REVIEW OF PALAEOBOTANY AND
ELSEVIER
Review of Palaeobotany and Palynology98 (1997) 187-205
PALYNOLOGY
Eucalyptus (Myrtaceae) pollen and its potential role in investigations of Holocene environments in southwestern Australia E.J.
Pickett ", J.C. Newsome b,,
a Department of Geography, University of Western Australia, Perth, WA. 6009, Australia b Biological Science, School of Environmental and Biological Sciences, Murdoch University, Perth, W.A. 6150, Australia Received 10 July 1996; accepted 11 March 1997
Abstract
This study investigated the potential for using suites of pollen morphological characters to identify pollen of
Eucalyptus (Myrtaceae) to species in modern and fossil material. Characteristics were examined in the modern pollen from nine species of Eucalyptus that occur in an area around Walpole, on the south coast of southwestern Australia. Success was achieved in separating the modem pollen into defined pollen types. These character suites were then applied to fossil Eucalyptus pollen in a nearby Holocene sequence, Boggy Lake, to assess their usefulness in distinguishing fossil pollen into types. Considerable success was achieved with this, although up to 50% of the pollen grains could not be allocated to the pollen types for a variety of reasons. However, the study demonstrated that the fossil records of Eucalyptus pollen are potentially more environmentally informative than has previously been thought, at least for Holocene studies in this region. The use of suites of characters may be applicable to the identification of Eucalyptus species elsewhere, and also to other genera within the Myrtaceae. © 1997 Elsevier Science B.V.
Keywords: Eucalyptus; Myrtaceae; pollen; Western Australia; Holocene
1. Introduction
The southwest of Western Australia is a large geographical region, but is one in which relatively few palynological investigations of Quaternary palaeovegetation and environments have been undertaken. Exceptions are the pioneering pollen work of Churchill (1968) and a small number of more recent studies (e.g. Backhouse, 1993; Newsome and Pickett, 1993). The region has a Mediterranean climate with a number of dry
* Corresponding author. Fax: +61-9360-6303; e-mail:
[email protected] 0034-6667/97/$17.00 © 1997Elsevier Science B.V. All rights reserved. PII 0034-6667 ( 97 ) 00028-6
months each year, and as a result many wetlands are seasonal. Despite this, there are many swamp and lake sites in the region that are suitable for palynological investigation. Although many of these are no older than Holocene (e.g. Churchill, 1968; O'Connor, 1986; Backhouse, 1993; Newsome and Pickett, 1993) there are also a number of sites with records that extend back into the Late Pleistocene. One concern with regard to palynological investigations in this region is that due to characteristics of the flora and the vegetation, the pollen records themselves may not readily yield information about palaeoconditions. The region is dominated by sclerophyll plant communities. This is in con-
188
E.J. Pickett, J.C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 18~205
trast to southeastern Australia which has, at least in parts, a much greater variety of vegetation types including rainforest, sclerophyll communities and grasslands, and where palynological studies have already provided a great deal of information on past vegetation and environments (reviewed in Wasson and Donnelly, 1991). Although the sclerophyll communities of southwestern Australia are ftoristically diverse (e.g. Marchant, 1973; Hopkins et al., 1983; Hopper, 1992) they are, on the whole, structurally dominated by Eucalyptus and other Myrtaceae. The dominant trees of the forest and woodland vegetation types are commonly Eucalyptus species, although other genera may be important locally, e.g. Melaleuca in wetlands. Where forests or woodlands do not dominate, usually as a result of edaphic factors, then Eucalyptus species may be a much less conspicuous part of the vegetation, but other Myrtaceae remain an important component. This is the case in the coastal and inland shrublands. The Myrtaceae family produces distinctive pollen grains, characterised by their syncolpate or parasyncolpate appearance. A number of studies on pollen of the Myrtaceae family, or parts of it, have been undertaken (Pike, 1956; Gadek and Martin, 1981; Martin and Gadek, 1988; Chalson and Martin, 1995). There is much variety in the pollen features of the group, superimposed on the basic form, but at the generic and species levels separation of taxa can be difficult or impossible. Because of these characteristics, and in view of the dominance of the Myrtaceae family in the vegetation across the region, it is possible that some environmental (and climatic) changes might not be recorded in fossil pollen records. This is most likely to be of concern where such shifts are subtle, as in the Holocene (Wasson and Donnelly, 199l). Differentiation of Myrtaceae pollen in fossil assemblages, where attempted, has generally been limited to the separation of major taxonomic groups such as Eucalyptus, Melaleuca and Leptospermum. Few studies have attempted more detailed subdivision of Myrtaceae pollen, and these have concentrated on separation of Eucalyptus pollen types (e.g. Churchill, 1968; Dodson, 1974, 1977; Dodson and Wilson, 1975; Ladd, 1979a,b). It remains a pressing concern to determine the
extent to which the Myrtaceae can be differentiated into groups based on their pollen morphology. The most important question is whether Eucalyptus pollen might be identifiable to species, or groups of species. Churchill (1956) suggested that pollen of the Eucalyptus species occurring in one area of southwestern Australia could be separated using morphological characteristics. He produced a pollen identification key for these species, which he also applied to fossil material (Churchill, 1956). From this he reconstructed Holocene vegetation and climates based primarily on changes in the Eucalyptus pollen record. However, the full details of his taxonomic study have never been published, and his palaeoclimatic conclusions, although widely quoted, are not supported by other evidence. In view of contemporary interest in palaeoclimates there was a need to reappraise this earlier work (Churchill, 1956, 1968) and clarify the interpretations (Pickett, 1990; Newsome and Pickett, 1993). This study presents the pollen taxonomic aspects of that reappraisal. The pollen of the nine Eucalyptus species that occur in a small region of southwestern Australia was examined to see if the species could be distinguished and if so, to assess the extent to which they are recognisable in fossil sequences from the same area.
