| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Invited Special Paper |
Ecosystem Management, University of New England, Armidale, NSW, 2351 Australia
Received for publication April 19, 2004. Accepted for publication February 15, 2005.
ABSTRACT
Rainforests in tropical Australia occupy a very small, discontinuous area (<1% of the continent), yet they are floristically diverse (c. 2800 vascular species) with high endemicity. There is a distinctive Gondwanan and autochthonous element, and some of the world's ancestral links to the basal angiosperms are uniquely found here. The rainforests can be evergreen or deciduous, but there is a distinct dry season with intermittent drought years. With these characters, the evolutionary pressures on species may be very different to that experienced elsewhere. Sexual systems of 1113 tree species (83 families) from northern Australia were compared with published accounts from the paleo- and neotropics. Hermaphroditic systems dominated all tree floras, and within all floras but Australia dioecy was the most common unisexual system. In tropical Australia, however, significantly more monoecy than dioecy occurred at landscape and community levels. Incorporating phylogeny revealed that sex and fruit types are significantly clustered. The Euphorbiaceae and Sapindaceae contributed c. 50% of the monoecious taxa. Inefficient pollinators (e.g., beetles) may have favored the maintenance of monoecy at the expense of dioecy in the Australian tropics although <1% of the flora has been studied for pollinators and none of the monoecious tree species.
Key Words: Australian rainforest • meta-analysis • monoecy • phylogenetic conservatism • pollinator efficiency • supertree
The causes and maintenance of high species diversity in tropical rainforests have been a focus in the study of plant sexual systems (e.g., Ashton, 1969
; Bawa, 1992
; Stacy, 2001
). One approach to understanding the evolution and distribution of plant sexual systems has been to compare and contrast the types of systems presented in different floras. This work has mostly concentrated on woody species in the neotropics (see studies cited in Arroyo and Uslar, 1993
; Chazdon et al., 2003
; see Table 2), and from this a strong pattern has emerged that the majority of species have hermaphrodite flowers, fewer species have dioecious and even fewer have monoecious systems.
|
) although it is important to consider that a skew toward dioecy in a local flora may in fact reflect the level of dioecy in the pool of families as much as, or more than local selective factors (Renner and Ricklefs, 1995
).
The reproductive ecology of the Australian tropical flora however, is poorly understood with <15 studies in which sexual systems and floral visitors have been examined (see Table 4). Although species diversity on a per site basis may not be higher in Australian rainforests than in rainforests of southeast Asia or some neotropical regions (see Gentry, 1988
), the Australian tropical rainforest flora (north of 19° S) has certain attributes that in concert distinguish it from elsewhere, such that currently accepted generalizations about life history characters (e.g., sexual systems) may not apply. For example, the climate experienced by rainforests in northern Australia is more variable than most equatorial rainforests because there is a distinct dry season with intermittent drought periods—this combination of water stress may have a major influence on flowering behavior (e.g., see Numata et al., 2003
; Borchert et al., 2004
). The flora has a distinctive Gondwanan and autochthonous element that is exemplified with many unique representatives (e.g., species from the Myrtaceae and Proteaceae). Moreover, there is a high representation of relictual angiosperms (e.g., Austrobaileyales, Magnoliales, Laurales, Canellales, Piperales). The uniqueness of the species assemblage is also reflected by high generic diversity with c. 1060 vascular genera (with 66% monotypism) and over 2800 species of rainforest plants of which more than half are endemic to Australia (QRSHERB.DBF, 2003
).
|
). Possible explanations are that the Australian rainforests, as ancient remnants, have lost species and/or failed to speciate, resulting in monotypic genera. On the other hand, some genera are especially widespread and species-rich in northern Australia (Hyland et al., 2003
; e.g., 46 spp. in Syzygium [Myrtaceae], 33 spp. in Cryptocarya [Lauraceae], 22 spp. in Elaeocarpus [Elaeocarpaceae]). Hypotheses explaining mechanisms of speciation in the Australian tropics require careful evaluation of pollination mechanisms. A useful starting point would be to examine the types of sexual systems presented in the flora and their phylogenetic distribution followed by detailed studies on monophyletic groups, which are likely to be more informative than broad studies (see Renner and Ricklefs, 1995
). In this study, the sexual systems and fruit types of the tree flora of northern Australia are examined at the landscape and population levels and compared with other tropical biomes as a precursor to more detailed studies. MATERIALS AND METHODS
Rainforest types in Australia
Rainforests in the Australian context, and as summarized by Adam (1992)
, are comprised of a forest with a closed canopy (70–100% foliage cover, although some species may be deciduous), in contrast with the more extensive Eucalyptus-dominated sclerophyll forests and woodlands, which have more open canopies and pendant leaves. Rainforest types in Australia are usually further divided by geographical position (e.g., tropical, subtropical, warm temperate, cool temperate) or more usefully by structural types (e.g., mesophyll vine forest, after Webb, 1978
and see Adam, 1992
). In this study the term tropical rainforest is used broadly to cover many different structural types of rainforest (see Webb, 1978
) found north of the Tropic of Capricorn. The subtropical rainforests referred to in this study are lowland closed-canopy forests found south of the Tropic of Capricorn, but not beyond southern New South Wales. The terms dry rainforest or monsoonal forest (the vine forests of Webb, 1978
) are used by some authors to describe deciduous and semi-evergreen vine thickets found in tropical and subtropical areas of Australia. Adam (1992)
estimates that there is c. 2 million hectares of rainforest remaining in Australia—this is an area less than 1% of the continent. Further discussion of rainforests types and composition can be found in Adam (1992)
.
