• angiosperms;
  • Cretaceous;
  • eudicots;
  • flowers;
  • Magnoliidae;
  • paleobotany;
  • phylogeny;
  • pollen


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix

Abstract  Over the past 25 years, discoveries of Early Cretaceous fossil flowers, often associated with pollen and sometimes with vegetative parts, have revolutionized our understanding of the morphology and diversity of early angiosperms. However, few of these fossils have been integrated into the increasingly robust phylogeny of living angiosperms based primarily on molecular data. To remedy this situation, we have used a morphological dataset for living basal angiosperms (including basal eudicots and monocots) to assess the most parsimonious positions of early angiosperm fossils on cladograms of Recent plants, using constraint trees that represent the current range of hypotheses on higher-level relationships, and concentrating on Magnoliidae (the clade including Magnoliales, Laurales, Canellales, and Piperales) and eudicots. In magnoliids, our results confirm proposed relationships of Archaeanthus (latest Albian?) to Magnoliaceae, Endressinia (late Aptian) to Magnoliales (the clade comprising Degeneria, Galbulimima, Eupomatia, and Annonaceae), and Walkeripollis pollen tetrads (late Barremian?) to Winteraceae, but they indicate that Mauldinia (early Cenomanian) was sister to both Lauraceae and Hernandiaceae rather than to Lauraceae alone. Among middle Albian to early Cenomanian eudicots, we confirm relationships of Nelumbites to Nelumbo, platanoid inflorescences and Sapindopsis to Platanaceae, and Spanomera to Buxaceae. With the possible exception of Archaeanthus, these fossils are apparently not crown group members of living families but rather stem relatives of one or more families.

The past four decades have seen unprecedented growth in knowledge of the Early Cretaceous fossil record of angiosperms and its implications for their origin and early evolutionary radiation. This progress began with recognition that the first dispersed pollen of angiosperms was far less diverse and modern than expected (Scott et al., 1960; Hughes, 1961; Brenner, 1963), contradicting the earlier view (elaborated by Axelrod, 1960, 1970) that angiosperms were already highly diversified when they first entered the Cretaceous fossil record. Instead, studies of dispersed pollen and leaves (Fig. 1) showed a progressive diversification of morphological types through the mid-Cretaceous (Doyle, 1969; Muller, 1970; Wolfe et al., 1975; Brenner, 1976; Doyle & Hickey, 1976; Doyle et al., 1977; Hickey & Doyle, 1977; Walker & Walker, 1984; Hughes, 1994). This began in the Hauterivian and Barremian stages with monosulcate angiosperm pollen, distinguished from the monosulcate pollen of other seed plants by its reticulate-columellar exine structure, which was joined in the Aptian and early Albian by simple angiosperm leaves with irregular pinnate to palmate venation. Tricolpate pollen, the basic type for the vast clade known as eudicots, appears near the Barremian–Aptian boundary in Northern Gondwana (Brenner, 1976; Doyle et al., 1977; Doyle, 1992) but is rare in Laurasia before the early Albian (Kemp, 1968; Hochuli et al., 2006; Heimhofer et al., 2007). The middle and late Albian diversification of tricolpate pollen coincides with the appearance of more diverse leaf types, including palmately veined cordate and peltate leaves, palmately lobed “platanoids,” and pinnately lobed and compound leaves called Sapindopsis. Tricolporate pollen appears in the late Albian and triporate pollen in the middle Cenomanian. Most early studies of these fossils emphasized general trends and their consistency with hypotheses on character evolution derived from studies of living plants and refrained from systematic comparisons with living taxa, except to argue that the record supported the view that the monosulcate groups then assigned to “Magnoliidae” (in the sense of Takhtajan, 1966, 1980), now recognized as a paraphyletic basal grade of angiosperms, were relatively plesiomorphic.


Figure 1. Stratigraphic sequence of major angiosperm pollen and leaf types in the Potomac Group of eastern North America (modified from Doyle & Hickey, 1976, with permission from C. Beck [editor]), with correlations of plant-bearing localities and fossil taxa in other geographic areas.

Download figure to PowerPoint

This situation changed dramatically in the 1980s, thanks to discoveries of Cretaceous flowers and fruits (Friis, 1983; Crane et al., 1986; Friis et al., 1986), mostly in the millimeter size range (mesofossils) and preserved as lignite or charcoal, which offer far greater opportunities for systematic comparisons (summarized by Crane et al., 1995; Friis et al., 2006). Dispersed pollen and leaves are still important for a broad picture of early angiosperm evolution, thanks to their broader geographic and stratigraphic sampling, but there is no question that flowers should be easier to relate to modern clades. Not only do they have many more characters, but they often have pollen in the stamens or on the stigma, allowing them to be correlated with the dispersed pollen record. In fact, many Cretaceous flowers have been assigned to modern taxa, such as Chloranthaceae, Lauraceae, Calycanthaceae, Platanaceae, and Buxaceae (Crane et al., 1986, 1989; Friis et al., 1986; Drinnan et al., 1990, 1991; Pedersen et al., 1991, 1994; Eklund et al., 1997, 2004). However, Crepet et al. (2004) questioned most proposed systematic relationships of Early Cretaceous angiosperms, largely on the grounds that they had not been tested by phylogenetic analyses of living and fossil plants.

The present study is part of a project intended to remedy this situation, by using a morphological cladistic dataset to integrate fossils into the phylogeny of living angiosperms. This task has become easier as relationships among living taxa have become more firmly established by analyses of ever-increasing numbers of DNA sequences, now including nearly complete plastid genomes (Jansen et al., 2007; Moore et al., 2007). Since 1999 (Mathews & Donoghue, 1999; Parkinson et al., 1999; Qiu et al., 1999, 2006; Soltis et al., 1999, 2000, 2005; Barkman et al., 2000; Graham & Olmstead, 2000; Zanis et al., 2002), such analyses have continued to reveal the same rooting of the angiosperm tree, among the so-called ANITA lines—the New Caledonian genus Amborella, Nymphaeales, and Austrobaileyales—and the same major clades within the remaining angiosperms (Mesangiospermae of Cantino et al., 2007). It has been argued that such results mean that morphological cladistic studies are now obsolete (Scotland et al., 2003). However, even if the phylogeny of living plants can be reconstructed entirely from molecular data, we still need morphological data if we wish to place fossils on the molecular phylogenetic tree (Wiens, 2004), and this is worth doing for several reasons. Even if fossils do not affect the relationships that we infer, they can still provide unique evidence on character evolution, the sequence of events in the origin of modern clades, their age, and their distribution, ecology, and diversity in the geologic past.

Some general remarks on terminology and the phylogenetic logic for placement of fossils on trees may be useful (Doyle & Donoghue, 1993; Doyle et al., 2008b). The crown group contains all the living members of a clade, their most recent common ancestor, and any fossils derived from this ancestor. All groups seen in molecular analyses are crown groups. The stem lineage is the line leading from the base of the crown group to its common ancestor with its living sister group. A fossil attached to the stem lineage is a stem relative. A fossil is inferred to be a stem relative if it has the derived state (a synapomorphy) of the crown group in one character (or characters) but the ancestral (outgroup) state in another. A fossil is inferred to belong in the crown group if it has not only synapomorphies of the crown group as a whole but also a synapomorphy of some subgroup of the crown group.

The starting point for our present project was a morphological dataset for basal angiosperms that we analyzed alone and in combination with sequences of three genes (Doyle & Endress, 2000): 18S nrDNA, rbcL, and atpB (Soltis et al., 2000). One goal of this study was to determine how much morphology and molecular data agree and disagree about relationships, by comparing trees obtained when the two kinds of data were analyzed separately and together. In most cases where morphology and DNA gave different results, the combined analysis agreed with the molecular tree, as expected from the far greater number of molecular characters. For example, it placed Nymphaeales in the basal ANITA grade, rather than with monocots. However, there were a few exceptions. For example, the first branch above the ANITA grade—the basal line in mesangiosperms, or the sister group of all other mesangiosperms—was Chloranthaceae, which are notable for their unusually simple flowers, and the sister group of Lauraceae was Hernandiaceae, rather than Monimiaceae or Monimiaceae plus Hernandiaceae. Another goal was to reconstruct early morphological evolution in angiosperms, by using parsimony to plot (optimize) characters on the tree obtained from the combined dataset. For example, this analysis showed that the ancestral carpel was of the ascidiate type, as in the ANITA lines and Chloranthaceae (Endress & Igersheim, 2000), and not the plicate type, as in Degeneria and some Winteraceae, which was considered primitive by Bailey & Swamy (1951) and many subsequent authors.

Since 2000 we have been improving this dataset by adding and redefining characters, splitting taxa into more homogeneous units, and adding new taxa. Additions and improvements concerning pollen have been presented in Doyle (2005), leaves in Doyle (2007), and floral phyllotaxis in Endress & Doyle (2007). Most of the dataset was published in Endress & Doyle (2009), devoted to reconstruction of the flower in the most recent common ancestor of extant angiosperms and its initial modifications. This included all floral characters and others needed to evaluate the implications of morphology for placement of the living aquatic genus Ceratophyllum and the controversial Early Cretaceous fossil Archaefructus (Sun et al., 1998, 2002; Friis et al., 2003), but omitting characters that cannot be scored in these taxa, such as wood anatomy. Ceratophyllum was not included by Doyle & Endress (2000) because half its characters were missing or problematic and it was expected to have little effect on reconstructing character evolution, but we added it in Endress & Doyle (2009) because of its potential relevance to the placement of fossils and recent discussions of the importance of the aquatic habit and reduced flowers in early angiosperm evolution (Sun et al., 2002; Friis et al., 2003; Coiffard et al., 2007; Rudall et al., 2007). Doyle et al. (2008a) also used this dataset in an analysis of Early Cretaceous fossils that have been interpreted as monocots. In the present paper we have further modified the scoring of some floral characters in Laurales and Piperales based on re-evaluation of older data and newer information (Wanke et al., 2007; Staedler & Endress, 2009) and increased the number of states in the carpel number character to better express the diversity of conditions in living and fossil taxa.

The ideal way to integrate fossils into the phylogeny of living plants might be a “total evidence” analysis of living and fossil taxa together using both morphological and molecular data, with the fossils scored as unknown for the latter. Considerable work will be required to amass an enlarged molecular dataset and resolve the problems of combining molecular data on species with morphological data on higher taxa, so we have used a temporary approximate approach, analyzing the morphological dataset with the arrangement of living taxa fixed using “backbone constraint” trees based primarily on molecular data (called a “molecular scaffold” approach: Springer et al., 2001; Manos et al., 2007). The fossils then attach to this backbone tree wherever it is most parsimonious in terms of morphology. Essentially this assumes that the relationships among living taxa are so robust that addition of fossils would not affect them. This is likely considering the strong statistical support for most molecular clades, as measured by bootstrap analysis, and the small number of characters preserved in fossils, but it should be tested with future total evidence studies.

In Doyle et al. (2008a) and Endress & Doyle (2009) we used two backbone trees, chosen to represent the range of current hypotheses on relationships. Most relationships within major clades are strongly supported by molecular data. The greatest uncertainties concern the arrangement of the five major clades that make up the mesangiosperms: Magnoliidae in the monophyletic sense of Cantino et al. (2007), comprising Magnoliales, Laurales, Canellales (including Winteraceae), and Piperales (including Aristolochiaceae); monocots; eudicots; Chloranthaceae; and Ceratophyllum. Relationships among these lines vary among analyses depending on the genes or analytical methods used, presumably because there was very little molecular divergence between splitting events. One backbone tree (designated J/M) assumed relationships found in analyses of nearly complete plastid genomes (Jansen et al., 2007; Moore et al., 2007), with Chloranthaceae sister to Magnoliidae, and with monocots sister to Ceratophyllum and eudicots. The other (D&E) was an updated version of the arrangement found in the combined analysis of Doyle & Endress (2000), with Ceratophyllum added in its most parsimonious position based on morphology. In this tree, Chloranthaceae and Ceratophyllum form a clade that is sister to other mesangiosperms, in which eudicots are basal and monocots are linked with magnoliids. Some molecular support for these relationships of Chloranthaceae and Ceratophyllum has been found in analyses of cp rDNA ITS, nuclear, and mitochondrial genes (Antonov et al., 2000; Duvall et al., 2006, 2008; Qiu et al., 2006). In both trees, we kept relationships within these clades the same, with new taxa added and others rearranged based on more recent molecular data (e.g., transfer of Piperales to the magnoliid clade), except for the sister group relationship of Lauraceae and Hernandiaceae, the case in which morphology most strongly overrode molecular data in the analysis of Doyle & Endress (2000). Use of these two backbone trees had relatively little effect on inferences on character evolution or the positions of fossils.

Considering our results to date, Endress & Doyle (2009) analyzed the position of Archaefructus (Sun et al., 1998, 2002), an aquatic plant with dissected leaves and simple flowers from the Barremian or Aptian of China. Sun et al. (2002) concluded that Archaefructus was more basal than all living angiosperms (i.e., a stem relative), but this was questioned by Friis et al. (2003), who argued that it was instead a reduced member of the crown group. Our analysis (Endress & Doyle, 2009) indicated that the most parsimonious position of Archaefructus within the angiosperms was sister to the highly reduced family Hydatellaceae, formerly considered monocots but shown by Saarela et al. (2007) to be basal Nymphaeales, or to Ceratophyllum, depending on the backbone tree used. A related analysis of seed plant phylogeny (Doyle, 2008) concluded that a crown group relationship to Hydatellaceae was more parsimonious than a position on the angiosperm stem lineage. Doyle et al. (2008a) evaluated the positions of Early Cretaceous fossils that have been interpreted as monocots, most of which were questioned by Gandolfo et al. (2000). This analysis confirmed the monocot affinities of Aptian and Albian monosulcate pollen of the classic Liliacidites type, with finer sculpture at the ends of the grain, but other pollen types once associated with Liliacidites have been found in flowers that appear not to be monocots. Albian pollen that Walker & Walker (1984) called “Liliaciditesminutus was associated by Friis et al. (1994a) with Virginianthus, which they assigned to Calycanthaceae; according to our analysis, Virginianthus is sister either to Calycanthaceae or to the remaining Laurales. The pollen genus Similipollis, lumped with pollen of the Liliacidites type by Doyle (1973) and Walker & Walker (1984) but segregated by Góczán & Juhász (1984) and dissociated from monocots by Doyle & Hotton (1991), was associated by Friis et al. (1997) with Albian flowers called Anacostia, which our analysis nested within the basal order Austrobaileyales. Aptian and Albian pollen of the coarsely reticulate Pennipollis type and associated stamens and carpels, which Friis et al. (2000) compared with Alismatales, were inferred to represent an extinct outgroup of Chloranthaceae, or of Chloranthaceae and Ceratophyllum if these two taxa are sister groups.

