Sporophyte frequency in British dioicous liverworts

ABSTRACT Introduction. The extent and causes of variation in sporophyte production among British dioicous liverworts are assessed. Methods. Reproductive attributes of dioicous species are taken from Paton’s The Liverwort Flora of the British Isles (1999). The extent to which sporophyte production is potentially restricted by low levels of sex expression, unbalanced phenotypic sex ratios, and probable spatial segregation of the sexes, is quantified. Patterns of variation in sexual reproduction are also characterised for species associated with different biogeographical zones in Europe. Results. Of the 188 British liverwort species that are dioicous (62%), sporophytes are unknown in 65, rare in 58, occasional in 50 and frequent in 15. Some species are not or rarely sexually reproductive, but major causes of restricted sexual reproduction are sexual imbalance and probable isolation of the sexes. Among species with skewed sex ratios, female bias is substantially more common than male bias. Sporophyte frequency is relatively high among species widely distributed in temperate and boreal Europe, compared with taxa associated with other climatic zones, including the oceanic fringe. Conclusions. Restricted sporophyte production implies that asexual reproduction and clonal longevity are critical for the persistence of many liverworts. Further insights will require field surveys of patterns of sex expression and sporophyte distribution, together with information on reproductive plasticity, genotypic sex ratios, and levels of genetic diversity and structure.


Introduction
In bryophytes, the sporophyte generation is of particular significance for the production of spores via meiosis, as well as dispersal capacity and gene flow (Korpelainen et al. 2005;Patinõ and Vanderpoorten 2018), yet sporophyte frequency varies considerably among species. Approximately 60% (mosses) to 70% (liverworts) of bryophyte species are dioicous (Vanderpoorten and Goffinet 2009), and sporophyte production is more often frequent in monoicous species than in dioicous species (Longton and Schuster 1983). In dioicous taxa, comparatively high levels of genetic variation have been observed in some populations with frequent sexual reproduction (e.g. Wyatt et al. 2005). By contrast, low frequency of sexual reproduction may be associated with high levels of genetic differentiation between populations (Kyrkjeeide et al. 2014).
Major causes of restricted sporophyte production are the absence of sex expression, or in dioicous species, skewed sex ratios and spatial segregation of the sexes (Longton 1990). In dioicous bryophyte species, segregation of sex chromosomes at meiosis generally leads to expectation of a balanced sex ratio among spore progeny, although there are examples of bias in spore sex ratio (de Jong et al. 2018). Adult sex ratios in dioicous species are often female biased (Bisang and Hedenäs 2005), and these have been compared with genotypically determined sex ratios (e.g. Bisang et al. 2015;Ekwealor et al. 2017;Bisang et al. 2020). In the moss Drepanocladus lycopodioides, Bisang et al. (2017) found a balanced meiotic sex ratio, but a strongly female-biased genetic sex ratio in adult populations. Expressed male bias also occurs in some cases, and sex ratios may vary between populations of the same species, as has been observed in the liverworts Marchantia inflexa (Fuselier and McLetchie 2004) and Scapania undulata (Holá et al. 2014).
Other features of sexual reproduction in dioicous bryophytes that limit the success of sporophyte production include the need for females and males to coexist in close spatial association due to a highly restricted, water-mediated fertilisation range of mostly less than 1 m (Glime and Bisang 2017), although longer fertilisation distances have been determined in some species. Pressel and Duckett (2019) recently recorded fertilisation across distances of up to 19 m in the weedy liverwort Marchantia polymorpha subsp. ruderalis. In addition, there is the cost of sporophyte burden on females (e.g. Bisang and Ehrlén 2002) so that low fertilisation rates could result in female bias (de Jong et al. 2018). This may arise from an intersex performance hierarchy in growth rate where non-sporophytic females outperform males, which in turn outperform sporophytic females (Rydgren et al. 2010). Ceratodon purpureus is the classic example (Kollar et al. 2021). In some perennial species at least, males appear to be more limited than females by suboptimal habitat and climatic conditions (e.g. Bowker et al. 2000;Blackstock 2015). There may also be sexual differences in spore germination patterns (e.g. McLetchie 1992) and physiology (e.g. McLetchie and Puterbaugher 2000).
Most species of bryophytes are known to produce sporophytes somewhere (Longton and Schuster 1983), but in dioicous species they may be absent or rare in parts of their range at local, regional, and in some cases, continental scales. In Pleurozium schreberi, for instance, sporophytes are frequent in the boreal zone but rare in temperate regions (Longton 1976), suggesting that climate regime influences reproductive activity. It has further been suggested that historical factors may have left a signature on disrupted sex ratios, notably the Pleistocene climate deterioration in the northern hemisphere (Longton and Schuster 1983;Schuster 1983). The extent to which different causes of variation in sporophyte frequency operate at a regional scale in dioicous species has not hitherto been quantified.
In the present investigation, regional data on major causes of variation in sporophyte frequency were assessed among the dioicous British liverwort species. Following Gemmell (1950), several studies have analysed reproductive data from major Floras covering British bryophytes. Of these, Longton (1992) and Laaka-Lindberg et al. (2000) assessed sporophyte frequency in relation to rarity for mosses and liverworts, respectively. The two studies revealed similar trends, notably that a higher proportion of dioicous than of monoicous species fail to produce sporophytes in Britain, that a higher proportion of non-sporophytic species are rare relative to those that are sporophytic, and that among sporophytic species a higher proportion of monoicous than of dioicous species are rare. In liverworts, species that do not produce either spores or asexual propagules are most likely to be rare, but production of gemmae and other asexual diaspores was not found to be associated with rarity or sexual system.
More recently, Paton's The Liverwort Flora of the British Isles (1999) has, unlike its predecessors, provided consistent information on sex expression and sex ratio imbalance at a national scale, as well as sporophyte frequency, for virtually all taxa. In the current study, these reproductive traits are quantified for all dioicous species with the aim of comparing their contributions to variation in sporophyte frequency in Britain. Additionally, sporophyte frequencies in Britain are compared among groups of species with different biogeographical and climatic affinities in Europe.

