SEARCH

SEARCH BY CITATION

Keywords:

  • climate change;
  • germination niche;
  • seed germination;
  • south-west Australia;
  • temperature gradient plate;
  • vulnerability

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

In obligate seeding species, the germination niche is crucial for colonization and population survival. It is a high-risk phase in a plant's life cycle, and is directly regulated by temperature. Seeds germinate over a range of temperatures within which there is an optimum temperature, with thresholds above and below which no germination occurs. We suggest that abrupt changes in temperature associated with a warming climate may cause a disconnect between temperatures seeds experience and temperatures over which germination is able to occur, rendering obligate seeding species vulnerable to decline and extinction. Using a bidirectional temperature gradient system, we examined the thermal constraints in the germination niche of some geographically restricted species from the low altitude mountains of the Stirling Range, southern Western Australia, including seedlots from lowland populations of four of these species. We demonstrated that high temperatures are not a limiting factor for germination in some restricted species, signifying a lack of relationship between geographic range size and breadth of the germination niche. In contrast, we identified other restricted species, in particular Sphenotoma drummondii, as being at risk of recruitment failure as a consequence of warming: seeds of this species showed a strong negative relationship between percentage germination and increasing temperature above a relatively low optimum constant temperature (13°C). We found some ecotypic differences in the temperature profiles between seeds collected from montane or lowland populations of Andersonia echinocephala, and while specific populations may become more restricted, they are perhaps at less risk of extinction from climate warming. This seed-based approach for identifying extinction risk will contribute tangibly to efforts to predict plant responses to environmental change and will assist in prioritizing species for management actions, directing limited resources towards further investigations and can supplement bioclimatic modelling.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Climatic variables influence reproductive traits such as seed germination and seedling growth (Fenner & Thompson 2005). Although soil moisture availability is critical for seedling emergence and survival, temperature is arguably the most important climatic variable as it synchronizes germination to environmental conditions most suitable for seedling establishment, such as high likelihood of adequate water availability (Bell 1994; Fenner & Thompson 2005). Temperature affects seed responses through regulating dormancy loss in both dry and moist dormant seeds, and germination rate in non-dormant seeds (Probert 2000).

Germination and seedling establishment are the highest-risk phases in the life cycle of plants (Harper 1977; Silverton & Charlesworth 2001). Consequently, mechanisms that minimize the risk will be under strong selection pressure (Meyer et al. 1997), with natural selection favouring seed germination patterns that increase the probability of successful seedling establishment. For example, although temperature is not considered to be a limiting factor for recruitment in Mediterranean-type ecosystems (Lloret et al. 2004), germination generally coincides with early winter when temperatures are low (typically 10–15°C; Bell 1999) and rainfall is reliable (Mott 1972; Bellairs & Bell 1990; Bell & Bellairs 1992; Bell et al. 1995). This maximizes the period of root development prior to summer drought, thereby increasing seedling survival (Bell 1999). For many species, temperature-mediated germination cues are important to ensure population persistence. Consequently, environmental change, including anthropogenic climate warming, might be expected to have significant effects on plant distribution and survival as the timing of this key life-history transition of seedling emergence, may no longer be optimal (Lloret et al. 2004; Gworek et al. 2006).

Sensitivity to climate change for a given species will depend on its ecological niche and geographic distribution (Thuiller et al. 2005; Broennimann et al. 2006), life-history characteristics, and genetic and phenotypic plasticity (Aitken et al. 2008). Germination niche breadth may influence the ecological breadth and geographic ranges of species (Donohue et al. 2010) and species with a broad germination niche might be expected to have larger geographic ranges than species with narrow germination niches, although this remains unconfirmed (Thompson & Ceriani 2003). Most species can tolerate some short-term variability in climate through phenotypic plasticity, although predicted responses to global warming generally involve shifts in either seasonality (Jump & Peñuelas 2005; Bradshaw & Holzapfel 2008) or, for alpine species, altitude (Gworek et al. 2006; Parolo & Rossi 2008), rather than the shifts in thermal tolerance that would be necessary for long-term persistence. In addition, those species most vulnerable to a warming climate may be those either with restricted niches, in regions with high anomalies in climate, or with physical barriers to migration (Broennimann et al. 2006; Malcolm et al. 2006). Plants with limited dispersal ability, low reproductive rates (Jump & Peñuelas 2005), obligate seeding (Bell 2001; Bond & Midgley 2001), and those already restricted to the physical limits of their range, such as to higher latitudes and upper montane locations (Walther et al. 2002; Parmesan & Yohe 2003; Pauli et al. 2003; Root et al. 2003), are presumed to be most threatened by climate change.

