Shallow but striking genetic structure in the highly connected New Zealand sand dollar (Fellaster zelandiae)

Marine biogeography

Phylogeographic structure in marine species can shed light on how physical environmental processes shape biodiversity (Emerson and Hewitt 2005; Beheregaray 2008; Kumar and Kumar 2018), which is of particular importance as species distributions shift in response to global warming (Banks et al. 2010; Lo Brutto et al. 2011; Halbritter et al. 2015). While traditional genetic approaches, such as sequencing one or a few markers, have helped us understand broad-scale patterns of marine biogeography (e.g. Palumbi 1996), newer, high-resolution genomic tools can provide powerful and detailed insights into cryptic diversity, timelines of species divergence, migration, and dispersal abilities (Rocha et al. 2013; Fraser et al. 2016; Vaux et al. 2023).

Global analyses of marine biogeography have found distinct endemicity of marine biota in 30 different realms (Briggs 1995), with ~40% of species being unique to each realm (Costello et al. 2017). New Zealand has been identified as one such unique biogeographic realm, with the large ocean basins acting to separate many New Zealand species from other regions. Towards the end of the Oligocene (c. 23 Ma), the Zealandia plate was thought to be partially ‘drowned’ (Waters and Craw 2006; Knapp et al. 2007), which is suggested to have reset many New Zealand ecosystems (Knapp et al. 2007; Gordon et al. 2010). Trans-Tasman dispersal and resettlement from Australia has been inferred from genetic data to occur in some species whose larval stages can last 5 months or more (Ovenden et al. 1992; Chiswell et al. 2003; Waters et al. 2007; Thomas et al. 2021); however, trans-Tasman dispersal for shorter larval stage organisms are less likely to succeed (Ross et al. 2009). The resulting unique biota and considerable isolation of New Zealand makes it an intriguing location for biogeographic research.

New Zealand marine hydrological processes and geographic patterns are particularly influenced by the Cook Strait (Chiswell et al. 2021), separating the North and South Islands, with fast tidal flows and deep bathymetry causing biogeographic disjunctions on the upper South Island in several shallow water marine taxa (Goldstien et al. 2006). This oceanographic separation is relatively new on geological timescales, with the estimated final formation of the Cook Strait being c. 500,000 years ago (Trewick and Bland 2012). A land bridge connected the two islands both prior to the Strait formation (Alloway et al. 2007) and during Pleistocene sea regressions, including during the last glacial period c. 20,000 years ago. This transient isolation caused differentiation of population structures in Hector’s and Maui dolphins (Alvarez-Costes et al. 2025), and so the Cook Strait area may be the epicentre of both north–south and east–west genetic breaks. This central location may also have the most genetic mixing from recent recolonisation efforts of these different populations.

Further south, the Chatham Rise off the eastern coast of the South Island’s Banks Peninsula forces a convergence of currents and pushes the Southland Current offshore (Nodder et al. 2003; Sutton 2003; Hopkins et al. 2010). Shorter larval development windows of species such as the mollusc Lunella smaragda, and brachiopods Terebratella sanguinea and Liothyrella neozelancia can result in genetic isolation between stretches of coast in New Zealand (Ostrow 2004; Chiswell and Booth 2008; Arranz et al. 2021), whereas species with longer larval stages such as the crustacean Jasus edwardsii, the sea urchin Centrostephanus rodgersii, and molluscs Scutus breviculus and Austrolittorina antipodum show an absence of population subdivision across New Zealand (Ovenden et al. 1992; Waters et al. 2007; Thomas et al. 2021). These differences raise questions about how ubiquitous the influences of abiotic and ecological forces are versus hydrological and geographic factors in shaping biodiversity in New Zealand.

