Seed Morpho-Anatomy and Germination Enhancement of the Australian Native Species Lomandra longifolia Labill. and L. hystrix (R.Br.) L.R. Fraser & Vickery

: Lomandra species are an important understory component of many Australian native ecosystems, contributing to the ﬂoristic richness and stabilizing soils. However, a limited understanding of their germination biology currently hinders their efﬁcient use in seed-based restoration and ornamental plant production. The present study investigated Lomandra longifolia and L. hystrix diaspore morpho-anatomy and evaluated different mechanical and/or chemical treatments (nicking, leaching, smoke water and gibberellic acid [GA 3 ]) and under light or dark conditions to enhance germination. Embryos of both species were small and linear with a low embryo to seed ratio (<0.45). Germination rates of both species were signiﬁcantly hastened by leaching seeds in running water for 36 h as compared to a non-leached seed. The results suggest that pre-treating both Lomandra species by leaching could maximize the effectiveness of seed used by resulting in faster, more uniform and, therefore, reliable germination of these species. Finally, seeds of L. longifolia had low ﬁnal germination (<40%), with a high presence of viable but dormant seeds. The ecological cues that promote germination in nature for both species should be further examined.


Introduction
Lomandra is a genus within the family Asparagaceae [1] that generally consists of small perennial herbs with a rhizomatous growth habit that often form tussocks [2]. Lomandra longifolia Labill. (spiny-head mat-rush) is one of the most widely distributed species in Australia [3] (Figure 1) and is particularly common in the south-east region of Australia [4]. It is highly adaptable and can grow in a wide range of habitats, such as on hillsides of dry forests, alongside creeks and coastal headlands. Lomandra hystrix (R.Br.) L.R. Fraser and Vickery (creek mat-rush) has a more confined natural distribution than L. longifolia and is found in the coastal regions of Queensland and New South Wales. Both species are commonly used as ornamental plants as well as for seed-based restoration projects [5]. These species are important understory components of many Australian ecosystems, providing shelter, breeding sites and food resources for native wildlife [4]. They have multiple benefits for seed-based restoration, which include their contribution to floristic richness and their ability to stabilize soils to prevent erosion due to their fibrous root system [6]. Large quantities of L. longioflia seeds are used annually to restore degraded Australian bushlands and disturbed plant communities due to human activity such as mining and construction [6]. Both species can be propagated using freshly produced diaspores (seed encased within the pericarp; hereafter, referred to as seeds) when sown in autumn conditions in Australia.
Even though Lomandra spp. are an important ecological, ornamental and restoration component in Australia, their seed biology has been poorly studied, hindering efficient seed use. It is well known that many Australian native species have diverse dormancy seed treatments to overcome possible PD at optimum alternating temperatures in li dark conditions. This information will provide an understanding of the dormancy anisms that are preventing germination and will help in developing improved see mination protocols for L. longifolia and L. hystrix. Improved germination would als to increased use of these highly beneficial species in land regeneration projects and r the costs associated with seed wastage.