The focus of the present paper is on fossils that appear to belong to the magnoliid clade and the eudicots. Most of these are flowers described by Friis, Crane, and coworkers, but they include one dispersed pollen type, Walkeripollis, which has been related to Winteraceae (Canellales; Walker et al., 1983; Doyle et al., 1990a, 1990b). We emphasize that this task would have been impossible without the highly detailed, rigorous, and consistent descriptions of the floral fossils. We have restricted our survey to the Early Cretaceous and earliest Late Cretaceous (early Cenomanian) because of the systematic scope of our extant dataset, which includes basal eudicots, basal monocots, and more basal groups (“magnoliids” in the old paraphyletic sense). Younger beds show the radiation of core eudicots, or Pentapetalae of Cantino et al. (2007), and greatly expanded taxon sampling would be needed in order to avoid incorrectly forcing pentapetalous taxa into more basal groups. We have used the same approach that we took in Doyle et al. (2008a) and Endress & Doyle (2009), introducing fossil taxa to the same backbone trees based primarily on molecular data. Prominent Early Cretaceous taxa that we have not yet treated include fossils related to Chloranthaceae, such as Couperites and flowers with Asteropollis pollen (Pedersen et al., 1991; Friis et al., 1994b), which were included in a morphological analysis with much more limited outgroup taxon sampling by Eklund et al. (2004), and somewhat similar floral structures of more problematic affinities, such as Appomattoxia (Friis et al., 1995).

1 Material and methods

  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix

1.1 Data for Recent plants

A list of Recent taxa in our dataset is provided in Appendix I; references to taxon-specific studies on their morphology and studies on internal relationships that we consulted in order to estimate ancestral states for variable characters may be found in Endress & Doyle (2009). Sources of data on fossil taxa and interpretations of their characters are presented in the following section. In contrast to Doyle & Endress (2000), where we considered putative fossil relatives in scoring some taxa, but as in Doyle et al. (2008a) and Endress & Doyle (2009), all extant taxa are defined as crown groups, that is, they do not include potential fossil outgroups, and fossils were not considered in scoring their characters.

Definitions of characters and their states are presented in the Appendix I, along with documentation and argumentation for changes in the dataset made since Doyle & Endress (2000) that were not covered in Endress & Doyle (2009) and the additional modifications made since the latter paper. In Doyle & Endress (2000) we avoided including autapomorphies, on the grounds that they have no effect on inferred relationships, but we have changed this policy in view of the present and future use of this dataset for placement of fossils, because features that are found in only one extant taxon may be synapomorphies that would unite it with fossil taxa.

1.2 Fossil taxa and basis for character scoring

For ease of comparison with Endress & Doyle (2009) and Doyle et al. (2008a), we have included Archaefructus (interpreted as having an inflorescence of unisexual flowers), the Pennipollis plant, Anacostia, Liliacidites, and Virginianthus in the data matrix (Appendix I, Table 1).

Table 1.  Data matrix
inline image

1.2.1 Archaeanthus  Compressions of elongate floral axes with numerous multiovulate follicles in the fruit stage (Archaeanthus linnenbergeri Dilcher & Crane), bilobed leaves (Liriophyllum kansense Dilcher & Crane), isolated perianth parts (Archaepetala beekeri and Archaepetala obscura Dilcher & Crane), and bud-scales (Kalymmanthus walkeri Dilcher & Crane), which Dilcher & Crane (1984) interpreted as parts of the same plant based on association and possession of similar resin bodies, from the Dakota Formation at Linnenberger Ranch, Kansas, USA. Some of these fossils were figured earlier by Dilcher et al. (1976) and Dilcher (1979). Dilcher & Crane (1984) considered this locality latest Albian or earliest Cenomanian in age, but studies of carbon isotope stratigraphy by Gröcke et al. (2006) indicate that the Rose Creek flora in Nebraska is latest Albian, and the same may be true for the Archaeanthus bed, based on regional transgressive relations (D.L. Dilcher, pers. comm., 2009).

No stem anatomy is preserved but, as noted by Dilcher & Crane (1984), the stoutness of the floral axis implies that it had cambial activity. Scars on the stem indicate alternate phyllotaxis and presence of a sheathing leaf base. We use the bilobed character of the bud-scales, which resemble the stipular bud-scales of Magnoliaceae, to score the plant as having paired cap-like stipules. The bizarre bilobed leaves are too highly modified to be assigned to either our obovate-elliptical or ovate shape state. We follow Dilcher & Crane in interpreting the resin bodies as diagenetically altered contents of oil cells; they noted that oil cells are often preserved in leaves of Tertiary Lauraceae. Only carpels are known attached to the floral axis, but the presence of a large sheathing bract (calyptra) and the perianth organization can be inferred from scars on the floral axis. We use the occurrence of two size classes of detached tepals to score the tepals as differentiated into sepaloid and petaloid. Between the perianth scars and carpels is a zone with what Dilcher & Crane interpreted as scars of numerous spiral stamens, but we find the scars too obscure to score their phyllotaxis. Although it is often impossible to distinguish ascidiate and plicate carpels without anatomical or developmental data (Endress, 2005), the two distinct adaxial crests imply that the carpels were most likely plicate. However, we consider the identification of an extended stigma and papillae uncertain. Many carpels show ventral dehiscence (and often contain sediment), but we hesitate to characterize the fruit wall. Dilcher & Crane considered ovule curvature uncertain but probably anatropous; we find the shape of the seeds more consistent with an anatropous than an orthotropous curvature. 41/142 = 28.9% of characters scored.

1.2.2 Endressinia  Based on a branching axis with attached leaves and several flowers (Endressinia brasiliana Mohr & Bernardes-de-Oliveira), preserved partly by impression and partly by iron oxide permineralization, described by Mohr & Bernardes-de-Oliveira (2004) from the Crató Formation in the Araripe Basin, Brazil. Based on palynological and ostracode correlations with better-dated coastal sequences, the age of this unit is probably late Aptian (see Mohr & Bernardes-de-Oliveira, 2004).

The leaves are too incomplete to determine whether they were obovate-elliptical or ovate, but they have pinnate venation and entire margins and contain round bodies reasonably interpreted as oil cells. Because some flowers appear to be solitary, but one has two buds lower on the axis, we interpret the inflorescence condition as either solitary or botryoids (0/1). The numerous outer floral organs pass from poorly preserved tepals of unclear phyllotaxis to better-preserved staminodes with marginal glands toward the center of the flower, and there are numerous free carpels with a distinct ventral slit, which we score as plicate, but no preserved ovules. No stamens were recognized, but this is probably due to poor preservation of the outer organs. The glands on the staminodes, which Mohr & Bernardes-de-Oliveira (2004) compared with those of Eupomatia, correspond to our character 69. 29/142 = 20.4% of characters scored.

1.2.3 Mauldinia (=Prisca?)  Lignitized and fusainized (charcoalified) inflorescence fragments and bisexual flowers (Mauldinia mirabilis Drinnan, Crane, Friis & Pedersen) described by Drinnan et al. (1990), and fusainized wood (Paraphyllanthoxylon marylandense Herendeen) described by Herendeen (1991) from the Potomac Formation at Mauldin Mountain, Maryland, USA. Herendeen associated the wood with the inflorescences based on identical anatomical features in the first-formed wood of P. marylandense and inflorescence axes of Mauldinia. The palynoflora at Mauldin Mountain falls in lower Zone III of the Potomac sequence, dated as earliest Cenomanian by palynological correlations with other areas (Doyle & Robbins, 1977). Retallack & Dilcher (1981) described more poorly preserved but apparently related axes, Prisca reynoldsii Retallack & Dilcher, from the Dakota Formation at the Linnenberger Ranch in Kansas, which as discussed above may be latest Albian. They associated Prisca with simple, pinnately veined leaves that they called Magnoliaephyllum sp., but because the evidence for association with the flowers from Maryland is too indirect we have not incorporated these leaves into our reconstruction.

The inflorescences have lateral bilobed cladodes bearing five sessile flowers, which we interpret as modified cymes. The whole inflorescence is therefore either a thyrse or a thyrsoid; which of these it is cannot be determined because it is not known whether or not the main axis terminated in a flower. Most floral characters are very clearly preserved: notably, two whorls of three tepals, the inner distinctly larger than the outer; a shallow hypanthium; stamens with paired basal glands and dehiscence by two upward-opening flaps, assumed to imply two microsporangia; one superior, glabrous carpel containing a single pendent seed; an inner layer of columnar cells in the carpel wall, resembling the endocarp of most Laurales; and endosperm in the seed. The outer two whorls of stamens are introrse, the innermost latrorse-extrorse, as in many Recent Lauraceae; as in Lauraceae we score them as introrse. No pollen was found in the stamens. Paracytic stomata are known on various organs; Drinnan et al. (1990) interpreted resin bodies in the tepals as evidence of oil cells. Data of Herendeen (1991) indicate that the wood had predominantly solitary vessels with mixed perforations, reduced fiber pits, narrow rays, and very sparse parenchyma. 57/142 = 40.1% of characters scored.

1.2.4 Walkeripollis  Dispersed pollen tetrads (Walkeripollis gabonensis Doyle, Ward & Hotton) described by Doyle et al. (1990a) from sample 2963 (939–944 m) in the N’Toum No. 1 well, Gabon, from the upper part of Elf-Aquitaine palynozone C-VII. Using palynological correlations with better-dated sequences elsewhere in Africa–South America, Doyle et al. (1977, 1982) dated this zone as early Aptian, but more recent data suggest that it may be late Barremian (Doyle et al., 1990a; Doyle, 1992).

The light microscope, scanning electron microscope, and transmission electron microscope (TEM) observations of Doyle et al. (1990a) allow all pollen characters to be scored unambiguously. 15/142 = 10.6% of characters scored.

1.2.5 Nelumbites  Impressions of peltate leaves (Nelumbites extenuinervis Upchurch, Crane & Drinnan), dispersed perianth parts with characteristic rows of dark spots, molds of three round floral receptacles, and rhizomes described by Upchurch et al. (1994) from the Potomac Formation at Quantico, Virginia, USA. Upchurch et al. interpreted these remains as parts of one plant based on the dominance of the leaves in the bed, where they were apparently preserved in situ in a pond environment (cf. Doyle & Hickey, 1976; Hickey & Doyle, 1977). The Quantico bed is barren palynologically, but based on the leaf flora and regional geologic relations it appears to correlate with the upper half of Subzone II-B of Brenner (1963), which is thought to be late middle to early late Albian in age (Doyle & Hickey, 1976; Doyle & Robbins, 1977; Upchurch et al., 1994). Beds with similar Nelumbites leaves at Mount Vernon and White House Bluff (Ward, 1895; Berry, 1911) are dated palynologically as middle Subzone II-B, probably late middle Albian (Doyle & Hickey, 1976; Hickey & Doyle, 1977).

For the purposes of this analysis we accept that the rhizomes and tepals belonged to Nelumbites, leading us to interpret the habit as rhizomatous and the flowers as having a perianth consisting of more than one whorl or series (based on variation in size) of either petaloid or both petaloid and sepaloid tepals. No carpels are preserved, but the floral receptacles are covered with numerous closely spaced pits similar to those in Nelumbo, each of which we assume contained a single free carpel. We assume that separation of the carpels would preclude presence of an extragynoecial compitum. 20/142 = 14.1% of characters scored.

1.2.6 West Brothers platanoid  Lignitized male and female heads and flowers (Platananthus potomacensis Friis, Crane & Pedersen and Platanocarpus marylandensis Friis, Crane & Pedersen) described by Crane et al. (1986) and Friis et al. (1988) from the Potomac Formation at the West Brothers clay mine, Maryland, USA. The male and female heads were associated based on similar floral organization and identical pollen in the stamens and on the female flowers. Doyle et al. (1975) presented light microscope and TEM observations on a mass of similar pollen (presumed anther contents) from the same bed, identified as Tricolpites cf. micromunus (Groot & Penny) Burger. The West Brothers pollen flora (Brenner, 1963; Doyle, 1969) was assigned by Doyle & Hickey (1976) to upper Subzone II-B, considered late Albian (Doyle & Robbins, 1977).

As in Recent Platanus, we cannot determine whether the heads were more likely modified botryoids or racemes, but we assume from the lack of visible grouping of flowers that the lateral units were single flowers rather than cymes. The flowers have a single whorl of five stamens or five carpels, but in contrast to some Late Cretaceous platanoids (Magallón et al., 1997) the number and phyllotaxis of the tepals are uncertain. However, there is clearly more than one perianth cycle or series. The stamens have a peltate connective apex, although more weakly differentiated than that of Platanus, and valvate dehiscence by an H-shaped slit. The shape of the anthers in cross-section is similar to that in Platanus, with pollen sacs at the limit between our protruding and embedded states (scored 0/1). The carpels are presumed to be plicate, based on the fact that they have a distinct line running down the ventral side, but they differ from Platanus in having a sessile stigma. We score fruit characters as unknown because the carpels are still attached and immature. No seeds are known. 44/142 = 31.0% of characters scored.