Fire-prone, species-rich Mediterranean regions of the world with their cool wet winters and hot dry summers have been recognized as potentially highly vulnerable to climate change (Allen 2003; Cowling et al. 2004). Increasing temperatures and changes in amount and variability of rainfall may also alter fire frequency, further endangering this rich plant diversity (Pittock 2003; Pitman et al. 2007). Under projected climate change scenarios, there are suggestions that an almost total collapse of flora endemic to the south-west region of Western Australia will occur (Hughes et al. 1996; Pouliquen-Young & Newman 2000; Malcolm et al. 2006; Fitzpatrick et al. 2008). Similar negative impacts on native flora have been predicted for species in the Cape Floristic region of South Africa (Midgley et al. 2002; Hannah et al. 2005; Broennimann et al. 2006). However, these predictions are based on the response of adult plants to edaphic conditions.

Few investigations have considered the effect of climate warming on seed traits or the germination niche, despite the insights that can be provided by studying the influence of temperature on seed dormancy and germination in the context of environmental change (Ooi et al. 2009). Investigation of germination temperatures for seeds of species from south-west Western Australia has been substantial, but mostly based on germination at a limited range of constant temperatures (Bell et al. 1995). For these species, temperature optima are generally low, indicative of winter temperatures when seedling establishment and survival is likely to be optimal (Bell 1999). However, germination and early seedling responses to temperature for species from Western Australia's low southern mountains have not previously been investigated. Species from this region are likely to be at elevated risk from climate change as there is no opportunity for further upward or southward migration and this region has a greater predicted rise in temperature compared with the rest of south-west Australia (Hope et al. 2006). We use these species for closer examination because of their physical limits to distribution spread and their likelihood of facing extinction risk. They provide a case study for assessing direct physiological constraints on recruitment, for example the width of the germination niche, in particular the width of the temperature window for germination and growth. Recruitment constraints may provide a measure of susceptibility to climate change: species that only germinate over a narrow temperature window are potentially most at risk. Temperatures outside that window represent no, or limited, opportunity for germination.

We report on the thermal limits for germination and early seedling growth for 10 perennial plant species endemic to southern Western Australia, with six species restricted to the summits of the major mountain range in the region (the Stirling Range; maximum elevation <1100 m a.s.l.). We determined (i) whether a restricted geographic distribution and mountain habitat can explain thermal constraints in the germination niche; (ii) whether thermal constraints for early seedling root growth follow the same pattern as for germination; and (iii) whether the niche limits are different between lowland and montane populations. Because niche characteristics can be powerful indicators of species sensitivities to climate change, we discuss our data in the context of developing predictors of extinction risk under climate warming.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Species and site details

Seeds of 10 perennial species (Table 1) were collected from the Stirling Range (34°22′ S; 118°15′ E; the major mountains of southern Western Australia. Six of the species are restricted to peaks of the Stirling Range. The mountains are predominantly sandstone and quartzite with thin layers of slate and phyllite (Semeniuk 1993) and the peaks of the range are more than 1400 km from the nearest similar mountain peaks. Consequently, these mountains are effectively islands in an otherwise largely flat landscape. The study species are all found within the threatened thicket vegetation community which grows on the upper levels of all the major peaks. Most species rely on sexual reproduction for population maintenance; six of the 10 species rely on formation of a soil-stored seed bank, with the remaining four species holding seed reserves in the plant canopy (Appendix S1). All species generally depend on disturbance (fire) for regeneration, an adaptation for survival in fire-prone Mediterranean ecosystems.

Table 1.  Seed mass and germination profile data for seeds from 10 species from south-west Western Australia: optimum constant temperature regime for germination and radicle growth, slope of the linear logistic regression of germination at constant temperatures above the optimum (Fig. 1), optimum alternating temperature regime for germination, and the mean time to germinate (MTG) at the optimum constant and alternating temperature regime
SpeciesPopulation (harvest)Seed mass (mean ± SE) (mg)Optimum constant germinationRadicle growth rate (mm day−1) at (in parentheses) the optimum temperature (°C)Optimum alternating germination
Temperature (oC)MTG (days)Germination (%)Viability (%)SlopeDay/night temperature (°C)MTG (days)Germination (%)Viability (%)
  • Seed collected from either montane (M) or lowland (L) populations. Calothamnus crassus was collected from the same population in two different years, as indicated. Kunzea montana was collected from two montane populations, as indicated.

