Same, same, but different: dissimilarities in the hydrothermal germination performance of range-restricted endemics emerge despite microclimatic similarities

Tetratheca spp. are herbaceous shrubs notable for their allopatric speciation over short distances in Western Australia. We compared the germination responses of four species of Tetratheca under combined water and temperature stress with the environmental conditions found where these plants grow. Despite similar environments, their germination niches are subtly different.


Introduction
Rare and threatened species are often considered at most risk of decline, especially when their distribution is rangerestricted or geographically isolated, and where significant dispersal barriers exist due to their specific environmental requirements such as unusual substrates, stochastic moisture regimes or extreme conditions between refugia (Byrne et al., 2019).However, short-range endemic species, those that occur only in very specific habitats (Lavergne et al., 2004), may not be necessarily associated with numerical rarity, nor intrinsic extinction risk (Murray et al., 2002).Nevertheless, conserving short-range endemic species is challenging where their only known habitat is vulnerable to anthropogenic alteration (Lande et al., 1999), especially where their capacity to disperse in the face of changing conditions is limited.In situations where whole biogeographic regions and taxonomic groups are defined by species with ranges ∼10-100 km 2 (Markey et al., 2012), specialization to cryptic environmental conditions is often assumed to drive speciation and persistence of diverse short-range endemic taxa (Lavergne et al., 2004).Under such conditions, investigating the link between biogeography and the environmental conditions that are critical to species' persistence can be particularly revealing of population dynamics, species distributions and management initiatives underpinning conservation and restoration (Pironon et al., 2018).Understanding species' physiological thresholds is required to inform management actions, particularly in the face of current global change impacts driven by increasing average temperatures and evaporation, and decreasing rainfall (Tomlinson et al., 2021;Speißer et al., 2022).Moreover, it is important to quantify the impact of environmental change on species' performance at key stages of the life cycle as population persistence can be constrained by ontogenetic-specific requirements (Warren et al., 2011;Merow et al., 2014;Pascual et al., 2022).
Seed germination is a crucial stage of the life cycle of plants during which new seedlings are highly susceptible to environmental stressors (Larson and Funk, 2016;Gremer, 2023).Recruitment strategies, such as 'bet-hedging', where average fitness is sacrificed to minimize variance in fitness, and 'risk avoidance', where fitness is maximized by preventing germination in unfavourable conditions, can help mediate against seedling losses by optimizing germination to milder, and more favourable periods for seedling establishment (Baskin and Baskin, 2014;Duncan et al., 2019;Gremer, 2023).Ecophysiological constraints on germination therefore represent one of the main contributors to plants' geographical distributions (Fernández-Pascual et al., 2017;del Vecchio et al., 2020), with success or failure at this critical point profoundly impacting all following life stage transitions (Donohue et al., 2010).
Temperature and water stress are two major environmental factors that regulate the success of seed germination, particularly in water-limited ecosystems (Duncan et al., 2019).Both temperature and water stress interact to define a bivariate space within which seeds can germinate upon the alleviation of seed dormancy (Bradford, 2002;Onofri et al., 2018), but quantitative studies accounting for hydrothermal interactions are often understudied for short-range endemics (Rajapakshe et al., 2020;Tomlinson et al., 2021).The hydrothermal time concept establishes thresholds for temperature, as well as moisture availability, that regulate the germination of a seed population over time (Bradford, 2002;Onofri et al., 2018).Within specific temperature and moisture thresholds, seeds will accumulate hydrothermal time and progress towards germination (Bradford, 2002).The germination limits for temperature are defined between the minimum (T b , base) and maximum (T c , ceiling) temperatures, and the base water potential threshold (Ψ b ) for moisture availability (Bradford, 2002).These thresholds are defined by both the dormancy state of the seeds and the specific physiological tolerance to both temperature and water stress for any given species (Alvarado and Bradford, 2002;Baskin and Baskin, 2014;Lewandrowski et al., 2017).Quantifying the hydrothermal germination niche, the hygric and thermal requirements necessary for seed germination, can provide important insights into the roles of these environmental variables in structuring populations and recruitment events, as well as shaping biogeographical patterns of rarity and endemism; critical information when planning future conservation actions such as translocation or assisted migration (Rajapakshe et al., 2020).
Banded ironstone formations (BIFs) are widely distributed in the semi-arid and arid Yilgarn region of Western Australia, characterized by shallow, skeletal soils, emergent rock surfaces and often substantial topographical variability at small spatial scales (Gibson et al., 2012).The seeds of rangerestricted taxa living on shallow-soil habitats such as outcrop surfaces can have narrower hydrothermal germination norms (i.e.germinate under a narrower range of hygric and/or thermal conditions) compared to widely occurring sympatric species (Donohue et al., 2010).However, there is a scarcity of data on how closely related short-range endemic species vary in their adaptation to hydrothermal stress, especially where non-overlapping distributions can be quite close together (<50 km) but are separated by different landforms and vegetation communities (Gibson et al., 2010;Yates et al., 2011;Elliott et al., 2019), essentially forming isolated islands of endemism (Byrne et al., 2019).
Several studies have highlighted the importance of assessing the role of different environmental variables on the germination response of range-restricted taxa and the implications for their distributional extent (Vincent et al., 2015;Cochrane, 2020).We quantified the hydrothermal germination niches of a guild of closely related Tetratheca Sm. species, each highly restricted (area of occupancy <50km 2 ) to geographically separate BIF ranges.Though the species occur on isolated, yet adjacent ranges (Fig. 1), we hypothesized that due to the highly localized range of each species, each species may display subtle differences in their hydrothermal germination envelope that reflects subtle site-specific differences in their  microclimate envelope.We also expected the environmental factors most strongly associated with each species' biogeography to reflect hydrothermal germination niches that are more aligned with cooler and reliably wetter germination windows that reflect the unique capacity for BIFs to retain water for longer periods compared to the surrounding landscape (Byrne et al., 2019).