1.2.7 Sapindopsis  Compressions and impressions of pinnatifid to nearly compound leaves (Sapindopsis variabilis Fontaine) and lignitized male and female inflorescences and stamens (Aquia brookensis Crane, Pedersen, Friis & Drinnan and Platanocarpus brookensis Crane, Pedersen, Friis & Drinnan), described by Crane et al. (1993) from the Potomac Formation at Bank near Brooke, Virginia, USA. Sapindopsis leaves from Brooke were described by Fontaine (1889), Berry (1911), Doyle & Hickey (1976), Hickey & Doyle (1977), and (with cuticle analysis) Upchurch (1984). Hickey & Doyle (1977) noted axes bearing small heads in the leaf bed and speculated that they might belong to Sapindopsis. Crane et al. (1993) convincingly associated the inflorescences with the leaves based on cuticular similarities, and the female heads with the stamens based on adhering pollen. Palynological correlations place Brooke in lower Subzone II-B, thought to be middle Albian (Doyle & Hickey, 1976; Doyle & Robbins, 1977).

Several leaf characters are difficult to score because of the unusual pattern of leaf dissection, with the bases of the leaflets (more precisely lobes) more or less decurrent onto the rachis. Because the size of the leaflets increases toward the leaf apex, these leaves differ from the ternately dissected leaves of many Ranunculales, in which the basal lobes or leaflets are larger. Because dissected ranunculalean leaves resemble ovate, palmately veined simple leaves in overall shape and major venation, we score them as ovate and palmately veined. By similar reasoning, because the whole leaf in Sapindopsis is wider toward the apex, as are the individual leaflets, we score it as obovate-elliptical in shape and pinnate in major venation. Upchurch (1984) described Sapindopsis as having both paracytic and laterocytic stomata but did not indicate which type was more common, so we score stomatal type as uncertain (0/1). Our scoring of inflorescence structure follows the reasoning used for the West Brothers platanoid. Floral organization is known only for the female flowers, but it is clearer than in the West Brothers platanoid, with at least two whorls of tepals, five in the inner whorl, and five carpels. Stamens have a long filament and lack a peltate connective extension; the degree of pollen sac protrusion is unclear because the anthers had dehisced. The mature carpel dehisced along a ventral slit, so we assume it was plicate; it contained one pendent, orthotropous seed, but we cannot exclude the possibility that there was a second abortive ovule, as in Platanus. Seed anatomy is uncertain, but because there is no evidence of a palisade exotesta, we score the exotesta character as (0/2). 56/142 = 39.4% of characters scored.

1.2.8 Spanomera  A consensus of lignitized and fusainized flowers and inflorescences (Spanomera marylandensis Drinnan, Crane, Friis & Pedersen and Spanomera mauldinensis Drinnan, Crane, Friis & Pedersen) described by Drinnan et al. (1991) from the Potomac Formation at West Brothers (late Albian) and Mauldin Mountain (early Cenomanian), Maryland, respectively. Some of the association of parts, especially in Spanomera marylandensis, is based on presence of the distinctive pollen, which corresponds to the dispersed species Striatopollis vermimurus (Brenner) Srivastava and Striatopollis paraneus (Norris) Singh.

Preservation of Spanomera marylandensis is more fragmentary than that of Spanomera mauldinensis, but the two species are similar in most characters that are known in both; they differ in the number of tepals and stamens in the male flowers and coarseness of the pollen sculpture. The inflorescence is a botryoid with lateral male flowers and a terminal female flower. Following Drinnan et al. (1991), we interpret the flowers as basically dimerous, as in extant Buxaceae (giving the appearance of four stamens opposite four tepals), but with two anterior tepals and two stamens in double positions in Spanomera mauldinensis. Thickness of the pollen nexine is near the limit between our thin and thick states and difficult to judge because of the obliqueness of the TEM sections (scored 1/2). The two plicate carpels, which have an extended stigma with apparently unicellular papillae flanking a pronounced ventral slit, are fused only basally, less than in Recent Buxaceae but more than in any Recent taxa scored as apocarpous in our dataset. Seeds not known. 55/142 = 38.7% of characters scored.

1.3 Analyses

As in Doyle et al. (2008a) and Endress & Doyle (2009), most analyses were carried out with the arrangement of Recent taxa held constant using the two backbone constraint trees described in the Introduction and explained in detail in Endress & Doyle (2009). The D&E tree is based primarily on the combined analysis of morphology, 18S nrDNA, rbcL, and atpB by Doyle & Endress (2000), but with changes in topology and addition of taxa based on more recent analyses, with Ceratophyllum the sister group of Chloranthaceae and the two groups sister to the remaining mesangiosperms. The J/M tree incorporates relationships among mesangiosperms found in analyses of nearly complete chloroplast genomes (Jansen et al., 2007; Moore et al., 2007), with Chloranthaceae sister to Magnoliidae, Ceratophyllum sister to eudicots, and the two latter clades linked with monocots, but with the same relationships within clades as in the D&E tree.

Positions of fossils were evaluated by adding them individually or together to the extant dataset and carrying out parsimony analyses with PAUP (Swofford, 1990), with one of the two backbone trees, random addition of taxa, and TBR branch swapping. The strength of the relationships obtained and the relative parsimony of alternative arrangements were evaluated by searching for trees one, two, and sometimes more steps longer than the most parsimonious trees and by moving taxa manually with MacClade (Maddison & Maddison, 2003). Character evolution and characters supporting particular relationships obtained were evaluated with MacClade. When characters are described as unequivocal synapomorphies, this means that the position of the character state change is unequivocal, not necessarily that it occurs only once on the entire tree.

2 Results and Discussion

  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix

Results are presented in terms of trees (Figs. 2–8) in which the fossil under discussion is placed in its most parsimonious position on the D&E backbone tree, or in one of its most parsimonious positions if there are two or more. The thickest lines indicate all branches to which the fossil can attach in the set of most parsimonious trees, and successively thinner lines indicate branches to which it can attach in trees that are one and two steps less parsimonious.


Figure 2. One of three most parsimonious trees (1017 steps) obtained after addition of Archaeanthus to the D&E backbone constraint tree of Endress & Doyle (2009) (1015 steps). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Archaeanthus. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Archaeanthus linnenbergeri reproduced from Dilcher & Crane (1984), with permission from D.L. Dilcher and Missouri Botanical Garden.

Download figure to PowerPoint


Figure 3. One of seven most parsimonious trees (1015 steps) obtained after addition of Endressinia to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Endressinia. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Endressinia brasiliana reproduced from Friis et al. (2006), with permission from E.M. Friis and Elsevier.

Download figure to PowerPoint


Figure 4. Single most parsimonious tree (1018 steps) obtained after addition of Mauldinia (=Prisca?) to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Mauldinia. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Mauldinia mirabilis reproduced from Friis et al. (2006), with permission from E.M. Friis and Elsevier.

Download figure to PowerPoint


Figure 5. Single most parsimonious tree (1016 steps) obtained after addition of Walkeripollis to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Walkeripollis. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Scanning electron microscopy image of Walkeripollis gabonensis reproduced from Doyle et al. (1990a), with permission from Botanical Society of America.

Download figure to PowerPoint


Figure 6. Single most parsimonious tree (1015 steps) obtained after addition of Nelumbites to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Nelumbites. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Nelumbites extenuinervis reproduced from Doyle & Hickey (1976), with permission from C. Beck (editor).

Download figure to PowerPoint


Figure 7. One of two most parsimonious trees (1022 steps) obtained after addition of Sapindopsis and the West Brothers platanoid to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for West Brothers platanoid and Sapindopsis. *Branch not found in the other most parsimonious tree, in which the two fossils form a clade. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Platanocarpus brookensis (Sapindopsis) reproduced from Friis et al. (2006), with permission from E.M. Friis and Elsevier.

Download figure to PowerPoint


Figure 8. Single most parsimonious tree (1017 steps) obtained after addition of Spanomera to the D&E tree (Endress & Doyle, 2009). Thicker lines indicate all most parsimonious (MP), one step less parsimonious (MP + 1), and two steps less parsimonious (MP + 2) positions for Spanomera. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales. Drawing of Spanomera mauldinensis reproduced from Friis et al. (2006), with permission from E.M. Friis and Elsevier.

Download figure to PowerPoint

In the present analyses, use of the two different backbone trees never affected the number or topology of most parsimonious trees, in contrast to the situation for Virginianthus and Liliacidites in Doyle et al. (2008a) and Archaefructus in Endress & Doyle (2009). However, in some cases not all the trees that were one or two steps less parsimonious were the same with the two backbones. We mention differences among suboptimal trees found with the J/M backbone in the text. In general, placement of the fossils included in this study is less ambiguous than it was for the taxa treated in Doyle et al. (2008a) and Endress & Doyle (2009), and the relationships inferred are closer to those proposed by the authors who originally described them.

2.1 Magnoliidae

The first four fossils appear to belong to Magnoliidae in the monophyletic sense of Cantino et al. (2007), made up of the APG II (2003) orders Magnoliales, Laurales, Canellales, and Piperales. This should not be confused with Magnoliidae in the older paraphyletic sense of Takhtajan (1966, 1980), which also included the more basal ANITA lines, Chloranthaceae, and Ceratophyllum, or of Cronquist (1981), which included the APG II eudicot order Ranunculales. Our results confirm the growing impression, much of it gained from studies of fossil flowers, that the magnoliid clade was radiating vigorously alongside the eudicots, which are more conspicuous in the pollen and leaf records.

2.1.1 Archaeanthus  This fossil, from near the Albian–Cenomanian boundary in Kansas, was the first mid-Cretaceous flower to be described in detail. In 1976, Dilcher et al. reported axes with numerous follicles but no other floral parts, which they associated with peculiar bilobed leaves (Liriophyllum) in the same beds. More abundant and better-preserved material described by Dilcher & Crane (1984) showed that Archaeanthus was a bisexual flower with close correspondences to Magnoliaceae: below the follicles there is a zone with what appear to be scars of numerous stamens, scars of three cycles of three perianth parts, and a scar on the pedicel where Magnoliaceae have a calyptrate bract derived from two fused stipules, and there are similar detached tepals and bilobed bracts in the same bed. Despite these similarities, Dilcher & Crane (1984) hesitated to assign Archaeanthus to any living group, because its individual features are scattered across many different taxa, and because of these doubts Crepet et al. (2004) also considered its relationships uncertain. In the meantime, Archaeanthus has been used in several molecular dating studies as providing a minimum age for crown group Magnoliales (Magallón & Sanderson, 2001) or the clade comprising all Magnoliales except Myristicaceae (Doyle et al., 2004; Richardson et al., 2004).

Our analysis unambiguously associates Archaeanthus with Magnoliaceae (Fig. 2). It has three most parsimonious positions, one as the sister group of Magnoliaceae as a whole and two in the crown group, sister to either Liriodendron or to Magnolioideae (other Magnoliaceae, =Magnolia sensu lato of recent authors, e.g., Judd et al., 2008). Unequivocal synapomorphies of Archaeanthus and Magnoliaceae are the sheathing leaf base, bilobed stipules, and elongate receptacle. Other features that support a position in this part of the tree are more than one whorl or series of carpels (a synapomorphy of Magnoliales and Laurales), solitary flower (a synapomorphy of either Magnoliales plus Laurales, or of Magnoliales other than Myristicaceae—this is a case of equivocal character optimization), calyptra (a synapomorphy of Magnoliaceae or of Magnoliales other than Myristicaceae), and plicate carpels (a synapomorphy of mesangiosperms other than Chloranthaceae and Ceratophyllum on the D&E tree). The presence of more than two whorls (or series) of three tepals may be retained from the first angiosperms (Endress & Doyle, 2009). Archaeanthus differs from living Magnoliaceae in having a sessile stigma (i.e., no style), which supports a position on the stem lineage of the family. Its equally parsimonious position linked with Liriodendron is supported by the bilobed apex of the leaves (which however are otherwise different), whereas its position with Magnolioideae is based on the dehiscent fruit. Its next-best positions, which are two steps worse, are also in Magnoliales, linked with either Degeneria (which also has dehiscent fruit) or Galbulimima (with an elongate receptacle).

These results illustrate the general principle that scattered convergences with different groups do not necessarily preclude establishment of a relationship with one group—it is the combination of characters that counts. They also confirm use of Archaeanthus as a calibration point in molecular dating studies; specifically, it provides a minimum age of latest Albian (ca. 100 MY: Gradstein et al., 2004) for the node connecting Magnoliaceae with Degeneria, Galbulimima, Eupomatia, and Annonaceae.

2.1.2 Endressinia  This plant, represented by a shoot bearing several leaves and flowers from the late Aptian of Brazil, was assigned by Mohr & Bernardes-de-Oliveira (2004) to Magnoliales, based on such features as pinnate venation, oil cells in the leaves, solitary flowers, multiparted androecium with inner staminodes, and numerous plicate carpels. Mohr & Bernardes-de-Oliveira compared glands on the staminodes with those of Degeneria, Galbulimima, and Eupomatia, which serve as food bodies for pollinating beetles (Endress, 1984). Such glands also occur in the basal annonaceous genus Anaxagorea (Scharaschkin & Doyle, 2006).

Our analyses confirm that Endressinia is related to Magnoliales, specifically the clade consisting of Degeneria, Galbulimima, Eupomatia, and Annonaceae; all seven positions of the fossil immediately below and within this clade are equally parsimonious (Fig. 3). The only unequivocal synapomorphy uniting these taxa is the presence of glands on the stamens or staminodes, but other features are supportive of this position over others, such as more than two whorls or series of stamens (a synapomorphy of Magnoliales and Laurales on the D&E tree but potentially a retention from lower on the J/M tree), more than one whorl or series of carpels (synapomorphic for Magnoliales and Laurales: cf. Archaeanthus), and plicate carpels (cf. also Archaeanthus). Although Degeneria, Galbulimima, Eupomatia, and Annonaceae are the only Magnoliales with inner staminodes, these may or may not be a synapomorphy, as they are also ancestral in Laurales and may be a synapomorphy of Magnoliales plus Laurales, with a loss in Magnoliaceae (and Myristicaceae, which we argued might have lost any staminodes as a side-effect of unisexuality and we therefore scored as unknown). The features of Degeneria and Galbulimima that Mohr & Bernardes-de-Oliveira (2004) cited as differences from Endressinia (such as single carpel and elongate receptacle) do not rule out relationships with these taxa, as they are autapomorphies. With the D&E backbone tree, the four next most parsimonious placements of Endressinia, which are one step worse, are also in or sister to Magnoliales, but with the J/M backbone positions sister to Laurales or to Magnoliales plus Laurales have the same score. The more ambiguous position of Endressinia within Magnoliales and its weaker association with the order, relative to Archaeanthus, may be a result of the smaller proportion of characters preserved (20.4% vs. 28.9%).

