Read an Excerpt
CHAPTER 1
Relationships of Angiosperms to Other Seed Plants
Seed plants are of fundamental importance both evolutionarily and ecologically. They dominate terrestrial landscapes, and the seed has played a central role in agriculture and human history. There are five extant lineages of seed plants: angiosperms, cycads, conifers, gnetophytes, and Ginkgo. These five groups have usually been treated as distinct phyla — Magnoliophyta (or Anthophyta), Cycadophyta, Coniferophyta, Gnetophyta, and Ginkgophyta, respectively. Cantino et al. (2007) used the following "rank-free" names (see Chapter 12): Angiospermae, Cycadophyta, Coniferae, Gnetophyta, and Ginkgo. Of these, the angiosperms are by far the most diverse, with ~14,000 genera and perhaps as many as 350,000 (The Plant List 2010) to 400,000 (Govaerts 2001) species. The conifers, with approximately 70 genera and nearly 600 species, are the second largest group of living seed plants. Cycads comprise 10 genera and approximately 300 species (Osborne et al. 2012; Fragnière et al. 2015). Gnetales consist of three morphologically disparate genera, Gnetum, Ephedra, and Welwitschia (~90 species total) that are so distinctive that each has been placed in its own family (Gnetaceae, Ephedraceae, and Welwitschiaceae). There is a single living species of Ginkgo, G. biloba. Each of these extant lineages has a rich fossil history (T. Taylor et al. 2009; Friis et al. 2011); we cover the fossil record of the angiosperms in more detail in Chapter 2 and also in those chapters focused on specific angiosperm clades (Chapters 4–10).
There are also many extinct lineages of seed plants (Crane 1985; Decombeix et al. 2010; E. Taylor and T. Taylor 2009; T. Taylor et al. 2009). Although extant gymnosperms appear to be monophyletic (below and Chapter 2), all gymnosperms (living and extinct) together are not monophyletic. Importantly, several fossil lineages, Caytoniales, Bennettitales, Pentoxylales, and Glossopteridales (glossopterids), have been proposed as putative close relatives of the angiosperms based on phylogenetic analyses (e.g., Crane 1985; Rothwell and Serbet 1994; reviewed in Doyle 2006, 2008, 2012; Friis et al. 2011). These fossil lineages, sometimes referred to as the para-angiophytes, will therefore be covered in more detail later in this chapter. Another fossil lineage, the corystosperms, has been proposed as a possible angiosperm ancestor as part of the "mostly male hypothesis" (Frohlich and Parker 2000), but as reviewed here, corystosperms usually do not appear as close angiosperm relatives in phylogenetic trees.
The seed plants represent an ancient radiation, with the first seeds appearing near the end of the Devonian (~370 million years ago; mya). By the Early to Middle Carboniferous, a diversity of seed plant lineages already existed (e.g., Cordaites and walchian conifers; Thomas 1955; Bhatnagar and Moitra 1996; Kenrick and Crane 1997; Davis and Kenrick 2004; T. Taylor et al. 2009). Heterospory, prerequisite to evolution of the seed, evolved in parallel in different major clades, including lycophytes, water ferns (e.g., Marsilea), sphenophytes, and aneurophytes, and seed-like structures, with a retained endosporic megagametophyte nearly surrounded by an integument-like covering, occur in some lycophytes (e.g., Lepidocarpon). Lepidocarpon is not considered a true seed, but is an example of convergence. Importantly, phylogenetic analyses that include the five clades of living seed plants show that they indeed form a clade, indicating that all have inherited seeds from a common ancestor — and that these seeds did not evolve in parallel. Phylogenetic analyses including extinct seed plants also place these groups in the same clade as extant seed plants (see below). Thus, analyses support the hypothesis that fossil and extant seed plants (Spermatophyta) had a single origin.