Study species, seed collection and processing
The distributions of Tetratheca Sm. (Elaeocarpaceae) species in the semi-arid and arid regions of Western Australia (WA) represents an intriguing display of short-range endemism and diversification, where a number of unique species are confined to rocky habitats for reasons that are not obvious (Di Virgilio et al., 2018;Byrne et al., 2019).Tetratheca species in the region very rarely inhabit more than one outcrop (Yates et al., 2011), and neighbouring outcrops often support different species that have been separated for substantial periods of time (Butcher et al., 2007).
We compared four species of Tetratheca that are geographically restricted to BIF habitats in the Yilgarn region of WA in this study (Fig. 1, Supplementary Table S1).The distribution of each species is restricted to one of four BIF ranges (Fig. 1) that are within ∼100 km of each other, though separated by different edaphic conditions and floristic communities (Butcher et al., 2007;Gibson et al., 2010).Tetratheca aphylla F.Muell.subsp.aphylla inhabits the relatively deep skeletal soils of lower slopes, hill crests and cliffs of the Helena and Aurora Range (Yates et al., 2011).Tetratheca erubescens J.P.Bull is found in rock fissures and crevices associated with cliffs, hill crests and steep slopes of the Koolyanobbing Range (Krauss and Anthony, 2019).Tetratheca harperi F.Muell.inhabits hill crests, cliffs and cliff slopes of the Mt.Jackson Range (Yates et al., 2011).Tetratheca paynterae Alford subsp.paynterae is restricted to fissures on steep cliffs and tors of the Windarling Range (Ladd et al., 2019).All four taxa are of high conservation concern and have been gazetted as threatened flora in Western Australia (Herbarium, 1998).
Seeds were harvested from wild populations of all four species in 2019 (license numbers SW019800 and DRF7789) and once collected were cleaned, processed (Supplementary Table S2) and stored at 15 • C and 15% relative humidity (controlled environment room) at the Kings Park Science laboratories, Western Australia for 15 months.These conditions both enhance seed longevity and suppress physiological dormancy loss, maintaining the physiological state of the seeds, including their dormancy status, for several years following collection (Turner et al., 2009;Turner et al., 2013).Prior to experimentation, seeds were assessed by X-Ray (MX-20 digital X-Ray cabinet, Faxitron, USA), to identify filled seeds.with these anatomical features identified by the presence of uniform white shading of internal tissues.Seeds deviating from this visual appearance were determined to be non-viable and discarded (Erickson and Merritt, 2016).

