Bougoure, J., Brundrett, M., Brown, A., & Grierson, P. (2008). Habitat characteristics of the rare underground
orchid Rhizanthella gardneri. Australian Journal of Botany, 56(6), 501-511. DOI: 10.1071/BT08031
Australian Journal of Botany
Link to publication in the UWA Research Repository
Post print of work supplied. Link to Publisher's website supplied in Alternative Location.
Copyright owners retain the copyright for their material stored in the UWA Research Repository. The University grants no end-user
rights beyond those which are provided by the Australian Copyright Act 1968. Users may make use of the material in the Repository
providing due attribution is given and the use is in accordance with the Copyright Act 1968.
If you believe this document infringes copyright, raise a complaint by contacting firstname.lastname@example.org. The document will be
immediately withdrawn from public access while the complaint is being investigated.
Download date: 02. Sep. 2017
Habitat characteristics of the rare underground orchid, Rhizanthella
Ecosystems Research Group, School of Plant Biology M090, University of Western
Australia, Crawley, WA 6009, Australia.
Department of Environment and Conservation, Kensington, WA. 6151, Australia.
Corresponding author. Email: email@example.com
, via a connecting mycorrhizal fungus, for the purpose of
carbohydrate and nutrient acquisition. Here, we quantify key climate, soil and
vegetation characteristics of known R. gardneri habitats to provide baseline data for
monitoring of known R. gardneri populations, better understand how R. gardneri
interacts with its habitat, and to identify possible new sites for R. gardneri
introduction. We found that the habitats of the two known R. gardneri populations
show considerable differences in soil chemistry, Melaleuca structure and Melaleuca
‘Melaleuca’ in this paper refers to the various species of the Melaleuca uncinata complex that
productivity. Multivariate analyses showed that both MDS and PCA ordinations of
soil chemical characteristics were very similar. Individual sites within populations
were relatively similar in all attributes measured while overall Northern and Southern
habitats were distinct from each other. These results suggest that R. gardneri can
tolerate a range of conditions and may be more widespread than previously thought,
given that there are extensive areas of Melaleuca thickets with similar habitat
characteristics across south-western Western Australia. Variability within the habitats
of known R. gardneri populations suggests translocation of this species into sites with
similar vegetation may be a viable option for the survival of this species.
Rhizanthella gardneri R.S.Rogers (western underground orchid) is a critically
endangered orchid that is entirely subterranean and mycoheterotrophic, meaning that
it has no ability to photosynthesise and limited capacity to independently access soil
nutrient pools (Dixon and Pate 1990). Rhizanthella gardneri and the only other
member of the genus, R. slateri (Rupp) M.A.Clem. & P.J.Cribb (eastern underground
orchid), are unusual plants in that they remain subterranean during flowering (Fig. 1a)
– a unique trait even amongst mycoheterotrophic species (Leake 1994). It is thought
that R. gardneri is linked, via a mycorrhizal fungus (Thanatephorus gardneri
Warcup), to an autotrophic shrub (Melaleuca), to form a tripartite relationship of
nutrient and carbohydrate exchange (Warcup 1985, 1991; Dixon and Pate 1990). If R.
nutrition, climate and biotic interactions probably most important. While there are
comprehensive climate records for all known R. gardneri sites, the nutritional status
and productivity of R. gardneri habitats, i.e., the capacity of the habitat to maintain
nutrient and carbon supply under field conditions, remains unknown.
become increasingly apparent. Monitoring over the past 26 years has identified a suite
of intensifying threats to the survival and proliferation of R. gardneri including; small
population sizes, changes in rainfall distribution (Australian Bureau of Meteorology
2008), limited capability for recruitment due to habitat fragmentation, isolated
habitats in excessively cleared landscapes, possible encroachment by salinisation,
increased weed species competition, unsuitable fire management and herbicide aerial
spray drift (Brown et al. 1998). Populations of R. gardneri have purportedly been in
decline since monitoring began in 1979, with anecdotal evidence suggesting that
decreased sightings of R. gardneri are correlated to a decline in health of the
associated Melaleuca habitat. Declining Melaleuca ‘health’ has been associated with
casual observations of a reduced litter layer, diminishing ground cover of other
species, decreased density of thickets, limited recruitment and the yellowing of
foliage. However, there have been no quantitative or qualitative measurements of
either the structure or nutritional status of known R. gardneri habitats, or to what
extent habitat that is deemed healthy differs from that in decline. Consequently, we
assessed vegetation structure, productivity and soil nutrients of known R. gardneri
habitats. We sought to determine key similarities and differences among habitats to
help define prerequisites and limitations to the survival of R. gardneri and to identify
new R. gardneri populations and potential sites for R. gardneri introduction.
Field sites analysed in this study represent habitat of all known live R. gardneri
individuals during 2007. Rhizanthella gardneri has been found in only two areas of
south-west Western Australia; three sites within 30 km of Corrigin in the northern
wheatbelt (Northern Population) and three sites approximately 80 km east of
Ravensthorpe in the south-eastern wheatbelt (Southern Population) (Fig. 1c). The
Northern and Southern populations are 300 km apart with many areas of apparently
similar habitat between these sites; however, there are no records of R. gardneri from
these intervening areas. The six sites where R. gardneri is known to occur are
between one and eight hectares in area and are generally within vegetation remnants
directly adjacent to agricultural land. Two sites are located on private property
(Dallinup Creek North and South); two sites are in protected Nature Reserves
(Babakin and Sorenson’s Reserve); and two sites are located on Unallocated Crown
Land (Oldfield River and Kunjin). None of the sites have been directly affected by
fire in the past 30 years. However, similar surrounding vegetation that has been burnt
in the past 30 years has rapidly and successfully regenerated.
dry summer months (Australian Bureau of Meteorology 2008). Seasonal temperature
fluctuations at Northern sites range from 5-15
C in winter and 15-32
C in summer.