These results indicate that the crown group of Magnoliales existed by the late Aptian, specifically increasing the minimum age of the node connecting Magnoliaceae with Degeneria, Galbulimima, Eupomatia, and Annonaceae to ca. 115 MY (Gradstein et al., 2004), and they also have implications for the early ecological history of angiosperms. Based on an analysis of the ecophysiology of living members of the ANITA grade and Chloranthaceae in a phylogenetic framework, Feild et al. (2004) concluded that the first angiosperms were physiologically restricted to dark and disturbed habitats in the wet forest understory. More recently Feild et al. (2009) argued that this “xerophobic” ecology persisted into the magnoliid clade. However, the small size of the leaves of Endressinia (ca. 4 cm) compared to most living Magnoliales and the abundant sedimentological and paleobotanical evidence for a semiarid climate in the Aptian–Albian of Brazil and equatorial Africa (Brenner, 1976; Doyle et al., 1982; Doyle, 1999; Mohr & Bernardes-de-Oliveira, 2004) suggest that at least some members of the currently wet-adapted clade Magnoliales had adapted to drier conditions by the late Aptian.

2.1.3 Mauldinia (=Prisca?) Retallack & Dilcher (1981) first described probable remains of this group from near the Albian–Cenomanian boundary in Kansas, as Prisca, which they interpreted as an elongate floral axis bearing numerous follicles with orthotropous ovules. However, Drinnan et al. (1990) found similar but better-preserved axes in the basal Cenomanian at Mauldin Mountain in Maryland, with well-preserved bisexual flowers borne on bilobed cladodes, interpreted as cymes with several fused bracts, which they called Mauldinia and assigned to the Lauraceae (a relationship accepted by Crepet et al., 2004). If Prisca and Mauldinia are related, this implies that the supposed “floral axes” of Prisca were inflorescences, the “follicles” were cladodes, and the “ovules” were unicarpellate fruits (Drinnan et al., 1990). The flowers have a typical lauraceous organization: two whorls of three tepals, stamens in trimerous whorls with two basal glands and flap dehiscence, and one uniovulate carpel. However, they differ from most living Lauraceae in having inner tepals that are larger and more petaloid than the outer ones, and endosperm in the mature seed, which is used up in Lauraceae. Herendeen (1991) convincingly associated fossil wood from this locality with the inflorescences. If Prisca and Mauldinia represent one genus, its name should be Prisca (Retallack & Dilcher, 1981), but until their equivalence is established we will use the name Mauldinia, which is based on better-preserved remains.

It might be expected that Mauldinia would associate with Lauraceae alone, but our analyses (Fig. 4) indicate that it is instead the sister group of both Lauraceae and Hernandiaceae, which also have one carpel but are more apomorphic in many characters. Unequivocal synapomorphies of the three taxa are solitary vessels (a reversal from the grouped vessels of more basal Laurales), inflorescences with lateral cymes (thyrses or thyrsoids), whorled tepals, whorled stamens, and one carpel. Features of Mauldinia that favor a position below both Lauraceae and Hernandiaceae are lack of well-developed paratracheal parenchyma in the wood, perianth of both sepaloid and petaloid tepals (as in more basal Laurales), superior ovary, and endosperm in the seed. Other apomorphies of Mauldinia arose lower in Laurales: long, narrow stamen filaments and pendent ovule below Monimiaceae; introrse anthers below the “atherosperm” clade consisting of Siparunaceae, Atherospermataceae, and Gomortega; and hypanthium in the ancestor of Laurales as a whole. Other features are also derived within Laurales, but their point of origin is equivocal: basal stamen glands (present in Atherospermataceae and Gomortega but not in Siparunaceae); flap anther dehiscence (present in the atherosperm clade but not in Monimiaceae); one ovule per carpel (all Laurales except Calycanthaceae, but the ancestral state for the order is equivocal); and woody endocarp (Romanov et al., 2007; scored as unknown in the achenes of Calycanthaceae). The next most parsimonious positions for Mauldinia are four steps worse; these include not only a sister group relationship to Lauraceae but also to the clade consisting of Monimiaceae, Lauraceae, and Hernandiaceae, to Hernandiaceae, and to Gomortega.

These results imply that Lauraceae and Hernandiaceae were derived from a common ancestor that had typically lauraceous flowers, except for the petaloid inner tepals. This is not implausible, because the floral features by which Hernandiaceae consistently differ from Lauraceae (polymerous androecium, one or two whorls of stamens) are inferred to be more apomorphic. The sessile character of the flowers in Mauldinia is an autapomorphy, suggesting it is not directly ancestral to the two living families. If these conclusions are correct, Mauldinia could not be assigned to Lauraceae in a cladistic classification and would instead have to be placed in an extinct family. Such a family, Priscaceae, was proposed for Prisca by Retallack & Dilcher (1981), although for different reasons.

This conclusion should be viewed with caution, because most of the features that place Mauldinia below Lauraceae and Hernandiaceae occur in some Lauraceae. Most obviously, most Lauraceae resemble Mauldinia in having a superior ovary. We scored Lauraceae as inferior because Rohwer & Rudolph (2005) reconstructed this as the ancestral state in the family, being found in the basal genus Hypodaphnis and members of the next most basal “Cryptocarya group”; they inferred that reversals to a superior ovary occurred within the Cryptocarya group (Beilschmiedia and derivatives) and in “higher” Lauraceae. Similarly, Actinodaphne, Caryodaphnopsis, Dahlgrenodendron, and Licaria have scanty axial parenchyma (Herendeen, 1991), and the inner tepals are larger than the outer ones in Anaueria, Caryodaphnopsis, Cassytha, Nothaphoebe, Williamodendron, and some species of Alseodaphne, Aniba, Dehaasia, and Persea (Drinnan et al., 1990; Rohwer, 1993). Drinnan et al. (1990) also stated that Cassytha has endosperm in the seed, but this appears to be based on an older misinterpretation (Rohwer, 1993). These taxa are relatively nested in the phylogeny of Rohwer & Rudolph (2005), implying that their seemingly primitive conditions are reversals. However, with the interesting exception of Caryodaphnopsis, the different reversals do not occur in the same taxa, so they are unlikely to be evidence that Mauldinia is nested within the crown group. This question should be addressed by future analyses with expanded taxon sampling of Recent Lauraceae.

Since 1990, Mauldinia-type inflorescences and isolated flowers have been reported from Albian and Upper Cretaceous rocks of many geographic areas (Herendeen et al., 1994; Eklund & Kvaček, 1998; Eklund, 2000; Frumin et al., 2004; von Balthazar et al., 2007; Viehofen et al., 2008), indicating that the Maryland species was just one representative of a major radiation at this time. Future resolution of the relationships of these fossils to living clades in higher Laurales should elucidate the evolution of this ecologically important group through the Late Cretaceous.

Pollen is notably absent in the stamens of Mauldinia, but this is not surprising, because Lauraceae and Hernandiaceae have a highly reduced, delicate exine, consisting of a thin granular layer and a tectum reduced to spines, which is rarely fossilized (Muller, 1981; Hesse & Kubitzki, 1983; an exception was reported by Herendeen et al., 1994). The fact that Lauraceae are abundant in the Tertiary leaf record but essentially invisible in the pollen record has long been noted (cf. Drinnan et al., 1990).

These analyses assume that Lauraceae and Hernandiaceae are closest extant relatives, but this is a subject of conflict between morphological and molecular data. Based on morphological data, this relationship is one of the strongest among basal angiosperms, with 100% bootstrap support in the morphological analysis of Doyle & Endress (2000). With the present dataset the two families have 17 unequivocal synapomorphies, and the alternative topologies are 15–16 steps worse. In contrast, molecular analyses agree that Monimiaceae, Lauraceae, and Hernandiaceae form a clade, but the sister group of Lauraceae is usually either Monimiaceae or Monimiaceae plus Hernandiaceae (Doyle & Endress, 2000; Renner & Chanderbali, 2000). However, the bootstrap support for both molecular arrangements is weak, and in our combined analysis (Doyle & Endress, 2000) the morphological data predominated, with 100% bootstrap support for the Lauraceae-Hernandiaceae clade. Whether this conclusion holds up in combined analyses with more molecular data remains to be seen. If the present dataset is analyzed using a backbone tree in which Monimiaceae are sister to Lauraceae, the most parsimonious position of Mauldinia is sister to the whole clade consisting of Lauraceae, Monimiaceae, and Hernandiaceae; a sister group relationship with Lauraceae is two steps worse, as are the three other positions relative to Lauraceae, Monimiaceae, and Hernandiaceae.

It is instructive to consider the implications of Mauldinia in connection with Virginianthus (Friis et al., 1994a), a middle Albian flower with numerous spiral tepals, stamens, and inner staminodes on the rim of a deep hypanthium and carpels inside, as in Calycanthaceae, the sister group of all other extant Laurales. Friis et al. (1994a) assigned this fossil to Calycanthaceae, but they noted it was more plesiomorphic than the living family in having monosulcate rather than disulculate pollen. The analysis of Doyle et al. (2008a) found two most parsimonious positions for Virginianthus, as a stem relative of either Calycanthaceae (including Idiospermum) or the remaining Laurales; the same result is found with the modifications in the present dataset. Virginianthus thus provides a minimum age of middle Albian for crown group Laurales. Together Virginianthus and Mauldinia indicate that considerable diversification had occurred in Laurales by the end of the Albian, and that flowers in the order had evolved from the inferred ancestral type with numerous spiral tepals, stamens, and carpels to a “lauraceous” type with tepals and stamens in trimerous whorls, stamens with basal glands and flap dehiscence, and only one carpel.

2.1.4 Walkeripollis  Most Early Cretaceous dispersed angiosperm pollen types, especially the monosulcates, have such widespread characters that they cannot be placed phylogenetically, but a few have distinctive advances that suggest relationships with particular modern clades. One consists of boat-shaped reticulate monosulcate pollen with finer sculpture at the ends of the grain, corresponding to Liliacidites in a strict sense, which phylogenetic analysis confirms are most likely monocots (Doyle et al., 2008a). Another consists of permanent tetrads of monoporate (ulcerate) grains, including a reticulate type described by Walker et al. (1983) from the Aptian-Albian of Israel and assigned to the vesselless magnoliid family Winteraceae, which also have reticulate, monoporate tetrads, and somewhat older grains with smaller tectal perforations from near the Barremian–Aptian boundary in Gabon, named Walkeripollis by Doyle et al. (1990a, 1990b).

In our analysis (Fig. 5), the most parsimonious position of the older and less clearly winteraceous species, W. gabonensis, is sister to Winteraceae. However, other positions (12 with the D&E backbone, two with J/M) are only one step worse—among more basal lines with round, reticulate monosulcate single grains, and (with the D&E backbone) with Lactoris in the Piperales, which has porate tetrads that differ from W. gabonensis in having a continuous tectum, intermediate (rather than columellar) infratectum, no supratectal spinules, no sculpture on the aperture membrane, and no endexine. This rapid “decay” of the link with Winteraceae must be a function of the small number of characters (10.6% of our dataset), and it highlights how useful it would be to find Walkeripollis in a fossil flower. However, this analysis may underestimate the evidence for a relationship with Winteraceae, as Walkeripollis and living Winteraceae are also similar in that the sculpture on the pore forms a ring around a central thin area. We did not include this character in our dataset because it would only be readily applicable in the two taxa in the dataset with a round pore (Lactoris, Winteraceae) and would therefore be uninformative.

Although these results support the relationship of Walkeripollis with Winteraceae proposed by Walker et al. (1983) and Doyle et al. (1990a, 1990b), other aspects of a phylogenetic analysis presented by Doyle et al. (1990b) are obsolete. This study, which included the two species of Walkeripollis, two other fossil pollen types (Afropollis, Schrankipollis), Winteraceae, and “Illiciales” (Illicium, Schisandraceae), concluded that Illiciales were nested within the clade including Walkeripollis and Winteraceae, which led to the suggestion that the thin area at the proximal pole in Schisandraceae is a vestige of the proximal thinning in a tetrad ancestor. This scheme is now contradicted by overwhelming molecular evidence that the two living taxa are widely separated, with Winteraceae in Canellales in the magnoliid clade, and Illicium and Schisandraceae in Austrobaileyales in the basal ANITA grade.

Both Liliacidites and Walkeripollis show that although pollen grains have only a few characters, these may be enough to suggest where they belong in the phylogeny of living angiosperms, and that dispersed pollen can still help to fill out the picture of early angiosperm diversity based on flowers. Walkeripollis is also significant in implying that Canellales, and therefore the crown magnoliid clade, had begun to diversify by the latest Barremian.

2.2 Eudicots

Probably the most important case in which a fossil pollen type can be unambiguously related to a major angiosperm clade concerns the eudicots (Eudicotyledoneae of Cantino et al., 2007), which include some 95% of the species formerly called dicots. The clearest synapomorphy of this clade is tricolpate pollen, which occurs only in eudicots and is inferred to be ancestral there (cf. Doyle, 2005). It appears near the end of the Barremian in Africa (Doyle et al., 1977, 1982; Doyle, 1992) but is rare in Laurasia until the early Albian (Kemp, 1968; Hochuli et al., 2006; Heimhofer et al., 2007); an exception is a report from the late Barremian of England (Hughes & McDougall, 1990). Some Albian eudicots are the first abundant angiosperms in the leaf record. The fossils treated here are from the middle Albian to earliest Cenomanian interval of the Potomac Group, when tricolpate pollen was becoming common.