Microclimatic conditions of seed sources
The hydrothermal correlates of the realized niche of each study species were quantified using species occurrence data sourced from records maintained by the Herbarium of Western Australia (Herbarium, 1998), supplemented by extensive unpublished surveys by stakeholders from the mineral extraction industry.These known locations were intersected with spatial data describing elevation; aspect and slope (Gallant et al., 2011;Gallant and Austin, 2012a;Gallant and Austin, 2012b); edaphic data describing clay, sand and silt percentage composition Viscarra Rossel et al. (2018a, 2018b, 2018c) and soil depth interpolated from national soil data provided by the Australian Collaborative Land Evaluation Program (ACLEP), endorsed through the National Committee on Soil and Terrain (NCST; www.clw.csiro.au/aclep).We incorporated these spatial data into soil types using a soil textural triangle (Gerakis et al., 1999) to inform the 'micro_global' algorithm in the 'NicheMapR' package (Kearney and Porter, 2017) to estimate microclimatic details that are most relevant to seed germination (Tomlinson et al., 2020).These specific param-eters were chosen to capture the most extreme hydrothermal dynamics experienced by each species, where high water availability at the warmest temperatures may offset thermal stress, and warmer temperatures in the wettest months may promote more rapid germination, given that winter is relatively cold in this region (Supplementary Fig. S1).Higher average soil water availability was thought to represent a less challenging germination environment generally.These data were then averaged between sunrise and sunset at each location to generate daily temperatures and water stress levels 2.5 cm below the soil surface.We used these microclimate estimates to model a 3D environmental niche model built around the annual average soil water potential (SWP), the SWP in the warmest quarter (hottest SWP) and the soil temperature in the wettest quarter (wettest temperature) for each species separately using a Monte Carlo approach, using the 'hypervolume' function of the 'hypervolume' R package (Blonder et al., 2018).The construction function for the hypervolume was a Gaussian kernel density estimation (KDE), where the bandwidth was estimated using a Silverman estimator.The resulting 3D hypervolumes (Fig. 2) were used to define a range of experimental hydrothermal regimes.Hypervolumes of each of the four species were also compared in n-dimensional space to assess overlap using the 'hypervolume_overlap_statistics' function of the 'hypervolume' package (Blonder et al., 2018).The findings of this comparison were used to guide our expectations of likely species hydrothermal performance, in that species with high microclimatic overlap should have highly congruent population performance means (Carscadden et al., 2020).

Germination requirements and hydrothermal thresholds
Prior to incubation, seeds were surface sterilized with a solution of 2% (w/v) calcium hypochlorite (Ca[OCl] 2 ) infused with detergent (Tween 80) for 30 minutes under vacuum (−70 kPa), followed by rinsing with sterile deionized water three times.Seeds were placed inside 90-mm plastic Petri dishes that were lined with moist (9 ml of water per Petri dish) 84-mm germination papers (Advantec, Dublin, CA, USA).The Petri dishes were sealed with plastic wrap to prevent desiccation and covered with aluminium foil to minimize the impact of light, which has a potential confounding effect on germination (Bell, 1994;Ruiz-Talonia et al., 2018).Before commencing hydrothermal germination experiments, seeds were exposed to a warm stratification treatment that, based on previous research, was assumed to be appropriate for dormancy alleviation in all test species (Elliott et al., 2019).These seeds were stratified at 30 • C for 28 days, then removed from Petri dishes and dried at 15 • C and 15% RH (cool dry storage) for 1 week prior to starting the hydrothermal experiments.A random subsample of seeds was also immediately moved from 30 • C to 15 • C while maintaining their hydration status to confirm the effectiveness of this stratification approach.Seeds were maintained for 4 weeks at 15 • C to assess germination after the application of the warm stratification treatment (Supplementary Table S2).
The germination response of seeds to hydrothermal stress was assessed with three replicates of 15 seeds per species, which were incubated at one of 30 hydrothermal stress regimes (water stress levels of 0.0, −0.2, −0.4,−0.6, −0.8 and − 1.2 MPa at temperatures of 10, 15, 20, 22 and 25 • C).These hydrothermal regimes encompass a broad range of the hydrothermal conditions constructed for the habitats of the test species (see above; Fig. 2).Seeds were placed on 84-mm germination paper (Advantec, Dublin, CA, USA) moistened with either distilled water or polyethylene glycol 8000 (PEG) solution (of different concentrations) to generate the range of water potentials outlined above for each incubation temperature (Michel, 1983).These solutions had 2% (v/v) PPM (Plant Preservation Mixture, Austratec Pty Ltd, Bayswater, Victoria Australia) added to minimize microbial contamination during stratification and incubation treatments.Nine millilitres of this solution was added per 90-mm Petri dish for all species, except Petri dishes containing T. aphylla subsp.aphylla.These were instead irrigated with the addition of 1 μM karrikinolide (KAR 1 -Flematti et al. (2004); Chiwocha et al. (2009)) for the duration of the incubation period as preliminary data found an improved germination response for this species following the application of KAR 1 (Supplementary Table S2).All Petri dishes were sealed with plastic wrap and aluminium foil.
During the incubation period, iButton data loggers (Maxim Integrated™, San Jose, USA) were placed in the middle of each set of Petri dishes, which were subjected to the same treatment regime to record the temperature that seeds were exposed to within each incubator while the germination trial was underway (Supplementary Table S3).Germination was identified when the radicle emerged >2 mm from the testa and was scored 3 days per week for 56 days.

