinea where the mangroves, especially Rhizophora apiculata, immediately abut the fringing main- land coral reefs and macroalgae normally part of the coral reef community such as Caulerpa racemosa (Forrskal) J. Agardh, Halimeda macroloba Descaisne and H. opuntia (Linnaeus) Lamouroux grow on the aerial roots (King 1990b). A comparable situation occurs at North-West Cape in Western Australia. Another group of algae associated with mangroves are free-living or unattached forms. These include the free-living Gracilaria species which have been described for New Zealand shores by Chapman (1975) and which are certainly more widely spread especially in Australia. In eastern Australia the fucalean Hormosira banksii (Turner) Descaisne occurs as a free-living population (King 1981a, b) and similar populations occur in New Zealand. These populations are comparable to the ‘ecads’ or ‘forms’ of Ascophyllum, Fucus and Pelvetia found in the salt marshes of the northern hemisphere, in that they exhibit only vegetative reproduction and have a compact densely branched form. Recent papers from Australia and the region just to the north which either list algae associated with mangroves and their distribution, or from which such information can be derived, are given in Table 1. Further information can be extracted from the taxonomic works discussed below, and some detail is embedded in Post’s contributions. Distribution on a local scale, often depicted as zonation on pneumatophores or stilt roots, has been the basis on many descriptive papers in other parts of the world, and an example is provided in the work of Chihara and Tanaka (1988). Various hypotheses involving factors such as salinity, desiccation, and tolerance of sediments in suspension have been put forward to explain the abundance and distribution patterns observed. Beanland and Woelkerling (1983) for example have shown that the degree of shading by the tree canopy could be correlated to changes in frequency distributions of different algae epiphytic on the pneu- matophores of Avicennia marina, even though species diversity, total cover and biomass do not. TAXONOMY AND SYSTEMATICS The taxa most characteristic of mangrove communites in Australia are Bostrychia, Caloglossa, Catenella and Stictosiphonia. The global taxonomic treatment of Bostrychia (including Stictosiphonia) and Caloglossa by Post (1936a) has been considerably modified as further collections have become available, new species have been described, and new techniques (culture studies, chemotaxonomy) have been applied, but a classical morpho- logical and anatomical treatment still forms the basis against which the newer data are tested (King and Puttock 1989, 1994b). The genus Catenella is in need of further study (Min-Thein and Womersley 1976). Taylor-Wood (1990) and Mostaert (1994) have under- Proc. LINN. SOC. N.S.W., 115, 1995 R.J. KING 155 taken detailed ultrastructural studies of Bostrychia and Caloglossa respectively using freeze substitution techniques but the data so far are not useful in resolving taxonomic prob- lems. TABLE 1 Papers on mangrove macroalgal floristics or from which distributional data can be derived, for Australia and coasts to the immediate north (Published since the review of Saengeret al. 1977). Region and scope of work References AUSTRALIA general reviews Saenger et al. (1977), King (1981b), Hutchings and Saenger (1987) Rhodophyta King and Puttock (1994) species list and ecological accounts Queensland Cribb (1979), New South Wales King (1981b), King and Wheeler (1985) Victoria Davey and Woelkerling (1980, 1985) South Australia Beanland and Woelkerling (1982, 1983) BRUNEI species list Kig, Puttock and Booth (1992) INDONESIA species list Tanaka and Chihara (1988a, b) ecology of macroalgae Chihara and Tanaka (1988) PAPUA NEW GUINEA species list and ecological notes King (1990b) Bostrychiaand Stictosiphonia The only broad treatment of the taxonomy of the genus Bostrychia since Post (1936a) is that of King and Puttock (1989) who reassessed the genus recognising Bostrychia Montagne and the resurrected and emended genus StictosiphoniaJ.D. Hooker et Harvey. Eleven species of Bostrychia and six species of Stictosiphonia were recognised and Bostrychia pinnata J. Tanaka and Chihara was shown to be a species widely distributed in tropical Australia (King and Puttock 1986). Subsequent publications include a compara- tive study of the spermatangia in Bostrychia (King and Puttock 1991) and further details of the species B. pilulifera Montagne which was formerly inadequately known (King et al. 1991). Although the major study was based essentially only on morphological and anatomical observations it did include full descriptions of all life-history stages available. In addition multivariate analysis was used to investigate the polymorphic species B. tenella (Lamouroux) J. Agardh which is now regarded as including two subspecies B. tenella ssp. tenellaand B. tenellassp. flagellifera (Post) R.J. King et Puttock (King e¢ al. 1988). Since the publication of King and Puttock (1989), West et al. (1992a) have described anew species, Bostrychia bispora, from Darwin, Australia, based on features of reproduction and development in culture. Maggs (1988) has discussed the validity of recognising life-history variants of this nature as separate species. Bostrychia bispora is closely related to B. montziana (Sonder ex Kuetzing) J. Agardh but is recognised by asexual reproduction by bisporangia and no sexual reproduction, and its susceptibility to infection by the para- sitic red alga Dawsoniocolax bostrychiae (Joly and Yamaguishi-Tomita) Joly and Yamaguishi- Tomita. These characteristics cannot be determined on herbarium material so that many sterile collections attributed to B. moritziana may be properly this species (King and Puttock 1994a). West et al. (1992b) have questioned the separation of B. moritziana with monosiphonous lateral branches from B. radicans (Montagne) Montagne with only polysiphonous branches (see King and Puttock 1989). They consider that these species cannot be separated reliably in the field: in culture their morphology can be changed from one form to another, and both are susceptible to infection by Dawsoniocolax bostrychi- Proc. LINN. SOC. N.S.W., 115, 1995 156 MANGROVE MACROALGAE ae whereas other ecorticate species of Bostrychiaare not. Molecular systematic approaches may be useful in resolving these and related issues of systematics and biogeography. The genus Stzctosiphonia is separated from Bostrychia on the basis of the number of tiers of pericentral cells per axial cell, and the pattern of development of subsequent cor- tication. This separation is partially supported by the distribution of the polyols D-sorbitol and D-dulcitol. These two isomeric hexitols are generally considered to be characteristic of the subfamily Bostrychiordeae which is comprised of Bostrychia and Stictosiphonia (Kremer 1976). Investigations by Karsten and Kirst (1989a, b), indicated that this might not be always the case. A subsequent re-evaluation of the polyols showed that in the three species of Stictosiphonia examined only D-sorbitol was present, whereas species of Bostrychiaalmost all contained both D-sorbitol and D-dulcitol (Karsten et al. 1990). Since that time there have been published a number of exceptions to this simple story, including reports of some populations of B. radicans from the east coast of the USA lacking D-dulcitol, and S. tangetensis from South Africa containing both D-sorbitol and D-dulcitol (Karsten e¢ al. 1992b, 1993, 1994b; West et al. 1992b). Caloglossa Until the publication of King and Puttock (1994b) there was no overview of the genus Caloglossa since the taxonomic synthesis of Post (1936a) and the subsequent exami- nation of the morphology and ecology of the genus (Post 1943). Post (1936A) recognised six species, and although she did not provide detailed descriptions of the taxa consider- able information can be extracted from her key. In addition to the six species she also recognised a number of varieties and forms. Subsequently three new taxa were described (C. leprieunt var. angusta Jao, C. saigonensis Tanaka and Pham Hoang Ho and C. oga- sawaraensis var. latifolia Kumano), but none of these has been widely adopted. The first two would be accommodated within C. lepreurii (Montagne) J. Agardh as used by Post (1936a), and the last in C. beccarii (Zanardini) De Toni. King and Puttock (1994b) have provided a monographic treatment of the genus in which eight species are recognised. In this the prostrate species formerly placed in C. adnata (Zanardini) De Toni is recognised as two species, C. adhaerens R.J. King and Puttock and C. bengalensis (G. Martens) R.J. King and Puttock. No detailed analysis of the range of variation of the widespread and morpho- plastic species C. leprieun (Montagne) J. Argardh was undertaken in the study but four subspecies were recognised. The type of detailed study of this complex required, includ- ing crossing experiments using Australian material, is being undertaken by Mitsunobu Kamiya with Professor Y. Hara at Tsukuba University in Japan. West et al. (1994) have recently recognised a new apomeiotic species, Caloglossa apomeiotica West and Zuccarello, closely related to C. leprieurii from North America. ECOPHYSIOLOGICAL STUDIES The algae of the Bostrychia — Caloglossa — Association present ideal subjects for the investigation of tolerance to salinity change and desiccation. They grow in environments where there are strong physical gradients related to tide at specific localities, and most of the species also occur over a wide range of salinity along estuaries. For similar reasons ecophysiological research has often been undertaken on saltmarsh and mangrove phanerogams (Adam 1990). Published Australian studies have been concerned largely with the genus Caloglossa. Mosisch (1993) studied the effects of salinity on the distribution of C. lepriewrn, demonstrating its euryhalinity and the presence of distinct salinity eco- types. In a broader study of six Caloglossa species, Karsten andWest (1993) investigated the growth pattern and photosynthesis-irradiance curves in relation to salinity and light, and demonstrated an adaptation to shaded habitats for all species. Members of the genus Bostrychia have been much studied in relation to their Proc. LINN. SOC. N.S.W., 115, 1995 R.J. KING 157 capacity to osmoregulate, and special attention has been given to the role of compatible solutes, organic osmolytes such as polyols, amino acids and quaternary ammonium com- pounds. Compatible solutes balance the high ionic concentration in the vacuole and thus prevent the cytoplasmic enzymes being exposed to high ionic concentrations which would be damaging. In a study of Bostrychia scorpioides Karsten and Kirst (1989a) showed that along with concentrations of Na*, Kt and CI, D-sorbitol and D-dulcitol increased with increasing salinity. Although the polyols acted as compatible solutes turgor regula- tion was not complete. Karsten and Kirst (1989a) suggested that they could be a prerequi- site to long term survival in desiccating environments. Desiccation tolerance is a feature of Bostrychia species (Post 1963c). In a subsequent study Karsten and Kirst (1989b) reported on the role of inorganic ions and D-sorbitol in the maintenance of the internal osmotic potential of cells of B. radicans. The role of both D-sorbitol and D-dulcitol in Bostrychiaand Stictosiphonia have been more widely investigated since (Karsten et al. 1990, 1992b, 1993, 1994b; West et al. 1992b). An extensive study of compatible solutes in the genus Caloglossa has established that mannitol is a major low molecular weight carbohydrate in that genus (Karsten et al. 1992a). The data were obtained with both "C NMR and HPLC techniques. The result was surprising since mannitol has been regarded as typical of brown algae (Reed et al. 1985, Wright et al. 1987) and some prasinophytes and haptophytes (Dickson and Kirst 1987) and had not been demonstrated convincingly in any red alga. The variation in mannitol con- tent of field collections of Caloglossain the Sydney (Australia) region support the interpre- tation that it serves as a compatible solute (Karsten et al. 1992a), and in culture the intracellular mannitol concentration in six species of Caloglossa has been shown to be directly propotional to the external salinity (Karsten and West 1993). There is some evi- dence that genetic differences as well as environmental conditions play a role in the accu- mulation of mannitol. Further studies (Karsten et al. 1994a) on Caloglossa lepreurii from different localities using “C-NMR and 'H-NMR reaffirm the role of mannitol. The compound 4-hydroxypro- line betaine was reported for the first time in Caloglossa but was present in only low concen- trations in marine specimens and absent from freshwater plants. Digeneaside, the main photosynthetic and reserve product of the Ceramiales (Kirst 1980), was also reported, but the concentration of this and 4hydroxyproline betaine was not regulated by changes in salinity and they are thought to be unlikely to have a major role in osmoregulation. An osmotic role for the heteroside floridoside has recently been established in Catenella nipae Zanardini from eastern Australia (Karsten et al. 1995). While quantitative changes in organic osmolytes may be sufficient to account for osmotic adjustment in the long term, in macroalgae they are too slow to be of significance in environments with rapid changes in salinity. Water movement into or out of cells result- ing in swelling or shrinkage of cells is characteristic of immediate response of algae to osmotic shock. It is the separation of the plasmalemma from the cell wall with plasmolysis which causes damage. The concomitant swelling of the cell wall can prevent this separa- tion as Fischer (1984) suggested for Caloglossa leprieuri. Mostaert and King (1993) have shown that in hypersaline conditions up to 5 times seawater concentration the cell wall cross-sectional area of Caloglossa from a marine habitat increased from 5.7% to 38.2% of the area occupied by cell contents and cell wall within one hour and in plants from a fresh- water habitat from 14.2% to 47.9%. The magnitude and pattern of change in cell wall thickness was consistent with these changes in cell wall having a critical role in osmotic acclimation. PRODUCTIVITY AND BIOMASS An important role in community productivity for mangrove macroalgae has been Proc. LINN. SOC. N.S.W., 115, 1995 158 MANGROVE MACROALGAE assumed rather than demonstrated. Biomass is generally low and measured photosynthe- sis rates of Bostrychia, Caloglossa and Catenella (Karsten and West 1993; Mosisch 1993) are consistent with a range of benthic rhodophytes (King and Schramm 1976). Studies on the free-living Hormosira in Botany Bay indicate a conservative production rate of 440g m? y’ though this figure based on seasonal change in biomass fails to take into account loss of fixed carbon by exudation, tissue decay or grazing (King 1981la). A reassessment at the same site in spring 1992, 17 years after the initial observations, indicates that biomass val- ues were virtually unchanged. Larkum (1981) made a first attempt to quantify production for the entire Botany Bay ecosystem (Sydney, New South Wales) including mangroves but provided no specific estimate of the primary production of mangrove epiphytic algae. It appears that the macroalgae make their major trophic contribution to mangrove ecosystems through detrital food chains, since hervbivores are not conspicuous whereas detritivores are often abundant. Whether this is due to the physical environment or whether the particular algae have chemical defence mechanisms against herbivores has not been investigated. Even in ecosystems where grazing molluscs are abundant as in parts of eastern Australia where large numbers of gastropods are present there are no obvious grazing effects (King 1981a). This brief overview of mangrove macroalgal studies, emphasising recent studies in Australia, highlights a number of shortfalls in our present knowledge. Taxonomic studies have now advanced to the stage where further progress requires the application of newer techniques including culture studies, crossing experiments and the techniques of molec- ular genetics, in order to resolve relationships between taxa, and to be able to delimit taxa in such morphoplastic species. In the ecological area the factors controlling abundance and distribution are not well understood, though significant progress has been made in understanding response to salinity change. There is also a need for studies on the contri- bution of the algae to productivity of estuarine and mangrove areas, and an assessment of the way in which this productivity enters the food web. ACKNOWLEDGEMENTS I wish to thank colleagues who have coliected mangrove algae and/or participated in expeditions. In particular I thank Paul Adam (UNSW), Ian Price (Townsville), Bill Kilkeary, Freya Dawson and Alaric Fischer (Darwin) and Di Walker (Perth) and Dr W,]. Woelkerling (for collections from Western Australia). I especially thank the intrepid members of the Japanese Mangrove Expedition to northern Australia (1991) — Y. Hara, T. Horiguchi, T. Kato, J. Tanaka, M. Kamiya and K. Ishida, and those participating in sub- sequent expeditions in Baja California, Mexico (1992) and southern Japan (1993). Funding has been provided by the University of New South Wales, the Australian Research Council, the Joyce Vickery Foundation of the Linnean Society of New South Wales, and most generously by Monbusho (Japan). I am especially grateful to Associate Professor K. Barrow, Professor Y. Hara, Mitsunobu Kamiya, Dr U. Karsten, and Professor J. West for interesting discussions and helpful comments, and long term research collaboration. Finally Iam pleased to acknowledge technical support throughout this work B. L.J. Kertesz, D. Nicholls, and C.J. Puttock and to thank especially the students who have con- tributed so much to this project, particularly Don Fortescue, Lance Islip, Anika Mostaert, Arne Schoeler, Eleni Taylor-Wood and Mark Wheeler. Proc. LINN. SOC. N.S.W., 115, 1995 R.J. KING 159 References ApaM, P., 1990 — Saltmarsh Ecology Cambridge University Press: Cambridge. ALLENDER, B.M., 1980. — Dictyotopsis propagulifera (Phaeophyta) — an algal enigma. Phycologia 19: 234-236. BEANLAND, W.R., and WOELKERLING, W.J., 1982. — Studies on Australian mangrove algae: II. Composition and geo- graphic distribution of communities in Spencer Gulf, South Australia. Proceedings of the Royal Society of Victoria 94: 89-106. BEANLAND, W.R., and WOELKERLING, W.J., 1983. — Avicennia canopy effects on mangrove algal communities in Spencer Gulf, South Australia. Aquatic Botany 17: 303-313. CHAPMAN, V,J., 1975. — Mangrove Vegetation. Cramer: Lehre. CHIHARA, M.,and TANAKA, J., 1988. — Species composition and ecology of macroalgae in mangrove brackish areas of East Indonesia. Jn: OGINO K. and CHIHARA, M., (eds) Biological system of Mangroves. A Report of East Indonesian Mangrove Expedition 1986. 7-20. Ehime University:Japan. Criss, A.B., 1979. — Algae associated with mangroves in Moreton Bay, Queensland. Jn: BAILEY, A., and STEVENS, N.C., (eds) , Northern Moreton Bay Symposium: 63-69. Royal Society of Queensland: Brisbane. Davey, A., and WOELKERLING, W.]., 1980. — Studies on Australian mangrove algae I. Victorian communities: com- position and geographic distribution. Proceedings of the Royal Society of Victoria 91: 53-66. Davey, A., and WOELKERLING, W.J., 1985. — Studies on Australian mangrove algae III. Victorian communities: structure and recolonisation in Western Port Bay. Journal of Experimental Marine Biology and Ecology 85: 177- 190. Dickson, D.M.J., and Kirst, G.O., 1987. — Osmotic adjustment in marine eukaryotic algae: the role of inorganic ions, quaternary ammonium, tertiary sulphonium and carbohydrate solutes: II. Prasinophytes and hapto- phytes. New Phytologist 106: 657-666. FISCHER, H., 1984. — Turgor regulation in Caloglossa leprieurt (Montagne) J. Agardh (Delesseriaceae: Rhodophyta). Journal of Experimental Marine Biology and Ecology 81: 235-239. HUTCHINGS, P., and SAENGER, P., 1987. Ecology of Mangroves. University of Queensland Press: St. Lucia. KarsTEN, U., and Kirst, G.O., 1989a. — Incomplete turgor regulation in the ‘terrestrial’ red alga Bostrychia scorpi- oides (Huds.) Mont. Plant Science 61: 29-36. KarsTEN, U., and kirst, G.O., 1989b. — The effect of salinity on growth, photosynthesis and respiration in the estuarine red alga Bostrychia radicans Mont. Helgolander Meeresuntersuchungen 43: 61-66. KarsTEN, U., and WEST, J.A., 1993. — Ecophysiological studies on six species of the mangrove red algal genus Caloglossa. Australian Journal of Plant Physiology 20: 729-739. KarsTEN, U., KING, R.J., and Kirst, G.O., 1990. — The distribution of D-sorbitol and D-dulcitol in the red algal gen- era Bostrychia and Stictosiphonia (Rhodomelaceae, Rhodophyta) — a re-evaluation. British Phycological Society Bulletin 25: 363-366. KarsTEN, U., WEST, J.A., MOSTAERT, A.S., KING, R.J., BARROW, K.D., and Kirst, G.O., 1992a. — Mannitol in the red algal genus Caloglossa (Harvey) J. Agardh. Jowrnal of Plant Physiology 140: 292-297. KarsTEN, U., WeEsT, J.A., and ZUCCARELLO, G., 1992b. — Polyol content of Bostrychia and Stictosiphonia (Rhodomelaceae, Rhodophyta) from field and culture. Botanica Marina 35: 11-19. KarsTEN, U., WEST, J.A., and GANESAN, E.K., 1993. — Comparative physiological ecology of Bostrychia radicans Mont. (Ceramiales, Rhodophyta) from freshwater and marine habitats. Phycologia 32: 401-409. KarsTEN, U., BARROW, K.D., MosTarrt, A.S., KING, R.J., WEST, J.A., and Kirst, G.O., 1994a. — °C and 'H-NMR studies on digeneaside in the red alga Caloglossa leprieurtt — a re-evaluation of its osmotic significance. Plant Physiology and Biochemistry 32: 1-5. KarsTEN, U., WEST, J.A., ZUCCARELLO, G., and Kirst, G.O., 1994b. — Physiological ecotypes in the marine alga Bostrychia radicans (Ceramiales, Rhodophyta) from the east coast of the USA. Journal of Phycology 30: 174- 182. KarsTEN, U., BARROW, K.D., MosTarrT, A.S., and KING, R.J., 1995 — The osmotic significance of the heteroside floridoside in the mangrove alga Catenella nipae (Gigartinales; Rhodophyta) in eastern Australia. Estuarine, Coastal and Shelf Science 239-247. KING, R,J., 1981a. — The free-living Hormosira banksii (Turner) Descaisne associated with mangroves in temperate eastern Australia. Botanica Marina 24: 569-576. KING, R.J., 1981b. — Mangroves and saltmarsh plants. Jn: CLAYTON, M.N., and KING, R.J., (eds), Marine Botany: an Australasian Perspective. 308-328. Longman-Cheshire: Melbourne. KING, R.J., 1990a. — Ecology of marine plants. Jn: CLAYTON, M.N., and KING, R,J., (eds), Biology of Marine Plants: 245-249. Longman-Cheshire: Melbourne. KING, R.J., 1990b. — Mangrove macroalgae in Papua New Guinea. Botanica Marina 33: 55-62. KING, R.J., and Briccs, B., 1988. — Valerie May: fifty years of phycology. Telopea 3: 273-279. KING, R,J., and PUTTOCK, C.F., 1986. — Bostrychia pinnata]. Tanaka et Chihara in Australia. Bulletin of the National Science Museum, Tokyo., Ser B 12: 17-24. KING, R,J., and PUTTOCK, C.F., 1989. — The morphology and taxonomy of Bostrychia Montagne and Stictosiphonia). Hooker et Harvey (Rhodomelaceae, Rhodophyta). Australian Systematic Botany 3: 1-73. KING, R.J., and Putrock, C.F., 1991. — A comparative study of spermatangia in Bostrychia Montagne. Japanese Journal of Phycology 39: 143-150. KING, R.J., and PUTTOCK, C.F., 1994a. — Macroalgae associated with mangroves in Australia: Rhodophyta. Botanica Mamina 37: 181-191. KING, R.J., and PUTTOCK, C.F., 1994b. — Morphology and taxonomy of Caloglossa (Delesseriaceae, Rhodophyta). Australian Systematic Botany 7: 89-124. KING, R.J., PUTrock, C.F., and BooTu, W., 1992. — The mangrove macroalgae of Brunei Darussalam. Wallaceana 65: 1-3. Proc. LINN. SOC. N.S.W., 115, 1995 160 MANGROVE MACROALGAE KiNG, R.J., Putrock, C.F., and de PauLa, E.J., 1991. — The morphology of Bostrychia pilulifera Montagne (Rhodomelaceae, Rhodophyta). Japanese Journal of Phycology 38: 11-16. KING, R.J., PUTTOCK, C.F., and VICKERY, R.S., 1988. — The Bostrychia tenella complex. Phycologia 27:10-19. KING, R,J., and SCHRAMM, W., 1976. — Photosynthetic rates of benthic marine algae in relation to light intensity and seasonal variations. Marine Biology 37: 215-229. KING, R,J., and WHEELER, M.D., 1985. — Composition and geographic distribution of mangrove macroalgal com- munities in New South Wales. Proceedings of the Linnean Society of N.S. W. 108: 97-117. Kirst, G.O., 1980. — Low mw carbohydrates and ions in the Rhodophyceae: quantitative measurement of florido- side and digeneaside. Phytochemistry 19: 1107-1110. KREMER, B.P., 1976. — “C -assimilate pattern and kinetics of photosynthetic '*CO¢- assimilation of the marine red alga Bostrychia scorpioides. Planta 129: 63-67. LARKUM, A.W.D., 1981. — Marine primary productivity. In: Clayton, M.N. and King, R.J., (eds), Marine Botany: an Australasian Perspective. 369-385. Longman-Cheshire: Melbourne. Maccs, C.A., 1988. — Intraspecific life history variability in the Florideophycidae (Rhodophyta). Botanica Marina 31: 349-361. Min-THEIN, U., and WoMERSLEY, H.B.S., 1976. —Studies on southern Australian taxa of Solieriaceae, Rhob- doniaceae and Rhodophyllidaceae (Rhodophyta). Australian Journal of Botany 24: 1-166. MosiscuH, T.D., 1993. — Effects of salinity on the distribution of Caloglossa lepriewrii (Rhodophyta) in the Brisbane River, Australia. Journal of Phycology 29: 147-153. MostaerT, A.S., and KING, R.J., 1993. — The cell wall of the halotlerant red alga Caloglossa leprieurii (Montagne) J Agardh (Ceramiales, Rhodophyta) from freshwater and marine habitats: effects of changing salinity. Cryptogamic Botany 4: 40-46. MosTaerT, A.S., 1994. — Cell Structure and Physiology of the Red Alga Caloglossa leprieurii (Montagne) J. Agardh. Sydney: University of New South Wales, PhD Thesis, unpublished. Post, E., 1936a. — Systematische und pflanzengeographische Notizen zur Bostrychia-Caloglossa Assoziation. Revue algologique9: 1-84. Post, E., 1936b. — Zur Verbreitung von Catenellain Neuseeland. Hedwigia 76: 189-190. Post, E., 1937. — Weitere Daten zur Verbreitung des Bostrychietum I. Hedwigia 77: 11-14. Post, E., 1938a. — Weitere Daten zur Verbreitung des Bostrychietum II. Hedwigia 78: 202-215. Post. E., 1938b. — Zur Okonomie des Bostrychietum. Planta. 28: 743-744. Post, E., 1939a. — Bostrychia tangatensis spec. nov., eine neue Bostrychia der Ostafrikanischen Mangrove. Archiv fur Prostistenkunde 92: 152-156. Post, E., 1939b. — Weitere Daten zur Verbreitung des Bostrychietum III. Archiv fur Prostistenkunde 93: 6-37. Post, E., 1941. — Bostrychia hamana-tokidai spec. nov., eine neue sudjapanische Bostrychia. Beth. bot. Zbl. B 61: 200- 210. Post, E., 1943. — Zur Morpholgie und Okologie von Caloglossa. Ergebnisse der Sunda-Expedition der Notgemeinschaft der deutschen Wissenschaft 1929/30. Archiv fur Protistenkunde 96: 123-220. Post, E., 1955a. — Weitere Daten zur Verbreitung des Bostrychietum IV. Archiv fur Protistenkunde 100: 351-377. Post, E., 1955b. — Weitere Daten zur Verbreitung des Bostrychietum V. Bericht der Deutschen botanischen Gesellschaft 68: 205-215. Post, E., 1957a. — Weitere Daten zur Verbreitung des Bostrychietum VI. Archiv fur Protistenkunde 103: 489-506. Post, E., 1957b. — Fruktifikationen und Keimlinge bei Caloglossa. Hydrohiologia 9: 105-125. Post, E, 1959. — Weitere Daten zur Verbreitung des Bostrychietum VII. Archiv fur Protistenkunde 103: 492-496. Post, E. 1961. — Bostrychia flagellifera Post in Japan. Bull. Res. Cown. Israel Sect.D10: 101-115. Post, E., 1962. — Murrayellopsis dawsonu gen. et spec. nov. Post aus einem GoldfischNest. Schriften des Naturwissenschaftlichen Vereins fur SchleswigHolstein 33: 4pp. Post, E., 1963a. — Murrayellopsis dawson gen. et spec. nov. Post aus einem marinen GoldfischNest. Naturunssenschaften 2: 49. \ Post, E., 1963b. — Zur Verbreitung und Okologie der Bostrychia-Caloglossa- Assoziation. International Revue der Gesamten Hydroliologie 48: 47-152. Post, E., 1963c. — Bostrychia—nicht tot zu kriegen. Botanica Marina5: 9-18. Post, E., 1963d. — Bostrychia radicans im Susswasser Westafrika. Revue algologique 6: 270-281. Post, E., 1964a. — Murrayellopsis dawsonii gen. et spec. nov. Post aus einem marinen GoldfischNest. Hydrobiologia 23: 274-280. Post, E., 1964b. — Bostrychia hamana-tokidai Post. Hydrobiologia 23: 544-560. Post, E., 1964c. — Bostrychietum aus dem National Park von Melbourne. Revue algologique 7: 242-255. Post, E., 1964d. — Bostrychia tangatensis Post. Hydrobiologia 24: 584-601. Post, E., 1965a. — Calglossa beccariiim Golf von Mexico. Hydrobiologia 26: 184-188. Post, E., 1965b. — Bostrychia scorpioides im tropischen Westafrika. Hydrobiologia 26: 301-306. Post, E., 1966a. — Caloglossa stipitatain Florida. Hydrobiologia 27: 317-322. Post, E., 1966b. — Neues zur Verbreitungsoekologie neuseelandischer and mittelamerikanischer Bostrychia- Caloglossa-Assoziation. Revue algologique8: 127-150. Post, E., 1966c. Caloglossa ogasawaraensis in Westafrika. Hydrobiologia 27: 317-322. Post, E., 1966d. — Bostrychietum auf den Philippinen. Hydrobiologia 27: 344-352. Post, E., 1967a. — Zur Okologie des Bostrychietum. Hydrobiologia 29: 263-287. Post, E., 1967b. — Zur Verbreitung und Okologie von Dictyotopsis propagulifera W. Troll. Revue algologique 8: 279-291. Post, E., 1968a. — Bostrychia kelanensis Grun. Hydrolnologia 31: 81-150. Post, E., 1968b. — Zur Verbreitungs-Okologie des Bostrychietum. Hydrobiologia 31: 241-316. Proc. LINN. SOC. N.S.W., 115, 1995 RJ. KING 161 REED, R.H., DAVIDSON, I.R., CHUDEK, J.A., and FosTER, R., 1985. — The osmotic role of mannitol in the Phaeophyta: an appraisal. Phycologia 24: 35-47. SAENGER, P., SPECHT, M.M., SPECHT, R.L., and CHAPMAN, V,J., 1977. — Mangal and Coastal Salt-marsh Communities in Australia. In: CHAPMAN, V.J., Wet Coastal Ecosystems: 293-345 Elsevier: Amsterdam. TANAKA, J., and Chihara, M., 1988a. — Macroalgae in Indonesian mangrove forests. Bulletin of the National Science Museum, Tokyo, Series B 14: 93-106. TANAKA, J., and Chihara, M., 1988b. — Macroalgal flora in mangrove brackish areas of east Indonesia. Jn: OGINO K., CuHara, M., (eds) Biological System of Mangroves. A Report of East Indonesian Mangrove Expedition 1986. 21- 34. Ehime University: Japan. TayYLOR-Woop, E., 1990. — Structure of the red algal genus Bostrychia. Sydney: University of New South Wales, BSc (Hons) thesis, unpublished. WEST, J.A., ZUCCARELLO, G.C., and CALUMPONG, H.P., 1992a. — Bostrychia bispora sp. nov. (Rhodomelaceae, Rhodophyta), an apomictic species from Darwin, Australia: reproduction and development in culture. Phycologia 31: 37-52. WEST, J.A., ZUCCARELLO, G.C., PEDROCHE, F.F., and KArsTEN, U., 1992b. — Marine red algae of the mangroves in Pacific Mexico and their polyol content. Botanica Marina 35: 567-572. WEST, J.A. ZUCCARELLO, G.C., PEDROCHE, F.F., and KARSTEN, U., 1994. — Caloglossa apomeiotica sp. nov. (Ceramiales, Rhodophyta) from Pacific Mexico. Botanica Marina 37: 381-390. WRIGHT, P.J., CLAYTON, M.N., CHUDEK, J.A., FOSTER, R., and REED, R.H., 1987. — The carbohydrate altritol in Bifurcariopsis capensis, Hormosira banksii and Xiphophora chondrophylla (Fucales, Phaeophyta) from the southern hemisphere. Phycologia 26: 429-434. Proc. LINN. SOC. N.S.W., 115, 1995 anach es! + veto ea site inne vil ra a Tn Oe i 7 cw. f i ¢ ‘itis ‘ iP 7 P4 t i ‘ ery Th acltach et 9 A wit mi See i? arg or Costes iy P " { fh Ral ig PS fA hod Pe TH dd Tee ee S| be ersey ‘xa ‘eta, imal 8 ian sik ‘ i ; 3 \ ‘ ,- 5 , reet 3% uv aN ai wie Sd + ad ’ » Usenet ien ¥ eet ek ae | Vana is a Un ak ache. inte Yd so Redd) incernue Atareved be eR 1H oe TS bl ‘ ow aeser Y's OY) Pa Teas , aly a | h Prat: i sas VWetagpihin® HoAons len. el eae ie Pau aiadtieeor ff dren Goble Procy, Z 5 Lees ene Way mee) i eTe§ yu lity Ba A y an ee ey ee we tr ‘i We Ehyexi ihe, arate cahgelegtgiar LS ee Prer.F., -e mien ery) i haben ; we Pest. 8. thas, 4 Slates bes co) hy Got Con oer Peggecigly ete iA eRe in Frey_c AP. ees berm } eS Prerejew Zeeks ch ond Mies 26 GH aM. : gi a a ates fe mn Kin ait t Pig " tive *. ae Spr, Bar meen peageet #V Lee phe ie aid niiteinmesihaiTilen { hie and Val — wi. eet Hime y a4 Prom 8, tari (ema ar Grain oar, Bh a LA hea 2, Petes Hs Fini, P View i 0 oe eee poe Pi 7 , f ini ’ Nally : : 4 Alar, To. 1W0tm = Poe 8 ve ied Hynes hier ~y s ne a Rep). en iu © diver airy rl CURA’ sade BG it =r W Tatt: tite ee =e | 3 oer, M1. (ini ved eenris aS. PM teotoid ye Ft Od) uy Roar, BK — tut Wer in aeigs OM glagis dex Neiman , Pgdeaibedaghe SI ou ih; LT eivi f ee CR Las fl The Pollen Morphology of Some Co-occurring Species of the Family Myrtaceae from the Sydney Region JANE M. CHALSON and HELENE A. MARTIN CHALSON, J.M., & MARTIN, H.A. The pollen morphology of some co-occurring species of the family Myrtaceae from the Sydney Region. Proc. Linn. Soc. N.S.W. 115: 163-191 (1995). Myrtaceous pollen is superficially very similar so that the identification of dispersed grains is notoriously difficult. Given the importance of the family in the Australian vegeta- tion, the identification of dispersed grains is extremely important for palynological studies on the history of the vegetation. This study examines means of a more precise and reliable identification of myrtaceous pollen. Areference set of 20 species were analysed, from within an area used for palynolog- ical analysis and a historical vegetation study. A number of gross morphological and fine detail characters were scored. There are sufficient diagnostic character states to separate all but two of the species in which there is overlap. The success of this approach depends upon the inclusion of fine-detail characters. The methods employed here use tables and keys as aids to identification but punch cards or acomputer program would be equally suitable. Indeed, with a larger reference set, a computer program would be essential. Jane M. Chalson, 46 Kilmarnock Road, Engadine, N.S.W. 2233. Helene A. Martin, School of Biological Science, University of New South Wales, P.O. Box 1, Kensington, N.S.W.; manuscript received 29 February 1992, accepted for publication 20 October 1993. INTRODUCTION Early studies on Myrtaceae pollen (e.g. Erdtman, 1952 and Pike, 1956) concluded that the family has more or less uniform pollen. Pike (1956) surveyed some 300 species from 71 genera and used the slight differences she observed to construct a tentative key to the genera. The key separates out some single genera but most of the end groups contain a number of genera, and one group contains 15 genera. Moreover, some genera, e.g. Angophora and Eucalyptus occur in 3 of these multi-generic groups. Clearly, this approach is unsatisfactory for the practical purposes of identifying dispersed pollen with a parent taxon. These earlier studies, however, concentrated on gross morphology. In a more detailed study, McIntyre (1963) found that of 5 genera of Myrtaceae pre- sent in New Zealand, 3 could be distinguished individually. However, identification of species is difficult and usually not possible except for genera with only one species. This contrasts with Australia because the problem of identifying myrtaceous pollen is much simpler in New Zealand, which has only 18 native species. In a study undertaken mainly for taxonomic purposes, Gadek and Martin (1981) examined 28 species of the subtribe Metrosiderinae. They found that the gross morpho- logical characters, e.g. syncolpy or parasyncolpy, presence or absence of a polar island, may not be taxonomically reliable at the species level. The characters most useful in assess- ing taxonomy were those of fine detail, e.g. the structure of the angle of the amb, exine structure and sculpture, form and type of colpi and apocolpia. Using these characters, some species were distinctive enough to be identified from all others in that study. The identification of dispersed myrtaceous pollen is of paramount importance for palynological studies into the history of the Australian vegetation. Altogether there are some 47 native genera and 1300 species of this family in Australia (Beadle, 1981). Eucalyptus is dominant over much of the Australian continent outside of the arid Proc. LINN. SOC. N.S.W., 115, 1995 164 POLLEN MORPHOLOGY OF MYRTACEAE Eremaean Zone and the small, discontinuous area of rainforest. Although not dominant in the Eremaean Zone, Eucalyptus is far from uncommon there. In rainforest, other gen- era in the family may be important. Vegetation units are defined in terms of their domi- nants. If past vegetation, as deduced from pollen spectra, is to be defined with any sort of precision, then the problem of identifying dispersed myrtaceous grains must be tackled. There are some Quaternary palynological studies that differentiate myrtaceous pollen into groups. The most successful of these is that of Churchill (1957, 1961 and 1968) who distinguished 7 species of Eucalyptus and a group of 2 species which are closely related and hybridize freely. Churchill (1957) used ten character states of both gross morphology and fine detail. These included equatorial diameter, polar axis, sexine sculpture, pres- ence of polar islands, depth of colpi, concavity of sides of amb, comparative thickness of the sexine and nexine and several pore characters. This study, using many more charac- ters than any other, has been very successful. However, this approach has been ignored by subsequent workers in the field. Dodson (1974) used 4 characters to place 27 species of 8 genera into 13 groups. Two of the characters are those of fine detail, viz. pore type and nature of the colpus. The other 2 are those of gross morphology, viz. presence or absence of a polar island and equatorial diameter. The 12 Eucalyptus species in that study were placed in 4 groups and one of these is a single species. There are 3 other single species groups. In subsequent works, Dodson (1977, 1979, 1982) continued this approach. Binder (1978), using the same 4 characters of Dodson, divided 30 species into 9 groups, 3 of which are Eucalyptus groups. Only 2 of these groups are single species, and they are not eucalypts. Rose and Martin (in prep.) use much of the same characters to define 4 groups within the Myrtaceae, 2 of which are Eucalyptusand Angophora. Ladd (1979) divides the Eucalyptus species in his area of study into 5 groups without any indication as to how this division has been achieved. Not a single groups is exclusive, i.e. at least one species in each group is found in another group. Most other Quaternary studies attempt little subdivision of the myrtaceous pollen, e.g. Colhoun et al. (1982) with 2 groups, one of Eucalyptus/Melaleuca species, the other of Leptospermum species. In summary, the previous studies show that subdivision and identification of