  • Figure in parentheses is the optimum constant temperature for radicle growth determined from fitting a quadratic equation to the raw data (Appendix S4). BK, Bluff Knoll; EB, East Bluff.

Allocasuarina decussataM5.25 ± 0.20121198698−0.990.83 (19)15/1522.68083
L6.14 ± 0.2552123.52691−0.590.42 (19)9/1337.33090
13/2229.92386
22/2637.12388
Andersonia echinocephalaM0.26 ± 0.0341330.182100−1.160.60 (17)11/643.1100100
7/1631.89595
11/1136.49595
L0.16 ± 0.0201526.238100−0.610.31 (15)19/935.17090
Banksia browniiM44.09 ± 1.5501525.27092−0.580.71 (17)9/1147.34060
L27.49 ± 1.1171523.94367−0.231.00 (17)11/2243.24073
Calothamnus crassusM (2001)0.18 ± 0.0601321.59597−0.350.63 (16)15/922.8100100
M (2006)0.16 ± 0.0501323.2100100−0.540.63 (16)14/626.795100
Deyeuxia drummondiiM0.35 ± 0.0151510.49090−0.281.32 (17)22/1512.2100100
Eucalyptus megacarpaM4.01 ± 0.1992314.59598−1.311.47 (24)20/241578100
L3.47 ± 0.1992115.492100−0.771.65 (23)18/2214.893100
Gastrolobium leakeanumM7.09 ± 0.2421912.28888−0.520.78 (19)11/2418.6100100
Kunzea montanaM (BK)0.23 ± 0.0591521.39798−0.150.70 (23)26/1517.6100100
22/2018.0100100
M (EB)0.23 ± 0.0651523.28082−0.110.92 (24)24/1814.3100100
Sphenotoma drummondiiM0.03 ± 0.0101327.77088−1.130.38 (14)11/933.67080
Velleia foliosaM1.39 ± 0.0521021.983100−0.221.07 (20)13/1313.5100100
9/2616.2100100

While the Stirling Range experiences a typically Mediterranean climate, the mountains have a strong influence on the local environment. Average annual rainfall at the Park Ranger Station (34°24′ S; 118°03′ E) is 551 mm. Rainfall averages vary greatly with elevation, with the highest annual rainfall (about 1000 mm) estimated to be on the eastern peaks (Courtney 1993). The exposed higher peaks exhibit more extreme temperatures with extended drizzle through summer and even occasional snow during winter. Day temperatures are usually 5°C cooler on the peaks than on the plains (Tmax 27°C in January; Tmin 15°C in July) and the eastern peaks are often shrouded with low cloud. Night-time temperatures are usually warmer on the peaks than on the plains (Tmax 13°C in January; Tmin 6°C in July) as cold air drains down the slopes (Courtney 1993).

Three species (B. brownii, A. decussata and E. megacarpa) also occur outside the Stirling Range and seedlots from coastal populations, c. 100 km south of the Stirling Range, were included in this investigation. On the coast, the annual rainfall averages 929.8 mm (1877–2007) although for more recent recording (1971–2007) averages just 875.5 mm. Rainfall is concentrated mainly between May and September (autumn–winter). February and July are the hottest and coolest months, respectively, with mean maximum temperatures of 22.9 and 15.4°C and mean minimum temperatures of 15.7 and 8.1°C, respectively (Australian Government Bureau of Meteorology: Albany station 35°1′48′′S; 117°52′48′′E; elevation 3 m).

Seeds were collected between December 2000 and April 2007. Collections made prior to 2007 were dried at 15°C and 15% relative humidity and stored at −20°C in sealed laminated foil bags until use in April 2007. Bags were allowed to warm to laboratory temperature (about 20°C) before opening. More recent collections were similarly dried, but not stored at −20°C prior to experiment set-up.

Mean seed mass (after drying as above, then equilibrated to laboratory temperature, approximately 20°C) was determined to the nearest µg by weighing 20 individual seeds of Banksia brownii, Eucalyptus megacarpa, Velleia foliosa, Allocasuarina decussata, Gastrolobium leakeanum and Deyeuxia drummondii or five replicates of 10 seeds each, for the smaller-seeded Calothamnus crassus, Andersonia echinocephala, Kunzea montana and Sphenotoma drummondii.