Fellaster zelandiae population dynamics

Fellaster zelandiae (Gray, 1885) (‘kina papa’ in te Reo Māori) is an irregular echinoid (Clypeasteroida) that acts as a benthic–pelagic coupler and bioturbator, and is endemic to Aotearoa (New Zealand). Fellaster zelandiae is the only extant species in its genus, and has populations distributed around New Zealand. Classical taxonomic study has determined that F. zelandiae populations around New Zealand belong to a single species with no notable morphological differences (Durham 1955). Fellaster zelandiae are broadcast spawners whose planktotrophic larval stages develop for several weeks in the water column until they metamorphose into juveniles. For such species, larval dispersal time is an important factor in the distribution and genetic connectivity of populations between geographically distinct regions (Strathmann 1978). Generally, the longer a larva spends in the pelagic period, the greater is the potential connectivity of populations (Strathmann 1974; Shanks 2009). In many echinoids, the length of the larval period is strongly influenced by water temperature (Bosch et al. 1987; Hoegh-Guldberg and Pearse 1995).

Fellaster zelandiae has a larval development period of 28 days or more (Bunckenburg 2003), which is longer than other sand dollar species that range from 5 to 15 days, and rarely last more than 3 weeks (Caldwell 1972; Emlet 1986a; Chen and Chen 1993). We hypothesise that the prolonged time spent in the water column give F. zelandiae a greater chance to disperse compared to other sand dollar species, and could enhance gene flow among populations. Although other genetic studies of regular and irregular echinoderms have shown cryptic speciation (Coppard et al. 2013; Bronstein et al. 2017), or considerable phylogeographic structure (Lessios et al. 1999; Coppard and Lessios 2017), no previous genetic work has been undertaken on F. zelandiae. Determining the extent of genetic connectivity in F. zelandiae will help improve our understanding of the processes driving biogeographic patterns in marine species with medium-to-long larval periods. As Fellaster zelandiae occurs throughout coastal New Zealand from 35 to 46°S, and across diverse environmental gradients (hydrology, primary productivity, temperature and pH), genetic research could also provide insights into any local adaptation and phenotypic plasticity.

Aims

All previous ecological work has assumed the uniformity of Fellaster zelandiae populations without considering spatial patterns in phenotypic plasticity or genetic diversity (Bunckenburg 2003; Grkovic and Copp 2013; Karelitz et al. 2017). Here, we set out to test the following questions: (1) where does Fellaster zelandiae sit phylogenetically among Clypeasteroida; (2) how homogeneous is the genetic structure of Fellaster zelandiae across its geographic range; and (3) do classical population breaks seen in coastal taxa apply to Fellaster zelandiae? We tested these questions by genetic analysis of 15 populations encompassing the biogeographic range of Fellaster zelandiae.

Field sampling

Nine geographic regions were identified by hydrological and geographical features around New Zealand (Fig. 1a) on the basis of previous work to model larval connectivity (Chiswell and Rickard 2011). Fifteen locations of interest with potential populations of Fellaster zelandiae were identified across New Zealand on the basis of recorded observations. Intact living adult sand dollars were collected by free-diving at nine subtidal sites, by hand from four intertidal sandy environments, and by dredge at two benthic sites (Fig. 1b). Sampling in Regions 1, 2, 5, 7 and 9 was successful, with live individuals collected from at least two distinct locations for each region. In Regions 3, 4 and 8, only one site was successfully sampled, with failed sampling attempts across more of each region due to poor sea conditions. At Region 6, we found evidence of populations (broken tests washed up onshore) but were unable to collect fresh samples.

Fig. 1.

(a) New Zealand currents: Tasman Front (TF), North Cape Eddy (NCE), East Auckland Current (EAUC), West Auckland Current (WAUC), East Cape Eddy (ECE), D’Urville Current (DC), East Cape Current (ECC), Wairarapa Eddy (WE), Westland Current (WC), Southland Current (SC), Antarctic Circumpolar Current (ACC) and Subantarctic Front (SAF). New Zealand water masses: subtropical water (pink), subtropical front (purple), coastal water (green), Southland front (yellow) and subantarctic water (blue). (b) Distribution of genetic samples taken of Fellaster zelandiae with nine geographic regions of interest, and locations of interest. Successful sequencing of >10 samples (purple), 6–10 samples (orange) or 1–5 samples (red).