The first seed-like structures, observed in the Late Devonian to early Carboniferous, apparently ancestral to true seeds, had free integumentary lobes and lacked a micropyle; the pollen-receptive structure, lagenostome, was formed by the nucellus rather than the integument. The fusion of integumentary lobes, except for a micropylar channel, led to the formation of true seeds as in lyginopterid seed ferns.
The fossil record of conifers dates to the Late Carboniferous and that of true cycads to the Early Permian. Available data indicate that by the Permian (~299–251 mya), at least three (cycads, conifers, Ginkgo) of the five extant lineages of seed plants had probably diverged (Kenrick and Crane 1997; Donoghue and Doyle 2000). In contrast, the angiosperms are relatively young — their earliest unambiguous fossil evidence is from the Early Cretaceous (~130 mya) although molecular dating methods infer older dates for their origin (see Chapter 2).
Relationships among the lineages of extant seed plants, as well as the relationships of living groups to fossil lineages, have been issues of longstanding interest and debate. A topic of particular intrigue has been the closest relatives of the angiosperms. Angiosperms are responsible either directly or indirectly for the majority of human food and account for a huge proportion of photosynthesis and carbon sequestration. They have diversified to include 350,000–400,000 species in perhaps 130–170 myr and now occur in nearly all habitable terrestrial environments and many aquatic habitats. Understanding how angiosperms accomplished this is of fundamental evolutionary and ecological importance.
At some point, nearly every living and fossil group of gymnosperms has been proposed as a possible ancestor of the angiosperms (e.g., Wieland 1918; Thomas 1934, 1936; Melville 1962, 1969; Stebbins 1974; Meeuse 1975; Long 1977; Doyle 1978, 1998a,b; Retallack and Dilcher 1981; Crane 1985; Cronquist 1988; Crane et al. 1995; reviewed in Doyle 2006, 2008, 2012; Friis et al. 2011). Among extant seed plants, the relationship between angiosperms and Gnetales has received considerable attention.
Ascertaining the closest relatives of the angiosperms is not only of great systematic importance but also critical for assessing character evolution. For example, the outcome of investigations of character evolution among basal angiosperms, including studies focused on the origin and diversification of crucial angiosperm structures (e.g., floral organs, endosperm, vessel elements), may be influenced by those taxa considered their closest relatives. The effect of outgroup choice on the reconstruction of character evolution within angiosperms is readily seen via the widespread use of Gnetales as an outgroup for angiosperms. As reviewed below, for nearly two decades beginning in the 1980s, Gnetales were considered by many to represent the closest living relatives of the angiosperms. The use of Gnetales as an angiosperm outgroup profoundly influenced character-state reconstruction within the flowering plants (see "The Anthophyte Hypothesis" section).
Clarifying relationships among seed plants, both extant and fossil, has been extremely difficult. Factors that have contributed to the difficulties in phylogeny reconstruction of seed plants (living and extinct) include the great age of these groups and the considerable morphological divergence among them, as well as the extinction of many lineages. The tremendous morphological gap among extant and fossil seed plant lineages has complicated and ultimately compromised efforts to reconstruct relationships with morphology because of homoplasy and uncertainty about the homology of structures (e.g., Doyle 1998a, 2006, 2012; Donoghue and Doyle 2000; Soltis et al. 2005b, 2008b; Friis et al. 2011).
Although progress has been made in elucidating relationships among extant seed plants using DNA sequence data, relationships remain problematic. Even with the addition of more taxa and more genes representing all three plant genomes, issues remain. Resolution of relationships among extant seed plants with DNA sequence data has also been difficult because some lineages have relatively short branches (e.g., angiosperms or Pinaceae), whereas other clades (e.g., Gnetales) have long branches. This problem is further compounded by the presence in most analyses of long branches to the sister group of seed plants (monilophytes). In groups such as the angiosperms and conifers, more taxa can be added to break up long branches, but this is not possible across seed plants as a whole given the considerable extinction that has occurred. Another concern given the ancient divergences in seed plants is multiple substitutions per site leading to saturation of base substitutions. Hence, whereas the use of morphological characters has been criticized in seed plant phylogeny (and in a global sense by Scotland et al. 2003), DNA has its own problems and certainly has not been a consistent solution to resolve relationships among extant seed plants (see Burleigh and Mathews 2004, 2007; Mathews 2009; Mathews et al. 2010; Soltis et al. 2005b, 2008b).