Species selection
The primary source of data on British liverwort species with a dioicous sexual system is Paton (1999), but taxonomy and nomenclature follow Blockeel et al. (2021). The handful of native dioicous species that have been newly recorded post Paton are included; these are Anastrophyllum alpinum (Long et al. 2006), Clevea hyalina (Long et al. 2003), Conocephalum salebrosum (Sweykowski et al. 2005), Herbertus norenus (Bell et al. 2012) and Moerckia hibernica (see Blockeel et al. 2014). In all, 188 native dioicous taxa are covered, representing 62% of the British liverwort flora. Subspecies of Gymnocolea inflata and Marchantia polymorpha are included, but varieties ignored. Numbers of taxa referred to in tables and text include species and subspecies. Species recorded exhibiting both dioicous and monoicous morphs in Britain or in other parts of their European range are excluded. In addition, seven species that were recently introduced in Britain are treated separately from the overall assessment of presumed long-established native taxa, including Heteroscyphus fissistipus (Hodgetts 2020) and Tricholepidozia lindenbergii (Blackstock et al. 2019), which are not included in Paton (1999).

Reproductive attributes
For each of the 188 native dioicous species and subspecies, as well as the seven introduced species, frequency of sex expression and sporophyte production, and where appropriate sex ratio bias, were compiled from Paton (1999), or for more recently recorded species, extracted from Blockeel et al. (2014) and other published information. Each reproductive trait was assigned to one of four frequency classes, namely unrecorded (U), rare (R), occasional (O) or frequent (F), which were treated consistently when qualified by terms such as 'usually', 'often' and 'very'. The frequency classes are relative terms that are not quantified in Paton (1999) or other sources but apply across the distribution of each species in Britain, although not necessarily in other parts of their range. In the present context, they are used to indicate how often reproductive structures had been observed on gametophytic plants.
Sex expression is used to indicate the frequency that female and/or male reproductive structures have been recorded in Britain. In the case of Plagiochila asplenioides, sex expression is recorded as rare in Paton (1999), but recent unpublished field and herbarium data indicate that both sexes are expressed occasionally (personal observations). For species reported to exhibit skewed sex ratios, only one sex may be recorded or one sex is recorded as occasional or frequent and the other sex recorded as rare.
Species recorded with one sex rare and the other as very rare are not included in this category.