Seed Lots
Lomandra longifolia and L. hystrix seeds were obtained from a commercial see plier (Native Seeds and Land Repair, Maleny, QLD, Australia). Lomandra longifolia were collected from the suburb of Redlands, part of the Brisbane metropolitan a south-east Queensland during December 2017, while L. hystrix seeds were collected the suburb of Caloundra, Sunshine Coast Region in south-east Queensland during ary 2017. Both seed batches had >90% viability as determined by the supplier. After ery, seeds were stored in a seed storage cabinet at 15 ± 1 °C temperature and 15 relative humidity until used. Seed age was 16 months for L. longifolia and 27 months hystrix when used. Identifying the mechanisms that control dormancy and germination, together with finding new ways to hasten seed germination and seedling emergence, could be of great importance to achieve cost-effective usage of Lomandra seeds. For example, studies have shown that a combination of smoke water and stratification can increase seed germination to 50% in Lomandra preisii (Endl.) Ewart. [7], which could be associated with an overcoming of PD. Moreover, leaching of inhibitors from L. longifolia seeds with five cycles of soaking and rinsing combined with warm stratification significantly improved its germination [14]. Similarly, leaching seeds with running tap water or pericarp removal could overcome dormancy in L. sonderi and could achieve germination of ca. 20% of the seeds. Similar treatments increased germination from 40 to 80% in L. drummondii [2]. The authors related the increases in germination with the removal of germination-inhibiting chemicals found in the tissues surrounding the embryo. Gibberellic acid (GA 3 ) and smoke water have also been used successfully in promoting germination in L. sonderi [2]. Gibberellic acid is known to stimulate endosperm weakening and stimulate embryo germination [15]; smoke water has been shown to promote germination in a wide range of Australian native species [7,16,17]. Furthermore, scarification and seed nicking (a small cut through the tissues surrounding the embryo) of the pericarp and seed coat can be effective in overcoming PD or morpho-physiological dormancy (MPD) for several species. These treatments can relieve mechanical restrictions of the fruit tissues and/or seed coat, allowing embryo growth [18,19]. This can be particularly helpful in seeds that have a low embryo growth potential.

Seed Fill, Weight and Morpho-Anatomy
Microscopic investigation of seed internal morpho-anatomy is useful to characterize the seed and help elucidate germination biology aspects. Such analysis can determine embryo size, shape and location within the seed to determine the embryo to seed ratio (E:S) and size of the endosperm and/or cotyledons [10]. Documentation of embryo characteristics can also be used in determining the presence of morphological dormancy (MD) or MPD associated with embryo development [20]. Underdeveloped embryos have differentiated organs and tend to have low E:S ratios [21]. Embryos in seeds with MD need to undergo a growth period prior to radicle protrusion [10] (for example, in [22]). There are few studies that focus on morpho-anatomy characterization of L. longifolia and L. hystrix seed (such as [5,6]). This information could provide helpful insights into explaining the dormancy and slow germination rates observed for both species.
The current study describes the seed morpho-anatomy of both Lomandra species and determines methods for elevating seed germination to enable a more cost-efficient use of these seeds in seed-based restoration projects and ornamental plant production. Thus, the objectives were to (1) identify relationships between germination, seed fill and seed morphoanatomical structures (embryo and seed size); and (2) develop methods to speed up the rate and increase the final seed germination by investigating chemical and mechanical seed treatments to overcome possible PD at optimum alternating temperatures in light or dark conditions. This information will provide an understanding of the dormancy mechanisms that are preventing germination and will help in developing improved seed germination protocols for L. longifolia and L. hystrix. Improved germination would also lead to increased use of these highly beneficial species in land regeneration projects and reduce the costs associated with seed wastage.

Seed Lots
Lomandra longifolia and L. hystrix seeds were obtained from a commercial seed supplier (Native Seeds and Land Repair, Maleny, QLD, Australia). Lomandra longifolia seeds were collected from the suburb of Redlands, part of the Brisbane metropolitan area in south-east Queensland during December 2017, while L. hystrix seeds were collected from the suburb of Caloundra, Sunshine Coast Region in south-east Queensland during January 2017. Both seed batches had >90% viability as determined by the supplier. After delivery, seeds were stored in a seed storage cabinet at 15 ± 1 • C temperature and 15 ± 3% relative humidity until used. Seed age was 16 months for L. longifolia and 27 months for L. hystrix when used.

