Potential recruitment events that may have occurred in more recent times would not
have been evident from the 2003 aerial photographs utilised in Chapter 8, since there
is a variable time delay between recruitment and the plants being large enough to
distinguish on an aerial photograph. This delay is likely to be approximately 10 years
- based on known growth rates of young plants and aerial photograph resolution. As a
potential recruitment year (1993) identified in the hierarchical cluster analysis used in
section 8.2.2 was able to be determined from the aerial photographs there was an
opportunity for on-ground confirmation of recruitment. Aging juvenile plants
presumed to have recruited recently could identify safe sites.
To determine if there had been more recent recruitment events at Dowd Morass in the
past 10 years or so, 50 x 50 m quadrats were established in the field in February 2006
in three vegetation communities with potential recruitment conditions. These areas
were based on previous vegetation classification carried out at Dowd Morass by Elisa
Raulings (pers. comm.): Swamp Scrub, Reeds and Open Water/Bare Sediment. By
definition, Reed communities encompass a wide range of herbaceous tussock-forming
species including Phragmites, Juncus, Paspalum and Baumea. Three quadrats were
placed in each vegetation community in three widely separated areas of Dowd Morass
Each quadrat was searched for seedling or juvenile plants and the numbers and
position (hummock/hollow) of these plants were recorded in a contingency table.
Seedlings were identified by the characteristic of M. ericifolia to have an alternate
pattern of the first few leaves and the rounded tips to these leaves; no other plant
species in the wetland has these characteristics. Annual growth counts were carried
out on all identified juveniles to determine approximate age; annual growth is readily
identifiable on most Melaleuca species as new growth emerges from dormant buds
leaving the basal section of the new stem covered in bracts.
Figure 8.1 Sample sites for identification of possible recent recruitment of Melaleuca ericifolia in Dowd Morass. Sample sites are marked with a red cross. (Source –
Wetlands Ecology Unit, Monash University)
Environmental data (salinity, pH, moisture content and organic matter content) were
obtained for soil in the hummocks and in hollows of the reed communities. Soil cores,
(5 cm wide x 10 cm deep) were taken laterally into the top, middle and base of 6
hummocks and from the surrounding substrate adjacent to each hummock in each 50
x 50 m quadrat. Samples were taken from all three of the quadrats in the reed
communities; in all, 72 soil cores were taken. Soil cores were processed within 48
hours of collection. Samples were processed using Australian standard methods for
pH (Department of Sustainable Natural Resources, n.d. - A), salinity (Department of
Sustainable Natural Resources, n.d - B) and water content (Standards Association of
Australia, 1977). To determine pH, approximately 50 g of each sample was air dried
in tins. A 1:5 soil:water suspension was prepared with deionised water and the
container mechanically shaken for 1 hour at 15 rpm. A hand-held pH/EC/TDS meter
(Hanna Instruments, Westlab Laboratory Supplies, Ballarat, Victoria) was used to
measure pH. To obtain water content approximately 80 g of each sample were
weighed prior to drying and then oven-dried at 110
C for 16 hours. All samples were
weighed after drying to determine moisture content, expressed as percentage dry soil
weight (MC%). Electrical conductivity (EC) followed the methods of Hatton (2004)
and was measured using a TPS-LC81 salinity meter, with a k = 1 conductivity cell
and a 1:5 soil:water suspension. The suspension comprised 10 g of powdered dry soil
and 50 mL of deionised water. Reference solution consisted of 0.01 M KCl, which
has a known EC of 1.41
1 was used to calibrate the salinity meter. Final EC
μS cm) were converted to salinity values of mg/l
using a conversion factor
of 0.6 (MDBC, 1990). Organic matter content was determined using the Weight Loss
on Ignition method, temperature 500
C for 15 hours (Mullins and Heckendorn 2005)
8.2.2 Statistical analysis
Data were analysed with Analysis of Variance (ANOVA) with the SPSS (version 12;
http://www.spss.com/, verified September 2006) and Systat (version 5;
, verified September 2006) computer packages. Percentage
data were arc-sine transformed before analysis. One-way and three-way fully
orthogonal ANOVA designs were used for analysis. Since all factors were considered
as “fixed”, treatment effects were calculated with reference to the MS residual (error)
term (Zar 1999). Post-hoc tests used Bonferonni-corrected probability values.