2. Materials and methods
The area chosen for the investigation lies on the south coast of Western Australia and is centred on the small town of Walpole (ca. 350 km southsoutheast of Perth). The area has a mixture of forest and woodland types in which a number of Eucalyptus species are dominant, as well as some heath and swamp communities (Beard, 1981; Wardell-Johnson et al., 1989). The area also holds a number of sites that are potentially suitable for palynological investigation, most notably Boggy Lake, a small shallow lake situated 4 k m southwest of Walpole (36°01'S, 116°38'E). This site was first investigated by Churchill (1968) and subsequently re-examined (Pickett, 1990; Newsome and Pickett, 1993). The
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
Boggy Lake record extends from the mid-Holocene (pre-4500 yr B.P.) to the present day. Therefore there is a high probability that Eucalyptus pollen in the fossil record is derived from species still found in the wider region. If the modern pollen of the area can be differentiated to species then it may also be possible to differentiate the fossil Eucalyptus pollen from this and nearby sites, and thus provide more detailed palaeoecological information from these sequences. Brooker and Kleinig (1990) describe ten species of Eucalyptus which may be found today in the area under consideration. O f these, E. calophylla, E. diversicolor and E. marginata are tall trees that dominate forests of the most southwesterly areas of Western Australia. E. guilfoylei and E. jacksonii are tall trees restricted to the Walpole area, as is the small tree E. ficifolia. E. cornuta and E. megacarpa are generally medium-sized trees with a more scattered distribution within the forests. E. patens is a medium to tall tree with a widespread distribution in the southwest. These nine species of Eucalyptus were chosen for the detailed investigation of pollen morphology. One other species, E. occidentalis was omitted due to its limited occurrence in the area. Pollen from the selected taxa was obtained from herbarium specimens at the Western Australian Herbarium and Murdoch University. Material for each species was obtained from a minimum of three specimens, if possible from across the geographical range of the species (sources are listed in Appendix A). Some samples were found to lack developed pollen once processed. Pollen was prepared by treating the macerated anthers according to standard preparation techniques ( K O H treatment and acetolysis) and stained using safranin in aqueous solution prior to dehydration and mounting in Silicone oil (4000 cs). Measurements and observations were made with a light microscope on a minimum of 20 grains per sample. Light photomicrographs were taken at × 1000 magnification using an oil immersion lens. Fossil material was prepared as for modern material but with the addition of H F treatment after acetolysis. This should not seriously affect the grain sizes (Faegri and Iversen, 1989, pp. 78
189
and 234) and modern and fossil material should be comparable.
3. Morphological characters examined
Eucalyptus pollen is typically isopolar, radially symmetric, tricolporate, triangular in polar view and oblate. Grains are usually parasyncolpate with a distinctive apocolpial field, although grains may approach the syncolpate form in some species. The endexine and ectexine layers are usually distinguishable and of similar thickness. The choice of morphological characters for detailed examination was guided largely by previous studies of the Myrtaceae (Pike, 1956; Gadek and Martin, 1981; Martin and Gadek, 1988; Chalson and Martin, 1995). Some of the characteristics used are those defined by Chalson and Martin (1995) and are described in detail there. A further important consideration was that the characters chosen should preserve well enough to be observable in the fossil material. Terminology follows Punt et al. (1994) unless stated otherwise. 3.1. Grain shape The grain walls may be straight, convex or concave in polar view. The amb angle of all Eucalyptus pollen examined was rounded, but additionally could be slightly or distinctly notched (Chalson and Martin, 1995).
3.2. Apocolpial field, apocolpial edges and polar islands (Fig. 1; Chalson and Martin, 1995) The apocolpial field may be arcuate or angular in shape, or intermediate between these. Apocolpial edges may range from smooth through irregular to rough or broken. Presence of ectexine over the apocolpial field was noted, and if present its size, shape and surface patterning were recorded. These close fitting or irregular patches of ectexine are termed polar islands (see, for example, Chalson and Martin, 1995). Grains lacking a polar island may have an unpatterned apocolpial field, or may have granules on this area: the presence of granules was also noted.
190
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 18~205
3.3. Grain size Three measurements were determined for each grain examined (Fig. 1; Chalson and Martin, 1995). The most useful was the size, in polar view, from the pore to the midpoint between pores on the opposite side. This was referred to as the A/B (apex/base) measurement. Also measured were wall thickness at the midpoint between pores in polar view and pore height.
3.4. Pore characteristics (Fig. 2) Eucalyptus pollen typically possesses a distinct vestibulum formed by the splitting of the endexine and ectexine in the region of the pore (Fig. 2). The floor of the vestibulum may be flat, concave or convex, while the roof may be unthickened, slightly thickened or distinctly thickened (Chalson and Martin, 1995). 3.5. Surface pattern A number of surface pattern types were recognised. These were psilate, scabrate (irregular elements less then ca. 1 lam), granulate (regular elements to ca. 1 ~tm), verrucate (regular elements > 1 gm, broader than they are high) and gradations between these. Surface patterning may extend over the whole mesocolpium to the colpial edges, or the margo colpi may be unpatterned. Changes in surface patterning on the margo colpi may be sharp or gradual. These features were also recorded.
Pollen morphology also varied between and to differing extents within species and is summarised in Table 2. Only the most commonly occurring character state is listed if 80% or more of grains examined were of this type. If the most common character state occurred in less than 80% of grains then the percentage is given as is the range of character states encountered (excluding character states occurring in less than 5% of grains). If only a range is given then no single character state was dominant. Details of the morphological study, including percentage of grains having any particular character, are provided in Appendix B. The characteristics of the nine species are described below. 4.1. Eucalyptus calophylla R. Brown ex Lindley
(Fig. 4) This is a large grained Eucalyptus type (Fig. 4), varying in size (A/B) from 21.5 to 36.0 gm. The grain sides are straight to concave in polar view, occasionally convex. The amb is rarely slightly notched at the base of the vestibulum. Grains are parasyncolpate with a relatively small, but well defined, arcuate apocolpial field. Apocolpial edges are usually irregular. The vestibulum floor is flat to convex, while the vestibulum roof varies in thickness and may be unthickened or thickened. Pore height ranges from 2.0 to 4.0 gm. Exine patterning can be granulate/scabrate, granulate/ verrucate or verrucate, but the margo colpi are unpatterned.
4.2. Eucalyptus cornuta Labillardiere (Fig. 4) 4. Characteristics of the modern pollen Most species examined showed some variation in size attributes both within and between samples (Table 1). Mean wall thickness was between 1 and 2 gm for all species and mean pore height ranged from 2.3 to 3.7 Jam. Neither of these characters was particularly useful for separating pollen of the different species. Mean grain size (A/B) appeared more useful with both E. ficifolia and E. calophylla having mean grain sizes greater than the other species examined (Fig. 3).
Grains tend to be either concave or straight triangular in polar view, ranging in overall size from 18.0 to 23.0gm. The amb is frequently slightly or distinctly notched at the base of the vestibulum. Grains are parasyncolpate with a distinct angular or arcuate/angular apocolpial field, with scattered granules in 75% of grains. The apocolpial edges are commonly rough, but may be irregular or broken. The vestibulum floor is usually flat but ranges from flat/convex to concave: the roof is generally thickened or slightly thickened and overall pore height is 3.0 to 4.5 tam. Most
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
A
C
191
B
D
E
Walthi l ckness ~ / ~ _
F
G
GrainsizeA/B
H Fig. 1. Types of apocolpial field, colpial edges and grain measurements. (A) Arcuate apocolpial field. (B) Angular apocolpial field. (C-G) Types of colpial edge: (C) smooth; (D) irregular; (E) rough; (F) and (G) broken. (H) Measurements determined for each grain in polar view.