dis- persed myrtaceous pollen may be successful if the fine detail characters are used, together with those of gross morphology. Greater success has been achieved using a larger number of characters. This paper explores the feasibility of specific identification of myrtaceous pollen in the reference set. Variation within a species has been assessed as well as the definition of diagnostic character states of a species or group of species. The reference set of 20 species comprising those found in the Jibbon Swamp area of the Royal National Park, with a few specimens not actually growing in the region but included as possible occupants of the area in the past. A palynological study of Jibbon Swamp and the history of the vegetation (Chalson, 1983) relies on the identification of the myrtaceous pollen. METHOD Dry anthers have been collected from herbarium specimens in either the John T. Waterhouse Herbarium, University of N.S.W. (UNSW) or the National Herbarium, Royal Botanic Gardens, Sydney (NSW) (Table 1). The anthers have been treated according to Erdtman’s acetolysis method as modified by Chanda (1966). The treated pollen has been mounted in safranin/glycerine jelly. This mounting medium is preferred because the slight expansion it causes, when compared with silicone oil, allows the layers of the exine to be seen more clearly. Pollen grains which have been mounted in glycerine jelly for a long period of time may swell (Anderson, 1960) for this reason most of the pollen slides Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 165 used were mounted within the last two years. Some slides from the modern reference pollen collection at the Botany School, University of N.S.W. were included, but only after selection to preclude swollen grains. Photomicrographs have been taken using a Zeiss microscope, xl00 lens (n.a.1.3) under oil immersion and through a green filter. Observation has been carried out with the same set-up without the green filter. Measurements are based on at least 20 grains; the mean measurement is enclosed in parenthesis between the range observed. The measurements are in micrometers (jm). Herbarium specimens used for obtaining pollen for analysis are listed in Table 1. All pollen slides have been placed in the collection at the Botany School, University of N.S.W. Palynological terms are defined in Kremp (1965) unless otherwise stated. GENERALISED POLLEN MORPHOLOGY The grains are free, radiosymmetric, isopolar or almost isopolar, oblate, triangular, angulaperturate and tricolporate. Two layers in the exine are usually distinguishable, the endexine and ektexine, both more or less of equal thickness (Fig. 1). The equatorial apertures are complex and are formed by a splitting of the endexine and ektexine to form a vestibule. In equatorial view the vestibule appears as a meridional colpus in the ektexine when viewed from the surface, and is elliptical to slit-like in shape. The endexine viewed through the equatorial colpus appears granular if the inner surfaces of the vestibule are roughened. The endexine appears to have an indistinct equatorial col- pus. In polar view the roof and the floor of the vestibule are extensions of the ektexine and endexine layers, respectively, of the exine. Whilst the floor is usually the same thickness as the rest of the endexine, the roof may be thicker, thinner or the same width as the ektex- ine elsewhere on the grain (Fig. 10). Most grains settle in the polar aspect due to the extreme oblate shape, thus all obser- vations have been made from the polar view and the equatorial endocolpus is not visible. If the ektexine is markedly thickened or patterned over the vestibule, the equatorial col- pus may not be visible, and the whole of the pore may be partially obscured, or indistinct. Very small vestibules with little separation of the ektexine and endexine may also be indis- tinct in polar view. The morphology of the polar region varies greatly both between species and, in some groups, within species. Grains may be syncolpate, parasyncolpate, brevicolpate or syndemicolpate (Fig. 6). The apocolpium may be covered with ektexine (a polar island) or an expanded membrane (Gadek and Martin, 1981) similar to that of the colpi (Fig. 5). Exine patterning is generally fine, most species being scabrate or smooth. Patterning often changes along the colpi edges and over the pore areas. Abnormal grains are found in all species. There are two common forms of abnor- mality. Some grains have 2, 4 or 5 pores instead of the normal 3 (Fig. 12S) which does not appear to affect the other character states of the grain used for identification, and thus these grains can be ascribed to a species if they have the correct orientation. The other common abnormality occurs when grains have very thick exine. These grains are usually smaller than the normal range and have strongly concave or convex walls. These grains are not identifiable. THE CHARACTERS The pollen grains range from concave triangular through straight triangular to con- vex triangular (Fig. 3). The angle of the amb may be sharply rounded or rounded, and there may be a notch at the base of the vestibule (Fig. 4). The polar region displays a variety of character states. The colpi may not meet at the Proc. LINN. SOC. N.S.W., 115, 1995 166 POLLEN MORPHOLOGY OF MYRTACEAE pole, or only one colpus may bifurate (Fig. 9). The apocolpium may be covered with ektexine (a polar island) similar to elsewhere on the grain, more highly patterned than elsewhere or thinner and less patterned than elsewhere on the grain. The apocolpium may, alternatively, be an expanded membrane similar to that of the colpi and either smooth or with scattered granules on the surface (Fig. 5C). Polar islands may either fit closely into the apocolpium (i.e. have the same shape and almost the same size) or they may be smaller than the apocolpium and have an irregular shape (Fig. 5A, B). The pore area is similarly complex and variable. In all cases some degree of splitting of the endexine and ektexine is visible. However, visibility may be impaired by thickened ektexine over the roof of the vestibule, rough inner surfaces of the vestibule, continuation of the exine patterning over the vestibule roof, small size of the vestibule and/or strong colouring of the exine. The floor of the vestibule is a continuation of the endexine else- where on the grain. It may be concave, flat or convex and is generally the same thickness as the endexine elsewhere on the grain (Fig. 10). The roof of the vestibule is smooth or pat- terned on the outside. Although the roof is a continuation of the ektexine elsewhere on the grain, the surface pattern usually changes over the immediate pore area. The thick- ness of the roof may either taper gently towards the exopore, round off abruptly or be truncate at the exopore edge. The inner surfaces of the vestibule are either rough or smooth (Fig. 10). The measurement of pore height is external, it includes both the floor thickness and roof thickness. The measurement of pore width is internal as the inclusion of wall thickness would make the measurements too variable to be of use (Fig. 2). Grains must be viewed in true polar view otherwise the character states associated with the pore area will not be seen clearly. The elements of the exine patterning are generally less than 1 wm., i.e. scabrate (Erdtman, 1952). However, different shaped elements are clearly visible, even at this small size, and they are described as smooth, undulate (Fig. 11B), scabrate, verrucate/scabrate, granulate/scabrate, granulate, verrucate, vermiculate or rugulate (Fig. 11). The term scabrate is reserved for a very rough irregular form of scabrate pat- terning. There is variation in the distribution of exine patterning. Patterning may be found evenly distributed over the whole grain, more commonly it is absent or reduced around the pore area. In addition, patterning may be restricted to the equatorial mesocol- pal areas or it may be most pronounced along the edges of the colpi. Colpi may have gaps where the colpus is not visible and normal ektexine is found occupying the gap instead (Figs 8, 9). Some colpi have thickened borders, others have broken, rough, irregular or smooth edges (Fig. 7). Broken edges have pieces of normal ektexine closely associated with the edges of the colpi but not attached to the mesocolpal region. Rough edges are continuously uneven but have no detached sections of ektexine. Irregular edges have infrequent, irregular sharp bends and smooth edges follow even curves or only bend at bifurcation points between colpi and poles. The thickness of the wall is also measured (Fig. 2). DESCRIPTION OF THE SPECIES Angophora hispida Grains have slightly concave sides and rounded ambs. The amb has a notch at the base of the vestibule. Colpi are usually arcuate with a distinct polar membrane approxi- mately 5-10j.m. The apocolpium rarely contains a polar island which is found on one side of the grain only. The edges of the colpi are broken. The colpi do not always extend from the pole to the pore but often have gaps and other irregularities around the immediate polar area (Table 7, Figs 8, 9). The exine is very thick. The surface is smooth with a slightly undulatory appearance in the optical section in the mesocolpal areas (Fig. 12A). Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 167 The pore areas protrude from the rest of the grain. The vestibule is small, and not always distinctly visible, and the inside surfaces are obviously roughened. The floor is flat, the roof is thickened and tapers abruptly at the exopore opening with rounded edges (Fig. 10D). The outside surface is smooth. Dimensions are in Table 3. A. hispidais distinguished from most species by its large apex/base measurement. It does not have the grooved pattern of A. costata or E. eximia. More difficulty is experienced with E. gummifera. While some of the grains may be identified with their respective species there is overlap of all the character states (Table 3). Thus it is not possible to assign all the grains to a species. Angophora costata Grains have straight to slightly convex sides. The amb is rounded and notched at the base of the vestibule. Colpi are arcuate, syncolpate or with a very small polar membrane of about 1-24.m wide. The colpi may have gaps or very small polar islands (Table 7) and irreg- ular edges (Fig. 7B). The exine has a smooth unthickened band, about 6m wide, follow- ing the colpi. The mesocolpal areas are clearly patterned with vermiculate grooves (Fig. 11A). Grains seen in oblique polar view are rarely deformed, giving the impression of a strong rigid wall (Fig. 12B). The amb has a small distinct vestibule with rough inner surfaces. The floor of the vestibule is concave in polar view. The ektexine over the vestibule is conspicuously thick- ened, but thins abruptly towards the rounded edges of the exopore and is smooth on the outer surface (Fig. 10D). Dimensions are in Table 3. This species is clearly distinguished from Angophora hispidaand Eucalyptus gummifera by the distinctive, grooved pattern in mesocolpal regions (Table 3). A costata is distin- guished from E. eximia by its apex/ base measurement. Eucalyptus botryoides Grains have straight to concave sides. The amb is rounded and notched at the base of the vestibule. Colpi are straight and angular (Fig. 6C) with rough margins, enclosing a polar island 4-9,.m wide. The polar island closely fits the apocolpium. Exine patterning is finely scabrate and distributed evenly over the entire grain except the pore areas which are smooth. The amb generally has a narrow vestibule, a rough inner surface is visible on some of the specimens with wider vestibules. The floor of the vestibule is straight to slightly con- vex. The ektexine over the vestibule is very conspicuously thickened and is smooth on the outside. The ektexine thins abruptly at the exopore and has rounded edges (Fig. 10B). Dimensions are in Table 3. This species is distinguished from the Angophora/bloodwood group by its fine scabrate pattern on the exine. It differs from the other species of the Eucalyptus group by a large polar island with the same patterning as that of the ektexine elsewhere on the grain (Table 4). This species is distinguished from the Melaleuca group by its large polar island, scabrate patterning, angular colpi and rough colpi edges (Table 2). Eucalyptus camfieldii Grains have straight to slightly concave sides and rounded ambs. Colpi are straight and angular with rough edges (Fig. 7C); the bifurcations enclose a large, triangular polar membrane 8-94.m wide, which lacks ektexine. The ektexine has a faint scabrate pattern which is more pronounced on the equatorial mesocolpium. The less patterned ektexine forms a band approximately 54m wide along the colpi edges. Most of the grains are seen in an oblique angle indicating a less extreme oblate shape than is the norm for the Myrtaceae species studied here. The polar region is frequently depressed probably because the lack of ektexine makes it less rigid. With the depression of the polar mem- brane, the surrounding areas tend to cave in around it, making clear viewing of the polar Proc. LINN. SOC. N.S.W., 115, 1995 168 POLLEN MORPHOLOGY OF MYRTACEAE region difficult. The amb has a distinct vestibule, the inner surfaces of which appear to be rough- ened. The floor of the vestibule is straight or slightly concave in pores in true polar view, but may appear convex if viewed at an oblique angle. The ektexine over the vestibule is conspicuously thickened and smooth on the outer surface. The depressed polar region is characteristic. Dimensions are in Table 3. This species is separable from the Angophora/bloodwood group by its scabrate pat- tern. It differs from the rest of the Eucalyptus group by having both a smooth polar mem- brane and a thickened vestibule roof (Table 4). Eucalyptus eximia Grains have concave to straight sides, the amb is rounded and notched at the base of the vestibule. Colpi are accurate with very rare gaps (approximately 7.5% of the sample) which are less than 1m long (Table 7). The sides of the colpi are rough and irregular and not thickened (Fig. 12D). Frequently the apocolpium has a small polar island which may be irregular in shape (Fig. 5B) or may closely fit the apocolpium. The apocolpium is approximately 48m across with ektexine granules when a polar island is absent. The ektexine is rugulate or rugulate/scabrate. Ambs have medium sized, indistinct vestibules with rough inner surfaces. The floor of the vestibule is flat, the roof is thickened (Fig. 12C) when compared to the rest of the ektexine and is patterned on the outside. The roof tapers abruptly towards the exopore and has rounded ends (Fig. 10D). Dimensions are in Table 3. The vermiculate grooves separate this species from most other Myrtaceae pollen. Angophora costatais much larger and Eucalyptus obstans has angular colpi and no colpi gaps. Eucalyptus gummifera Grains have straight to slightly concave sides, the amb is rounded and notched at the base of the vestibule. Colpi are usually arcuate enclosing an apocolpium up to 10m in diameter, however, the polar region is very variable (Fig. 9). The colpi have broken edges (Fig. 7D), they may be continuous, or one or more may have a gap from 1 to 54m long (Fig. 12F), where the colpus appears to end and normal ektexine occupies the area. There is also variation in the way the colpi meet at the pole. There may be a polar island or a par- tially separated island or an island not separated at all (Fig. 9). This variation is discussed further below under intraspecific variation. The exine is undulate in the equatorial meso- colpal region (Fig. 11B). There is a small indistinct vestibule with roughened inner surfaces while the floor is straight to slightly convex (Fig. 12E). The ektexine over the vestibule is thickened and smooth on the outer surface. The roof tapers abruptly towards the exopore and has rounded edges. Dimensions are in Table 3. This species is morphologically close to Angophora hispida and most character states overlap between species (Table 3). Thus while some of the grains may be identified with their respective species, it is not possible to assign all of the grains confidently. Eucalyptus haemastoma Grains have straight to slightly concave sides, the amb is rounded and notched at the base of the vestibule. Colpi are straight and angular with rough edges (Fig. 12H). The colpi bifurcate to form a large apocolpium approximately 5-8m wide. A large polar island fills the apocolpium almost entirely. The polar region is best observed in fully expanded grains because the expansion allows the colpi and the edge of the polar island to be seen clearly. The ektexine of the polar island is not the same as the rest of the grain, itis thinner and has a rugulate pattern, or it may consist only of granules (Fig. 5C). Exine is smooth to faintly scabrate. The endexine is approximately half the thickness of the ektexine. The pattern is uniform over the grain excluding the pore area. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 169 There is a large and indistinct vestibule with rough inner surfaces. The floor of the vestibule is flat, the roof thickens between the edge of the vestibule and the exopore open- ing where it thins abruptly (Fig. 12G), the edge is rounded. The outer surface is smooth. Dimensions are in Table 3. This species is separable from the Angophora/bloodwood group by its smaller size. It differs from the rest of the Eucalyptus group by a combination of the notched amb at the base of the vestibule, angular colpi, an exine pattern which extends over all of the grain excluding the pore area and a small pore height (Table 3). Eucalyptus luehmanniana Grains have straight to slightly concave or convex sides. The amb is rounded. Colpi are arcuate, the edges are rough and enclose an apocolpium 5-8ym wide. The polar region is one of two types; either the apocolpium contains a polar island 3-6 4m wide (which has thinner ektexine with granules (Fig. 12J)); or the colpi merge into an area of strong rugulate patterning of normal ektexine (Fig. 11G). The polar island is found on one side only and no specimens with polar islands in both sides were observed. Exine is smooth or faintly scabrate (Fig. 121). The patterning is most pronounced at the colpal edges. There is a small indistinct vestibule with rough inner surfaces. The floor is straight and the roof is thickened when compared to the rest of the ektexine and tapers abruptly to rounded edges at the exopore. Dimensions are in Table 3. This species is distinguished from the Angophora/bloodwood group by its smaller size. It differs from the rest of the Eucalyptus group by rounded angles, arcuate colpi and exine patterning which is most pronounced at the colpal edges (Table 4). Eucalyptus obstans Grains have straight to slightly concave sides, the amb is rounded and notched at the base of the vestibule. Colpi are straight and angular, enclosing a large polar island which fits the apocolpium closely and is 7-9,.m in diameter (Fig. 5A). The edges of the colpi are broken towards the pole, gradually becoming rough towards the pores. The ektexine of the polar island is thinner than elsewhere and may be covered with scattered granules. Exine patterning is vermiculate, the grooves run radially outwards from the polar region (Fig. 12L). The patterning is more pronounced at the colpal edges. The amb has an indistinct vestibule with roughened inner surfaces. The floor is straight or slightly convex, although if viewed obliquely, it may appear concave. The roof of the vestibule is conspicuously thickened compared to the rest of the ektexine (Fig. 12K), and the whole of the pore has a heavy protruding look. The roof thins abruptly to rounded edges at the exopore. The outside of the roof is smooth. Dimensions are in Table 3. This species is distinguished from all the other species in the Eucalyptus group by the radial grooves of the vermiculate exine pattern (Table 4). Eucalyptus sieberr Grains have straight sides (Fig. 3C), the amb is rounded and notched at the base of the vestibule. Colpi are curved enclosing an apocolpium approximately 44.m diameter. The edges of the colpi are rough. Exine has a fine scabrate pattern (Fig. 12N). The ektex- ine around the colpi is smooth. There is a conspicuous vestibule, the inside surface is smooth. The floor is straight or almost so. The ektexine is the same thickness over the vestibule as that around the entire grain, thinning gradually to the edges of the exopore (Figs 10A, 12M), and is smooth over the vestibule. On some of the grains the areas surrounding the polar region cave in over the weak, polar membrane so that it is obscured. Dimensions are in Table 3. The species is distinguished from the rest of the Eucalyptus and Angophora species by a thin roof and smooth inner surfaces of the vestibule (Table 4). Proc. LINN. SOC. N.S.W., 115, 1995 170 POLLEN MORPHOLOGY OF MYRTACEAE Leptospermum arachnoides Grains have concave sides (Fig. 3B) and sharply rounded ambs. Colpi are straight and do not meet at the pole (Figs 6D, 120). The colpi extend between half and two thirds the distance between the pore and the pole. The edges of the colpi are smooth and not thickened. The endexine is smooth and narrow. The ektexine is thicker (approx. 1.2m) and has a strong verrucate pattern (Fig. 11C). The verrucae are 1.5 to 24m in diameter over most of the grain but become smaller in both diameter and thickness (and thus ver- rucate/scabrate) towards the amb (Fig. 11D). The vestibules are difficult to distinguish due to the intense patterning of the ektex- ine over the amb. The vestibule is small and the inside surfaces are very slightly rough- ened. The floor of the vestibule is flat and the roof thins slightly towards the exopore and has a smaller element size of the ektexine patterning. Dimensions are in Table 6. This species is separated from the rest of the Leptospermum group by its verrucate pat- terning and by its colpi which do not meetat the pole (Table 6). Leptospermum trinervium Grains have straight to slightly concave sides and sharply rounded ambs. Colpi are straight, angular and syncolpate (Fig. 6A). The edges are not thickened and are rough. The apocolpium is less than 0.54.m wide. The ektexine has a granular pattern, granules < or = 1wm wide (Fig. 11E). Ambs have small distinct vestibules with smooth inner surfaces. The floor of the vestibules is straight or almost straight and the ektexine over the vestibule is the same thickness as elsewhere on the grain. The outer surface of the pore has a granular pattern. The roof of the vestibule distinctly curves inwards at the exopore and does not thin. The pore is so small that its structure is distinct only on the best specimens. Dimensions are in Table 6. This species is separated from the rest of the Leptospermum group by a combination of rough colpi edges and smooth inner surfaces of the vestibule (Table 6). Leptospermum polygalifoium Grains have concave sides and sharply rounded ambs (Fig. 4A). Colpi are straight, angular and syncolpate, the edges are not thickened and are smooth and there is usually no apocolpium. The ektexine has a fairly rough and irregular scabrate pattern over the entire grain. Ambs have small vestibules with rough inner surfaces. The floor of the vestibule is slightly convex. The roof of the vestibule is the same thickness as the ektexine elsewhere on the grain and has the same scabrate patterning of the ektexine. The roof of the vestibule distinctly curves inwards at the exopore opening and does not thin. Dimensions are in Table 6. This species is distinguished from Leptospermum arachnoides by its syncolpy and the scabrate patterning on the exine. This species is separated from the rest of the Leptospermum group by the combination of smooth colpi edges (Fig. 7A) and rough inner surfaces of the vestibule (Table 6). Leptospermum juniperinum Grains have concave sides (Fig. 3A) and sharply rounded ambs. The colpi are excep- tionally difficult to see and are not visible on all specimens. Colpiare straight, angular and usually syncolpate. The edges are not thickened and are smooth. Colpi are variable, some- times they have gaps or do not meetat the pole. The exine is smooth. Vestibules are difficult to distinguish but good specimens reveal a convex floor slightly narrower than the endexine elsewhere on the grain. The roof of the vestibule is thin and may thin further towards the exopore or it may remain the same thickness. When visible, the inside surfaces of the small vestibule are smooth. Dimensions are in Table 6. Proc. LINN. SOC. N.S.W., 115, 1995 J.-M. CHALSON AND H.A. MARTIN 171 This species is distinguished from the other members of the Leptospermum group by its smooth exine pattern (Table 6). Leptospermum laevigatum Grains have straight to concave sides and the ambs are sharply rounded. Colpi are straight, angular and usually syncolpate with the edges rough and not thickened (Fig. 12P). An apocolpium may expand slightly to 24m towards the pole. Gaps in the colpi may be present but they are rare. Ektexine has a granular/scabrate pattern with rounded and regular elements over the grain but not extending across the immediate area of the vestibule. There is a narrow vestibule with a concave floor. The roof is the same thickness as the ektexine elsewhere and smooth on the outside. The roof does not thin towards the exopore and is truncate. The inner surfaces of the vestibule are rough (Fig. 10C). Dimensions are in Table 6. This species is separated from the rest of the Leptospermum group by a combination of rough edges of the colpi and rough inner surfaces of the vestibule (Table 6). Melaleuca armillans Grains have concave sides and rounded ambs (Fig. 4B). Colpi are arcuate (Fig. 6B) enclosing a small apocolpium, up to 2um wide, the edges of the colpi are smooth. The colpi have distinct, narrow, thickened borders (<1,1m wide). The exine is smooth. The amb has a distinct vestibule with conspicuously roughened inner surfaces. The floor of the vestibule is straight or slightly convex. The ektexine over the vestibule is gen- erally the same thickness as elsewhere on the grain but on some specimens it appears to become thinner towards the exopore. There is considerable variation of the vestibule roof. On some grains the ektexine covering the vestibule tends to project outwards on one side of the exopore which sug- gests that the roof of the vestibule is somewhat fragile and subject to deformation. The floor of the vestibule, however, maintains its shape. Dimensions are in Table 5. This specimen is most like Melaleuca ericifolia but may be distinguished by the more concave sides and larger, more conspicuous pore (Table 5), smooth exine and lack of a notch in the amb. Melaleuca ericifolia Grains have straight to slightly concave sides, the amb is rounded and notched at the base of the vestibule (Fig. 4C). Colpi are curved, enclosing a small apocolpium or rarely a polar island about 1m diameter. The colpi have rough edges and thickened margins about 1m wide. The exine is smooth around the colpal region and is scabrate (Fig. 11H) in the equatorial mesocolpal regions. There is some variation in the intensity of the pat- tern between grains. Ambs have a distinct vestibule, the inside surfaces are roughened. The floor of the vestibule is straight or slightly concave. The roof of the vestibule is the same thickness as the ektexine elsewhere on the grain. This feature is quite variable with the orientation of the grain, and may not be seen in oblique views. Dimensions are in Table 5. This species is most like Melaleuca armillaris but is distinguished by the scabrate exine pattern and the notches at the base of the vestibule (Table 5). This species is also like M. thymifolia but has thickened colpi margins. Melaleuca nodosa Grains have straight or slightly concave sides. The amb is rounded. Colpi are straight or slightly curved, and the edges are distinctly roughened. The bifurcations of the colpi enclose a large polar island 7-8,.m diameter. Exine has a very fine, distinct, granu- lar/scabrate pattern (Fig. 11F). Proc. LINN. SOC. N.S.W., 115, 1995 Nee, POLLEN MORPHOLOGY OF MYRTACEAE Ambs have a distinct vestibule. The separation of the two layers of the exine is rela- tively narrower than that seen on other species. The inside surfaces of the vestibule are roughened. The floor is straight or slightly concave; this variation is partly due to the aspect of viewing. The roof has a tendency to curve inwards at the edges of the exopore. The ektexine over the vestibule is unthickened and is finely scabrate, like the rest of the exine (Fig. 12Q). The colpi are unusually distinct in this species. This species is most like Melaleuca quinquenervia but is distinguished by its larger polar island (Table 5). Melaleuca quinquenervia Grains have straight to slightly concave sides and rounded ambs. Colpi are angular and enclose a distinct polar island 3-5um diameter (Fig. 12R). The polar island has the same ektexine as the rest of the grain and fits the apocolpium closely (Fig. 5A). The colpi have smooth edges and have thickened margins up to 11m wide. The exine is smooth. Ambs have a distinct vestibule with roughened inner surfaces. The floor of the vestibule is straight to convex. The roof of the vestibule is the same thickness as the ektex- ine elsewhere on the grain and has a smooth outer surface. Dimensions are in Table 5. This species is most like Melaleuca nodosa but is distinguished by its smaller polar island (Table 5). Melaleuca thymifolia Grains have concave sides, the amb is rounded and may have notches at the base of the vestibule. Colpi are straight and are either syncolpate or enclose an apocolpium, with a maximum diameter of lum. There is no visible thickening to the edges of the colpi which are irregular. The exine has a faint scabrate pattern all over the grain. The ambs have a distinct vestibule with the inner surfaces roughened. The floor of the vestibule is flat or slightly concave. The roof thins towards the exopore and is never thicker then the ektexine elsewhere on the grain. The outer surface of the pore has a faint scabrate pattern. Dimensions are in Table 5. This species is most similar to Melaleuca ericifolia but the sides of the amb are markedly concave and there are thickened margins to the colpi (Table 5). This species is also similar to Melaleuca armillaris but the colpi are straight and the visible polar mem- brane is 30m Angophora/bloodwoods A* Apex/base > 30um B Colpi edges broken Angophora/bloodwoods B* Colpi edges rough C Roof of vestibule thickened Eucalyptus Cc* Roof of vestibule not thickened D Colpi edges thickened Melaleuca D* me edges not thickened Colpi edges smooth F Colpi arcuate Eucalyptus sieben | Sh Colpi angular Leptospermum E* Colpi edges rough or irregular G Large polar island 6-84.m Melaleuca nodosa G* ae polar island Vestibulum roof thins gradually towards pore opening Melaleuca thymifolia H* Vestibulum roof is truncate at pore opening = Leptospermum KEY TO THE SUBDIVISION OF THE ANGOPHORA/ BLOODWOOD GROUP A Exine pattern of vermiculate grooves Angophora costata A* Exine pattern smooth or slight undulations B Edges of colpi irregular, vestibule large and clear Eucalyptus eximia B Edges of colpi rough, vestibule small and indistinct Eucalypatus gummifera Angophora hispida Proc. LINN. SOC. N.S.W., 115, 1995 A* J.M. CHALSON AND H.A. MARTIN 175 KEY TO THE SUBDIVISION OF THE EUCALYPTUS GROUP Roof of the vestibule not thickened Eucalyptus sieberr Roof of vestibule thickened B Exine pattern vermiculate Eucalyptus obstans B* _ Exine pattern scabrate or smooth C Angle of amb rounded D Colpi angular Eucalyptus camfieldii D* Colpi arcuate Eucalyptus leuhmanniana C* Notch at base of vestibule E Polar island 4-9y.m, with the same ektexine as the rest of the grain Eucalyptus botryoides E* Polar island with rugulate/granulate ektexine or polar membrane with granules Eucalyptus haemastoma KEY TO THE SUBDIVISION OF THE MELALEUCA GROUP Colpi straight/angular B Polar island 3-5u.m, colpi edges thickened Melaleuca quinquenervia B* No polar island, colpi edges not thickened Melaleuca thymifolia Colpi arcuate C Polar island 7-8.m, colpi edges not thickened Melaleuca nodosa C* Nopolar island or rare polar island <2u.m, diameter, colpi edges thickened D Exine pattern scabrate in mesocolpal areas, smooth around colpi, colpi edges rough Melaleuca ericifolia D* — Exine smooth, colpi edges smooth Melaleuca armillaris KEY TO THE SUBDIVISION OF THE LEPTOSPERMUM GROUP Exine patterning verrucate Leptospermum arachnoides Exine patterning smooth or scabrate B Colpi do not meet at the pole C Exine smooth Leptospermum juniperinum C* Exine scabrate or verrucate Leptospermum arachnoides B* —_Colpi meet at pole D Edges of colpismooth E Inside surfaces of vestibule rough, exine scabrate Leptospermum polygalifolium E* Inside surfaces of vestibule smooth, exine smooth Leptospermum junipernum D* Edges of colpi rough F Inside surfaces of vestibule rough, apex/base measurement 17.6 to 20.64m Leptospermum laevigatum F* Inside surfaces of vestibule smooth, apex/base measurement 13.7 to 15.7m Leptospermum trinervium ACKNOWLEDGMENTS We wish to thank Barbara Wiecek, Peter Stricker, Dr. Chris Quinn and the late John T. Waterhouse for help with the identification of specimens. The production of photographs was done by Paul Gadek and Ross Arnett. Barbara Wiecek and Ian Chalson helped with the pre- paration of the diagrams. Proc. LINN. SOC. N.S.W., 115, 1995 POLLEN MORPHOLOGY OF MYRTACEAE 176 (MSN) ‘6L61°X!'66 ‘PIEMOH “WL (MSNQ) ‘GT89MSNN ‘6L6L ALI ‘Uotsinoxy AB0[007 [e1gua4yy (MSN) ‘EL TOSMSNN ‘6L61 X'S “J9PleysIUIM “WY (MSND2) ‘O9ESMSNN ‘GL6L EX FL ‘gsnoyiaieM ‘J [pue uum” [5 (MSN/) ‘L661 X66 ‘T9XeId “A (MSNN) ‘P6OGIMSNN ‘1861 ST ‘slurs ’s (MSNA) ‘8961 Ar 91 ‘ueudeyD ‘ay (MSNA) ‘O8L4EMSNN ‘8Z61 ‘Aen “Sv (MSNM) ‘8Z661TMSNN ‘G86L XE GT “JopreyssiUlM "WY (MSN/1) ‘6L61 SPIe4193UIM AY (MSNM) ‘GETFIMSNA ‘S86 X Ol ‘UsuUeM “Sg pure Ya0aIM “| (MSN/) ‘POST HLT ‘Te Xeld “Cd (MSNN) ‘F96L''6I ‘lXPIa “A (MSNN) ‘O8EMSNN‘L961'“0L ‘asnoyiaiem Lf (MSNNQ) ‘6S8MSNN ‘6961 ‘esnoysaiem ‘Lf (MSN) ‘PIeywey ‘f (MSNN) ‘TSIMSNDA ‘696 THES ‘ssnoysare A Lf (MSNA) ‘8S6L' FTE ‘TXeld ‘ (MSNN) ‘bLEGMSNN ‘6861 18s ‘esnoysaieM Lf (MSNN) ‘GSP LMSNN ‘¢8 dy ‘uospeyD ‘Wf (MSNM) ‘6961 WSs ‘UMOIg WY (MSNN) ‘L961'°*'6s ‘AsnoysaieM “Lf (MSNN) ‘TIS6MSNN. ‘6261 “XO “A (MSNN) ‘FLOLMSNN ‘L661 “Xt'ST “VepIeg ‘O uondsa][0D yoRag UIeIOM “HILL 381 US axe TS, WIWS Yd TEN [PAoy ‘yoery, AapIeyy WION Peoy Jo SULT S [Jog WOLF WIR"O “py IID s UeMOg peoyx WON (19ATY OO) “1U) ‘yaor1p ue y,—sduremg SUuOI2; 01 YORIT, ATUeI, “SAY PLIP{ YON asnosag &] uoo0se7] Ay 29 Id TEN [ehoy Ad TEN ‘SW PN ‘\uiog s8pLiquozney 0} 9NOI Ua [Te II1y “Id TEN [eAoy ‘e]jowRNeM 0} JJouIN) “Iq suasays Weg IIS aseyy 1es8-suLny JeA PUO| “Yd TEN [ehoy ‘youn [nqunie\ Ad JEN YIeID ooresury ajoomeay{ JsdIOJ S$ YIUIAT suapiey vay ‘Tyemsung qd TEN [edoy ‘durems uoqai{ “AMY oyloeg ‘wreysurngo ‘QUOUZT pure eyeAIep] UVaMIIg “‘aqqty Sig “Yq TeN [ehoy “py eusspung ‘yor.n Aare Ayyes0T wa|]od 606 9906 6606 1606 6606 8606 L606 6106 9606 6106 1606 $906 6506 6906 8006 OL06 6606 1906 TéL1 6906 0906 9906 L106 8906 ‘ON OPUS TENW “Yy (seuyeeg) wnywsinav) wnusadsojdaT wptusg wnuusdruntl wnusadsoidaT “qstjeg wnzyofyosQod wnusadsojdaT uosdwoy J, ‘[ (yytuIg) wneauu wnusadsoiqaT Jouyiaey saprouyrviv wnusadsoiqaT wpruig vyofrudys vananpayy DACA LS (Ae) viwauanbuinb vanavpaywy YIUIg (IaUIIIeD “xd JOS) vsopou vanavjayy yg v77ofroUa vInaD/aW. YU (IauIIeD Xa JOS) supp vIn DAW uosuyo[ "T uagars srad(qvong IH 128 uosuyo[] sunjsqgo snadtvony ‘]INe| ‘y vunwuvuyany srad{qvong YIUIS DUojspUaDYy sradQDvINT uapreyy wppaefuns snadipvony yturg saprodyog sniqQqvong “IYIOPT (taujIae‘y xa ‘[OS) viafewwns snjqQqvong Taneyps viwexa srqgavong uaylig (tauq1ae)) 7p)809 vioydosuy [exer (ytuig) vprdsry vioydosuy satveds ‘T66[ ‘UspseH jae sono y Ne [Py ‘SsaTeM YINOS MON Ur aie SUONIOT [TV ‘Saye YINOS MAN JO AtISIIATUL) ‘SIDUIIDG [LIIBOTOIg JO [OOYISG “UONIIT[OD UIaT]Og ULIpoyP 9) UL aquINU aprs UaT]Og ‘Kpnys sof pajoajas suamwads [ TIVE wnuiagsoigaT DINIDIAW snggappong poompooygq /noydasuy dnoay 1995 115, Proc. LINN. SOC. N.S.W., 177 J.M. CHALSON AND H.A. MARTIN (33) (F§) (€°9) (6°S8) youHsipul saa0o13 6606 PT6G PLAYS O'LE-G'E6 “‘unIpaul oye [NITULIOA ysnoi ‘endo DIUIXd “T (FS) (IF) (18) (09g) younsipul 66CT 6SF'E 8 66S 0'0F-0'EE ‘yeus ayeynpun usyO1q niafnuuns “iy (3°) (g°¢) (¥°¢) (€ 1) younsipul PrOSG 696F 886s €'G€-G'96 ‘yews ajeynpun uoyoiq ppigsey “Y (Car) (3'F) (39) (0°4¢) SaA0013 6606 66'S 6967 0'°8&-0'6E Jeajo ‘]feus SETNITULTOA senso DIDISOI “VW (ueouwl) (ueau) (ueaul) (ueaw) winyiInqnsaa ulayed dUIXa tdjoo sataadg uid asuel urn asuel ul asuel ud asues alg jedjososayyj josaspqy UIPIM [TEM IYsI9Y 310d UIPIM 210g aseq /xady paipnys sawads poompoog /e1oydosuy ay) fo uajog ay} fo uoseunquory ¢ TIVE (re2]2) (a}e[NITULIDA) (ayenoe) younNsIpul ysno1 0°83-0'61 ayeiqeos ou 10 sah 64 peuaypiuy ysno1 Jepn8ue ou s0sak g sapadssndQvenziaiQ G youNsIput ayePNITWULIAA ysno1 10 usyxoiq afqetiea Jo reap ysnoi L'9F-G'ES Jo ayeynpun sak 0-0 pauayoipy JoIpNsainm = puevayenoie ou Josak F poompooyq /nioydosuy ¢ reajo woours 0°€3-0'°81 ayeiqeos ou F uly tmoous ayenose sak I ea a pousyoryy ayeiqeos aqdeul Ie] Jepn3ue Teajo ysnol GUc-L 91 jutejio yoous ou JO sak 8-0 uy) =-n8aimsoy8nor = lo ayend1e_ = ou IO sak G pongynay Zz » (JepnueIs8 qoouls “yoouls ‘a}8oN.I9A) yoUNsSIpUl Jo y8nol GBSL sl aearqeos /ayenuess ou ie uly) ysnoi 10 yJoous Jejnsue ou G wmnusaqsoiqgayT | winjnqnsaA = wmnynqnsaajo (wi suri) Sutusayjed purest (ui) ozs = aynqnsaa a8pea jedjoo idjog = aynqnsea —_ satgads aweu dnoiy aoeyins apisuy = aseq /xady ouixy Jejog umidjosody jojooy joadky, joaseq jo‘on 7e YDION (dno13 ayy 0) uondaoxa aj8uls) , parpnys s¢noss seareysA ay fo uajog ay) fo uosenquoy Z AIA I, Proc. LINN. SOC. N.S.W., 115, 1995 POLLEN MORPHOLOGY OF MYRTACEAE 178 (8°9) purest Uo dUuIxa}{9 (61) (€°GZ) (16) 81-6 ayenuess ur SoG sa8pe [edjoo PLES E6 £GC-6'E ysno1 pausyormy 6-Z purjstrejod = ayejnorw9A =v paduNOUOd s10U Iepnsue sod SUDISGO “J sajnues3 YIM (8'T) JULIqUIDUI IO vaIe G3-0'I (69) ajejnues3 /ayepnsn1 ayeiqeos (¢'3s) (0'F) 8L6¢S YpIM ayedyoot21q yurey GPES-L1Z 6 E'S y8no1 pausyoim IO 9-¢ purysi sejod 0} yJoours ure18 I9AO UdAd Iepnsue sod DULOJSDUADY “FY (GT) vole 0°6-0'T (9'F) ayepnuess /ayen3n1 ayeiqeos (FS) (Fg) 6S-6'S UPI aredjoona1q JUIey sa8pa [edjoo CFS-9'6I1 6'E-6'S ysno1 peuayomp IO 9-¢ purysi rejod 0} Moows 3e paounouo.d a10ul arenoie ou pupluuDuyan) “J (8'F) (41) (0°93) (G'T) 69-66 gueIquigul 6641 saspea : F'LE-G'ES 0'S-0'T ysno1 peuayormp sejod yoours a}eiqeos edjoo 72 yJoours iejnsue ou upparfuno 5 (9°F) (8'T) (9°GZ) (0°) 69-66 64 064 T ; 0'LE-0'FS 6'S-0'T y8no1 poueyory) pueyjst 1ejod ayeiqeos UIVIB IJAO UIAI iejnsue sod saprokiyog “iy (¥°g) (€1) (0°03) (ST) 8 L-6'E sueIquigul CLOT saspa 0'S3-0'81 0'S-0'T yoouws uly) iejod yyoous aeiqeos [edjoo 72 yyoows arenoie sod maga “5 (wi uesut) (wi uvaut) (wi uvout) aynqnsaa wmidjosody = (wiv uevaut) yUtayed aurxa djop quire ut satvadg aseq/xody iY4819y a10g UIpIM a10g jo jooy ssoUyory J, ULUONRLIRA YDION aynqnsaa jo uiraned aovyins 19UU] JUIX] vou a10d 190 10U >», ‘parpnys savadg smddyeony ay) fo uanpog ay fo uoswpquuor) p ATAV 115, 1995 Proc. LINN. SOC.N.S.W., 179 J.M. CHALSON AND H.A. MARTIN (8'F) 6'S-6'§ (6°81) (L°3) 9069 LI FE06 (G°g) 69-6 F (8°81) (9°3) 9°06-L91 6606 (SF) 69-66 (G81) (LS) 906-L'91 PE-0'6 (€'F) 6766 (8°81) (0°3) 9069 LI 66OT (09) 8 L6'¢ (6°81) (Gg) GéoL 91 66-06 (ui uvout) (wi uvou) aseq /xady YIPIM (uevour) 1yBIay 910g (T'T) GT-O'T sah ayeiqeos (0'T) GTOT ou qjoours (TT) TOT aqeiqeos ou /xejnues3 (11) Vo0T sak ayeiqeos (¢'1) SLOT ou qoours ‘quie ul (ui7 urou) YDION usaned dUIXY ou G-g SoA 8-1, SOA G> del ou “WIRIp purjsi rejog ou sok ou sod sod 1djoo jo uIs.1eUL pousyoIy LL Jepn3sai1 1ysrens woouis Jemnsue ysno1 ayenoie ysno1 ayenoie qoouis ajenoie sa8pa odeys tdjop ‘parpnys sarvagg eonaelaW ay7 fo uajjog ay) fo wosupquor) G ATAV 9ARIUOD aaeouod AN ysis 07 148re.ns aaevouos ANySiys 0114 3rens aaevouod ANYySiys 07 14 8te.ns 9ARIUOD quire Jo sapig pyofrucy) W pusauanbunb ‘wr DSOpoUu "JA Dyofiaua "WW SUD]]NULD "TY satvadg Proc. LINN. SOC. N.S.W., 115, 1995 POLLEN MORPHOLOGY OF MYRTACEAE 180 (OL) (LT) (63) aaeou0d O'1-0'T (€'F1) pepunor 06-G'T 66-G'T Anystys UIe.IB JO |v 19A0 LGI-L'€1 Andniqe wy qjoouis ysno1 ajedjoouds 07 143rens ayeiqeds /repnueis WNULGULLY “TJ (ST) (G'S) (3'F) aavou0d 0'-0'1 (€'61) 660% 6F-6% AnySys sai0d 1¥ WOOUIS 07 9'0¢-9'LT ajeounn YN 4010 (0.0) ysno1 ysno1 ayedjoouks O11YSIeENS — ayerqvos /reNuLIs wnjoswan) “T (LT) (S'3) ayedjoo1a1q (on) (€ FL) Sutiadey 06ST 66-OT pure sde3 a1ei CLOT LYEUL SI Ayyenpei3 X9AUO09 qoours qoours ajedjoouds JAvDU09 qoouis wnuuagrunt “7 (€1) (33) (8S) 03-01 (3°S1) pepunor 6607S 6 FSS ures JO wo OLI-L'E1 Andniqe xaauoo Apysys ys8no.1 toous ayedjoouds JAvdU0d [TW 419A0 alerqeos —-wnzpof2vs(od “7 (61) (G3) (8°¢) OST (0°03) Pepuner 6606 6 FG's aeaqeos /repnuer1s G6é69 LI Andniqe yey y8no1 qoouis ajedjoo191q dARIUOD 0} Q]BON.LIOA SPPLOUYIDAD “T (ui1l ueout) a3pa (wil ueout) (ui! uvaut) 1djoo jo umidjosody quie jo (wi! ura) saiads aseq /xady a10doxy 1yS3IaH WPM saspy Sapls ssauyory J. IOOY saoevyins Surusayed anqnsa,, JQUUI QUIX] arnqnsaA, ‘parpnys savady unuitadsoydaq ay7 fo uaqpog ayy fo wosunduoy OQ ATdV Proc. LINN. SOC. N.S.W., 115, 1995 181 J.M. CHALSON AND H.A. MARTIN GSS Gs Gs Gg myvjsoo “y oar G3 GZ G38 oor pprdsry “W GL C36 Druaxa “sy c0 0'€ 06 06 09 0°81 piafiuund “7 ‘ade1a0y Gs ggg oral 0902 G6 GL I6L1 Gs 0ST GLE 0'Sh 1902 GZ 0ST Cor Gs LT 6606 G G18 GL, 906 ‘Deafrmunsd “sq es a ee ee RE ees ( SE nl ee % % % % % % seeds ag 34 «V8 pur gg ‘B1y ‘A ‘A6 “B1y d‘O‘d ‘v6 34 Od ‘ve 3 a9 ‘314 nedjoonoaig puvyst 1ejod 1djoo jo sueiquisu 1ejod sdve3 jedjory aad 109 1djop [Tews uONIUTJap 100g uONIUTJop 100g aja,durosuy a nS ED ete, ee Oh SP ee ee OUR eer ee ee ures13 9U0 Ul 194980} BULLINIIOC 5 ‘qnory poompoojg/e1oydosuy ay) u2 uoyoun, oyfrvadsnaquy 1 vIav L$ Proc. LINN. SOC. N.S.W., 115, 1995 182 POLLEN MORPHOLOGY OF MYRTACEAE Amb Cc Ek S En P | A Exo ———— y M Fig. 1. Generalised Myrtaceae pollen grain, polar view. Ek. Ektexine; En, Endexine; A, Apocolpium; Exo, Exopore; Amb, The amb; C, Colpus; S, Sides of amb; P I, Polar island; V, Vestibule; M, Mesocolpal region. Pore Width Wall Thickness Apex /Base Pore Height Fig. 2. Measurements made on pollen grains. A B C Fig. 3. Sides of pollen grains: A, concave; B, convex; C, straight. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 183 AN A A B C Fig. 4. Angle of the amb and notch at the base of the vestibule: A, sharply rounded angle; B, rounded angle; C, rounded angle with a notch at the base of the vestibule. A C Fig. 5. Types of apocolpia: A, apocolpium with closely fiiting polar island; B, apocolpium with small irregular polar island; C, apocolpium with granules on the polar membrane. Proc. LINN. SOC. N.S.W., 115, 1995 184 POLLEN MORPHOLOGY OF MYRTACEAE ABA ANS Fig. 6. Types of colpi: A, syncolpate; B, parasyncolpate with arcuate colpi; C, parasyncolpate with angular colpi; D, brevicolpate; E, syndemicolpate. { A B Cc D Fig. 7. Types of colpal edges; A, smooth; B, irregular; C, rough; D, broken. Fig. 8. Colpal gaps; A, 1 gap; B, 2 gaps; C. 3 gaps. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 185 INPSBRA DESL Fig. 9. Other colpal irregularities associated with the Angophora/bloodwood group: A-D, incomplete def inition of apocolpium, with or without colpal gaps; E-G, poor definition of colpi around apocolpium. Proc. LINN. SOC. N.S.W., 115, 1995 186 POLLEN MORPHOLOGY OF MYRTACEAE R oF ee Sm Unth FI V C Fl Rgh Fl Th Ro C Fig. 10. The pore area: A, Eucalyptus sieberi; Roof tapers gently to the exopore edge and is unthickened. Vestibule large and inner surface smooth. Floor of vestibule flat. B, Eucalyptus botryoides, Thickened roof over small vestibule rounds off abrupt- ly at exopore edge. Floor of vestibule convex, C, Leptospermum laevigatum; Roof over small vestibule is unthickened and is truncate at exopore edge. Floor of vestibule concave. D. Angophora costata; Roof rounds off abruptly at exopore edge. The amb is notched at the base of the vestibule. Vestibule roof is thickened and has rough inner surfaces. The vestibule floor is concave. C, Colpus; Ek, Ektexine; En, Endexine; Exo, Exopore; Fl, Floor of vestibule; N, Notch at the base of the vestibule; R, Roof of vestibule; Rgh, Rough inner surfaces of vestibule; Sm, Smooth inner surfaces of vestibule; Th, Thickened roof of vestibule; Unth, Unthickened roof of vestibule; V, Vestibule. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 187 C D Fig. 11. A-D. Types of exine pattern. Optical section with plan view below. Scale bar indicates 11m; A, vermiculate patterning of Angophora costata; B, undulate patterning of Eucalyptus gummifera; C, verrucate patterning of Leptospermum arachnoides; D, verrucate/scabrate patterning of Leptospermum arachnoides. Proc. LINN. SOC. N.S.W., 115, 1995 188 POLLEN MORPHOLOGY OF MYRTACEAE Fig. 11. E-H. E, granulate patterning of Leptospermum trinervium, F, granulate/scabrate patterning of Melaleuca nodosa; G, rugulate patterning of Eucalyptus luehmanniana; H, scabrate patterning of Melaleuca enicifolia. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 189 : ee SS se O 10.20 | ium Fig. 12. A—K. All photographs are from untouched negatives. A, Angophora hispida, showing roughened inner sur- faces of vestibule and undulate exine pattern. B, Angophora costata, showing thickened pore roof and concave vestibule floor. C, Eucalyptus eximia, showing slightly thickened pore roof and small vestibule. D, Eucalyptus eximia, showing rough colpal edges. E, Eucalayptus gummifera, showing flat vestibule floors. F, Eucalyptus gummifera, show- ing colpal gaps. G, Eucalyptus haemastoma, showing thickened pore roof. H, Eucalyptus haemastoma, showing rough colpal edges and angular colpi. I, Eucalyptus luehmanniana, showing faint scabrate patterning in mesocolpal areas. J. Eucalyptus luehmanniana, showing membrane with granules in the apocolpium. K, Eucalyptus obstans, showing greatly thickened vestibule roof. Proc. LINN. SOC. N.S.W., 115, 1995 190 POLLEN MORPHOLOGY OF MYRTACEAE Fig. 12. L-S. L, Eucalyptus obstans, showing broken edges of colpi and vermiculate grooves in the exine. M, Eucalyptus sieberi, showing thin pore roof. N, Eucalyptus sieberi, showing scabrate exine pattern. O, Leptospermum arachnoides, showing brevicolpate morphology and verrucate exine pattern. P, Leptospermum laevigatum, showing syncolpate morphology and granular/scabrate exine pattern. Q, Melaleuca nodosa, showing a large polar island and thin-roofed vestibule. R, Melaleuca quinquenervia, showing a smaller polar island and thin tapering roof of the vestibule. S, Eucalyptus haemastoma, abnormal four-pored grain. References ANDERSON, S.T., 1960. — Silicone oil as a mounting medium for pollen grains. Danmarks Geol. Unders. iv Raekke 4, 1-24. BEADLE, N.C.W., 1981. — The Vegetation of Australia. Cambridge University Press, Cambridge, U.K. Ist. edn. BINDER, R.M., 1978. — Stratigraphy and pollen analysis of a peat deposit, Bunyip Bog, Mt. Buffalo, Victoria. Monash Publications in Geography No. 19, pp. 52. CHALSON, J.M., 1983. — Palynology and Paleocology of Jibbon Swamp, Royal National Park, N.S.W. BSc. (Hons.) Thesis, Botany School, University of N.S.W. CHANDA, S., 1966. — On the pollen morphology of the Centrolepidaceae, Restionaceae and Flagellariaceae with special reference to taxonomy. Grana Palynol. 6, 355-415. CHURCHILL, D.M., 1957. — An Investigation of Some Pollen Bearing Sediments from South Western Australia. BSc. (Hons.) Thesis, University of W.A. ——— , 1961. — The Tertiary and Quaternary Vegetation and Climate in relation to the Living Flora in South Western Australia. PhD. Thesis, University of W.A. , 1968. — The distribution and prehistory of Eucalyptus diversicolor. F. Muell, FE. marginata Donn. ex Sm., and E. calophyllaR. Br. in relation to rainfall. Aust. J. Bot. 16, 125-51. COLHOUN, E.A., van de Geer, G., and Mook, W.G., 1982. — Stratigraphy, pollen analysis and palaeoclimatic inter- pretation of Pulbeena Swamp, Northwestern Tasmania. Quat. Res. 18, 108-126. Cookson, I.C., and Pike, K.M., 1954. — Some dicotyledonous pollen types from Cainozoic deposits in the Australian region. Aust. J. Bot. 2, 197-219. Proc. LINN. SOC. N.S.W., 115, 1995 J.M. CHALSON AND H.A. MARTIN 191 Dodson, J.R., 1974. — Vegetation history and water fluctuations at Lake Leake, South-eastern South Australia. I. 10,000 B.P. to Present. Aust. J. Bot. 22,'719-41. , 1977. — Late Quaternary paleoecology of Wyrie Swamp, South-eastern South Australia. Quat. Res. 8, 97- 114. , 1979. — Late Pleistocene vegetation and environments near Lake Bullenmerri, Western Victoria. Aust. J. Ecol. 4, 419-427. , 1982. — Modern pollen rain in South-eastern new South Wales, Australia. Rev. Paleobot. Palynol. 38, 249- 268. ERDTMAN, G., 1952. — Pollen Morphology and Plant Taxonomy, Angiosperms. The Chronica Botanica Co. Waltham, Mass. U.S.A. Ist edn. GabDEK, P.A., and Martin, H.A., 1981. — Pollen morphology in the subtribe Metrosiderinae of the Lepto- spermoideae (Myrtaceae) and its taxonomic significance. Aust. J. Bot. 29, 159-84. HARDEN, G,]., 1992. — Flora of New South Wales Vol. 2. University of New South Wales Press, Sydney. Ist edn. KremP, G.O.W., 1965. — Morphologic Encyclopedia of Palynology. University of Arizona Press, Tucson, U.S.A. 2nd. edn. Lapp, P.G., 1979. —A short pollen diagram from rainforest in highland eastern Victoria. Aust. J. Ecol. 4, 229-237. McINTYRE, D.J., 1963. — Pollen morphology of New Zealand species of Myrtaceae. Trans. R. Soc. N.Z. 2 (7), 83-107. PIKE, K.M., 1956. — Pollen morphology of Myrtaceae from the south-west Pacific area. Aust. J. Bot.4, 13-53. Rose, S.L. and Martin, H.A., (in prep.). — A comparison of the vegetation and surface pollen spectra of the Thirlmere Lakes region, New South Wales. Proc. LINN. SOC. N.S.W., 115, 1995 = .*% Cae iA » a i aren’) wh jag +o >) ee ial ’ 1 == ; ath 08. (hE a ses Ging 194 ues wn ; Jt. : } ha ay ’ @ lipped Bow "ion : byes, Ltuleoricy fl Wy oe ein |) Mi The Veer ad Uptieions 9 pemeliy th hrek Linivereny od i—>G its (weet qariow) bul “ rit ebeperln K, fie yireatlor: (pra resus ne de One. do, ti Mock. W writes cf iplaons ry | PigA Aremertiay Fegeer, tag.) Bal . “yD, a | Lase@ Se 5 TOA CAP, Bh ORM SF *! agri ies iM. -i086 =< fen LNT-£19 Jake) ry. tay OP ba } AV aa, Gare i ne ‘ his as » pies a ies heen Ty NN, epee My. wil Toth, MSA BSE Cle ; = } i ry Ruta Orel FP é eh VN erh ewe f. empresa (uaa, ee a circ oril sulaces Tony bie ti % \ Stratigraphic Palynology of the Murray River Valley in New South Wales HELENE A. MARTIN HELENE A. MarTIN. Stratigraphic Palynology of the Murray River Valley in New South Wales. Proc. Linn. Soc. N.S.W. 115: 193-212 (1995) The Murray River region, from Cohuna to Albury, has a complex stratigraphy which has a bearing on groundwater quality. The basement is Palaeozoic, Early Permian and Mid Permian in age, and the last-named is confined to the Oaklands-Coorabin coal- fields. The Cainozoic sediments range from the late Eocene, the oldest, found in the west of the study area, to Pleistocene, the youngest, in the east. Low salinity groundwater is found in the Tertiary sands and gravels. Helene A. Martin, School of Biological Science, University of New South Wales, P.O. Box 1, Kensington. Australia 2033; manuscript received 22 September 1993, accepted for publication 21 September 1994. KEYWORDS: Murray Basin, Murray Valley, Permian, Cainozoic, palynostratigraphy INTRODUCTION The Murray River system can be traced back to Eocene time, at least 50 million years. Most of the major tributaries have existed from Eocene time also (Stephenson and Brown, 1989). The position of the mouth of the Murray River has been controversial. One school of thought proposes that the river extended across the present line of the Mt. Lofty Ranges before these ranges were uplifted, along what is now the course of the Broughton River and emptied into Spencers Gulf (Williams and Goode, 1978). The other school of thought claims that the outlet of the Murray River has always been east of the Mt. Lofty Ranges (Stephenson and Brown, 1989), although the location of the mouth was variable and intimately associated with the marine transgressions of the Murray Basin. For a fasci- nating account of river history and this controversy, see Stephenson and Brown (1989). The upstream, non marine part of the Murray River has a similar long history. A study of the pre-Tertiary basement contours under the southern Riverina Plain in Victoria (Macumber, 1978) suggest that a drainage system ancestral to the Murray River was in existence in Eocene time. From Oligocene to early Miocene time, it is not possible to demonstrate the presence of a co-ordinated drainage system because the Murray River Valley was then probably little more than a swamp. In late Miocene times, a co-ordinated system re-appeared, and the palaeo-Murray flowed into a deep marine embayment in the vicinity of Cohuna (Macumber, 1978). The study area (Fig. 1) extends along the river from the southeastern edge of the Murray Basin near Cohuna and Wakool in the west, to Albury and Holbrook in the east. The deeper sediments in the west become shallower with distance upstream. The deep- est, oldest Tertiary sediments are late Eocene and the shallowest, at the eastern end of the study area, are late Pliocene-Pleistocene. Tertiary sediments form the cover beds of the Permian Oaklands - Coorabin coal basin. Coal was mined from1917 to 1920, when production ceased because of water prob- lems. Intermittent production of coal continued from 1934 to 1958, and the colliery closed in 1959 (Bembrick, 1975). There is recent interest in these coal deposits, but reopening the mines would raise serious environmental issues as the mine waters have a high salt content. The valley sediments are important for groundwater. Interest in groundwater Proc. LINN. SOG. N.S.W., 115, 1995 194 PALYNOLOGY OF THE MURRAY VALLEY e° 36078 36558 © 36102 536201 ¢ Yallakool * Deniliquin © 36587 36 @s65e5 @ 36586 0 36283 @ 36588 36582 Tt | < e 36584 locum ® Mathura Lake Urana co 30497 363990. me —— — Limits of the Oaklands-Coorabin Coalfield Jerilderie \ a B ) T , @ Town § 96638 Sonne © Bore, reported in Martin (1984b) 3 @ Bore, this study Games = Main faults of the Ovens Graben AS doula 36292 NOAK 12 e eco 2 | ¢ Daysdale @ Culcaim a OAK 13 @ \ OAK 11 7: *®OAK7°°° co 1 30763 Holbrook @ rerrh \ te 2 OAK 9 NE . @ 36394 \ © 36390 036292) a 36314 id soo) Ee §,36306 @-. # Hopefield i “oak 10 36352 $36295 : ae 26354 ) 36308 Shoat + gttowlong * Lorowa @ 36376 3641 Albury 25281 ¢ Fig. 1. Locality map. The limits of the Oaklands - Coorabin Coalfields are from Bembrick (1975). began before the turn of the century when gold mining encountered water in the auri- ferous ‘deep leads’ in Tertiary sands. As the area became more populated and settled, bores were drilled to the upper aquifers for stock and domestic water supplies (Williams, 1989). This paper reports the stratigraphic palynology of the Tertiary sediments and Permian basement. Bores sunk for exploration of the coal fields have been resampled and the results are reported here also. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 195 MATERIALS AND METHODS Both core and cutting samples have been used in this study. The possibility of con- tamination is greater with cuttings, both from carry down with the circulating mud and from cavings, but with proper drilling and sampling procedures, reliable samples may be obtained. For investigative drilling, the mud is circulated until it is clean of the coarse frac- tion and this practice greatly reduces contamination. If there is contamination it can be detected, either in the sediments themselves or in the preparations. A number of bores penetrate both Tertiary and Permian sediments and the amount of Tertiary contamina- tion in the Permian may be assessed. Usually there is no or very little contamination unless sampling has occurred close to the contact. Barren samples may occur anywhere in the sequence and this would not be possible with appreciable contamination. While the possi- bility of contamination can not be ruled out completely, cuttings produce consistent pat- terns repeated in bore after bore, and this consistency would not be possible with appreciable contamination. There is thus reasonable confidence that these samples pro- duce reliable results (Martin, 1984a). Preparation techniques used hydrochloric and hydrofluoric acids to remove the mineral material, controlled oxidation with cold Schultz solution, and potassium carbon- ate to clear the residues. The residues were mounted in glycerine jelly. Bores with five digit numbers are those sunk by the New South Wales Department of Water Resources. Bores sunk in the course of coalfield exploration are prefixed Oaklands (Oak) or Coorabin (Coo). GEOLOGY The basement in the eastern region consists of the lower-mid Palaeozoic Lachlan Fold Belt. These rocks are intensely folded, faulted and partly metamorphosed. Extensive but discontinuous Early Permian glacio-marine deposits form a thin veneer over much of this region. In the southern part, the glacial mud flows and tills suggest a close proximity to a glaciated land mass to the south (Brown, 1985; Brown and Stephenson, 1986). In the Oaklands-Coorabin Basin, the Early Permian is disconformably overlain by the three fluvial sequences of the Late Permian Coorabin Coal Measures (Brown,1985). The dominant structural feature in the area is the Ovens Valley Graben, a northwest- southeast trending structure thought to have commenced subsidence in the Early Permian. The Early Permian sediments are found both in and out of the graben, but they are thicker in the graben. The Late Permian sequence is confined to the graben (Yoo, 1982), and are disconformably overlain by a mid Triassic unit (Morgan 1977). The oldest Tertiary unit in this region is the late Eocene to ? mid Miocene Olney Formation that was deposited in fluvio-lacustrine, meandering-channel and extensive flood plan environments. It consists of grey coloured sands, silts and clays, which are fre- quently carbonaceous, and peaty coals (Brown and Stephenson, 1986). Wood is frequent- ly encountered in the bores. The Olney Formation is found in the western part of the study area. The Olney Formation is unconformably overlain by the Lachlan Formation that is equivalent to the Calivil Formation, and is late Miocene to Pliocene age (Williams, 1989). In the valley and where the pre-Tertiary basement is shallow, the Lachlan Formation over- lies basement rocks. The sands and gravels of the Lachlan Formation consist mainly of quartz, and the clays are predominantly grey. There are minor carbonaceous clays. The upper part of the Olney Formation and the lower part of the Lachlan Formation may be difficult to distinguish apart on the evidence of lithologies alone. Marine regression in the mid-late Miocene caused entrenchment of the drainage system as well as subaerial weathering that produced the distinctive Mologa Surface (Macumber, 1978). Transgression during the late Miocene caused sediments to be Proc. LINN. SOC. N.S.W., 115, 1995 196 PALYNOLOGY OF THE MURRAY VALLEY deposited further and further into the highland river tracts, thus increasing the alluvial fill in the valleys. This static reworking of incoming sediments may have contributed to the concentration of gold in some of the sands and gravels forming the ‘deep leads’ (Williams, 1989). The Shepparton Formation overlies the Lachlan Formation and is Pliocene in age (Brown and Stephenson, 1986). The lithology varies widely between the extremes of clay and gravel. The sands are quartzose, with only the upper part containing rock fragments representative of the present catchment rocks. The Formation is characteristically brown and yellow in colour. This Formation reflects a change in river morphology and possibly in climate. It has been deposited by leveed streams which meander within the alluvial zone, causing a build up of sediments. There are soil horizons at intervals (Williams, 1989). The Quaternary Coonambidgal Formation (Brown and Stephenson, 1989) has inset terraces, rather than being topographically higher than the underlying Shepparton Formation. The streams of this time carried relatively low annual discharges, though the discharge could be relatively high during flood events (Williams, 1989). The main aquifers are the quartz sands in the Lachlan Formation which yield low salinity water (Williams, 1989). PALYNOSTRATIGRAPHY Appendix | presents the palynological zones and ages assigned to the bores of this study. Permian The spores and pollen identified in selected samples are presented in Appendix 2 and the ranges of diagnostic species are shown in Fig. 2. Where specific diagnostic species are lacking, the assemblage may be assigned to the Early or mid Permian on general char- acteristics. For example, monosaccates (Barakarites, Plicatipollenites and Potonieisporites) and striated bisaccates (Protohaploxypinus and Striatopodocarpidites) are found throughout the sequence, but the monosaccates are abundant and bisaccates uncommon in the Early Permian, whereas the monosaccates are infrequent and the striated bisaccates (e.g., Protohaploxypinus, Striapodocarpidites) more common in the Mid Permian. Earlier studies of Permian palynostratigraphy place the upper part of the sequence in the Late Permian. Price (1983), reviews the history of Permian palynostratigraphy and places the upper part of the sequence in the Middle Permian. This study follows Price (1983). The oldest assemblages are stage 3a, Early Permian, and the youngest, upper stage 5b (see Appendix 1). Figs. 3 and 4 present two cross sections through the Oaklands- Coorabin Basin. Cainozoic The palynology on numerous bores in the eastern non marine section of the basin correlates reasonably well with the zonation of Stover and Partridge (1973, 1982), con- structed for the Gippsland Basin. There is one exception: the Upper N. asperus Zone, of latest Eocene-earliest Oligocene in the Gippsland Basin is not recognisable here by its original diagnosis. In the Murray Basin, the Late Eocene Middle N. asperus Zone is suc- ceeded by the Oligocene P. tuberculatus Zone. The thick sections of the Oligocene-Early Miocene P. tuberculatus Zone may be subdivided using several quantitative events which have proved useful for correlation, at least on a local scale (Martin, 1984a; 1984b; 1984c). The late Miocene-Pleistocene palynostratigraphy follows Martin (1987). Fig. 5 pre- sents the palynostatigraphy of the Cainozoic and Appendices 3 and 4, the spores and pollen identified in selected bores. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 197 Middle N. asperus Zone, late Eocene Nothofagus spp. are abundant, and most of the pollen is the brassi type. Gymnosperms may be equally or more abundant at some levels, particularly Phyllocladidites mawsoni (see Appendix 3). Species whose ranges terminate at the top of the Middle N. asperus Zone, viz. Proteacidites leghtoni, P. reticulatus and Triorites magnificus, are found here. The last named is restricted to the Middle N. asperus Zone (Stover and Partridge, 1973; 1982). P. tuberculatus Zone, Oligocene-early Miocene Nothofagus spp. are abundant, similar to the Middle N. asperus Zone, but the diversity of species is lower and the diagnostic species of the latter are not present. The P. tubercula- tus Zone is divided into three parts, viz, the A subdivision, with a greater abundance of Phyllocladidites mawsonii and/or Nothofagidites flemingii (early Oligocene); the B subdivi- sion, lacking the diagnostic features of the A and C subdivisions (mid-late Oligocene); and the C subdivision with the upper Nothofagidites flemingii acme and/or an increase in the Myrtaceae content so that it almost equals or exceeds Nothofagus spp. (latest | AGE |stace SPECIES Late | Permian | | | U5 Permian Early Permian LA Dulhuntyispora stellata Secarisporites bullatus Mehlisphaeridium cf. M. fibratum Interradispora versus Didecitriletes ericanus Dulhuntyispora dulhuntyi Dulhuntyispora parvithola Acanthotriletes (Microbaculisporites) villosus Bascanisporites undosus Praecolpatites sinuosus 3b Pseudoreticulatispora (Verrucosisporites) pseudoreticulata Granulatisporites trisinus Acanthotriletes tereteangulatus Marsupipollenites triradiatus 2) 2 a] SS 3a] 8|2 Se RS) Late as Carbonif- 8] 3 erous 8 1G a = 1 Fig. 2. Ranges of Permian diagnostic species, from Price (1983), Kemp et al., (1977) and McMinn (1985). L, lower. U, upper. Proc. LINN. SOC. N.S.W., 115, 1995 198 Oligocene-early Miocene age) (Martin, 1984a; 1984b). P. tuberculatus Zone, A subdivision, early Oligocene Phyllocladidites mawsonii is abundant and in this respect, these assemblages resemble the Middle N. asperus Zone, but they lack the diagnostic species of the latter. Nothofagidites flemingii may be unusually abundant (the lower N. flemingiiacme). 36635 | 8 A Mid N. asperus @ Tertiary « Mid Permian < Early Permian OAK 18 OAK 15 OAK 5 B Upper 6a Upper 5a 3b PALYNOLOGY OF THE MURRAY VALLEY OAK 4 OAK 12 A' 3a Scale Upper 5a @ g -00 Fig. 3. Section A-Al through the Oaklands — Coorabin Basin. For the location of the section, see Fig. 1. 36399 OAK 17 B 36638 Lower 5a 3a Lower 5b ° +50 Scale (m) +100 OAK 2 Upper 5a OAK 16 OAK 14 Lower 5c Upper 5b OAK 1 Lower 5b OAK 11 OAK7 co1 B' 3a Mid N. asperus Upper 5a Upper 4a @ Tertiary « Mid Permian 4 Early Permian Fig. 4. Section BB} through the Oaklands — Coorabin Basin. For the location of the section, see Fig. 1. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 199 P. tuberculatus Zone, B subdivision, mid-late Oligocene These assemblages lack the diagnostic features of both the A and C subdivisions. The Nothofagus brassii type is very common. Occasionally Haloragacidites harrisii (Casuarinaceae) is abundant as well. P. tuberculatus Zone, C subdivision, early Miocene Acacia first appears in the Early Miocene (Stover and Partridge, 1973), and although it may be found in earlier sediments elsewhere, it defines the base of this subdivision in this study area. The Myrtaceae/ Nothofagus ratios below this level are low, and above it, may be high. The upper N. flemingii acme, if present, is concurrent with the increase in Myrtaceae. T. bellus Zone, latest early Miocene-?late Miocene Nothofagus, Myrtaceae or gymnosperms may be the most abundant group. The diag- nostic species Symplocotpollenites austellus (Symplocos) and Triporopollenites bellus ( Gardinia= ‘Randia’ chartacea type) define the base of the zone. PALYNOLOGICAL EPOCH ZONATION PLEISTOCENE | Asteraceae/Poaceae Upper M rtaceae ae PLIOCENE - fe z Nothofagus Lower Myrtaceae MID —_- — MICS T. bellus Million years C subdivision Upper N. flemingii acme P. tuberculatus OLIGOCENE B subdivision Lower N. flemingii acme A subdivision LATE Mid N. asperus EOCENE Fig. 5. Cainozoic palynostratigraphic scheme, from Stover and Partridge (1973) and Martin (1987). Proc. LINN. SOC. N.S.W., 115, 1995 200 PALYNOLOGY OF THE MURRAY VALLEY The lower Myrtaceae phase, late Miocene Myrtaceae and/or Casuarinaceae are the most abundant taxa, Nothofagus is absent, but the gymnosperms and some rainforest angiosperms are present. The Murray River Valley is similar to the Lachlan River Valley (Martin, 1987) in this respect. The Nothofagus phase, ? early Pliocene. Some Nothofagus is present, but only the menziesii and fusca types. The Nothofagus brassii type is absent. Gymnosperms may be unusually abundant. The upper Myrtaceae phase, ?mid-late Pliocene. This phase is essentially similar to the lower Myrtaceae phase. If the Nothofagus phase cannot be identified, then the upper and lower Myrtaceae phases cannot be distinguished apart. Asteraceae-Poaceae floras, Pleistocene Asteraceae and/or Poaceae increase substantially. In the river valleys of the western slopes and Lake George in the eastern highlands, Asteraceae is usually more abundant than Poaceae. There are very few gymnosperms and the rainforest element is absent or very reduced. Polyporina granulata and Tubulifloridites pleistocenicus are frequently present in the Asteraceae-Poaceae floras. In this study, P. granulatais also found in the upper Myrtaceae phase and T. pleistocenicus has not been recorded (Appendix 4). Figs. 6-8 present cross sections through the late Cainozoic sequence and Fig. 9, a pollen diagram. In this study, the Nothofagus phase is more diffuse and less of the discrete horizon than in the Lachlan River Valley (Martin, 1987). 36350 OAK 10 36351 36352 6 © ) 20 @ Tertiary < Early Permian 40 Scale (m) 3a Nothofagus phase Mytaceaeiphase Early Permian bY Nothofagus phase C subdivision T. bellus } 2 4} 3a Early Permian Fig. 6. Cross section C-C! th rough the Cainozoic sediments. For location see Fig. 1. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 201 36356 36355 D 36354 D' Myrtaceae phase Pliocene (0) Early Permian } Nothofagus phase 20 Scale (m) 40 i Tertiary <1 Early Permian Early Permian Fig. 7. Cross section p-p! , near Mulwala. For location, see Fig. 1 36311 36306 36295 36303 36394 E Myrtaceae phase Myrtaceae phase Myrtaceae phase Pliocene Nothofagus phase ? Nothofagus phase } Pliocene Saif i) Myrtaceae phase lothofagus phase }w aN Nothofagus phase 20 Early Permian Tertiary Early Permian < Early Permian re) Fig. 8. Cross section E-E!, through Hopefield. For location, see Fig. 1. Where there is insufficient evidence for a zone or age determination, an approximation of the age is based on general characteristics of the assemblages. Proc. LINN. SOC. N.S.W., 115, 1995 202 PALYNOLOGY OF THE MURRAY VALLEY DISCUSSION The palynology shows a complex stratigraphy in the Murray River Valley. The Early Permian basement may be as shallow as 63 m, and the Tertiary sediments may be as deep as 260 m. The stratigraphy is extremely important to groundwater quality. Low salinity waters are only found in the Tertiary sands and gravels. Water in the basement strata has higher salinities. The Permian sequence recorded here is in general agreement with previous reports. Triassic palynofloras previously reported by Morgan (1977) have not been found and samples from sediments thought to be Triassic in age proved barren. The Eocene to Miocene sequence is similar to that found elsewhere in the Murray Basin. The vegetation of this time was ‘semi-swamp’ forest, similar to that of the lowland flood plains of New Zealand (Cockayne, 1958). The lowlands on the flood plains were sub- jected to prolonged flooding. Pools of water were frequent, but patches of dry ground were always present. Much of the ground was saturated for long periods and thick peat was common. Most likely, Nothofagus grew on the dry ground and the gymnosperms, espe- cially Dacrycarpus and Lagarostrobus, grew in the swamps (Martin, 1993). The most easterly occurrence of the early-mid Miocene strata are found in Bores 36350 and 36351 (Fig. 6). These bores are some 10-15 km north of the Murray River and indicate that the course of the river at that time was to the north of the present river. Where evidence exists, it shows a widespread hiatus in the late Miocene (Martin, 1987). Bores 36350 and 36351 (see Fig. 6) have the C subdivision of the P. tuberculatus Zone and the T. bellus Zone, respectively, at approximately the 130 m level. The T. bellus Zone is approximate 15 million years and the Nothofagus phase, approximately 5 million years, is found at the 120 m level. There is thus a section missing, the result of erosion and/or non deposition. This pattern is very similar to that of the Lachlan River Valley (Martin 1987). It is thought that this hiatus is the result of lowered sea levels when the % of total pollen count a Bore 36311 MAJOR POLLEN GROUPS ) 20 40 ZONATION i ee eS op Bore 36306 204 | eg Caen Ae [eae pees Myrtaceae phase E 604 i a c P m 2 [se Leet ——— —- — B0- | a Cc Pp Intermediate Lf. FT a -— —— uy 5 bone stages Cc \AP | | 7 | sere c fe ke Lt | Nothofagus phase | | | | | | ———— | | . 6 | Myrtaceae Asteraceae Permian basement 140-4 Spores | Gymnosperms | Casuarinaceae Nothofagus Poaceae Fig. 9. Composite pollen diagram from two bores. See Fig. 1 for locations. C, Cyathea. A, Araucariaceae. P, Podocarpus type. m, menziesii type. f, fusca type. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 203 rivers cut down and the sediment was transported out of the valleys. Haq et al. (1987) show that the low sea level stage in the late Miocene at approximately 10 million years was the lowest for the Tertiary. The late Miocene-Pliocene sequence is similar to that seen in the other river valleys of the western slopes of the Eastern Highlands (Martin, 1987; 1991). The Nothofagus phase, if present, is an excellent marker horizon. It occurs near the base of the alluvial fill, but not directly above the basement. The lower Myrtaceae phase beneath the Nothofagus phase is not common. The expression of the Nothofagus phase is variable. Here in the Murray River Valley it is a more diffuse horizon. In the Lachlan River Valley, the well defined horizon with Nothofagus frequencies decreasing with distance downstream, suggests a migration of Nothofagus down the valley from refugia in the eastern highlands, during a brief period when the climate was wetter (Martin, 1987). In the Murray River Valley, the highest fre- quencies of Nothofagus at the base of the sequence (124 m level in Fig. 9) indicate the ini- tial migration into the valley, and the lower frequencies higher in the sequence suggest that small stands remained much longer here than in the Lachlan River Valley. This varia- tion is in accord with the geographic difference: the higher/more southerly regions were more suitable for Nothofagus. The upper and lower Myrtaceae phases are similar to those of the Lachlan River Valley (Martin, 1987). The palaeovegetation was most likely a mosaic with some wet scle- rophyll forest, i.e. a eucalypt canopy with appreciable ferns, some rainforest taxa (the gymnosperms, Cupaneae, Quintinia, Symplocos, Tasmannia and Gardenia), and few herba- ceous taxa. Dry sclerophyll forest was present also, as shown by Acacia, Banksia, Dodonaea, Hakea, Micrantheum, Monotoca, Haloragis /Gonocarpus and Haloragodendron (Stephano- colpites oblatus). These genera may include some mesophytic species, but they are more prominent in the drier habitats. In the late Pliocene-Pleistocene, the rainforest element disappears entirely. There are few ferns, gymnosperms are rare or absent, and the shrub/herbaceous element has increased considerably, particularly the Asteraceae (Tubulifloridites spp.). ACKNOWLEDGMENTS Iam indebted to the New South Wales Department of Water Resources for materials and financial assistance. I wish to thank Mr. R.M. Williams, Department of Water Resources, for information and assistance. References BALME, B.E. and Henelly, J.P.F. 1956a. — Monolete, monocolpate, and alete sporomorphs from Australian Permian sediments. Australian Journal of Botany 4: 54-67. , 1956b. — Trilete sporomorphs from Australian Permian sediments. Australian Journal of Botany 4: 240-260. BEMBRICK, C.S. 1975. — Murray Basin. In MARKHAM, N.L. and BASDEN, H. (eds.) The Mineral Deposits of New South Wales. Government Printer, New South Wales, pp.555-570. Brown, C.M. 1985. — Murray Basin, southeastern Australia: stratigraphy and resource potential - a synopsis. BMR Report 264: 1-24. , and STEPHENSON, A.E., 1986. — Murray Basin, southeastern Australia: subsurface database. BMR Report 262: 1-60. COcKAYNE, L., 1958. — The Vegetation of New Zealand . Third (Reprint) Edn. London: H.R. Englemann and J. Cramer. Cookson, I.C., and PIKE, K.M., 1954a. — The fossil occurrence of Phyllocladus and two other podocarpaceous types in Australia. Australian Journal of Botany 2: 60-68. , 1954b. — Some dicotyledonous pollen types from Cainozoic deposits in the Australian region. Australian Journal of Botany 2: 197-219. FosTER, C.B., 1979. — Permian plant microfossils of the Blair Athol Coal Measures, Baralaba Coal Measures, and basal Rewan Formation of Queensland. Geological Survey of Queensland Publication 372, Palaeontological Paper 45, 1-244. GERMERAND, J. H., Hoppinc, C. A., and MULLER, J., 1968. — Palynology of sediments from tropical areas. Review of Palaeobotany and Palynology, 6: 189-348. Proc. LINN. SOC. N.S.W., 115, 1995 204 PALYNOLOGY OF THE MURRAY VALLEY Haq, B.Q., HARDENBOL, J. and VAIL, P.R., 1987. — Chronology of fluctuating sea levels since the Triassic. Science 235: 1156-1166. Harris, W. K., 1965. — Basal Tertiary microfloras from the Princetown area, Victoria, Australia. Palaeontographica Abt. B, 75-106. Kemp, E.M., BALME, B.E., HELBy, R.J., KYLE, R.A., PLAYFORD, G., and PricE, P.L., 1977. — Carboniferous and Permian palynostratigraphy in Australia and Antarctica; a review. BMR Journal of Geology and Geophysics 2: 177-208. MacumBER, P.G., 1978. — Evolution of the Murray River during the Tertiary period: evidence from northern Victoria. Proceedings of the Royal Society of Victoria 90, 43-52. MarTIN, H.A., 1973a. — Palynology of some Tertiary and Pleistocene deposits, Lachlan River Valley, New South Wales. Australian Journal of Botany , Supplement 6, 1-57. , 1973b. — Upper Tertiary palynology in southern New South Wales. Geological Society of Australia, Special Publication 4, 35-54. , 1984a. — The use of quantitative relationships and palaeoecology in stratigraphic palynology of the Murray Basin in New South Wales. Alcheringa, 8: 253-272. , 1984b. — The stratigraphic palynology of the Murray Basin in New South Wales. - II. The Murrumbidgee area. Journal and Proceedings of the Royal Society of New South Wales 117, 35-44. , 1984c. — Stratigraphic palynology of the Murray Basin in New South Wales III. The Lachlan area. Journal and Proceedings of the Royal Society of New South Wales 117: 45-51. , 1987. — The Cainozoic history of the vegetation and climate of the Lachlan River Region, New South Wales. Proceedings of the Linnean Society of New South Wales 109, 214-257. , 1991. — Tertiary stratigraphic palynology and palaeoclimate of the inland river systems in New South Wales. In WILLIAMS, M.A.J., De DEKKER, P., and KERSHAW, A.P., (eds). The Cainozoic of Australia: a re- appraisal of the evidence. Special Publication of the Geological Society of Australia 18, 181-194 , 1993 a. — The palaeovegetation of the Murray Basin, late Eocene to mid Miocene. Australian Journal of Systematic Botany 6: 491-531. , 1993 b. — Monotoca-type (Epacridaceae) pollen in the late Tertiary of southeastern Australia. Australian Journal of Botany 41: 709-720. , and McMInN, A., 1993 . — Palynology of Sites 815 and 823; the Neogene vegetation history of coastal northeastern Australia. Proceedings of the Ocean Drilling Program, Scientific Results 133: 115-125. McMInw, A., 1985. — Palynostratigraphy of the Middle Permian coal sequences of the Sydney Basin. Australian Journal of Earth Sciences 32, 301-309. Morcan, R., 1977. — Microfloras from the Oaklands area, southeastern Murray Basin, New South Wales. Report of the Geological Survey of New South Wales GS1975/130. (unpubl.). POCKNALL, D.T., and Crospik, Y.M., 1982. — Taxonomic revision of some Tertiary tricolpate and tricolporate pollen grains from New Zealand. New Zealand Journal of Botany 20, 7-15. Price, P.L., 1983. — Permian palynostratigraphy for Queensland. In Permian Geology of Queensland. Geological Society of Australia, Queensland Division pp.186-211. STEPHENSON, A.E., and BROwN, C.M., 1989. — The ancient Murray River System. BMR Journal of Australian Geology and Geophysics 11: 387-395. Stover, L.E., and Partridge, A. D., 1973.