Australian biome data
Sex system information was collected for 1113 tree species in the rainforest systems in tropical northern Australia (North of 19° S across Queensland, [Qld], Northern Territory [NT] and Western Australia [WA] Appendix 1, see Supplemental Data with online version of this article). The initial list of tree species was compiled from Hyland et al. (1999)
, with additional species being added later using Hyland et al. (2003)
and expert opinion (Appendix I). Each species was determined as native or endemic to Australia using Hyland et al. (2003)
. Introduced species (c. 32) were not considered further. The type of sex system for each species was determined from field and herbarium collections, published taxonomic treatments, expert opinion, or from the Flora of Australia series, Morley and Toelken (1983)
, or Hyland et al. (2003)
. All species were classified into one of three sex systems: hermaphroditic, dioecious, or monoecious. The hermaphrodite category includes all flowers bisexual on a plant (hermaphroditic, ;mf) including all monostylous and heterostylous hermaphroditic species. In the dioecious category, female (
) and male (
) flowers are found on separate individuals and the category includes androdioecious species. In the monoecious category, male and female flowers are found as separate flowers on the same individual. Andromonoecious and polygamomonoecious (bisexual as well as unisexual staminate and pistillate flowers on the same individual) species are included in this category. For some species, the type of unisexual system was poorly known or there was disagreement in the literature about the type of sex system. This occurs, for example, when a species is poorly represented by fertile collections, and temporal monoecy is mistaken for dioecy. It can also occur when flowers are cryptically polymorphic, e.g., when morphologically monomorphic flowers facilitate mistake pollination, but the flowers are actually cryptically dimorphic (see Cox, 1988
). It more often occurs because many species exhibit a continuum between sexual states (e.g., Lloyd, 1972
), a phenomenon recognized by Darwin (1877)
and studied by many (see Richards, 1986
for a review). The sex system of species, when unclear in the literature, was clarified in most cases with herbarium or fresh-population collections.
Fruit type can be correlated with sex system, particularly the dioecious habit (e.g., Bawa, 1980
) but correlations between these traits can reflect shared evolutionary history (discussed later) or ecological convergence. To look for any correlation between fruit type and sexual system, data about fruit types (dry-indehiscent, dry-dehiscent, fleshy) were collected for the Australian data and analyzed with sex system using 
2 contingency analyses and phylogenetic clustering techniques (discussed later). Fruit information was derived from the Flora of Australia series, Morley and Toelken (1983)
, taxonomic treatments, or from Hyland et al. (2003)
.
Australian population data
Upland and littoral communities
Species composition and abundance data for five sites of remnant rainforest (complex notophyll vine forest on basalt) on the Atherton Tableland (690–800 m a.s.l.), and five sites of rainforest on coastal siliceous sands (mesophyll vine forests on coastal sands) between Cape Tribulation and Cardwell (<100 m a.s.l.), north Queensland, Australia, were collated from detailed reports on floristics by Hopkins et al. (1996
, 1999)
. The first five sites in each report were used. In each site Hopkins and workers (Hopkins et al., 1996
, 1999
) used 20 x 20 m quadrats to record composition and abundance class. They used a semiquantitative scale of 0–3, where 0 = species absent; 1 = occasional, individuals present but uncommon, constituting less than 5% of the total number of individuals in a stratum and less than 5% of total importance in terms of basal area or canopy cover; 2 = common, individuals common, constituting 5–25% of the total number of stems or 5–25% of the importance in terms of basal area or canopy cover; and 3 = dominant, individuals numerically or structurally dominant in a stratum and constituting more than 25% of the importance in terms of basal area or canopy cover irrespective of the number of individuals. For each quadrat, I collated abundance classes and allocated a sex system (see earlier) to each tree species (excluding palms and other monocots and gymnosperms). Thus the data set consisted of five sites from each of two different rainforest communities and data from each site comprised tree floristics, sex system type, and abundance for a total of 0.2 ha.
Australian data
Other studies
Two additional Australian data sets were included for comparative purposes. Hansman (2001)
, in a study of floral biology of six sites within tropical dry rainforest and savanna communities in north Queensland, provided a list of sexual systems for 160 dry rainforest species including 56 tree species. A further five species of tree are listed without allocation to a sexual system. The 61 tree species and their sexual system were collated (and unknown sex system information was augmented from the Australian biome data set—see previous section). Russell-Smith and Lee (1992)
collected sexual system data for 84 tree species as part of an assessment of patch characteristics for 394 monsoonal rainforest patches (<2.5-ha patches). They did not distinguish between hermaphrodites and monoecious species but combined these species into a class called cosexual. Their work is included here for comparative purposes.
Data sets from other tropical regions
A comparison was made of sexual systems in trees from the tropics of northern Australia (see earlier) with published accounts of tree data from the neotropics—Florida (Tomlinson, 1974
), Mexico (Bullock, 1985
), the Caribbean (Flores and Schemske, 1984
), Costa Rica (Bawa, 1974
; Bawa et al., 1985
), Colombia (van Dulmen, 2001
), seven dry forests in central America (Gillespie, 1999
, includes shrubs) and Venezuela (Sobrevila and Kalin-Arroyo, 1982
), and with three old world communities, Malaysia (Ashton, 1969
), New Caledonia (Carpenter et al., 2003
) and Nigeria (Jones, 1955
and see Bawa and Opler, 1975
). Data from an Australian subtropical rainforest (Adam and Williams, 2001
) were also included.
Data analyses
Approach
Increasingly, phylogenetic independent contrasts (PICs) are being used to correct for the lack of independence in the analysis of life history traits and correlations between two or more characters. In very few tropical studies, however, have PIC analyses been used (e.g., Carpenter et al., 2003
; Chazdon et al., 2003
) to investigate trait correlation among species, although this is becoming more feasible with phylogenetic trees derived from molecular studies providing a framework for mapping character traits. Other methods attempt to minimize non-independence by working at the generic or family level. A contrary viewpoint is that phylogenetic correction is sometimes unnecessary, e.g., when trait abundance is being compared and contrasted among floras in attempts to examine adaptive evolution (Westoby et al., 1995
). In addition, when a large data set (one that is comprised of diverse species across a wide range of genera and families) is used, the differences among genera and families can account for a relatively large proportion of the total variation in traits (e.g., Carpenter et al., 2003
), making a PIC analysis unlikely to alter conclusions from analyses of independent observations (TIPs) where strong correlations have been found (Weathers and Siegel, 1995
). With these constraints and viewpoints in mind, PICs were used to investigate the distribution of characters against phylogeny within the Australian database but not among floras. Here species were treated as TIPs. TIP analysis was also used for the analysis of fruit types and sex systems, but in an attempt to minimize the effect of phylogeny, a separate analysis was undertaken by allocating the number of evolutionary events (EEs) (see Thomson and Brunet, 1990
) rather than species to a sexual system and fruit type. For example, a genus with 12 species comprised of eight hermaphrodites, three dioecious, and one monoecious species is given one datum point for each sexual system. This approach, however, may also overstate the number of traits because one of the EEs attributed to each genus is likely to be a pre-existing character. The two approaches are compared here.