192
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187205
A
with an arcuate apocolpial field: apocolpial edges range from smooth to irregular. The ectexine is predominantly unthickened over the vestibulum with the floor concave or flat: pore height ranges from 2.0 to 4.0 ~tm. Exine patterning is usually psilate. 4.4. Eucalyptus ficifolia F. Mueller (Fig. 4)
B
This is a large grained Eucalyptus type, ranging in size from 23.0 to 37.5 ~tm. Grains are usually straight triangular in polar view. Grains are typically parasyncolpate with a distinct arcuate apocolpial field; ca. 20% of grains appear to have a small, irregular, loose fitting polar island. Apocolpial edges are predominantly irregular or smooth/ irregular. Pore height ranges from 2.0 to 5.0 ~tm. The vestibulum floor is usually flat or convex: the roof can be unthickened or thickened. Surface pattern is most commonly psilate, ranging to granulate/psilate: the margo colpi are unpatterned. 4.5. Eucalyptus guilfoylei Maiden (Fig. 5)
C
Fig. 2. Types of pore. (A) Floor of vestibulum concave, roof thickened. (B) Floor of vestibulum flat, roof unthickened. (C) Floor of vestibulum convex, roof slightly thickened.
grains are psilate or granulate/psilate. When patterning is present, the margo colpi are unpatterned. 4.3. Eucalyptus diversicolor F. Mueller (Fig. 5)
Grains are straight to straight/convex triangular in polar view and range in size from 18.0 to 24.0 gm (Fig. 5). All grains are parasyncolpate
The grains are usually concave triangular in polar view with grain size ranging from 13.5 to 21.0 ~tm. Usually grains are parasyncolpate with an arcuate apocolpial field although in some grains this is arcuate/angular. Some grains (30%) in one sample have granules on the apocolpial field. The apocolpial edges are usually irregular but range from smooth/irregular to rough. The vestibulum floor is usually flat but can be flat/concave, while the roof is generally unthickened or infrequently slightly thickened. Pore height ranges from 1.5 to 3.5 gm. Exine pattern may be granulate, granulate/psilate or granulate/scabrate: the margo colpi are unpatterned. 4.6. Eucalyptus jacksonii Maiden (Fig. 5)
Grains are generally straight triangular in polar view, but occasionally may be straight/concave: size ranges from 13.0 to 17.0 gm. The apocolpial field can be arcuate or angular, but is most often arcuate: 45% of grains have granules on the apocolpial field. The apocolpial edges are usually rough or irregular/rough. There is a distinct and
193
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187~05 Table 1 Pollen grain size variability in the Eucalyptus species examined. All measurements are in gin Species
Grain size (A/B)
Wall thickness
Pore height
Range
Mean _+SE
Range
Mean _+SE
Range
Mean _+SE
2 all
1.0-1.5 1.5 2.0 1.0 2.0
1.20_+0.06 1.72_+0.04 1.46 -+0.05
2.0-3.5 2.0-4.0 2.0-4.0
2.95_+0.10 3.11-+0.10 3.03_+0.07
26.0-36.0 21.5-26.5 21.5-36.0
29.35_+0.59 23.92-+0.28 26.64_+0.54
1 2 all
0.5-1.5 1.0-2.0 0.5-2.0
1.02_+0.06 1.20_+ 0.07 1.11 _+0.05
3.0-4.5 3.0-4.0 3.0-4.5
3.73_+0.09 3.53-+0.09 3.64-+0.06
18.0 22.0 19.0 23.0 18.0-23.0
19.64+0.24 21.03-+0.23 20.33_+0.19
1
2 3 all
1.0-1.5 1.0-2.0 2.0-2.5 1.0-2.5
1.05_+0.04 1.35_+0.08 2.10_+0.05 1.53_+ 0.06
2.0 2.5 2.0-4.0 2.5-4.0 2.0 4.0
2.18-+0.06 2.85_+0.12 3.38_+0.09 2.82_+0.08
18.0 22.0 19.0 23.0 18.0-24.5 18.0 24.5
19.95-t-0.26 21.03-t-0.24 21.53_+0.35 20.84-t-0.18
E. Jicifolia
1 2 3 all
1.0 2.5 1.0-3.0 1.5-2.5 1.0-3.0
1.83-t-0.09 2.01-+0.10 1.97_+0.07 1.93 _+0.05
2.5-4.5 2.0-5.0 3.5-5.0 2.0-5.0
3.35-+0.11 3.39_+0.16 3.99-+0.11 3.58_+0.08
28.0-33.0 23.0-31.0 27.5-37.5 23.0-37.5
30.03_+0.36 26.68_+0.57 31.90-+0.60 29.53_+0.40
E. guilfoylei
1 2 3 all
0.5 1.5 1.0-2.0 1.0-1.5 0.5-2.0
1.03-+0.05 1.40-+0.07 1.21 _+0.05 1.21 -+0.04
1.5 3.0 2.0 3.0 1.5 3.5 1.5-3.5
2.13_+0.08 2.43_+0.09 2.43+0.10 2.32-+0.05
18.0 21.0 17.0 1 9 . 0 14.0-17.5 14.0-21.0
19.90_+0.20 17.91_+0.17 15.92-+0.21 17.91 -+0.24
E. jacksonii
1 2 all
1.0-1.5 1.0 2.0 1.0 2.0
1.30_+0.04 1.38_+0.06 1.34+0.04
3.5-4.5 3.5-4.5 3.5 4.5
3.66_+0.23 3.83_+0.06 3.75_+0.06
15.0-16.5 13.0 1 7 . 0 13.0-17.0
15.81 _+0.13 15.32_+0.19 15.57_+0.12
E. marginata
1 2 3 all
1.5-2.0 1.0 2.0 1.0-2.0 1.0-2.0
1.56_+0.03 1.63_+0.10 1.27_+0.07 1.48 -+0.05
3.0 4.0 1.5-4.0 2.0 3.5 1.5 4.0
3.37+0.04 2.83_+0.14 2.43+0.08 2.84_+0.08
15.5 1 7 . 5 17.0-21.0 15.5 1 8 . 5 15.5-21.0
16.65-+0.15 18.90_+0.26 17.16_+0.77 17.57_+0.17
1
2 3 all
1.0-2.5 1.0-1.5 1.0-1.5 1.0 2.5
1.79_+0.08 1.12_+0.03 1.23 _+0.04 1.41 _+0.37
2.0-3.0 1.5-3.5 1.5-3.5 1.5-3.5
2.43_+0.09 2.15_+0.09 2.37_+0.09 2.32-+0.05
20.0-23.0 20.0-22.0 17.5 22.0 17.5 23.0
21.62+0.19 21.45+0.16 19.42+0.24 20.84-+0.17
1
1.0 2.0
1.43_+0.08
2.5 3.5
2.98+0.07
19.0-23.0
20.70+0.21
E. calophylla
E. cornuta
E. diversicolor
E. megacarpa
E. patens
Sample
1
l a r g e v e s t i b u l u m in w h i c h t h e f l o o r is flat a n d t h e roof prominently thickened. Pore height ranges b e t w e e n 3.5 a n d 4.5 g m . T h e g r a i n s a r e p a t t e r n e d o n t h e m e s o c o l p i u m e x c e p t f o r t h e m a r g o colpi, a n d u s u a l l y t h e r e is a d i s t i n c t b o u n d a r y b e t w e e n the two. Exine patterning varies from granulate/scabrate to granulate/verrucate.