Biogeography
The set of 188 native dioicous liverworts was allocated to phytogeographical zones based on their distribution in northern and western Eurasia, originally defined for British bryophytes by Hill and Preston (1998). The zonation was devised to assemble species with similar climatic requirements. Allocation of species to different zonal groups follows the updated version of Hill et al. (2007), and additional information for more recently recorded species was obtained from other sources, mainly Blockeel et al. (2014). The groups cover nine major biomes associated with different latitudinal zones and seven eastern-limit longitudinal zones; the zonal boundaries are mapped in Hill and Preston (1998).

Statistical analysis
Sporophyte frequencies of the representative liverworts were compared statistically and graphically within the biome and eastern-limit geographical classes. Chi-square contingency table analyses were undertaken in both cases to test for associations between biogeographical zones and sporophyte frequency and assess the null hypothesis of independence between variables.

Results
A table of all 188 dioicous native species of British liverworts is presented in Appendix Table A1, with species divided into eight groups based on reproductive attributes. Frequency of sex expression and sporophyte production are itemised, and each species is also assigned to its biome and eastern-limit phytogeographical groups.
The seven introduced dioicous liverworts are also included in a separate group. Data from Appendix Table A1 groups are summarised in Tables and Figures  incorporated in the main text.

Overall variation in sporophyte frequency
Among the 188 native dioicous liverwort species in Britain, sporophyte production is unknown, rare or occasional in similar proportions, but in only relatively few cases is it recorded as frequent (Table 1).

Variation in sex expression, sexual bias and sporophyte frequency
Approximately 10% of British dioicous liverworts do not or rarely form sexual structures, but the reproductive group that contributes most (34%) to absence of sporophytes is that composed of species with only expressed females known (Table 2).
Of the 30 species with only one sex recorded, female-only species outnumber male-only species by 2.8 to 1. Among the 26 species with both sexes recorded in biased sex ratios, female bias predominates by a ratio of 2.3:1. In half of the latter species sporophytes are unrecorded, but they are rare in 11 species and are occasional in two cases.
Among the 56 species (30% of the dioicous taxa) with expressed sexual bias, the overall ratio of female to male bias is 2.5:1.
Of the remaining 60% of the dioicous species, with both sexes expressed occasionally or frequently, 50 species produce sporophytes rarely or in a few cases not at all, and 63 species are sporophytic occasionally or in 15 cases frequently (see Table 2).

Introduced species
In four of the seven introduced species, both sexes have been recorded (see Appendix Table A1). These include Lophocolea bispinosa, L. semiteres and Tricholepidozia lindenbergii, which have each also been recorded with sporophytes in Britain, and T. tetradactyla, which is non-sporophytic. Only female plants of Heteroscyphus fissistipus and L. brookwoodiana have been observed in Britain. The aquatic Riccia rhenana has not been reported with fertile thalli.

Biogeographical relationships
For the chi-square contingency table analyses, taxa with sporophytes occasional and frequent were combined because there are only few (15) British dioicous species with frequent sporophyte production. In addition, certain zones with relatively few representatives were combined, as indicated in the captions for the bar graphs in Figures 1 and There are significant associations between biogeographical zonation and sporophyte frequency among both the latitudinal biomes (chi-square = 19.83, df = 10, p < 0.05) and eastern-limit zones (chi-square = 20.04, df = 8, p < 0.05).
Among the biomes, the boreo-temperate zone has a relatively high proportion of species with occasional or frequent sporophyte production in Britain but fewer species that do not produce sporophytes than taxa in other zones (see Figure 1).
Among the eastern-limit zones, the hyperoceanic and oceanic categories have relatively large proportions of taxa that are non-sporophytic, while the European zone has a relatively high proportion of species that produce sporophytes occasionally or frequently, compared with taxa in other zones (see Figure 2).