With the Australian community data (Atherton and littoral sites), the counts of each sexual system were collated for each site, and a 2 contingency analysis was used to explore differences among the five sites within both community types. The distribution of sex systems was independent of site location for both Atherton (2 = 5.63, P = 0.68) and the littoral community (2 = 5.19, P = 0.74). The abundance class data from the five quadrats within each community type were then amalgamated to give an overall count of each sex system type for 145 tree species in a total of 1 ha of complex notophyll vine forest on the Atherton Tableland and 126 tree species in a total of 1 ha of mesophyll vine forests in the littoral community.
The Australian monsoonal (Russell-Smith and Lee, 1992
), New Caledonian (Carpenter et al., 2003
) Nigerian (Jones, 1955
), Venezuelan (Sobrevila and Kalin-Arroyo, 1982
), and Caribbean data (Flores and Schemske, 1984
) were only used for qualitative comparisons and were excluded from the analyses because of uncertainties with overall sample size (Nigerian data) or the exact number of species in a sex-system category (monsoonal, New Caledonian, Venezuelan, and Caribbean data, see Table 2).
A 2 contingency analysis was used to explore differences among the remaining seven non-Australian data sets. A goodness of fit (G test) was used to compare the ratios of sexual systems (hermaphroditic: dioecious: monoecious) in the Australian data with combined non-Australian floras.
Phylogenetic reconstruction and character mapping
A supertree of the Australian biome data was constructed using the Phylomatic database (Webb and Donoghue, 2002
, and see Chazdon et al., 2003
), an automated process that builds a hypothetical tree based on the most recent all-angiosperm, three-gene tree (Soltis et al., 2000
), to which strict consensus trees are attached. Insufficient information on relationships between or within taxa is presented as a polytomy. Branch lengths were not known for all relationships and were given unit length. The supertree database used here was revision L20020928 containing 37 sources of data (Webb and Donoghue, 2002
). Two supertrees were built: the first containing all species (1113 species supertree) and the second at the generic level that represented unique character traits (445 taxa supertree). The 1113 species supertree was not used further owing to the large number of polytomies. A NEXUS version of these trees is available at http://www-personal.une.edu.au/
cgross/allsppconsres.nex and http://www-personal.une.edu.au/cgross/phylomaticgenonly.nex.
Traits (sex types, fruit types) were traced as binary character states over the 445 taxa supertree using MacClade version 4.03 (Maddison and Maddison, 2001
). Polytomies were randomly resolved to create 10 trees, and then character states were then traced onto these phylogenetic trees to assess the level of clustering in the phylogeny for each trait.
The clustering of traits (phylogenetic conservatism) was examined by determining the number of steps required for each character (summed cost of all changes, gains and losses) for each character state in the 445 taxa supertree. Polytomies were set as soft to represent uncertain resolution and so that the polytomy is resolved in the most parsimonous way for that character. (Hard polytomies, on the other hand, assume for any character exhibited among a group of taxa, that that character has evolved independently for each taxa, e.g., if six taxa in a group are all dioecious, it assumes that dioecy has evolved six times instead of once). For each character state, the state was reshuffled among the 445 taxa in 1000 random resolutions of the supertree. Character tracing was then used to compare the actual number of steps with the number of steps in the 1000 trees based on randomly reshuffling the character states. If the actual number of steps ranked within the lowest 5% of the 1000 random resolutions, the character was considered significantly phylogenetically clustered.
The concentrated-changes test in MacClade version 4.03 (Maddison and Maddison, 2001
) was used to examine whether observed associations among traits (e.g., dioecy and fleshy fruits) reflect phylogenetic correlations (see also Chazdon et al., 2003
). This test examines whether gains (0 to 1 changes in a binary character) in one character (e.g., dioecy) are more concentrated than expected by chance on those branches of the tree that are reconstructed to have a particular distribution of state in the second character (e.g., fleshy fruit). This test can only be used on resolved trees (i.e., without polytomies), and thus the 10 randomly resolved trees (see before) were used so that a range of probabilities could be reported.
RESULTS
The type of sex system presented was resolved for 1100 of the 1113 tree species in the Australian biome data set (Table 1, Appendix I). Fruit type and sex system type were together resolved for 1096 species. The tree species were distributed over 83 families and 408 genera. High levels of endemism occur at the species level (ca. 66%) (Appendix I). Nearly 20% of genera are monotypic (ca. 77% endemic), and ca. 55% of species are the sole representative of the genus in the study area.
|
|
2 = 0.58, P = 0.74). These data were combined for comparison with overseas data (discussed later). There was no significant difference in the abundance (mean number of abundance classes) of the three sex systems at the population level (Fig. 2) for Atherton (F2,142 = 0.54, P = 0.58) or for the littoral community (F2,123 = 2.71, P = 0.08).
|
2 analysis of the Australian biome data collected here and Hansman's (2001)
data for dry rainforests in the same region, the distribution of sex systems between the two data sets did not vary significantly at the species level (2 = 4.02, P = 0.13) or EEs level (2 = 0.56, P = 0.75) and monoecy was the predominant unisexual system. The sex system data from subtropical Australia (Adam and Williams, 2001
) also has more monoecious than dioecious species (Table 2).