4. 7. E u c a l y p t u s m a r g i n a t a D o n n e x S m i t h (Fig. 5) G r a i n s v a r y in size b e t w e e n 15.5 a n d 21.0 gin a n d a r e c o n c a v e o r s t r a i g h t t r i a n g u l a r in p o l a r view. G r a i n s are u s u a l l y p a r a s y n c o l p a t e w i t h a n a r c u a t e a p o c o l p i a l field, b u t t h e y c a n s o m e t i m e s be a r c u a t e / a n g u l a r o r m o r e r a r e l y a n g u l a r . A b o u t
194
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187 205 E. ficifolia E. calophylla E. diversicolor E. patens E. megacarpa E. cornuta E. marginata E. guilfoyei E. jacksonii
7. I
a
12
14
I
16
I
I
i
I
I
I
I
i
i
i
18
20
22
24
26
28
30
32
34
36
i
38
i
40
Grain Size A/B (kam) Fig. 3. Range and mean grain sizes (± SE) of the nine Eucalyptus species examined.
40% of grains examined had granules on the apocolpial field. The apocolpial edges are most commonly irregular, but range through to rough. The vestibulum floor is flat or convex, rarely concave and the vestibulum roof is distinctly thickened. The pore height ranges from 1.5 to 4.0 lam. Surface pattern can be granulate/scabrate, granulate/verrucate or verrucate but the margo colpi are unpatterned.
4.8. Eucalyptus megacarpa F. Mueller (Fig. 5) Grains are most commonly straight-sided triangular in polar view but range from convex to straight/concave. Size ranges from 17.5 to 23.0 ~tm. The grains are parasyncolpate. The apocolpial field can be angular or (less commonly) arcuate and is always large, occasionally with granules. The apocolpial edges are predominantly irregular but range from smooth/irregular to irregular/rough. The vestibulum floor tends to be concave to flat but can sometimes be convex, while the roof is usually unthickened or slightly thickened. Pore height varies from 1.5 to 3.5 gm. Exine patterning was found to be very variable between samples and was commonly psilate or granulate/scabrate over the mesocolpium. Some
grains from one sample were granulate/verrucate or verrucate. 4. 9. Eucalyptus patens Bentham (Fig. 5) Grains are straight triangular in polar view, ranging in size from 19.0 to 23.0 gm. They are parasyncolpate with an arcuate apocolpial field and irregular apocolpial edges. The vestibulum floor ranges from flat to concave, and the vestibulum roof is unthickened. The pore height ranges from 2.5 to 3.5 gm. Most grains have a granulate/scabrate surface pattern but are unpatterned on the margo colpi. The above descriptions and Tables 1 and 2 show that intraspecific variability differs in the taxa examined, and that the variability is greater for some characters than for others. This intraspecific variability has previously been noted in Eucalyptus pollen (e.g. Martin and Gadek, 1988) and contributes to the difficulty of separating species on the basis of pollen morphology. All of the characters examined may be variable within some species, and some species seem to show great variation in a number of characters. Eucalyptus megacarpa has particularly variable pollen. Its surface patterning shows a number of
St 55%-Cx
Cv 65%-St
St/Cx 70%-St St/Cx 50% (St Cx) St 73% St/Cx
St
St
St
1
2
1 2
1
2
3
St 60%-St/Cx
St Cv 70% St/Cv
E. marginata
Cv Cv
2 3
St St 64%-St/Cv
Cv
1
3
E. jacksonii
E. guilJbylei
E. ficifolia
E. diversicolor
E. cornuta
2
St 55% (St/Cv-Cx) Cv 45%-St
1
E. calophylla
Wall shape
Morphology
Sample
Species
10% sl
10% sl
75% sl 15% ds 45% sl
20% sl
Amb notch
Table 2 Pollen morphology of the Eucalyptus species examined
Ang 55%Arc/Ang
95% granules
Ir
Ro Ir/Ro 50% (Ir Ro) lr lr/Ro 75% (Ir-Br) Ir
Arc
Arc 75%Arc/Ang Arc Arc 75%Arc/Ang Arc Arc-Ang
Arc-Ang Arc Arc
20% small 10% granules
30% granules
45% granules 45% granules
60% granules 45% granules
Ir 65%-Sm/Ir Ir 55%Ro
Ir
Arc
10% small
Sm/Ir 50% (Sm-Ir/Ro) Ir
Arc
Ir Ir 70% (Sm-Ir/Br) Sm 77% Ir
Ro 55% (Ir/Ro Br) Ir-Br
25% small
Arc
Arc Arc
Arc-Ang
Ir
Arc
55% granules
Ir
Arc
Fl 55% (Cx-Cv)
F1 F1-Cx
FI FI
FI 65% FI/Cv FI
FI
Fl 70% (Cx-Cv) FI 70% (F1/Cv-Cx)
FI-Cx
F1 60%-Fl/Cv
Cv Cv
FI 75% FI/Cx
F1 50% Cv
FI 55% (FI/Cx-FI/Cv) Cx
Floor
Apocolpial edges
Polar Island
Shape
Vestibulum
Apocolpial field
Th
Th Th
Th Th
Unth Unth SI Th
Unth
Th 60%-Unth
Unth-Th
Unth
Unth
Unth Unth
Th-SI Th
Th 60%-S1 Th
Th
Unth
Roof
unpatt
Verr 45%, Gr/Verr 40%, Gr 15% Gr/Verr 60%, Gr/Sc 40%
unpatt unpatt
unpatt unpatt unpatt unpatt unpatt unpatt
unpatt unpatt unpatt
unpatt unpatt
unpatt unpatt unpatt
Ps Gr/Ps
Ps Ps Ps Ps 75%, Gr/Ps 25% Ps 70%-Gr/Sc Ps-Gr/Sc
Gr Gr/Ps Gr/Sc
Gr/Verr-Gr/Sc Gr/Sc 75%, Gr/Verr 25% Verr/Sc Verr Gr 60%-Gr/Sc Gr/Sc 60%-Verr
unpatt
Margo colpi
Type
Pattern
oo
St
Cx 50%-St
St
2
3
l
Amb notch
40% granules
lr
Arc
Ang 55%Arc/Ang, large
Ang, large
Sm/Ir 67% (Sin Ir/Ro) lr 75% (Sm/Ir lr/Ro) Ir
Arc-Ang, large
FI Cv
Cx Cv
F1 70%-Cv
Cv 67%-F1/Cv
Floor
Apocolpial edges
Polar Island
Shape
Vestibulum
Apocolpialfield
Unth
Sl Th Unth
Unth 75% SI Th
Unth
Roof
patt
Margo colpi
Gr/Sc
Ps
unpatt
patt
Ps Verr patt
Gr/Sc
Type
Pattern