At each site, more than the requisite number of individuals was collected initially (i.e. 70), and 27 healthy-looking individuals were selected from the pool (Nazareno et al. 2017). Suitability was judged on the basis of symmetry, a lack of predation marks and full spine coverage of the tests. Unhealthy individuals and surplus healthy individuals were returned to their beds. Collected samples were preserved immediately in 95% EtOH, and then stored at −80°C within a few days of collection. Tissue samples were taken for genetic analysis by opening the oral surface and removing 3 g of soft tissue from the intestinal and gonadal area.

DNA extraction and QC

Tissue was subsampled and rinsed three times with ultrapure H₂O to remove ethanol. DNA was extracted using a Qiagen DNeasy 96 Blood and Tissue Kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol, yielding ~20 μL of DNA of varying concentrations per sample. DNA was used in both of the following procedures.

COI

Genotyping by sequencing (GBS)

Extracted DNA was cleaned using MEGA quick-spin plus Fragment DNA purification kit (iNtRON Biotechnology, Gyeonggi, South Korea) to remove degraded DNA fragments. RNAse A (Sigma–Aldrich, Burlington, MA, USA) was added into cleaned solutions to remove RNA contamination.

Animal ethics

Fellaster zelandiae were sampled following jurisdictional requirements for biological sample collection and were euthanised following best ethical practices. These organisms were collected under New Zealand MPI Collection Special Permits SP644 and SP823.

COI data

Smart Model Selection in PhyML found that the best substitution model was HKY85+G+I (Hasegawa et al. 1985), with an estimated 0.593 proportion of invariable sites, four substitution-rate categories, and an estimated 0.919 gamma shape parameter. A consensus tree was built from 40 COI sequences: 34 sequences of 34 known species of sand dollar and 6 sequences of Fellaster zelandiae from 6 different populations (Fig. 2). Clades were resolved similarly to previous literature (Lee et al. 2023) on the basis of bootstrap values (BV) and posterior probabilities (PP). Fellaster zelandiae was resolved as a sister genus to the genus Arachnoides (BV = 83, PP = 1.00), and then to three species of the genus Clypeaster (BV = 99, PP = 0.78). Percentage relatedness (Table S3) of base reads showed the same genus relationship, with Fellaster relatedness to Arachnoides being the greatest (85.8–86.6%), followed by the relatedness to Clypeaster (81.5–84.5%). The relatedness of all other genera to F. zelandiae ranged from 79.5 to 82.8%. These results were supported by the Bayesian inference MCMC analysis run using MrBayes, which provided strong posterior probability support for the sister relationships (0.99–1.00). Divergence times of Fellaster to Arachnoides genera were calculated as 3.3–3.6 Ma (Table S4), with divergence between Fellaster and all other species being 3.9–4.4 Ma. The next-closest relatives to Fellaster zelandiae were Clypeaster virescens (16.4% different) and Clypeaster japonica (15.6% difference), helping confirm that the genus Fellaster sits within the family Clypeasteridae.

Fig. 2.

Phylogenetic relationship of 33 species of sand dollars (Echinodermata: Irregularia) from 618-bp COI sequences. Nodal supports are shown as bootstrap values (BV) as a percentage (above) and posterior probabilities (PP) from 0 to 1 (below). BV below 70% and PP below 0.7 are not shown.

Analysis of COI sequencing data found little evidence of speciation or major genetic divergences among the populations of Fellaster zelandiae, because the p-distance of the Tauranga site was 0.045 and p-distances of the five other F. zelandiae populations were 0.012–0.032 (Table S5). The values of percentage relatedness of the six F. zelandiae specimens were all >99%, indicating no evidence of cryptic speciation. All but one sample of F. zelandiae shared near-identical COI haplotypes: Ahipara, New Plymouth and Te Waewae Bay shared identical haplotypes (from Regions 1, 1, and 9 respectively), the Timaru sample (Region 8) had one G–A substitution, and the sample from Hikuwai (Region 2) had a single G–A substitution in a different region. Only the sample from Tauranga was distinct, with five substitutions along the sequence (A–G, C–T, G–A, G–A, T–C) different from all other samples, including the Hikuwai sample in the same region.