— Tertiary and Late Cretaceous spores and pollen from the Gippsland Basin southeastern Australia. Proceedings of the Royal Society of Victoria 85: 237-286. STOVER, L.E., and PARTRIDGE, A.D., 1982.— Eocene spore-pollen from the Werillup Formation, Western Australia. Palynology 6: 69-95. TRUSWELL, E.M., SLUITER, I.R. and Harris, W.K., 1985. — Palynology of the Oligocene-Miocene sequence in the Oakvale-1 corehole, western Murray Basin, South Australia. BMR Journal of Geology and Geophysics 9: 267- 295. WILLIAMS, G.E., and Goong, A.D.T., 1978. — Possible western outlet for an ancient Murray River in South Australia. Search 9: 442-447. WILLIAMS, R.M., 1989. — Groundwater Resources of the unconsolidated sediments associated with the Murray River between Albury and Corowa, N.S.W. Water Resources Technical Services Report 89: 1-62. Yoo, E.K., 1982. — Geology and coal resources of the northern part of the Oaklands Basin. Records of the Geological Survey of New South Wales 18: 114-116. Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 205 APPENDIX | The Palynological Zonation Bores arranged W-E, then N-S. For location , see Fig. 1. Bore and Depth Palynological Zonation Age 36558 168-190 m Bsubdivision, P. tuberculatus Zone Mid-late Oligocene 248-266 m A subdivision, P. tuberculatus Zone Early Oligocene 36582 119-166 m B subdivision, P. tuberculatus Zone Mid-late Oligocene 170-171 m A subdivision, P. tuberculatus Zone Early Oligocene 36587 143m B/C subdivision, P. tuberculatus Zone Late Oligocene-early Miocene 217m Bsubdivision, P. tuberculatus Zone Mid-late Oligocene 235m Asubdivision, P. tuberculatus Zone Early Oligocene 280 m Middle N. asperus Zone Late Eocene 36588 157-173 m B subdivision, P. tuberculatus Zone Mid-late Oligocene 36585 132-136.5 m Upper N. flemingii acme, P.tuberculatus Zone Late Oligocene-early Miocene 190-204 m Bsubdivision, P.twberculatus Zone Mid-late Oligocene 208-213 m Lower N. flemingit acme, P.tuberculatus Zone Early Oligocene 222-226.5m ? Stage 4 Early Permian 36586 150-196 m B subdivision, P. tuberculatus Zone Mid-late Oligocene 199-201 m Asubdivision, P. tuberculatus Zone Early Oligocene 209-254 m Middle N. asperus Zone Late Eocene 36584 139.5m Stage 4a Early Permian Oaklands 15 114.3m Bsubdivision, P. tuberculatus Zone Mid-late Oligocene 166.1 m Upper stage 5a Mid Permian 36635 100-101 m B subdivision, P. tuberculatus Zone Mid-late Oligocene 100-109 m A subdivision, P. tuberculatus Zone Early Oligocene 118-119m Mid N. asperus Zone Late Eocene Oaklands 18 107.1-116.15 m B subdivision, P. tuberculatus Zone Mid-late Oligocene Oaklands 5 218.2m Upper stage 5a Mid Permian 317.0m Stage 3b Early Permian Oaklands 12 135.0m Stage 3a Early Permian Oaklands 4 212.4m Upper stage 5a Mid Permian 246.9 m Stage 4a Early Permian 256.0 m Stage 3b Early Permian Oaklands 2 272-278.9 m Upper stage 5a Mid Permian Oaklands 16 257.5-268.2 m Upper stage 5b Mid Permian Oaklands 1 224.0m Lower stage 5b Early Permian Proc. LINN. SOC. N.S.W., 115, 1995 206 Oaklands 14 198.1-235.6 m 36638 173-174m Oaklands 11 98.4m 103.3-117.6m Oaklands 17 176.5-179.5 m Bore 36356 81.4-82.9 m Bore 36350 128.6-130.1 m 137.8-139.3m 145.3m Oaklands 13 96.6-98.4 m Oaklands 7 90.8 m 110.5m 36399 51-55m Oaklands 6 107.6m 118.3 Oaklands 9 100.6m Bore 36390 103.6m Oaklands 10 32.9m 121.9m Coorabin 2 63m 36351 101.2-108.8 m 127.1-130.1m 139.3-146.9 m 152.4m 36392 119.5-121.0m 125.6-127.1 m 130.1-131.7m 36354 46.3-50.9 m 92.0-103.6m 137.8-139.3 m 36355 98.1-110m 3925] 32-34 m 96-96.3 m PALYNOLOGY OF THE MURRAY VALLEY Lower stage 5c Stage 3a ? Upper stage 5a Lower stage 5b Late P. tuberculatus/T. bellus Zone Stage 3a ? Mid N. asperus Zone Mid N. asperus Zone Upper stage 4a Lower stage 5a Lower stage 5a 2 Upper Myrtaceae phase Nothofagus phase Stage 3b Myrtaceae phase T. bellus Zone Stage 3a 2 Stage 3a Upper Myrtaceae phase ? Nothofagus phase Proc. LINN. SOC. N.S.W., 115, 1995 Mid Permian Early Permian Mid Permian Mid Permian Mid Permian Pliocene ? Mid Miocene Early Permian Early Permian Late Eocene Late Eocene Mid Permian Mid Permian Mid Ppermian Mid Permian Early Permian Early Permian Pliocene Late Miocene-early Pliocene Early Permian Late Miocene-Pliocene Mid Miocene Early Permian Early Permian Early Permian Early Permian Early Permian Pliocene ?Late Miocene-early Pliocene Early Permian Early Permian ? Pleistocene Early Permian Coorabin 1 86.6-89.9 m 36352 99.7-101.2 m 111.4m 36394 89-92 m 101.2-102.7 m 36311 37.4-39.2 m 110.6-124.3 m 136.5-138.1 m 36306 68.0-69.5 m 96.9-98.4 m 136.5-138.1 m 36295 120.4-121.9m 126.5-131.1m 36303 66.4-69.5 m 36281 42.0-76.0 m 83.5-88.5 m 36376 45-46 m 101.5m 36416 27.28m 25281 52.461 m 25357 53.0-53.6 m 30763 Culcairn 79.5-81 m 36292 Holbrook 45-47.5m H.A. MARTIN Stage 3a Stage 3a 2 ? Nothofagus phase Lower Myrtaceae phase Upper Myrtaceae phase Nothofagus phase ? Upper Myrtaceae phase Nothofagus phase Lower Myrtaceae phase Nothofagus phase 2 Upper Myrtaceae phase Nothofagus phase Upper Myrtaceae phase Nothofagus phase Asteraceae /Poaceae Mytaceae phase Myrtaceae phase Myrtaceae phase 207 Early Permian Early Permian Early Permian ? Late Miocene-early Pliocene Late Miocene Pliocene Late Miocene Pliocene Early Permian Pliocene Late Miocene-early Pliocene Late Miocene Late Miocene-early Pliocene Early Permian Pliocene Pliocene Late Miocene-early Pliocene Pliocene Late Miocene-early Pliocene Pleistocene Late Miocene—Pliocene Late Miocene-Pliocene Late Miocene-Pliocene ?Pliocene Proc. LINN. SOC. N.S.W., 115, 1995 208 PALYNOLOGY OF THE MURRAY VALLEY APPENDIX 2 Permian spores and Pollen in selected samples. References where descriptions may be found are as follows: 1. Balme and Hennelly (1956b). 2. Foster (1979). 3. Price (1985). 4. Truswell et al (1977). 5. Balme and Hennelly (1956a). Bore Coo Oak Oak Oak Oak Oak Oak Oak Oak 1 4 4 4 7 11 14 16 Ob 86.6 212.4 246.9 256.0 110.5 103.3 198.1 257.5 283.7 to to to to Depth (m) 89.9 117.6 235.6 268.2 Acanthotriletes (Microbaculisporites) villosus 1,3 + A. tereteangulatus 2 + + Alisporites splendens 2 + Alisporites sp + + + + Apiculatisporis cornutus 2 + Bipartitispores cf. Verrucosisporites trisectus 2 + Bascanispontes undosus 1 + + Barakarites rotatus 2 + + + + + B. pluriaenus 2 + B. scissa2 + + + Brevitriletes levis 2 + Cannanoropollis cf. C. janakii 2 + Calamosporis diversiformis 1 + Circulisporis parvus 2 + Cyadopites follicularis 2 + Dictyotriletes aules 2 + Didecitriletes ericanus 2 + te + + Dulhuntyispora dulhuntyi3 + D. parvithola 2, 3 + + D. stellata 3 + Granulatisporites micronodosus 1 + cr + G. quadruplex 2 + + Horrditriletes curvibaculosus 2 + Indotriradites splendens 2 + + Interradispora versus 3 + Lewotriletes directus 1 + + + Marsupipollenites triradiatus 2 + + + + + + Maculatisporites amplus 2 aP Mehlisphaeridium cf. M. fibratum 2 + Micrhystridium sp 2 + Osmundacidites senectus 2 + Plicatipollenites gondwanensis 2 + + + + + + + Potonieisporites sp 4 Praecolpatites sinuosus 2 + + + + + + Protohaploxypinus amplus 2 + Plimpfidus 2 + + P. varius + Protohaploxypinus spp + + + + + Punctatisporites gretensis | + + + + Pseudoreticulatispora (Verrucosisporites) pseudoreticulata 2,3 + + + + + + Scheuringipollenites ovatus 2 + + Striatoalneites multistriatus 2 a + + Secarisporites bullatus 2 + Striatopodocarpidites spp. + + + Striapollenites saccatus 2 + Verrucososporites hamatus 5 + Vitreisporites signatus 2 + + + i A + + + + + Proc. LINN. SOC. N.S.W., 115, 1995 H.A. MARTIN 209 APPENDIX 3 Late Eocene-Oligocene spores and pollen in selected bores. References where taxa may be found are as follows: 1, Stover and Partridge (1973). 2, Germerrad et al. (1968). 3, Martin (1973a). 4, Harris (1965). 5, Cookson and Pike (1954). 6, Stover and Partridge (1982). 7, Mildenhall and Pocknall (1989). Bore 366586 Bore 36558 168- 186 248- 261- 173, 190 252 266 Depth (m) 150- 195- 199- 209- 217- 253- 151 196 201 211 219 254 Spores Baculatisporites disconformis | + 0.6 Crassotrilites vanraadshooveni 2 + Cyathea paleospora 3 1.3 1.8 0.6 13) 1.2 6.0 0.6 1E3, Cyatheaceidites annulatus 1 0.7 + + Cyathadites australis 1 0.6 Dictyophyllidites concavus 4 + 0.6 Gleichenia circinidites 3 13 0.6 + + 0.6 Ischyosporites gremius | 0.6 0.7 Klukisporites lachlanensis 3 0.6 0.6 Laevigatosporites ovatus 3 0.6 0.6 12. 1.2 Latrobosporites crassus 4 + 0.6 Lycopodiumsporites sp. + + Matonisporites ornamentalis | + Peromonoletes densus 1 7 P. vellosus 1 0.6 Polypodiidites sp. 3 + 0.6 0.6 Rugulatisporites mallatus | + + 1.3 R. trophus 1 + 0.6 Stereisponites sp. (Sphagnum) 0.6 Todispontes sp. 3 0.6 Verrucosisporites cristatus 1 0.6 + 0.6 + V. kopukuensis | 0.6 + 0.6 + 0.7 + 0.6 + Gymnosperms Araucaniacites australis 3 1.3 7 + 0.6 0.6 Cupressaceae 3 0.6 Dacrycarpites australiensis 3 0.6 4.5 2.3 2.4 6.4 3.3 12 3.0 3.1 1.3 Ephedripites’ notensis 3 + Lygistepollenites florinii 1 7.6 1.3 2.3 8.6 3.8 4.7 17, 2.4 2.5 5.1 Microcachryidites antarcticus 3 1.8 2.0 0.6 0.6 0.6 Parvisaccites catastus | 0.7 Phyllocladidites mawsonii 1 76 17.9 28.8 21.1 15.3 | 0.6 FOS: Diy qld P. palaeogenicus 5 0.6 12 0.6 0.6 Podocarpidites spp. 4.4 8.3 4.9 8.0 5.1 4.7 4.6 7.8 2.5 1.9 Trisaccites micropteris 5 0.6 0.6 0.6 1.3 6.6 + Angiosperms Banksieaeidites elongatus 1 + Cunoniaceae, tricolplate form + 0.6 Cupaniecdites orthoteichus 1 0.7 Cyperaceae 3 0.6 Ericipites crassiexinous | 1.3 Gephyrapollenites calathus 1 0.6 Granodiporites nebulosus 1 + Haloragacidites harrisii 1 9.5 5.8 6.8 One 7 13%5s 13) 25:6) 11220 6.8 5.1 Tlexpollenites anguloclavatus 1 1.9 1.9 1.2 0.7 | 0.6 + + 0.6 Lihacidites spp. 1.3 0.6 0.6 0.7 | 0.6 Malvacipollis subtilis/diversus | 0.6 a2. 2.4 1.9 1.3 0.6 1e2 0.6 Milfordia sp. 0.6 Myrtacecdites eucalyptoides 3 0.6 0.6 0.6 M. parvus 3 5.7 2.6 0.6 0.6 3.6 1.2 Myrtaceae unidentified 2.6 2.3 IES 0.7 | 4.1 0.6 0.6 Nothofagidites asperus 1 1.3 2.9 1.2 N. brachyspinulosus 1 1.6 0.6 1.8 1.9 1.3 IE? 0.6 N. emaradus | 492.4 340 25.3 11.0 205 22.7 |20.8 29.3 19.9 30.4 N. falcatus 1 0.6 0.6 1.3 0.6 12 eS) N. flemingii 1 2.5 5.8 3.7 1.8 1.3 3.3 1.2 3.6 3u7 1.9 N. goniatus 1 0.6 + 1.2 Proc. LINN. SOC. N.S.W., 115, 1995 210 Depth (m) N. incrassatus 1 N. vansteenisi2 Perfotricolporites digitatus 2 Periporopollenites demarcatus | P. vesicus 1 Polyorificites oblatus 3 Proteacidites annularis 1 P. cf. beddoesii | P. cf. latrobensis 1 P. ivanhoensis 3 P. obscurus 1 P. cf. pseudomoides 1 P. recavus | P. rectomarginis | P. reticulatus 1 P. stipplatus 1 P. subscabratus 3 P. tuberculatus 1 Procteacidites spp Quinitinapollis psilatispora7 Sparganiaceaepollenites barungensis 3 S. sphaericus 3 Tetracolporites palynius 6 Tnicolporites cf. T. angurium | T. leuros 1 T. adelaidensis 6 T. substriatus/paenestriatus 3, 1 Tnorites magnificus 1 Tricolpate-tricolporates Summary Spores Gymnosperms Casuarinaceae Myrtaceae Nothofagus Ratios P. mawsoniui/gymnosperms N. flemingii/ Total Nothofagus Inferred Age Proc. LINN. SOC. N.S.W., 115, 1995 150- 151 0.6 7.6 3.8 1.3 12.6 51.7 0.05 195- 196 5.1 0.6 22.3 5.8 5.8 44.5 0.34 0.13 PALYNOLOGY OF THE MURRAY VALLEY Bore 366586 199- 209- 217- 253- ADI NI As aay 2.4 1.3 0.7 13.6 6.7 5.8 6.7 0.6 2.6 1.3 0.6 2.3 0.6 0.7 0.6 1.8 0.6 0.7 1.2 + 0.6 1.2 0.6 1.3 0.6 0.6 + 1.3 0.6 12 0.6 1.3 0.6 0.6 + 0.6 1.8 0.6 351 3.7 5.8 2.7 3.6 0.6 1.8 9.4 28.0 51.4 36.4 32.0 6.8 6.1 13.5 11.3 2.9 1.2 1.3 0.7 43.2 24.3 30.8 30.0 0.64 0.56 0.59 0.48 0.08 0.07 0.42 O11 1.2 0.6 0.6 2.9 0.6 5.8 4.4 11.0 25.6 4.1 35.4 0.14 0.03 Bore 36558 186- 248- 261- 190 252 266 5.4 10.5 25.3 + 1.2 1.3 + 0.6 12 0.6 0.6 0.6 - + 0.6 0.6 1.3 + 0.6 + 0.6 0.6 + 0.6 0.6 0.6 0.6 + 0.6 + 0.6 + + U2 5.0 2.5 10.2 1.8 4.4 123 43.4 20.9 12.0 6.8 5.1 4.8 1.2 0.6 40.7 36.5 59.5 + 0.77 0.54 0.09 0.01 0.03 Early Oligocene Eocene Oligocene H.A. MARTIN 211 APPENDIX 4 Upper Tertiary and Pleistocene spores and pollen in selected samples. References where descriptions may be found are as follows: 1, Stover and Partridge (1973). 2, Martin (1973a). 3, Cookson and Pike (1954a). 4, Truswell et al. (1985). 5, Martin (1973b). 6, Harris (1965).’7, Martin and McMinn (1993). 8, Martin (1993b). 9, Pocknall and Crosbie (1982). 10, Cookson and Pike (1954b). Bore Oak 10 36351 36394 36376 36416 Depth (m) 32.9 121.9 }101.2 127.1 | 89.0 101.2 45.0 101.5 27.0 Spores Baculatisporites disconformis 1 4.0 Cyathea paleospora 2 14.5 16.0 | 13.4 20.0 | 10.8 13.3 25.0 24.4 0.8 Cyatheacidites annulatus | 0.8 Eg 2.0 0.8 Cyathidites sublilis 1 0.8 Deltoidospora inconspicua 2 8.5 0.8 3.4 0.8 2.0 1.5 Dictyophylhdites concavus 1.0 Gleicheniidites circinidites 2 0.8 + 0.8 2.0 Klukisporites lachlanensis 2 1.6 0.8 1.0 2.8 1.0 2.3 Matonisporites ornamentalis 1 1.7 1.6 0.8 + 0.9 1.5 Laevigatosporites ovatus 2 6.0 2.4 2.5 1.7 1.0 1.0 1.5 1.7 Lycopodium sp. 0.8 Osmundaceae sp 1,2 0.8 7. Polypodiidites sp. 0.8 0.8 9.5 1.5 Rugulatisporites cf R. trophus 1 0.8 cf. Pteris 7 1.0 Verrucosisporites sp. 0.8 + 0.9 Gymnosperms Araucariacites australis 2 8.8 a 0.8 9.8 0.8 Cupressaceae 2 2.6 0.8 1.7 0.8 Dacrycarpites australiensis 2 2.6 0.8 0.8 1.0 0.9 3.0 3.0 Lygistepollenites florini 1 0.8 0.8 1.0 + Microcachryidites antarcticus 1 1.6 1.5 Phyllocladidites palaeogenicus 3 0.8 5.9 6.7 4.9 0.9 1.0 5.8 Podocarpidites spp. 0.8 8.0 5.0 4.2 5.9 3.8 4.0 13.0 Angiosperms Acacia myniosporites 2 0.8 + Banksieaeidites elongatus 2 0.8 1.0 0.8 Canthiumidites (Triporopollenites) bellus 1,9 0.8 Chenopodipollis chenopodiaceoides4 0.8 1.0 0.8 Cupanieidites orthoteichus 1 0.8 Cyperaceae 2 2.6 0.8 2.5 1.0 2.0 1.5 Dodonaea sphaerica 2 0.8 0.9 Ericipites sp. 0.8 + 2.0 0.9 1.0 0.8 Graminidites media 2 1.7 0.8 2.9 0.9 1.5 5.8 Haloragacidites haloragoides 2 0.8 0.8 1.7 1.0 0.9 1.0 + 3.4 H. harrisii (Casuarinaceae) 2 7.7 1.7 5.0 4.9 3.8 14.0 8.4 16.0 Milfordia hypolaenioides 2 0.9 Monotoca8& 1.0 Micrantheum spinyspora 2 0.8 3.4 Myrtaceidites eucalyptoides 2 1.7 5.6 2.5 0.8 1.0 8.5 11.0 1.7 M. mesonesus 2 1.7 6.7 1.7 4.9 2.8 2.0 M. parvus 2 8.5 7.2 9.2 657 || 1227 5.7 2.5 Myrtaceae unidentified 20.5 10.4 | 21.8 10.0 8.8 15.2 9.0 2.3 8.4 Myriophyllum 10 0.8 Nothofagidites asperus 1 12.8 2.5 3.3 6.9 4.6 N. brachyspinulosus 1 5.8 N. emarcidus 1 6.7 Polyporina granulata 2 0.8 0.8 0.8 P. symphyonemoides 1 + + Proteacidites cf. Hakea 1.0 0.8 P. wanhoensis 2 3.3 P. subscaboratus 2 1.0 2.3 0.8 Proteacidites sp. 1.7 0.8 - Quintinia psilatispora 2 0.9 1.0 1.5 Stephanocolpites oblatus 2 2.5 Proc. LINN. SOC. N.S.W., 115, 1995 212 PALYNOLOGY OF THE MURRAY VALLEY Bore Oak 10 36351 36394 36376 36416 Depth (m) 32.9 121.9 |101.2 127.1 | 89.0 101.2 45.0 101.5 27.0 Symplocotpollenites austellus 1 0.8 Tasmannia (Drimys) tetradites 2 0.8 cf. Tricolpites reteculatus 9, 10 2.5 Tnicolponites geranioides 5 T. mataurensis9 8.4 T. pelargonioides 5 Tubulifloridites spp. 2 0.8 21.8 Unidentified pollen types 5.9 17.6 Summary Spores 32.5 2.5 Gymnosperms 6.0 Casuarinaceae 7.7 16.0 Myrtaceae 32.5 12.6 Nothofagus Asteraceae 0.8 21.8 Poaceae 2 5.8 Cyperaceae 2.6 Plio- Pearly i Late Pearly Pleisto- cene Pliocene | Miocene : Miocene Pliocene | Pliocene cene Pliocene Pliocene Proc. LINN. SOC. N.S.W., 115, 1995 W.H. Harvey in New South Wales Letters by the phycologist W.H. Harvey, written in New South Wales in 1855 S.C. DUCKER Ducker, S.C. W.H. Harvey in New South Wales. letters by the phycologist W.H. Harvey, written in New South Wales in 1855. Proc. Linn. Soc. N.S.W. 115: 213-223 (1995). The Irish Professor of Botany visits New South Wales in 1855 and writes a letter to his sister, Hannah Todhunter, and to the Tasmanian botanist, Robert Campbell Gunn. He describes Sydney, the people, and the Botanic Gardens. S.C. Ducker, Botany School, University of Melbourne, Parkville, Victoria, Australia 3052; manuscript received 17 September 1993, accepted for publication 17 December 1993. KEY WORDS: W.H. Harvey, New South Wales, Sydney 1850, botany, phycology. INTRODUCTION The life of W.H. Harvey and his travels to the antipodes have been narrated in dif- ferent publications (Fisher 1869; Praeger 1913; Webb 1966; Ducker 1988). So much is known about the man and his time through his letters, because Harvey was a most prolific letter writer. All his charm, his power of observation, his almost universal interests and most of all his love of life and nature are mirrored in his correspondence. In 1988 I pub- lished all the letters known to me about his travels to Australia and the Pacific. During a visit to Dublin in 1989 a new and recently-acquired bunch of Harvey letters was brought to my notice. They were letters addressed to members of his family in Dublin and were at the time forwarded to his sister-in-law in New York. Her initials are written on several of the let- ters. The provenance of the letters must have been clear to some recent descendants of this New York sister-in-law and they returned the letters to the Trinity College Library, which kindly gave me permission to publish them. As little is known of Harvey’s impressions of Sydney in 1855, two letters written by the traveller in New South Wales are reproduced here. Annotations are included only where necessary, as the majority of references to people and plants are covered in the ear- lier letters. The first letter is written to Harvey’s sister Hannah Harvey Todhunter (TCD 3640/24) with whom he made his home for many years. The majority of letters sent by Harvey while on his journey to the southern hemisphere were addressed to Hannah and were in the form of a diary. The recipient of the second letter (from the Mitchell Library, Sydney) was Ronald Campbell Gunn (1808-1881), the resident botanist at Launceston, Tasmania. Gunn had inspired Harvey to visit Tasmania, and for this Harvey was grateful. In 1844, Gunn sent a packet of exciting seaweeds, including the beautiful and rare red algae Claudia elegans, to Professor William Hooker at Kew. The parcel was passed by Hooker to Harvey, who thus became aware of the rich Australian algal flora. While visiting Georgetown and Launceston during his Australian travels, Harvey became friendly with Gunn. Both letters were written prior to Harvey’s departure for the Pacific Islands. Harvey was consulted by Augustus Gregory about Ferdinand Mueller’s suitability as an accompa- nying botanist for the scientific expedition to be mounted by the Imperial Government, and this important information is described in these letters. If anyone ever warranted trust and praise, it was Mueller. Little did anybody know, however, what hardships awaited members of the Gregory expedition in their traverse of the northern parts of Australia. Proc. LINN. SOC. N.S.W., 115, 1995 214 W.H. HARVEY IN NEW SOUTH WALES The rich harvest of plants collected by Mueller, including many new taxa, was early proof of his qualities. THE LETTERS NewCastle — New South Wales — May 15” 1855 — My dear Hannah, As I have nothing particular to do this evening, I shall begin a journal letter, in con- tinuation of my proceedings since I left Melbourne, from which I wrote last — I parted from friends Robinson & Unthank at the Queen’s wharf, Melbourne on the 1“& wentina steam tender on board the “Telegraph’’ Steam ship lying in Hobson’s Bay, which set sail about 2 hours afterwards & in 3 hours more we reached the Heads of Port Phillip of which I took a sunset farewell, on a fine evening with a smooth sea. — The Telegraph is the best ship on the line & we went at a spanking rate. — I soon ‘turned in’ as I never sit up long at sea & when I rose next morning was told, that we had passed Wilson’s Promontory about 2 in the morning. — The Coast was in sight most of the day, which was very fine & warm. — About 4 P.M. we approached the dividing point between Victoria & New South Wales & passed the Cape shortly after sunset. — The shore is sandy with wooded mountains behind, at no great distance. — There is a bright light on the end of the Cape, a narrow sand spit running into the sea. — The night was fair — the moon at full & we went on rapidly. — About 8 passed the ‘Wonga Wonga’' Steamer & let off rockets. — Next morning 3” opened in heavy rain & it rained heavily the greater part of the day, being what is called a ‘black North Easter’. — We steamed close along the shore, but saw little save sandstone cliffs alternating with sandy beaches, black cloudy sky & bare open country. —I dare say that in sunshine it would have looked more inviting & possibly had I been ashore, I should have found it good botanizing ground. — In the afternoon, we passed the opening of Botany Bay — the rain ceased, the clouds continued — but the pas- sengers came on deck to watch for Port Jackson, which we reached in about an hour after- wards. — The coast here is very grand & bold, reminding you of Kilkee’ cliffs 2 to 300 ft high, quite perpendicular, with surf lashing their bases — a lighthouse is built on the South Head, I suppose at nearly 300 ft above sea. — the North Head is equally bold — The entrance is wider than I had anticipated & I wonder Cook’ should not have been tempted further in — for once within the Heads, the prospect is one of extreme beauty. — The Port itself is of vast extent, with deep water almost throughout & in most places close up to the shore, so that large vessels lie alongside natural quays — at all times of tide. — There are numerous arms & inlets & bays, stretching from all sides & their shores being moderately high & steep & well wooded, with occasional projecting rocks are very picturesque. — The views are varied very pleasingly & there are many & picturesque villas on the shores — small & large — silver strands. — green lawns, shrubberies &c — One of the most grotesque (rather than picturesque) trees is the Norfolk Island Pine,’ of which there are many large specimens all round Port Jackson & in the city of Sydney. — This tree is so perfectly formal & regular in its growth that it looks more like an artificial than a natural production & con- stantly reminds me of the little conical & (comical) green trees which are attached to the childs toy, called a sheep cot. — They have exactly this character when seen from a dis- tance against the sky. — When you approach them, I admit, you lose the sense of formality in the beauty of the foliage. — It was quite dark when we reached the wharf & dark before I landed. — & the rain again set in. — A slushy & steep walk brought me to Petty’s Hotel,’ on the top of Church Hill where I took up my quarters very comfortably. — 4” of May. — Rain continues this morning. — Nevertheless after breakfast, I walked to the Botanic Gardens & called on M' Moore’ the Curator. — The Garden is within '/, Proc. LINN. SOC. N.S.W., 115, 1995 S.C. DUCKER 215 mile of the town & within the limits of the ‘City’. — It is part of the Gov' Demesne and con- tains about 30 acres of beautifully diversified ground, fronting the bay. — The soil is very bad & shallow & difficult to cultivate, as it burns up quickly in dry weather, whereof there is much in this latitude. Last year many valuable shrubs & trees perished. — It was an unusu- ally hot & dry summer — Just now all is green & juicy — as they have had very heavy rains for the last 6 weeks — There are not many flowers or small things in the garden, but its shrubberies are very beautiful & there are many Conifers cultivated. — Some of the larger Norfolk pines are over 100 feet high — also some fine Cunninghamia species & several species of Dammara’ (broad leaved tropical pines) introduced by M" Moore. Many sub- tropical shrubs & some tropical ones do well but not so well as in other gardens near Sydney. — Mangoes bear fruit but are generally stolen before they are quite ripe. — The commoner Bananas ripen, but not the better kinds. — The grounds are chiefly laid out with broad walks & grass plots for promenade — but there is a small portion set aside for systematic arrangement both on Linnaen & Nat. Systems. — Moore gives an annual course of Elementary Lectures which he says are not well attended, nor much appreciat- ed. — I should not expect them to be in a Colony where education is at such a low state as itis here — Among the prominent features of the garden, I must not omit 2 noble Ferns — Platycerium Alcicorné & P. Grande the former common in the neighbourhood of Sydney — the latter an importation from the North. — You have no idea of their grandeur from the small specimens seen with you in cultivation. — They either stick on trees or rocks in huge masses. — In the garden is a monument to poor Cunningham’ the Botanist — & former curator, but just now in the midst of a pond of muddy water — After a couple of hours spent at the garden I returned to town to get my luggage (20 packages!!) landed & stored & saw them safely placed in a store at the Gardens, where I pay no storage. — Then through muddy streets & occasional showers hither & thither to make acquaintance with the town. — Sydney is built on a long tongue of land between two arms of the harbour & stands on hills & in vallies — Some of the streets are therefore steep. — Some are crooked & some are very broad or very narrow — The common building stone is a light coloured sand stone — easily worked & therefore the houses are usually fronted with cutstone. — Some are plaistered & built of brick — every one follows his own fancy in architecture, so that uniformity is the exception not the rule — big & little houses being constantly in juxta position — On Church Hill, where the hotel is, are three Churches of the three creeds of England, Scotland & Rome — The English Church is the oldest in the Colony & a very funny looking building — I took it at first for a Monkery belonging to the RC. Church adjoining — There is a round tower at one end — some off buildings & a little semi dome at the other (for the chancel) — A new Church of Gothic architecture & a handsome building is being built at a little distance, to replace the present one. — Beautiful views of the harbour with its various coves & arms may be had from various heights in the town. — Gov' House, an extensive partly castellated, partly Palatial building stands in a well kept, but not extensive ornamental grounds facing the harbour & commanding a fine view of the Bot. Garden — & neighbouring points — with some of the shipping &c — Shops most- ly good & English looking & many of the private houses handsome — Norfolk Island Pines are frequently planted before the doors in many parts of the town & have a quaint effect. — The streets are lighted by gas, but dimly the lamps few & far between. — On the whole I do not agree with the Lady I met at Eagle Hawk Neck, that ‘Dublin is nothing to Sydney’ — In the evening I called on Dr Bennett,” Author of Wanderings in New South Wales & a surgeon in large practice — having also a noble library of general literature & particu- larly of Illustrated works — a great many are Natural History to which he is attached as an Amateur & has contributed to various periodicals & authors much valuable information on the animals & plants seen in his travels. He it was who procured the first living Nautilus seen in England — & first kept Ornithorhynci as pets &c. &c. His passion is books on which he spares no expense. Consequently every hole & corner of the house where a book can be stuck up or piled, there it is. The shelves are two or three deep of vols — Big folios Proc. LINN. SOC. N.S.W., 115, 1995 216 W.H. HARVEY IN NEW SOUTH WALES lie on tables & chairs — & there is scarcely more room than to turn about in the parlours. I was received very kindly & spent a pleasant evening, looking at pictures & talking of fifty different things. 5" May — It rained heavily almost the whole day which I spent chiefly at the botanic gardens arranging my luggage for the next campaign. When it rains here it pours & the streets are scarcely passible in many places. 6" — Was Sunday — showery in the morning — but a small glimpse of sky behind the clouds. Later in the day it rained heavily & sometimes furiously, & I was caught in the afternoon & glad to take refuge in an omnibus — into which shortly after got in 3 wet chi- namen. I went to Church in the morning to the queer old building close at hand & heard avery low churchman — in the afternoon I walked some two miles to what I was told was a Pussey''-cat Church, & was punished for my gadding by finding only a bald service with half a dozen persons in Church, a christening & a Churching — but no sermon — & a tremen- dous shower aforesaid. I ascertained afterwards that the regular evening service is at 7°C. to which I did not return. 7*— At last a fine day — but cloudy — The air close & warm. After breakfast I called & presented letters to the Gov’ Sir W™ Denison’ — who received me very kindly — & after a few introductory words says “Well now, what can J do for you?’ whereupon I up & told him my plans that I wished to go to some of the Northern parts, where I looked for letters to the Port Officers to supply me with boats &c. But said ‘I have just heard that there is a Missionary ship going to visit some of the Islands in the Pacific & I think I shall try for a pas- sage in her for the cruize’. ‘Why not go in H.M.S. Herald’, saith he, ‘she is going the same route & will be surveying & knocking about among the islands &c’. So he gave me a letter to the Captain Denham & ordered me aman of war’s boat & crew to take me on board the Herald, now lying at Watson’s bay near the harbour’s mouth. After seeing the Gov’ I went in search of Mr Gregory the head of an exploring Expedition soon to be sent to the North and I counted to talk to him about Dr. Mueller, who wishes to go as botanist I was directed to ‘6 Hunter St’ So seeing No 5 on the door, I enquired at the next house if it was No 6? — ‘No — this is No 7 — Then pray where is No 6? — Oh — no 6 is moved to the other end of the street (quarter of a mile off), where, after a search I found it — but did not find Mr G. —I then returned to Hotel for Lunch & thence went to the Wharf, to go on board H.MLS. Juno, in whose cutter I was to be sent on board the Herald. Politely received by the gun room officers, & after a delay — we set out in the cutter for the Herald. By this time there was a strong wind blowing & clean against us & our distance was 7 or 8 miles — so we had many tacks to make before we reached the vessel & then it was nearly dusk. When we got there we ascertained that Captain Denham was living with some friends on shore, at one of the villas on the harbour — & so we again set [ed. note: Fig. 1 shows the manuscript from this point] off (the Middy & I) in the boat to seek him. We landed at two or three wrong villas but at last found the right one. In our scramble through the bush I first made acquain- tance Epacris Grandiflora © Crowea saligna’ (well known green house plants) in a native state. Both are very lovely as they spring up among the mossy rocks. A curious green-flow- ered Orchis (Pterostylis) was common in the grass. At last we found Capt Denham to whom I gave the Govn letter — & had a conversation — In which I ascertained that the Herald’s cruize was not likely to be such as would suit me, as she would probably be out 8 or 9 months, with small chance of communicating during the time, & that much of her time would probably be spent out of sight of land — sounding on deep banks. So there was an end of this scheme. It was not quite dark & on our way back through the woods to the boat we were attracted by many large & very luminous fungi, which shed a broad glow of light among the grass & rotten leaves. I gathered some & found them to be Agarics (Mushrooms) 3 or 4 inches in diameter, with a flattish, wavy, pale slate colour or whitish cap, very numerous thickly set & decurrent gills, & a solid, curved & frequently eccentric stalk. The light was very white, like ghost-like moonlight & was so strong that I could see the time by my watch by it. I brought them home & they retained their lustre till decomposure Proc. LINN. SOC. N.S.W., 115, 1995 S.C. DUCKER 217 set in. The strongest light appeared to be when the fungus was in ts best condition & fully grown, as neither the young nor the old was so luminous. The substance was soft & watery. I have since found the same Agaric abundant in other places. To Dr. Bennett’s in the evening where I slept on the sofa, the night being wet. 8". Called on Mr Gregory” this morning & had a satisfactory talk about Dr Mueller, who will I expect get the app‘ of Botanist to the Expedition. I then went to the Botanic Gardens & afterwards took a walk in the outer demesne to the end of a projecting point, which is called Mrs Macquarie’s chair. An inscription cut in the rock declares that the road round the chair was called ‘Mrs Macquarie’s road’ by order of Macquarie esq Governor — & that she planned it &c. By the way, this is only one of a great many inscriptions attribut- ing deeds to the same L. Macquarie Esq who appears to have been very found of seeing himself ‘stuck up’. I gathered a few small plants in flower, including some of the smaller Orchids (Pterostylis & Acianthus) & found a solitary specimen of the remarkable fungus called Aseroe'® — like a Cuttle fish — (of which I have a diagram at home). It is a fine crim- son colour. I have dried it & it makes a tolerable specimen — though much of its beauty is lost. also gathered a fresh batch of luminous fungi with which I amused the folk staying at the Hotel, & who though passing the best part of their lives in the bush, had never seen the like before. The Hotel was chiefly frequented by Squatters as the cattle & sheep farmers are termed. They are (so far as I have seen) a frank, open-tongued but not ungentlemanly set of men — Many of them pleasant & well-informed — under an uncouth exterior often possessing many agreeable qualities. They hold their lands or ‘runs’ by a slender tenure from the Gov‘ & many of them have 150,000 to 200 or 300,000 acres in a run. 9" May — A thick fog this morning — it cleared away at 9°C this morning & proved to be a bright warm day. I waited from 10 to 12°C — two mortal hours — on.a Rev. Mr Boyce,” superintendant of the Wesleyan Missions who had appointed to see me about my proposed tour in the Pacific Islands, but who did not make his appearance. So I left a note for him & then went to the Bot. Gardens where Mr Moore proposed to drive [ed. note: Fig. 1 shows the manuscript to this point] me to the Heads of P' Jackson, to see a little of the Bush at this season. We went out by a road bordering the Botany Bay marshes, where Sir Joseph Bankes first botanized, & returned by another road that skirts the sinuosities of Port Jackson. The latter is a most picturesque drive, abounding in beautiful views — & some of the Sydney magnates have beautiful villas large & small, along this road. We botanized in the bush, both as we went along & after we had put up the vehicle at the Inn, near the Heads. There was not much in flower — but Epacrideae, particularly the common E. grandiflora were very gay as was a bright yellow Bossiaea & sundry white Leptospermums — Among the crags by the seashore a tropical looking thing (Morinda jasminoides) was cov- ered with bright orange berries. A green flowered Spurge Laurel (Daphne Indicus) * was in flower & fruit — & a few scrubby palms (Corypha australis) '° were met here & there in the jungle. I filled my collecting book with one thing or another but do not consider it a profitable collection. However the day was a very pleasant one & may be marked white. 10" — I saw Mr Boyce (the Missionary) this morning & in a few words arranged with him for a passage on board the John Wesley”— a schooner of 236 tons, fitted up like a yacht & belonging to the Mission [being] employed in carrying supplies among their stations & conveying the Missionaries from post to post. As I am going in the next cruize you will probably like to know where we set off, & where we are going in the next cruize you will probably like to know where we set off, & where we are going & how long it will take. Know then that we propose setting out about the first of next month June, & expect to be 4 months among the Islands. We first go to Auckland, New Zealand where we stay only a week; — then sail for Tonga Tibou & visit each of the islands of that group staying a few days at each; we then go the Fejee Islands & visit every one of those islands in succession; then return to Tonga which is the Wesleyan Metropolitan See, & so back again to Sydney. I have seen the vessel, which is most comfortably fitted up & greatly praised by Nautical men. She was built at Southampton of Teak, cost £6000. I have a cabin to myself of fair size & there is Proc. LINN. SOC. N.S.W., 115, 1995 = {See a E Z, ‘ Z \. Z : = ie fa 28 © Coe > = Ys) SS Ne . ia 5 Re a0) iS 5 = a bS 7) : = , ~~ 7X « gE s | me ee ot Wn . 7 =a : FS es ~ ee 4 = r ~. = . 