Shape: St = straight sides; C x - convex sides; Cv = concave sides. Amb: sl = slightly notched below amb; ds = distinctly notched below amb. Polar island: percentage of grains having polar islands or granules on the apocolpial field. Colpi: Arc = arcuate apocolpium; Ang = angular apocolpium. Colpial edges: Sm = smooth; Ir = irregular; Ro =rough; B r = broken. Vestibulum floor: Fl=ftat; Cx=convex; Cv =concave. Vestibulum roof: S1 Th=slightly thickened; Th=thickened; U n t h = unthickened. Pattern: Ps =psilate; Gr =granulate; Sc = scabrate; Verr =verrucate; p a r t - patterned on the margo colpi; unpatt ~ unpatterned on the margo colpi.
E. patens
St 76% St/Cv
1
Wall shape
E. megacarpa
Morphology
Sample
Species
Table 2 (continued)
t~
4
ao
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
0
197
10 lam
Fig. 4. Eucalyptus pollen. (A, B) E. calophylla and (C, D) E. ficifolia showing the small apocolpial field and relatively large grain size in these two species. (E, F) E. cornuta showing slightly notched amb and thickening of the vestibulum roof.
-r
"1"1
m
t~
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
distinct forms with psilate, granulate/scabrate, granulate, granulate/verrucate and distinctly verrucate grains being found. However, patterning, when present, occurs over the whole of the mesocolpium which is unusual in pollen from the other species. Other taxa also exhibit some variation in surface patterning; for example E. marginata surface patterns range from granulate/scabrate, through intermediate types, to verrucate, but here there seems to be much more of a gradation of characters than in E. megacarpa. Some species examined show a high degree of intraspecific constancy in morphology, e.g.E, diversicolor, and can more readily be characterised.
5. Definition of pollen types The variation within the taxa examined and the resultant overlap in characters between species was such that the species could not generally be separated on the basis of a single character. However, some success was achieved by using a suite of characters to define and separate the pollen taxa. In defining character suites for each species or taxon the full range of characters exhibited was considered, except that those which occurred in 10% or less of grains were not regarded as typical for that species, and were excluded from the analysis. Including infrequently occurring or atypical characters (often only occurring in 1-2% of grains), would widen the character ranges of taxa such that separation of species or groups would be more difficult or impossible. Since a suite of characters was used to define pollen, and the likelihood of a grain exhibiting atypical characters across the whole spectrum of features examined is very small, exclusion of these atypical characters would only marginally increase the risk of misallocating pollen to a group. Using the characters shown in Tables 1 and 2
199
and Appendix B a number of pollen types were defined. The most important attributes which were used to define each type are summarised in Table 3. 5.1. E. calophylla type This includes pollen from both E. calophylla and E. ficifol&. These have large pollen which can be differentiated from most grains of other taxa on the basis of grain size, A/B (Fig. 3). N o grains of E. ficifolia, and very few (less than 8%) of E. calophylla were less than 23 jam in size. Small (23-24 jam), psilate grains of E. ficifol& could potentially be confused with the low percentage of E. diversicolor pollen having grains of this size unless they also had a distinctly thickened vestibulum roof. E. calophylla and E. ficifolia overlap considerably with each other in most characters, including grain size, except that E. ficifolia pollen may be psilate whereas E. ¢alophylla pollen is not. In practice the two were grouped together as E. calophylla type. This was defined as all patterned grains greater than 23 jam, psilate grains greater than 24 jam, and psilate grains greater than 23 jam if also having a distinctly thickened vestibulum roof. 5.2. E. marginata type
This includes pollen from both E. marginata and E. jacksonii. Both species can be differentiated from most other taxa on the basis of surface pattern (ranging from granulate through to verrucate), thickened vestibulum roof, irregular to rough apocolpial edges and grain size ( < 17 jam in E. jacksonii and <21 jam in E. marginata). There is some overlap of characters between the two species. E. jacksonii tends to be smaller, have rougher edges to the apocolpia, consistently has a flat, not convex floor to the vestibulum and is not usually scabrate/verrucate in surface pattern.
Fig. 5. Eucalyptus pollen. (A, B) E. guilfoylei showing irregular/rough apocolpial edges. (C) E. diversicolor showing psilate grain surface and arcuate apocolpial field. (D, E) E. jacksonii showing large vestibulum with a fiat floor and prominently thickened roof. (F, G) E. marginata showing concavegrain with distinctlythickened vestibulum roof, granular/verrucatepatterning on the mesocolpium and unpatterned margo colpi. (H, I ) E. patens showingarcuate apocolpialfieldwith irregularedges, granular/scabratepatterning on the mesocolpiumand unpatterned margo colpi. (J, K) E. rnegacarpa showing large arcuate/angular apocolpial field, unthickened vestibulum roof and psilate grain surface.