Seed Fill, Weight and Morpho-Anatomy
Seeds of both species were examined by using an X-ray machine (Faxitron MX-20 Imaging system, Lincolnshire, IL, USA) to determine seed fill percentage. Seed samples (5 replicates of 25 seeds per species) were exposed to 18 Kv of X-ray tube voltage for 20 s and images were captured using Bioptics software (Olympus, Tokyo, Japan) at 2× magnification of resolution. The percentage of filled seeds was determined by counting the number of seeds that had a full-sized endosperm and embryo. Filled seeds had a white color, and damaged or unfilled seeds were indicated by black areas inside the seed. The percentage of filled, partially filled (seeds with parts of their endosperm and/or embryo missing) and unfilled seed was determined. Partly filled and unfilled seeds were considered non-viable. Filled seeds with intact and healthy-looking embryos were considered viable. To determine the 100-seed mean weight, 5 samples of 100 seed from each species were randomly selected and weighed.
Seed anatomical structure was determined by photographic analysis by using a light microscope (Olympus SZX7, Mornington, TAS, Australia) with a digital camera attached. Seed and embryo size-specifically, length, width and area-were measured using CellSens software. To determine the E:S ratio, seeds were dissected longitudinally, and the embryo length was divided by seed length [23]. Embryo development was classified based on its anatomy (size and shape) according to Martin [21]. The presence of a developed or underdeveloped embryo was evaluated to identify if MD was present [24].

Germination Stimulation Using Mechanical and Chemical Treatments
Eight mechanical and/or chemical treatments were used viz. seed leaching, seed nicking, chemical treatment with smoke water at three different concentrations (Regen 2000 Smokemaster, batch no. 11957R, Tecnica, Bayswater, VIC, Australia), GA 3 (90% gibberellin A3, Sigma-Aldrich, lot BCBD6798V, St. Louis, MO, USA) or a combination of treatments (Table 1). Prior to treatments, all seeds were surface sterilized in 2% (v/v) sodium hypochlorite (NaOCl) solution for 10 min [25] with two drops of Tween 20 (Labchem, Zelienople, PA, USA) added as a surfactant. Seeds were then washed four times with sterile water and blotted dry. To undertake leaching, seeds of each species were transferred to several mesh ball infusers (diameter of 5 cm) and placed individually into 250 mL glass beakers for 36 h under running, turbulent cold tap water (ambient from main town water supply). To nick seeds, a small cut on the embryo end of the seed was undertaken using a scalpel blade ( Figure 2). Chemical treatments (5 mL) were applied to each Petri dish (9 cm diameter) containing two Whatman No. 1 filter papers. Previcure ® (2% v/v; Bayer Crop Science) was added to the Petri dishes to inhibit fungi growth [19,26]. Then, following the addition of seed, each Petri dish was sealed with Parafilm to prevent evaporation of solutions. This was undertaken in a laminar air flow hood to reduce possible microbial contamination. All treatments were applied under light (with a 12/12 h day/night photoperiod) or dark conditions to simulate the seed being placed on the soil surface (light) or seed burial (dark). For seeds imbibed under light, the 12/12-h photoperiod used had a light intensity of 100 µmol m −2 s 1 (produced by cool white, fluorescent tubes). Dark conditions were achieved by wrapping the Petri dishes with two layers of aluminum foil. Petri dishes were placed in an incubator (TRIL-750 Illuminated Refrigerator Incubator, Thermoline, Wetherill Park, NSW, Australia) using a matching 12/12-h thermoperiod of 20/10 ± 1 • C. The thermoperiod was selected from earlier studies carried out on both Lomandra species (unpublished data) and from the published literature (maximum germination for L. longifolia was at 20 • C and for L. hystrix it was at 15 • C [1]).