Experimental design

A bidirectional temperature gradient plate (Model GRD1, Grant Instruments, Cambridge, UK) was used to provide either 12 constant temperatures or 30 alternating and six constant temperatures ranging from 6°C to 32°C. When using the temperature gradient plate in one direction (12 constant temperatures), two replicates of 30 seeds each were used per collection per temperature, with the exception of S. drummondii, G. leakeanum and A. echinocephala (lowland population; L) that had two replicates of 20 seeds each. An incubator was used to obtain the lower limit of 5°C. When seed availability was limiting (because of the rare and threatened nature of some species), two replicates of 30 seeds (B. brownii), or three replicates of 10 seeds each (V. foliosa and D. drummondii), were placed in germination incubators at six constant temperatures (5, 10, 15, 20, 25 and 30°C). Subsequently, the temperature gradient plate was used to obtain a range of diurnally alternating temperatures ranging from approximately 5°C to 35°C. Because of space constraints on the temperature gradient plate, or seed availability, one 50-mm Petri dish per temperature per collection contained 10 (S. drummondii, A. echinocephala), 15 (B. brownii), 20 (K. montana, E. megacarpa, C. crassus, D. drummondii, G. leakeanum and A. echinocephala) or 30 seeds (V. foliosa and A. decussata).

Seeds were sown on the surface of 1% w/v water agar on 50-mm Petri dishes except for seeds of K. montana that were incubated on two layers of Whatman No. 2 filter paper over thick germination paper. The filter papers were moistened with 2 mL of 10−7 M butenolide solution (3-methyl-2Hfuro[2,3-c]pyran-2-one), the major germination stimulant in smoke; previous studies on K. montana germination have indicated a requirement for smoke stimulation (Cochrane unpubl. data, May 2002). The butenolide used was isolated, purified and identified from smoke-saturated water derived from burned Passerina vulgaris Thod. and Themeda triandra L. as described by van Staden et al. (2004). Other dormancy-breaking treatments used included the addition of gibberellic acid to the agar for S. drummondii (as GA3 at 25 mg L−1) and GA3 and butenolide (at 250 mL L−1) for A. echinocephala; these treatments have been found to produce optimal results for those species (Cochrane unpubl. data, December 1998). Seeds of G. leakeanum have a water-impermeable seed coat (physical dormancy sensuBaskin & Baskin 2001) and required scarification of the seed coat prior to imbibition and radicle emergence. This was achieved by removing a small portion of the seed coat with a scalpel along the longitudinal axis of each seed. Actual temperatures experienced by seeds on the temperature gradient plate were determined by placing five bead thermistors connected to a data logger (Grant Instruments Squirrel 1000 Series) on the surface of the germination medium in the Petri dishes in each corner and in the middle of the temperature gradient plate.

All environments received irradiance for 12 h day−1. Germination was scored every other day and defined as visible radicle emergence. Germination tests ran for up to 40 (constant temperatures) or 55 days (alternating temperatures). Non-germinated seeds were subjected to a cut-test to determine potential viability; percentage germination was calculated after subtracting the number of empty seeds from the number of seeds sown.

Radicle growth rates

To determine the temperature at which radicle growth extension rate (RGER) was maximal, 12 germinated seedlings from each collection germinated under constant temperature conditions were transferred to an agar plate held at the same temperature and under identical light conditions either on the temperature gradient plate or in incubators. Root length was measured to the nearest mm with an eye piece reticule calibrated with a stage micrometer under the microscope (small seedlings) or with vernier callipers (large seedlings) every 1–2 days for 5 days (at least three times per 5-day period) and the mean radicle growth rate RGER calculated for each temperature.

Statistical analysis and data presentation

Statistical analysis was carried out using GenStat 11.1 (VSN International Ltd., UK). T-tests of log10 transformed data (for normality, confirmed by the Shapiro–Wilk test) were used to compare the mass of seeds of the same species collected from different populations, or from the same population in different years. Mean time to germinate (MTG) was calculated for the optimum constant and optimum alternating temperature regime (highest percentage germination or most rapid if more than one regime resulted in the same maximum percentage germination) according to the equation:

  • image

where: n = number of seeds germinated between scoring intervals, d = the incubation period in days at that time point and N = total number of seeds germinated.