2.2.1 Nelumbites  One of the most distinctive angiosperm groups in Albian leaf floras, from the Potomac Group to Kazakhstan and Siberia (Vakhrameev, 1952; Samylina, 1968), consists of peltate leaves with palmate venation, variously described as Nelumbites or Menispermites (for justification of the name Nelumbites, see Upchurch et al., 1994). These leaves usually occur in fine-grained pond deposits, which together with their local abundance and variation between flat and funnel-shaped, as in floating and emergent leaves of Nelumbo, led Doyle & Hickey (1976) and Hickey & Doyle (1977) to argue that they were aquatics ecologically similar and possibly related to Nelumbo. They differ from leaves of Nelumbo in their smaller size and more irregular venation. Nelumbo was formerly assigned to or linked with Nymphaeales (e.g., Takhtajan, 1980; Cronquist, 1981), but it has tricolpate pollen, suggesting it is a eudicot, and this has been confirmed by molecular data, which place it in the near-basal eudicot order Proteales. At the Quantico locality in Virginia, Upchurch et al. (1994) associated Nelumbites leaves with casts of pitted floral receptacles that resemble those of Nelumbo but are round rather than flat on top, and with detached tepals and rhizomes. Presumably conspecific leaves were described earlier from classic Virginia localities such as Mount Vernon and White House Bluff (Ward, 1895; Berry, 1911).

Our analysis (Fig. 6) confirms that the Quantico Nelumbites plant is most closely related to Nelumbo, based on peltate leaves and carpels sunken in pits. Its palmate venation supports a position in the eudicots, while more than one whorl or series of carpels is a synapomorphy of Proteales (reduced to one carpel in Proteaceae)—the inferred ancestral condition in eudicots is one whorl of 2–5 carpels. The absence of chloranthoid teeth (an inferred loss) is another synapomorphy of Proteales with the D&E backbone tree, but this is equivocal with the J/M backbone. However, it is only one step less parsimonious to associate Nelumbites with Brasenia (Cabombaceae, Nymphaeales), which also has peltate leaves and more than one whorl of carpels. Four other positions in Proteales and Nymphaeales are two steps worse. As with Walkeripollis, this instability is presumably a function of the small number of characters preserved (13.4%). More data on floral morphology or pollen of the fossils could test these conclusions. In the context of other Proteales, Nelumbites indicates that there were trends toward larger leaves, more regular venation, and the unique flat-topped receptacle in the Nelumbo line, as inferred by Upchurch et al. (1994).

2.2.2 West Brothers platanoid, Sapindopsis  Probably the most common angiosperm megafossils in the Albian are palmately lobed leaves with palmate venation, termed platanoids because of their similarities to Platanus, including palinactinodromous major venation (where lateral primary veins diverge in close succession rather than from one point) and (in later Albian forms) percurrent tertiary venation. Such leaves appeared in the middle Albian but became particularly widespread in the late Albian and early Cenomanian throughout Laurasia, from Kansas (Lesquereux, 1892) to Kazakhstan (Vakhrameev, 1952), especially in sandy stream levee facies, suggesting a riparian habitat like that of Platanus today (Doyle & Hickey, 1976; Hickey & Doyle, 1977; Crabtree, 1987). Cuticle studies (Upchurch, 1984) revealed additional evidence for a relationship to Platanus. The existence of plants related to Platanus was also suggested by the occurrence in the same beds of axes bearing small, round heads (Hickey & Doyle, 1977) and confirmed by discoveries that such heads consisted of unisexual flowers, as in Platanus, but with larger tepals and stamens and carpels in fives (Crane et al., 1986; Friis et al., 1988; Pedersen et al., 1994). Of these we consider the species from the late Albian West Brothers locality (Crane et al., 1986; Friis et al., 1988). Unfortunately, these remains have not been associated with a particular leaf type; the West Brothers bed contains both lobate platanoid leaves and members of the Sapindopsis group (Upchurch, 1984; Friis et al., 1988).

Sapindopsis is first represented in the middle Albian (Brooke locality) by leaves that appear at first sight to be pinnately compound but are actually pinnatifid, with the blade of the “leaflets” running down onto the rachis, and then in the late Albian by truly compound leaves. Hickey & Doyle (1977) suggested that Sapindopsis was related to the platanoids based on leaf architectural similarities and the occurrence of platanoid-like heads in the same beds. Subsequently, Upchurch (1984) showed that the cuticle structure of both pinnatifid and compound Sapindopsis leaves is similar to that of Potomac platanoids. Because pinnately compound leaves are most common in the subclass Rosidae (Hickey & Wolfe, 1975; Doyle, 2007), Hickey & Doyle (1977) suggested that Sapindopsis might represent more primitive relatives of Rosidae (cf. Wolfe et al., 1975), and that its similarities to the platanoids therefore supported a relationship between Rosidae and “Hamamelidae” (to which Platanus was then assigned).

This situation became clearer when Crane et al. (1993) conclusively associated pinnatifid Sapindopsis leaves at Brooke with heads and showed that these heads consisted of unisexual flowers similar to those of the platanoids, except for longer stamen filaments and a non-peltate connective apex. This strengthened the view that Sapindopsis was related to Platanus, but it cast doubt on its role as a “link” between Platanus and Rosidae, whose common ancestor would not be expected to have had such apomorphic features as heads and unisexual flowers. This reasoning has been amply confirmed by molecular evidence for the distant separation of Platanus (in the basal eudicot order Proteales, with Nelumbo and Proteaceae) and Rosidae (nested within the core eudicots, or Pentapetalae of Cantino et al., 2007). A relationship of both Sapindopsis and the West Brothers platanoid to Platanaceae was accepted by Crepet et al. (2004).

Our analyses strongly link both the late Albian platanoid from West Brothers and the middle Albian Sapindopsis species from Brooke with Recent Platanus, whether they are included in the analysis individually or together (Fig. 7). Their next most parsimonious positions are five or six steps worse with both backbone trees (five for Sapindopsis and the two fossils together, six for the West Brothers platanoid). The same unequivocal synapomorphies supporting these relationships are found with both backbone trees, except that with the D&E backbone loss of chloranthoid teeth is a synapomorphy of Proteales as a whole, whereas the course of evolution of this character is equivocal with the J/M backbone.

Synapomorphies uniting the West Brothers platanoid and Platanus are heads, sessile flowers, unisexual flowers, polymerous androecium (pentamerous in the West Brothers platanoid, unknown in Sapindopsis, and varying between trimerous and tetramerous in Platanus), one stamen whorl, short stamen base (i.e., reduced filament), and H-valvate anther dehiscence. Peltate connective apex may be an additional synapomorphy, but this is uncertain because states differ in Proteaceae (truncate apex) and Nelumbo (extended apex) and the sequence of character change is equivocal. Synapomorphies preserved in the West Brothers platanoid that link it with both Platanus and Proteaceae are thick nexine and orthotropous ovule.

Sapindopsis alone is linked with Platanus by heads, sessile flowers, unisexual flowers, polymerous perianth (unknown in the West Brothers platanoid, varying between trimerous and tetramerous in Platanus), and H-valvate anther dehiscence. It is united with both Platanus and Proteaceae by a shift from ovate to obovate-elliptical leaf shape (leaves are unknown in the West Brothers platanoid), as well as thick nexine and orthotropous ovule.

When both fossils are added to the analysis, they have two most parsimonious arrangements: either Sapindopsis is basal and the West Brothers platanoid is sister to Platanus, or the two fossils form a clade (the third arrangement, with Sapindopsis sister to Platanus, is two steps worse). In the first case, the two fossils and Platanus are united by the same five synapomorphies that are shared by Sapindopsis and Platanus alone, whereas the West Brothers platanoid is linked with Platanus by reduced filament and peltate connective apex. In the second case, the three taxa are linked by the same five synapomorphies plus polymerous androecium and one whorl of stamens, and the two fossils are united by small pollen size and sessile stigma (both homoplastic characters with the first arrangement). Reduced filament and peltate connective apex are either separate advances in the West Brothers platanoid and Platanus or synapomorphies that were reversed in Sapindopsis. In both cases all three taxa are linked with Proteaceae by obovate-elliptical leaves, thick nexine, and orthotropous ovule. Both arrangements imply that there was an intriguing shift from two dimerous whorls of stamens on the stem lineage of Proteales (and possibly of eudicots as a whole: cf. Drinnan et al., 1994; Endress & Doyle, 2009) to one whorl of five stamens (and carpels) in Albian members of the platanaceous line, followed by reduction to three or four stamens in living Platanus. This is a scenario that might not be suspected without fossil evidence. Scenarios for perianth evolution are similar but more ambiguous, partly because of uncertainty concerning perianth organization in the fossils.

These results further strengthen the well-established view that the platanaceous line was far more diverse in the Cretaceous than it is today. The diversity of fossil stem relatives on the line leading to Platanus is an argument for retaining the family Platanaceae rather than sinking Platanus into Proteaceae, as proposed by APG II (2003). In terms of phylogenetic nomenclature (Cantino et al., 2007), it might be useful to define Platanaceae as an apomorphy-based clade with heads (or perhaps unisexual flowers) homologous with those of living Platanus. It is intriguing that such highly divergent early lines as Nelumbites, the platanoids, Sapindopsis, and Proteaceae (represented by pollen from the Cenomanian onward: Ward & Doyle, 1994; Sauquet et al., 2009) all belonged to the near-basal eudicot order Proteales, and that they made up such a conspicuous portion of the Albian angiosperm flora. Hopefully future studies will reveal further links among these groups and clarify the transformations that led to their present morphologies.

It should be noted that not all palmately lobed “platanoid” leaves in the mid-Cretaceous are necessarily related to Platanus. Using cuticle and venation features, Upchurch & Dilcher (1990) showed that some such leaves from the Dakota Formation (late Albian or early Cenomanian), described as Pabiania, more likely represent extinct members of the Laurales. Similar leaves occur at the Quantico Nelumbites locality (Upchurch et al., 1994).

2.2.3 Spanomera  This taxon is based on floral remains of two very similar species from the late Albian and early Cenomanian of Maryland (West Brothers and Mauldin Mountain) that Drinnan et al. (1991) compared with Buxaceae, a family that molecular data firmly place among basal eudicots. Spanomera had unisexual flowers with a dimerous or derived plan, borne in bisexual inflorescences with decussately arranged male flowers and a terminal female flower, as in Buxus, and two basally fused carpels with a long ventral slit (carpels in extant Buxaceae vary between two and three, and they are more fused: von Balthazar & Endress, 2002a, 2002b). They produced tricolpate pollen with distinctive reticulate-striate to rugulate sculpture (i.e., a tendency for parallel alignment or a contorted, “vermiculate” pattern of the muri that make up the tectum), known in the dispersed state as Striatopollis paraneus and Striatopollis vermimurus. Some tendency to striate sculpture occurs in the usually reticulate pollen of Buxus and has been considered plesiomorphic there (Köhler & Brückner, 1989), and other Buxaceae are more apomorphic in having polyforate pollen, usually with “crotonoid” sculpture (Gray & Sohma, 1964), so we scored the family as uncertain (0/1) for the striate character.

Our analysis (Fig. 8) links Spanomera with Buxaceae, based on unisexual flowers and introrse anthers. Synapomorphies that unite both Spanomera and Buxaceae with Trochodendraceae (Trochodendron, Tetracentron) are botryoids (monotelic inflorescences), striate muri, eusyncarpy, and ventral fruit dehiscence. The two next best positions of Spanomera, which are two steps less parsimonious, are sister to Trochodendraceae and to both Buxaceae and Trochodendraceae (with both backbone trees). These relationships imply that the long, sessile stigmatic crest of Spanomera is derived rather than primitive (as was suggested by Drinnan et al., 1991), as is the short stamen filament. As noted above, dimerous tepals and two whorls of dimerous stamens may be ancestral for eudicots as a whole. The extended stamen connective of Spanomera and Buxaceae may be either ancestral or derived within eudicots.

The clade represented by Trochodendraceae, Buxaceae, and Spanomera probably also includes the core eudicots (Gunneridae, including Gunnerales and Pentapetalae of Cantino et al., 2007). This would imply that botryoids, reticulate-striate pollen, and at least basal eusyncarpy may have been basic features for core eudicots, which were variously modified or reversed within the clade, or else that they were parallelisms rather than synapomorphies in Trochodendraceae and the line leading to Buxaceae.

Striate tricolpate pollen of the Spanomera type was a common element in Albian pollen floras, and similar but more fragmentary flowers with striate pollen (Pedersen et al., 2007) have been reported in mesofossil floras from Portugal (probably early Albian: Heimhofer et al., 2007; see discussion in Doyle et al., 2008a). Similar pollen extends back to the early Aptian of Gabon (Doyle et al., 1977; Doyle, 1992, 1999) and Egypt (Penny, 1988), but its presence in Laurasia has not been confirmed before the early Albian (Hochuli et al., 2006; Heimhofer et al., 2007). No leaf fossils have been associated with this group, but optimization of characters in our dataset on the phylogeny predicts that the Spanomera plant would have ovate, palmately veined leaves with chloranthoid teeth, like “Populuspotomacensis in the Potomac Group, which has been compared with Trochodendraceae (Crabtree, 1987; Magallón et al., 1999). Buxus and Didymeles, the specialized Madagascan sister group of Buxaceae, have colpi with two or more ora, which suggests that Spanomera, with simple tricolpate apertures, was basal to Didymeles and Buxaceae, as does the lesser degree of carpel fusion than in Buxaceae (cf. Drinnan et al., 1991); Didymeles has only one carpel and would have been scored as unknown for carpel fusion (von Balthazar et al., 2003).