Data Collection and Analysis
For both species, each treatment was replicated 4 times, and each replication had 25 seeds per Petri dish. A completely randomized design was used. All seed germination tests were run for 60 days. Petri dishes were examined for germination twice weekly. Germinated seeds (radicle protrusion ≥ 2 mm) were recorded and removed. Seeds germinated under dark conditions were observed under a green safety light (Lion 24 LED Magnetic work lamp, covered with a green plastic sheet) in a dark room. Cumulative germination over time for the different seed treatments for each species was determined using a nonlinear regression model fitted with the drm function in package drc [27] using R software, version 3.5.3 [28]. A three-parameter log-logistic model was used [29]. The germination rate index (GRI) was determined according to Maguire [30] (Equation (1)). Germination data (final germination percentage, percentage of dormant seeds and GRI) for each species were analyzed using a two-way factorial analysis of variance (ANOVA). When significant differences were identified, a Tukey's honest significance difference (HSD) test was used as a post-hoc analysis to identify significant differences between treatment means.
Equation (1): Germination rate index (GRI; Maguire [30]) i: Day of germination count G i : Percentage of seeds germinated in day i G i−1 : Percentage of seeds germinated the previous count day

Seed Fill and Seed Morpho-Anatomy
Lomandra longifolia had a 100% and L. hystrix had a 99% seed fill (Figure 3a,b, respectively). The seed cross-sectional area was 7.5 mm 2 for L. longifolia and 11.8 mm 2 for L. hystrix ( Table 2). Embryos of both species were fully differentiated and seemed to be fully developed. They were small and linear and located in the basal part of the seed (Figure 4). Both species had a large proportion (>70%) of the seed consisting of endosperm tissue surrounding the embryo (Figures 3c and 4). The E:S ratio was 0.4 for both L. hystrix and L. longifolia ( Table 2). The 100-seed weight was 900 mg for L. hystrix and 860 mg for L. longifolia. Table 2. Seed characteristics as determined from 10 randomly selected seeds of Lomandra longifolia Labill and Lomandra hystrix (R.Br.) L.R. Fraser and Vickery. An endosperm was present in both species, and the embryo type was linear. The E:S ratio, seed and embryo length, area, perimeter and width were measured. Mean ± SE.

Germination Enhancement Using Mechanical and Chemical Treatments
Leaching significantly increased (p ≤ 0.001) the GRI for both species in comparison to untreated seeds and other pre-treatments regardless of light conditions (Table 3). Leached seeds were also the first to start germination ( Figure 5). For L. longifolia, seeds leached in darkness had a higher GRI (1.0 ± 0.1% day −1 ) as compared to leached seeds under light (0.5 ± 0.1% day −1 ). Leached seeds for L. hystrix incubated under light conditions had GRI