GBS

In total, 40,903 SNPs were detected with UNEAK for 166 Fellaster zelandiae from across the entire range of New Zealand, with a mean sample depth of 2.7 (Supplementary Fig. S1). The Standard Hardy–Weinberg filter removed 178 SNPs (Supplementary Fig. S2) for a total of 40,725 SNPs used for analyses. The MAF distribution was skewed to the left, indicating good diversity of the samples (Supplementary Fig. S3). F_ST values were extremely low (greatest value = 0.025), indicating a very high genetic similarity (Supplementary Fig. S4, S5). Observed heterozygosity values were also very low (H_O = 0.150) and less than expected values (H_E = 0.222) indicating potential population substructure (Table S1).

Plots of the first four principal components (Supplementary Fig. S6) show no major population differences, but two subtle distinct clusters were differentiated by the first principal components (Fig. 3). Overall, population variation was low, with the first two principal components accounting for only 2 and 1% of variation. Cross-entropy values fitted to clusters (K) from LEA showed no appreciable decrease in value at any cluster greater than 1, indicating relative genetic homogeneity (Supplementary Fig. S7). Ancestry coefficients plotted as pie charts in geographic regions for K = 2, 3 and 4 indicated subtle breaks between north-eastern (Regions 2 and 3), north-western (Region 1), central (Regions 4 and 5) and southern (Regions 7, 8 and 9) regions (Fig. 4a, 5a, 6a). The distinction of the north-eastern population is the most prominent, occurring in each of the three assignment tests. Notation of the same groupings was made in a PCA plot to visualise the magnitude of the genetic difference (Fig. 4b, 5b, 6b). PCA clusters confirmed the distinction of the north-eastern population, while showing that the north-western and central distinctions were less prominent. Admixture clusters (K = 2, 3, 4) plotted by latitude were determined relevant in a subpopulation context to support the minor differences among geographic regions (Fig. 4c, 5c, 6c), because each additional cluster aligned with prominent geographical or hydrological features around New Zealand.

Fig. 3.

Site location and first two principal components from principal component analysis of 166 genetic sequences for Fellaster zelandiae colour-coded to location of the specimen.

Fig. 4.

Analysis of population genetics under an assumption of two subpopulations. (a) Population clusters of genetic samples plotted at each location, with distinctions between north-eastern (red) and southern (blue) subpopulations seen in the geographic breaks indicated by the dashed line. (b) Comparison of principal component analysis for Fellaster zelandiae, with differences between population groups highlighted in the coloured ovals of each population. (c) Subpopulation clustering (shading) compared to prominent hydrology (arrows), indicative of potential larval dispersal and discrete subpopulation. (d) Admixture plot with the clustering set to K = 2 for each of the 166 GBS sequences, resolving the population clusters of all samples by location.

Fig. 5.

Analysis of population genetics under an assumption of three subpopulations. (a) Population clusters of genetic samples plotted at each location, with distinctions among north-eastern (red), central (yellow) and southern (blue) subpopulations seen in the geographic breaks indicated by the dashed line. (b) Comparison of principal component analysis for Fellaster zelandiae, with differences among population groups highlighted in the coloured ovals of each population. (c) Subpopulation clustering (shading) compared to prominent hydrology (arrows), indicative of potential larval dispersal and discrete subpopulation. (d) Admixture plot with the clustering set to K = 3 for each of the 166 GBS sequences, resolving the population clusters of all samples by location.

Fig. 6.

Analysis of population genetics under an assumption of four subpopulations. (a) Population clusters of genetic samples plotted at each location, with distinctions among north-eastern (red), north-western (orange), central (yellow) and southern (blue) subpopulations seen in the geographic breaks indicated by the dashed line. (b) Comparison of principal component analysis for Fellaster zelandiae, with differences among population groups highlighted in the coloured ovals of each population. (c) Subpopulation clustering (shading) compared to prominent hydrology (arrows), indicative of potential larval dispersal and discrete subpopulation. (d) Admixture plot with the clustering set to K = 4 for each of the 166 GBS sequences, resolving the population clusters of all samples by location.

