Introduction
Many studies have highlighted the beneficial effects of smoke and smoke products on the germination of seeds in Mediterranean habitats and other fire-prone environments (Måren et al., Reference Måren, Janovský, Spindelböck, Daws, Kaland and Vandvik2010; Moreira et al., Reference Moreira, Tormo, Estrelles and Pausas2010). These trends have been reported in South Africa (Brown et al., Reference Brown, Jamieson and Botha1994), around the central Mediterranean basin (Crosti et al., Reference Crosti, Ladd, Dixon and Piotto2006), in California (Keeley and Fotheringham, Reference Keeley and Fotheringham1998) and in different regions of Australia (Dixon et al., Reference Dixon, Roche and Pate1995; Ooi et al., Reference Ooi, Auld and Whelan2006). Many genera that have been found to be highly responsive to smoke and smoke products are also well represented in the biodiverse island of New Caledonia, which is located 1300 km north-east of Australia. New Caledonia is characterized by a mild tropical climate and is internationally recognized as one of only 34 biodiversity hotspots, with over 3300 species of vascular plants within a land area of only ~18,000 km2, including the most basal extant angiosperm – Amborella trichopoda (Mittermeier et al., Reference Mittermeier, Robles Gil, Hoffman, Pilgrim, Brooks, Goettsch Mittermeier, Lamoureux and Da Fonseca2004; Pillon and Munzinger, Reference Pillon and Munzinger2005). Over 74% of species are endemic and this unusually high level of endemism can be explained, at least in part, by the presence of nutrient impoverished ultramafic soils (Jaffré et al., Reference Jaffré, Morat, Veillon and MacKee1987). The flora occurring on these substrates is highly distinct and localized in terms of species richness and endemicity compared to other vegetation communities in New Caledonia, and in many cases has more affinities to Australian taxa than to local Pacific species (Jaffré et al., Reference Jaffré, Rigault, Dagostini, Tinel-Fambart, Wulff and Munzinger2009). Indeed, many of the dominant families on these ultramafic soils are also significant elements of the Australian flora, with many having undergone intense radiation on these soils (Setoguchi et al., Reference Setoguchi, Osawa, Pintaud, Jaffré and Veillon1998; Howarth et al., Reference Howarth, Gustafsson, Baum and Motley2003). Plants in these soils typically have very low growth rates and may also be prone to fire during the dryer season (July to November) due to their sclerophyllous nature (Jaffré et al., Reference Jaffré, Rigault and Dagostini1998).
Little is currently known about the germination requirements of New Caledonian taxa, as few studies have so far been reported (Fogliani et al., Reference Fogliani, Bouraïima-Medjebi, Medevielle and Pineau2004). In addition, virtually nothing is known about the responsiveness of New Caledonian species to fire or the by-products of fire, such as smoke. Many of the families common in the Australian flora, known for their smoke responsiveness, are also remarkably common in New Caledonia. This includes species of Ericaceae (Rokich et al., Reference Rokich, Dixon, Sivasithamparam and Meney2002), Rutaceae (Dixon et al., Reference Dixon, Roche and Pate1995), Cyperaceae (Thomas et al., Reference Thomas, Morris and Auld2003), Dilleniaceae (Dixon et al., Reference Dixon, Roche and Pate1995) and Goodeniaceae (Dixon et al., Reference Dixon, Roche and Pate1995). Given the rapid development of mining operations on these highly speciose ultramafic soils, there is a pressing need to understand the underlying seed ecology of New Caledonia species and to develop the appropriate restoration framework to reconstruct these damaged environments once mining has ceased. In this short communication we report the effect of smoke water on the germination of two endemic New Caledonian scrubland species occurring on ultramafic soils: Hibbertia pancheri (Dilleniaceae) and Scaevola montana (Goodeniaceae) to determine whether smoke responsiveness is found in an aseasonal tropical context.
Materials and methods
Seed collection
Mature seeds of H. pancheri (Dilleniaceae) were collected on 24 April 2010, near Prony in the South province of New Caledonia (22°17′51″S, 166°42′31″E) at an elevation of 30 m. The fleshy aril was removed. Mature S. montana (Goodeniaceae) drupes were collected on 23 December 2010, in the Tontouta valley in the South province (21°56′01″S, 166°17′38″E) at an elevation of 150 m. The fleshy mesocarp was removed in order to expose the endocarp. In this paper the word drupe will be used to define the endocarp and the seeds it contains. All treated seeds were then stored at ambient room conditions (approximately 26°C and 50% relative humidity) prior to experimentation. Experiments and germination studies started on 12 May 2010 (H. pancheri) and on 27 December 2010 (S. montana), and ran for a total of 57 and 58 weeks, respectively.