The contingency table generally followed Zar (1999). However, due to values of zero
appearing in all but one of the categories all values were increased by a value of 1.
8.3.1 Recruitment sites determined in field inspections
The field inspection showed that juveniles of M. ericifolia were identified only in reed
areas (Table 8.1, Figure 8.2). This classification is, by necessity, fairly wide and
covers a range of herbaceous plant genera including Phragmites, Juncus, Paspalum and Baumea all of which form hummocks. Within these areas further discrimination
could be made between hummocks and sediments, with juveniles only found on
herbaceous plant hummocks.
Figure 8.2 View of reed community at the western end of Dowd Morass (see Figure
8.1). Hummocks in the foreground are composed of Paspalum distichum (Water
Couch), those in the distance Phragmites australis (Common Reed). Arrows point to
juvenile M. ericifolia.
Figure 8.3 Close up of juvenile M. ericifolia recruit on a hummock composed of
Juncus krausii (Sea Rush) at Dowd Morass.
Juvenile M. ericifolia plants occurring on hummocks varied in height from 80 cm to
140 cm with only several of the larger plants exhibiting ramet production. Age was
determined to be between 12 – 13 years based on averaged annual growth counts
(data not shown).
Characterisation of hummocks and surrounding sediments identified significant
differences between moisture, salinity and pH levels within the various levels of the
hummocks and the surrounding sediments (Figures 8.4 –8.6). Soil moisture, salt and
pH in the lower portions of hummocks and immediately adjacent sediment samples
were not significantly different (N= 18, P = 0.88. 0.16 and 0.12, respectively). There
were significant differences in all three variables between the upper and middle levels
of the hummocks (N = 18, P < 0.001) and the middle levels and the hummock
bases/sediments (N = 18, P < 0.001).
Salinity concentrations in hummocks overall were below 11 g L
, but decreased to 8
in the middle levels of the hummock and averaged about 5 g L
at the base of
the hummocks and in adjacent non-hummock sediments. The middle and lower level
of the hummocks were at or below the water surface, and therefore salinity in these
levels would have been partially or totally diluted, whereas raised salinity levels of the
upper hummock level could have been influenced by capillary transport through
There were significant differences in pH between the upper and lower portions of the
individual hummocks. The upper portions of hummocks were the only positions that
had pH above 5.0; middle portions of the hummocks had a mean pH of 4.0, but
ranged from 3.2 – 5.0. Lower portions of the hummock and sediments had a mean pH
of just over 3.0, ranging from 2.3 – 4.0.
The mean moisture content in the two upper portions of the hummocks was 325 –
450% of dry weight. Mean moisture contents were generally less than 100% of dry
weight in the hummock bases and in adjacent, non-hummock sediments (Figure 8.4).
The very high moisture contents in the middle and upper portions of the hummocks
may have been related to the relatively high organic matter content and capillary
action of the organic matter. The lower portions of the hummocks and the surrounding
substrates had relatively higher levels of mineral earths and consequently lower soil-
Table 8.1 Contingency table of potential recruitment sites for Melaleuca ericifolia at
selected locations at Dowd Morass.
Hummock 1 44
Hollow 1 1
45 2 49
Hummock 1.88 42.2
Chi-square = 14.6
Degrees of freedom = 2
Probability = 0.001
0 100 200 300 400 500 600 700 Top M iddle Base Sediment % Moisture content
Figure 8.4 Moisture content of substrata at three contrasting positions on hummocks
and surrounding sediments in reed beds at Dowd Morass (error bars = Standard Error,
n = 18).
0 2 4 6 8 10 12 14 16 Top M iddle Base Sediment Salinity (g L -1 )
Figure 8.5 Salinity of substrata at three contrasting positions on hummocks and
surrounding sediments in reed beds at Dowd Morass (error bars = Standard Error, n =
a b c c a b b a
0 1 2 3 4 5 6 7 Top M iddle Base Sediment pH
Figure 8.6 pH of substrata at various positions on hummocks and surrounding
sediments in reed beds at Dowd Morass (error bars = Standard Error, n = 18).