Phylogeny of Fellaster zelandiae

Our results place the genus Fellaster sister to the Australian genus Arachnoides, with an estimated divergence c. 3.4 Ma (Table S4) that vastly post-dates the formation of the deep ocean basin of the Tasman sea (Hayes and Ringis 1973; Weissel and Hayes 1977). Physical distance separating these two species coupled with both the need for long dispersal time for trans-Tasman transport (Chiswell et al. 2003) and the short larval window of Arachnoides placenta (Aung, 1975; Gonzalez-Bernat et al. 2013) suggests that these two species have become isolated and genetically distinct reasonably recently in geological timescales. An extinct south-eastern Australian species, Fellaster incisa, exists in the fossil record from the Pliocene (5.3–2.6 Ma) (Sadler and Pledge 1985; Holmes 1987). This species had a more southern distribution than the genus Arachnoides (Foster and Philip 1980) and it is possible that Fellaster incisa diverged from Arachnoides by geographical and ecological separation in Australia, and then dispersed to New Zealand by a chance dispersal event (e.g. by storms, zoochory or rafting on buoyant seaweed; Gillespie et al. 2012). The extinct Fellaster incisa may hence be ancestral to F. zelandiae, which was present in some Pliocene fossil records of New Zealand (Farquhar 1894). Recent trans-Tasman dispersal has also been inferred for another echinoderm found in New Zealand and Australia, Centrostephanus rodgersii (Thomas et al. 2021), although that species has much longer-lived echinopluteus transversus larvae thought to be for long-distance transport (Soars et al. 2009). Dispersal of Fellaster larvae remains a somewhat less plausible explanation for a past trans-Tasman dispersal event, given the shorter-lived larvae, and the length of time for passive particles to move across the Tasman Sea. For instance, 85% of satellite drifters released in Australia (Tasmania) arrived in New Zealand after 18 months (Cresswell 2000), indicating that the pace of movement in water currents is likely to be too slow for the development window of Fellaster, unless past currents were more rapid, or the pelagic larval period was longer.

Biogeography and connectivity of Fellaster zelandiae

Differentiation among populations of Fellaster zelandiae across New Zealand was low (Fig. 3), but the subtle structure between northern and southern populations suggests some level of reduced mixing, possibly driven by oceanography. Principal component analysis (Fig. 3) found that of these subpopulations, all specimens of the North Islands’ north-eastern and eastern coasts (Regions 2 and 3) were distinct from those in the rest of the country. Although other geographic differences emerged as the number of clusters increased, it is hard to definitively determine whether one cluster is more accurate than another. Rather, the display of greater cluster numbers can be thought of as a representation of minor allele differences, which may, over time, diverge. Presently, the genetic differences seen across all populations of F. zelandiae are remarkably low (Fig. 3, S4, Table S2), possibly owing to F. zelandiae being a species with a planktonic larval duration and dispersal that drives genetic homogeneity.

Two other echinoids in New Zealand, the sea urchin Evechinus chloroticus (‘kina’ in te Reo Māori) and Centrostephanus rodgersii, have also shown subtle genetic variation. However, only E. chloroticus’ dispersal across the entire country is useful for comparison, because the range of C. rodgersii is confined to the warmer waters (>15°C) in northern New Zealand (Pecorino et al. 2013). Original surveys using gel electrophoresis of five enzyme loci indicated that E. chloroticus had little or no variation among six populations (Mladenov et al. 1997), and the 4–6 week larval period was suggested to allow high gene flow (Barker 2013). However, subsequent analysis of microsatellite loci in the same species found restricted gene flow between the North and South Islands, as well as differences within and among fjords in Fiordland and between ecotypes of open coast and fjord populations (Perrin 2002). These differences were attributed to restricted larval dispersal in sheltered populations and to hydrology, as seen directly in Lamare (1998). The small differences seen in E. chloroticus used to infer subtle population differences may be taken into consideration here for F. zelandiae as these two species follow similar larval development patterns and timing.