Discussion
In approximately a third of the dioicous liverwort flora sporophytes have never been recorded in Britain, and they have rarely been observed in a little under a further third. Among the more commonly sporophytic species, sporophyte production is mostly recorded as occasional and rarely frequent. Similar proportions of sporophyte frequency have been reported for British dioicous mosses (Longton 1992). Low levels of sporophyte production run counter to expectation in a temperate-oceanic region where climatic conditions are likely to be favourable for fertilisation and sporophyte development. For many dioicous liverworts in Britain reproduction and dispersal appear to be wholly or largely through asexual systems, although immigration from external sources may also play a role. At a global scale, Laenen et al. (2016) found that geographical range in liverworts is correlated with production of asexual diaspores, rather than sexual system, suggesting that further investigation of the dispersal capacity of gemmae and other vegetative propagules is needed.
Although allocation of species to different reproductive groups in this account is based on the long history of exploration of the British bryophyte flora, rather than structured survey, the data for overall frequency of sporophytes are likely to be reliable in the great majority of cases. However, it would be useful to have more precision on the distribution of reproductive plants and sporophyte production of dioicous liverworts. This is especially so in leafy liverworts in which the development of protective structures may be deceptive, notably in the case of perianths that can form in the absence of fertilisation within various genera, including Lophozia, Plagiochila and Scapania (Paton 1999).
Sex expression, sex ratios and potential spatial segregation of the sexes Low levels of sex expression and particularly skewed sex ratios are associated with failure or restriction of sporophyte production in British dioicous liverworts. It also appears highly likely that spatial segregation of the sexes plays an important role among species with commonly expressed females and males, but in those which have infrequent sporophyte production. It is possible that fertilisation failure is partially limited by other factors, such as sexual incompatibility or phenological differences between the sexes, but no evidence is available for liverwort species in Britain.
Species with only non-reproductive plants, for example Marsupella arctica, and those with both functional sexes recorded rarely in Britain, for example Diplophyllum taxifolium, account for 10% of the dioicous liverwort flora. Of the latter, sporophytes have been observed rarely in four of the 11 species. The majority of the species are upland or montane in Britain, although the lowland aquatic Riccia fluitans is a notable exception.
Among the 56 species (30%) of British dioicous liverworts with reported sexual disparity, there are substantially more cases of overall female bias than of male bias (2.5F:1M) in Britain, which is in accordance with the commonly observed trend across bryophytes (Bisang and Hedenäs 2005). These data for liverwort phenotypic sex ratios may not reflect genotypic sex ratios that have been assessed, using genetic markers to determine the sex of non-reproductive individuals, for only a restricted range of bryophytes including the desert moss Syntrichia caninervis (Ekwealor et al. 2017) and the fen species Drepanocladus lycopodioides (Bisang et al. , 2020. Of the 30 species in the reproductive groups with only one observed sex functional, female-only taxa (22) make the largest single contribution to species lacking sporophytes in Britain. In another 26 species the frequency of one expressed sex (most often female) exceeds that of the other, and some of these are rarely and two species occasionally sporophytic. At a national level, there thus appears to be a considerable impact of unbalanced sex ratios on sporophyte production. Although direct evidence is lacking, it seems probable that spatial segregation of the sexes, as well as sexual disparity, prevents sporophyte production in a number of species, such as Metzgeria consanguinea and Orthocaulis atlanticus (both with female bias), whereas in Saccogyna viticulosa there is female prevalence and rare males that are found in mixedsex, usually sporophytic, populations (Blackstock 2020).