7 eas 4 o o) aN e < \ ee - “an a7 Ne ‘o 2 5 ‘ Y Se \ . SQ < S ~Y o) oe ie EG 5 ‘ ee C) | es ‘ ’ =A os : v ATOR TPIT, lg = ie os jae | ge eee cates eee } rr oa — : go . ee swiss sscti rat gitar & pybvsshf% Youn Hw : ~ Cintas fw 2h — Studien reret Ls een ake ties mae tia iat ms i ie: fetaubst omer aoe Satta % | toeonsd Lie ts heats ey trerapriat ; PL eth KM vr shuld (bi 7 waaay ite} pees EP EEG tye, ty HED aang fake ae our wa ta) "a SRE 4 prvmhly dos. fnart shal! CPLeee 2) ~~ par anay Sea das os | ty aK von inet 1 Coker eat ater iel 24) pe ne orf allkind: aS VANDER aa tay Wy 1) [hee tris Vi vt § eek panties aa ts a si : ~ thal 5 “d ipillgg Pe rate Algo rs) va Ber CH PNG REA TENE HY nd ome iB Sry ot ic : > is : 7 3 Hadi ipod eude es foaribtncerceSeeaiemdigant ans 5 chninatisniy? wana enter? wh Rerdrrin 312 gehurtinsy oa hom bre dlasoiiars grit otk bexllol wat or Kile 3 , ae, pearly 2 at Leys, iwat om fos via-scoet it ott Wie Ws ERE Oe) 380 pi we th Peas * wit Remeitys Seats PA Ne "SE Lesinciteniare saree? ha Wee) Hupees. 1 2G entry citecmaA bewetorw igi swnd-denl, (BOHLEL wadgon'y sinus po atten ed goons wets pri a ead aizile hemi aa Liced ) lege pip ah sSiechinwraennenniar Bnei ke chesney Ct fern “y M oxy 2 punt padding fares 11 wilh At ro ge um feat were wit 2. ia ido 215, al reine he Titra aia vat wo iit ae 3 ert Litbtary, ‘Stale Lilyaiy of Mew Santhy Wales tataay { “ niherslly af Mefhourns fraverteeth es ofieaiy aidign WuPlgis so vertiane = nae we the aadixn palatine at i i sai ies nae Roarehrom tnd me cet oe oo a3 Occurrence of Pathogenic Thelohania (Microsporida: Nosematidae) in the Australian Freshwater Crayfish, Cherax quadricarinatus (Decapoda: Parastacidae) G.P. SEMPLE (Communicated by J.R. MERRICK) SEMPLE, G.P. Occurrence of pathogenic Thelohania (Microsporida: Nosematidae) in the Australian freshwater crayfish (Cherax quadricarinatus (Decapoda: Parastacidae). Proc. Linn. Soc. N.S.W. 115: 225-231 (1995). Thelohania was a major pathogen in experimental Cherax quadricarinatus stocks, introduced to North America, infecting late embryos, larvae, juveniles and adults. Spores encysted in heart, limb and abdominal muscles, small numbers of spores were also detect- ed in ovarian and neural tissues. The vigorous but incomplete host response included melanization in most tissues and concentration of spores in the tips of gill filaments; there was no apparent immune reaction in neural tissues. Evidence for direct transmission from parent to embryo is presented. Trials with asymptomatic groups indicated that thermal stress (3-6h exposure to temperatures above 32°C) would induce thelohaniasis symptoms within hours. Regression of these symptoms was also observed when some infected individuals were maintained at lowered tempera- tures (27-31°C). C. quadricarinatus populations free of Thelohania survived protracted exposure to temperatures of 35-37°C. G. Semple, 36 Malvern Avenue, Merrylands, Australia 2160, (formerly Department of Fisheries and Allied Aquacultures and Alabama Agricultural Experimental Station, College of Agriculture, Auburn University, United States of America 36849-5419); manuscript received 15 December 1993, accepted for publication 23 March 1994. KEY WORDS: Thelohania; microsporidian; transmission; environmental stress; Cherax quadricarinatus, Parastacidae; Australia. INTRODUCTION After the plague fungus (Aphanomyces astaci), microsporidial protozoans of the genus Thelohania are responsible for the most important disease problems in freshwater crayfish throughout the world (Alderman and Polglase, 1988); Thelohania species are known from North America (Sprague, 1950), the United Kingdom (Cossins, 1973), Europe and Russia (Voronin, 1971; Vey and Vago, 1973) and Australasia (Quilter, 1976; Mills, 1983). Levels of incidence reported in wild populations have ranged from 0.3-38% in Australia (Herbert, 1988; O’Donoghue et al., 1990) and 10-30% in western Europe (Alderman and Polglase, 1988). Endoparasitic microsporidians infect a range of crayfish species, including some of economic importance; however, their taxonomy is still confused and life cycles are poorly known (Langdon, 1990) — current knowledge is summarized below. Identification and Classification Examination of scrapings or squashes of infected tissues, using a compound light microscope, is required to ascertain if spores and pansporoblasts are present (Merrick and Lambert, 1991: 125). Stains such as Giemsa, haematoxylin and eosin, will highlight spore recognition features (Herbert, 1988). Individual mature Thelohania spores are pyri- form (3.00-3.69 wm length: 2.00-2.35 zm width) refractile in transmitted light and phase bright; 6-7 polar filament windings and 3 layers in the spore wall can be observed with elec- tron microscopy (O’Donoghue et al., 1990). Each sporocyst contains a maximum of 8 Proc. LINN. SOC. N.S.W., 115, 1995 226 PATHOGENIC THELOHANIA IN FRESHWATER CRAYFISH spores, but dimensions of the spore masses (xenomas) which then develop vary consider- ably — up to 2mm in length and 7-80u.m in width; the xenomal wall is reported to be smooth — without invaginations or septae (Herbert, 1988; O’ Donoghue e¢ al., 1990). Although the validity of the genus has not been questioned the familial designation, and numbers of species recognized within the genus, have been subjects of controversy. Usually assigned to the family Thelohaniidae, Thelohania was referred to Nosematidae in recent revisionary studies (Hazard and Oldacre, 1975); but it is below the generic level that most taxonomic confusion exists. Langdon (1990) noted that many species have been assigned to Thelohania and infect vertebrates as well as other invertebrate groups. From the Thelohania material recorded from crayfish several distinct species have been described; but it is still not clear if the Thelohania group parasitizing crayfish comprises a small number of widespread species, each infecting many populations of various crayfish species, or a larger number of isolated parasites adapted to particular hosts (Sprague and Couch, 1971; Kelly, 1979). Life Cycle s ‘In the absence of detailed information Thelohania probably follows the general microsporidian cycle. This general pattern involves initial ingestion of spores by the cray- fish; once in the intestine the spore extrudes a polar filament which penetrates the gut wall. The spore cytoplasm is believed to be liberated through the filament and having pen- etrated the intestinal epithelium it enters an intestinal muscle fibre. Asexual division with- in the muscle produces schizonts, each of which then divides repeatedly to produce a pansporoblast containing 8 sporoblasts within a membrane. Spores may be liberated singly or in the pansporoblast — more muscle cells are subsequently infected by individ- ual spores, so the infection spreads. Although primarily restricted to muscle, spores have been found in the nervous system, the connective tissue surrounding the gut, ovary, devel- oping eggs and in the haemocoel (Johnson, 1977; Alderman and Polglase, 1988). Generally crayfish show little reaction to this intracellular parasite, aside from an inflammatory response associated with muscle cell rupture; most parasites apparently remain as pansporoblasts within the converted muscle cells. Death is considered inevitable, resulting from increasing destruction of muscle — especially buccal and heart muscle; however, survival times vary widely. Infected individuals may take 5 or 6 months to show clinical symptoms and then take a further 12 months to die (Alderman and Polglase, 1988). Thelohania only appears to be transmitted by the ingestion of spores or infected tis- sues containing them (Alderman and Polglase, 1988). Fish and invertebrate secondary hosts or intermediate stages have been suggested (Chartier and Chaisemartin, 1982; Herbert, 1988; Langdon, 1990) but no evidence supporting these ideas has been present- ed; the possibility of transovarian transmission has also been raised, but observations to date have not confirmed it (Alderman and Polglase, 1988). There is no known treatment for thelohaniasis in crayfish and the only control mea- sure is to regularly monitor the population — removing any specimens showing symp- toms. Factors suggested as influencing outbreaks of this disease include crayfish population density, high temperatures and low pH (Alderman and Polglase, 1988; Langdon, 1990). The objectives of studies reported here, on the northern Australian endemic Cherax quadnicarinatus, were: to document initial symptoms and host immune reactions to Thelohania infection; to investigate the evidence for possible alternative modes of trans- mission; and to observe host-parasite interaction under selected thermal conditions. Proc. LINN. SOC. N.S.W., 115, 1995 G.P. SEMPLE 227 MATERIALS AND METHODS Experimental stocks were derived from a consignment of 20,000 Cherax quadricari- natus juveniles (approx. 10mm TL) obtained from commercial suppliers in Queensland and held in isolation in Alabama. These imported stocks were housed in two newly con- structed 0.125 ha ponds, filled and topped up with bore water, and monitored for eight months. A sample of 1400 adults (>30g) was removed from the ponds to hatcheries and the trials reported were done on the progeny produced. The hatcheries were monitored, twice daily, for a period of eight months. Any mortality or disease was investigated by the author with the assistance of the Disease and Parasite Laboratory, Department of Fisheries, Auburn University. All tissue samples were examined with a compound light microscope and Thelohania spores identified from fresh and stained smears. No classification below the generic level was attempted. Reference samples of diseased tissue were retained in preser- vative, but other diseased remains were disposed of by incineration. Observations of reactions to initial infection and the occurrence of spores are listed and results of laboratory trials, in which infected individuals were subjected to high tem- peratures, are summarized in Table 1. RESULTS Identification, Initial Infection Microscopic observation of mature Thelohania spores from abdominal muscle showed that they were uniform in size (3.0m x 2.0 wm) and lightly basophilic, whereas sporonts and sporoblasts were eosinophilic. Examination of frozen tail muscle revealed only spores; no intact pansporoblasts, as observed in fresh tail muscle and other tissues. During the initial stages of infection C. guadncarinatus appeared normal and the first observable external symptom was the development of white or grey streaks in the anterior ventral abdominal musculature. This discolouration corresponded to a concen- tration of spores and the formation of xenomas. The abdominal infection (muscle destruction) then spread posteriorly; however, examinations of a number of adults revealed spores, pansporoblasts and xenomas in mus- cles of the heart, chelipeds and pereiopods. Small numbers of spores were also present in connective layers surrounding the ovaries and neural tissues. Immune responses observed included melanization in most tissues and the concentration of spores in distal portions of gill filaments. This latter reaction of aggregation and melanization in gill fila- ments was conspicuous in larvae. As the trials reported below demonstrate, Thelohania also infects embryos, larvae and juveniles of C. quadricarinatus. Environmental Stress Newly fertilized eggs when removed from healthy females and incubated in isola- tion, did not survive to hatching at any temperature; despite all precautions total mortali- ty resulted from fungal infections. By contrast, a large number of eggs (totalling approximately 30,000) removed at blastula, early gastrula or later stages had hatching rates of 30%-98% if maintained at 18-31°C. Furthermore, up to 60% of late stage embryos (incubated in isolation) survived prolonged exposure (>3h) to temperatures of 32-36°C. Batches of larvae (from eggs incubated in isolation) had survival rates of 60%-100% after 3-6h exposure to temperatures of 32-36°C, but larvae maintained below 31°C had less than 5% mortality. Likewise juveniles exposed to high temperatures had survival rates of up to 40%, whereas 80-95% of those below 31°C survived (Table 1). Most of the mortality associated with the artificial incubation trials involving late Proc. LINN. SOC. N.S.W., 115, 1995 PATHOGENIC THELOHANIA IN FRESHWATER CRAYFISH 228 “WISITeqrIuUeD o[eULaJ OF ONp ATTeIOUI % OT pomoys sa;duresqns aurog# ‘saimjeiodura JUdIIPIp 0} Np UOHLINP UI UONLLILA VPI ~ “ystp Wad ut pourejureu ‘Juoied wioy paypeiep = parejosy 4 -(uondsosqe yJoA) aseajar oj dn ssoy ay se payepngyeo st it 9eAIET Ur ‘SuTyiey 0} UdyEI st} sa8eIs OAIquIa /33a Ut :sueds aun JUdIAFIp O} JIeJaI sanyea Aye08 || x shed 8 "HOMTEIOW ——- Do L808 006 ‘syse|qoiodsued ‘sa.1ods 99.4 ‘pasrasqo 1s.17 su10}d uAs Joye SIY FHI-E pPaip [ye ‘aseasout ainjyessdura} Jo s1y 9G-E UTYIIM soTeulay surojduuds padojaaap uonepost ur pjay sareuray asayT,"%OOTAMEMOW = D0. 8E-oE € petieg SLTNaV skeq 09 (0g) "%06-¢ AITRIOW Dol&-066 000‘0€ (uoneimeA) “TL Dol, G06 ce) WID0¢ 0} BSeRT9 ‘avarry 10g se suroiduns °% 001-09 MEOW =—-- OSIUO-E 000°6 SATINAAN (0g) poveyost ‘QRAIL] PayepOs! pure poyoene yIoq ul %G-9 Aupers0 | Dol G-066 000‘0€ so[euloy UO she GFF aseajai OL ‘saystp nad ut % 0F-9 Ape: parejos] ‘uraysds A10yeTNII Ul s9.10ds JaJq Sv [[aM Sv S9BRIS ea ur s}se{qoi1odsued aatsua}xa pur saseys ApIea ut s}sepqo1odsue 6L-GT Aeq jo Sajpung ‘paseaivap asuodsai yoo F [re} pue JuUsUIIAOUI Zaseis sv a}1yM AJa}9]duI09 Udy) ‘papnoys suledaq xeIOY) pue spadijayo 09E=o68 (0€) YL asus ‘spodorasad ‘uawiopae Jo sapsnyy -% 001 0} dn Arper0py] ® SIU 9-¢ 000‘¢ SaTeUlaj UOC AVA ‘saysip unad ur % 04-2 Aye0W (OL) parepos] i # (Wusyeqruued ajeuray auIOS) %OT-g AfeIOU aseI9Ay OolE-o8T 000°L soTewoy UO SARC 96-6 eo ae = LN oS 2 Se ee a Se eae SS Se es SSS a ee yoy OF, ‘soysip ined ut % 00 T-OF ANTE ON porejos] ‘Sa8eIs 1918] SOANAWA ursjsejqoiodsued aaisua} xa pure sa8e}s Aj1ea ul sysejqo1odsued jo FL O1G Arq sa]pung ‘paseaidap JUaUIAAOW SB a}1YM AT9}9|duI09 Udy papnoya payuowsid sada auredaq xeioy) pur spadyayp ‘spodotaiad ‘uau0pae jo sapsny o9E-0E (02) —snydneuiysog “pap sapewy [Te UayM %0OT-09 ANTeIOW| @ SIY 9-E 000% Soyeuloy UO AOVLS ALVI %OOT ATerIOWI (OL) porepos] %O1-G Anes Dol€-066 000°L Sofeulay UO skeq S11 = a = 6 es So ae ee. an ek SS = ce a eee yoiey Of, ‘wyeap 0} 10ud UOLIOYS F-0 eq uo padoyaaap 18ung {(ystyM) pammoyoosip auredaq UOLIOY) o9E-0BE (02) -parejosy — smadneu-Surumedg 's880 palvyost pue payoene Tog UI YOOTSeM ,AMPELOP, = —-—« ® SY O-E 000°% sayeulay UO soog uonemg SUOQHEAIISGQ =. Ua ALIAT, (sajdures-qns) uonwOT astis aydures ur Jaquimu [e}0 J, 066 I Unf 01 696 [ 429012C) powiad ay) durunp spor) ssags pousay) fo synsas puv spuauyna. ‘saquivs fo Lanmung T TIavy Proc. LINN. SOC. N.S.W., 115, 1995 G.P. SEMPLE 229 embryos at temperatures >33°C can be attributed to Thelohania, as spores occurred in muscle tissue. Samples of larvae which had been exposed to the general population and subjected to high temperatures (33°-36°C) , developed a very high incidence of thelohani- asis within 6-72 h. Clinical symptoms of thelohaniasis were often observed following rough handling or moulting. Up to 10% of larval, juvenile and adult C. quadricarinatus presented symp- toms 3-8 days after ecdysis. It was also noted that in some juveniles subjected to thermal stress the initial infec- tion symptoms began to regress when the animals were returned to temperatures below 31°C. The anterior abdominal cysts began to clear — from the original infection site first, then posteriorly. DISCUSSION Thelohaniasis is the most important disease of crayfish in Australia; it has been reported in South Australia, Victoria, New South Wales and Queensland (O’ Donoghue et al. 1990; Merrick and Lambert, 1991: 121) and at least three Cherax species may be infect- ed (Mills, 1983; Herbert, 1987; Semple, 1993, unpublished data). It is also acknowledged that previous reports of infection levels — largely based on individuals exhibiting clinical symptoms — are underestimates. Identification As mentioned previously the number of Thelohania species parasitizing crayfish is unknown. Little variation in mature Thelohania spores has been observed in Australia (Herbert, 1988; O’ Donoghue et al., 1990); however, some variation in spore size from dif- ferent hosts has been reported between continents (Cossins and Bowler, 1974; Quilter, 1976; Herbert, 1988). The determinations are based on fresh squashes and the presence of separate spores only in frozen material is considered due to lysis of the pansporoblast walls; this phenomenon of lysis as a result of the freezing process has also been reported in prawns (Owens and Glazebrook, 1988). Although the specific identity of the Thelohania studied was not determined, the evi- dence suggests that it was an Australian species imported with the experimental stocks. Reasons for drawing this conclusion include: no observation of any mortality in experi- mental stocks with Thelohania symptoms prior to transfer to hatchery; the new, isolated ponds supplied with bore water — in an area where thelohaniasis has not been reported; the thorough preparation of hatchery facilities where Thelohania had not previously been recorded. The probability of contamination from an external source is very low; further- more, the direct transmission mode (transfer of spores from adults to developing eggs), as demonstrated by these studies, indicates a mechanism to explain how the disease could have been present although undetected. Initial Infection, Transmission Immune reactions to Thelohania were observed in all stages of C. quadricarinatus from larvae to adults; however, host response was incomplete. The observation of Herbert (1988) that there was no apparent response to the small number of spores associated with nerve tissue was confirmed. The overall distribution of spores in adult tissues was also as reported by Herbert (1988) and is consistent with transmission by means of ingestion and invasion from the gut. Whilst the studies reported here do not corroborate this indirect transfer, character- istic of microsporidians, they do demonstrate a second direct transmission mode in Cherax quadncarinatus. This is the first time that Thelohania has been reported from developmen- tal stages in an Australian species. Proc. LINN. SOC. N.S.W., 115, 1995 230 PATHOGENIC THELOHANIA IN FRESHWATER CRAYFISH Another Thelohania species is known to be transferred directly in vertebrate host eggs (Post, 1987: 173); furthermore, Voronin (1971) reported Thelohania spores in embryos of an astacid crayfish. Microscopic examination did not reveal spores in C. guadn- carinatus oocytes, but they were present in the ovarian wall. Furthermore, it should be emphasized that detection of occasional spores in oocytes would be difficult because of the dense, granular nature of oocyte contents. Although spores were not actually observed in C. quadricarinatus ova, their occurrence in ovarian tissues and embryos (4 days old) indicates that they were almost certainly present. The only alternative explanation is that free mature spores infected eggs rapidly by directly entering ova at release, during fertilization or soon after. It is clear that Thelohania does not need an intermediate host for positive activation in Cherax quadricarinatus, how- ever, the trials reported do not preclude the possibility that, under some conditions, inter- mediate vectors may be involved in transmission. Previous arguments for the involvement of intermediate hosts have centred around the need for priming of the spore; however, the present research indicates the possibility that spores could be primed by a physicochemical environmental factor, such as tempera- ture or pH, without the need to enter another vector. The occurrence of Thelohaniaspores in other benthic invertebrates, such as simuliids (Chartier and Chaisemartin, 1982), does not mean that they are intermediates. Environmental Stress The detection of spores in fresh tissue squashes eliminates the possibility that the macroscopic clinical symptoms were due to lactic acidosis induced by thermal stress. Langdon (1990), in general comments about microsporidians, noted that: their incidence in crayfishes decreased with increased latitude; incidence increased with crowding and high temperatures. The implication was the incidence may be related to stressful conditions. These studies have confirmed this suggestion; however, it is not clear whether initiation of parasite activity is due to priming of spores, lowering of host immu- nity or both. The lack of response in previous infection trials with C. quadricarinatus (Herbert, 1988) may be, at least partly, due to the relatively low temperatures (18-24°C) at which experimental stocks were held. Other observations which support the hypothesis of environmental stress influenc- ing Thelohania activity include: the 10% incidence reported in C. quadricarinatus up to 8 days after ecdysis; infections, in small numbers of C. destructor, which developed after short exposure to lowered dissolved oxygen levels (Semple, 1993, unpublished data). The observation of regression of symptoms in animals returned to lowered temper- atures is significant in two ways. It is the first indication that this infection is reversible and secondly may assist in the development of control measures. Management The present studies have not fully elucidated the transmission of Thelohanza, but the results have several broad implications for crayfish conservation and culture in Australia. Firstly, of the Australian species carrying Thelohania, Cherax quadricarinatus and C. destruc- tor are extensively cultured and have been widely translocated. The possible impact of dis- ease outbreaks in cultured stocks on other indigenous parastacids, many of which have restricted distributions, is unknown (Merrick, 1993: 90-92) but should be investigated urgently. Secondly, the role of environmental stress should be considered when decisions are taken on culture at the margin of, or outside, a species range. In these marginal areas lim- iting environmental conditions (for that species) are likely to be encountered periodical- ly; this pathogen could be present (but asymptomatic) for many months and then be rapidly activated by a short period of extreme conditions. Proc. LINN. SOC. N.S.W., 115, 1995 G.P. SEMPLE 231 Thirdly, the direct transmission demonstrated means that screening or quarantine procedures should include examinations of nerve tissues, eggs and larvae as well as muscle; short duration thermal testing may also be useful. Finally, a key observation emerging from these studies is that this infection is not necessarily irreversible. The immune reactions observed from larva to adult may, in com- bination with manipulation of selected environmental factors, form the basis of a control strategy. ACKNOWLEDGMENTS Most of the research reported was undertaken while the author was a Rotary International Vocational Scholar and Honorary Associate of the Department of Fisheries and Allied Aquacultures, Auburn University, Alabama; Drs E.W. Shell, D.B. Rouse and B.L. Duncan and supporting staff, of Auburn University, are thanked for their assistance. The use of Macquarie University facilities, during completion of studies in Australia, is acknowledged. Appreciation is also expressed to Mr and Mrs N. Semple for support and Miss P. R. Davies for manuscript preparation. References ALDERMAN, D.J., and POLGLASE, J.L., 1988. — Pathogens, parasites and commensals. Jn: HOLDICH, D.M., and Lowery, R.S., (eds). Freshwater Crayfish: Biology, Management and Exploitation. Chapter 7, pp. 167-212. Croom Helm Ltd., London. CHARTIER, L., and CHAISEMARTIN, C., 1982. — Relationships between biotypes and autochothonos crayfish popula- tions (healthy or attacked by microsporidiosis): Austopotamobius pallipes (Lereboullet, 1858). Annales Station Biologique Bessen — Chadesse, 16: 42-125. Cossins, A.R., 1973. — Thelohania contajeant Henneguy. Microsporidian parasite of Austropotamobius pallipes Lereboullet — an histological and ultrastructural study. Freshwater Crayfish, 1: 151-164. Cossins, A.R., and Bow er, K., 1974. — An histological and ultrastructural study of Thelohania contajeani Henneguy, 1892 (Nosematidae), microsporidian parasite of Austropotamobius pallipes Lereboullet. Parasitology, 68: 81-91. HazarbD, E.I., and OLDACRE, S.W., 1975. — Revision of Microsporidia (Protozoa) close to Thelohania with descrip- tions of one new family, eight new genera and thirteen new species. Agriculture Research Service, United States Department of Agriculture, Technical Bulletin 1530: 90-104. HeErsert, B., 1987. — Notes on diseases and epibionts of Cherax quadricarinatus and C. tenuimanus (Decapoda: Parastacidae). Aquaculture, 64: 165-173. HERBERT, B., 1988. — Infection of Cherax quadricarinatus (Decapoda: Parastacidae) by the microsporidium Thelohania sp. (Microsporida: Nosematidae). Journal of Fish Diseases, 11: 301-308. JOHNSON, S.K., 1977. — Crawfish and freshwater shrimp diseases. Texas A and M University, Sea Grant Publication TAMU-SG-77-605: 1-18. KELLY, J.F., 1979. — Tissue specificities of Thelohania duorara, Agmasoma penaei and Pleistophora sp., microsporidian parasites of the pink shrimp, Penaeus duorarum. Journal of Invertebrate Pathology, 33: 331-339. LANGDON, J.S., 1990. — Major protozoan and metazoan parasitic diseases of Australian finfish. Post Graduate Committee in Veterinary Science, University of Sydney Proceedings (128) : 233-255. MERRICK, J.R., 1993. — Freshwater Crayfishes of New South Wales. Linnean Society of New South Wales, Sydney. MERRICK, J.R., and LAMBERT, C.N., 1991. The Yabby, Marron and Red Claw. Production and Marketing. J.R. Merrick Publications, Sydney. MILLs, B.J., 1993. — A review of diseases of freshwater crayfish, with particular reference to the yabbie, Cherax destructor. Department of Fisheries, South Australia, Fisheries Research Paper, (9): 1-18. O’ DONOGHUE, P., BEVERIDGE, I., and PHILLIPS, P., 1990. — Parasites and ectocommensals of yabbies and marron in South Australia. Central Veterinary Laboratories, South Australian Department of Agriculture, VETLAB Report. OWENS, L., and GLAZEBROOK, J.S., 1988. — Microsporidiosis in prawns from northern Australia. Australian Journal of Marine and Freshwater Research, 39: 301-305. Post, G., 1987. — Textbook of Fish Health. T.F.H. Publications, Neptune City, New Jersey. QUILTER, C.G., 1976. — Microsporidian parasite Thelohania contajeani Henneguy from New Zealand freshwater crayfish. New Zealand Journal of Marine and Freshwater Research, 10: 225-231. SPRAGUE, V., 1950. — Thelohania cambari n. sp., a parasite of North American crayfish. Journal of Parasitology (Supplement), 36: 46. SPRAGUE, V., and Coucu, J., 1971. — An annotated list of protozoan parasites, hyperparasites, and commensals of decapod Crustacea. Journal of Protozoology, 18: 526-537. Vey, A., and Vaco, C., 1973. — Protozoan and fungal diseases of Austrapotamobius pallipes Lereboullet in France. Freshwater Crayfish, 1: 165-179. VoRONIN, V.N., 1971. — New data on microsporidiosis of the crayfish Astacus astacus (L. 1758). Parazitologiya 5: 186-191. Proc. LINN. SOC. N.S.W., 115, 1995 7 SOE ‘ bux mietiiehne she a cap HE MEREHOF? tieelsciasi saan Bite wit ‘piston rein Aieowe orn Cx oe “tee wii. i ee it hay by entiver’ { Apaaphe te ¥: wt baal hake sieges lorsivanalhe ose A | byes ih sons is aan ria Wn Theo het we ied pf cal loon Fisk uasuaMaioen Poe, ety iy) baa ] wy, 2G ial ? SA big, ‘ 2 A ee seh age re dad TaiMogee raleer' Nr Rts IN LA ERISA rs Lea Seat Dye Bek or PHA eh apsindyytie AaSpyenpnl Ws heiguntrsare riath 2 teat rarest Bion shnailen 2 yf) rehome JO eeae(e tice SeONROn Ay oth el ise Seve x eee Eres bly cue oe aye. ra . r; wale Cala ait ae “ah He RAL Nis eres SVAN LeSttiyand minaret << ~ ‘HOE Bh ; Hrinyal | ali bee wets wlved anh) tadh new aeric si airy we Vel, MO Paley take + oe ae Mpa) . ware.) PPM CAG wenn? tor Bey Ati ulin e ANNA Ten El iva dy irk PowbelP ocahiaditde ste Tah A ee Se uy Sele Vr amin Aa caaeigd wt enmity jiichaprtnelont wre ndu iad kaliawnglt brant Hn gay 5 itimar cnethere als 4 Cola RA Aah sentna/lanit}aibtarstttem icant be omnon i ‘ cap nen yore tape An pry ety pF AN ME piwh;s nee th one (for pee he i js set yaairspa eh nema tes tonto via eobiieaialeanstly> hast er saroponin eee 2 ee ee First Records from Wellington Caves, New South Wales, of the Extinct Madtsoiid Snake Wonambi naracoortensis Smith, 19'76 JOHN D. SCANLON (Communicated by M. ARCHER) SCANLON, J.D. First Records from Wellington Caves, New South Wales, of the Extinct Madtsoiid Snake Wonamli naracoortensis Smith, 1976. Proc. Linn. Soc. N.S.W. 115: 233-238 (1995). Two trunk vertebrae of a species of Wonamli (Serpentes, Madtsolidae) from unknown deposits in the Wellington Caves system are the only specimens of this taxon known from New South Wales, and considered to be Pleistocene in age. The specimens are referred to the type species, W. naracoortensis Smith, 1976. Together with other known occurrences of the genus (in South Australia and Western Australia), this record supports an hypothesis of association with temperate climates during the Quaternary. Other speci- mens previously referred to Wonamli are reassessed with implications for the distribution of Wonambiand another large madtsoiid, Yurlunggur Scanlon, 1992. John D. Scanlon, Vertebrate Palaeontology Laboratory, School of Biological Sciences, University of New South Wales, N.S.W. 2052; manuscript rececved 8 March 1994, accepted for publication 19 October 1994, KEY WORDS: Serpentes; Madtsoiidae; Wonamli naracoortensis, Yurlunggur sp.; Quaternary; Distribution; Wellington Caves, N.S.W.; Vertebrae. INTRODUCTION The two specimens described here have somewhat different histories, although both were collected from unknown deposits in the Wellington Caves system, and both passed through the hands of Jack Mahoney (University of Sydney, Geology) as well as the collections of the Australian Museum (AM). During work by the author on the AM verte- brate fossil collection in 1987, several small lots of unregistered material came to light which, as indicated by handwritten labels, had been collected from Wellington Caves by Mahoney et al. during the 1960s. In one of these was a single large vertebra showing char- acteristic features of the extinct family Madtsoiidae (described as a subfamily of Boidae by Hoffstetter, 1961, but treated as a distinct family by McDowell, 1987 and Scanlon, 1992, 1993). A second specimen was located by Paul Willis in the collections of the Department of Geology and Geophysics, University of Sydney, in a batch of Wellington material which appears to have been collected prior to 1915, most probably in Gaden Cave, and later set aside from the rest of the collection by Mahoney (Willis et al., 1992). After comparison with the published descriptions of Wonambi naracoortensis Smith, 1976 and other madtsoi- ids, vertebrae of Wonamli from Naracoorte, and Tertiary material of related taxa from sev- eral northern Australian sites, there seems no reason to distinguish the Wellington specimens taxonomically from W. naracoortensis. FAMILY MADTSOIIDAE HOFFSTETTER, 1961 WONAMBI NARACOORTENSIS SMITH, 1976 The first vertebra has been registered as F72999 in the AM palaeontological collec- tion. The only data on the handwritten label with the specimen are “Stephenson and Mahoney 1962’. Neville Stephenson (pers. comm.) (now retired from the staff of the Proc. LINN. SOC. N.S.W., 115, 1995 234 WONAMBI NARACOORTENSIS FROM WELLINGTON CAVES University of Sydney, and living in the U.K.) recalls one occasion to which this could refer, but was not otherwise familiar with Wellington Caves and unable to provide any details of the site(s) from which the material came. Jack Mahoney died in 1985. If any further details of collection and stratigraphic provenance exist, they are not presently known. Other skeletal material with the same data include a dingo skull (Canis familias dingo) and sev- eral marsupial species, but it would be unwarranted at present to assume that all this mate- rial is of the same age or comes from the same deposit. The vertebra is a middle thoracic, within the size range of the holotype and paratypes of Wonambi naracoortensis from Victoria Cave (Smith, 1976, table 1). It is dam- aged but has most features represented nearly intact on one or both sides. Two paracoty- lar foramina, and a single distinct parazygantral foramen, are present on each side. The neural spine is broken near its base, and the hypapophysis is also practically missing (but can not have been very wide), the ventral surface of the centrum showing an area of damage suggesting a single shear fracture. The paradiapophysis on the right side is practi- cally absent, also apparently sheared off, but the bone surface is preserved almost intact on the left. The preservation appears appears similar to that of the Victoria Cave specimens examined (South Australian Museum P16166 and P16170b): no secondary mineralisa- tion or deep staining is apparent, and some of the broken faces appear quite fresh. Most of the neural spine being broken off, the internal sinus of the neural arch is exposed; it con- tains a network of struts like the long bones of birds. The base of the broken paradia- pophysis shows a more spongy structure as seen when these processes are only slightly worn; the break on the ventral surface shows that the base of the hypapophysis, while still containing spaces and channels, is more solidly constructed. The second specimen was located in the collections of the Department of Geology and Geophysics, University of Sydney, by Paul Willis, who recognised it as possibly Wonamhi and brought it to my attention. It has also been transferred to the AM collection and is registered as F92050. The original handwritten label with this batch of material (also including extinct, characteristically Pleistocene mammals) appears to have read ‘New Cave, Wellington’ which is considered most probably to refer to Gaden Cave (Willis et al., 1992). The age of the snake fossil is probably similar to that of the other material, thus Pleistocene. This is a smaller vertebra (Fig. 1) representing a more posterior part of the trunk, probably from a snake of about the same size (neural canal height, which appears to be less strongly dependent on vertebral position than other dimensions, is similar in the two specimens; Table 1). It is similarly preserved, but more complete than the other: only the neural spine and right postzygapophysis are broken off. The keel of the neural spine extends close to the anterior edge of the zygosphene. The paradiapophyses are conspicu- ously wider than the zygapophyses, their smooth surfaces almost intact; the strongly con- vex diapophysis and flat parapophysis are additionally set off from each other by a constriction of the posterior edge. There is a distinct concavity in the dorsal edge of the diapophysis (in lateral view), which appears to be typical and rather distinctive for this species. Distinct paracotylar, parazygantral, zygantral, upper and lower lateral, and sub- central foramina are present. There are also regions of scattered small pits (as in the type material; Smith, 1976, p. 43 and fig. 2). These occur lateral and dorsal to the paracotylar and parazygantral foramina, dorsal to the diapophyses, at the base of the neural spine in the middle of the vertebra’s length; and across the midline, on the dorsal surface of the zygosphene, below the cotylar rim and posteriorly on the hypapophysis. In contrast to mid-trunk specimens, the haemal keel is well defined laterally by subcentral grooves or lymphatic fossae (La Duke, 1991; cf. Smith, 1976, fig. 2) and is not developed as ‘paired hypapophyses’ either as in the holotype or as in posterior trunk vertebrae of Yurlunggur camfieldensis (Scanlon, 1992, fig. 1). A weak but distinct narrow keel is present in the mid- line, strongest in the middle of the vertebra, between the subcentral foramina; these are Proc. LINN. SOC. N.S.W., 115, 1995 J.D. SCANLON 235 Fig. 1. Posterior trunk vertebra of Wonambi naracoortensis from Wellington Caves (?Gaden Cave), NSW (AM F92050). From top: lateral, posterior, anterior, ventral and dorsal views. Scale bar equals 1 cm. the largest of any on the vertebra, lie close to the midline and open anteriad. Table 1 gives linear measurements of both vertebrae, which allow comparison of proportions with the type series (Smith, 1976). Proc. LINN. SOC. N.S.W., 115, 1995 236 WONAMBI NARACOORTENSIS FROM WELLINGTON CAVES TABLE 1 Dimensions (in mm) of W. naracoortensis vertebrae from Wellington Caves, New South Wales. An asterisk indicates the measurement is affected by damage to the specimen, and would originally be greater. F72999 F92050 Length between zygapophyses 20.75 16.25 Maximum height (condyle-neural spine) 34.50* 24.05* Width across prezygapophyses 33.95* 26.95 Width across paradiapophyses 32.60* 30.20 Width across postzygapophyses 33.30 26.50 Minimum width of neural arch 25.70 18.00 Zygosphene width 13.00* 9.65 Zygantrum width 15.95 10.30 Condyle width 13.80 11.45 Condyle height (oblique) 12.20 8.65 Cotyle width 14.10 12.55 Cotyle height 10.85* 9.30 Centrum length (cotyle-condyle rim) 14.55 12.20 Neural canal height 3.95 3.85 Neural canal width 7.20 4.55 Neural spine — paradiapophysis 34.55* 25.95* Cotyle — zygosphene height 24.70 18.80 Zygosphene — condyle (oblique height) 31.40 23.60 (L/R) (L/R) Prezygapophyseal facet width 6.25 / 6.20 4.20 / 4.70 Prezygapophyseal facet length 11.55*/ 11.70 8.90 / 8.80 Postzygapophyseal facet width 6.20 / 6.40 4.50 / —* Postzygapophyseal facet length 10.70 /10.60 8.85 / —* Zygosphenal facet width 5.00 / 4.75* 3.50 / 3.60 Zygosphenal facet length 6.70* / 8.30 5.65 / 5.70 Zygantral facet width 4.60 / 5.00 3.00 / 2.95 Zygantral facet length 8.10 / 7.70 5.75 / 5.70 Paradiapophysis facet width 6.60 / —* 6.90 / 6.65 Paradiapophysis facet length 14.70 / —* 12.00 /12.00 OTHER PLIOCENE AND PLEISTOCENE RECORDS Pledge (1992) records about twenty vertebrae and some jaw fragments of large madtsoiid snakes (referred to Wonambi sp. cf. W. naracoortensis) from the Curramulka Local Fauna, Corra Lynn Cave, Yorke Peninsula, South Australia. This fauna is thought to be probably early Pliocene in age (i.e. several million years older than known W. nara- coortensis) ,and to be drawn from a forested habitat (rodents, bandicoots and grazing her- bivores are absent; Pledge, ibid.). When I examined some of the material in 1990, I concluded only that Wonamli was definitely present. One of the specimens figured by Pledge (SAM P26535, fig. 3a), from the middle to posterior trunk, has a high neural spine and is probably Wonamhn. However, the fauna may also include Yurlunggur, SAM P29908 (ihid., fig. 3b; posterior trunk vertebra) appears very similar to Yurlunggur camfteldensis (cf. Scanlon, 1992, fig. 1E). (But it is not possible to identify snake vertebrae positively from a single illustration; five views are required for adequate description.) The Curramulka madtsoiid material should be studied further, as it may indicate sympatry between large species of Wonambiand Yurlunggur. Pledge (1992) reports W. naracoortensis from the ‘Plio-Pleistocene Kanunka Local Fauna of the Katipiri Sands, Lake