When comparing the subtle difference seen in this study, Fellaster zelandiae populations from the north-west (Region 1) may connect to central (Regions 4 and 5), central may connect to the western coast (Region 6), the north-east (Region 2) may connect to the eastern coast (Region 3), and the south (Region 9) may connect to the south-east (Regions 7 and 8). However, unidirectional flow may prevent mixing of north-western (Region 1) and north-eastern (Region 2) groups, and geographic stretches of unsuitable substrate or habitat (gravel or mud and silt) in the fjords may be too large for larval dispersal and gene flow from central New Zealand (Region 5) and southern New Zealand (Region 9). The genetic separation of F. zelandiae on the eastern coast of the North Island (Regions 2 and 3) from other populations aligns with the prevailing East Auckland Current and East Cape Current (Fig. 4–6). The intensity and direction of these currents from north-west to east may reduce gene mixing with the north-west (Region 1). In all cluster analyses, the Chatham rise additionally acts as a barrier between north-eastern populations (Regions 2 and 3) and southern populations (Regions 7–9) as the Southland Current moves offshore at Banks Peninsula. Paired with poor sediment habitat of mud and gravel seen in the Kaikōura region, the north-eastern coast of the South Island may be a large geographic break between populations of F. zelandiae in the south and in the centre of the country. Our study is limited in determining the veracity of this idea because no DNA samples were able to be collected from the western coast of the South Island. Future work should look to gather genetic samples from this region to resolve such questions.

The strong connectivity in Fellaster zelandiae contrasts with the high phylogeographic structure observed in many New Zealand marine taxa, such as the green-lipped mussel, Perna canaliculus (Apte and Gardner 2002), three limpets of Cellana ornata, Cellana radians and Cellana flava (Goldstien 2005), two red algae of Agarophyton chilense and Agarophyton transtasmanicum (Huanel et al. 2020), and three species of whelks, namely, Buccinulum vittatum (Gemmell et al. 2018), Cominella maculosa (Dohner et al. 2018) and Cominella virgata (Walton et al. 2019). Many of these species have short to medium larval dispersal times of days to weeks, and have seen population structure form around the Cook Strait, between the west coast to east coast, from small geographic locations, and from environmental factors such as temperature (Waters and Roy 2004; Ross et al. 2009). Investigating F. zelandiae populations with clusters of K = 2–4 shows that these influences may be faintly present (Fig. 4–6), but the degree of difference is insufficient to quantify populations as distinct. The absence of much population subdivision in F. zelandiae across New Zealand aligns somewhat with species with very long larval dispersal duration, such as Jasus edwardsii (Ovenden et al. 1992), gastropod genera Scutus and Austrolittorina (Waters et al. 2007), and the sea star Patiriella regularis (Waters and Roy 2004; Ayers and Waters 2005). A similar geographic pattern has been seen in Patiriella regularis, where a low mean divergence of 0.96% between populations determined a north–south disjunction south of the Cook Strait (Ayers and Waters, 2005). These longer dispersal species often show little morphological variation, and genetic homogeneity.

The connectivity in Fellaster zelandiae contrasts with other echinoderm taxa, including the pantropical cidaroid sea urchin genus Eucidaris (Lessios et al. 1999), and with sand dollars: notably, the Atlantic genus Mellita (Coppard et al. 2013) the widespread genus Encope (Coppard and Lessios 2017), and the species Clypeaster subdepressus (Zigler et al. 2008). For the genera Mellita and Encope, phylogeographic analyses have found speciation along the coasts of the Americas as a result of poor gene flow. Fellaster zelandiae has a longer larval development time of 28 days than the development time in the other sand dollar genera (5–13 days; Emlet et al. 1987; Herrera et al. 1996), which provides for a greater potential dispersal distance. Similar levels of population connectivity to Fellaster zelandiae are seen only in one other sand dollar species, Clypeaster rosaceus (Zigler et al. 2008). Clypeaster rosaceus is one of the few species of echinoderm with a facultative planktotrophic larvae, providing a longer larval period than for typical echinoderm species (Emlet 1986b). Similarities in population homogeneity and larval dispersal times in Clypeaster rosaceus, Evechinus chloroticus and Fellaster zelandiae indicate that medium-to-long development times drive connectivity within regions.