Viability
Seed viability of S. montana was estimated using a tetrazolium test (ISTA, 2003). In contrast, this test could not be done on seeds of H. pancheri due to the very small size of the embryo ( < 300 μm). Three replicates of 20 drupes were cut transversally to determine the mean number of embryos per drupe. These embryos were then placed on a filter paper imbibed with a 1% chloride tetrazolium solution (TTC) for 24 h in darkness at room temperature to estimate their viability, based on the presence of uniform red colouration. Viability was expressed as the percentage of viable embryos per 100 drupes.
Drupe, seed and embryo characteristics
Endocarp, seed stricto sensu and embryo structures were observed with a scanning electron microscope and a stereomicroscope equipped with an ocular micrometer. On a total of 30 seeds, the size of the embyo and the embryo to seed ratio (E:S) were calculated according to Forbis et al. (Reference Forbis, Floyd and de Queiroz2002). Seeds were soaked in water for 24 h to facilitate the longitudinal cut, realized with a scalpel, to calculate the E:S.
Germination tests
For all germination tests, the seeds of H. pancheri and drupes of S. montana were disinfected in a solution of 1.5% sodium hypochlorite for 20 min and then rinsed five times in sterile distilled water. Three replicates of 50 seeds (H. pancheri) or 50 drupes (S. montana) received the following treatments: (1) 24 h of soaking in distilled water (control); (2) 24 h in a solution of diluted smoke water (1/10, commercial solution Regen™); and (3) 24 h in a solution containing gibberellic acid (GA3) at a concentration of 0.3 g l− 1. These soakings were all performed at room temperature. After soaking, the seeds of H. pancheri and drupes of S. montana were positioned on filter paper in Petri dishes, regularly moistened with distilled water, placed in incubators with daily light and temperature variation (12 h of light at 30°C and 12 h of dark at 20°C, conditions that occur in the natural habitat of the species in summer, a period with important rainfalls when a majority of species of ultramafic scrublands fruit) and checked for emergence every week. The average germination percentages obtained on three replicates of 50 seeds for H. pancheri were determined. For S. montana three replicates of 50 drupes, i.e. 100 seeds per replicate (two potential seeds per drupe), were monitored. Final germination rates of this species were determined on the mean number of seeds found per replicate.
Statistical analysis
Germination percentages were all arcsin square root transformed and then analysed with a one-way analysis of variance (ANOVA), followed by Tukey's honestly significant difference (HSD) test (P < 0.05). All analyses were performed with the statistical software R, version 2.10.1 (R Development Core Team, 2011). Non-transformed data appear in all figures.
Results
Viability
Drupes of S. montana contained two locules with 71.6 ± 3.1% containing a seed. TTC assessment found that 24.5 ± 3.5% of these seeds stained red (were viable). Based on these characteristics there were approximately 35 viable seeds per 100 drupes.
Embryo characteristics
The seeds of H. pancheri were found to be endospermous (Fig. 1B, C) with an underdeveloped embryo 0.2 ± 0.1 mm long. The E:S ratio between embryo and seed length was 0.08 ± 0.02. S. montana seeds were also found to be endospermous with a linear-type embryo, 1.3 ± 0.1 mm long. The E:S ratio in S. montana was much higher than that of H. pancheri, 0.69 ± 0.11.
(A) Seeds of H. pancheri. (B) Scanning electron microscope observation of a longitudinal section of a seed of H. pancheri. (C) Stereomicroscope observation of a longitudinal section of a seed of H. pancheri. (D) Fruits and drupes of S. montana. (E) Scanning electron microscope observation of a transverse section of a drupe of S. montana. (F) Stereomicroscope observation of a longitudinal section of the endocarp and seed of S. montana. (A colour version of this figure is available online at http://www.journals.cambridge.org/ssr)
Germination
Hibbertia pancheri
Germination of H. pancheri seeds began after 50 d incubation and continued for the next 320 d (Fig. 2A). The best treatment for stimulating germination of H. pancheri seeds was soaking in a solution containing a commercial brand of smoke water (Regen™) for 24 h prior to water incubation; germination following smoke-water treatment reached 88 ± 7.1% after 400 d (Fig. 2A). Seeds exposed to this treatment also germinated more rapidly (P < 0.01): ~58% germination after 100 d of incubation (compared to < 20% for the other two treatments assessed). In comparison, untreated seeds germinated well (70.0 ± 5.8%) although the rate of germination was more gradual. Seeds soaked in a GA3 solution also germinated to a high degree (60.7 ± 6.2%) but there was no significant difference (P>0.05) in germination between the water control and GA3-treated seeds in terms of either total germination or germination rate.
Effects of smoke water and GA3 on the germination of H. pancheri (A) and S. montana (B) seeds. Control seeds have not received any soaking. Standard errors are indicated for each monitoring event.