Kanunka’, South Australia. However, Tedford et al. (1992) place the Kanunka Fauna in the lower Pompapillina Member of the Tirari Formation, and consider it to be approximately 3.4 million years old (Pliocene). I have not seen this material, but consider it to be possibly either Wonamli or Yurlunggur. McNamara (1990) figured a vertebra from the Late Pleistocene Wyandotte Local Proc. LINN. SOC. N.S.W., 115, 1995 J.D. SCANLON 237 Fauna, northeastern Queensland, which he correctly identified as comparable to Wonamhi and referred to Wonamli cf. W. naracoortensis. The fauna is regarded as ‘typically Pleistocene’ and comparable to many similar-aged deposits of southern Australia (McNamara, ibid.). The single madtsoiid vertebra (NMV P186652) was well illustrated (tbid., fig. 4G-K) , and has been kindly provided on loan by G. McNamara. There is a question as to its correct generic assignment; the specimen lacks the most distinctive characters (neural spine and paradiapophyses) and the haemal keel is more similar to those of posterior trunk vertebrae in Yurlunggur (Scanlon, 1992, fig. 1D-E) than to available specimens of Wonamidn. Provisionally I refer it to Yurlunggur sp. If this identifi- cation is confirmed, it represents the latest known record of Yurlunggur, which is other- wise known from possible late Oligocene to middle Miocene of northern and central Australia (Scanlon, 1992; pers. obs.). Lydekker (1888) listed six ‘large’ snake vertebrae in the British Museum (Natural History) collections (BM(NH) 42682) from Pleistocene deposits in the Wellington Caves system; these were regarded as pythonine, differing from Morelia but ‘agreeing closely’ with Nardoa schlegelii (= Bothrochilus boa) except in size. Particularly considering their large (but unspecified) size, it could have been thought that these specimens might represent an earlier record of Wonamli, which has larger vertebrae than most pythons; but in fact the specimens under this number are of a large elapid snake (not yet identified further). DISCUSSION Wonambhi naracoortensis was the first non-pythonine boid (sensu lato) species described from Australia, now recognised as a member of the extinct Gondwanan family Madtsoiidae. Other madtsoiids are now also known from a number of Tertiary faunas in northern Australia (Yurlunggur camfteldensis Scanlon, 1992 from Bullock Creek, N.T.; species of Alamitophis and Patagoniophis from Murgon, Qld; and other species of Yurlunggurand Wonambifrom local faunas at Riversleigh, Qld.; Scanlon, 1992, 1993 and in prep.). The present record brings the number of localities for W. naracoortensis to five, all in southern Australia: Victoria Cave and Henschke’s Cave, Naracoorte, southern South Australia (Smith, 1976; Barrie, 1990); Mammoth Cave and Koala Cave, southwestern Western Australia (fide Molnar, 1982); and now Wellington Caves, eastern New South Wales. These are all Quaternary deposits in limestone cave systems, mostly Late Pleistocene in age but with older and younger remains also present. In the case of Naracoorte, the other reptile species present do not appear to differ significantly from those now occuring in the same area, and include several elapids but no pythons (Smith, 1976); few reptile remains have yet been identified from the other sites (Dawson, 1985; Molnar, 1982). W. naracoortensis is thus represented by more complete material and from a wider geographic range than any other madtsoiid snake, but its biology is still a matter of specu- lation rather than firm inference. Its broad southern Australian distribution in the Pleistocene can be compared with that of the extant Tiger snake, Notechis scutatus (Peters) (Elapidae), and suggests tolerance of cool climates, whereas Yurlunggur, apparently pre- sent in the south earlier but by the Pleistocene persisting only in the north, may have been less cold-tolerant. Large madtsoiids coexisted with pythons of the genus Morelia in the Oligo-Miocene of northern Australia (Smith and Plane, 1985; Scanlon, 1988, 1992; Kluge, 1993), and at Riversleigh, pythons but not madtsoiids are found in cave deposits (pers. obs.). The Carpet python Morelia spilota (Lacépede) is today broadly distributed across southern Australia, but it has not been recorded from the cave deposits where Wonamli occurs. Whatever the environmental or biotic factors that drove the last madtsoiids to extinction, pythons appear to be the most direct beneficiaries of their demise. Proc. LINN. SOC. N.S.W., 115, 1995 238 WONAMBI NARACOORTENSIS FROM WELLINGTON CAVES ACKNOWLEDGMENTS I would like to thank Neville Pledge (South Australian Museum) for loan of Wonamlhi paratypes; John and Julie Barrie for loan of specimens, and discussion of Wonambi and the Madtsoiidae; Greg McNamara for loan of the Wyandotte specimen; Bob Jones for supervision and assistance at the Australian Museum; Neville Stephenson for information; Paul Willis for finding the second specimen and videotaping material in the British Museum; and Mike Archer and Glen Ingram for reading and constructively criti- cising the manuscript. References BARRIE, J., 1990. — Skull elements and additional remains of the Pleistocene boid snake Wonambi naracoortensis. Memoirs of the Queensland Museum 28 (1): 139-151. Dawson, L., 1985. — Marsupial fossils from Wellington Caves, New South Wales: the historic and scientific signifi- cance of the collections in the Australian Museum, Sydney. Records of the Australian Museum 37 (2): 55-69. HOFFSTETTER, R., 1961. — Nouveaux restes d’un serpent Boidé (Madtsoia madagascariensis nov. sp.) dans le Crétacé Supérieur de Madagascar. Bulletin du Muséum nationale d Histoire naturelle, sér. 2, 33: 152-160. KuuGE, A.G., 1993. — Aspidites and the phylogeny of pythonine snakes. Records of the Australian Museum Supplement 19: 1-77. La Duke, T.C., 1991. — The fossil snakes of Pit 91, Rancho La Brea, California. Natural History Museum of Los Angeles County Contributions in Science 424: 1-28. LYDEKKER, R., 1888. — Catalogue of Fossil Reptiles and Amphibians. Vol. 1. British Museum, London. McDowELL, S.B., 1987.— Systematics. 1-50 in R.A. Seigel, J.T.C. Collins & S.S. Novak (eds), Snakes: Ecology and Evolutionary Biology. MacMillan, New York. McNamara, G., 1990. — The Wyandotte Local Fauna: a new, dated, Pleistocene vertebrate fauna from northern Queensland. Memoirs of the Queensland Museum 28 (1): 285-297. MOoLnak, R.E., 1982. — Cenozoic fossil reptiles in Australia. 227-233 in The fossil vertebrate record of Australasia, P.V. Rich and M. Thompson (eds). Monash Univ. Offset Printing Unit, Clayton, Vic. PLEDGE, N.S., 1992. — The Curramulka Local Fauna: a new Late Tertiary fossil assemblage from Yorke Peninsula, South Australia. The Beagle, Records of the Northern Territory Museum of Arts and Sciences9 (1), 115-142. SCANLON, J.D., 1988. — The snakes of Riversleigh. Riversleigh Notes 3: 7-8. SCANLON, J.D., 1992. — A new large madtsoiid snake from the Miocene of the Northern Territory. The Beagle, Records of the Northern Territory Museum of Arts and Sciences 9: 49-59. SCANLON, J.D., 1993. — Madtsoiid snakes from the Eocene Tingamarra Fauna of eastern Queensland. Kaupia - Darmstadter Beitrage zur Naturgeschichte 3: 3-8. SMITH, M.J., 1976. — Small fossil vertebrates from Victoria Cave, Naracoorte, South Australia. IV. Reptiles. Transactions of the Royal Society of South Australia 100: 39-51. SMITH, M.J. and PLANE, M. 1985. — Pythonine snakes (Boidae) from the Miocene of Australia. Bureau of Mineral Resources Journal of Australian Geology and Geophysics 9: 191-195. TEDFORD, R.H., WELLS, R.T. and BARGHOORN, S.F., 1992. — Tirari Formation and contained faunas, Pliocene of Lake Eyre Basin, South Australia. The Beagle, Records of the Northern Territory Museum of Arts and Sciences 9: 173-193. WILLIs, P.M.A., Davies, S.M. and OSBOURNE, R.A.L., 1992. — Important vertebrate fossils from the palaeontologi- cal collections of the Department of Geology and Geophysics, University of Sydney. Jowrnal and Proceedings of the Royal Society of New SouthWales 125: 113-118. Proc. LINN. SOC. N.S.W., 115, 1995 Lower Permian Fauna from Manning Facies Rocks along the Peel-Manning Fault System, Glenrock Station, Southern New England Orogen JAMES STRATFORD and JONATHAN AITCHISON STRATFORD, J.M.C., and AITCHISON, J.C. Lower permian fauna from Manning facies rocks along the Peel-Manning fault system, Glenrock Station, New England orogen. Proc. Linn. Soc. N.S.W. 115: 239-246 (1995). A fragmentary macrofossil assemblage consisting of Eurydesma, Keenia and Peruvispirawas collected from Manning facies sediments in the Peel-Manning Fault System at Glenrock Station. The sedimentary succession containing the faunal assemblage has been variably tectonised, but indicate Lower Permian, Fauna II affinities. Volcaniclastic sediments in which the fossils are found are similar to the adjacent rocks of the Devonian Gamilaroi terrane and the presence of the Lower Permian sediments has revised previous mapping of the area, extending the width of Peel-Manning Fault System in this area. J-M.C. Stratford and J.C. Aitchison, Department of Geology and Geophysics, University of Sydney, New South Wales 2006; manuscript received 26 July 1994, accepted for publication 16 November 1994. KEYWORDS: Bivalve, gastropod, New England orogen, Permian. INTRODUCTION The New England orogen in northeastern New South Wales is an amalgum of mid to late Paleozoic terranes. Paleontological evidence provides age constraints vital to understanding the sequential development and relative depositional settings of different rock units which have later been juxtaposed along faults within tectonically complex areas. Furthermore this evidence provides constraints which enable definition of the tim- ing of relationships between adjacent terranes in orogenic collages (Coney et al., 1980). Strata within sedimentary basins which developed in response to transtension and transpression during a period of highly oblique plate convergence in the Early Permian record a history of terrane displacement along strike-slip faults within the New England orogen (Aitchison and Flood, 1992). These sediments are typically assigned to the Manning Group (Mayer, 1972). Numerous discrete sedimentary basins, each with a simi- lar sedimentary facies, appear to have developed along major strike-slip faults in New England during the Early Permian. Sediments in many of these basins were never con- tiguous with similar facies in adjacent basins, thus in a pure lithostratigraphic sense it is inappropriate to refer to these rocks as a single orogen-wide ‘Manning Group’. The term ‘Manning facies’ is more appropriate. Diamictites deposited out of high-density mass flows are the characteristic unit of the Manning facies sedimentary rocks, these vary from basin to basin, both in terms of up-section trends in clast composition and the clast-matrix relationships (Aitchison and Landis, manuscript in preparation). Glenrock Station is located 90 km north of Scone on the Ellerston-Barry Road in the New England region of northeastern NSW. Three distinctly different tectonostrati- graphic terranes, the Weraerai, Gamilaroi and Djungati terranes (Flood and Aitchison, 1992), and a younger sedimentary assemblage derived from these terranes are present in Proc. LINN. SOC. N.S.W., 115, 1995 240 LOWER PERMIAN FAUNA FROM NEW ENGLAND Peel Manning Fault Zone Figure 2 Tertiary Volcanics Undifferentiated Peel -Manning Fault Zone rocks Djungati terrane Siluro-Devonian Weraerai terrane EarlyCambrian 151°25' Southern New England Orogen 151°28' Manning Facies Permian Gamilaroi terrane Devonian Gamilaroi terrane limestone Emsian-Eifelian Gamilaroi terrane felsic volcanics Devonian Fig. 1. Simplified geological map of Glenrock Station showing the distribution of the different rock units and the position of the fossil locality map Fig. 2. Proc. LINN. SOC. N.S.W., 115, 1995 J. STRATFORD AND J. AITCHISON 241 the Glenrock district (Fig. 1). The Gamilaroi terrane is a Devonian intra-oceanic island arc assemblage of volcaniclastic sediments, tuffaceous mudstones intercalated with dolerite and gabbro sills, pillow basalts and felsic tuffs (Aitchison and Flood, 1994). Distinctive red cherts of Silurian to Upper Devonian age, minor basalts and volcaniclastic sediments repeated by imbricate thrust faults reminiscent of an subduction complex form the Djungati terrane (Aitchison et al., 1992a). The Djungati terrane is separated from the Gamilaroi terrane in the southwest by strands of the Peel-Manning Fault System (PMFS). The PMFS is a fundamental structural discontinuity within the southern New England orogen. Rocks associated with the Weraerai terrane, a dismembered Early Cambrian ophiolite (Aitchison et al., 1992b; Aitchison et al., 1994; Aitchison and Ireland, 1994) are commonly associated with the PMFS; these have undergone widespread serpentinisation and structural dismemberment during its emplacement at high structural levels. The PMEFS is imbricate at Glenrock station. The main trace of the fault at Glenrock is marked by a km-wide zone of tectonic melange with various blocks, some of mappable size, enveloped in serpentinite and mudstone matrix. Rapidly-deposited Lower Permian marine sedimentary rocks of the Manning facies are common along the PMEFS and it is from a zone of these rocks that we report a new discovery of fossils. The fossils occur in poorly-sorted volcaniclastic conglomeratic sandstones. The main component of these sandstones are clasts of dolerite up to 12 mm long which are set in a framework of medium sand-sized plagioclase and pyroxene grains derived from the same doleritic source as the clasts. The dolerites are most probably sourced from the adja- cent Gamilaroi terrane. Similar volcaniclastic sandstones also occur within the Gamilaroi terrane sediments indicating that Devonian sedimentation was partially autocannibalistic and making discrimination between unfossiliferous Gamilaroi terrane and Manning facies rocks difficult. Elsewhere (in the Barry and Pigna Barney districts adjacent to Glenrock) altered serpentinite clasts in the sediments are found within coarse grained Manning facies rocks, indicating that some detritus in these sediments originated from the Weraerai terrane. Red chert is a common clast lithology within Manning facies sedi- mentary rocks (Aitchison and Flood, 1992) elsewhere in the New England orogen, includ- ing Glenrock, though it is absent from sedimentary rocks near the fossil site. This suggests that the Djungati terrane was nota significant source in this section of the basin at the time the fossiliferous sediments were deposited. Due to their appreciable volcanic component, these coarse volcaniclastic sand- stones were originally mapped as part of the Gamilaroi terrane. Identification of the Permian fauna has extended the previously mapped distribution of the Manning facies at Glenrock, increasing the recognised width of the PMFS and demonstrating the impor- tance of fossil age control when mapping highly dismembered fault zones. PALEONTOLOGY Material was collected from one site (GR 471950 on the Glenrock 9134-IS, 1:25 000 topo sheet; Fig. 2). The fossils occur as casts in extremely hard blue-grey conglomeratic sandstone. Shell material is usually incomplete and has largely been removed by weather- ing, though occasionally the spiral internal moulds of the gastropods remain (Fig. 3G). The sediment which infills some of the gastropods differs from the surrounding matrix (Fig. 3H) indicating that some of this material is redeposited. This interpretation might also explain the predominance of single Ewrydesma valves which were disarticulated at the time of deposition. The remaining shell material was dissolved in dilute HC1 before latex casts of the fossils were made. Most specimens have been tectonically altered to some degree and are often quite distorted. Tectonic distortion of the fossils means that considerable care had to be taken with identifications, which were only possible to genus levels. Proc. LINN. SOC. N.S.W., 115, 1995 242 LOWER PERMIAN FAUNA FROM NEW ENGLAND KEY Manning Facies Gamilaroi terrane Gamilaroi terrane limestone Djungati terrane Weraerai terrane Undifferentiated fault zone rocks Fig. 2. Geological map indicating the position of the Permian fossil locality within the Peel-Manning Fault System. All specimens are catalogued with the Australian Museum (AMF). Class PELECYPODA Family EURYDESMIDAE Reed, 1932 Eurydesma sp. indet. (Fig 3 A-C) Type species. Eurydesma cordata Morris, 1845 Diagnosis. see Runnegar (1970a) Discussion The material is fragmentary. Shells are equivalved, sub-circular to slightly elongate oval in shape. All specimens are relatively small (Table 1), the maximum dimension to close to 45 mm in height. Original calcareous shell material is rare. Shell thickness is 1 and 2mm, the valves are moderately inflated. Umbones are low and broad, centred on circular specimens, pointing slightly towards the anterior on more elongate shaped shells. Beaks are inrolled, the left valve apparently more so than right valve. Ligament areas are up to 7 mm in width on posterior Proc. LINN. SOC. N.S.W., 115, 1995 J. STRATFORD AND J. AITCHISON 243 H Fig. 3. Fossils from the Lower Permian Manning facies at Glenrock Station. A-C, Eurydesma sp. A, Latex cast of byssal notch right valve, AMF 92620, x1.5. B, Latex coast of dorsal view, AMF 92619, x1.3. C, Latex cast of lateral view of left valve, AMF 92617, x2. D, E. Keeniasp. D, Latex cast of apical view, AMF 92627, x1.5. E, Latex cast of apertural view, AMF 92621, x1.7. F, Peruvispirasp., AMF 92628, x5. G, Natural internal mould of Keenia sp. AMF 92625, x2.8. H, Transverse section through Keeniasp. showing different, lighter coloured sediment infilling, AMF 92626, x1.5. ventral margin. Dental process are well developed with a byssal notch present on right valve. Muscle scars are not observed. The enlarged dental process present in the right valve, characteristic umbones, liga- ment area and thick shell are all characteristic of Ewrydesma. However, the actual species is difficult to determine due to the small size of the specimens. The moderate inflation is most similar to E. hobartense, though tectonic compression is the most likely reason for the elongate dimensions of some specimens and it may also be responsible for the apparent lack of inflation. Eurydesma hobartense ranges from the Allandale Fauna through to Fauna III (Runnegar, 1970b). Proc. LINN. SOC. N.S.W., 115, 1995 244 LOWER PERMIAN FAUNA FROM NEW ENGLAND TABLE 1 Measurements forEurydesma sp. from the Manning facies sedimentary rocks at Glenrock. Eurydesma Length (mm) Height (mm) Thickness (mm) AMF 92614 37 36 13 AMF 92615 35 28 10 AMF 92616 31 37 9 AMF 92617 25 min. 24 7 AMF 92618 23 25 5 min. TABLE 2 Measurements for Keenia sp. from the Manning facies sedimentary rocks at Glenrock. Keenia Height (mm) Width (mm) Apical angle (°) AMF 92621 19 21 99 AMF 92622 20 22 127 AMF 92623 9 13 118 AMF 92624 19 15 112 Class GASTROPODA Family EUPHALIDAE De Konick Keenia sp. indet. (Fig. 3D, E) Type species. Keenia platyschismoides Etheridge Jr., 1902 Diagnosis. see Dickins (1957) Discussion Large turbiniform shells with adpressed whorls and a flat spire (Table 2). Whorls are much wider than high and moderately convex on the upper surface, though the basal whorl has increased convexity. The aperture is ovate. Ornament is present only on lower part of the shell and consists of growth lines parallel to the edge of the aperture, becoming very prominent in the last 2 to 5mm approaching the outer lip. Comments on the shape and possible affinities of these forms must be tempered with the likelihood of tectonic distortion. The apical angle of these specimens varies from 115-125° which is most similar to K. canarvonensis (Dickens, 1957), espeically when com- pared to the lower 95-105° apical angles of K. platyschismoides and K. ocula, or the higher 130-145° angles of the broader whorl forms such as K. minor (Fletcher, 1958). Family PLEUROTOMARIDAE D’ Orbigny Peruvispira sp. indet. (Fig. 3F) Type species. peruvispira delicata Chronic, 1949 Diagnosis. see Dickins (1961) Discussion Single very small specimen 4 mm in height and 3 mm wide. Tall turbinate spire with tightly coiled whorls and an apical angle approximately 60°. Traces of characteristic selin- zone barely visible on the last whorl. Proc. LINN. SOC. N.S.W., 115, 1995 J. STRATFORD AND J. AITCHISON 245 AGE Previous estimates for the Manning facies sedimentary rocks at Glenrock from fau- nal and floral remains indicate Upper Carboniferous to Lower Permian affinities (Offler, 1982). On the basis of the fauna reported herein, these rocks are considered to be Lower Permian, possibly Sakmarian. Manning facies faunas previously described from the New England region are usu- ally associated with Fauna IT (Runnegar, 1969), correlated with Fauna II of the Bowen Basin (Dickins et al., 1964). Locality 37 from north of Attunga (Runnegar, 1970b) is the only published locality of Manning facies rocks in which the faunal elements from Glenrock, Eurydesma, Peruvispiraand Keenza, are also reported together. Fauna II suggests Lower Permian affinites for the Glenrock strata, with a range, based on the presence of Eurydesma in eastern Australia, from upper Asselian [Allandale Fauna] to upper Artinskian [Ulladulla Fauna] (Runnegar, 1970b). Manning facies sedimentary rocks from the nearby Pigna Barney region, south of the Glenrock Permian rocks, contain a brachipod-bryozan dominated fauna of Tastubian (lower-middle Sakmarian) age (Dongal, 1994). Close spatial correlation between these localities supports assignment of the Glenrock asemblage to fauna II (Sakmarian). ACKNOWLEDGMENTS J.M. Dickens from AGSO provided invaluable assistance with identification of the fossils. The manager B. MacNaughton, owners and staff at Glenrock Station are thanked for their hospitality. JCA acknowledges field support from ARC, Thanks also to P. Flood for his helpful review. References AITCHISON, J.C. and FLoop, P.G., 1992. — Early Permian transform margin development of southern New England orogen, eastern Australia (eastern Gondwana). Tectonics, 11: 1385-1391. AITCHISON, J.C., FLoop, P.G., and SPILLER, F.C.P., 1992a. — Tectonic settting and paleoenvironment of terranes in the southern New England orogen as constrained by radiolarian biostratigraphy. Palaeogeography Palaeoclimatology and Palaeoecology, 93: 31-54. AITCHISON, J.C., IRELAND, T.R., BLAKE, M.C.Jr., and FLOop, P.G., 1992b. — 530 Ma zircon age for ophiolite from the New England orogen: Oldest rocks known from eastern Australia. Geology, 20: 125-128. AITCHISON, J.C. and FLoop, P.G., 1994. — Gamilaroi Terrane: a rifted Devonian intra-oceanic island arc assem- blage, NSW, Australia. Jn: Smellie, J. (ed). Volcanism associated with extension at consuming plate margins., Geological Society of London Special Publication 81: 155-168. AITCHISON, J.C. and IRELAND, T.R., 1994. — Age profile of ophiolitic rocks across the Late Paleozoic New England Orogen, NSW, Australia: Implications for tectonic models. Australian Journal of Earth Sciences, 41: 11-23. AITCHISON, J.C., BLAKE, M.C. Jr., FLoop, P.G. and JAyKo, A.S., 1994. — Comparison of ophiolitic rocks from the New England Orogen of eastern Australia. Tectonics. 13: 1135-1149. CONneEY, P.J., JONES, D.L. and MONGER, J.W.H., 1980. — Cordilleran suspect terranes. Nature, 288: 329-333. DICKINS, J.M., 1957. — Lower Permian pelecypods and gastropods from the Carnarvon Basin, Western Australia. Bulletin of the Bureau of Mineral Resources, Geology and Geophysics Australia, 41: 1-55. DICKINS, J.M., 1961. — Eurydesma and Peruvispira from the Dwyka Beds of South Africa. Palaeontology, 4: 138-148. DICKINS, J.M., MALONE, E.J. and JENSEN, A.R., 1964. — Subdivision and correlation of the Permian Middle Bowen Beds, Queensland. Report of the Bureau of Mineral Resources, Geology and Geophysics Australia, 70. DONGAL, G., 1994. — Age-diagnostic Early Devonian (Pragian and early Emsian) conodonts and other fauna from the south-eastern end of the Tamworth terrane. Abstracts © Programme, Australasian Paleontological Convention-94, Sydney: p.26. FLETCHER, H.O., 1958. — The Permian gastropods of New South Wales. Records of the Australian Museum, 24: 115- 164. FLOOD, P.G. and AITCHISON, J.C., 1988. — Tectonostratigraphic terranes of the southern part of the New England orogen, eastern Australia. In: Kleeman, J.D. (ed). New England orogen: Tectonics and Metallogenesis, 7-10, University of New England, Australia. Mayer, W., 1972. — Palaeozoic sedimentary rocks from southern New England: a sedimentological evaluation. (unpubl.) Ph.D thesis, University of New England. OFFLER, R., 1982. — Geochemistry and tectonic setting of igneous rocks in the Glenrock Station area, N.S.W. Journal of the Geological Society of Australia, 29: 443-455. RUNNEGAR, B.N., 1969. — The Permian faunal succession in eastern Australia. Special Publication of the Geological Society of Australia, 2: 73-98. Proc. LINN. SOC. N.S.W., 115, 1995 246 LOWER PERMIAN FAUNA FROM NEW ENGLAND RUNNEGAR, B.N., 1970a. — Eurydesma and Glendella gen. nov. (Bivalvia) in the Permian of eastern Australia. Bulletin of the Bureau of Mineral Resources, Geology and Geophysics Australia, 116: 83-118. RUNNEGAR, B.N., 1970b. — The Permian faunas of northern New South Wales and the connection between the Sydney and Bowen Basins. Journal of the Geological Society of Australia, 16: 697-710. Proc. LINN. SOC. N.S.W., 115, 1995 Diversity, Distribution and Conservation of Freshwater Crayfishes in the Eastern Highlands of New South Wales JOHN R. MERRICK MERRICK, J.R. Diversity, distribution and conservation of freshwater crayfishes in the east- ern highlands of New South Wales. Proc. Linn. Soc. N.S.W. 115: 247-258 (1995). The biology of most of the indigenous parastacid crayfishes inhabiting the high- lands of New South Wales is poorly known, but many species have very limited ranges. Analysis of these distributions in relation to the National Parks and State Forests shows that most species are protected in reserves; however, around and within these areas a large number of potential polluted sites have been identified. The effects of clearing, longstanding salmonid stocking and the potential problems of recent widespread introductions of non-indigenous Cherax species for aquaculture are discussed. Recommendations for conservation and future management include: biological programmes to provide data on environmental preferences, interactions between indige- nous and non-indigenous crays and influences of introduced salmonids; restoration and maintenance of riparian zones; surveys of polluted sites in, or adjacent to, very small ranges with initiation of remedial measures where necessary; more active policing of quarantine measures at aquaculture facilities; development of comprehensive, workable and enforce- able policy on translocations; implementation of eel control measures at large impound- ments on eastern drainages. John R. Merrick, Graduate School of the Environment, Macquarie University, North Ryde, Australia 2109; manuscript received 16 August 1994, accepted for publication 15 February 1995. KEY WORDS: Conservation, Australia, freshwater crayfishes, highlands, Parastacidae, aquaculture, Cherax, Engaeus, Euastacus, interactions. INTRODUCTION The Australian cray fauna is second only to North America in numbers of species and diversity; taxonomic investigations are continuing but approximately 100 species, assigned to nine genera, are currently recognised. The majority of these endemics is con- centrated in the south-eastern part of the continent — especially in highland areas — and the biology of most species is very poorly known (Merrick, 1993: 5, 35-36). New South Wales Crayfish Fauna In New South Wales the indigenous crayfish fauna comprises at least 27 species — two in the genus Engaeus, four in Cherax and 21 in Ewastacus; further new species and sub- species of this latter group have been recognised but formal descriptions have not yet been published. In addition, three other Cherax species have been widely introduced for aquaculture (Merrick, 1993: 39). Table 1 below lists all freshwater crayfishes currently recognised in this state together with remarks on known ranges. Clearly the New South Wales fauna is dominated by Ewastacus species which almost all inhabit highlands or streams near elevated areas; other Euastacus species dominate south-eastern Queensland and Victorian highlands. After revisionary studies this genus of spiny crays has become the largest in the family Parastacidae (Morgan, 1986, 1988, 1989, 1991). Hence this paper will be primarily considering the conservation of Euastacus species; the investigation is focussed on New South Wales but similar situations exist else- where. Several additional environmental modifications are known to affect the Cherax species inhabiting coastal lowlands (Merrick, 1993: 94) but will not be considered further here. Proc. LINN. SOC. N.S.W., 115, 1995 248 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS TABLE 1. Freshwater crayfishes indigenous to New South Wales (based on Merrick, 1993: 40-82). Genus, Species * Cherax cuspidatus C. destructor C. neopunctatus C. rotundus # Engaeus cymus Distribution, Remarks Extended range, mostly coastal lowlands Wide inland range, extends to foothills Extended coastal lowland range Extensive range, includes both coastal plain and highlands Extensive range, to altitudes >1,000 m E. orientalis Extensive range, variety of highland habitats Euastacus alienus Very restricted range (upper Karuah R. system) E. aquilus Very restricted range (east of Armidale) E. armatus Extensive range, mostly inland lowlands but to >700 m E. australasiensis Limited range (Illawarra, Blue Mtns, Gosford districts) E. bidawalus Limited range (150-400 m altitude) E. brachythorax Limited range (Tuross R. basin) E. claytoni Extensive range E. clydensis Limited range including coastal lowlands E. crassus Extensive range E. hirsutus Very limited range (Illawarra coastal streams) E. keirensis Very restricted range (Mt Keira) E. neohirsutus Very restricted range (inland of Coffs Harbour) E. nobilis Very limited range (Parramatta, Hawkesbury R. systems) E. polysetosus Restricted range (upper Hunter, Manning R. systems) E. reductus Very restricted range (upper Hunter R. system) E. simplex Restricted range (upper Macleay, Clarence R. systems) E. spinifer Limited range (Blue Mtns, Illawarra) E. spinosus Restricted range (Hastings, Camden-Haven R. systems) E. sulcatus Limited range, altitudes >300 m E. suttoni Extensive range, altitudes >680 m E. valentulus Extended coastal range, lowlands to 600 m altitude * Three additional non-indigenous Cherax species (C. albidus, C. quadricarinatus, C. tenuimanus) have been widely introduced for commercial culture. # These are land crayfishes but occur in highland areas. Natural Ranges, Reserves, Aquaculture and Pollution Many Euastacus species have very small, discrete distributions and a number of long- term, environmental factors that have contributed to this zoogeographic pattern are dis- cussed by Dodson et al., (1992) and Ross et al., (1992). For a description of the wild highland aquatic habitat see Merrick and Schmida (1984: 15, 17). These small, remote ranges make these crayfishes both susceptible to local disasters and difficult to monitor (Merrick, 1993: 15). Although the eastern highlands are where the majority of National Parks and State Forests are sited, the substantial development and environmental modification in eastern New South Wales has only been partially bal- anced by the declaration of new reserve areas. These crays, which are generally slow to mature with low fecundities, are in the same areas where trouts have been cultured and stocked for many years; these salmonids are limited to the highlands by thermal tolerances. In addition there has been considerable expansion of other types of freshwater aquaculture in recent years (Merrick, 1992). A recent review shows that 105 farms are registered to produce 10 species of freshwater cray- fish in all except the far north-western river basins (e.g. Paroo River); many of these facili- ties are producing non-indigenous crays — either to New South Wales or the particular area of the state. Most facilities are very small with low production and part-time manage- ment. About 20 farms, concentrated in the south-east, produce four species of salmonid (Merrick, 1992). Recognition of the degradation of Australian freshwater environments has coincid- ed with the political acknowledgement of the importance of environmental issues — in particular the maintenance of biodiversity and ecologically sustainable development (Ecologically Sustainable Development Working Groups, 1991). Problems in freshwaters Proc. LINN. SOC. N.S.W., 115, 1995 J.R. MERRICK 249 have been developing for many years (Lake, 1978) but the recent widespread blue-green algal blooms have elevated the issues into a high priority category (Creagh, 1992; Hart, 1992). Anumber of water quality problems have been recognised throughout New South Wales (Water Resources Commission, 1986) and many potentially polluted sites identi- fied; some 60,000 sites are considered to require investigation and 7,000 of these possibly require remedial action (Grant, 1992). Biology and Management Although not as well known as vertebrate groups it has become increasingly clear that crayfishes are of prime importance in many aquatic ecosystems, in terms of biomass and energy transfer (Hogger, 1988). Fortunately, there are no records of introductions of exotic (non-Australian) crayfishes (Thomson ¢é¢ al., 1987; Holdich, 1993) and, although no endemics are considered endangered at present a number of Ewastacus species have been classified as threatened (Horwitz, 1990). There are few data on the ecology of rela- tively undisturbed wild populations, but even less is known about the interactions of native crays with non-indigenous crays or other introductions such as fishes (Courtenay, 1990; Merrick, 1992). As with other environments, management decisions relating to aquatic habitats have often been taken on the basis of the better known vertebrates — usually fishes. Recent research has provided some information on habitat preferences and require- ments of at least the larger or dominant fish species (Mallen-Cooper, 1989); but there are still no data on the relative importance of snags (submerged logs, fallen branches or roots protruding from under banks) or other environmental factors for invertebrates. The objectives of this paper are: to compare known ranges of native crays in New South Wales highlands with the distributions of reserves and registered aquaculture facili- ties; to discuss these patterns, generally, in relation to known polluted sites and water qual- ity problems; to summarise the data on apparent habitat preferences; to identify high priorities for immediate research and discuss potential problems; to combine existing biological and environmental data in the formulation of recommendations for the con- servation of the crayfish fauna. MATERIALS AND METHODS Natural Ranges, Reserves, Aquaculture and Pollution Species ranges are illustrated together with the extent of National Parks, State Forests and other reserves as well as the distribution of culture facilities (Figures 1, 2). Key data on aquaculture facilities are summarised (Tables 2, 3) and comments on polluted sites are also included. Biology and Management Aspects of the biology and environmental requirements of native crays particularly relevant to conservation management are summarised. Management problems and high priority areas for urgent attention are identified and options considered. Finally detailed recommendations, both for immediate and progressive long-term implementation, are formulated. RESULTS Natural Ranges, Reserves, Aquaculture and Pollution Figures 1 and 2 demonstrate several general points. Firstly, the ranges of most cray Proc. LINN. SOC. N.S.W., 115, 1995 250 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS Seepiind National Parks, Nature and Timber Reserves, Hh | N wa and State Forests of eastern New South Wales | | | Glen Innes v_ > S Port Macquarie is Wauchope Stroud “ALS og ¥ Newcastle y¥ SYDNEY f YG ~ 2% Wollongong 100 200 kms Fig. 1. National Parks, Nature and Timber Reserves and State Forests of eastern New South Wales (stippled) with natural ranges of 21 endemic Euastacus species marked (outlined). Range outline around Sydney includes reported distributions for four species. Based on Forestry Commission of New South Wales (1991), Riek (1969) and Merrick (1993). Proc. LINN. SOC. N.S.W., 115, 1995 J.R. MERRICK 251 Lamington Plateau y/ \ i “@2 Stanthorpe ~. y Seager Brunswick MURRAY - DARLING DIVISION = —~ / % Heads Q la © e pi South-East Coast Division 29 \ . 718 ~ Ballina 1 Tweed River —Glenn : o 2 Brunswick River fe Innes A Thy 1) Ph 3 Richmond River Si ecea Oo ad 4 Clarence River sre Bees 5 Bellinger River / a : 2 Gratton 6 Macleay River S900 an 5 7 Hastings River “9 re 8 Manning River yee ee hoes s 9 Karuah River / ; Sent! ian ols Harbour 10 Hunter River | ccleM TE inon 11 Macquarie - Tuggerah Lakes i Armidale 3% 12 Hawkesbury River 6 cent 13. Sydney Coast - Georges River ‘eine eres : 14 Wollongong Coast f: 15 Shoalhaven River 27 3 @—e% — Pon Macquarie 16 Clyde River - Jervis Bay 3 s 17 Moruya River 33°. oe LY ~~ Wauchope 18 Tuross River 7. cata 8 ele! 19 Bega River se? / 20 Towamba River Barrington Tops Bier 21 East Gippsland / mesahes ge) 22 Snowy River A 10 sat sos gail enatd a e : ee - Murray-Darling plvislon fh 34 Cy aes Bauicieancn: 23 Upper Murray River oe 25 Murrumbidgee River . o> Newcastle 26 Lake George [ g He Aaa 27 Lachlan River he ~Newnes:i / SOUTH - EAST COAST DIVISION 29 Border Rivers assets Plateau 31 Gwydir River y, 32 Namoi River Oo . eesiord 33 Castlereagh River 12 @.." “Ge— SYDNEY 34 Macquarie - Bogan asta 5 *43 e 27 ont See e Oty ase *"[.— Wollongong i, a iste awe Kiama ff pa : 5 N Jervis Bay e * * 0 25 ce aN 23 ? Bombala. 