We assume from the homogeneity of the GBS results that the lack of variation is due to ongoing gene flow. However, an alternative is that New Zealand sand dollar populations suffered population extirpations sometime in the past, and have had a recent population expansion. Habitat loss, such as what may occur during sea level drop associated with the last glacial maxima c. 20,000 years ago, could have affected Fellaster zelandiae population ranges and sizes in the past. However, seeing the similarities in the population homogeneity of Evechinus chloroticus and Fellaster zelandiae, two species with very similar larval developments and planktonic larval durations but distinct habitat requirements within shallow coastal ecosystems, makes the scenario of local extinction and then recolonisation for both species less likely.

Future genetic connectivity

The extent of the genetic homogeneity of Fellaster zelandiae as an echinoderm species that lives across such a large gradient may give insight into how gene flow occurs in long larval-dispersal species compared with shorter larval-dispersal species such as tropical echinoderms, and how gene flow may ameliorate climate change. Changing environmental conditions in the global oceans have predicted, but largely unquantified, negative implications for the survival of many marine species. Future ocean conditions (FOC) are expected and include changes to pH, temperature, primary productivity and hydrology, and Fellaster zelandiae populations currently occur in a gradient of all these variables. Phenotypic plasticity currently seen in this species may posit high adaptation potential of this species, so long as gene flow remains the same.

One impact of FOC on Fellaster zelandiae may be a decrease in larval thermal windows and shortening of the dispersal time because of the positive effects of moderate warming and increases in water column productivity predicted for New Zealand coastal waters (Law et al. 2018). This has been reported for other species that encounter warmer (García et al. 2015a, 2015b; Cui et al. 2024) and more productive environments (Emlet 1986a). Fellaster zelandiae currently experiences a 28-day larval phase, in which time it could disperse across regions along the coast; however, studies of F. zelandiae larvae raised in FOC decreased the development window (Dixon-Anderson 2025). Recently, potential larval dispersal (PLD) modelling across New Zealand for a 30-day dispersal window (similar to the development window of F. zelandiae) found a maximal dispersal distance of 400 km, but a mean dispersal distance of only 30 km (Michie et al. 2024). Larval connectivity among populations of different geographic regions can occur in 30-day dispersal models, but these predictions are relative to the strength and direction of the currents (Chiswell and Rickard 2011). In a few locations, such as the eastern coast of the South Island, larval dispersal occurred in a biased direction (i.e. northward), consistent with other marine taxa (Collins et al. 2010). In calm location embayments, eddies and weak currents may confine dispersal to a small region, resulting in the majority of recruitment being local. Thus, currents are a key driver in range sizes and changes to genetic connectivity in F. zelandiae. Because Michie et al. (2024) found that 7- and 14-day PLD had limited larval exchange among modelled locations, the shortening of larval windows for F. zelandiae may have important consequences on its population biology. This may indicate that genetic connectivity among populations that live at either extreme of these environments may weaken and diverge. Because this research has found the subtle hydrographic and geographic breaks that currently provide the genetic diversity (Table S1), we expect these breaks to widen and reduce the gene flow that occurs across them. As the majority of locations were seen to self-seed in each duration of model (Michie et al. 2024), it can be inferred that most populations are not reliant on external spawning for sustaining populations. This is a positive outlook, because no populations of F. zelandiae are in danger of extinction from genetic recession, merely that genetic homogeneity is at risk. Re-assessing the population genetics of F. zelandiae in the future will elucidate how larval ranges and biogeographical processes have been altered over geographic and temporal stretches.