As stressed by other investigators, a complete understanding of seed plant phylogeny is not possible without the integration of fossils. Many investigations have attempted this integration (e.g., Crane 1985; Doyle and Donoghue 1986; Doyle 1996, 1998a,b, 2001, 2006, 2008, 2012; Frohlich 1999; Donoghue and Doyle 2000; Hilton and Bateman 2006; Magallón 2010); we cover these analyses in more detail later in this chapter. Despite these efforts, the integration of fossils into studies of seed plant phylogeny remains an area where more research is needed. Seed plant relationships and the closest relatives of the angiosperms have been the focus of many reviews (e.g., Crane 1985; Doyle and Donoghue 1986; Doyle 1996, 1998a, b, 2001, 2006, 2008, 2012; Frohlich 1999; Donoghue and Doyle 2000; Mathews 2009; Friis et al. 2011) and continue to spawn new analyses (e.g., Hilton and Bateman 2006; Doyle 2008, 2012; Magallón 2010; Mathews et al. 2010). We will consider seed plant relationships in general (living and extinct), but a major focus of this chapter is discussing the closest relative(s) of the angiosperms.
Phylogenetic Studies: Extant Taxa
We first review the considerable effort devoted to reconstructing the phylogeny of living seed plants. Given the immense debate regarding the relationships of Gnetales, we also provide a brief history of the placement of Gnetales relative to the angiosperms. We then focus on cladistic analyses that include fossil as well as extant seed plants.
Molecules and morphology have so far yielded different conclusions about the relationships of Gnetales and angiosperms. Whereas analyses of morphology have consistently placed Gnetales sister to angiosperms (but see review by Rothwell et al. 2009), molecular data support alternative placements (see below). We are strong advocates for the use of morphology in phylogenetic analyses. However, based on the morphological characters so far used, the coding employed, and analyses now available, one could legitimately conclude that to this point seed plants represent an example in which cladistic analyses of morphological characters alone have failed to resolve major relationships in congruence with molecular-informed analyses.
Placement of Gnetales
A close relationship of angiosperms and Gnetales was first proposed by Wettstein (1907) and by Arber and Parkin (1908) based on several shared features: vessels, net-veined leaves (present in Gnetum as well as angiosperms), and "flower-like" reproductive organs (Fig. 1.1) (see also reviews by Doyle 1996; Frohlich 1999). However, the reasoning that Wettstein (1907) and Arber and Parkin (1908) each used to explain the close relationship of Gnetales and angiosperms differed dramatically. Wettstein (1907) proposed that Gnetales were ancestral to the angiosperms based on the view that the formerly recognized angiosperm group Amentiferae, a group that included wind-pollinated taxa such as Juglandaceae, Betulaceae, and Casuarinaceae, are the most "primitive" living angiosperms. We stress throughout that which extant group exhibits the most "primitive" morphological traits and which is sister to all others are not equivalent, but these statements are often confounded. We can infer ancestral character states via character-state reconstruction using the best estimate of phylogeny, as we have done throughout (Chapter 6). Wettstein maintained that the distinctive inflorescences (termed aments) of Amentiferae, consisting of simple, unisexual flowers, are homologous with the unisexual strobili of Gnetales. Arber and Parkin (1908) also proposed a close relationship of angiosperms and Gnetales, but, in contrast, argued that the reproductive structures of Gnetales are not primitively simple, but reduced, derived from ancestral lineages having more parts.