2.3 Summary

Like our previous studies (Doyle et al., 2008a; Endress & Doyle, 2009), these analyses show that it is possible to integrate Early Cretaceous fossils into the phylogeny of living angiosperms, even when they do not fit into any modern family, and that this can provide novel insights on the age, former diversity, biogeography, and morphological evolution of living clades. These results can be summarized with a representative most parsimonious tree found when all 13 taxa treated in this study, Doyle et al. (2008a), and Endress & Doyle (2009) are included in the analysis, using the D&E backbone tree (Fig. 9). This analysis produced 2856 most parsimonious trees, which is the product of the two most parsimonious arrangements for the West Brothers platanoid and Sapindopsis and the numbers of positions of all other fossil taxa added individually. In other words, addition of several fossils had no effect on the inferred positions of any individual fossil relative to the extant tree.


Figure 9. One of 2856 most parsimonious trees (1045 steps) obtained after addition of all eight taxa treated in this study plus Archaefructus, Anacostia, the Pennipollis plant, Virginianthus, and Liliacidites to the D&E tree (Endress & Doyle, 2009). Colors distinguish taxa that appear in the Barremian–early Albian and middle Albian–early Cenomanian intervals. Aust, Austrobaileyales; Ca, Canellales; Chlor, Chloranthaceae; Magnol, Magnoliales; Nymph, Nymphaeales; Piper, Piperales.

Download figure to PowerPoint

Figure 9 shows broad congruence between the Early Cretaceous fossil record of angiosperms and molecular trees based on extant plants, and it is therefore consistent with the concept that the angiosperm flora was diversifying through the Early Cretaceous (Doyle, 2001; Friis et al., 2006). Taxa known from the Barremian, Aptian, and early Albian part of the record (Archaefructus, Anacostia, the Pennipollis plant, Endressinia, Walkeripollis, Liliacidites) tend to be attached lower in the tree (an average of 6.17 nodes above the basal node) than those that are not known before the middle Albian to early Cenomanian interval (Archaeanthus, Virginianthus, Mauldinia, Nelumbites, Sapindopsis, the West Brothers platanoid, Spanomera: average 9.29 nodes above the base), all of which are nested within the magnoliid and eudicot clades. However, the existence near the Barremian–Aptian boundary of fossils representing several major clades (Archaefructus, possibly Nymphaeales; Pennipollis, the chloranthoid line; Walkeripollis, Magnoliidae, Canellales; Liliacidites, monocots; the first tricolpate pollen, eudicots) indicates that significant splitting of basal angiosperm lineages and morphological divergence had occurred by this time. We still have no clear mega- or mesofossil records from the earliest phases of the angiosperm radiation, which is represented in the Valanginian and Hauterivian by pollen records only (Trevisan, 1988; Hughes, 1994; Brenner, 1996; Friis et al., 2006). Future more comprehensive studies of this sort promise to provide additional new insights on the age of clades and the course of evolution in basal angiosperm groups that could not be gained from studies of living taxa alone.


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix

Acknowledgements  We wish to thank De-Yuan HONG and Yin-Long QIU for inviting JAD to present the symposium talk on which this paper is based, the Journal of Systematics and Evolution and Shenzhen Botanical Garden for travel expenses, the NSF Deep Time Research Collaboration Network (RCN0090283) for facilitating collaboration, David DILCHER and Paula RUDALL for use of unpublished and in press results, Else Marie FRIIS for providing drawings of fossils, and Pat HERENDEEN and Gary UPCHURCH for helpful comments on the manuscript.