Germination modelling
The germination responses of each species, except T. aphylla subsp.aphylla, which failed to germinate beyond 27% in any given treatment, were characterized by a hydrothermal timeto-event (HTTE) model (Onofri et al. (2018)).HTTE models are well-grounded in physiological theory and work on the assumption that germination does not occur outside certain temperature or water stress thresholds and parameterize the hydrothermal germination niche by a function of time (Onofri et al. (2018)).Preliminary assessments of a global model compared five defined HTTE models with selection based on graphical inspection of goodness of fit and Akaike's Information Criterion (AIC; Akaike (1974)).The model selected was a six-parameter log-logistic cumulative distribution function (HTTEM; Mesgaran et al., 2017, Onofri et al., 2018, Onofri et al., 2022), defined as: With three distinct sub-models: Where germination proportion is dependent upon time (t), temperature (T) and base water potential (Ψ ); the model is computed by the reciprocal of time to the 50th percentile of the germinated fraction ( 1 t g ) GR50 HTTE , maximum germination proportion (G; while accounting for possibility that Pmax HTTE may not reach 1 under any condition); median base water potential limiting to germination in the seed lot  After global model selection, we constructed a speciesspecific model, and determined if the delineation identified responses that were significantly different to the 'global model' via an F-test using the anova function in the 'stats' package (R Development Core Team, 2022).Post hoc comparisons of parameter estimates were conducted with the compParm function in the 'drc' package (Ritz and Streibig, 2012) with significance established at P < 0.05.However, we do not report on comparisons for b, as it is a shape parameter that is regarded as independent from the environmental covariates, or σ Ψ b , due to insignificant fit and parameterization in the global model.After the hydrothermal time distributions were modelled, using equation 1.3, hydrothermal germination responses were plotted at 15, 30 and 60 days to quantify the changes in germination behaviour for each species.Maximum germination proportions (G max ) and thresholds for T b and T c , as well as Ψ b were derived from the predictions at each increment, defining the breadth of the hydrothermal germination niche as it changes over time.

Niche overlap in microclimatic conditions
The highest degree of microclimatic niche overlap (70%) was between T. paynterae subsp.paynterae and T. aphylla subsp.aphylla and between T. paynterae subsp.paynterae and T. harperi, with all four species sharing at least 50% of their 3D microclimatic niche space with each other (Table 1; Fig. 2).The percentage of unique space a species had was complex, as different species nested within each other's realized niches in different ways.The main patterns were that T. erubescens consistently had the lowest percentage of unique space (11-23%) compared to the other three (i.e. had a realized niche that mostly nested within the other three).This was followed by T. aphylla subsp.aphylla (10-18%) compared to T. paynterae subsp.paynterae and T. harperi, except when compared to T. erubescens.Compared to T. erubescens, T. aphylla subsp.aphylla had a higher percentage (45%) of unique space that mainly related to having a broader envelope of soil temperature in the wettest quarter (wettest temperature), where the known occurrence locations are found.Leaving T. paynterae subsp.paynterae and T. harperi to both retain ∼30% of the combined niche space uniquely to themselves (Table 1), that related to also having a broader microclimatic envelope across all three hydrothermal conditions, albeit both dominating the opposite ends of this envelope.Tetratheca harperi uniquely occupied microclimates characterized by higher temperatures than any other species, while T. paynterae subsp.paynterae occupied microclimates characterized by the coolest and wettest conditions (Fig. 2).