In a 2 analysis of seven non-Australian floras, the distribution of sex systems at the species level is independent of biome/forest location (2 = 10.82, P = 0.54). These data were amalgamated within sex systems so that the ratio of hermaphrodite to dioecious to monoecious sex systems could be derived and used for the null hypothesis to test against the Australian data. The non-Australian combined data are represented by 61.4% hermaphroditic, 25.1% dioecious and 13.5% monoecious species, thus a ratio of 4.5 : 1.9 : 1. The Australian landscape ratio however is 3.56 : 1 : 1.32 and significantly different (G = 89.79, p < 0.001) from the non-Australian tropical sex-system ratio.
EEs could only be derived for two non-Australian data sets (La Selva, Bawa et al., 1985
; Colombia, van Dulmen, 2001
). The distribution of sex systems at the EEs level did not vary significantly between these floras (2 = 0.21, P = 0.90). These non-Australian data are represented by a ratio of 5.0 : 2 : 1. Using the ratio of sex system data for EEs in two Australian communities (that did not differ significantly, see above) of 3.0 : 1 : 1.4, a G test showed the two ratios to differ significantly (G = 28.32, P < 0.001).
The distribution of sex systems was investigated at the species and EEs level, but the two approaches did not yield significantly different results for the Australian data. This may be explained in part by ca. 55% of the 408 Australian genera tested being represented by only one species in this study (and so the species count and EEs count were equal on ca. 55% of occasions). There was a phylogenetic skew in the Australian data, however, as ca. 72% of the monoecious species belonged to just five families, the Sapindaceae, Euphorbiaceae, Malvaceae, Moraceae, and Phyllanthaceae (Table 1, Appendix I).
An analysis of the distribution of fruit types (Table 3) at the species level within sex systems showed that the overall distribution of fruit types is dependent on the type of sex system present (2 = 141.26, P = 0.0001). Dry, dehiscent fruits (commonly capsules) were the most common fruit type in monoecious species (Table 3), whereas fleshy fruits were the most common fruit type in hermaphrodites and dioecious species (Table 3). Dry, indehiscent fruit types were poorly represented overall (Table 2). When EEs were used instead of species in the analysis (Table 3), again the overall distribution of fruit types was dependent on the type of sex system present (2 = 19.47, P = 0.0006); however, this time hermaphrodites had almost equal percentages of dry-dehiscent and fleshy fruits. Monoecious species still had a high percentage of dry, dehiscent fruit types, and dioecious species still had a high percentage of fleshy fruit types (Table 3).
|
The high levels of biodiversity in the Australian tropics has received little attention from pollination ecologists (Table 5). There has been one community study (Hansman, 2001
, in which floral biology was described but floral visitors were not) and <15 autecological studies (Table 5) to date with very few studies in the last 10 yr. Very few workers have actually discerned pollinators from the general spectrum of floral visitors. Thus, the only general trend that can be commented upon is that insect pollination and in particular beetle pollination is prevalent, a point made by Irvine and Armstrong (1990)
.
|
Monoecy can be considered a broad term for bisexual individuals that have at least some unisexual flowers (e.g., strict monoecy and flowers, gynomonoecy ;mf and flowers, andromonoecy ;mf and flowers). The proportions of flower types with different sex systems, along with the spatiotemporal aspects of monoecy, can make this sex system difficult to diagnose, particularly in poorly studied floras as found in tropical Australia. This uncertainty was reduced in the current study by identifying the sexual system after examining herbarium collections for most of the monoecious species. The higher proportion of monoecious to dioecious sex systems in the Australian tropical tree flora was unexpected given the opposite trend that has been detected in other non-Australian systems (see Table 1 and Bawa and Opler, 1975
for a review).
Strict monoecy (equal ratio of : flowers) was not the most common form of monoecy encountered. Male biased monoecy (c. more : flowers found on an individual) and andromonoecy ( and 
flowers found on an individual) were the common conditions. Further studies are required to determine whether the hermaphrodite flowers on the andromonoecious species are functionally bisexual. The anthers may well be nonfunctional but enhance attractiveness to pollinators as suggested by Bawa (1977)
for Cupaniopsis guatemalensis (Sapindaceae). Within the herbarium specimens and field collections made for the Australian Sapindaceae, the male and female flowers overlapped temporally on an individual although there were generally fewer females than male flowers on most specimens (Gross, unpublished data). Detailed phenological measurements of flowers on individuals are now required to explore the sex biases on individuals in more detail. Interestingly, the floral size differences reported for neotropical, andromonoecious species (Bawa and Beach, 1981
, p. 260) were not observed among the monoecious systems in the present study.
Within the Australian data, ca. 50% of the monoecious species belong to the Euphorbiaceae and the Sapindaceae. The high proportion of monoecy in the Australian tropical Sapindaceae (89%) contrasts Bawa's assertion (p. 27 in Bawa, 1980
, but see p. 60 in Bawa, 1977
) that the Sapindaceae contain few monoecious taxa (and then these species may be functionally dioecious as has been documented for Cupaniopsis guatemalensis [Bawa, 1977
]). Certainly, as Bawa (1977)
explains, the sex systems in the Sapindaceae can appear ambiguous with incomplete field collections. Bullock (1985)
also comments that determining the sexual system in the Sapindaceae is problematic. The generic affinities that the Australian Sapindaceae share with southeast Asia (ca. 70%) may be echoed in the lability of the sexual systems, as Rao (2001)
points out that the Sapindaceae in Malaya, which comprises 60 species, shows transgradation between monoecious and dioecious forms, or polygamous or polygamodioecious states.
Analysis of fruit type revealed that fleshy fruits dominate the tree flora of northern Australia for both dioecious and hermaphroditic species, and dry fruits are correlated with monoecy (Table 3). Overall ca. 55% of the tree flora has fleshy fruits. The correlation between fleshy fruit type and dioecy has been documented elsewhere to varying degrees (Bawa, 1980
; Givnish, 1980
; Muenchow, 1987
) but rarely has the correlation between monoecy and dry fruits been demonstrated or discussed (Flores and Schemske, 1984
; Chazdon et al., 2003
). Flores and Schemske (1984)
found a significant relationship between monoecy and dry fruits among species of trees and shrubs but, owing to a lack of comparative data, they did not speculate on the association except to note that ecological characters such as habitat or pollination system may be of primary evolutionary importance for monoecious taxa. Further investigations of seed type may be useful here too because dry fruits may have seeds with fleshy appendages.