Germination Enhancement Using Mechanical and Chemical Treatments
Leaching significantly increased (p ≤ 0.001) the GRI for both species in comparison to untreated seeds and other pre-treatments regardless of light conditions (Table 3). Leached seeds were also the first to start germination ( Figure 5). For L. longifolia, seeds leached in darkness had a higher GRI (1.0 ± 0.1% day −1 ) as compared to leached seeds under light (0.5 ± 0.1% day −1 ). Leached seeds for L. hystrix incubated under light conditions had GRI of 3.7 ± 0.1% day −1 as compared to control of 1.7 ± 0.2% day −1 (Table 3). However, the GRI for leached seeds incubated under light was significantly higher (p < 0.0001) than in darkness (2.4 ± 0.2% day −1 ). Appl. Biosci. 2023, 2, 7 of 3.7 ± 0.1% day −1 as compared to control of 1.7 ± 0.2% day −1 (Table 3). However, the GRI for leached seeds incubated under light was significantly higher (p < 0.0001) than in darkness (2.4 ± 0.2% day −1 ). Leaching or nicking (mechanical treatments) did not improve the final germination percentage for either species when compared to the untreated seeds ( Figure 5). Lomandra longifolia had a maximum germination of 37.0 ± 1.9% (Figure 5a; leached seeds incubated in darkness). Furthermore, untreated L. longifolia seeds under darkness had a significantly higher final germination percentage (26.0%) as compared to seeds exposed to light (9%; p ≤ 0.01). In L. longifolia, the germination percentages for leached seeds were significantly higher as compared to nicked seeds (7% higher under light and 13% higher under darkness; p ≤ 0.01). Lomandra hystrix had a maximum of 86.0 ± 2.6% germination (Figure 5b; leached seeds incubated under illuminated conditions), and unlike L. longifolia, no significant differences (p > 0.05) were observed for untreated seeds of L. hystrix incubated under Leaching or nicking (mechanical treatments) did not improve the final germination percentage for either species when compared to the untreated seeds ( Figure 5). Lomandra longifolia had a maximum germination of 37.0 ± 1.9% (Figure 5a; leached seeds incubated in darkness). Furthermore, untreated L. longifolia seeds under darkness had a significantly higher final germination percentage (26.0%) as compared to seeds exposed to light (9%; p ≤ 0.01). In L. longifolia, the germination percentages for leached seeds were significantly higher as compared to nicked seeds (7% higher under light and 13% higher under darkness; p ≤ 0.01). Lomandra hystrix had a maximum of 86.0 ± 2.6% germination (Figure 5b; leached seeds incubated under illuminated conditions), and unlike L. longifolia, no significant differences (p > 0.05) were observed for untreated seeds of L. hystrix incubated under light or dark conditions. In L. hystrix, untreated seeds, leached seeds and nicked seeds had significantly higher final germination percentages (>53%) than for the remainder of the treatments (<40%; p ≤ 0.05). Table 3. Germination rate index (GRI) for Lomandra hystrix (R.Br.) L.R. Fraser and Vickery and Lomandra longifolia Labill. incubated under complete darkness (24 h) or light/dark (12/12-h photoperiod) conditions following chemical and/or mechanical treatments. All seeds were incubated for 60 days at 20/10 • C with a 12/12-h matching thermoperiod. Treatments were as follows: leaching with running tap water for 36 h, nicking (small cut through pericarp and seed testa), smoke water (SW)-SW1:50, SW2:100, SW3:200 mL L −1 ; gibberellic acid-GA 3 : 289 µM; SW2 + nicking, GA 3 + nicking and control (untreated seeds). Means ± SEM were calculated using 4 replications, 25 seeds per replication. Treatments that had a significantly higher GRI than the control are in bold, and treatments that resulted in zero germination are denoted by a dash.

Light Dark Light Dark
Leaching Treatment with smoke water, at all three concentrations (SW1:50, SW2:100 and SW3:200 mL L −1 ) and in combination with nicking, gave a significantly lower final germination percentage for both species in light and dark conditions when compared to untreated seeds (≤37% for L. hystrix and ≤1% for L. longifolia; p ≤ 0.05). Lomandra hystrix and L. longifolia seeds treated with a combination of GA 3 and nicking had significantly higher final germination in darkness as compared to light conditions (p ≤ 0.05; Figure 6). Under light, GA 3 and SW3 gave a significantly lower final germination of <3% than other treatments for L. hystrix and L. longifolia (p ≤ 0.005). In L. longifolia, significantly higher (p ≤ 0.05) final germination percentage occurred in dark conditions as compared with light conditions, for leaching, nicking, GA 3 and the control (24, 18, 18, 17% higher, respectively).
After 60 days, ≥50% of L. longifolia seeds remained ungerminated (Figure 6a). These seeds were considered dormant as they were filled (determined by X-ray) and firm. No significant differences were observed in dormant seeds of L. longifolia between treatments and control (p ≥ 0.05). For L. hystrix (Figure 6b), the number of dormant seeds was significantly higher for seeds treated with smoke water (≥80% dormant seeds in all concentrations), SW2 + nicking (≥63%) and GA 3 + nicking (>40%) (Figure 6). Moreover, for GA 3 and nicking + GA 3 treatments in this species, dormancy was significantly higher in light conditions when compared to darkness (98 vs. 56% and 87 vs. 41% respectively: p ≤ 0.005) ( Figure 6). Seed death occurred mostly in L. longifolia, with the highest number observed in SW2 + nicking (51% dead seeds in light conditions) and GA 3 + nicking (>40% for light and dark conditions) ( Figure 6). In contrast, a few seeds (≤8%) of L. hystrix were dead at the end of the experiment.