Hydrothermal time-to-event germination modelling
The global HTTE model resolved significant fits for each parameter (P < 0.001), excluding σ

Hydrothermal threshold variation over a 60-day window
Maximum germination (G max ) proportion after a 15-day window was 0.12 in T. erubescens, in contrast to 0.04 and 0.05 in T. harperi, and T. paynterae subsp.paynterae, respectively.These proportions increased progressively in all species, although variation was characterized by overall lower germination performance in T. paynterae subsp.paynterae (G max < 0.21, after 60 days), in contrast to T. erubescens and T. harperi (G max > 0.5 after 60 days).The lower germination performance of T. paynterae subsp.paynterae was also matched by the narrowest hydrothermal ranges for germination in T. paynterae subsp.paynterae (see T b , T c and Ψ b , Fig. 3).While all species were constrained to germinate between 10 and 25 • C, T. harperi had the broadest range over time (Fig. 3).There was progressive broadening of Ψ b thresholds in all species, with T. erubescens characterized by the lowest Ψ b over the time-course (refer to Fig. 3 for all T b , T c and Ψ b parameter estimates).

Discussion
In natural environments, successful seedling establishment is reliant on the duration of the window of opportunity for germination (Gremer, 2023).Hydrothermal stress is a major selective pressure that restricts that window of opportunity, particularly in semi-arid ecosystems where water availability can be extremely variable from year to year due to the dynamic nature of these extreme environments (Huang et al., 2016).To our knowledge, our results are the first to characterize the hydrothermal germination niche for Tetratheca species, using a contemporary HTTE modelling approach (Onofri et al., 2018).The cool thermal thresholds quantified from our hydrothermal time modelling are consistent with reports in the region (Dalziell et al., 2022) that all four species tend to germinate during the cooler winter months, and the HTTE models explained slowed germination, along with a gradual widening of hydrothermal thresholds over a 60-day germination window.Despite substantial overlap in the realized microclimatic niche estimated from known occurrence locations, subtle interspecific differences were evident in the hydrothermal niche consistent with the specific biogeography of each species.