Binary logistic regression analysis (logit link function) was used to assess the effect of increasing temperature above the optimum (constant temperatures only) on the proportion of seeds that germinated (linear model). Origin 6.1 (Microcal) was used to draw contour plots showing germination on the bidirectional temperature gradient plate. χ2 tests (maximum likelihood) were used for pairwise comparison of the proportion of seeds that germinated under the optimum constant or optimum alternating temperature regimes within each seedlot, or of the proportion of seeds that germinated from seedlots collected from different populations within species. For each seedlot, the optimum and limiting temperatures for radicle growth were determined by fitting a quadratic relationship to the raw data.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Seed mass

Seed mass ranged between 0.03 mg for S. drummondii and 44.09 mg for the montane population of B. brownii (Table 1). Mean seed mass for A. decussata was significantly greater for seeds from the lowland populations compared with seeds from the montane populations (< 0.01). Conversely, for B. brownii, E. megacarpa and A. echinocephala, seeds from the montane populations had significantly greater mean seed mass than seeds from the lowland populations (< 0.05). In the case of C. crassus, mean seed mass was significantly greater in 2001 compared with 2006 (< 0.01). There was no difference in the mean mass of seeds collected from two different montane populations of K. montana.

Germination at constant temperatures

The optimum constant germination temperature (i.e. temperature where the proportion of seeds that germinated was greatest) ranged from 10°C to 23°C for the 16 seedlots studied (Table 1; Appendix S2). There was little or no difference in the optimum germination temperature determined for seeds collected from montane or lowland populations of A. decussata (21°C), A. echinocephala (13–15°C), B. brownii (15°C) and E. megacarpa (21–23°C), or from different montane populations of K. montana (15°C). Similarly, the optimum germination temperature for the older (2001) and more recent (2006) seedlots of C. crassus did not differ (13°C).

The proportion of seeds that germinated at the optimum germination temperature was generally high (>80%; Table 1; Appendix S2). However, there was significantly lower germination of seeds from the lowland populations of A. decussata, A. echinocephala and B. brownii (≤43%) compared with seeds of the same species from the montane populations over the same time period (≤ 0.005).

For some seedlots, for example those of K. montana, there was a relatively wide range of constant temperatures where high levels of germination were attained (13–29°C; Appendix S2). However, for other seedlots, for example those of A. echinocephala and S. drummondii, germination was confined to a narrow range of temperatures (11–15°C; Appendix S2). Logistic regression analysis indicated significant relationships between percentage germination and temperature as it increased above the optimum (< 0.001; Fig. 1). The slope of this relationship varied between species, ranging from −0.11 (K. montana; relatively small effect of increasing temperature above the optimum on percentage germination) to −1.13 (S. drummondii; large effect of increasing temperature; Table 1).

image

Figure 1. The results of logistic regression analysis showing the linear relationship between final germination and temperature above the optimum constant temperature for seeds of A. decussata, A. echinocephala, B. brownii, and C. crassus and D. drummondii, E. megacarpa, G. leakeanum, K. montana, S. drummondii, and V. foliosa. Solid lines and symbols correspond to seedlots from montane populations unless otherwise indicated in the legend; dashed lines and hollow split symbols correspond to seedlots from lowland populations unless otherwise indicated. Arrows indicate 0% germination.

Download figure to PowerPoint

Seeds placed at suboptimal temperatures were slower to germinate. They often had a longer delay before germination commenced and a higher MTG, and generally exhibited lower total percentage germination at the time when the tests were terminated (Appendix S2).

Seedling radicle growth rates and temperature

The quadratic relationship between radicle growth rate and temperature for each seedlot clearly showed how radicle growth occurred over a broad range of temperatures for some species, and a narrow range for others (Appendix S3). Temperatures most favourable for radicle growth rates for the 10 species ranged from 15 (A. echinocephala (L)) to 24°C (K. montana (EB population)) (Table 1). In general, the optimum temperature for radicle growth was higher than that for germination for all seedlots with the exception of A. decussata, A. echinocephala (L) and G. leakeanum (Table 1). For A. echinocephala (L) and G. leakeanum, the temperature most favourable for radicle growth did not differ from that for germination. For montane populations of both A. decussata and A. echinocephala radicle growth rate was twice that of seedlings from lowland populations at the optimum temperature. For most species, the temperature at which radicle growth rate peaked showed little difference to that for optimum germination. The seedlot with the greatest difference between the optimum temperature for germination and that for radicle growth was V. foliosa. Radicle growth rate was fastest at 20°C (1.07 mm day−1) and relatively slow (0.49 mm day−1) at the temperature which produced optimum germination (10°C).

