Is the sword moss (Bryoxiphium) a preglacial Tertiary relict?

Highlights

• The extant distribution of the moss Bryoxiphium is a striking case of Tertiary range fragmentation.

• Alaska was recurrently colonized from East Asia.

• Macaronesia was colonized twice from North American and East Asian ancestors.

• Icelandic populations were founded about two million years ago from Macaronesian ancestors.

Abstract

The disjunction of floras between East Asia, Southeast North America, West North America, and Southwest Eurasia has been interpreted in terms of the fragmentation of a once continuous mixed mesophytic forest that occurred throughout the Northern Hemisphere due to the climatic and geological changes during the late Tertiary. The sword moss, Bryoxiphium, exhibits a distribution that strikingly resembles that of the mesophytic forest elements such as Liriodendron and is considered as the only living member of an early Tertiary flora in Iceland. These hypotheses are tested here using molecular dating analyses and ancestral area estimations. The results suggest that the extant range of Bryoxiphium results from the fragmentation of a formerly wider range encompassing North America and Southeast Asia about 10 million years ago. The split of continental ancestral populations is too recent to match with a continental drift scenario but is spatially and temporally remarkably congruent with that observed in Tertiary angiosperm relict species. The timing of the colonization of Iceland from Macaronesian ancestors, about two million years ago, is, however, incompatible with the hypothesis that Bryoxiphium is the only living member of an early Tertiary flora of the island. Alaska was recurrently colonized from East Asia. The ability of Bryoxiphium to overcome large oceanic barriers is further evidenced by its occurrence on remote oceanic archipelagos. In particular, Madeira was colonized twice independently from American and East Asian ancestors, respectively. The striking range disjunction of Bryoxiphium is interpreted in terms of its mating system, as the taxon exhibits a very singular pattern of spatial segregation of the sexes.

Introduction

Cooling climates in the later part of the Tertiary period (65–2 million years ago; hereafter Ma) forced large assemblages of warm temperate to subtropical biotas to retreat from medium to high latitude circumboreal distributions southwards to large refugial regions in East Asia, Eastern and Western North America, and Southwest Eurasia, which preserved the warm wet climate that they needed (Wen, 1999, Milne and Abbott, 2002). The floras concerned are considered as Tertiary relict and have long captured the interest of botanists (Tiffney, 1985a, Tiffney, 1985b, Wen et al., 2010). In angiosperms, over 100 genera exhibit disjunct distributions between two or more of these regions (Wen, 1999, Milne and Abbott, 2002, Milne, 2006), with disjunctions between Eastern North America and Eastern Asia being by far the most common (Xiang et al., 2000, Donoghue and Smith, 2004). The last Tertiary land connection between Eurasia and North America was the Bering land bridge, which experienced significant climate cooling from 15 Ma onwards, initiating divergence and speciation between North American and Eurasian species of many Tertiary relict genera (Milne and Abbott, 2002).

The tulip tree, Liriodendron, offers one of the most striking examples of the fragmentation of a once continuous mixed mesophytic forest that occurred throughout the Northern Hemisphere following climatic and geological changes during the late Tertiary and Quaternary (Wen, 1999, Xiang et al., 2000). Liriodendron currently occurs in Eastern North America (L. tulipifera) and Southeast Asia (L. chinense), but unequivocal fossils dating back to the Miocene in Idaho, upper Pliocene in Europe, middle and late Miocene in Japan (Xiang et al., 2000), and upper to middle Miocene in Iceland (Denk et al., 2005) attest of a previously much wider, and more continuous range.

Similar striking disjunctions are commonly observed in mosses, and are hence sought to be interpreted by a common process, namely vicariance (Shaw, 2001). However, molecular dating evidence revealed that patterns that are spatially congruent with an ancient vicariance scenario might in fact be temporally congruent with a more recent long-distance dispersal hypothesis (for review see Villarreal and Renner, 2014). In fact, bryophytes appear as extremely efficient dispersers based on world-scale patterns of floristic turnover (Patiño et al., 2014), correlations between floristic patterns and wind connectivity (Muñoz et al., 2004), experimental evidence (Lönnell et al., 2012, Lönnell et al., 2014, Sundberg, 2013) and substantial allele sharing among trans-oceanic populations (Szövényi et al., 2008, Shaw et al., 2014, Shaw et al., 2015a, Shaw et al., 2015b). Furthermore, bryophytes may be prone to biotic dispersal across continents and oceans (Lewis et al., 2014). Altogether, these observations support the hypothesis that, in spore-producing plants (van Zanten and Pócs, 1981, Wolf et al., 2001) and other mobile organisms (Sanmartín and Ronquist, 2004, Yoder and Nowak, 2006, Perrie and Brownsey, 2007, Sanmartín et al., 2007, De Queiroz, 2014), dispersal has obscured any signature of vicariance. In fact, recent re-analysis of the only study inferring geographic vicariance due to continental drift in mosses (McDaniel and Shaw, 2003) yielded much younger ages for the relevant split that are incompatible with a continental drift scenario (Villarreal and Renner, 2014).

The sword moss Bryoxiphium, which exhibits a strikingly disjunct distribution spanning East Asia, North and Central America, Iceland, Mauritius, and Macaronesia, appears as one of the best remaining candidates for a vicariant origin in bryophytes. As Steere (1937) suggested, ‘B. norvegicum is an old species. It is, to judge from its present distribution, undoubtedly a member of a once widespread flora, which extended over the Northern Hemisphere in late Cretaceous or nearly Tertiary times. The great uniformity of appearance and lack of variation within each species are also marks of old groups. One is reminded here of some of the conservative and unchanging phanerogams, as Ginkgo, Sequoia, or Liriodendron, which may be contemporary in geological age with this loss. The fossil distribution of the tulip tree resembles the present distribution of Bryoxiphium, and the two may be members of the same flora, and a relic of an epoch with a much more equable climate’ (Steere, 1937, p. 350). Given the extremely old stem age of 137 [62–219] Ma inferred for Bryoxiphium in the context of a dating phylogenetic analysis of the entire moss phylum (Laenen et al., 2014), it is tempting to see in the highly disjunct distribution of Bryoxiphium, which, as Steere (1937) noticed, is strikingly reminiscent of the past distribution of Liriodendron, the signature of such a relictual origin. In this context, Bryoxiphium has been considered to be the only living member of an early Tertiary flora in Iceland (Lawton, 1971). As Schuster (1983) noted indeed, and in contrast to the global equilibrium of higher plant species distributions with present climate (Araujo and Pearson, 2003), bryophytes are able to persist in microhabitats where a suitable microenvironment endures, long after the general climate of the region has changed. As Anderson (1963) suggested, ‘the diminutive bryophytes, therefore, potentially offer a better clue to the solution of phytogeographical problems than do many vascular plants’.

Here we employed phylogeographic reconstructions based on phylogenetic analyses, molecular dating analyses, and estimations of ancestral ranges, to infer the historical biogeography of Bryoxiphium and test the hypothesis that its disjunct distribution results from the fragmentation of the once continuous mixed mesophytic forest that occurred throughout the Northern Hemisphere until the late Tertiary.