Microclimatic niche and germination strategy
The high degree of overlap between the microclimatic niches and the hydrothermal optima of the Tetratheca species (Fig. 2, Fig. 3) suggests that similar selection pressures have optimized their hydrothermal niches, despite their geographic and genetic isolation (Byrne et al., 2019).The locations where these species occur are less stressful at critical times by comparison to the surrounding landscape, with SWPs higher than −1.2 MPa and soil temperatures between 5 and 25  rainfall period for the region (Supplementary Figure S1).As demonstrated by the HTTE models, the three modelled Tetratheca species require long periods of hydrothermal time accumulation to germinate, with germination constrained to comparatively cooler and wetter conditions (Table 2; Fig. 3).
Despite the narrow temperature range (10-25 • C), all species included in the analysis demonstrated remarkably wide base water potentials for germination, suggesting that a proportion of the seed population may be capable of germinating into significant water stress.Collectively, these observations suggest that different outcrop-endemic Tetratheca species may be differently adapted to cope with limited water availability.This interpretation is at odds with that which may be drawn when considering the high degrees of microclimatic niche overlap between the species (see above).However, despite the similarities present in the microclimatic niche, each species inhabits some unique microclimatic niche space, which correspond with their dissimilarities in hydrothermal niches.For example, T. erubescens inhabits the core of the microclimate envelope shared by all species, defined by moderately broad thermal conditions and SWPs (Fig. 2), and germinates relatively well across a wide combination of water stress and thermal stress (Fig. 3).It is also the youngest species in a phylogenetic sense (Byrne et al., 2019), and thus may not have specialized to microclimatic conditions unique to Koolyanobbing Range.Where T. harperi occurs uniquely, this microclimatic space is principally defined by warmer temperatures (Fig. 2), and the species has the broadest tolerance to thermal stress of any of the species studied (Fig. 3).Finally, T. paynterae subsp.paynterae occurs more in cooler and wetter microclimate space than any of the other species (Fig. 2) and has the lowest germination success of the three species that we characterized here, and only at cooler temperatures and lower water stress compared to either T. harperi or T. erubescens.The more limited germination success could be attributed to seeds of T. paynterae subsp.paynterae that remained in a conditional dormancy state at the time of the germination study, which explains the limited and narrow optimal conditions for germination, in contrast to the other species (Baskin and Baskin, 2014).
By capturing the time element of the germination window, we have shown differences in how divergent species respond to ephemeral conditions.The modelled Tetratheca species require hydrothermal time accumulation between 186 and 500 MPa • C days to complete germination of the seed population, which could be attributed to dormant proportions in the seed cohort (Baskin and Baskin, 2014), or being intrinsically slow-germinating species.Slow germination has been reported in many taxa endemic to mesic habitats of Western Australia (Vincent et al., 2015;Cochrane, 2020;Dalziell et al., 2022).However, slow germination is potentially risky as it could compromise seedling survival in microhabitats if they fail to retain moisture for a sufficient period following rainfall episodes.This supports contentions that the study species may have contracted to microhabitats on outcrops that can retain high soil moisture for sufficient time to support germination beyond what would normally be expected in surrounding environments (Gibson et al., 2010;Yates et al., 2011).The extended hydrothermal germination requirement may be a common characteristic of floristic endemism in the inselberg communities of the Yilgarn region (Gibson et al., 2012), and likely influences the biogeography of other BIF-restricted species such as Ricinocarpos brevis, Darwinia masonii and Lepidosperma gibsonii (Byrne et al., 2008).While the hydrothermal requirements for seed germination are largely unexplored for many BIF-restricted species, understanding these dynamics could provide critical insight into the constraints and requirements driving recruitment events in these floristically diverse ecosystems.

Seed dormancy and 'bet-hedging'
The specialization of Tetratheca species to slow germination in relatively cool and wet conditions is also likely enhanced by the presence of physiological dormancy, as seeds are cued for germination via warm stratification as soil temperatures gradually drop, leading into the cooler relatively wetter winter season (Turner et al., 2006;Just et al., 2023).Previous studies on several BIF-endemic species suggest that in physiologically dormant, range-restricted species, the windows of opportunities for seedling establishment are rare, unpredictable and perhaps only occur every few years (Yates et al., 2011;Elliott et al., 2019).Although not an essential trait in the formation of a soil seedbank (Gioria et al., 2020), dormant seeds are likely to persist in the soil seedbank while conditions are not conducive to germination and seeding recruitment, and seed dormancy can leave ungerminated or dormant seeds exposed to long periods between recruitment events (Long et al., 2015;Miller et al., 2019).Indeed, warm stratification is a relatively uncommon dormancy release treatment (Just et al., 2023), allowing seeds to retain a slow germination pattern in cool moist conditions as gradual dormancy release occurs in response to an initial warmer, moist soil environment.This gradual drop in soil temperature acts as an initial environmental filter to cue seeds for seasonal germination, rather than opportunistic, aseasonal germination (Vincent et al., 2015).As such, any given hydrothermal window could stimulate the germination of only a proportion of the available soil seed bank, minimizing the risk of unsuccessful germination in dynamic and unpredictable environments.While we cannot conclusively demonstrate this for the Tetratheca species studied here, other short-range endemic species in the region, such as Darwinia masonii and Lepidosperma gibsonii possess physiological seed dormancy and form long-lived soil seed banks (Elliott et al., 2019;Miller et al., 2019).The mechanisms underpinning the alleviation of seed dormancy under natural conditions are often difficult to unravel (Merritt et al., 2007), and additional research is necessary to understand the nuances of dormancy loss and germination stimulation in these species more broadly.