3 % Salmonids culture 22 Dai iics tes NT et, 20 e Crayfish culture aide: 2 ee es ons Fig. 2. Drainage Divisions and Basins of eastern New South Wales with approximate localities of most aquacultural facilities marked (N.S.W. Fisheries, unpublished data). Some salmonid farms also produce crayfish. Key to facili- ties: (¢) crayfish culture; (*) salmonid culture. Proc. LINN. SOC. N.S.W., 115, 1995 252 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS species are included in, or at least partially overlap, reserves. Secondly, the majority of aquaculture facilities are sited on, or east of, the dividing range. Facilities in eastern drainage basins are clustered in three areas: the northernmost cluster extends from just south of the border to the Grafton area; the largest cluster extends from Bellingen south to Port Stephens; a third, smaller group extends from west of Sydney to inland of Kiama. Finally, water quality problems have been identified in a number of areas — including upper reaches. Tables 2 and 3 demonstrate that crayfish farms comprise the biggest single category of facility (about 40% of the total) — with another 25% producing some crays in combi- nation with other organisms. Over 50% of all cray farms are east of the Dividing Range (Figure 2) and a majority produce two or more species. While 49 farms produce C. destruc- tor only, another 41 culture some of this species. Redclaw (Cherax quadricarinatus) is pro- duced by 40 farms but only eight are solely dependent on this species. Of the 20 farms registered to produce marron only three culture it exclusively. The seven Euastacus farms are all within or adjacent to the natural range of the particular species. Trout and salmon farms comprise another significant category (about 14%) of aquaculture facilities. TABLE 2. Registered aquaculture facilities in New South Wales (1992) grouped on the basis of organisms cultured — oyster leases not included (based on N.S. W. Fisheries, unpublished data). Categories of Organisms Cultured No of Farms* Crayfishes 65 Crayfishes + Native Fishes 16 Crayfishes + Native Fishes + Salmonids 3 Crayfishes + Salmonids 3 Crayfishes + Native Fishes + Aquarium Fishes 7 Crayfishes + Native Fishes + Shrimps 2 Crayfishes + Native Fishes + Shrimps + Mussels 3 Crayfishes + Native Fishes + Mussels + Prawns 1 Crayfishes + Shrimps + Mussels 2 Crayfishes + Mussels 1 Crayfishes + Prawns 1 Crayfishes + Shrimps 1 Prawns 8 Prawns + Oysters 4 Prawns + Native Fishes 1 Crab 1 Native Fishes 17 Native Fishes + Salmonids 2 Salmonids 13 Native Fishes + Aquarium Fishes 3 Native Fishes + Aquarium Fishes + Salmonids 1 Aquarium Fishes 8 Total farm licences 163 * A large majority of facilities (126) have a hatchery capacity, but whether it is currently in use is unknown (Rayns, 1992, personal communication); 37 farms are grow-out areas only. Although many polluted sites have been identified, data on exact locations, or risk status, are not freely available. A state-wide register is being compiled by the Environment Protection Authority (EPA) of New South Wales and Local Governments have lists of known polluted sites in their areas — for reference in response to specific requests (McFarland, 1993, personal communication). Proc. LINN. SOC. N.S.W., 115, 1995 J.R. MERRICK 253 TABLE 3. New South Wales crayfish farms and trout farms (1992) grouped on the basis of species cultured (based on N.S. W.Fisheries, unpublished data). Combination of Species per Farm No of Farms* (a) Crayfishes Cherax destructor 49 Cherax destructor + C. quadricarinatus 16 Cherax destructor + C. quadnicarinatus + C. tenuimanus Cherax destructor + C. quadricarinatus + C. tenuimanus + C. albidus Cherax destructor + C. quadricarinatus + C. albidus Cherax destructor + C. quadricarinatus + C. albidus + C. rotundus Cherax destructor + C. quadricarinatus + C. rotundus Cherax destructor + C. quadricarinatus + C. rotundus +E. valentulus Cherax destructor + C. quadricarinatus + E. valentulus Cherax destructor + C. tenuimanus Cherax destructor + C. tenuimanus + C. albidus Cherax destructor + C. tenuimanus + E. armatus Cherax destructor + C. rotundus Cherax destructor + C. rotundus + C. cuspidatus Cherax destructor + E. armatus C. quadricarinatus C. quadricarinatus + C. tenuimanus C. quadricarinatus + Euastacus sp. C. quadricarinatus +E. spinifer C. tenuimanus C. albidus Total farms producing crayfishes LO a ce cn © OTE 9 a I © — S Or (b) Salmonids Oncorhynchus mykiss 15 Oncorhynchus mykiss + Salmo trutta 2 Oncorhynchus mykiss + Salmo salar Oncorhynchus mykiss + Salmo trutta + Salmo salar — Oncorhynchus mykiss + Salmo trutta + Salmo salar + Salvelinus fontinalis 3 Total farms producing salmonids 22 A detailed survey of all suspected polluted sites in the highlands was beyond resources for this study; however, preliminary investigations revealed the following: (a) (b) (c) (d) (e) register records are still incomplete — for example, EPA has no information on the southern highlands of the State; among identified sites the risk varies enormously — some cattle dip sites are considered low risk; the major pollutant from agricultural land is the metal Arsenic — wherever a tannery has been established there is residual pollution (usually Arsenic) (McFarland, 1993, personal communication); over 1,100 sites are currently registered as mines in New South Wales and this does not include many small workings, in the highlands, long since abandoned (Woodside, 1993, personal communication); comparisons of Ewastacus distributions with EPA records did not reveal any significant threats in most highland areas, but there are some polluted ‘hot spots’ in the E. australasiensis and E. spinifer ranges (McFarland, 1993, personal communication). Proc. LINN. SOC. N.S.W., 115, 1995 254 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS Biology and Management Although current data are scant a number of environmental requirements and life cycle features appear to be similar in a number of Ewastacus species; these criteria, directly relevant to management, are drawn from Merrick (1993) and summarised below. Highland areas known to maintain abundant endemic cray populations generally have the following characteristics. The waterway is shallow (depths 0.05 — 5.0m) with a rock, gravel and/or sand substrate; logs and leaf detritus are also present. Flow is usually continuous for most of the year, although rates may be low; dissolved Oxygen levels are high (>5 p.p.m.), pH 7.0 — 8.5, salinity <3,000 p.p.m. and turbidity low. Water tempera- tures range from 5° — 30°C and there is a low incidence of large, carnivorous fishes (Murray cod, trout cod, trouts, eels). Banks may be gently sloping or steep (undercut) and are shaded by native vegetation (most frequently rainforest or wet sclerophyll forest). In addition to low growth rates and fecundities other biological features which con- tribute to slow recruitment include: an annual breeding (not all mature females spawn each year); a long incubation interval (4-6 months, often at low temperatures); an extended larval period (up to 2 months, sometimes in high temperatures); very limited dispersal or movement by young at release and significant predation by larger individuals on juveniles. DISCUSSION Comments on the present studies and the recommendations are framed in the con- text of the following general observations. Firstly, the distribution patterns of crays in the eastern highlands are similar to those now being documented for other invertebrate groups, such as gastropods and oligochaetes (Dyne, 1991; Stanisic, 1994) and show strong correlations with the occurrence of rainforest (Dyne, 1991). Secondly, recent systematic studies on other invertebrate groups have revealed enormous biodiversity in environ- ments previously considered depauperate and a complexity of faunal interactions unsus- pected earlier (Campbell and Brown, 1994; Ponder, 1994). Thirdly, the system of existing and planned reserves is very limited and unlikely to be adequate to maintain current bio- diversity (Pressey and Griffith, 1992). Fourthly, that with a few exceptions, conservation on a species-by-species basis is not feasible; the emphasis has to be on the conservation and management of systems or biotic/habitat complexes (Saunders, 1993). Finally, that even in the absence of comprehensive baseline data, some conservation measures can and should be taken in the immediate future. National Ranges, Reserves, Aquaculture and Pollution Detailed analyses of individual species ranges with respect to reserves and other features are beyond the scope of this paper and it should be noted that the only Ewastacus species to have been translocated to any extent is E. armatus (Harris and Battaglene, 1989). This has the broadest of all Euastacus distributions and is not considered at the same degree of risk as species with very restricted ranges in eastern drainage basins. Whilst the overlap of many highland cray ranges with existing National Parks and State Forests is reassuring in some ways, it should be emphasised that these invertebrates can still be effected by factors outside the reserves. The most important threat to Euastacus species appears to be clearing — for dairying or forestry — with the attendant changes in water quality and eutrophication (Merrick, 1993: 93); however, agriculture and man- made pollutants are also emerging as significant problems. Despite the fact that most salmonid introductions have not succeeded (Clements, 1988: 289) the continued efforts of many Government, Acclimatisation Society and Club hatcheries — with repeated releases at many highland sites over a century or more — must have had considerable biotic impacts (Merrick and Rimmer, 1984; Clements, 1988: Proc. LINN. SOC. N.S.W., 115, 1995 J.R. MERRICK 255 278-289). There is now evidence that predation by trout can have local negative effects on small native fish species (Tilzey, 1976). Brown and rainbow trout quickly established in many areas, with the latter species being produced in the largest numbers in recent years (Clements, 1988: 138-140); over 20 farms are sited close to highland cray ranges. Crayfish culture has expanded significantly in the south-east in recent years but the very features making Cherax destructorand C. quadricarinatus attractive to culture, such as frequent breeding and rapid growth, make them a potential threat. Overseas experience has repeatedly demonstrated displacement of native species by introductions that are aggressive, physiologically tolerant, grow quickly and breed frequently (Hogger, 1988; Lowery and Holdich, 1988; Holdich and Rogers, 1992). There are a few observations on the effects of the introduction of Cherax albidus to south-western Western Australia (Austin, 1985); but no studies have been published on the widespread introduction in eastern Australia of Cherax destructor, C. quadricarinatus or C. tenuimanus. The various potential environmental problems associated with translocations and aquaculture have been well documented elsewhere (Courtenay, 1990; Merrick, 1992), but the disturbing feature of this situation is the possibility that cray culture and the presence of trouts could act synergistically, with other factors, to contribute to the decline of isolated, indigenous populations. Over 80 facilities are sited adjacent to highland cray ranges. Although less modified than most areas National Parks and State Forests could still contain possibly point sources of pollution and are subject to the effects of previous mis- management elsewhere in the catchment. The implementation of Total Catchment Management Plans should, in future, minimise problems but old abandoned mine-sites with tailings or slag heaps are a continuing cause for concern. They are often close to streams and heavy rain can cause sudden leaching of high concentrations of metal, or other inorganic ore-processing chemicals, into the waterway. The numerous, long-term effects of mining on aquatic resources, together with the widespread impacts from indi- vidual mines, have been well documented (American Fisheries Society Water Quality Section Committee, 1988). Standards required for water quality parameters should be based on the recent recommendations of the Australian and New Zealand Environment and Conservation Council (1992). Although major parts of the natural ranges of E. australasiensisand E. spiniferare now urban and industrial areas associated with Sydney, with all the attendant pollution prob- lems, reduced populations of both species should persist. Fortunately, these Ewastacus are present in long-established National Parks ringing the metropolitan area and many popu- lations inhabit small, discrete, local catchments; these habitats are buffered, to some extent, from the environmental modification occurring in adjacent drainages. Biology Biological knowledge of all highland crays is totally inadequate for sustainable man- agement; baseline date on all aspects of life cycles are required. Interspecific interactions may be very important, although Tasmanian studies suggest that sympatric indigenous species segregate within the same habitat and have relatively little contact (Richardson and Swain, 1980). Another aspect of interspecific interaction which needs attention is that of preda- tion. Measures for controlling eel populations may be necessary. With increasing impoundment of eastern streams large new habitats are being created that are favourable for eels — especially the large, long-finned eel (Anguilla reinhardtii). Eels are particularly effective predators on crays (Cukerzis, 1988); for a description of a local situation where eels have become a dominant predator in impoundments, with a summary of active and passive population control options, see the study of Merrick and Rimmer (1984). Preferred environments for these crays can only be maintained if bank areas remain intact. In many cases these zones are rainforests and recent research has shown that this community type is very badly effected by fire (Friederich, 1991). Proc. LINN. SOC. N.S.W., 115, 1995 256 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS Management Recommendations The initiatives suggested below are designed: to be implemented (where appropri- ate) both within reserve boundaries and elsewhere; to provide further time in which to concentrate on identified priority research areas; to complement and supplement exist- ing initiatives, such as Landcare, Fishcare or National Parks and Wildlife Service Biodiversity Conservation Strategies. The order of listing does not indicate relative impor- tance or imply a necessary sequence — all measures can be concurrent. 1. Biological research programmes should be initiated immediately on all crays 10. native to the highlands. These studies will have three general aims: (a) to provide baseline biological data; (b) investigate habitat preferences and interactions between native and non-indigenous crays; (c) examine interactions between native crays and the introduced salmonids. . Where possible restoration of aquatic habitat should commence; further de- snagging, channelisation or impoundment of headwaters should not be permitted. Where an existing impoundment structure is redundant or unsound its removal should be considered. . Surveys of water quality and potential polluted sites — particularly in or adjacent to small ranges — are needed. Where problems are detected measures should then be initiated to reduce them to within recommended limits. . Prohibit stocking or any non-indigenous aquatic species to headwaters where an endemic cray with a limited range is known to occur; where stockings of mobile non-indigenous organisms, such as trouts, are required downstream any re- stocking above natural barriers should cease. . Restoration of cleared or damaged stream banks should commence immedi- ately; riparian strips of natural vegetation should be maintained (at least 100m wide). There are several aspects to this: i controlling dieback where it occurs in existing riparian stands; ii replanting with natives and continued weeding of exotics; where rainforest is involved an additional buffer zone of sclerophyll timbers may be necessary to minimise edge affects; iii eliminating access by livestock to riparian zones and watercourses; this will require fencing in some areas; iv _ active control of feral organisms — especially large animals such as horses, pigs and goats; Vv monitoring and effective protection of riparian zones from bushfires. Strict controls on all recreational activities are required and all fisheries for larger species should be converted to a sport category with appropriate regula- tions — such as those applied now to the Murray crayfish (Euastacus armatus). These regulations may have to vary between different areas to suit particular species or populations. . Further culture of non-indigenous cray species should be discouraged; and a more effective and comprehensive translocation policy developed. . Frequent inspections of all aquaculture facilities should be mandatory — checks on disease occurrence and control procedures, waster-water treatment and security measures such as screening or fencing should be especially rigorous. . Passive and active eel control measures should be instituted on large impound- ments on eastern drainages. Using a combination of watercourse and riparian features (physicochemical parameters, habitat diversity, food resources, types and occurrence of preda- tors) as well as other biological data (physiological tolerances, life cycle) a predictive system should be developed. This assessment would be designed to provide a quantitative value indicating the relative suitability of the particular habitat for stocking, or re-stocking, with a particular crayfish. Proc. LINN. SOC. N.S.W., 115, 1995 J.R. MERRICK 257 These initiatives are long-term, will require continued active input and/or monitor- ing and may involve a re-allocation of existing resources. Although specifically formulated to ensure the survival of crayfish stocks, the total programme involves the maintenance of entire watercourses and bank zones together with the biotic assemblage inhabiting those areas. Inaction or delay could result in irreplaceable losses. ACKNOWLEDGMENTS Dr N. Rayns (formerly New South Wales Fisheries, Sydney), Mr R. McFarland, Environment Protection Authority, New South Wales, and Dr D. Woodside, Zoological Parks Board of New South Wales are thanked for providing unpublished data. Appreciation is also expressed to Miss P.R. Davies, Graduate School of the Environment, and Mr J. Cleasby, School of Earth Sciences, Macquarie University for assis- tance with preparation of the manuscript and illustrations respectively. References AMERICAN FISHERIES SOCIETY WATER QUALITY SECTION COMMITTEE. 1988. — Effects of surface mining on aquatic resources in North America. Fisheries 13: 19-22. AUSTIN, C.M. 1985. — Introduction of the yabbie, Cherax destructor (Decapoda: Parastacidae) into southwestern Australia. Western Australian Naturalist 16: 78-82. AUSTRALIAN AND NEW ZEALAND ENVIRONMENT AND CONSERVATION COUNCIL. 1992. — Australian Water Quality Guildelines for Fresh and Marine Waters. A.N.Z.E.C.C., Canberra. CAMPBELL, A.J. and BROowN, G.R. 1994. — Distribution of parasitoids of scarab larvae in relation to remnant vegeta- tion: a preliminary analysis. Memoirs of the Queensland Museum 36: 27-32. CLEMENTS, J. 1988. — Salmon at the Antipodes. A History and Review of Trout, Salmon and Char and Introduced Coarse Fish in Australasia. J. Clements, Ballarat. CourTENay, W.R. Jnr. 1990. — Fish introductions and translocations and their impacts in Australia. Jn: POLLARD, D.A. (ed.) Introduced and Translocated Fishes and their Ecological Effects. Pp. 171-179. Australian Society for Fish Biology Workshop: Bureau of Rural Resources Proceeding 8. CrEAGH, C. 1992. — What can be done about toxic algal blooms? Ecos 72: 14-19. CUKERZIS, J.M. 1988. — Astacus astacus in Europe. Jn: HOLDICH, D.M. and Lowery, R.S. (eds). Freshwater Crayfish: Biology, Management and Exploitation. Pp. 309-340. Croom Helm Ltd., London. DODSON, J., FULLAGAR, R. and HEaD, L. 1992. — Dynamics of environment and people in the forested crescents of temperate Australia. Jn: DODsON, J. (ed.) The Naive Lands. Pp. 115-159. Longman Cheshire, Melbourne. Dyne, G.R. 1991. — Earthworm fauna of Australian rainforests. In: WERREN, G. and KERSHAW, P. (eds). The Rainforest Legacy. Australian National Rainforests Study. Volume 2. Pp. 335-343. Australian Heritage Commission, Special Australian Heritage Publication Series 7. ECOLOGICALLY SUSTAINABLE DEVELOPMENT WORKING Groups. 1991. — Final Report — Fisheries. Australian Government Publishing Service, Canberra. FORESTRY COMMISSION OF NEW SOUTH WALES. 1991. — New South Wales State Forests and National Parks. Map 1: 2,000,000. Second Edition. Forestry Commission of N.S.W., Sydney. FRIEDERICH, R. 1991. — Management of rainforest in national parks and equivalent reserves in northern New South Wales. In: WERREN, G. and KERSHAW, P. (eds). The Rainforest Legacy. Australian National Rainforests Study. Volume 3. Pp. 216-230. Australian Heritage Commission, Special Australian Heritage Publication Series 7. GRANT, C. 1992. Insights from experience in the management of contaminated sites in New South Wales. Fulbright Symposium Paper, Contaminated Sites in Australia: Challenges for Law and Public Policy, University of New South Wales, Sydney — 28-29 August. Harris, J.H. and BaTTAGLENE, S. 1989. — The translocation of freshwater fishes in south-eastern Australia. Australian Society for Fish Biology Pre-Conference Workshop, Introduced and Translocated Fishes and their Ecological Effects. (abstract only). Hart, B.T. 1992. — Ecological condition of Australia’s rivers. Search 23: 33-37. Hoccer, J.B. 1988. — Ecology, population biology and behaviour. Jn: HOLDICH, D.M. and Lowery, R.S. (eds). Freshwater Crayfish: Biology, Management and Exploitation. Pp. 114-144. Croom Helm Ltd., London. Ho .picu, D.M. 1993. — A review of astaciculture: freshwater crayfish farming. Aquatic Living Resources 6: 307-317. Ho pic, D.M. and Rocers, W.D. 1992. — Crayfish populations in the British Isles: farming, legislation, conserva- tion and management. Finnish Fisheries Research 14: 23-32. Horwitz, P. 1990. — The conservation status of Australian freshwater crustacea. Australian National Parks and Wildlife Service Report Series (14): 1-121. Lake, P.S. 1978. — On the conservation of rivers in Australia. Australian Society for Limnology Newsletter 16: 24-28. Lowery, R.S. and Ho.picu, D.M. 1988. — Pacifastacus leniusculusin North America and Europe, with details of the distribution of introduced and native crayfish species in Europe. Jn: HoLpICcH, D.M. and Lowery, R.S. (eds). Freshwater Crayfish: Biology, Management and Exploitation. Pp. 283-308. Croom Helm Ltd., London. Proc. LINN. SOC. N.S.W., 115, 1995 258 FRESHWATER CRAYFISHES IN EASTERN HIGHLANDS MALLEN-Cooper, M. 1989. — Fish passage in the Murray-Darling Basin. Jn: MURRAY-DARLING BASIN COMMISSION. Proceedings of the Workshop on Native Fish Management. Pp. 123-125. MDBC, Canberra. MERRICK, J.R. 1992. — Prospects for Aquaculture in New South Wales. Report 4 — Ecologically Sustainable Development. Total Environment Centre, Sydney. MERRICK, J.R. 1993. — Freshwater Crayfishes of New South Wales. Linnean Society of New South Wales, Sydney. MERRICK, J.R. and RIMMER, M.A. 1984. — Reservoir fishes and water quality monitoring systems. Baseline studies 1981-1983. Metropolitan Water Sewerage and Drainage Board Report. MERRICK, J.R. and SCHMIDA, G.E. 1984. — Australian Freshwater Fishes. Biology and Management. J.R. Merrick, Sydney. Morea, G.J. 1986. — Freshwater crayfish of the genus Ewastacus Clark (Decapoda, Parastacidae) from Victoria. Memoirs of the Museum of Victoria 47: 1-57. MorcGan, G.J. 1988. — Freshwater crayfish of the genus Euwastacus Clark (Decapoda, Parastacidae) from Queensland. Memoirs of the Museum of Victoria 49: 1-49. Morcean, G.jJ. 1989. — Two new species of the freshwater crayfish Euastacus Clark (Decapoda: Parastacidae) from isolated high country of Queensland. Memoirs of the Queensland Museum 27: 555-562. MoreGan, G.J. 1991. — The spiny freshwater crayfish of Queensland. Queensland Naturalist 31: 29-36. PONDER, W. 1994. — Austraian freshwater molluscs: conservation priorities and indicator species. Memoirs of the Queensland Museum 36: 191-196. PRESSEY, R.L. and GRIFFITH, S.J. 1992. — Vegetation of the coastal lowlands of Tweed Shire, northern New South Wales: plant communities, species and conservation. Proceedings of the Linnean Society of New South Wales 113: 203-243. RICHARDSON, A.M.M. and Swain, R. 1980. — Habitat requirements and distribution of Engaeus cisternarius and three sub-species of Parastacovdes tasmanicus (Decapoda: Parastacidae) , burrowing crayfish from an area of south-western Tasmania. Australian Journal of Marine and Freshwater Research 31: 475-484. RiEK, E.F. 1969. — The Australian freshwater crayfish (Crustacea: Decapoda: Parastacidae) with descriptions of new species. Australian Journal of Zoology 17: 855-918. Ross, A., DONNELLY, T. and Wasson, R. 1992. — The peopling of the arid zone: human-environment interactions. In: DoBsON, J. (ed.) The Naive Lands. Pp. 76-114. Longman Cheshire, Melbourne. Saunders, D. 1993. — Habitat reconstruction: the revegetation imperative — an animal perspective. Conserving Biodiversity. Threats and Solutions. 29th June — 2nd July. University of Sydney. Programme. (abstract only). STANISIC, J. 1994. — The distribution and patterns of species diversity of land snails in eastern Australia. Memoirs of the Queensland Museum 36: 207-214. THOMSON, J.M., LONG, J.L. and Horton, D.R. 1987. — Human exploitation of and introductions to the Australian fauna. In: DyNE, G.R. and WALTON, D.W. (eds). Fauna of Australia. Volume 1A. General Articles. Pp. 227-249. Australian Government Publishing Service, Canberra. Tizzey, R.D.J. 1976. — Observations on interactions between indigenous Galaxiidae and introduced Salmonidae in the Lake Eucumbene catchment, New South Wales. Australian Journal of Marine and Freshwater Research 27: 551-564. WATER RESOURCES COMMISSION, NEW SOUTH WALES. 1986. — Water and the Natural Resources of New South Wales. W.R.C.N.S.W., North Sydney. Proc. LINN. SOC. N.S.W., 115, 1995 Isonome Mapping: Graphic Analysis of Patterns of Species Distribution I. M. BREWER (NEE PIDGEON) BREWER, I.M., Isonome mapping: graphic analysis of patterns of species distribution. Proc. Soc. Linn. N.S.W. 115: 259-279 (1995). Fine scale analysis of species distribution and abundance is important in under- standing processes of vegetation change. A novel mapping technique based on simple quadrat-sampling was devised in 1941 for this purpose and used to describe and analyse the structure and patterns of species distribution in two vegetation types on Hawkesbury sand- stone of the Hornsby Plateau. In moist shrubland and in the understorey of adjacent Eucalyptus woodland at two sites, relative densities of shrubby species were calculated from the total numbers of plants of individual species recorded in rectangular (9.1 x 0.9 m) quadrats arranged in a grid. For each species with a sufficiently high density, the variation in its relative density across the grid was mapped as contour lines of equal percentage value called ‘isonomes’ (from 7so, equal, and nome, distribution). In isonome maps for individual species, mostly complex systems of isonomes with one or more centres of high relative den- sity emerged. By superimposing isonome maps of individual species, the composite pattern of species distribution over the area revealed a complex social structure in which the cen- tres of high relative density formed a mosaic, around the margins of which there was over- lapping of the lower-value isonomes. Graphic analysis by isonome mapping has provided information on sandstone vegetation not previously reported: e.g. sociology of woodland and scrub communities, patterns of occurrence and density of species across sharp eco- tones, and specific patterns have generated an hypothesis of temporal change operating at a small scale. In the application of isonome mapping techniques to other vegetation types, the investigator has to choose the appropriate size and spacing of the rectangular quadrats, so that variation in relative density of species across the grid will generate discernible pat- terns. This paper is of historic interest, not only as the first quantitative method devised to show the pattern of species distribution and abundance in a community, butalso as the first quantitative analysis of sandstone vegetation. It is also a record of species composition of pristine communities, devoid of introduced species, in urban fringe, pre-development veg- etation of Sydney. IM. Brewer, 18 Eastbourne Road, Darling Point, NSW 2027, Australia; manuscript received 15 November 1994, accepted for publication 15 February 1995. KEYWORDS: isonome mapping, species distribution, Hawkesbury sandstone vegetation. INTRODUCTION The twentieth century has seen the evolution of new methods and techniques for the classification, description and quantitative analysis of vegetation. The basis of quantitative investigations in plant sociology is the quadrat method, originally proposed for pastures by Stapledon (1912). Statistical methods of analysis gave information on the floristic composition of the community, or on density and type of distribution of species (random, over- and under-dispersion). With random dispersion, ecologists and agronomists had a valuable tool for studying changes in plant populations (Blackman, 1935; Ashby, 1935; Clapham, 1936). The arrival of Ashby in 1938, as Professor of Botany at Sydney University, and his fascination with the complexity of sandstone vegetation stimulated studies by Beadle in arid western NSW and that of Pidgeon on the sandstone vegetation. It also led to the collaboration with Pidgeon (now I.M. Brewer) in 1939 in a rigorous statistical analysis (using random-sampling techniques and rectangular strip-quadrats) of the effects of over-grazing on vegetation around Broken Hill (Pidgeon and Ashby, 1940). Proc. LINN. SOC. N.S.W., 115, 1995 260 ISONOME MAPPING OF SPECIES DISTRIBUTION Further investigations planned in applied ecology (see Pidgeon & Ashby, 1940) were relinquished when Ashby established and directed the Australian Liaison Bureau to maximize science in the war effort, and the Botany Department began its co-operation with the CSIR (later CSIRO) Food Preservation Research Laboratory which involved some of the teaching and research staff in solving problems related to fruit storage and marketing. The complex Hawkesbury sandstone vegetation had not been analysed by quantita- tive techniques prior to 1941, as until then there was no quantitative method whereby the structure or pattern of a community could be determined. In 1941, Ashby and Pidgeon devised a novel ‘isonome’ mapping technique, based on simple quadrat sampling in grids, to record patterns of distribution and abundance of species. It was used in 1941 by Pidgeon on Hawkesbury sandstone and sand dune vegetation. The technique was first outlined by Pidgeon and Ashby (1942), but the data have been recorded only in the D.Sc. thesis (Pidgeon, 1942), and have not been explored, expanded, or related to information since published on the flora of the central coast, NSW. From the 1950's, different approaches to describing and analysing vegetation were developed: objective methods (Goodall, 1961), and quantitative techniques (Grieg- Smith, 1983). Ordination (Bray & Curtis, 1957) was followed by computer-based methods of association analysis (Williams & Lambert, 1959, and others) and multivariate analysis (Gauch, 1982; James & McCulloch, 1990). These powerful and sophisticated techniques are now used almost exclusively in analysis of vegetation. The research described in this paper was done more than fifty years ago, during tenure (1937-41) of a Linnean Macleay Fellowship at the University of Sydney. The war and other priorities interrupted the publication of a number of quantitative studies in the thesis (Pidgeon, 1942). Data on isonome mapping of sandstone vegetation might well have been left buried in the thesis, but fora number of reasons it is offered for publication. It isa contribution of historical interest, not only as the original quantitative method devised to analyse the structure and pattern of species distribution and abundance in a plant community, but also as the first quantitative analysis of plant communities on Hawkesbury sandstone. There was a lapse of thirty years before other quantitative studies, using computer-analy- sis, of the vegetation of the central coast of NSW were published: Siddiqi et al. (1972), Burrough et al. (1977), Buchanan & Humphreys (1980). Another motive for offering this paper is the historic aspect of the data. Plant communities like species, can become endangered. With the spread of urbanisation, this has indeed happened to many communities on Hawkesbury sandstone, which, like the two sites investigated in 1941, are now lost to streets and housing. The complete floristic composition of these communities, given in the Appendix (a total of 110 species, includ- ing densities) provide a record of typical urban fringe sandstone communities in the 1940s, prior to development. The species composition of these pristine communities, par- ticularly the absence of introduced species now common in urban fringe vegetation, is remarkable. This paper also re-evaluates the isonome mapping technique, which although superseded by other techniques using computer-analysis, nevertheless has some intrinsic merits. The technique may interest ecologists who wish to summarize, in a simple way, the sociology of plant communities. By examining the vividly descriptive isonome maps, ecol- ogists can readily obtain information on distribution and abundance of species, often more rapidly than from computer analysis. ISonome mapping may also reveal graphic evidence oi temporal change, as in the analysis of sandstone vegetation, where some shrub species, abundant in scrub, were present in woodland as isolated, low-value isonomes, interpreted as ‘relics’. Analysis of patterns of species distribution and abundance may point to important ecological processes operating at small scales, e.g. changes in local density and distribu- Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 261 tion of obligate-seeder woody shrubs in relation to fire history (R. Whelan pers. comm.). As patterns are not always apparent by inspection, techniques for quantifying fine-scale patterns in vegetation are needed. Isonome mapping is one simple technique of those available to ecologists. Previous studies of the vegetation of the central coast of New South Wales have been largely in terms of its variation across a range of habitats (Pidgeon 1937, 1938, 1940, 1941, 1942; Davis 1936, 1941; Siddiqi et al. 1972; Burrough et al. 1977; Buchanan 1980; Buchanan and Humphreys 1980; Outhred et al. 1985; Thomas and Benson 1985). LOCALITY DESCRIPTION Terrain and Soils Two sites were selected on Hawkesbury Sandstone in the south-eastern portion of the uplands of the Hornsby Plateau (Fig. 1), the boundaries of which are described by Bembrick et al. 1980. Broken Bay 33°40" © NY Narrabeen / i Va YY SSS Port Jackson \ Fig. 1. Location of Sites A and B in Kilometres the south-eastern section of the Hornsby Plateau. Proc. LINN. SOC. N.S.W., 115, 1995 262 ISONOME MAPPING OF SPECIES DISTRIBUTION The plateau, of fairly uniform elevation (180-210m ASL), is composed mainly of Hawkesbury Sandstone, a highly cemented and indurated quartz sandstone with thin interbedded shales and pebble lenses (Standard, 1969). The terrain of flat-topped divides and V-shaped valleys is typified by structural benches and a blocked and stepped appear- ance, resulting from the coarse jointing and horizontal bedding of the sandstone (Healy, 1972). Some sandstone strata and their intercalated shales are relatively impervious (Corbett, 1972). Where these underlie slopes of low angle, drainage is poor or impeded, thus determing the type of vegetation. Soils derived from Hawkesbutry Sandstone are low in nutrients, especially phospho- rus (Beadle, 1962) and nitrogen (Hannon, 1956) and range on the plateau from light-tex- tured sandy lithosols to yellow podzolics, with discontinuous patches of acid peats. Vegetation The structural and floristic range of the vegetation of the plateau has been described previously by Pidgeon, (1938, 1940); Buchanan, (1980); Buchanan and Humphreys, 1980; Outhred et al., (1985); Thomas and Benson, (1985). The vegetation of the plateau is predominantly dry sclerophyll woodland with an open canopy (Beadle and Costin, 1952; Specht, 1970), mainly of Eucalyptus spp. and shrublands dominated by sclerophyllous shrubs, many of which also occur in the under- storey of adjoining woodlands. Differences in lithology, micro-topography, drainage and soils determine the distri- bution of two main types of shrubland: dry and moist scrub (Pidgeon, 1938). Dry scrub is found mainly on well drained skeletal soils, resting on sandstone close to the surface. Moist scrub occurs in habitats with impeded drainage, caused by impervious beds of mas- sive sandstone or impervious sub-soils weathered from shale lenses. The soils are intermit- tently moist, occasionally water-logged but the surface is sandy without much organic matter (Pidgeon, 1938). Sometimes perched on benches, the majority of moist scrubs are found in areas of low slope (Buchanan, 1980). The dominance of indicator species, main- ly moisture-tolerant shrubs especially Hakea teretifolia, readily identify these communities. Many shrubs are common to both dry and moist scrub (Pidgeon, 1938) and the two communities may be in proximity (see Fig. 2). Study Sites A characteristic feature of the sandstone uplands is the recurring pattern of Eucalyptus woodland on well-drained soils interspersed with areas of moist scrub. In these adjacent communities, two sites of apparently similar floristic composition were selected for quantitative analysis: Site A, at east Gordon, and Site B, about 10 km to the east in the vicinity of Elanora Heights (Fig. 1). Selected close to roads for easy access; both sites have since been absorbed by suburban development. Figures 2 and 3 show moist scrub in areas of low slope. The boundary between scrub and woodland was quite sharp at Site B (Fig. 3) but less clearly defined at Site A. Grids at Site A were laid 40m apart in typical moist scrub and woodland, whereas the grid at Site B traversed the abrupt shrub-tree junction. METHODS Sampling Technique To sample the vegetation at each site, a grid of rectangular quadrats was used. In dense vegetation, itis easier to count individual plants across the narrow width of a rectan- gle than in squares and, as Clapham (1932) showed, they may estimate density of plants with less variance than square quadrats of the same area (see Pidgeon and Ashby 1940). Tapes were laid in parallel lines 15 feet (4.6m) apart across each site. Every 30 feet Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 263 (9.1m) along the tapes, and for 3 feet (0.9m) to one side, the numbers of individual plants of each species were counted. In this way, data were collected at each site for rows of con- tiguous quadrats (8.3m?) arranged in columns 12 feet (3.7m) apart (see Fig. 4A). y S aS Sat Fig. 2. Moist scrub at Site A, East Gordon, with dry scrub on the ridge. (Photograph: I.M. Pidgeon, 1941). . eS & Avi st ‘ Fig. 3. Abrupt junction between moist scrub and woodland at Site B, Elanora Heights. (Photograph: I.M. Pidgeon, 1941). Proc. LINN. SOC. N.S.W., 115, 1995 264 ISONOME MAPPING OF SPECIES DISTRIBUTION Isonome Mapping Technique In each quadrat, the number of individual plants of each species was totalled and the relative density (relative abundance or importance value) was calculated as a percent- age (e.g. 40 individuals of a species out of a total of 200 plants in the quadrat represent a frequency of 20%). The percentage frequencies for individual species in every quadrat can then be plotted on separate maps of the site, drawn to scale (Fig. 4A). For each species, the variation in its relative density across the grid was mapped by lines drawn to connect areas with the same percentage values, in the same way as contour maps are constructed. These lines, termed ‘isonomes’ (from iso, equal, and nome, distribu- tion) show the pattern of distribution for that species over the site, with one or more cen- tres of high relative density emerging (Fig. 4B). The method of construction for an isonome map is fully explained in Results: Site A, Moist Scrub. By superimposing isonome maps for all species in sufficient abundance to generate discernible patterns, a composite pattern of species distribution over the area may be obtained. The patterns in the isonome maps are obviously dependent on (a) quadrat size and spacing (how these are chosen for a particular type of vegetation is explored in the Discussion) and (b) variation in density of a given species over the grid. RESULTS Isonome maps were prepared (in 1941) for twenty-three species selected to illus- trate various features of the floristic analysis of mainly shrubby species (Table 1 and Figs. 4 9). Overall, one hundred and ten species were recorded (see Appendix). Some species were restricted to moist scrub, others to woodland, while a third group was common to both (see Table 2). TABLE 1 Species represented in isonome maps and density per 85m2. * denotes the site and vegetation type of species represented in Figs. 4-9, x denotes less than 1 per 85m*, (S) seedlings. SITEA SITEB Moist Woodland Moist Woodland Species Scrub Scrub Actinotus minor (Sm.) DC. 21 77* 100 20 Angophora hispida (Sm.) Blaxell 27 16 60* 3 Baeckea diosmifolia Rudge 236* 2 22 Banksia ericifoliaL.