A review of the pollination ecology of the Australian tropical flora reveals a dearth of studies for any life form (Table 5) and particularly for monoecious species. In several of these studies plants had specific floral visitors (particularly insects) associated with their flowers. However, it is important to point out that, while the floral visitors may specialize on the flowers of a particular plant species, it is not necessarily a corollary that these same visitors are efficient pollinators or actually pollinate the flowers (e.g., Bullock, 1994
; Navarro, 1999
). For example, Gross and Mackay (1998)
found various levels of pollinator efficiency and pollen robbing among the nine regular floral visitors to the tropical shrub Melastoma affine. In addition, the celebrated Myristica inspida, with its oligolectic weevils, is not afforded efficient pollination services (Armstrong and Irvine, 1989a
, b
).
An important study was undertaken by Hansman (2001)
who compared the floral biology of species in dry rainforests (160 species) and adjacent savanna (33 species) in north Queensland. Although floral visitors were not discerned, she notes that the morphologies of flowers suggest pollination by small, generalist insects and that pollen to ovule ratios indicate obligate outbreeding in ca. 94% of species (N = 97 species, various life forms).
Monoecy is very often studied as an evolutionary step towards dioecy (e.g., Freeman et al., 1997
; Renner and Won, 2001
; Dorken et al., 2002
) and rarely are the maintenance, stabilizing factors, or advantages of monoecy considered in detail (but see Bawa and Beach, 1981
; Dorken et al., 2002
; Harrison and Yamamura, 2003
). If dioecy has so many advantages (reviewed in Thomson and Brunet, 1990
), then why is monoecy more prevalent than dioecy in the tropical tree flora of northern Australia? Or alternatively, what inefficiencies in the mating systems of the monoecious species have retarded the progression to dioecy? Nearly 50% of the monoecious tree species in the Australian tropics are comprised of species from just two families, the Sapindaceae and the Euphorbiaceae. These would seem worthwhile groups on which to focus research efforts, particularly as both families have both monoecious and dioecious species. However, in Australia, the reproductive ecology of species in these families has been neglected, and thus it is difficult to formulate hypotheses for specific species. Broad testable hypotheses on the stabilizing mechanisms of monoecy in the north Australian tropics are proposed here to stimulate further research.
Inefficient pollinators and the reproductive assurance hypothesis
There are several features that occur in the Australian tropical flora that in concert present a difference between the Australian tropical flora and tropical floras elsewhere. First, flowers are often small (<5 mm wide) and inconspicuous with pale perianths and shallow receptacles (Fig. 1, Armstrong and Irvine, 1990
; Gross, personal observation). Second, bat and bird pollination on the whole is infrequent (but more common in some families such as the Myrtaceae and Proteaceae, (e.g., Crome and Irvine, 1986
; Gross, personal observation), and there are no large social bees such as Apis dorsata or A. cerrana (that occur in nearby Asia) for long-distance pollen dispersal (the long-distance dispersal capabilities of the carpenter bees, Xylocopa spp., are unknown). Third, generalist pollinators or inefficient pollinators may be more common than specialists although there are very few studies to date (Table 5). Specialized and reliable pollen vectors, such as those found using dioecious species elsewhere (Renner and Feil, 1993
), may be uncommon within the Australian tropics (e.g., House, 1993
). Small insects, particularly beetles (Armstrong and Irvine, 1990
), flies, small bees, and thrips are the predominant floral visitors to trees in the northern (Table 5) and central eastern Australian tropics (Williams and Adam, 1994
, 1997
). These vectors may only be opportunistic (e.g., House, 1993
) or faithful for short distances (e.g., thrips, Momose et al., 1998
and Sakai, 2002
, but see Appanah and Chan, 1981
), or time periods and may be inefficient in that they spend long periods on a flower consuming pollen (e.g., beetle pollinators) even though they may be constant floral visitors to oliogophilic plant species. The fitness of male function, compared with female function, may be disproportionately retarded by such visitors. A monoecious strategy, in which pollen can potentially be both transferred and collected during one visit, would be advantageous when the pollinator pool is unreliable and/or inefficient. Reliance on inefficient pollinators may result in pollination deficits, which may explain why so many of the monoecious species examined offered more staminate than pistillate flowers, in both andromonoecious (e.g., Guioa lasioneura and Ganophyllum falcaltum, both Sapindaceae) and monoecious species (e.g., Sarcopteryx reticulata and Diploglottis smithii, both Sapindaceae). Furthermore, the five species of Sapindaceae for which Hansman (2001)
determined pollen to ovule ratios all produced exceedingly high amounts of pollen per flower (28 900–105 700 per ovule, three ovules per flower in each species), indicating that pollen transfer is inefficient and that these species are probably obligately outcrossing. Finally, the interference between pollen removal and pollen deposition in plants with small flowers and unspecialized insects may in fact be a selective force favoring monoecy (Bawa and Beach, 1981
), or as Renner and Feil (1993)
noted (and in view of the fact that dioecy often evolves from a monoecious pathway, e.g., Lloyd, 1980
; Renner and Ricklefs, 1995
), the evolution of dioecy in zoophilous species is only possible with reliable pollinators. Hence the high levels of monoecy reported in this study may not reflect particularly successful pollination strategies even with, perhaps, some reproductive assurance from geitonogamous pollination, but it may be better than no pollination at all, particularly as many of these rainforest species occur at low densities. Future studies could focus on the pollination ecology of monoecious species, particularly in the Sapindaceae, and compare and contrast the work with dioecious congeners. In addition, it will also be important to determine how pollen is transferred and collected by floral visitors, the spatial arrangement of plants, and the degree to which incompatibility mechanisms are operating in flowers.