Discussion
This study aimed to characterize seed morpho-anatomy and to determine seed treatments that could enhance the germination of two Australian native Lomandra species. The seed traits that influence germination of L. longifolia and L. hystrix had not been addressed in detail previously and there is a lack of published protocols on how to improve germination for both important seed-based restoration and ornamental species. The current study showed that seeds of both species had PD at the time of sowing, probably due to tissues surrounding the small embryo imposing a mechanical constraint to its protrusion and/or the presence of chemical inhibitors (as shown to occur in other Lomandra species). Further results indicate that germination rates of both Lomandra species can be significantly improved by leaching seeds under running tap water. Having a deep understanding of seed biology is crucial for seed-based restoration success [31]: it enables seed treatments to be optimized for improving seed performance [32] and can result in faster, more uniform and, therefore, reliable germination. Additionally, understanding seed germination timing will be crucial for seed use under future climate change [33].

Seed Morpho-Anatomy
The seed fill percentage in both seed lots studied was 100%, suggesting that other factors besides seed viability or seed fill are limiting germination. Both species presented small (E:S = 0.4), linear and basal embryos (Figure 4) surrounded by a thick layer of endosperm. Even though the embryos were small, microscopic evaluation showed that embryos in this current study were fully differentiated and seemed to be fully developed,

Discussion
This study aimed to characterize seed morpho-anatomy and to determine seed treatments that could enhance the germination of two Australian native Lomandra species. The seed traits that influence germination of L. longifolia and L. hystrix had not been addressed in detail previously and there is a lack of published protocols on how to improve germination for both important seed-based restoration and ornamental species. The current study showed that seeds of both species had PD at the time of sowing, probably due to tissues surrounding the small embryo imposing a mechanical constraint to its protrusion and/or the presence of chemical inhibitors (as shown to occur in other Lomandra species). Further results indicate that germination rates of both Lomandra species can be significantly improved by leaching seeds under running tap water. Having a deep understanding of seed biology is crucial for seed-based restoration success [31]: it enables seed treatments to be optimized for improving seed performance [32] and can result in faster, more uniform and, therefore, reliable germination. Additionally, understanding seed germination timing will be crucial for seed use under future climate change [33].

Seed Morpho-Anatomy
The seed fill percentage in both seed lots studied was 100%, suggesting that other factors besides seed viability or seed fill are limiting germination. Both species presented small (E:S = 0.4), linear and basal embryos (Figure 4) surrounded by a thick layer of endosperm.
Even though the embryos were small, microscopic evaluation showed that embryos in this current study were fully differentiated and seemed to be fully developed, consisting of a well-defined primary axis and cotyledon. In contrast, Ruiz-Talonia et al. [14] identified MPD in L. longifolia due to the presence of an underdeveloped embryo. In their study, a combination of leaching and warm stratification was needed to overcome dormancy and improve germination. This present study evaluated the use of stored seeds of L. Longifolia and L. hystrix (for 27 and 16 months, respectively). Seed age can influence seed dormancy status as seeds can undergo after-ripening, dormancy cycling [10] or overcome MD or MPD during storage. Therefore, there is a possibility that embryo growth occurred during storage, overcoming MD. Although, it is important to note that Baskin et al. [22] proposed that a seed with a small embryo in relation to the endosperm does not strictly mean that the embryo is underdeveloped. Some species, such as in the genus Nymphaea (Nymphaeaceae [18]) and Drosera anglica (Huds.) LePage and W.Bldw. (Droseraceae [22]), have small embryos with low E:S ratios (e.g., 0.24 ± 0.01 for D. anglica). However, these embryos did not exhibit growth inside the seed before germination could occur. Therefore, MD was not present. Consequently, to identify the presence of MD or MPD in both of our study species, further studies should be undertaken with freshly collected seeds to determine whether embryo growth occurs within the seed prior to them becoming germinable.