Figure 8.7 Organic matter content of various positions on hummocks and in adjacent
sediments in reed beds at Dowd Morass (error bars = Standard Error, n = 18).
c b c a a a b c
8.4.1 Recruitment sites
The distribution of successfully recruited seedlings and juvenile plants of M. ericifolia across Dowd Morass was limited to specific vegetation types and specific
microtopographic positions on the vegetated hummocks. Spatial and temporal
variation in the abiotic conditions of safe sites, such as is found in wetlands that
contain hummocks, can lead to habitat segregation based on the germination
requirements of different species (Bell and Clarke 2004). Moisture, duration of
inundation, salinity and temperature all play a major role in recruitment success as has
been shown for several wetland species including Melaleuca halmaturorum (Mensforth and Walker 1996; Mineke 2002; Kellogg 2003) and others. The
recruitment requirements of M. ericifolia are moderate moisture availability and low
salinity levels (Chapter 6 and 7). Other species found at Dowd Morass presumably
have different recruitment requirements, and therefore can utilize other safe site
conditions, leading to strong patterning in the vegetation (Bell and Clarke 2004;
Hatton 2004, Salter et al. 2007). It is presumed that edaphic conditions and possibly
competition with mature M. ericifolia precludes establishment of seedlings under
existing M. ericifolia plants.
Suitable safe sites for M. ericifolia, identified in this study, occurred on hummocks
formed by the herbaceous plants in the reed community. The present conditions found
on these hummocks were not within the range of tolerance for germination and
hypocotyl hair formation for M. ericifolia. However, established juveniles of M.
ericifolia, when age counts were carried out, corresponded to a period in the wetland
when conditions were temporarily favourable to recruitment, a situation induced by
dilution or flushing of hummock substrates created by above-average rainfall and
flooding with fresh water (Chapter 8). This implies that M. ericifolia, like many
species that occur in brackish-water situations, needs periodic freshwater conditions
Hummocks are widely identified as suitable safe sites for wetland plants as they
moderate the extreme environmental conditions found in lower strata of wetlands,
particularly highly saline conditions, permanent water-logging and acid sulfate soils
(Ehrenfeld 1995; Karofeld 1998; Roy et al. 1999; Nungesser 2003). Soil pH values
identified in this study would most probably be toxic to most plant recruits and adults
(Sammut et al. 1995). Higher elevations of the hummocks identified in this study,
moderated pH values to within a range that is conducive to seedling growth. Higher
elevations of hummocks in this study at Dowd Morass, while not appreciably
different in salinity levels at the time of sampling, have been shown to have
significantly reduced salinity levels when subjected to higher than normal
precipitation and/or flooding (Coppolino 2007).
The upper layers of hummocks, which are generally raised above the prevailing water
level, provide advantages to seedlings as they are better aerated and provide warmer
rooting zones (Roy et al. 1999). Both may be critical factors in rapid establishment
before unsuitable condition return (Cousens et al. 1988; Cornett et al. 2000). The
suitability of the upper portions of the hummocks for plant growth was clearly
demonstrated by Raulings et al. (2006) at Dowd Morass, with 90% of M. ericifolia
seedlings surviving when planting into the tops of hummocks, while only 10 % of
seedlings survived after 3 months in the less favourable lower elevations of the
The importance of hummocks to the restoration of wetlands is recognised in several
large-scale projects being carried out at Dowd Morass and in wetlands of coastal
Louisiana (Conner 2002; Bruland and Richardson 2005; Boon et al. 2007). In the case
of Dowd Morass, artificial hummocks have been created to determine their potential
use in rehabilitation where biotic conditions have altered considerably over time: from
freshwater to brackish (Boon et al. 2007). In other situations such as at the Hemet/San
Jacinto Wetlands in southern California and coastal wetlands in Louisiana, artificial
hummocks have been installed in created wetlands to increase the patterning of
vegetation in a wastewater treatment system (Bruland and Richardson 2005;Thullen et al. 2005).