Specialized pollinators, plants at low density, and infrequent crop hypothesis
An exception to the previously described hypothesis may be present in cases where specialized pollinators have to travel long distances for floral resources (e.g., Ficus spp., Nason et al., 1996
). Harrison and Yamamura (2003)
postulate that large infrequent crops in figs may in fact stabilize monoecy. In their study, monoecious species have specialized pollinators, occur at low densities, and produce large infrequent crops. Because the density of fruit-bearing trees is low, the pollinators must disperse long distances, a phenomenon that has been demonstrated in other figs (e.g., Nason et al., 1996
). It thus becomes advantageous for the plant to benefit from both pollen receipt and wasp eggs (eventual pollen vectors). Most species of Ficus are monoecious in northern Australia (Appendix I), and for this hypothesis to be plausible for species other than figs, the species would need to depend on their reproduction benefiting from reproduction in the insect. This phenomenon has been reviewed recently (Sakai 2002
) and demonstrated in Eupomatia laurina (Eupomatiaceae) in tropical Australia (Armstrong and Irvine, 1990
). This strategy would involve large crops of flowers with high levels of synchrony among individuals but with partitioning of sexual function within individuals (spatial and temporal monoecy). Greater knowledge on the phenological patterns in tropical Australia is integral to testing this hypothesis and understanding the benefits of monoecy.
These are exciting hypotheses to test, and challenging ones, because there are no studies on the pollination ecology of any monoecious tree in tropical Australia, despite the high representation of this sexual strategy. The Euphorbiaceae and particularly the Sapindaceae, provide great opportunities for the study of the maintenance of monoecy in tropical Australia and adjacent New Guinea, which shares many botanical affinities with Australia.
|
|
1 The author thanks A. Graham (CSIRO-Atherton) for making several reports available on the floristics of rainforests sites on the Atherton Tableland and the coastal areas. D. Dixon (DNA), B. Hyland (QRS), B. Gray (QRS), R. Elick (QRS), D. Halford (BRI) S. Worboys (JCU), A. Irvine (Atherton), B. Jackes (JCU), L. Jessup (BRI), A. Ford (CSIRO), P. Wilson (NSW) are thanked for sharing information. The Directors of the NSW, QRS and NE herbaria are thanked for kindly allowing me access to specimens and to their library materials. The Australian National Herbarium–Atherton (QRS) is particularly thanked for allowing me access to specimens, their database QRSHERB.DBF and to the arboretum. W. Cooper of Topaz generously collected specimens. D. Mackay helped with fieldwork (i.e., balanced on precarious ladders) and photography. R. Mackay is thanked for fieldwork and hospitality, E. Mackay for field assistance, and S. Simpson and P. Adam for useful comments. J. Bruhl, D. Mackay, A. Sjöström, I. Telford, B. Jackes, and L. Newstrom are thanked for stimulating discussions and for comments on the manuscript. This work was supported by funds from the University of New England to C. L.Gross. 
LITERATURE CITED
Adam P. 1992 Australian rainforests. Oxford University Press, Melbourne, Australia
Adam P. G. Williams 2001 Dioecy, self-compatibility and vegetative reproduction in Australian subtropical rainforest trees and shrubs. Cunninghamia 7: 89-100
Appanah S. H. T. Chan 1981 Thrips: the pollinators of some dipterocarps. Malaysian Forester 44: 234-252
Armstrong J. E. A. Irvine 1988 Beetle pollination in Australian tropical rainforests. Proceedings of the Ecological Society of Australia 15: 107-113
Armstrong J. E. A. Irvine 1989a Flowering, sex ratios, pollen-ovule ratios, fruit set and reproductive effort of a dioecious tree, Myristica insipida (Myristicaceae), in two different rain forest communities. American Journal of Botany 76: 74-85[CrossRef][ISI]
Armstrong J. E. A. Irvine 1989b Floral biology of Myristica insipida (Myristicaeae) a distinctive beetle pollination syndrome. American Journal of Botany 76: 86-94[CrossRef][ISI]
Armstrong J. E. A. K. Irvine 1990 Functions of staminodia in the beetle-pollinated flowers of Eupomatia laurina. Biotropica 22: 429-431[CrossRef]
Arroyo M. P. Uslar 1993 Breeding systems in a temperate Mediterranean-type climate montane sclerophyllous forest in central Chile. Botanical Journal of the Linnean Society 111: 83-102[CrossRef]
Ashton P. S. 1969 Speciation among tropical forest trees: some deductions in the light of recent evidence. Biological Journal of the Linnean Society 1: 155-196
Bawa K. S. 1974 Breeding systems of tree species of a lowland tropical community. Evolution 28: 85-92
Bawa K. S. 1977 The reproductive biology of Cupania guatemalensis Radlk. (Sapindaceae). Evolution 31: 52-63[CrossRef][ISI]
Bawa K. S. 1980 Evolution of dioecy in flowering plants. Annual Review of Ecology and Systematics 11: 15-39
Bawa K. S. 1992 Mating systems, genetic differentiation and speciation in tropical rain forest plants. Biotropica 24: 250-255[CrossRef][ISI]
Bawa K. S. J. H. Beach 1981 Evolution of sexual systems in flowering plants. Annals of the Missouri Botanical Garden 68: 254-274[CrossRef][ISI]
Bawa K. S. S. H. Bullock D. R. Perry R. E. Coville M. H. Grayum 1985 Reproductive biology of tropical lowland rain forest trees. II. Pollination systems. American Journal of Botany 72: 346-356[CrossRef][ISI]
Bawa K. S. P. Opler 1975 Dioecism in tropical forest trees. Evolution 29: 167-179[CrossRef][ISI]
Bennett G. 1860 Gatherings of a naturalist in Australasia. J. van Voorst, London, UK