Scaevola montana
Germination of S. montana seeds was initially much slower than that of H. pancheri seeds, with germination across all three treatments assessed remaining below 10% for the first 230 d of incubation (Fig. 2B), although the first germination event was observed following only 43 d incubation. Between 230 and 370 d, germination across all three treatments increased at a significantly higher pace, going from < 10% up to 71%. The best treatment identified for S. montana seeds was soaking in a solution containing smoke water (71%, Fig. 2B) prior to water incubation (P < 0.05), with the water and GA3 treatments resulting in much lower germination (23–34%).
Discussion
Smoke water had a significant influence on the germination of both New Caledonian species assessed in this preliminary study. For S. montana seeds the total number of germinants was significantly higher after soaking in smoke water than in the other two treatments assessed, whereas for H. pancheri seeds smoke water significantly accelerated the rate of germination. Germination responses to smoke or smoke water have been clearly demonstrated for several species of Hibbertia and Scaevola in previous studies (Roche et al., Reference Roche, Koch and Dixon1997; Hidayati et al., Reference Hidayati, Walck, Merritt, Turner, Turner and Dixon2012). Roche et al. (Reference Roche, Koch and Dixon1997) found that the application of aerosol smoke to punnets sown with Hibbertia and Scaevola seeds resulted in higher germination for both genera. Interestingly, smoke-induced germination was also significantly enhanced when combined with a period of soil ageing (in punnetts), with seeds left for >12 months becoming both more germinable as well as more smoke responsive over time. For example, initial germination of Hibbertia commutata was 0.0% and 2.4% in untreated and aerosol smoke-treated seeds, but following 12 months' soil storage this had increased to 2.8% and 76.5%, respectively. Likewise, 0.0% and 3.3% of Scaevola fasiculata seeds germinated in response to the same treatments but, again, following 12 months' soil storage germination had also risen to 0.6% and 92.5% (Roche et al., Reference Roche, Koch and Dixon1997).
As with H. commutata and S. fasiculata, both H. pancheri and S. montana also appear to have some form of seed dormancy, as germination initially remained very low for several months. Following this initial lag phase, the pace of germination was observed to increase, although it was still quite staggered and prolonged, taking between 320 and 400 d for maximum germination to be achieved – similar in many respects to the time taken (>1 year) for high germination (>70%) to occur in H. commutata and S. fasiculata seeds. Seeds are defined as being dormant if they fail to germinate within 28 d (Baskin and Baskin, Reference Baskin and Baskin2004a, Reference Baskin and Baskinb) so, clearly, both H. pancheri and S. montana have dormant seeds, with germination only occurring as seeds gradually come out of dormancy. The smoke response is also closely linked to seed dormancy, with seeds of many species needing to lose dormancy first through afterripening or stratification (Merritt et al., Reference Merritt, Turner, Clarke and Dixon2007; Turner et al., Reference Turner, Merritt, Renton and Dixon2009) before germinating in response to the presence of smoke or smoke products. For H. pancheri, the type of dormancy is likely to be morphophysiological, as the seeds of this species possess tiny embryos < 300 μm in length (Fig. 1B, C) which likely grow within the seed before radicle emergence, as demonstrated recently for other species of Hibbertia (Hidayati et al., Reference Hidayati, Walck, Merritt, Turner, Turner and Dixon2012), as well as having some form of physiological dormancy that must be overcome through afterripening or stratification to empower the embryo to grow. S. montana, on the other hand, has a well-developed embryo (Fig. 1E, F), making it likely that physiological dormancy suppresses fresh seed germination. As with Hibbertia, dormancy is likely to be overcome through afterripening or stratification. Preliminary results using a ‘move along’ experimental approach (Baskin and Baskin, Reference Baskin and Baskin2003) did result in slightly higher germination in S. montana seeds, but further work is required to confirm these observations (A. Wulff, unpublished data).
While the results presented in this study are very preliminary at this stage, they do hint at the likely significance of fire, and in particular smoke and its components, on key elements of the New Caledonia flora. Understanding how smoke interacts with seed dormancy and seed ecology will no doubt improve current conservation efforts for those fragile and highly localized species that occur nowhere else on Earth. Indeed, New Caledonia is currently undergoing rapid mining expansion, centred on these ultramafic soils, the very places where much of the New Caledonian biodiversity is presently found. While the economic pressures to mine intensify, it is essential to gather as much knowledge as possible to aid restoration efforts in order to preserve these exceptionally diverse ecosystems in perpetuity.
Acknowledgements
We would like to thank Stephane McCoy, Thomas Le Borgne and Casimir Véa for seed collection and monitoring. We are also thankful to Jacques Rabier for scanning electron microscope observations and to two anonymous reviewers for providing constructive comments on an earlier version of the manuscript. We would also like to thank Vale Nouvelle-Calédonie, the Government of New Caledonia and the Agronomic Institute of New Caledonia for financial support.