£. var. ericifolia 15 9 196* Zhe B. serrataL.f. 28* g* 70* Dillwynia floribunda Sm. x 75* D. retorta (Wendl) Druce 49* 3 44 Epacris microphylla R.Br. 49* 1 1 E. pulchella Cay. 29 60* 90* Eucalyptus gummifera (Sol. ex Gaertn.) Hochr. 15* 5(S) 19 Grevillea sericea (Sm.) R.Br. 12 3 37* G. speciosa (Knight) McGillivray 11 62* 1 Hakea teretifolia (Salisb.) J. Britten 160* 12* 238* 39* Kunzea cajntata Reichb. 673% 19* 246* 14* Leptospermum trinervium (Sm.) J. Thompson 40* 31 9 38 L. squarrosum Gaertner 566* 15% 86 20 Leucopogon microphyllus R.Br. 85* 6* 90 13 Micrantheum ericoides Desf. ] 218* 29 Petrophile pulchella (Schrad.) R.Br. 235* 204* 98 74 Pimelea linifoliaSm. 25(S) 273* Pultenaea elliptica Sm. 2 8 125* 48* P. daphnoides Wendl. 27* P. retusaSm. 16* Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 265 TABLE 2 Species restricted to moist scrub, woodland, and common to both excluding densities of 5 or less per 85 m2. A & B denote the presence of species at either or both sites. For common species, (S) denotes higher densities in moist scrub, (W) in woodland. * denotes seedlings or young plants recorded in moist scrub. Authorities for binomials as in Harden (ed.) Flora of N.S.W. Vols. 1-4. MOIST SCRUB SITES SITES Allocasuarina distyla A Gompholobium minus B Aotus ericoides A B Grevillea speciosa A B Baeckea densifolia A Hakea sericea B B. diosmifolia A B Isopogon anethifolius B B. imbricata A B Leucopogon esquamatus A B Bauera rubiodes B Persoonia lanceolata B Conospermum enicifolium A Phyllota phylicoides B Drosera peltata A B Woollsia pungens Epacris microphylia A WOODLAND Acacia longissima B Lomandra glauca A Adiantum sp. A B Lomaitia silaifolia A B Billardieria scandens A Micrantheum ericoides A B Boronia ledifolia B Monotoca scoparia B Conospermum longifolium B Patersonia glabrata B Dillwynia flonbunda B Platysace linearifolia B D. retorta A B Poranthera corymbosa B Eucalyptus gummifera A B* Pultenaea daphnoides E. haemastoma A B* P. retusa B Gompholoblum virgatum B Ricinocarpus pinifolius B Goodenza stelligera A Styphelia viridis B Grevellea sericea A B Tetratheca ericifolia B Hakea propinqua A B Xanthosia pilosa B Hibbertia aspera A B X. tridentata A B Lasiopetalum ferrugineum B COMMON Actinotus minor A B H. teretifolia A B(S) Angophora hispida A B(S) Hemigenia purpurea B Banksia encifolia A B(S) Kunzea capitata A B (S) B. oblongifolia B (S) Lambertia formosa B B. serrata A B(W) Leptospermum squarrosum A B (S) Boronia pinnata B L. trinervium A B Bossiaea heterophylia A B (W) Leucopogon microphyllus A B(S) B. scolopendria A Petrophile pulchella A B Comesperma ericinum B Pimelea linifolia B*(W) Epacris pulchella A B(W) Pultenaea elliptica A B (S) Hakea dactyloides A B (S) Xanthorrhoea resinifera A B(W) Site A: Moist Scrub Figures 4A and 4B illustrate the method of construction of the isonome map for Leptospermum trinervium, one of the eight most abundant shrubs in moist scrub (Table 1, Site A). The grid map (to scale) in Fig. 4A shows relative densities of L. trinerviumin each quadrat. Using the same principle as in contour mapping, 5% intervals were used to draw lines (isonomes) around these quadrats enclosing relative densities from 30% to 5% (interpolation of isonomes between two recorded frequencies is routine). In other parts of the grid, relative densities from 8% and 4% (column 1, rows 1 and 2) and 5% (column 2, row 4) enabled additional 5% isonomes to be drawn. In column 7, row 2, the quadrat with 11% relative density was circumscribed by 10% and 5% isonomes, as the flanking quadrats were 1%. The 3% isonome circumscribes all quadrats with relative densities from 3% to 5%, and excludes quadrats with lesser values or no presence. Construction of isonomes is, of course, discretionary, but consistent when done by the same person. Isonome maps of the vegetation at this site show various levels of complexity. The Proc. LINN. SOC. N.S.W., 115, 1995 266 ISONOME MAPPING OF SPECIES DISTRIBUTION Leptospermum trinervium Isonome map of columns ——4 SMO Fig. 4A, B. Method of construction of isonome map for Leptospermum trinervium, one of the eight co-dominant shrubs at Site A in moist scrub. A: Plan to scale of the grid; relative densities shown as percentages in each quadrat. B: Isonomes with intervals 5%, and 3% (broken line) drawn as contours from data in A. (See text, Results). Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 267 Leptospermum Squarrosum Baeckea diosmifolia Hakea teretfolia —— Leucopogon icrophyll Petrophile poe RS pulchella --- Epacris 0) 30 ft i microphylla 9.14m Fig. 5. Site A, scrub: Isonome maps (intervals 5%) for the other seven co-dominant shrubs in moist scrub, Site A. Centres of high relative density are: Leptospermum 60%; Kunzea 55% 50%; Baeckea 45% 35% 20% 20%; Petrophile 40% 25%; Hakea 25% 20%; Leucopogon 20% 20%; Epacris 20%. Proc. LINN. SOC. N.S.W., 115, 1995 ISONOME MAPPING OF SPECIES DISTRIBUTION 268 BIO, OJMBIAY BAYEH Loe EEE 2 1334 wnsowenbs Me a a a ae OV 0 WNIAIABU/A wnuedsordey/Oz NYY wnuwadsojday eAydosoiw susedg Yor 7 wordy 102 LZ] snjAydosoiw uobodesne 7/62 | aenel winsowenbs he (e) WwnwJadsojda7 fe) S—— —, Odi t SIGHTS ae esate, VIRIAISTAHOC ON SO ae ee eyeyoind eyyudoned vi Opiaaal Byayoind aydonad Loa Ov ° fe) . ~ = BIOfWSOIp Payoaeg ik ee sea ise : p Beare Te. Lovie eq Fo ‘ t BIJOJJWSOIP BEYOoRG | ] ° he 9 Rees ejeyideo eazuny / OV pyeydeo eezuny ee | | | fe} | Z Apel Oo ° 0 0% oi fe) oO =“ io) O B. Patterns of distribution for shrubs in moist scrub, Site A, obtained by superimposing high value y sup P Fig. 6A, ant shrubs. B: Mosaic obtained by superimposing g. 5. A: Integrity of all high centres exceeding 40%, with small 30% isonomes in the four most abund isonomes from maps of species in Fig. 4B and Fi isonomes of 20% or more for the ei amount of overlap for ght co-dominant shrubs. 115, 1995 Proc. LINN. SOC. N.S.W., I.M. BREWER 269 A Hakea teretifolia Kunzea capitata Leptospermum Leucopogon squarrosum microphyllus B Micrantheum ericoides Dillwynia retorta Pultenaea daphnoides Cc Petrophile pulchella Actinotus minor Banksia serrata eee 5 Eucalyptus gummifera -------- 0 30 ft ue 9.14m Fig. 7. Site A, woodland: isonome maps for eleven species with different patterns of distribution (see text). 7A: Four scrub species with restricted distribution of Hakea5% 3%; Kunzeaand Leptospermum 10% 5%; Leucopogon 3%. 7B: Three woodland shrubs with decreasing centres of relative density: Micrantheum 35% 35%; Dillwynia 20%; Pultenaea 15%. 7C: Two species occurring in scrub and woodland; centres of relative density are: Petrophile 60% 35%, Actinotus 30% 10%; and two species in the tree stratum: Banksia 15%; Eucalyptus 10%. Proc. LINN. SOC. N.S.W., 115, 1995 270 ISONOME MAPPING OF SPECIES DISTRIBUTION isonome map for L. trinerviumis a simple pattern of distribution, with two centres of high frequency (30% and 10%) . Isonome maps (Fig. 5) for the other seven shrubs with the highest densities (Table 1), show mostly complex patterns which have several centres of high relative density varying from 60% to 20% (see legend, Fig. 5). Although Hakea teretifolia appeared to be the dominant shrub at both sites in moist scrub, at Site A its density was significantly less than that of Kunzea capitata and Leptospermum squarrosum (see Table 1). However, as recorded in the thesis, only 10% (approx) of the population of H. teretifoliawas less than 30cm high, whereas 50% (approx) of the population of both K. capitataand L. squarrosumwas in this category. By superimposing isonome maps for the eight shrub species (Figs. 4B and 5), the composite pattern over the whole area (1112-m*) may be examined. Figure 6B, showing relative densities of 20% and above, reveals a complex ‘social’ structure in which the centres of high relative density for the various species form a mosaic, around the margins of which there is overlapping of the lower-value isonomes. From the method of construction, it is clear that the collective number of relative densities cannot exceed 100% in any one place. The integrity of high centres greater than 40% and the small amount of overlap for isonomes of 30% are illustrated by Fig. 6A. Site A: Woodland In 500-m* of the adjacent Eucalyptus gummifera— E. haemastoma woodland, isonome maps (Fig. 7) were selected for eleven species with varying density (Table 1) and different patterns of distribution. Four shrub species Hakea teretifolia, Kunzea capitata, Leptospermum squarrosum and Leucopogon microphyllus (Fig. 7A) that were abundant in moist scrub but had low densities in woodland (Table 1), show restricted patterns of low-frequency isonomes (Fig. 7A). This contrasts with the complex isonome maps for these same species in moist scrub (Fig. 5). Three shrub species restricted to woodland (Table 2), Micrantheum ericoides, Dillwynia retorta and Pultenaea daphnoides (Fig. 7B) with less complex patterns had corresponding decreasing densities (Table 1). Petrophile pulchella (Fig. 7C) with similar high densities in scrub and woodland (Table 1) reveals a pattern similar in complexity to its isonome map in scrub (Fig. 5). In contrast, Actinotus minor (Fig. 7C), a herb also common to both habitats, but significantly less abundant than P. pulchella (Table 1) showsa simple pattern of distribution over part of the woodland. For two species in the tree canopy (Fig. 7C) the distribution of isonomes show no overlap. Site B At site B, the grid covered 933-m2 of moist scrub and 1510-m?’ of Eucalyptus gum- mifera — E. haemastoma woodland, with an abrupt tree boundary (Fig. 3). Isonome maps for twelve species of shrubs reveal various patterns of distribution (Figs. 8A, 8B, 9A, 9B). Two species, Hakea teretifolia and Kunzea capitata, (Fig. 8A) with high densities (Table 1) and complex patterns in moist scrub, show restricted distribution in the adjoin- ing woodland, as at Site A (Fig. 7A). Conversely, Pimelea linifolia (Fig. 8B) was restricted to woodland and occurred only at Site B where its isonome pattern is complex. Its occur- rence in moist scrub was solely as seedlings (Table 1). Epacris pulchella (Fig. 8B), though present only in the woodland at Site A (Table 1) occurred in both scrub and woodland at Site B, where the high centres of distribution on each side of the tree boundary are linked by lower-value isonomes. Pairs of species in three genera, Grevillea (Fig. 9A), Banksia and Pultenaea (Fig. 9B), have complementary patterns of distribution in scrub and woodland, with some overlap- ping: G. sericeainto scrub, B. ericifoliaand P. ellipiticainto woodland. The isonomes for two species are abruptly terminated at the treeline, confining Dillwynia floribunda to wood- land, and Angophora hispida to scrub (Fig. 9A). Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 271 DISCUSSION In these field trials of the isonome mapping technique, the priority was to deter- mine whether the selected size and spacing of the quadrats was appropriate for mapping isonomes in shrubland and woodland. Investigations at both sites were therefore limited to a floristic analysis. No soil samples (the obvious variable) were taken at this time. However, other field investigations by the author in similar areas indicated the impor- tance of soils and drainage patterns in determining the two vegetation types: woodland and moist scrub. Isonome Mapping Techniques Based on relative abundances of the component species, the isonome method of analysis of vegetation uses density as the principal attribute with which to describe vege- tation. Isonome mapping seeks to display variation in patterns of distribution of individual species by mapping contours from the variation of their relative density in sampling units across the grid. The maps generated are dependent not only on the distribution and den- sity of the indivuals of species mapped, but also on the spacing and size of the quadrats | 4A | | 15 25: \ «—— MOIST SCRUB ——_> |—__—_ WOODLAND —————> Hakea teretifolia Kunzea capitata Fig. 8A Site B: isonome maps for two shrubs dominant in moist scrub (see text). Intervals 5%, with 2% boundaries (dotted lines). Centres of relative density are: Hakea 30% 20% 15% (scrub), 20% (woodland); Kunzea 40% (scrub), 10% (woodland). Proc. LINN. SOC. N.S.W., 115, 1995 272 ISONOME MAPPING OF SPECIES DISTRIBUTION used. Clearly, if the size of these is so small that few individuals of few species are included in each sample, then relative densities will vary greatly over short distances, and it would be difficult to map isonome contours that are meaningful. Conversely, if the quadrats are very large, variation in relative density of species will be discerned only at large scales. The investigator has to choose the scales of variation that are of interest. In this case, interest was in variation of patterns of distribution in species, especially shrubs, between two structurally distinguishable types of vegetation: scrub and shrub-woodland. The area sampled in each had to be large enough to represent the range of variation in distribution of the common species in them, and the samples taken in them appropriately sized and spaced to explore that range of variation. The size of the samples, rectangular quadrats (9.1 x 0.9m), was large enough to contain individuals of anumber of species and they were numerous enough and appropriately spaced to map isonome contours within the types of vegetation. The isonome maps produced thus reflected the variation in relative densities of species within the types of vegetation and allowed comparison of patterns of variation in relative abundances between them. The quadrat size and spacing may be varied to suit the type of vegetation: the method was worked out on dense scrub vegetation (see Figs 2, 3). Other vegetation types <— MOIST SCRUB ——>|*#——— WOODLAND. ————> Pimelea linifolia Epacris pulchella 0 60 ft _ ee | 18.29 m Fig. 8B Site B: ecotone boundaries for two shrubs (see text). Isonome intervals 5% with 2% boundaries (dotted lines). Centres of relative density are:Pimelea 40% 35% 30% 25% (woodland) ; Epacris 30% 10% (woodland), 15% (scrub). Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 273 may require quadrats of different dimensions. For sand dune vegetation (see Pidgeon, 1942) it was more appropriate to use sampling units of half the length (compared with scrub and woodland), to compensate for the spatial zonation of vegetation from fore- dune to hind-dune. In an investigation of grassland, it is very probable that the quadrats would need to be smaller and closer together than were used in this analysis. A similar system of isonomes is obtained if absolute densities (actual numbers) instead of percentages (relative densities) of each species is plotted. Isonomes not calcu- lated on a percentage basis have a limited application; they are not so useful when com- paring relative densities of species, and cannot be used as in Fig. 6 for a composite pattern of the area (Pidgeon & Ashby, 1942). It is generally more useful to eliminate variations in absolute density of the plant population over the area investigated by reducing figures to percentage cover of individu- al species. However, where there are considerable bare areas, e.g. on coastal dunes, or in arid regions, the vegetation cover may be expressed as percentage cover of total quadrat area; this also estimates at the same time the amount of the bare area. Percentage cover of each quadrat would also be estimated for stoloniferous grasses and for cloned species. With distinct two-layered communities, it may be advantageous to collect the data from the separate strata, and construct two sets of isonome maps; where the ground layer is insignificant, it may be ignored altogether. Grevillea sericea Dillwynia floribunda Qa z 5] 8 = [ee] 5) c O 7) | Y Oo = Grevillea speciosa 0o Bott Angophora hispida 18.29 m Fig. 9A Site B: ecotone boundaries for four shrubs (see text). Intervals of 5% with 2% boundaries for Grevillea spp. Centres of relative density shown: G. sericea 20% (scrub) 10% 5% (woodland); G. speciosa 10% (exclusive to scrub); Dillwynia 25% 15% 15% (exclusive to woodland); Angophora 15% 10% (exclusive to scrub). Proc. LINN. SOC. N.S.W., 115, 1995 274 ISONOME MAPPING OF SPECIES DISTRIBUTION The labour involved in counting individuals of several species in numerous samples limits isonome mapping to small areas of less than a hectare, more than sufficient to map patterns of distribution. With the assistance of about 20 second-year ecology students in 1941, the field recordings in this study were completed in one day at each site. The work involved in constructing isonome maps would be reduced by computer graphics. Although random-sampling techniques are most frequently used in quantitative analysis, Buchanan and Humphreys (1980) also used a grid (of circular quadrats in columns and rows) traversing the sharp boundary between podzol and non-podzol soils at two sites on Hawkesbury sandstone, to analyse by modern techniques, the sudden floristic change which accompanies the change in soil type. Patterns of variation in distribution and abundance of species Tree seedlings which become established in moist scrub (see Appendix) during dry summer periods do not survive when the water table rises after prolonged wet periods. The lack of aeration affects their root systems, which penetrate deeper than the shrubby species. During a series of dry years, Buchanan (1980) also recorded, in dried swamps, the growth of Eucalypts to several metres, before the water table rose and killed them. Banksia serrata Pultenaea retusa = \ — f— = —— _—_——n fo) WOODLAND ————> <——— MOISTSCRUB — | Banksia erlcifolla Pultenaea elliptica 60 ft 18.29 m Fig. 9B. Site B: Distribution of four shrubs in relation to the ecotone; isonome intervals 5%. Centres of relative density are: Banksia serrata 10% 10% 5% 5% (exclusive to woodland); B. ericifolia 20% 20% 20%; Pultenaea elliptica 25% 20% 15%; P. retusa5% 5%, 5% (exclusive to woodland). Proc. LINN. SOC. N.S.W., 115, 1995 ILM. BREWER 275 In the distribution of species between scrub and woodland, there were consistent results across both sites; some species were restricted to woodland, some to moist scrub; others were in common to both vegetation types; in this third group, most had substan- tially higher densities in scrub or woodland (Table 2). With the more abundant species of shrub, isonome mapping illustrated variation in relative densities between woodland and moist scrub, again with a high degree of consistency between the two sites. Overall, the density of shrubs in scrub and woodland is in the approximate ratio of 2:1 at both sites (see Appendix). Isonome mapping in this study revealed patterns of variation in abundance and distribution of species across the ecotones between woodland and moist scrub (Site B) that were not apparent even to the keen observer: (1) The exact extent of the overlap for shrub species which intrude into woodland, or understorey shrubs which intrude into scrub (2) Complementary patterns of distribution between pairs of species of the same genera that were precisely defined in scrub and woodland (3) The abrupt termination, at the sharp shrub/tree ecotone of some species of shrub, exclusive to woodland or moist scrub Differentiation in distribution and density of species may be related to spatial varia- tion in soil or in the behaviour of fires. Buchanan (1980) observed that, at abrupt ecotones of moist shrubland and wood- land, an obvious change in understorey shrubs (not specified) coincided with an abrupt junction of the impervious sub-soil (less than 1m below the surface) with the well drained sandstone soil of the woodland; a more gradual change in the understorey shrubs takes place when the impervious layer thins or deepens beneath the trees. As the abundance of some species of understorey shrubs at Site B is closely associated with the abrupt ecotone (Figs. 8B, 9A) it may be inferred from Buchanan’s observations that the junction between the impervious sub-soil of the scrub with woodland soil at Site B was also abrupt, as was assumed, but not verified in 1941. Variations in vegetation on sandstone, related to variations in soil characteristics, have been observed by Pidgeon (1937, 1941), Siddiqi et al. (1972), Burrough et al. (1979) , Buchanan and Humphreys (1980) and noted by others, Thomas and Benson (1985), Outhred et al. (1985). Though no observations were made on the soils at Sites A and B, the differences between woodland and moist shrubland soils are generally well known, partic- ularly the well-drained woodland soils in contrast to the temporary waterlogging charac- teristic of the moist scrub soils (Pidgeon 1938, Beadle 1962, Buchanan, 1980). The overall distribution and abundance of species between woodland and moist scrub and the ex- clusive presence (or preference) for a particular habitat (Table 2) is more likely to be related to spatial variation in soils, than to spatial variation in the behaviour of fires. There was no evidence of recent fires at either site, confirmed by the dominance in moist scrub of shrubs killed by fire (Table 1). Fire is probably the main factor in determin- ing composition and abundance of obligate-seeder species, largely controlled by fire frequency and inter-fire periods (see Thomas & Benson, 1985; Benson, 1985; and Keith in Press). Recent research into population dynamics, starting with such observations, has been developed in studies of individual species (Bradstock & Myerscough, 1981; Auld, 1986). Isonome mapping provides an avenue for relating these population-level studies to the sociology of communities which an experienced ecologist can interpret, often more rapidly than from computer-analysis. The restricted, low-value isonomes of obligate-seeder shrubs in woodland, abun- dant in scrub (Figs. 7A, 8A), could be considered as ‘relics’ of a pyric succession. Alternatively, they may be due to shading out by taller growing species in the absence of fire. In summary, isonome maps presented in this floristic analysis of moist scrub and woodland have revealed, for the first time in graphic detail, the structure, sociology, distri- Proc. LINN. SOC. N.S.W., 115, 1995 276 ISONOME MAPPING OF SPECIES DISTRIBUTION bution and abundance of species in these sandstone communities, and supported some observations by other researchers: (i) The association between high density of occurrence and vegetation type has been confirmed. (ii) Patterns of occurrence and density across ecotones have been revealed. (iii) Patterns of positive association, not previously apparent, have been discovered. (iv) Specific patterns i.e. relics in woodland of some obligate - seeder woody shrubs, abundant in adjacent scrub, have generated an hypothesis of temporal change operating at a small scale. In an age when sophisticated computer-based statistical techniques are used to anal- yse the structure of plant communities, the simplicity of isonome mapping, in association with investigation of micro-environmental factors, offers some advantages which are worth further exploration. ACKNOWLEDGMENTS The constructive criticism and valuable advice by Professor R. J. Whelan, Department of Biological Sciences, University of Wollongong, and Dr P. J. Myerscough, School of Biological Sciences, University of Sydney, on the presentation of this paper, is gratefully acknowledged. I also acknowledge my indebtedness to the late Lord Ashby for his comments on an earlier draft and his endoresement to expand the 1941 data on the isonome mapping technique. I thank Dr W. T. Williams for reading an early draft, for his encouragement to publish, and his assurance that a program could be written for com- puting the data for graphic display. I thank Ms Jan Fragiacomo for her generous assistance in word processing. The field work and figures were completed when the author was a Linnean Macleay Fellow in Botany (1937-1941), at the University of Sydney. References Asuy, E. 1935 — The Quantitative Analysis of Vegetation (with an Appendix by W. L. Stevens). Annals of Botany 49: 779-802. AULD, T. D. 1986. — Population dynamics of the shrub Acacia suaveolens (Sm. Willd.: Fire and the transition to seedlings. Australian Jounal of Ecology 11: 373-85. BEADLE, N. C. W. 1962: — Soil Phosphate and the Delimitation of Plant Communities in Eastern Australia. Ecology 43: 281-8. BEADLE, N.C.W. and Costin, A. B. 1952. — Ecological Classification and Nomenclature. Proceedings of the Linnean Society of New South Wales’77: 61-82. BEMBRICK, C., HERBERT, C., SCHEIBNER, E. and STUNTZ, J., 1980. - Structural sub-division of the Sydney Basin In: A Guide to the Sydney Basin, Herbert, C. and Helby, R. (eds.) Geololgical Survey of N.S. W. Bulletin 26 Dept. of Mineral Resources. BENSON, D. H. 1985. — Maturation periods for fire-sensitive species in Hawkebury Sandstone vegetation. Cunninghamia 1: 339-341. BLACKMAN, G.E. 1935 — A Study by Statistical Methods of the Distribution of Species in Grassland Association. Annals of Botany 49: 749-77. BRADSTOCK, R.A. and MyERSCOUGH, P. J. 1981. — Fire effects on seed release and the emergence and establishment of seedlings in Banksia enicifolia L.£. Australian Journal of Botany 29 521-531. Bray, J. R. and Curtis, J. T. 1957. — An ordination of the upland forest communities of Southern Wisconsin. Ecological Monograph 27: 325-49. BUCHANAN, R. A. 1980. — The Lambert Peninsula, Ku-ring-gai Chase National Park. Physiography and the distri- bution of podzols, shrublands and swamps, with details of the swamp vegetation and sediments. Proceedings of the Linnean Society of New South Wales 104: 73-94. BUCHANAN, R. A. and Humpureys, G. S. 1980. — The vegetation on two podzols on the Hornsby Plateau, Sydney. Proceedings of the Linnean Society of New South Wales 104: 49-71. BURROUGH, P. A., BRowNn, L. and Morris, E. C. 1977. — Variations in vegetation and soil pattern across the Hawkesbury Sandstone plateau from Barren Grounds to Fitzroy Falls, N.S.W. Australian Journal of Ecology 2: 137-59. CLAPHAM, A. R. 1932. — The Form of the Observational Unit in Quantitative Ecology. Journal of Ecology. 20: 192-7. CLAPHAM, A.R. 1936 — Over-dispersion in Grassland Communities and the Use of Statistical Methods in Plant Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 277 Ecology. Journal of Ecology 24: 236-51. CorsBETT, J. R. 1972. — Soils of the Sydney Area. In: The City as a Life System? Ed. H. A. Nix. Proceedings of the Ecological Society of Australia 7: 41-78. Davis, C. 1936. — Plant ecology of the Bulli District I. Stratigraphy physiography and climate: general distribution of plant communities and interpretation. Proceedings of the Linnean Society of New South Wales 61: 279-85. Davis, C. 1941. — Plant ecology of the Bulli District II. Plant communities of the plateau and scarp. Proceedings of the Linnean Society of New South Wales 66: 1-19. GaucHu, M. G. 1982. — Multivariate Analysis in Community Ecology. Cambridge: Cambridge University Press. _ GoopALL, D.W. 1961 — Objective Methods for the classification of vegetation. IV. Pattern and minimal area. Australian Journal of Botany, 9: 162-196. GrEIG-SMITH, P. 1983 — Quantitative Plant Ecology, 3rd Ed. Blackwell Scientific Publications, Oxford. HANNON, N. J. 1956. — The Status of Nitrogen in the Hawkesbury Sandstone Soils and their Plant Communities in the Sydney District 1. The Significance and Level of Nitrogen. Proceedings of the Linnean Society of New South Wales 81: 119-42. HARDEN, G. J. 1990, 91, 92, 93. — (Ed.) Flora of New SouthWales Vols. 1, 2,3, 4. New South Wales University Press. Heay T. R. 1972 — Structure and Terrain of the Sydney Area. In: The City as a Life System. Ed. H.A. Nix. Proc. Ecol. Soc. Aust. Canberra. 7: 27-40. JAMES, F. C. and McCuLLocu, C. E. 1990. — Multivariate Analysis in Ecology & Systematics: Panacea or Pandora’s Box? Annual Review of Ecology & Systematics. 21: 129-166. KEITH, D. 1994. — Jn Press McINTOSH, R. P. 1967. — The Continuum Concept of Vegetation. Botanical Review 33: 130-87. OUTHRED, R. LAINSON, R., LAMB, R. and OUTHRED, D. 1985. — A Floristic Survey of Ku-ring-gai Chase National Park. Cunninghamia 1: 313-38. PIDGEON, I. M. 1937. — The Ecology of the Central Coastal Area of N.S.W. I. The Environment and genera features of the vegetation. Proceedings of the Linnean Society of New South Wales 62: 315-40. PIDGEON, I. M. 1938. — The Ecology of the Central Coastal Area of N.S.W. II Plant succession on the Hawkesbury Sandstone. Proceedings of the Linnean Society of New South Wales 63: 1-26. PIDGEON, I. M. 1940. — The Ecology of the Central Coastal Area of N.S.W. III. Types of primary succession. Proceedings of the Linnean Society of New South Wales 65: 221-49. PIDGEON, I. M. 1941. — The Ecology of the Central Coastal Area of N.S.W. IV. Forest types on soils from Hawkesbury Sandstone and Wianamatta Shale. Proceedings of the Linnean Society of New South Wales 66: 113- 37. PIDGEON, I,. M. 1942. — Ecological Studies in N.S.W. D.Sc. thesis University of Sydney, unpublished. PIDGEON, I. M. and Asupy, E. 1940. — Studies in Applied Ecology I. A statistical analysis of regeneration following protection from grazing. Proceedings of the Linnean Society of New South Wales 65: 123-43. PIDGEON, I. M. and Asusy, E. 1942. — A New Quantitative Method of Analysis of Plant Communities. The Australian Journal of Science 5: 19-21. STAPLEDON, R.G. 1912 — Pasture Problems: Drought Resistance. Journal of Agricultural Science 5: 132. SippIQl, M. Y., CAROLIN, R. C. and ANDERSON, D. J. 1972. - Studies in the Ecology of Coastal Heath in N.S.W. Proceedings of the Linnean Society of New South Wales 97: 211-24. SPECHT, R. 1970. — Vegetation. The Australian Environment 4th Ed. Leeper, G. W. (ed). CSIRO, Melbourne. STANDARD, J.C. 1969. — Sydney Basin: Hawkesbury Sandstone. In Geology of N.S.W. Ed. A.H. Packham. Journal of the Geological Society of Australia 16: 407-417. THoMas, G. and BENSON, D. H. 1985. — Vegetation Survey of Ku-ring-gai Chase National Park. National Herbarium of N.S. W. Royal Botanic Gardens, Sydney. WILLIAMS, W. T. and LAMBERT, J. M. 1959. — Multivariate methods in plant ecology: I Association analysis in plant communities. Journal of Ecology 47: 83-101. Proc. LINN. SOC. N.S.W., 115, 1995 278 ISONOME MAPPING OF SPECIES DISTRIBUTION APPENDIX Species and densities per 85m. sq. recorded in quantitative analysis of adjacent areas fmoist scrub and woodland, Sites A and B. x =recorded but less than 1 per 85 m. sq., + = cloned species not estimated, s = seedlings only Authorities for the binomials: Harden, Ed. (1990-93), Vols. 1-4, Flora of New South Wales SITEA SITEB Species Moist Woodland Moist Woodland Scrub Scrub Allocasuarina distyla 19 1 Acacia brownii 2, A. longissima 15 A. myrtifolia 4 A. suaveolens 5 2 1 x Actinotus minor 21 77 100 20 Adiantum sp. 3 31 1 7a Angophora hispida 27 16 60 3 Aotus ericoides 16 78 x Baeckea brevifolia 2 B. densifolia 8 B. diosmifolia 236 2 22 B. imbricata 36 1 a Banksia ericifolia var. ericifolia 15 9 196 22 B. oblongifolia 3 2 39 9 B. marginata 5 B. serrata 28 9s 70 B. spinulosa 3 5 5 Bauera rubioides 75 5 Billardieria scandens ] 11 x Boronia ledifolia 5 30 B. pinnata 2 19 12 Bossiaea heterophylla 1 42 7 27 B. scolopendria 6 17 3 1 Cassytha glabella x Caustis flexuosa x Ceratopetalum gummiferum x Choretrum candollei x Comesperma ericinum 23 18 Conospermum ericifolium 8 C. longifolium 14 Dampnrera stricta 1 4 Darwinia fasiculans 1 3 Dianella caerulea 2 D. laevis 1 3 Dillwynia retorta 49 3 44 D. floribunda x 75 Drosera peltata 25 3 7 x Epacris microphylla 49 1 1 E. pulchella 29 60 90 Eucalyptus eugenioides 1 E. gummifera 15 5s 19 E. haemastoma Is 5 6s 6 E. racemosa x E. sieberi x Gompholohium virgatum 1 16 G. minus 8 Goodenia stelligera 3 25 3 1 Gonocarpus teucrioides 2 1 Grevillea buxifolia 3 4 2 G. speciosa 1] 62 ] G. sericea 12 3 37 Hakea dactyloides 11 2 53 8 H. propinqua 5 8 H. sericea 13 H. teretifolia 160 12 238 39 Hemigenia purpurea ] 10 6 Hibbertia aspera 8 24 Proc. LINN. SOC. N.S.W., 115, 1995 I.M. BREWER 279 SITEA SITEB Species Moist Woodland Moist Woodland Scrub Scrub H. riparia 1 2 1 Isopogon anethifolius 4 24 3 Kunzea capitata 673 19 246 14 Lambertia formosa 1 17 18 Lasiopetalum ferrugineum x 16 Lepidosperma sp. 11 8 2 L. laterale 1 Leptospermum arachnoides x 1 L. squarrosum 566 15 86 20 L. trinervium 40 31 9 38 Leucopogon appressus 3 L. esquamatus 15 5 98 2 L. microphyllus 85 6 90 13 Lindsaea microphylla 3 x x Lomandra sp. + Lomandra sp. + + L. glauca + + Lomatia silaifolia 12 6 Micrantheum ericoides 1 218 29 Mirbelia rubiufolia 1 5 Monotoca scoparia 1 11 Olax stricta 2 Patersonia glabrata 1 35 Persoonia hirsuta 2 P. lanceolata 9 2 16 3 P. mollis 1 P. levis 4 4 Petrophile pulchella 235 204 98 74 Philotheca salsolifolia 1 Phyllota phylicoides 59 Pimelea linifolia 25s 273 Platylobium formosum 2 Platysace linearifolia 1 25 1 13 Poranthera corymbosa 10 Pteridium esculentum 1 Pultenaea elliptica 2 8 125 48 P. daphnoides 27 P. retusa 16 P. stipularis 2 Restio fastigiatus + Ricinocarpus pinifolius 1 6 Stylidium graminifolium 5 1 Styphelia tubiflora 2 S. viridis 4 9 Telopea speciosissima 2 Tetratheca ericifolia 1 2 8 Woollsia pungens 13 1 Xanthorrhoea resinosa 7 12 8 4] Xanthosia pilosa 20 X. tridentata 7 3 12 Xylomelum pyriforme 4 Zieria laevigata 5 1 Proc. LINN. SOC. N.S.W., 115, 1995 any 23m artoues ind evi cer “Cavetreen aes _ Spotter inn Pietra Ce eet. cattle fo wreggolian i diene am Didar Feukeus ee "Lane atte HE Le 7 ters 8s. - : drathersrene) erage Se Prewntirets Pa fete ce . oe i ew S Fo ae , fen Fens oe Tics aes eer Section 2 — general papers 151 163 193 213 225 233 239 247 259 KING, R. Mangrove macroalgae: a review of Australian studies (1993 Presidential Address). CHALSON, J.M. and Martin, H.A. The pollen morphology of some co-occurring species of the family Myrtaceae from the Sydney region. MARTIN, H.A. The stratigraphic palynology of the Murray Valley in NSW. DuckER, S. W.H. Harvey in New South Wales. SEMPLE, G.P. Studies of pathogenic Thelohania species in the Australian freshwater crayfish Cherax quadricarinatus. SCANLON, JOHN D. First records from Wellington Caves, New South Wales, of the extinct madtsoiid snake Wonambi naracoortensis. STRATFORD, J. AND AITCHISON, J. Lower Permian fauna from Manning facies rocks along the Peel-Manning fault system, Glenrock Station, Southern New England orogen. MERRICK, J.R. Diversity, distribution and conservation of freshwater crayfish in the eastern highlands of NSW. BREWER, I. Isonome mapping: graphic analysis of patterns of species distribution. The Linnean Society of New South Wales publishes in its Proceedings original papers and review articles dealing with biological and earth sciences. Intending authors should write to the editor (M. L. Augee, Biological Science, University of NSW, Sydney NSW 2052, Australia) for instructions on the preparation of manuscripts and procedures for submission. Manuscripts not prepared in accordance with the Society’s instructions will not be considered. Proc. LINN. SOC. N.S.W., 115, 1995 PROCEEDINGS OF LINNEAN SOCIETY OF NEW SOUTH WALES VOLUME 115 Issued 1 September, 1995 CONTENTS: Section 1 — Papers in honour of Peter Myerscough 3 5 17 25 35 45 61 77 89 109 121 135 Keith, D. Introduction to the special section on plant ecology in honour of Peter Myerscough. AuLb, T.D. and Tozer, M. Patterns in emergence of Acacia and Grevillea seedlings after fire. BEDWARD, M. Simple models of pattern and process. BRADSTOCK, R.A. Demography of woody plants in relation to fire: Telopea speciosissima. CLARKE, P.J. The population dynamics of the mangrove shrub Aegiceras coinieniatin (Myrsinaceae): fecundity, dispersal, establishment and population structure. DUNCAN, F. and BROwN, M.J. Edaphics and fire: an interpretative ecology of lowland forest vegetation on granite in northeast Tasmania. KEITH, D. How similar are geographically separated stands of the same vegetation formation? Amoorland example from Tasmania and mainland Australia. Lowman, M.D. Herbivory in Australian forests — a comparison of dry sclerophyll and rain forest canopies. Monrnris, E.C. Self-thinning in Ocimum basilicum grown at three pel fertility levels with and without mycrorrhizal inoculum. Morrison, D.A. Some effects of low-intensity fires on populations of co-occurring small trees in the Sydney region. PANNELL, J.R. Factors affecting seedling recruitment of Allocasuarina distyla and A. nana at burnt and unburnt sites. WELLS, A.G. Classificatory groupings of tidal river systems in the Northern Territory and Kimberley region of W.A. on presence/absence of mangrove species. Contents continued inside Printed by Southwood Press Pty Limited, 80-92 Chapel Street, Marrickville 2204 So > Slee aoe rot ‘ TRA Beret ao end : Eee pR aee ay fe ‘ na Nt Si hes i Senet LA Chest Or Lb ds tae Nahata hoay ier Pe)