Borchert R. S. A. Meyer R. S. Felger L. Porter-bolland 2004 Environmental control of flowering periodicity in Costa Rican and Mexican tropical dry forests. Global Ecology and Biogeography 13: 409-425[CrossRef][ISI]
Bullock S. 1985 Breeding systems in the flora of a tropical deciduous forest. Biotropica 17: 287-301[CrossRef][ISI]
Bullock S. H. 1994 Wind pollination of neotropical dioecious trees. Biotropica 26: 172-179[CrossRef][ISI]
Carpenter R. J. J. Read T. Jaffre 2003 Reproductive traits of tropical rain-forests in New Caledonia. Journal of Tropical Ecology 19: 351-365[CrossRef][ISI]
Chazdon R. L. S. Careaga C. Webb O. Vargas 2003 Community and phylogenetic structure of reproductive traits of woody species in wet tropical forests. Ecological Monographs 73: 331-348
Cox P. A. 1988 Monomorphic and dimorphic sexual strategies: a modular approach. In J. Lovett Doust and L. Lovett Doust [eds.], Plant reproductive ecology. Patterns and strategies. Oxford University Press, New York, New York, USA
Crome F. A. Irvine 1986 "Two bob each way": the pollination and breeding system of the Australian rainforest tree Syzygium cormiflorum (Myrtaceae). Biotropica 18: 115-125
Darwin C. 1877 The different forms of flowers on plants of the same species. J. Murray London, UK
Dorken M. E. J. Friedman S. C. H. Barrett 2002 The evolution and maintenance of monoecy and dioecy in Sagittaria latifolia (Alismataceae). Evolution 56: 31-41[CrossRef][ISI][Medline]
Flora of Australia. 1989–1999 vols. 3, 4, 8, 12, 16, 17b, 18, 22, 25, 28, 29. Australian Government Publishing Service, Canberra/CSIRO Melbourne, Australia
Flores S. D. Schemske 1984 Dioecy and monoecy in the flora of Puerto Rica and the Virgin Islands. Biotropica 16: 132-139[CrossRef][ISI]
Freeman C. D. J. Lovett-Doust A. El-Keblawy E. D. McArthur 1997 Sexual specialization and inbreeding avoidance in the evolution of dioecy. The Botanical Review 63: 65-92
Gentry A. H. 1988 Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annals of the Missouri Botanical Garden 75: 1-34[CrossRef][ISI]
Gillespie T. W. 1999 Life history characteristics and rarity of woody plants in tropical dry forest fragments of Central America. Journal of Tropical Ecology 15: 637-649[CrossRef][ISI]
Givnish T. 1980 Ecological constraints on the evolution of breeding systems in seed plants: dioecy and dispersal in gymnosperms. Evolution 34: 959-972[CrossRef][ISI]
Gross C. L. 1993 The breeding system and pollinators of Melastoma affine (Melastomataceae); A pioneer shrub in tropical Australia. Biotropica 25: 468-474[CrossRef][ISI]
Gross C. L. D. Mackay 1998 Honeybees reduce fitness in the pioneer shrub Melastoma affine (Melastomataceae). Biological Conservation 86: 169-178[CrossRef][ISI]
Hamilton A. G. 1897 On the fertilization of Eupomatia laurina. Proceedings of the Linnean Society of New South Wales 22: 48-55
Hansman D. J. 2001 Floral biology of dry rainforest in north Queensland and a comparison with adjacent savanna woodland. Australian Journal of Botany 49: 137-153[CrossRef]
Harrison R. D. N. Yamamura 2003 A few more hypotheses for the evolution of dioecy in figs (Ficus, Moraceae). Oikos 100: 628-635[CrossRef][ISI]
Hopkins M. S. A. W. Graham R. Jensen J. Maggs B. Bayly R. K. Hewett 1996 Succession and disturbance in the remnant rainforest type complex notophyll vine forest on basalt (Type 5b). II. The floristics of type 5b on the Atherton Tableland. A report prepared for the Wet Tropics Management Authority. CSIRO Division of Wildlife and Ecology, Tropical Forest Research Centre, Atherton. Report no. VM 1/0296– 7. CSIRO, Atherton, Australia
Hopkins M. S. A. W. Graham R. K. Hewett 1999 The distribution and floristic composition of rainforests on coastal sands in the Wet Tropics region. A report prepared for the Wet Tropics Management Authority. CSIRO Wildlife and Ecology, Tropical Forest Research Centre, Atherton. Report no. VM 1/0399–12. CSIRO, Atherton, Australia
Hopper S. 1980 Pollination of the rain-forest tree Syzygium tierneyanum (Myrtaceae) at Kuranda, Northern Queensland. Australian Journal of Botany 28: 223-237[CrossRef][ISI]
House S. M. 1989 Pollen movement to flowering canopies of pistillate individuals of three rainforest tree species in tropical Australia. Australian Journal of Ecology 14: 77-93[ISI]
House S. M. 1993 Pollination success in a population of dioecious rain forest trees. Oecologia 96: 555-561[CrossRef][ISI]
Hyland B. P. M. T. Whiffin D. C. Christophel B. Gray R. W. Elick A. J. Ford 1999 Australian tropical rain forest plants. Trees and shrubs. CSIRO, Collingwood, Victoria, Australia
Hyland B. P. M. T. Whiffin D. C. Christophel B. Gray R. W. Elick 2003 Australian tropical rain forest plants. Trees, shrubs and vines. CSIRO, Collingwood, Victoria, Australia
Ippolito A. J. Armstrong 1993 Floral biology of Hornstedtia scottiana (Zingiberaceae) in a lowland rain forest of Australia. Biotropica 25: 281-289[CrossRef][ISI]
Irvine A. K. J. E. Armstrong 1990 Beetle pollination in tropical forests of Australia. In K. S. Bawa and M. Hadley [eds.], Reproductive ecology of tropical forest plants. Man and the Biosphere Series, vol. 7, UNESCO, Parthenon Publishers, UK
Jones E. W. 1955 Ecological studies on the rainforest of Southern Nigeria. IV. The plateau forest of the Okomu Forest Reserve. Journal of Ecology 43: 564-594[CrossRef][ISI]
Jones R. E. F. H. J. Crome 1990 The biological web—plant/animal interactions in the rainforest. In L. J. Webb and J. Kikkawa [eds.], Australian tropical rainforests. Science–values–meaning. CSIRO, Melbourne, Australia