Seed Germination Ecology and Enhancement
In L. longifolia, seeds incubated in the dark had a significantly higher final germination percentage than those imbibed in light conditions (37 vs. 13%; Figure 5). This suggests that L. longifolia can germinate to a higher extent if buried in the soil. On the other hand, there was no significant difference between light and dark treatment in L. hystrix, although marginal improvements in germination were observed under light. Seeds of many species are sensitive to light intensity and quality, which is a mechanism to avoid plant competition [34]. Therefore, light detection by seeds can be an important germination cue [35]. In L. longifolia, germination inhibition by light could be related to the avoidance of germination near or at the soil surface, ensuring seeds are positioned at sufficient depth where moisture is more reliably obtained. This is a common seed adaptation to environments where moisture is limited [36], such as those where L. longifolia naturally grows. Moreover, temperature is moderated at greater soil depths [35], which may be an adapdation of L. longifolia to enable germination to occur in a wide range of climates ( Figure 1). However, burial at depth does not appear to be a requirement of L. hystrix seeds, possibly because this species has adapted to growing along watercourses and in rainforests, where water is usually more abundant and temperature fluctuations are less extreme.
Leaching significantly improved the GRI of both species ( Figure 5). The positive effects of leaching in this study are consistent with previous observations made for L. longifolia [14] and L. sonderi. Plummer et al. [2] suggested that water-soluble germination inhibitors located in the pericarp and embryo could inhibit germination in L. sonderi; germination inhibition was successfully overcome by removing the pericarp or by leaching seeds for 24 h in running tap water. This may be an adaptation to regions where occasional, but heavy, rainfall can leach out seed germination inhibitors and break down the seed coat tissues [26,37]. This mechanism ensures that conditions are suitable for germination and seedling establishment. On the other hand, Baskin et al. [18] suggested that leaching can also act as a stratification treatment. Periods of warm or cold stratification have been shown to alleviate PD in seeds [38]. For example, stratification at 26/13 • C or 33/18 • C for 4-or 8-weeks alleviated dormancy in L. preisii when seeds were germinated at 18/7 • C [7]. Moreover, warm stratification achieved >80% germination in Acanthocarpus preisii Lehm. (Asparagaceae) as compared with <20% when seeds were not stratified [39]. Further studies should be directed at identifying if the use of stratification treatments or wet/drying cycles [40] could be involved in overcoming dormancy of Lomandra seeds in the soil seedbank.
The improvement in germination of seeds with non-deep PD after scarification has been related to overcoming a mechanical barrier to embryo growth imposed by the tissues surrounding the embryo [10,41]. Likewise, seed nicking has also been shown to relieve embryo growth restrictions [19]. Although in this study, nicking did not significantly promote germination, a positive trend was observed ( Figure 5). In seeds presenting PD, the fully developed embryo has a low growth potential; therefore, it cannot overcome the mechanical constraints imposed by its surrounding tissues [10]. Once treatments such as nicking are performed, the embryo can gain sufficient expansive force to protrude through the surrounding tissues. In nature, embryos with low growth potential need cues from the environment to initiate internal chemical signaling which promote certain covering tissues to breakdown and increase the growth potential [10]. However, this resistance can also be weakened over time in the seeds' natural environment by the production of tissue-softening enzymes released by the embryo or by weakening through physical biotic factors such as temperature, fire, animal ingestion, seed burial and saprophytic fungi [37,42]. Embryo germination resistance also varies according to imbibition conditions such as temperature and light/dark interactions [10], and further studies could be undertaken to determine the influence these parameters can have on endosperm resistance.
Constant exposure to smoke water throughout the incubation period significantly inhibited germination of both species, with ≤1% final germination for L. longifolia and ≤20% final germination for L. hystrix ( Figure 6). This contrasts with the findings of Merritt et al. [43], where smoke water was reported to promote germination in other Lomandra species. For example, imbibing seeds in smoke water (1:10 [v/v]) for 24 h enhanced seed responses to warm stratification providing the highest germination for L. preisii seeds (ca. 50% [7]). In contrast, Vening et al. [44] reported that smoke water used in an agar-based germination medium at 1:10 (v:v) had no significant effect on germination in Australian native forbs from fire prone environments. However, seed sensitivity to smoke water can be a complex process [7], as active constituents of smoke water (such as karrikins) can vary between different stock solutions and species react differently [45]. Furthermore, Adkins et al. [45] found that caryopsis of wild oats (Avena fatua L.) had greater germination when exposed to smoke water for 7 days prior to incubation with distilled water as compared to caryopsis that received smoke water before and during incubation.
Considering the above, constant exposure of Lomandra seeds to smoke water in this current study could have caused germination inhibition. This may be due to the dual effect of smoke water on the seed germination process reported by Light et al. [46] for 'Grand Rapids' lettuce (Lactuca sativa L.). This study proposed that smoke water had an inhibitory component that enters the seed and a promotor component that remains in the seed, inactive until sufficient rainfall has leached out the inhibitor. Additionally, it is important to note that smoke has been proposed to act as a germination enhancer, rather than a dormancy breaker [7,47]. Therefore, smoke water effects might only act to enhance germination after dormancy has been overcome. Further studies on both Lomandra species could be undertaken by applying smoke water as a pre-treatment prior to incubation, but after seeds have been treated for dormancy.
Although leaching improved the GRI for L. longifolia, the final germination percentage achieved in all treatments was low (≤37% germination). This low germination percentages of L. longifolia, even when exposed to GA 3 , could suggest the presence of a deeper level of PD. In many species with intermediate or non-deep PD, germination is stimulated by GA 3 , while those with deep PD fail to germinate in GA 3 treatments that would normally promote germination [48]. To alleviate PD in L. sonderi, the pericarp needed to be removed from the seed (presumably to remove germination inhibition imposed by these tissues); then, GA 3 (145 µM) was applied (to relieve embryo dormancy [2]). Further studies on L. longifolia seeds should be undertaken such as excision of the pericarp and seed coat [42] and excision of the embryo [10]; and then, applying GA 3 to identify if the pericarp or endosperm are preventing GA 3 from reaching the embryo. Although GA 3 did not improve germination in both Lomandra species, there was a significant interaction between GA 3 and darkness, where seeds treated with GA 3 had higher germination in the dark as compared to light. This could be related to GA 3 interacting with the phytochrome system within seeds [37].