By the mid-1900s, most authors no longer considered Gnetales and angiosperms closest relatives. Bailey (1944b, 1953) noted that the vessels in the two groups are derived from different kinds of tracheids and hence are not homologous. In addition, Gnetales bear ovules directly on a stem tip, whereas in angiosperms, the ovules are produced within the carpel, the latter structure possibly representing a modified leaf. Views on the earliest angiosperms also changed, with Magnoliaceae and other angiosperms with large, strobiloid flowers considered most ancient, whereas the simple flowers found in Amentiferae were considered secondarily reduced rather than ancestrally simple (e.g., Arber and Parkin 1907; Cronquist 1968; Takhtajan 1969). This new view disrupted the link between Gnetales and angiosperms (via a basal Amentiferae) envisioned by Wettstein.
Issues became more complex when Eames (1952) proposed that the three lineages of Gnetales were not monophyletic. Eames considered Ephedra to be related to the fossil group Cordaites and conifers while Gnetum and Welwitschia were thought to be closer to another extinct lineage of seed plants, Bennettitales. Although morphology and DNA later confirmed the monophyly of Gnetales (below), the work of Eames (1952) shifted interest away from Gnetales as an angiosperm relative. Concomitantly, paleobotanists focused attention on fossils such as Caytonia and Glossopteridales as the closest relatives of angiosperms(Doyle 1996; Frohlich 1999; see below), further diverting attention from Gnetales as possible close relatives of the angiosperms. Gnetales re-emerged, however, as putative close relatives of angiosperms when cladistic approaches were first used to investigate seed plant relationships (below).
The Anthophyte Hypothesis
Seed plant relationships were first assessed by cladistic methodology using morphological characters in the 1980s. Several of these early studies included both extant and fossil taxa (e.g., Crane 1985; Doyle and Donoghue 1986). These studies revealed that the three morphologically disparate members of Gnetales (Ephedra, Gnetum, and Welwitschia) are monophyletic (illustrated in Fig. 1.1), a finding now well supported by both morphology and molecules. Only Nixon et al. (1994) found Gnetales not to be monophyletic. However, Doyle's (1996) subsequent reanalysis of the data used in Nixon et al. (1994) found a monophyletic Gnetales.
Early phylogenetic studies relying on morphological characters (Parenti 1980; Hill and Crane 1982; Crane 1985; Doyle and Donoghue 1986) recovered Gnetales as the closest living relatives of angiosperms (Fig. 1.2). Perhaps the best known is Crane (1985), which also included fossil seed plants. Crane (1985) recovered Gnetales as the sister group to angiosperms (Fig. 1.3). Subsequent phylogenetic analyses of morphological characters (e.g., Loconte and Stevenson 1990; Doyle and Donoghue 1992; Doyle 1994, 1996; Hilton and Bateman 2006), some of which also included fossils, continued to recover this Gnetales + angiosperm relationship (Fig. 1.2); as summarized by Rothwell et al. (2009; p. 296), "the anthophyte topology of the seed plant tree continues to be supported by morphological analyses of living and extinct taxa."
However, these same early cladistic studies often differed in the relationships suggested among extant seed plants (see Fig. 1.2). In morphological cladistic analyses, extant gymnosperms do not form a clade distinct from the angiosperms, and the positions of some lineages were unstable. Considering extant seed plant lineages, Crane (1985) found that cycads are sister to other extant seed plants and that conifers + Ginkgo form a clade that is sister to angiosperms + Gnetales (Fig. 1.2A). In contrast, the shortest trees of Doyle and Donoghue (1986) indicated that conifers + Ginkgo are sister to a clade in which cycads are the sister to angiosperms + Gnetales (Fig. 1.2B). Loconte and Stevenson (1990) found cycads followed by Ginkgo, then conifers, to be subsequent sisters to Gnetales + angiosperms (Fig. 1.2C).
(Continues…)
Excerpted from "Phylogeny and Evolution of the Angiosperms"
by .
Copyright © 2018 The University of Chicago.
Excerpted by permission of The University of Chicago Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