Germination at alternating temperatures

The contour plots (Figs 2,3) show the germination response of each seedlot across a range of both constant and alternating temperature regimes. The preference for low germination temperatures as seen in the constant germination experiments for A. echinocephala, C. crassus, S. drummondii and V. foliosa is also apparent as ‘high’ isopleths (light filling) towards the bottom-left corner of the contour plots (Figs 2,3). In contrast, E. megacarpa seeds showed a preference for higher germination temperatures (Fig. 3) while D. drummondii, G. leakeanum, K. montana and V. foliosa germinated across a wide range of both the constant and alternating bidirectional temperature gradient plate regimes (Fig. 3).

image

Figure 2. Contour plots, with points of equal percentage germination connected by germination isopleths, for seeds of A. decussata, A. echinocephala, B. brownii and C. crassus. The shading calibration on the right hand side of the figure is for percentage germination: light filling isopleths represents high germination, dark filling low or no germination. Seeds were collected from either montane or lowland populations, as indicated (Appendix S1). Constant temperatures are presented on the diagonal line from the bottom-left corner of the diagrams (lowest temperature approximately 6°C) to the top-right corner (maximum temperature approximately 34°C). All points above and below the diagonal line represent alternating temperature regimes, with greatest amplitude at the top-left and bottom-right corners of each graph.

Download figure to PowerPoint

image

Figure 3. Contour plots, with points of equal percentage germination connected by germination isopleths, for seeds of D. drummondii, G. leakeanum, E. megacarpa, K. montana, S. drummondii and V. foliosa. Presentation is as in Figure 2.

Download figure to PowerPoint

In general, the maximum proportion of seeds that germinated under a constant temperature regime (using incubators or the temperature gradient plate unidirectionally) was similar to the maximum proportion observed when the seeds were placed on the bidirectional temperature gradient plate (Figs 2,3), with a few exceptions (Table 1). There was greater maximum germination of seeds from both populations of A. echinocephala when they were sown on the bidirectional temperature gradient plate. For seeds from the montane population, there was 100% germination at 11/6°C compared with a maximum of 82% germination at a constant 13°C (χ2 = 6.89, < 0.01). For seeds from the lowland population, there was 70% germination at 19/9°C compared with 38% germination at 15°C, although this difference was not statistically significant (because of low seed numbers for the bidirectional temperature gradient plate experiment; > 0.05). Seeds from both populations of K. montana also showed a preference for alternating temperatures, with a higher proportion of germinating seeds and a faster rate of germination (lower MTG; Fig. 3, Table 1). For seeds from the East Bluff population, the difference in germination between the optimum alternating temperature regime (24/18°C) and the optimum constant temperature (15°C) when seeds were sown on the unidirectional temperature gradient plate, was significant (χ2 = 7.59, < 0.01). In contrast there was significantly lower maximum germination of B. brownii (M) seeds (χ2 = 4.52, < 0.05) and E. megacarpa (M) seeds (χ2 = 6.81, < 0.01) on the bidirectional temperature gradient plate compared with the constant temperature regimes (incubators or unidirectional temperature gradient plate for B. brownii and E. megacarpa, respectively) (Table 1).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Predictions of changes in species geographic distribution as a consequence of climate change assume that plants survive within their current range because the existing soil and climate conditions are optimal, and that shifts in distribution may be predicted under various climate change scenarios when conditions change (Thomas et al. 2004; Malcolm et al. 2006; Fitzpatrick et al. 2008). However, such modelling does not consider vulnerability (likelihood of a decrease in range) as a consequence of climate change at particular stages in a plant life cycle, or the implications of suboptimal conditions on life stage transitions. As germination and seedling establishment are the highest-risk phases for plants (Harper 1977; Silverton & Charlesworth 2001), germination temperature data, such as presented here, may help to identify species most at risk in a changing climate.

Under projected scenarios, by 2030 there will be a 0.5–2.1°C increase in summer temperatures, a 0.5–2.0°C increase in winter temperature and a 2–20% decrease in rainfall across south-west Western Australia (IOCI 2009). A 2°C rise in temperature has already been shown to impact on seed responses including seed production, seed mass, seedling emergence and establishment and soil seed bank dynamics (Williams et al. 2007; Hovenden et al. 2008; Hedhly et al. 2009; Ooi et al. 2009). Changes in recruitment can lead to population decline and alter vegetation community patterns (species composition and abundance), paving the way for species invasions (Parolo & Rossi 2008). Levels of recruitment may be affected by total germination and/or the rate of germination. Some species may increase these parameters as temperatures rise, while in others they may be reduced (Williams et al. 2007; Milbau et al. 2009). Rising temperatures may reduce the rate of dormancy loss in seeds of temperate or alpine species in the soil seed bank, particularly those needing cold stratification before germination occurs (Steadman & Pritchard 2003), or in arid zone species (Ooi et al. 2009). Perhaps species at most risk will be those with complex dormancy which need to go through seasonal cycles (Lolium rigidum: Steadman et al. 2004).