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix
  • Antonov AS, Troitsky AV, Samigullin TK, Bobrova VK, Valiejo-Roman KM, Martin W. 2000. Early events in the evolution of angiosperms deduced from cp rDNA ITS 2–4 sequence comparisons. In: LiuYH, FanHM, ChenZY, WuQG, ZengQW eds. Proceedings of the International Symposium on the Family Magnoliaceae. Beijing : Science Press. 210214.
  • APG [Angiosperm Phylogeny Group] II. 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141: 399436.
  • Axelrod DI. 1960. The evolution of flowering plants. In: TaxS ed. The evolution of life. Chicago : University of Chicago Press. 227305.
  • Axelrod DI. 1970. Mesozoic paleogeography and early angiosperm history. Botanical Review 36: 277319.
  • Bailey IW, Swamy BGL. 1951. The conduplicate carpel of dicotyledons and its initial trends of specialization. American Journal of Botany 38: 373379.
  • Barkman TJ, Chenery G, McNeal JR, Lyons-Weiler J, Ellisens WJ, Moore G, Wolfe AD, DePamphilis CW. 2000. Independent and combined analyses of sequences from all three genomic compartments converge on the root of flowering plant phylogeny. Proceedings of the National Academy of Sciences USA 97: 1316613171.
  • Berry EW. 1911. Systematic paleontology, Lower Cretaceous, Pteridophyta-Dicotyledonae. In: ClarkWB ed. Lower Cretaceous. Baltimore : Maryland Geological Survey, Johns Hopkins Press. 214508.
  • Brenner GJ. 1963. The spores and pollen of the Potomac Group of Maryland. Maryland Department of Geology, Mines and Water Resources Bulletin 27: 1215.
  • Brenner GJ. 1976. Middle Cretaceous floral provinces and early migrations of angiosperms. In: BeckCB ed. Origin and early evolution of angiosperms. New York : Columbia University Press. 2347.
  • Brenner GJ. 1996. Evidence for the earliest stage of angiosperm pollen evolution: A paleoequatorial section from Israel. In: TaylorDW, HickeyLJ eds. Flowering plant origin, evolution & phylogeny. New York : Chapman & Hall. 91115.
  • Buxbaum F. 1922. Vergleichende Anatomie der Melanthioideae. Feddes Repertorium, Beihefte 29: 180.
  • Buxbaum F. 1927. Nachträge zur vergleichenden Anatomie der Melanthioideae. Beihefte zum Botanischen Centralblatt 44(1/2): 255263.
  • Cantino PD, Doyle JA, Graham SW, Judd WS, Olmstead RG, Soltis DE, Soltis PS, Donoghue MJ. 2007. Towards a phylogenetic nomenclature of Tracheophyta. Taxon 56: 822846.
  • Carpenter RJ, Hill RS, Jordan GJ. 2005. Leaf cuticular morphology links Platanaceae and Proteaceae. International Journal of Plant Sciences 166: 843855.
  • Cheadle VI. 1942. The occurrence and types of vessels in the various organs of the plant in the Monocotyledoneae. American Journal of Botany 29: 441450.
  • Cheadle VI, Kosakai H. 1971. Vessels in Liliaceae. Phytomorphology 21: 320333.
  • Chen L, Ren Y, Endress PK, Tian XH, Zhang XH. 2007. Floral organogenesis in Tetracentron sinense (Trochodendraceae) and its taxonomic significance. Plant Systematics and Evolution 264: 183193.
  • Coiffard C, Gomez B, Thevenard F. 2007. Early Cretaceous angiosperm invasion of Western Europe and major environmental changes. Annals of Botany 100: 545553.
  • Cooke DA. 1983. The seedling of Trithuria (Hydatellaceae). Victorian Naturalist 100: 6869.
  • Corner EJH. 1976. The seeds of the dicotyledons. Cambridge : Cambridge University Press.
  • Crabtree DR. 1987. Angiosperms of the Northern Rocky Mountains: Albian to Campanian (Cretaceous) megafossil floras. Annals of the Missouri Botanical Garden 74: 707747.
  • Crane PR, Friis EM, Pedersen KR. 1986. Lower Cretaceous angiosperm flowers: Fossil evidence on early radiation of dicotyledons. Science 232: 852854.
  • Crane PR, Friis EM, Pedersen KR. 1989. Reproductive structure and function in Cretaceous Chloranthaceae. Plant Systematics and Evolution 165: 211226.
  • Crane PR, Friis EM, Pedersen KR. 1995. The origin and early diversification of angiosperms. Nature 374: 2733.
  • Crane PR, Pedersen KR, Friis EM, Drinnan AN. 1993. Early Cretaceous (early to middle Albian) platanoid inflorescences associated with Sapindopsis leaves from the Potomac Group of eastern North America. Systematic Botany 18: 328344.
  • Crepet WL, Nixon KC, Gandolfo MA. 2004. Fossil evidence and phylogeny: The age of major angiosperm clades based on mesofossil and macrofossil evidence from Cretaceous deposits. American Journal of Botany 91: 16661682.
  • Cronquist A. 1981. An integrated system of classification of flowering plants. New York : Columbia University Press.
  • Cutler DF. 1969. Anatomy of the monocotyledons. IV. Juncales. Oxford : Clarendon Press.
  • Dilcher DL. 1979. Early angiosperm reproduction: An introductory report. Review of Palaeobotany and Palynology 27: 291328.
  • Dilcher DL, Crane PR. 1984. Archaeanthus: An early angiosperm from the Cenomanian of the Western Interior of North America. Annals of the Missouri Botanical Garden 71: 351383.
  • Dilcher DL, Crepet WL, Beeker CD, Reynolds HC. 1976. Reproductive and vegetative morphology of a Cretaceous angiosperm. Science 191: 854856.
  • Doweld AB. 1998. Carpology, seed anatomy and taxonomic relationships of Tetracentron (Tetracentraceae) and Trochodendron (Trochodendraceae). Annals of Botany 82: 413443.
  • Doweld AB. 2001. Carpology and phermatology of Gomortega (Gomortegaceae): Systematic and evolutionary implications. Acta Botanica Malacitana 26: 1937.
  • Doyle JA. 1969. Cretaceous angiosperm pollen of the Atlantic Coastal Plain and its evolutionary significance. Journal of the Arnold Arboretum 50: 135.
  • Doyle JA. 1973. Fossil evidence on early evolution of the monocotyledons. Quarterly Review of Biology 48: 399413.
  • Doyle JA. 1992. Revised palynological correlations of the lower Potomac Group (USA) and the Cocobeach sequence of Gabon (Barremian-Aptian). Cretaceous Research 13: 337349.
  • Doyle JA. 1999. The rise of angiosperms as seen in the African Cretaceous pollen record. Palaeoecology of Africa 26: 330.
  • Doyle JA. 2001. Significance of molecular phylogenetic analyses for paleobotanical investigations on the origin of angiosperms. The Palaeobotanist 50: 167188.
  • Doyle JA. 2005. Early evolution of angiosperm pollen as inferred from molecular and morphological phylogenetic analyses. Grana 44: 227251.
  • Doyle JA. 2007. Systematic value and evolution of leaf architecture across the angiosperms in light of molecular phylogenetic analyses. Courier Forschungsinstitut Senckenberg 258: 2137.
  • Doyle JA. 2008. Integrating molecular phylogenetic and paleobotanical evidence on origin of the flower. International Journal of Plant Sciences 169: 816843.
  • Doyle JA, Donoghue MJ. 1993. Phylogenies and angiosperm diversification. Paleobiology 19: 141167.
  • Doyle JA, Endress PK. 2000. Morphological phylogenetic analysis of basal angiosperms: Comparison and combination with molecular data. International Journal of Plant Sciences 161(Supplement): S121S153.
  • Doyle JA, Hickey LJ. 1976. Pollen and leaves from the mid-Cretaceous Potomac Group and their bearing on early angiosperm evolution. In: BeckCB ed. Origin and early evolution of angiosperms. New York : Columbia University Press. 139206.
  • Doyle JA, Hotton CL. 1991. Diversification of early angiosperm pollen in a cladistic context. In: BlackmoreS, BarnesSH eds. Pollen and spores: Patterns of diversification. Oxford : Clarendon Press. 169195.
  • Doyle JA, Le Thomas A. 1996. Phylogenetic analysis and character evolution in Annonaceae. Bulletin du Muséum National d’Histoire Naturelle, section B, Adansonia 18: 279334.
  • Doyle JA, Robbins EI. 1977. Angiosperm pollen zonation of the continental Cretaceous of the Atlantic Coastal Plain and its application to deep wells in the Salisbury Embayment. Palynology 1: 4378.
  • Doyle JA, Biens P, Doerenkamp A, Jardiné S. 1977. Angiosperm pollen from the pre-Albian Cretaceous of Equatorial Africa. Bulletin des Centres de Recherches Exploration-Production Elf-Aquitaine 1: 451473.
  • Doyle JA, Endress PK, Upchurch GR. 2008a. Early Cretaceous monocots: A phylogenetic evaluation. Acta Musei Nationalis Pragae, Series B, Historia Naturalis 64(2–4): 5987.
  • Doyle JA, Hotton CL, Ward JV. 1990a. Early Cretaceous tetrads, zonasulculate pollen, and Winteraceae. I. Taxonomy, morphology, and ultrastructure. American Journal of Botany 77: 15441557.
  • Doyle JA, Hotton CL, Ward JV. 1990b. Early Cretaceous tetrads, zonasulculate pollen, and Winteraceae. II. Cladistic analysis and implications. American Journal of Botany 77: 15581568.
  • Doyle JA, Jardiné S, Doerenkamp A. 1982. Afropollis, a new genus of early angiosperm pollen, with notes on the Cretaceous palynostratigraphy and paleoenvironments of Northern Gondwana. Bulletin des Centres de Recherches Exploration-Production Elf-Aquitaine 6: 39117.
  • Doyle JA, Manchester SR, Sauquet H. 2008b. A seed related to Myristicaceae in the Early Eocene of southern England. Systematic Botany 33: 636646.
  • Doyle JA, Sauquet H, Scharaschkin T, Le Thomas A. 2004. Phylogeny, molecular and fossil dating, and biogeographic history of Annonaceae and Myristicaceae (Magnoliales). International Journal of Plant Sciences 165(Supplement): S55S67.
  • Doyle JA, Van Campo M, Lugardon B. 1975. Observations on exine structure of Eucommiidites and Lower Cretaceous angiosperm pollen. Pollen et Spores 17: 429486.
  • Drinnan AN, Crane PR, Friis EM, Pedersen KR. 1990. Lauraceous flowers from the Potomac Group (mid-Cretaceous) of eastern North America. Botanical Gazette 151: 370384.
  • Drinnan AN, Crane PR, Friis EM, Pedersen KR. 1991. Angiosperm flowers and tricolpate pollen of buxaceous affinity from the Potomac Group (mid-Cretaceous) of eastern North America. American Journal of Botany 78: 153176.
  • Drinnan AN, Crane PR, Hoot SB. 1994. Patterns of floral evolution in the early diversification of non-magnoliid dicotyledons (eudicots). Plant Systematics and Evolution Supplement 8: 93122.
  • Duvall MR, Mathews S, Mohammad N, Russell T. 2006. Placing the monocots: Conflicting signal from trigenomic analyses. Aliso 22: 7990.
  • Duvall MR, Robinson JW, Mattson JG, Moore A. 2008. Phylogenetic analyses of two mitochondrial metabolic genes sampled in parallel from angiosperms find fundamental interlocus incongruence. American Journal of Botany 95: 871884.
  • Edwards JG. 1920. Flower and seed of Hedyosmum nutans. Botanical Gazette 70: 409424.
  • Eklund H. 2000. Lauraceous flowers from the Late Cretaceous of North Carolina, U.S.A. Botanical Journal of the Linnean Society 132: 397428.
  • Eklund H, Kvaček J. 1998. Lauraceous inflorescences and flowers from the Cenomanian of Bohemia (Czech Republic, central Europe). International Journal of Plant Sciences 159: 668686.
  • Eklund H, Doyle JA, Herendeen PS. 2004. Morphological phylogenetic analysis of living and fossil Chloranthaceae. International Journal of Plant Sciences 165: 107151.
  • Eklund H, Friis EM, Pedersen KR. 1997. Chloranthaceous floral structures from the Late Cretaceous of Sweden. Plant Systematics and Evolution 207: 1342.
  • Endress PK. 1972. Zur vergleichenden Entwicklungsmorphologie, Embryologie und Systematik bei Laurales. Botanische Jahrbücher für Systematik 92: 331428.
  • Endress PK. 1980. The reproductive structures and systematic position of the Austrobaileyaceae. Botanische Jahrbücher für Systematik 101: 393433.
  • Endress PK. 1984. The role of inner staminodes in the floral display of some relic Magnoliales. Plant Systematics and Evolution 146: 269282.
  • Endress PK. 1986. Floral structure, systematics, and phylogeny in Trochodendrales. Annals of the Missouri Botanical Garden 73: 297324.
  • Endress PK. 1987. The Chloranthaceae: Reproductive structures and phylogenetic position. Botanische Jahrbücher für Systematik 109: 153226.
  • Endress PK. 2001. The flowers in extant basal angiosperms and inferences on ancestral flowers. International Journal of Plant Sciences 162: 11111140.
  • Endress PK. 2005. Carpels in Brasenia (Cabombaceae) are completely ascidiate despite a long stigmatic crest. Annals of Botany 96: 209215.
  • Endress PK. 2006. Angiosperm floral evolution: Morphological and developmental framework. Advances in Botanical Research 44: 161.
  • Endress PK, Doyle JA. 2007. Floral phyllotaxis in basal angiosperms: Development and evolution. Current Opinion in Plant Biology 10: 5257.
  • Endress PK, Doyle JA. 2009. Reconstructing the ancestral angiosperm flower and its initial specializations. American Journal of Botany 96: 2266.
  • Endress PK, Igersheim A. 1997. Gynoecium diversity and systematics of the Laurales. Botanical Journal of the Linnean Society 125: 93168.
  • Endress PK, Igersheim A. 1999. Gynoecium diversity and systematics of the basal eudicots. Botanical Journal of the Linnean Society 130: 305393.
  • Endress PK, Igersheim A. 2000. Gynoecium structure and evolution in basal angiosperms. International Journal of Plant Sciences 161(Supplement): S211S223.
  • Endress PK, Lorence DH. 2004. Heterodichogamy of a novel type in Hernandia (Hernandiaceae) and its structural basis. International Journal of Plant Sciences 165: 753763.
  • Feild TS, Arens NC, Doyle JA, Dawson TE, Donoghue MJ. 2004. Dark and disturbed: A new image of early angiosperm ecology. Paleobiology 30: 82107.
  • Feild TS, Chatelet DS, Brodribb TJ. 2009. Ancestral xerophobia: A hypothesis on the whole plant ecophysiology of early angiosperms. Geobiology 7: 237264.
  • Fontaine WM. 1889. The Potomac or Younger Mesozoic flora. US Geological Survey Monograph 15, Washington , DC .
  • Foster AS. 1963. The morphology and relationships of Circaeaster. Journal of the Arnold Arboretum 44: 299321.
  • Friedman WE. 2008. Hydatellaceae are water lilies with gymnospermous tendencies. Nature 453: 9497.
  • Friis EM. 1983. Upper Cretaceous (Senonian) floral structures of juglandalean affinity containing Normapolles pollen. Review of Palaeobotany and Palynology 39: 161188.
  • Friis EM, Crane PR, Pedersen KR. 1986. Floral evidence for Cretaceous chloranthoid angiosperms. Nature 320: 163164.
  • Friis EM, Crane PR, Pedersen KR. 1988. Reproductive structures of Cretaceous Platanaceae. Biologiske Skrifter Danske Videnskabernes Selskab 31: 155.
  • Friis EM, Crane PR, Pedersen KR. 1997. Anacostia, a new basal angiosperm from the Early Cretaceous of North America and Portugal with trichotomocolpate/monocolpate pollen. Grana 36: 225244.
  • Friis EM, Doyle JA, Endress PK, Leng Q. 2003. Archaefructus—Angiosperm precursor or specialized early angiosperm? Trends in Plant Science 8: 369373.
  • Friis EM, Eklund H, Pedersen KR, Crane PR. 1994a. Virginianthus calycanthoides gen. et sp. nov. —A calycanthaceous flower from the Potomac Group (Early Cretaceous) of eastern North America. International Journal of Plant Sciences 155: 772785.
  • Friis EM, Pedersen KR, Crane PR. 1994b. Angiosperm floral structures from the Early Cretaceous of Portugal. Plant Systematics and Evolution Supplement 8: 3149.
  • Friis EM, Pedersen KR, Crane PR. 1995. Appomattoxia ancistrophora gen. et sp. nov., a new Early Cretaceous plant with similarities to Circaeaster and extant Magnoliidae. American Journal of Botany 82: 933943.
  • Friis EM, Pedersen KR, Crane PR. 2000. Fossil floral structures of a basal angiosperm with monocolpate, reticulate-acolumellate pollen from the Early Cretaceous of Portugal. Grana 39: 226239.
  • Friis EM, Pedersen KR, Crane PR. 2006. Cretaceous angiosperm flowers: Innovation and evolution in plant reproduction. Palaeogeography Palaeoclimatology Palaeoecology 232: 251293.
  • Frumin S, Eklund H, Friis EM. 2004. Mauldinia hirsuta sp. nov., a new member of the extinct genus Mauldinia (Lauraceae) from the Late Cretaceous (Cenomanian-Turonian) of Kazakhstan. International Journal of Plant Sciences 165: 883895.
  • Gandolfo MA, Nixon KC, Crepet WL. 2000. Monocotyledons: A review of their Early Cretaceous record. In: WilsonKL, MorrisonDA eds. Monocots: systematics and evolution. Collingwood , Australia : CSIRO Publishing. 4451.
  • Góczán F, Juhász M. 1984. Monosulcate pollen grains of angiosperms from Hungarian Albian sediments I. Acta Botanica Hungarica 30: 289319.
  • GradsteinFM, OggJG, SmithAG eds. 2004. A geologic time scale 2004. Cambridge : Cambridge University Press.
  • Graham SW, Olmstead RG. 2000. Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 87: 17121730.
  • Gray J, Sohma K. 1964. Fossil Pachysandra from western America with a comparative study of pollen in Pachysandra and Sarcococca. American Journal of Science 262: 11591197.
  • Gröcke DR, Ludvigson GA, Witzke BL, Robinson SA, Joeckel RM, Ufnar DF, Ravn RL. 2006. Recognizing the Albian-Cenomanian (OAE1d) sequence boundary using plant carbon isotopes: Dakota Formation, Western Interior Basin, USA. Geology 34: 193196.