200
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
Table 3 Summary of the pollen attributes used to define Eucalyptus pollen types E. marginata
E. cornuta
E. guilfoylei
E. megacarpa
E. diversicolor
E. patens
E. calophylla
Size A/B
13.0-21.0
18.0-23.0
14.0-21.0
17.5-23.0
19.0-23.0
Vestibulum roof Surface pattern
Th Gr-Sc-Verr
SI Th-Th Ps-GrlPs
Uuth (-SI Th) GriPs, Gr, GrISc
Ps-Gr/Sc
18.0-24.5 usually < 23 Unth Ps
21.5-37.5 usually > 23 Unth-Th Ps-GrlSc, Grl Verr- Verr
Margo c o l p i Arab Apocolpial field
unpatterned
unpatterned
unpatterned
patterned
unpatterned
unpatterned unpatterned
Arc Arc/Ang
Ang (Arc)
Arc
Arc
Arc
Smllr-lrlRo
large Smllr-IrlRo
Cx St
Sm-lr St-StICx
lr
Cv
Sm Ir St Cv
Unth (-SI Th)
Unth GrISc
sl notched (60%)
Arc-Arc/Ang A r c i n g granules (75%) Apocolpial edges Ir-Ro Ir Br Grain shape St-Cv Cx-Cv
St (St/Cx)
The most important attributes are highlighted. Abbreviations are as in Table 2.
However, it was felt that the two species would be difficult to separate in fossil material and in practice would best be considered together as a single type.
grains have a vestibulum r o o f which is thickened whereas in E. g u i l J o y l e i the vestibulum roof is unthickened in the majority of grains.
5.5. E. cornuta t y p e 5.3. E. diversicolor t y p e This type can be differentiated from others on the basis of wall shape (straight to straight/ convex), smooth to irregular apocolpial edges, psilate surface pattern and unthickened vestibulum roof. The grain size is defined as being less than 23 gm to avoid confusion with psilate grains of E. f i c i f o l i a as described above. Some grains of the variable E. m e g a c a r p a m a y also be very similar but the large, frequently angular apocolpial field in the latter should allow the types to be distinguished.
5.4. E. guilfoylei t y p e This type can be distinguished from others on the basis of its concave wall shape, smooth/ irregular to irregular/rough apocolpial edges, granulate, granulate/psilate or granulate/scabrate surface pattern, unthickened to slightly thickened vestibulum roof, and grain size (<21 gin). It is quite similar to E. c o r n u t a but the latter frequently has a slightly notched amb, and the majority of
This type can be distinguished on the basis of psilate to granulate/psilate surface pattern, thickened or slightly thickened vestibulum r o o f and grain size (18-23 gm). The potential overlap with E. g u i l f o y l e i is described above. Some unpatterned grains of the rather variable E. m e g a c a r p a may be similar, but should be distinguishable on the basis of their large apocolpial field. Additionally E. c o r n u t a exhibits a slightly notched amb and thickened vestibulum r o o f in a majority of grains, whereas E. m e g a c a r p a is unnotched with an unthickened or only slightly thickened vestibulum roof. It is possible that a small proportion of grains of E. g u i l f o y l e i or E. c o r n u t a may not be satisfactorily distinguished to one or other of these types.
5.6. E. megacarpa t y p e This variable type can be distinguished from most others on the basis of smooth/irregular to irregular/rough apocolpial edges, presence of a large apocolpial field and patterning, when present,
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
on the whole of the mesocolpium. The potential problems with E. diversicolor, E. guilfoylei and E. cornuta are discussed above. 5. 7. E. patens type This type can be separated on the basis of straight walls, unthickened vestibulum roof, irregular colpial/apocolpial edges, granular/scabrate surface patterning with unpatterned margo colpi and grain size (< 23 ~tm). There is overlap in a number of characters with E. guilfoylei, but grain shape is distinctly different in the two types.
6. Eucalyptus pollen in the fossil record The character suites used to define the groups of contemporary Eucalyptus pollen were applied to fossil material in the Boggy Lake sequence. A significant proportion of grains, up to 50% per sample, could not be allocated to a particular Eucalyptus type. The most common problem was that the grain characteristics could not be sufficiently well observed due to poor preservation, folding or crumpling of the grain, or partial obscuration by matrix material. However, some grains, even when clearly observed, did not fit readily into one of the Eucalyptus groups and therefore were not assigned to one. Full details of the pollen record from Boggy Lake are presented elsewhere (Newsome and Pickett, 1993). In summary, the contribution of Eucalyptus pollen to the terrestrial pollen sum remains fairly constant throughout the sequence except for the two basal samples. This suggests that the overall importance of Eucalyptus in the vegetation has not changed significantly during the last ca. 4500 yr. Much of the remaining terrestrial pollen is probably quite local in origin and includes taxa such as Agonis type and Banksia type. Examination of the Eucalyptus pollen curves (Fig. 6) shows that there are marked fluctuations in the contribution of different Eucalyptus types to the total Eucalyptus pollen sum. Of note are changes in the proportion of E. marginata type and E. megacarpa type which are most common in the upper part of the sequence, and E. diversico-
201
lor type which is most strongly represented in the lower part of the sequence. E. calophylla type is rather variable throughout the sequence, and is not recorded at all in some sections. E. guilfoylei type and E. cornuta type are also rather variably represented, while E. patens type is not recorded. Initial studies by Pickett (1990) and Newsome and Pickett (1993) showed E. jacksonii as a distinctive type from E. marginata. However, this was based on a small sample of E. jacksonii pollen grains which is now considered to have been composed largely of aberrant grains. Grains of this type are included in the Eucalyptus undifferentiated curve in Fig. 6. The changes in the proportions of the Eucalyptus types throughout the sequence do not appear to be random, as trends of increasing or decreasing abundance over a series of samples are usually evident. However, in any interpretation of these changes, the large proportion of Eucalyptus grains which could not be allocated to a group is a cause for concern unless it is assumed that all Eucalyptus pollen suffers equally from degradation processes. These changes in the proportions of Eucalyptus pollen types are likely to be a reflection of changes in the composition of the surrounding vegetation, confirming that there have been changes in the composition of the Eucalyptus woodlands and forests in the region around the site during the latter part of the Holocene. Interpretation of the causes of the vegetation changes are, however, difficult. Some changes are in proportions of species typical of higher rainfall regions (e.g.E. diversicolor type, E. guilfoylei type) relative to species with wider ranges (e.g.E. marginata type) but it is questionable to draw climatic conclusions solely on the basis of these, particularly when types such as E. marginata may in fact include pollen of E. jacksonii, which is also found only in high rainfall areas. The character suites used to define pollen types in this investigation are not identical to those suggested by Churchill (1956) in his pollen key. Churchill (1956) emphasised that size differences could be used to differentiate a number of taxa, however, the current study found that there was too much overlap in size for it to be useful alone, except possibly for separation of E. calophylla and
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
202
Dep~ (cm)
o,':' /
/
/
/
Suct'aeo 1 ~ 2 0-
I I 100-
200-
1'o
~o~
Pollen sum: Total £ac~pm~ pollen
Fig. 6. Eucalyptus p o l l e n d i a g r a m f r o m B o g g y L a k e .