Thullen et al. (2005) recognised a wide range of benefits provided by hummocks,
apart from assisting plant recruitment, including: water quality enhancement,
mosquito management, enhanced plant decomposition, hydraulic control and the
meeting of wildlife management goals. However, the cost of construction of artificial
hummocks at Hemet/San Jacinto Wetlands was high and maintaining integrity of
created hummocks was problematic. This is likely to be the case with all artificially
created hummocks. An additional problem in coastal wetlands is the disturbance of
acid sulfate soils and exposure of these sediments to the air, which releases toxic
compounds negating the benefits of hummock construction (Boon et al 2007).
The benefits associated with the creation of artificial hummocks, particularly
microtopographic patterning, may overcome any short-term deleterious effects such
as acid sulfate soils by creating suitable safe sites for the colonisation of wetland
plants. Re-establishment of hummocks with their attendant influences on substrate
conditions may accelerate restoration in artificially created or rehabilitated wetlands
(Bruland and Richardson 2005).
Safe sites identified in this study, when combined with data from previous chapters
(6, 7 and 8) provide a strong basis for future management of wetlands of the
Gippsland Lakes. These data elucidate the specific conditions under which
germination and recruitment of M. ericifolia is possible. When combined with
historical aerial photography and climatic data, predictions can be made of when and
where recruitment will take place. A combination of all of this data can, at least in
theory, allow land managers to create the conditions under which recruitment and
establishment is most likely to succeed.
Chapter 9 General Discussion
The six experimental chapters (Chapters 3-8) in this thesis have investigated various
aspects of the growth and germination characteristics of Melaleuca ericifolia and the
implications of these characteristics for the management of this species in coastal
wetlands in southeast Australia.
Genetic studies described in Chapter 3 demonstrated unequivocally the clonal nature
of M. ericifolia and how clonal growth influences interactions between individual
clones (genets) of this species. Genetic studies were combined with historical aerial
photography to determine growth rate and longevity of individual genets. Mean
lateral growth was comparatively rapid at approximately 0.5 m per year although this
rate increased with the age of the plants. Clones occupying areas up to 45 m in
diameter were estimated to be approximately 46 - 52 years old. Clonality studies
provided background information on the reliance or lack of reliance on seed
recruitment in M. ericifolia.
Viability, or more precisely germinability, studies were described in Chapter 4.
Germinability in the clonal M. ericifolia was contrasted with germinability in the non-
clonal M. parvistaminea to determine whether there was a trade-off between seed
germinability and clonality in two sympatric species in the same genus. Germinability
was also compared within M. ericifolia across most of its range in south-eastern
Australia in relation to population size and human-induced alterations of habitat
quality, namely fragmentation and secondary salinisation.
The influence of salinity, temperature and light on germination of the seed of M. ericifolia was described in Chapter 5. Hypocotyl hairs, a little investigated feature of
newly germinated seedlings, were discovered serendipitously during germination
trials in Chapter 5 and further investigated in Chapter 6, particularly in relation to
their sensitivity to salinity, temperature and light.
Potential and actual safe sites for germination and recruitment events were
investigated in Chapters 7 and 8 using a range of techniques including historical aerial
photography, historical climate data, on-ground investigation of identified
germination sites and comparison of this data with germination and hypocotyl hair
data determined in previous chapters.
A more detailed description of the key findings for each of these components is
provided in the following sections.
Clonality in M. ericifolia
It was critical to identify the extent of clonality and the interaction of individual
genets of this species to fully understand the population structure and dynamics of M. ericifolia. No previous studies had identified the extent of individual clones of M. ericifolia but had, in some instances, made limited comment that the species was
clonal (e.g. Jeans 1996; Craven and Lepschi 1999; Holliday 2004). Several other
wetland species in south-eastern Australia, most notably Casuarina/Allocasuarina species and other Melaleuca species have been identified as being clonal but the
extent of their clonality has not been investigated (Carter et al. 2006a; Carter et al.
2006b). A mixture of laboratory-based molecular biology approaches was
complemented by an analysis of aerial photographs to demonstrate the extent of
clonality in M. ericifolia.