Limitations to model interpretations
While the applied dormancy alleviation treatments resulted in varied degrees of germination success across the studied species, the low germination rates seen for T. aphylla subsp.aphylla may be a result of physiological dormancy that was not adequately overcome.This suggests that an extended stratification period may be required, in addition to the stimulus applied (KAR 1 ).Moreover, the germination responses here may still encapsulate some component of seed dormancy, which would account for the generally much lower germination success that we also found for T. paynterae subsp.paynterae across all hydrothermal conditions (Fig. 3).In general, exposure to KAR 1 enhanced the germination success of these species, potentially indicating that exposure to smoke represents a critical environmental stimulus that defines the germination requirements of those taxa, particularly under conditions where disturbance events (e.g.fire) may occur (Turner et al., 2018).As such, dormancy alleviation and the hydrothermal niche require further attention to provide a comprehensive understanding of the drivers of in situ recruitment beyond the fundamental microclimatic elements that we have reported here.
The capacity of range-restricted taxa to persist in the face of changing conditions is dependent on the influence of interacting niche dimensions on population demographics (Pironon et al., 2018).Our study captured the performance norms of seeds collected under current climatic conditions but offers little direct insight into phenotypic flexibility in these norms.The extent to which climate change can affect the physiological limits and optima of range-restricted taxa in terms of altered rainfall patterns and temperature, as well as dormancy cycling, must also be quantified (Walck et al., 2011;Cochrane et al., 2014;Cochrane, 2017).However, a greater suite of species should be compared including both range-restricted taxa as well as those with wide distributions to gain useful insights into how hydrothermal germination niche affects patterns of species distribution.Collectively, this can help determine whether range-restricted species display attributes not typically observed with more broadly occurring sympatric species (Byrne, 2019).

Conservation implications
This study highlights the significant role of hydrothermal stress on germination responses that are likely to influence population dynamics and distribution patterns of BIF specialist Tetratheca species.The high degree of overlap between the microclimatic niches, coupled with existing phylogenetic evidence, supports suggestions that these are allopatric species (i.e.localized speciation on extended time frames; Byrne et al. (2019)) that have persisted on localized outcrops due to the relatively milder conditions they support compared to the surrounding semi-arid landscapes (Byrne et al., 2018).The short-range endemic Tetratheca species are associated with such habitats (Butcher et al., 2007;Byrne et al., 2018) is more aligned with cooler and reliably wetter regions in southwest Western Australia.This conserved germination strategy suggests that Tetratheca species endemic to BIFs may be more susceptible to environmental changes as the climate further warms and dries.Furthermore, other Tetratheca species endemic to granitic, sandstone and ironstone inselbergs throughout WA have unknown germination requirements, leaving uncertainties as to how they may respond to the unpredictable changes of climate likely to eventuate throughout the arid ecosystems of Australia (Lioubimtseva, 2004).

Figure 1 :
Figure 1: Distributional extents of the Tetratheca spp.studied here all fall within an area ∼20 × 20 km, located 370 km north-east of Perth, Western Australia (a).The study region is characterized by relatively uniform soil temperatures (b, c) and soil water potentials (d, e).Even in the coolest and wettest periods (b, d) soils in the region are relatively warm and very dry.The points on the maps (b-e) indicate the known populations of our four study species T. paynterae subsp.paynterae, T. harperi, T. aphylla subsp.aphylla and T. erubescens (north to south).

Table 1 :
Patterns of niche overlap between pairs of Tetratheca species.The Sorenson overlap indicates the proportion of the combined niche space that the two species share, while the Fraction Unique columns indicate what proportion of that combined niche space is occupied by only the indicated species (Sp. 1 or Sp. 2) (Onofri et al., 2022) defined under the conditions between which germination occurs (i.e. between T b and T c ; maximum temperature for germination) and is equal to 0 for temperatures T ≤ T b or T ≥T c , and water potentials Ψ < Ψ b − k T − T b .Models were fitted using the drcte function of the 'drcte' statistical package(Onofri et al., 2022)with hydrothermal time to event model coded in the 'drcSeedGerm' package(Onofri et al., 2022)using the R statistical environment (version 4.2.3) and RStudio Version 2023.03.0 (R Development Core Team, 2022).