The strong signal of female bias among a number of dioicous species suggests that males are relatively often disfavoured following capsule dehiscence. In certain cases, this may arise initially from sexual differences in spore survival and establishment, as found in the liverwort Sphaerocarpos texanus (McLetchie 1992). Sexual differences in habitat use, but often with considerable overlap, have been reported in some bryophyte studies (e.g. Bowker et al. 2000;Fuselier and McLetchie 2004;Bisang et al. 2020) but not in others (Bisang and Hedenäs 2005;Bisang et al. 2015), and phylogenetic history may also have an influence on reproductive patterns (Bisang et al. 2014). Other causes of female bias can be higher female growth rates (Rydgren et al. 2010;Kollar et al. 2021), lower levels of male sex expression (the shy male hypothesis), or greater male mortality (Ekwealor et al. 2017). Reproductive attributes may also change during successional development (Hedenäs et al. 2021). Additionally, sex ratio patterns can be influenced by non-sexual reproduction, particularly in species with specialised asexual propagules, such as Marchantia inflexa (Stieha et al. 2017;Brzyski et al. 2018).
Despite an overall regional bias in one expressed sex, variation among populations may occur (e.g. Fuselier and McLetchie 2004). In a study of Scapania undulata in streams in Finland, sexual parity and both male and female bias were observed among sites but with an overall excess of males, which may be a strategy counteracting sperm dilution in flowing water (Holá et al. 2014). In Welsh populations of Frullania tamarisci, balanced sex ratios were recorded in woodlands, while there was strong female bias in exposed coastal habitats (Blackstock 2015).
In the group of 50 species (27% of British dioicous liverworts) that are occasionally or often sex expressing with both sexes present, but among which sporophyte production is unrecorded in seven and rare in 43 species, it seems probable that spatial isolation of the sexes plays a major role in restricting sexual reproduction to low levels. Such segregation could take place at varying spatial scales. Included in this group are a number of peat bog liverworts, such as Cephalozia macrostachya, Mylia anomala, Odontoschisma fluitans and O. sphagni, among which regeneration takes place from subsurface flagelliform axes in some species (Duckett and Clymo 1988). There are also perennial stayers, such as Bazzania trilobata and Lepidozia cupressina, that are characterised by long-term persistence and low reproductive output, but with large numbers of small spores produced if and when they  Hill et al. (2007). Taxa in biomes 2 and 3, and 8 and 9, were combined due to small numbers of taxa in biome 3 (n = 3) and biome 9 (n = 10). No taxa occur in biome 6. Chi square = 19.83, df = 10, p < 0.05.
are sporophytic (Longton 1997). This is a group for which further insights on sexual segregation could be obtained from field studies that examine distribution of reproductive plants of both sexes within and among populations.
Of the other 63 dioicous British liverwort species (34%) with commonly fertile plants of both sexes, sporophyte production is mostly recorded as occasional and in 15 cases frequent. They comprise a rather heterogeneous group in terms of habitat, life-history traits, and distribution in Britain. It may be that British rarities included, such as Clevea hyalina, Petalophyllum ralfsii and Southbya tophacea, that do not produce asexual propagules, are at least partially dependent on spore dispersal for local persistence. It is also notable that this group includes a number of common ruderals, such as Marchantia polymorpha subsp. ruderalis and Solenostoma gracillimum, that may be dependent on high spore output tracking suitable but ephemeral habitats that are patchily distributed. Another example is Nardia scalaris, in which mixed-sex populations and sporophytes are frequent and widely distributed in north Wales (Blackstock 2006). Occasional sporophyte production is also associated with epixylic species, including Cephalozia catenulata, Cephalozia lunulifolia and Scapania umbrosa, and epiphytes, such as Metzgeria furcata, including those that lack asexual propagules (e.g. Frullania dilatata and F. tamarisci, which produces sporophytes frequently in western British woodlands ;Blackstock 2015). There is also a range of common species in the uplands that are often sporophytic, for example Diplophyllum albicans, Gymnomitrion crenulatum, Lophozia ventricosa and Marsupella emarginata.