Plant traits, thermal constraints and the germination niche

Here, we have considered first, whether or not the species relies on its seed for propagation and second, the temperature constraints for seed germination. First, our data show a relationship between seed mass and altitude for three of the seedlot pairs investigated (B. brownii, A. echinocephala and E. megacarpa). Seed mass is considered a key influencing factor for seed survival and dispersal (Harper 1977). Altitude has been shown to have an effect on seed mass with a trend for lowland species to have smaller seeds than upland species (Murray et al. 2003). As temperature and moisture decrease with altitude, seed size increases. Three of the four small-seeded species (A. echinocephala, C. crassus and S. drummondii with <0.3 mg seed mass) exhibited a ‘typical’ Mediterranean germination syndrome with low temperature requirements for germination (Bell & Bellairs 1992). Species with small seeds are reported to have narrower temperature optima for germination than larger seeded species (Bell et al. 1995) and often have more precise cues for germination (Keeley 1991) possibly because they are especially susceptible to drought (Daws et al. 2007; Jankowska-Blaszczuk & Daws 2007); high mean temperature or wide temperature alternations may coincide with risky periods for germination, when there is limited water availability. Of these three species, S. drummondii, displayed an extreme sensitivity to high temperature. Seeds of this obligate seeding species only germinated over a narrow range of cooler temperatures; increasing temperature above the optimum constant temperature had a pronounced negative effect on germination, with an absence of germination evident across a wide range of warmer temperature conditions. Current winter temperatures on the Stirling Range peaks with mean temperatures of 13 (day) and 6°C (night) correspond to the observed germination temperature optima. This already rare species may be at greatest risk of climate warming compared with the other species studied.

Similar to S. drummondii, the slope of the relationship between temperature above the optimum constant temperature and germination was high for another obligate seeder, A. echinocephala and this species also showed a preference for cooler temperatures for germination. However, there were some ecotypic differences in the temperature profiles between seeds collected from montane and lowland populations, with seeds from the lowland population, while more dormant, tending to germinate at higher temperatures (Fig. 2). Dormancy regulating chemicals such as the GA3 used to stimulate germination in both S. drummondii and A. echinocephala may have induced germination over a wider temperature range than would have occurred in its absence, but temperature ranges for germination in these species were nonetheless narrow for these seedlots.

We found no indication that smaller seeds of lowland populations of A. decussata, A. echinocephala and B. brownii had particularly different temperature requirements for germination than the larger seeds of their montane counterparts. Seeds from the lowland populations of these species did show evidence of slower germination, with ≤43% maximum germination at constant temperatures after 28, 35 and 37 days, respectively. However, an alternating temperature regime increased germination in A. echinocephala and the results for seeds from the montane population suggest that higher maximum germination may have been observed with lower night temperatures than achieved during this experiment. Ecotypic differences in temperature optima have been observed for other species. For example, Thurling (1966) reported that seeds of Cardamine showed considerable difference in germination response across an altitudinal gradient with high temperature (>25°C) inhibition of germination in lowland populations and low temperature inhibition (<10°C) in montane populations. Such species are clearly at less immediate risk, although specific populations (ecotypes) may decline. Species that have flexibility in their regeneration strategy tend to have higher germination temperature optima than obligate seeding species (Bell et al. 1995). For the species included in our study, the highest temperature optima for germination were observed for seedlots from two of the species which are able to regenerate by resprouting as well as by seed, A. decussata (21°C) and E. megacarpa (21–23°C).

The width of the germination niche for seeds of D. drummondii, G. leakeanum, K. montana and V. foliosa was greater than for either S. drummondii or A. echinocephala. The temperature tolerances of these species may be wider than the climatic envelope they currently occupy, or recruitment events may occur well into the warmer months when extended drizzle and precipitation from cloud cover may provide seed with sufficient moisture for successful germination and seedling growth. While these species may have the phenotypic plasticity to survive some climate warming, the fact that they already have a conservation listing and/or are restricted to higher altitudes within the Stirling Range, suggests that despite the broad tolerance with respect to germination temperature, their niche breadth is determined by other mature plant traits like dispersal. Furthermore, these, and other species that are already restricted to higher elevations, may face greater competition from species that migrate from lower altitudes as a consequence of climate warming (Gworek et al. 2006; Parolo & Rossi 2008).