The use of the genetic analysis technique, inter-simple sequence repeats was
particularly useful. Inter-simple sequence repeats can be used when there is little
available information on the genome of the species (Zietkiewicz et al. 1994) and
allow the investigator to access variation in numerous microsatellite regions dispersed
throughout the plant genome.
Individual clumps identified visually in the field by their characteristic dome shape
(Chapter 3, Figure 3.2) were confirmed to be individual genets (plants) on the basis of
the molecular analysis. Of particular note was the lack of intermingling of individual
genets. Lack of intermingling is a characteristic of the phalanx mode of clonal growth
(Lovett Doust 1981), in which the plant produces short and frequently branched
connections between ramets. The phalanx mode of growth is thought to be an
adaptation to low-nutrient, high-light habitats with a high degree of spatial and
temporal heterogeneity for other environmental conditions such as water regime and
salinity (Van Groenendael et al. 1997; Kleign and van Groenendael 1999).
The interconnectedness of the individual ramets, clearly illustrated in Chapter 3
(Figure 3.1), allows for the sharing of water and nutrients allowing the genet as a
whole to have wider ecological amplitude than plants that lack the clonal ability.
Extensive clonality, as identified in a range of other wetland species and in M. ericifolia, presumably allows the individual genet to effectively capture limited
resources and increase competitive ability (Rea and Ganf 1994; Barsoum 2002;
The use of aerial photographs, coupled with genetic studies as utilised in Chapter 3,
allowed ageing of individual clones and the determination of their extent and growth
rates. The eighteen clones identified in Chapter 4 were approximately 46 years old but
could have been as old as 52 years old based on the data shown in Chapter 7.
Individual genets of M. ericifolia ranged in size from 1,174 to 3,274 m
considerably larger than has been recorded in other Australian and overseas
Myrtaceae, (19 m
- 530 m
) but smaller than some clonal species from other families:
Lomatia tasmanica (1.2 km wide, Lynch et al. 1998), Pteridium aquilinum (1.2 km
wide, Parks and Werth 1993), Zostera marina (6,400 m
, Reusch et al. 1999) and
Populus tremuloides (43.3 ha, Kemperman and Barnes 1976). Additional plants of M. ericifolia identified from locations close to Dowd Morass, but not specifically studied,
were nearly twice as large (40 – 60 m wide) and were estimated to be between 80 and
100 years old. Growth rates of the various clones studied varied between 3.5 and 9 m
, an approximate lateral spread of ~0.5 m yr
9.2 Trade-off between sexual and asexual recruitment and impacts on germinability
Plants can be loosely categorised into two broad groups according to regeneration
strategy: obligate seed regenerators and rootstock regenerators (Lamont and Wiens
2003). There is commonly a trade-off between sexual reproduction and vegetative
reproduction because of the high energenic costs of seed production (Lovett Doust
1989). The diversion of resources into vegetative growth would by necessity reduce
sexual output; in contrast a limited opportunity for sexual reproduction could be offset
by greater probability of reproduction by vegetative means.
There is a complete range of regeneration strategies within the genus Melaleuca, ranging from fully obligate seed regenerators (M. parvistaminea), rootstock
regenerating species (M. halmaturorum, M. quinquenervia) to strongly regenerating
clonal species (M. cajaputi, M. ericifolia). Melaleuca parvistaminea and M. ericifolia are largely sympatric but segregated within individual wetlands: M. parvistaminea occurs on rarely inundated areas whereas M. ericifolia more in regularly inundated
areas. The co-occurrence of these two species in several wetlands in southern
Australia provided a unique opportunity to determine if there was in fact reduced
viability in the rootstock regenerating species M. ericifolia.
Viability of M. ericifolia seed was markedly lower, ranging from 0 – 38% across the
species range in Victoria and Tasmania, whereas the viability of M. parvistaminea seed was much higher and ranged between 70 – 80%. The finding of reduced viability
in M. ericifolia is consistent with other studies of species pairs that found reduced
viability in the clonal species. There were significant differences in seed weight with
M. ericifolia ranging between 15-29