Introduced species
The seven dioicous species introduced to Britain, and first recorded since 1950 (Blockeel et al. 2014), have persisted to some extent in gardens and a few in seminatural habitats. Most of the species have probably arrived on horticultural material, usually from Australasia. It is remarkable that in four cases both sexes have been imported; in three of these cases, the species are sporophytic in Britain (as is Heteroscyphus fissistipus in southwest Ireland). In the non-sporophytic Tricholepidozia tetradactyla female and male introductions are widely separated, indicating different founding events. The aquatic Riccia rhenana, probably introduced from aquarium sources, has not been reported with sex-expressing thalli. Further information is given in Paton (1999), Blackstock and Long (2002), Blockeel et al. (2014), Blackstock et al. (2019) and Hodgetts (2020).

Biogeographical relationships
From a latitudinal viewpoint, sporophyte production in dioicous liverworts in Britain is relatively frequent in species of the boreo-temperate zone, for example Metzgeria furcata and Scapania nemorea, which range over large parts of western and northern Europe, and include many bryophyte species that are common in Britain (Hill and Preston 1998). This contrasts with the large more northerly boreal-montane group, with sporophyte production unknown or rare in many species, for example Harpanthus scutatus and Mesoptychia bantriensis. These differences may be due to direct effects of climatic conditions or have arisen indirectly through habitat conditions in Britain and competitive interactions.
Among the more southerly groups, recent survey findings in Britain indicate that sporophyte frequency has increased recently in the southern-temperate epiphyte Metzgeria violacea (Blackstock 2016) and that male plants of the Mediterranean weed Lunularia cruciata have recently become more widely reproductive (Blackstock 2018) in Britain. Both species exhibit female Figure 2. Variation in sporophyte frequency among British dioicous liverworts in different eastern-limit zones. Eastern-limit categories as listed in Hill et al. (2007). Only one taxon occurs in each of eastern-limit categories 4 and 5, so these data were combined with eastern-limit category 3. Chi square = 20.04, df = 8, p < 0.05.
bias. These changes may be due to recent increases in annual temperatures in Britain, but further evidence is required.
Among the groups of species defined by eastern limits of their distribution in Eurasia, a relatively large proportion of the dioicous liverworts in the European zone produce sporophytes in Britain, while the hyperoceanic and oceanic elements have relatively large proportions of non-sporophytic taxa. A notable feature is the strong representation of hyperoceanic and oceanic taxa (17 species), with only one sex known in Britain (see Appendix Table  A1); examples include Harpalejeunea molleri and Metzgeria leptoneura (both female only) and Adelanthus lindenbergianus (male only). Historical suppression of sex expression or possible loss of one sex (commonly male) may have played a role in present-day absence of sporophyte production (Damsholt 2017; Blackstock 2020). Further insights on sexual history in these and other cases are likely to require population data on genetic diversity and structure, in the context of past colonisation and extraregional reproductive performance (Flagmeier et al. 2020;Boquete et al. 2022).

Acknowledgements
I owe a special debt of gratitude to Mrs Jean Paton for the thorough treatment of reproductive attributes in her Flora, on which this overview is based. I am also very grateful for editorial guidance, and especially for critical and constructive commentary from two reviewers, that helped considerably to improve this account. Particular thanks are also due to Dr Jane Stevens for help with statistical analysis and data presentation.

Disclosure statement
No potential conflicts of interest were reported by the author.

Funding
None.

Notes on contributor
Tim Blackstock worked for the Nature Conservancy Council and subsequently the Countryside Council for Wales in Bangor. His main focus was habitat ecology and conservation, with a specialist interest in bryology. Since retirement he has undertaken bryological projects, mostly on liverwort reproductive biology.

Supplemental material
Supplemental material for this article can be accessed here: https://doi.org/10.1080/03736687.2023.2188388.