Our data suggest that germination may be a more sensitive life stage to climate warming than seedling root growth, with temperatures for seedling growth at least as high as for germination in most species. In addition, as soil temperature is a primary driver for differential selection of plant growth rates with cold soils determining the growth limits of plants by inhibiting root initiation (Kaspar & Bland 1992), warming soils may alter the competitive balance between seedlings of different species. While there is some certainty regarding increasing temperature and our ability to predict direct physiological changes on individual species, changes in other variables and their interactions that affect seed germination and establishment such as rainfall, seasonality, fire and soil moisture remain uncertain (Williams et al. 2007; Berg et al. 2010) and further studies that integrate these factors are required.

Geographic range and germination niche

Our results did not support the hypothesis that germination niche breadth corresponded to breadth of geographic range: species with restricted ranges were neither more, nor less, sensitive to increased temperatures associated with climate change than widespread ones. A lack of consistency in geographic range size and germination niche was also reported by Thompson et al. (1999) and Thompson and Ceriani (2003) for the British flora, and suggests that factors (biotic and/or abiotic) other than thermal tolerance may confine species like V. foliosa and G. leakeanum to their montane habitat. Their wide germination niche may indicate significant phenotypic plasticity which may impact on their future persistence in a changing environment.

Climate manipulation experiments have demonstrated that some plant populations may be able to adapt to higher temperatures (Jump et al. 2008), although whether this is sufficient to ensure population survival is unknown. Adaptive capacity is likely to be low in unique and isolated locations such as in the Stirling Range (Pittock 2003). While gene flow will aid adaptation to warming climates in relatively contiguous populations, for isolated populations this would be more limited and may lead to reduced levels of climate-related genetic variation rendering them less able to adapt to any further changes (Jump & Peñuelas 2005). Rare or restricted species may already suffer from depressed levels of genetic diversity that can reduce a populations' ability to persist in a changing environment (Jump & Peñuelas 2005). Consequently, species lacking phenotypic (e.g. S. drummondii) or genotypic (e.g. B. brownii: Sampson et al. 1994) plasticity in seed traits may experience stress or even extinction during extended climate change. Delayed reproductive maturity will reduce the number of generations that can establish from seed in any given time period therefore long-lived woody perennials with long primary juvenile periods such as B. brownii and species with low levels of fecundity are not expected to undergo rapid evolutionary adaptation (Jump & Peñuelas 2005). The importance of phenotypic plasticity and/or genetic variation will undoubtedly hold important insights into germination niche theory.

Plant life-cycle transition stages that link one generation to the next, such as the germination-emergence transition stage, are likely to be highly vulnerable to environmental change. Where basic knowledge of germination limiting factors is lacking, tipping points or thresholds require identification in order to predict risk of species decline. In a broad sense, our study has helped to define the bounds of the germination niche, particularly for geographically restricted species for which there is already some conservation concern. We have used a novel approach to identify species that possess conservative germination niches and hence, are potentially susceptible to environmental change: with a slight upward shift in temperature, seeds of these species simply will not germinate.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Funding for a research visit to the Millennium Seed Bank, Royal Botanic Gardens, Kew, to conduct this study was provided through the international seed conservation partnership between Kew and the Western Australian Department of Environment and Conservation. Financial support for the Kew seed bank was provided by the Millennium Commission, The Wellcome Trust and Orange plc. The Royal Botanic Gardens, Kew is supported by grant-in-aid from Defra. We are grateful to a number of people for assistance with seed collection, including Todd Erickson, Tony Friend and Sarah Barrett. We acknowledge the comments of the editor and two reviewers whose suggestions greatly improved this manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  9. Supporting Information

Appendix S1. Collection data for 10 south-west Western Australian species used in the study.

Appendix S2. Germination progress curves at constant temperatures for seeds of different seedlots.

Appendix S3. The quadratic relationship between radicle growth rate and temperature at constant temperatures for each seedlot.

FilenameFormatSizeDescription
AEC_2211_sm_SupplInfor1.doc54KSupporting info item
AEC_2211_sm_SupplInfor2.tif7651KSupporting info item
AEC_2211_sm_SupplInfor3.jpg2997KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.