E. ficifolia from the other types. Some of the other character states suggested by Churchill (1956) concur with those found here but others are in conflict. He describes a number of characters as being species constant whereas the current study has revealed a greater amount of intraspecific variability. The differences may be partly a reflection of smaller pollen samples studied by Churchill (1956). Another concern is that Churchill (1956) apparently identified all of his fossil Eucalyptus pollen to species. This may have been a result of 'forcing' grains to fit the alternate choices in the dichotomous key. The current study, while showing that Eucalyptus pollen can be separated using suites of characters, was unable to replicate these results and a large proportion of grains always remained undifferentiated. This study therefore reinforces doubts in the
widely quoted palaeoclimatic interpretations of Churchill's earlier work (1968) since it fails to support the species separations which are the basis of those palaeoclimatic interpretations (see also Newsome and Pickett, 1993).
7. Discussion
The area centred on Walpole in the southwest of Western Australia supports ten Eucalyptus species, a situation which is unlikely to have changed much during the Holocene. The Eucalyptus species currently present in the area have pollen which is sufficiently distinctive to allow them to be differentiated to pollen types with some confidence, so long as suites of morphological characters, rather than single characters are used. The most useful characters found were size
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205
(A/B), characteristics of the apocolpial field including shape, size, and edges, surface patterning and its distribution on the mesocolpium and margo colpi, and the amount of thickening of the vestibulum roof in the pore. These characters could prove useful in separating Eucalyptus pollen in other areas. They may also be worth further investigation in other genera within the Myrtaceae family. Application of these characters to fossil Eucalyptus pollen from the Holocene sequence at Boggy Lake showed that much of the pollen could be allocated to a pollen type if the necessary features were preserved in the pollen grain. However, some grains do not fit satisfactorily into the defined categories. There are some limitations to this method of separating Eucalyptus pollen. Clearly the details of this study are restricted to the Walpole area of Western Australia, and temporal applicability is also limited to the Holocene, during which species present in the area are unlikely to have been greatly different from those present today. For other applications, attempts to separate Eucalyptus pollen into species types would be difficult, if not impossible, in areas having large numbers of species or for fossil sequences where major environmental shifts have occurred which could have been accompanied by significant changes in the actual
203
Eucalyptus species occurring. The method has potential uses in studies of the Holocene and could be very informative in high resolution studies where the interest lies in forest or woodland dynamics. A basic pre-requisite is that the area under study has a relatively small number of Eucalyptus species, and that these can be separated, to a satisfactory extent, on the basis of pollen morphology. Acknowledgements We wish to thank the staff of the Western Australian Herbarium for pollen reference material samples, the Western Australian Department of Conservation and Land Management for permission to work in Walpole-Nornalup National Park, and colleagues at the University of Western Australia for their ongoing critical interest in the project. Gordon Thomson provided advice and assistance with the photographs. The study was funded in part by a grant from the Australian Research Council, and part of the work was done while J.C. Newsome held a University Postdoctoral Research Fellowship at the University of Western Australia. We would also like to thank the two referees for their constructive comments on the original draft of the paper.
Appendix A Sources of the pollen mater&lexamined Species
Sample
Grains
Source
(#) E. calophylla
1 2
20 20
Middle Swan, coll. P. Ladd 1986. MU Canning River Foreshore, coll. M.L. Clark 1976. WA Herbarium 01319280
E. cornuta
1
20
West of Mt Manypeaks, coll. G.J. Keighery 1986. WA Herbarium
2
20
Crystal-Boggy Lake near Walpole, coll. J.W. Green 1956. WA Herbarium
1
20 20 22
Augusta, near Jewel Cave, coll. G.G. Smith 1961. WA Herbarium Karibul Gully, Mt Hopkins, south of Walpole, coll. D.M. Churchill 1956. WA Herbarium North of Blow Hole Dr. Torndirrup National Park, coll. G.J. Keighery and J.J. Alford 1987. WA Herbarium 01337017
20 20 20
Soho Block, Denmark, coll. Wildflower Group 1988. WA Herbarium Peaceful Bay, east of Walpole, coll. A.S. George 1973. WA Herbarium 01098918 Ficifolia Rd., 8 km ESE of Nornalup, coll. A. Strid 1982. WA Herbarium 01097857
E. diversicolor
2 3
~c/fo~
E.J. Pickett, J. C Newsome / Review o f Palaeobotany and Palynology 98 (1997) 18~205
204
20 20 20
N W of Nornalup, 1982. W A Herbarium Coll. C.A. Gardner 1974. W A Herbarium 01062999 Deep River, SW of Walpole, coll. J.W. Green 1956. WA Herbarium
1
20
2
20
Vicinity of Howe Road and Valley of the Giants Rd. junction, coll. P.R. Mawson 1984. MU(CSIRO) 0.9 k m along Tinglewood Lodge Road, west of Walpole, coll. P.R. M a w s o n 1984. M U ( C S I R O )
1
20 20 20
4 k m N of Margaret River townsite, det M.I.H. Brooker 1984. M U ( C S I R O ) M u r d o c h University Campus, coll. P. Ladd 1986. M U 3 km E N E of Mt Lesueur, N E of Jurien, coll. E.A. Griffin 1975. W A Herbarium 01391798
2 3
21 20 20
Scotts River Plains, coll. P. Ladd 1986. M U N o r m a n ' s Inlet, 40 k m E of Albany, coll. G.J. Keighery 1986. W A Herbarium 01428179 Mt Frankland, 29 km N of Walpole, coll. B.J. Conn and J.A. Scott 1989. W A Herbarium 01133152
1
20
Near K a n g a r o o Creek, Brookton Highway, coll. P. Ladd 1986. M U
E. guilJbylei
E. jacksonii
E. marginata
2 3
E. megacarpa
E. patens
1
M U = M u r d o c h University; M U ( C S I R O ) = Division of Forest Research specimens at Murdoch University.