Lloyd D. 1972 Breeding systems in Cotula L. (Compositae, Anthemideae). I. The array of monoclinous and diclinous systems. New Phytology 71: 1181-1194[CrossRef]
Lloyd D. 1980 The distributions of gender in four angiosperm species illustrating two evolutionary pathways to dioecy. Evolution 34: 123-134[CrossRef][ISI]
Maddison D. R. W. P. Maddison 2001 MacClade, version 4.03. Sinauer Associates, Inc., Sunderland, Masachusetts, USA
Momose K. T. Nagamitsu T. Inoue 1998 Thrips cross-pollination of Popowia pisocarpa (Annonaceae) in a lowland dipterocarp forest in Sarawak. Biotropica 30: 444-448[CrossRef][ISI]
Morley B. D. H. R. Toelken [eds.] 1983 Flowering plants in Australia. Rigby Publishers, Adelaide, Australia
Muenchow G. E. 1987 Is dioecy associated with fleshy fruit?. American Journal of Botany 74: 287-293[CrossRef][ISI]
Nason J. D. E. A. Herre J. L. Hamrick 1996 Paternity analysis of the breeding structure of strangler fig populations: evidence for substantial long-distance wasp dispersal. Journal of Biogeography 23: 501-512[CrossRef][ISI]
Navarro L. 1999 Pollination ecology and effect of nectar removal in Macleania bullata (Ericaceae). Biotropica 31: 618-625[CrossRef][ISI]
Numata S. M. Yasuda T. Okuda N. Kachi N. S. M. Noor 2003 Temporal and spatial patterns of mass flowerings on the Malay Peninsula. American Journal of Botany 90: 1025-1031
Osunkoya O. O. 1998 Population structure and breeding biology in relation to conservation in the dioecious Gardenia actinocarpa (Rubiaceae)—a rare shrub of North Queensland rainforest. Biological Conservation 88: 347-359[ISI]
Qrsherb.Dbf. 2003 Database from the Australian National Herbarium. CSIRO, Atherton, Australia
Rainforest Conservation Society of Queensland. 1986 Tropical rainforests of North Queensland. Their conservation significance. A report to the Australian Heritage Commission. Special Australian heritage publication series no. 3. Commonwealth of Australia. Australian Government Publishing Service, Canberra, Australia
Rao A. N. 2001 Some observations on dioecious flowering plants. In N. S. Rangaswamy [ed.], Phytomorphology Golden Jubilee Issue 2001: Trends in Plant Sciences. University of Delhi, Delhi, India
Renner S. J. Feil 1993 Pollinators of tropical dioecious angiosperms. American Journal of Botany 80: 1100-1107[CrossRef][ISI]
Renner S. S. H. S. Won 2001 Repeated evolution of dioecy from monoecy in Siparunaceae (Laurales). Systematic Biology 50: 700-712[CrossRef][ISI][Medline]
Renner S. S. R. E. Ricklefs 1995 Dioecy and its correlates in the flowering plants. American Journal of Botany 82: 596-606[CrossRef][ISI]
Richards A. J. 1986 Plant breeding systems. George Allen & Unwin Pub. Ltd., London, UK
Russell-Smith J. A. Lee 1992 Plant populations and monsoon rain forest in the Northern Territory, Australia. Biotropica 24: 471-487[CrossRef][ISI]
Sakai S. 2002 A review of brood-site pollination mutualism: plants providing breeding sites for their pollinators. Journal of Plant Research 115: 161-168[CrossRef][ISI][Medline]
Skutch A. F. 1932 The pollination of the palm Archontophoenix cunninghamii. Torreya 32: 29-37
Sobrevila C. M. T. Kalin-Arroyo 1982 Breeding systems in a montane tropical cloud forest in Venezuela. Plant Systematics and Evolution 140: 19-37[CrossRef][ISI]
Soltis D. E. P. S. Soltis M. W. Chase M. E. Mort D. C. Albach M. Zanis V. Savolainen W. H. Hahn S. B. Hoot M. F. Fay M. Axtell S. M. Swensen L. M. Prince W. J. Kress K. C. Nixon J. S. Farris 2000 Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133: 381-461[CrossRef]
Stacy E. A. 2001 Cross-fertility in two tropical tree species: evidence of inbreeding depression within populations and genetic divergence among populations. American Journal of Botany 88: 1041-1051
Stevens P. F. 2001 (and onwards) Angiosperm phylogeny website, version 4, May 2003 [and more or less continuously updated since], http://www.mobot.org/MOBOT/research/APweb/
Thomson J. D. J. Brunet 1990 Hypotheses for the evolution of dioecy in seed plants. Trends in Ecology and Evolution 5: 11-16
Tomlinson P. B. 1974 Breeding mechanisms in trees native to tropical Florida—a morphological assessment. Journal of the Arnold Arboretum 55: 269-290[ISI]
Van Dulmen A. 2001 Pollination and phenology of flowers in the canopy of two contrasting rain forest types in Amazonia, Colombia. Plant Ecology 153: 73-85[CrossRef][ISI]
Weathers W. W. R. B. Siegel 1995 Body size establishes the scaling of avian postnatal metabolic rate: an interspecific analysis using phylogenetically independent contrasts. Ibis 137: 532-542[ISI]
Webb C. O. M. J. Donoghue 2002 Phylomatic: a database for applied phylogenetics. Software distributed by author, available at website http://www.phylodiversity.net/phylomatic
Webb L. J. 1978 A general classification of Australian rainforests. Australian Plants 9: 349-363
Weber E. T. 1994 The reproductive biology of Elaeocarpus largiflorens C. White subsp. largiflorens (Elaeocarpaceae). Honours thesis, James Cook University of North Queensland, Queensland, Australia
Westoby M. M. R. Leishman J. M. Lord 1995 On misinterpreting the ‘phylogenetic correction.’. Journal of Ecology 83: 531-534[CrossRef]
Williams G. P. Adam 1994 A review of rainforest pollination and plant–pollinator interactions with particular reference to Australian subtropical rainforests. Australian Zoologist 29: 177-212
Williams G. A. P. Adam 1997 The composition of the bee (Apoidea: Hymenoptera) fauna visiting flowering trees in New South Wales lowland subtropical rainforest remnants. Proceedings of the Linnean Society of New South Wales 118: 70-95
| ||||||||||||||||||||||||||||