Conclusions
This study is one of the first to investigate techniques to enhance germination of the Australian native species L. longifolia and L. hystrix. The innate slow and initially low germination of both Lomandra seeds requires a significantly large number of viable seeds to be sown to achieve the plant density required for ornamental plant production or seed-based restoration projects. Results from this study show that both species had small, linear embryos and a high proportion of endosperm tissue. Slow germination of both species is most likely associated with the presence of at least one mechanism of PD present in the seed. Leaching seeds prior to incubation was the only treatment to significantly hasten seed germination in both Lomandra species. To determine the mechanisms by which leaching functioned on hastening germination and to correctly classify dormancy in both species, further studies are now needed on freshly collected seeds. Future areas of research include measuring embryo growth during incubation to test for MPD, undertaking warm stratification prior to germination incubation, and seed treatments with GA 3 . Understanding the factors that influence seed germination and pre-treating seeds accordingly, or ensuring these requirements are met in the natural environment (in the case of seed-based restoration), is crucial for the success and cost-efficient use of these seeds. It is also important to consider the scaling-up of treatments for large restoration projects or ornamental plant production and how that could affect seed tissues and the overall cost-effectiveness of the treatment. Moreover, the possibility of providing similar effects naturally in the field by sowing seeds when long periods of rain are forecasted should also be examined.