Climate conditions for the Southern sites are slightly milder with average winter
temperatures as low as 7-16
C during winter and 14-28
C during summer
(Australian Bureau of Meteorology 2008).
Lepschi thickets (Fig. 2b) whereas the Southern populations occur in thickets of M.
species of the M. uncinata complex (Craven et al. 2004). All of the above Melaleuca
species are taxonomically similar and belong to the M. uncinata complex, a group of
species widely distributed throughout southern Australia (Craven et al. 2004;
Broadhurst et al. 2004). The six sites, encompassing both Northern and Southern
populations, although primarily thickets of Melaleuca, also include more open areas
dominated by smaller Dryandra spp., large exposed rocks, areas of grasses and
sedges, Allocasuarina campestris (Diels) L. A. S. Johnson thickets or occasional
larger Banksia media R.Br. and/or mallee eucalypts. However, it appears that the
occurrence of R. gardneri plants is specific to the Melaleuca dominated patches
within each site.
sites are considerably sandier than the loamy soils at Southern sites, although they
often form hard surface crusts during prolonged dry spells, especially during summer.
A summary of the general site characteristics of the Northern and Southern
populations of R. gardneri, including historical information, is given in Table 1.
Aboveground structure and productivity of Melaleuca thickets
Three replicate plots of 10 m x 10 m were established within Melaleuca thickets at
each of the six sites. Shrub density, canopy cover, stem number and height were
measured for ten Melaleuca shrubs within each 10 m x 10 m plot. The biomass of
C for 48
hours, and specific leaf area (m
) was calculated for each shrub. Leaf area per
shrub was calculated as specific leaf area x total dry leaf mass.
C for 48 hours and then weighed.
coincide with R. gardneri flowering and when soil moisture was consistently greatest.
Duplicate soil cores (7.5 cm diameter) were sampled 10, 25, 50 and 100 cm in a south
westerly direction from three randomly selected Melaleuca shrubs from each of the
three plots at each of the six sites. Cores were bulked by position within each plot
after separating into 0-5 cm and 5-15 cm depths for measurement of root biomass.
Leaf litter depth (mm) was also recorded at all sampling points. Roots were separated
from soil by sieving samples (< 1 mm) under running water, and hand-sorting roots.
Root samples were then oven dried at 60
C for 48 hours and all roots < 2 mm
diameter were weighed. The gravimetric moisture content of sampled soils was
calculated after drying for 24 hours at 110
chemical attributes. The pH of 5 g fresh soil was measured after vigorously mixing
samples with deionised water (1:1) (Thomas 1996). Labile phosphorus was measured
using two techniques. First, labile inorganic P (Bray P
) was measured according to
the method of Bray and Kurtz (1945). Briefly, air dry soil was mixed with an
extraction solution (0.03 N NH
F and 0.1 N HCl), shaken for 45 seconds and filtered
immediately. Second, labile organic and inorganic hydroxide-extractable P fractions
were measured using the method described by Grierson and Adams (2000). Briefly,
10 g of soil in 50 mL of 0.1 M NaOH solutions were shaken for 16 hours after which
extractions were centrifuged and the supernatant removed via filtration. One aliquot
) of the filtered supernatant was acidified with HCl, to precipitate organic
matter, and re-filtered. A second aliquot (OH-P
) of the filtered supernatant was
). Phosphorus in all extracts was measured using a modified
ascorbic acid method of Kuo (1996). Organic P (OH-P
) was estimated as the
difference in P between OH-P
C isotope signatures, using an Automated Nitrogen Carbon Analyser-Mass
Spectrometer consisting of a Roboprep connected with a Tracermass isotope ratio
spectrometer (Europa Scientific Ltd., Crewe, UK) in the West Australian
standardised against a secondary reference of Radish collegate (3.167 % N,
41.51 % C,
that was in turn standardised against primary analytical
standards (IAEA, Vienna). Accuracy was measured at 0.07 %, while precision was
measured at 0.03 %, according to the stipulations for reporting analytical error in
stable isotope analysis outlined by Jardine and Cunjak (2005).
The best models for predicting component biomass of Melaleuca based on canopy
area as the independent variable were determined after testing a number of different
allometric models. The allometric models that best fitted the data were chosen by
examining residual distributions and maximum adjusted r
. The behaviour of the
models for small diameter and "out of sampled range" trees were also carefully
examined. The models tested were; y = a + bx and y = a + bx + cx
; where x is total
canopy area (m
), y is the mass of different components of sample shrubs, a, b and c
0.05 and Statview
5.0 software (SAS Institute 1996) was used for all statistical analyses.
among sites. Similarity of the soil characteristics of Northern and Southern sites was
calculated using principal components analysis and non-metric multidimensional
scaling ordinations (Primer software version 6.2, Primer-E Ltd, Clarke and Gorley
2006). Analysis of similarity (ANOSIM) was performed to determine if samples
within groups were more similar than between groups. Similarity percentages
(SIMPER) analysis was used to determine contribution of individual variables to
dissimilarity of groups. Data were normalised prior to analysis and Euclidean distance
was used to generate the resemblance matrix for data (Clarke and Gorley 2006).