Appendix B Percentage of grains exhibiting each character attribute E. calophylla Shape:
Amb: Apocolpial field:
Apocolpial edges:
Vestibulum floor:
Vestibulum roof:
convex straight/convex straight straight/concave concave notched slightly notched polar island granules arcuate arcuate/angular angular large smooth smooth/irregular irregular irregular/rough rough rough/broken broken concave flat/concave flat flat/convex convex thickened slightly thickened unthickened
5.0 10.0 40.0 22.5 22.5 7.5
100.0
7.5 10.0 82.5
7.5 27.5 25.0 40.0 42.5 7.5 50.0
E. E. E. E. E. E. cornuta diversicolor ficifolia guilfoylei jacksonii marginata 17.5 5.0 37.5 7.5 32.5 7.5 60.0 75.0 20.0 32.5 47.5
12.5 15.0 40.0 20.0 12.5 2.5 20.0 62.5 15.0 55.0 40.0 5.0
3.2 46.8 48.4 1.6
96.8 3.2
3.3 83.3 11.7 1.7
18.3 8.3 100.0
32.3 11.3 53.2 3.2
15.0 18.3 63.4 3.3
56.5 24.2 19.3
5.0 6.7 56.7 13.3 18.3 30.0 16.7 53.3
4.8 95.2
1.7 10.0 88.3
82.5 17.5
5.0 8.3 51.7 11.7 23.3
5.0
5.0
11.7 80.0 18.3 1.7
47.5 57.5 32.5 10.0
36.7 70.0 18.3 11.7
11.7 70.0 15.0 3.3
2.5 32.5 65.0
2.5 15.0 80.0 3.3 1.7 1.7 16.7 79.9
97.5
100.0
5.0 61.7 28.3
5.0 8.3 1.7 50.0 18.3 21.7 93.3 5.0 1.7
E. megacarpa 16.4 8.2 65.6 9.8
13.1 18.0 13.1 68.9 100.0 1.6 26.3 59.0 11.5 1.6
42.6 11.5 36.1
E. patens
20.0 80.0
100.0
15.0 85.0
9.8
35.0 40.0 15.0 5.0 5.0
26.2 73.8
5.0 95.0
E.J. Pickett, J. C. Newsome / Review of Palaeobotany and Palynology 98 (1997) 187-205 Surface pattern:
psilate granulate/psilate granulate 7.5 granulate/scabratepsilate granulate/scabrate 22.5 scabrate scabrate/verrucate granulate/verrucate 47.5 verrucate 22.5 Pattern, if mescolpium, not 100.0 present, on: margo colpi mesocolpium and margo colpi
40.0 47.5 2.5
95.2 4.8
60.0 8.3 1.7 13.3
1.7 35.0 31.7
2.5
20.0
4.9 11.5
10.0
16.7
31.6
60.0
31.7 1.7 19.9 10.0 16.7 100.0
31.1
90.0
10.0 100.0
37.7
37.5 100.0
205
100.0
100.0
100.0
6.6 8.2 100.0 100.0
References Backhouse, J., 1993. Holocene vegetation and climate record from Barker Swamp, Rottnest Island, Western Australia. J. R. Soc. West. Aust. 76, 53-61. Beard, J.S., 1981. Vegetation Survey of Western Australia-Swan. Explanatory notes to sheet 7. Univ. Western Australia Press. Brooker, M.I.H., Kleinig, D.A., 1990. Field guide to eucalypts. Southwestern and Southern Australia. Inkata Press, Melbourne, 428 pp. Chalson, J.M., Martin, H.A., 1995. 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. Churchill, D.M., 1956. An investigation of some pollen-bearing sediments from South Western Australia. Unpublished Hons. Thesis, Univ. Western Australia. Churchill, D.M., 1968. The distribution and prehistory of Eucalyptus diversicolor F. MUELL., E. marginata Donn EX SM., E. calophylla R. BR., in relation to rainfall. Aust. J. Bot. 16, 125 151. Dodson, J.R., 1974. Vegetation history and water fluctuations at Lake Leake, south-eastern South Australia. I. 10 000 BP to present. Aust. J. Bot. 22, 719 741. Dodson, J.R., 1977. Late Quaternary palaeoecology of Wyrie Swamp, southeastern South Australia. Quat. Res. 8, 97 114. Dodson, J.R., Wilson, I.B., 1975. Past and present vegetation of Marshes Swamp in south-eastern South Australia. Aust. J. Bot. 23, 123-150. Faegri, K., Iversen, J., 1989. Textbook of Pollen Analysis. IV edition by K. Faegri, P.E. Kaland and K. Krzywinski. Wiley, Chichester, 328 pp. Gadek, P.A., Martin, H.A., 1981. Pollen morphology in the subtribe Metrosiderinae of the Leptospermoideae (Myrtaceae) and its taxonomic significance. Aust. J. Bot. 29, 159 184. Hopkins, A.J.M., Keighery, G.J., Marchant, N.G., 1983. Species-rich uplands of south-western Australia. Proc. Ecol. Soc. Aust. 12, 15-26.
Hopper, S.D., 1992. Patterns of plant diversity at the population and species level in south-west Australian mediterranean ecosystems. In: Hobbs, R.J. (Ed.), Biodiversity of Mediterranean Ecosystems in Australia. Surrey Beatty, pp. 27 46. Ladd, P.G., 1979a. A Holocene vegetation record from the eastern side of Wilson's Promontory, Victoria. New Phytol. 82, 265-276. Ladd, P.G., 1979b. A short pollen diagram from rainforest in highland eastern Victoria. Aust. J. Ecol. 4, 229-237. Marchant, N.G., 1973. Species diversity in the Southwestern flora. J. R. Soc. West. Aust. 56, 23 30. Martin, H.A., Gadek, P.A., 1988. Identification of Eucalyptus spathulata pollen and its presence in the fossil record. Mem. Assoc. Australas. Palaeontol. 5, 311-327. Newsome, J.C., Pickett, E.J., 1993. Palynology and palaeoclimatic implications of two Holocene sequences from southwestern Australia. Palaeogeogr., Palaeoclimatol., Palaeoecol. 101,245 261. O'Connor, A., 1986. An evaluation of a diatom assemblage as a biological indicator of palaeoenvironmental change in southwest Western Australia. Hons. Thesis, Univ. Western Australia (unpubl.). Pickett, E.J., 1990. A palynological investigation of Holocene environments for southwestern Australia. Hons. Thesis, Univ. Western Australia (unpubl.). Pike, K.M., 1956. Pollen morphology of Myrtaceae from the south-west Pacific area. Aust. J. Bot. 4, 15-53. Punt, W., Blackmore, S., Nilsson, S., Le Thomas, A., 1994. Glossary of Pollen and Spore Terminology. Lab. Palaeobot. Palynol. Contrib. Ser. 1, LPP Foundation, Utrecht. Wardell-Johnson, G., Inions, G., Annels, A., 1989. A floristic classification of the Walpole-Nornalup National Park, Western Australia. Forest. Ecol. Manage. 28, 259-279. Wasson, R.J., Donnelly, T.H., 1991. Palaeoclimatic reconstructions for the last 30,000 years in Australia a contribution to prediction of future climate. CSIRO Div. Water Resour. Tech. Memo. 91/3.