  • Hamann U. 1975. Neue Untersuchungen zur Embryologie und Systematik der Centrolepidaceae. Botanische Jahrbücher für Systematik 96: 154191.
  • Hamann U, Kaplan K, Rübsamen T. 1979. Über die Samenschalenstruktur der Hydatellaceae (Monocotyledoneae) und die systematische Stellung von Hydatella filamentosa. Botanische Jahrbücher für Systematik 100: 555563.
  • Heimhofer U, Hochuli PA, Burla S, Weissert H. 2007. New records of Early Cretaceous angiosperm pollen from Portuguese coastal deposits: Implications for the timing of the early angiosperm radiation. Review of Palaeobotany and Palynology 144: 3976.
  • Herendeen PS. 1991. Lauraceous wood from the mid-Cretaceous Potomac group of eastern North America: Paraphyllanthoxylon marylandense sp. nov. Review of Palaeobotany and Palynology 69: 277290.
  • Herendeen PS, Crepet WL, Nixon KC. 1994. Fossil flowers and pollen of Lauraceae from the Upper Cretaceous of New Jersey. Plant Systematics and Evolution 189: 2940.
  • Hesse M, Kubitzki K. 1983. The sporoderm ultrastructure in Persea, Nectandra, Hernandia, Gomortega and some other lauralean genera. Plant Systematics and Evolution 141: 299311.
  • Hickey LJ, Doyle JA. 1977. Early Cretaceous fossil evidence for angiosperm evolution. Botanical Review 43: 1104.
  • Hickey LJ, Wolfe JA. 1975. The bases of angiosperm phylogeny: Vegetative morphology. Annals of the Missouri Botanical Garden 62: 538589.
  • Hochuli PA, Heimhofer U, Weissert H. 2006. Timing of early angiosperm radiation: Recalibrating the classical succession. Journal of the Geological Society, London 163: 587594.
  • Hoot SB, Zautke H, Harris DJ, Crane PR, Neves SS. 2009. Phylogenetic patterns in Menispermaceae based on multiple chloroplast sequence data. Systematic Botany 34: 4456.
  • Hufford L. 1996. Ontogenetic evolution, clade diversification, and homoplasy. In: SandersonMJ, HuffordL eds. Homoplasy. San Diego : Academic Press. 271301.
  • Hughes NF. 1961. Fossil evidence and angiosperm ancestry. Science Progress 49: 84102.
  • Hughes NF. 1994. The enigma of angiosperm origins. Cambridge : Cambridge University Press.
  • Hughes NF, McDougall AB. 1990. Barremian-Aptian angiospermid pollen records from southern England. Review of Palaeobotany and Palynology 65: 145151.
  • Igersheim A, Endress PK. 1997. Gynoecium diversity and systematics of the Magnoliales and winteroids. Botanical Journal of the Linnean Society 124: 213271.
  • Igersheim A, Endress PK. 1998. Gynoecium diversity and systematics of the paleoherbs. Botanical Journal of the Linnean Society 127: 289370.
  • Igersheim A, Buzgo M, Endress PK. 2001. Gynoecium diversity and systematics in basal monocots. Botanical Journal of the Linnean Society 136: 165.
  • Ito M. 1986. Studies in the floral morphology and anatomy of Nymphaeales III. Floral anatomy of Brasenia schreberi Gmel. and Cabomba caroliniana A. Gray. Botanical Magazine Tokyo 99: 169184.
  • Ito M. 1987. Phylogenetic systematics of the Nymphaeales. Botanical Magazine Tokyo 100: 1735.
  • Jansen RK, Cai Z, Raubeson LA, Daniell H, DePamphilis CW, Leebens-Mack J, Müller KF, Guisinger-Bellian M, Haberle RC, Hansen AK, Chumley TW, Lee SB, Peery R, McNeal JR, Kuehl JV, Boore JL. 2007. Analysis of 81 genes from 64 plastid genomes resolves relationships in angiosperms and identifies genome-scale evolutionary patterns. Proceedings of the National Academy of Sciences USA 104: 1936919374.
  • Jones EN. 1931. The morphology and biology of Ceratophyllum demersum. University of Iowa Studies in Natural History 13: 1155.
  • Judd WS, Campbell CS, Kellogg EA, Stevens PF, Donoghue MJ. 2008. Plant systematics: A phylogenetic approach. 3rd ed. Sunderland , Massachusetts : Sinauer Associates.
  • Keating RC. 2000. Anatomy of the young vegetative shoot of Takhtajania perrieri (Winteraceae). Annals of the Missouri Botanical Garden 87: 335346.
  • Keating RC. 2002. Anatomy of the monocotyledons IX. Acoraceae and Araceae. Oxford : Clarendon Press.
  • Kemp EM. 1968. Probable angiosperm pollen from British Barremian to Albian strata. Palaeontology 11: 421434.
  • Kessler PJA. 1993. Menispermaceae. In: KubitzkiK, RohwerJG, BittrichV eds. The families and genera of vascular plants, vol. II. Flowering plants. Dicotyledons. Magnoliid, hamamelid and caryophyllid families. Berlin : Springer-Verlag. 402418.
  • Kimoto Y, Tobe H. 2008. Embryology of Illigera and Sparattanthelium (Hernandiaceae, Laurales): family characteristics and relationships. International Journal of Plant Sciences 169: 391408.
  • Köhler E, Brückner P. 1989. The genus Buxus (Buxaceae): Aspects of its differentiation in space and time. Plant Systematics and Evolution 162: 267283.
  • Kong HZ. 2001. Comparative morphology of leaf epidermis in the Chloranthaceae. Botanical Journal of the Linnean Society 136: 279294.
  • Leins P, Erbar C. 1995. Das frühe Differenzierungsmuster in den Blüten von Saruma henryi Oliv. (Aristolochiaceae). Botanische Jahrbücher für Systematik 117: 365376.
  • Lesquereux L. 1892. The flora of the Dakota Group. US Geological Survey Monograph 17, Washington , DC .
  • Li J, Ledger J, Ward T, Del Tredici P. 2004. Phylogenetics of Calycanthaceae based on molecular and morphological data with a special reference to divergent paralogues of the nrDNA ITS region. Harvard Papers in Botany 9: 6982.
  • Liang HX, Tucker SC. 1995. Floral ontogeny of Zippelia begoniaefolia and its familial affinity: Saururaceae or Piperaceae? American Journal of Botany 82: 681689.
  • Lorence DH. 1985. A monograph of the Monimiaceae (Laurales) in the Malagasy region (southwest Indian Ocean). Annals of the Missori Botanical Garden 72: 1165.
  • Maddison DR, Maddison WP. 2003. MacClade 4: Analysis of phylogeny and character evolution, version 4.06. Sunderland , Massachusetts : Sinauer Associates.
  • Magallón S, Sanderson MJ. 2001. Absolute diversification rates in angiosperm clades. Evolution 55: 17621780.
  • Magallón S, Herendeen PS, Crane PR. 1997. Quadriplatanus georgianus gen. et sp. nov.: staminate and pistillate platanaceous flowers from the Late Cretaceous (Coniacian-Santonian) of Georgia, USA. International Journal of Plant Sciences 158: 373394.
  • Magallón S, Crane PR, Herendeen PS. 1999. Phylogenetic pattern, diversity, and diversification of eudicots. Annals of the Missouri Botanical Garden 86: 297372.
  • Manos PS, Soltis PS, Soltis DE, Manchester SR, Oh SH, Bell CD, Dilcher DL, Stone DE. 2007. Phylogeny of extant and fossil Juglandaceae inferred from the integration of molecular and morphological data sets. Systematic Biology 56: 412430.
  • Mathews S, Donoghue MJ. 1999. The root of angiosperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947950.
  • Metcalfe CR. 1987. Anatomy of the dicotyledons, second edition, vol. III. Magnoliales, Illiciales, and Laurales (sensu Armen Takhtajan). Oxford : Clarendon Press.
  • Metcalfe CR, Chalk L. 1950. Anatomy of the dicotyledons. Oxford : Clarendon Press.
  • Mohr BAR, Bernardes-de-Oliveira MEC. 2004. Endressinia brasiliana, a magnolialean angiosperm from the Lower Cretaceous Crato Formation (Brazil). International Journal of Plant Sciences 165: 11211133.
  • Moore MJ, Bell CD, Soltis PS, Soltis DE. 2007. Using plastid genome-scale data to resolve enigmatic relationships among basal angiosperms. Proceedings of the National Academy of Sciences USA 1104: 1936319368.
  • Muller J. 1970. Palynological evidence on early differentiation of angiosperms. Biological Reviews of the Cambridge Philosophical Society 45: 417450.
  • Muller J. 1981. Fossil pollen records of extant angiosperms. Botanical Review 47: 1142.
  • Oganezova GG. 1984. Morfologo-anatomicheskie osobennosti ploda i semeni nekotorykh predstaviteley podsemeystva Melanthioideae s. str. (Liliaceae) v svyazi s ikh sistematikoy i filogeniey. Botanicheskiy Zhurnal 69: 772781.
  • Oh IC, Denk T, Friis EM. 2003. Evolution of Illicium (Illiciaceae): mapping morphological characters on the molecular tree. Plant Systematics and Evolution 240: 175209.
  • Parkinson CL, Adams KL, Palmer JD. 1999. Multigene analyses identify the three earliest lineages of extant flowering plants. Current Biology 9: 14851488.
  • Pedersen KR, Crane PR, Drinnan AN, Friis EM. 1991. Fruits from the mid-Cretaceous of North America with pollen grains of the Clavatipollenites type. Grana 30: 577590.
  • Pedersen KR, Friis EM, Crane PR, Drinnan AN. 1994. Reproductive structures of an extinct platanoid from the Early Cretaceous (latest Albian) of eastern North America. Review of Palaeobotany and Palynology 80: 291303.
  • Pedersen KR, Von Balthazar M, Crane PR, Friis EM. 2007. Early Cretaceous floral structures and in situ tricolpate-striate pollen: New early eudicots from Portugal. Grana 46: 176196.
  • Penny JHJ. 1988. Early Cretaceous striate tricolpate pollen from the Borehole Mersa Matruh 1, North West Desert, Egypt. Journal of Micropalaeontology 7: 201215.
  • Prychid CJ, Rudall PJ. 1999. Calcium oxalate crystals in monocotyledons: A review of their structure and systematics. Annals of Botany 84: 725739.
  • Qiu YL, Lee J, Bernasconi-Quadroni F, Soltis DE, Soltis PS, Zanis M, Zimmer EA, Chen Z, Savolainen V, Chase MW. 1999. The earliest angiosperms: Evidence from mitochondrial, plastid and nuclear genomes. Nature 402: 404407.
  • Qiu YL, Li L, Hendry TA, Li R, Taylor DW, Issa MJ, Ronen AJ, Vekaria ML, White AM. 2006. Reconstructing the basal angiosperm phylogeny: Evaluating information content of mitochondrial genes. Taxon 55: 837856.
  • Renner SS, Chanderbali AS. 2000. What is the relationship among Hernandiaceae, Lauraceae, and Monimiaceae, and why is this question so difficult to answer? International Journal of Plant Sciences 161(Supplement): S109S119.
  • Renner SS, Hausner G. 2005. Siparunaceae. Flora Neotropica Monograph 95: 1247.
  • Renner SS, Murray D, Foreman D. 2000. Timing transantarctic disjunctions in the Atherospermataceae (Laurales): Evidence from coding and noncoding chloroplast sequences. Systematic Biology 49: 579591.
  • Retallack G, Dilcher DL. 1981. Early angiosperm reproduction: Prisca reynoldsii, gen. et sp. nov. from mid-Cretaceous coastal deposits in Kansas, U.S.A. Palaeontographica Abteilung B 179: 103137.
  • Richardson JE, Chatrou LW, Mols JB, Erkens RHJ, Pirie MD. 2004. Historical biogeography of two cosmopolitan families of flowering plants: Annonaceae and Rhamnaceae. Philosophical Transactions of the Royal Society of London, Series B 359: 14951508.
  • Rohwer JG. 1993. Lauraceae. In: KubitzkiK, RohwerJG, BittrichV eds. The families and genera of vascular plants, vol. II. Flowering plants. Dicotyledons. Magnoliid, hamamelid and caryophyllid families. Berlin : Springer-Verlag. 366391.
  • Rohwer JG, Rudolph B. 2005. Jumping genera: the phylogenetic positions of Cassytha, Hypodaphnis, and Neocinnamomum (Lauraceae) based on different analyses of trnK intron sequences. Annals of the Missouri Botanical Garden 92: 153178.
  • Romanov MS, Endress PK, Bobrov AVFC, Melikian AP, Bejerano AP. 2007. Fruit structure of Monimiaceae s.str. (Laurales). Botanical Journal of the Linnean Society 153: 265285.
  • Rudall PJ, Furness CA. 1997. Systematics of Acorus: Ovule and anther. International Journal of Plant Sciences 158: 640651.
  • Rudall PJ, Sokoloff DD, Remizowa MV, Conran JG, Davis JI, Macfarlane TD, Stevenson DW. 2007. Morphology of Hydatellaceae, an anomalous aquatic family recently recognized as an early-divergent angiosperm lineage. American Journal of Botany 94: 10731092.
  • Saarela JM, Rai HS, Doyle JA, Endress PK, Mathews S, Marchant AD, Briggs BG, Graham SW. 2007. Hydatellaceae identified as a new branch near the base of the angiosperm phylogenetic tree. Nature 446: 312315.
  • Samylina VA. 1968. Early Cretaceous angiosperms of the Soviet Union based on leaf and fruit remains. Journal of the Linnean Society (Botany) 61: 207218.
  • Sauquet H, Doyle JA, Scharaschkin T, Borsch T, Hilu KW, Chatrou LW, Le Thomas A. 2003. Phylogenetic analysis of Magnoliales and Myristicaceae based on multiple data sets: implications for character evolution. Botanical Journal of the Linnean Society 142: 125186.
  • Sauquet H, Weston PH, Barker NP, Anderson CL, Cantrill DJ, Savolainen V. 2009. Using fossils and molecular data to reveal the origins of the Cape proteas (subfamily Proteoideae). Molecular Phylogenetics and Evolution 51: 3143.
  • Scharaschkin T, Doyle JA. 2006. Character evolution in Anaxagorea (Annonaceae). American Journal of Botany 93: 3654.
  • Schneider EL, Carlquist S. 1996. Conductive tissue in Ceratophyllum demersum (Ceratophyllaceae). Sida 17: 437443.
  • Schneider EL, Williamson PS. 1993. Nymphaeaceae. In: KubitzkiK, RohwerJG, BittrichV eds. The families and genera of vascular plants, vol. II. Flowering plants. Dicotyledons. Magnoliid, hamamelid and caryophyllid families. Berlin : Springer-Verlag. 486493.
  • Scotland RW, Olmstead RG, Bennett JR. 2003. Phylogeny reconstruction: The role of morphology. Systematic Biology 52: 539548.
  • Scott RA, Barghoorn ES, Leopold EB. 1960. How old are the angiosperms? American Journal of Science 258-A(Bradley Volume): 284299.
  • Sokoloff DD, Remizowa MV, Macfarlane TD, Tuckett RE, Ramsay MM, Beer AS, Yadav SR, Rudall PJ. 2008. Seedling diversity in Hydatellaceae: Implications for the evolution of angiosperm cotyledons. Annals of Botany 101: 153164.
  • Soltis DE, Soltis PS, Chase MW, Mort ME, Albach DC, Zanis M, Savolainen V, Hahn WH, Hoot SB, Fay MF, Axtell M, Swensen SM, Nixon KC, Farris JS. 2000. Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133: 381461.
  • Soltis DE, Soltis PS, Endress PK, Chase MW. 2005. Phylogeny and evolution of angiosperms. Sunderland , Massachusetts : Sinauer Associates.
  • Soltis PS, Soltis DE, Chase MW. 1999. Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402: 402404.
  • Springer MS, Teeling EC, Madsen O, Stanhope MJ, De Jong WW. 2001. Integrated fossil and molecular data reconstruct bat echolocation. Proceedings of the National Academy of Sciences USA 98: 62416246.
  • Staedler YM, Endress PK. 2009. Diversity and lability of floral phyllotaxis in the pluricarpellate families of core Laurales (Gomortegaceae, Atherospermataceae, Siparunaceae, Monimiaceae). International Journal of Plant Sciences 170: 522550.
  • Staedler YM, Weston PH, Endress PK. 2007. Floral phyllotaxis and floral architecture in Calycanthaceae (Laurales). International Journal of Plant Sciences 168: 285306.
  • Staedler YM, Weston PH, Endress PK. 2009. Comparative gynoecium structure and development in Calycanthaceae (Laurales). International Journal of Plant Sciences 170: 2141.
  • Sun G, Dilcher DL, Zheng S, Zhou Z. 1998. In search of the first flower: A Jurassic angiosperm, Archaefructus, from northeast China. Science 282: 16921695.
  • Sun G, Ji Q, Dilcher DL, Zheng S, Nixon KC, Wang X. 2002. Archaefructaceae, a new basal angiosperm family. Science 296: 899904.
  • Swofford DL. 1990. PAUP: Phylogenetic Analysis Using Parsimony, version 3.0. Champaign , Illinois : Illinois Natural History Survey.
  • Takhtajan AL. 1966. Sistema i filogeniya tsvetkovykh rasteniy. Moscow : Nauka.
  • Takhtajan AL. 1980. Outline of the classification of flowering plants (Magnoliophyta). Botanical Review 46: 225359.
  • TakhtajanAL ed. 1985. Sravrintel’naya anatomiya semyan. Tom 1. Odnodol’nyye. Leningrad : Nauka.
  • TakhtajanAL ed. 1988. Sravrintel’naya anatomiya semyan. Tom 2. Dvudol’nyye. Magnoliidae, Ranunculidae. Leningrad : Nauka.
  • TakhtajanAL ed. 1991. Sravrintel’naya anatomiya semyan. Tom 3. Dvudol’nyye. Caryophyllidae – Dilleniidae. Leningrad : Nauka.
  • Taylor DW. 1991. Angiosperm ovules and carpels: Their characters and polarities, distribution in basal clades, and structural evolution. Postilla 208: 140.
  • Taylor ML, Osborn JM. 2006. Pollen ontogeny in Brasenia (Cabombaceae, Nymphaeales). American Journal of Botany 93: 344356.
  • Taylor ML, Gutman BL, Melrose NA, Ingraham AM, Schwartz JA, Osborn JM. 2008. Pollen and anther ontogeny in Cabomba caroliniana (Cabombaceae, Nymphaeales). American Journal of Botany 95: 399413.
  • Tillich HJ, Tuckett R, Facher E. 2007. Do Hydatellaceae belong to the monocotyledons or basal angiosperms? Evidence from seedling morphology. Willdenowia 37: 399406.
  • Tobe H, Keating RC. 1985. The morphology and anatomy of Hydrastis (Ranunculales): Systematic reevaluation of the genus. Botanical Magazine Tokyo 98: 291316.
  • Tomlinson PB. 1982. Anatomy of the monocotyledons VII. Helobiae (Alismatidae) (including the seagrasses). Oxford : Clarendon Press.
  • Tratt J, Prychid CJ, Behnke HD, Rudall PJ. 2009. Starch-accumulating (S-type) sieve-element plastids in Hydatellaceae: implications for plastid evolution in flowering plants. Protoplasma. doi: 10.1007/s00709-009-0067-2.
  • Trevisan L. 1988. Angiospermous pollen (monosulcate-trichotomosulcate phase) from the very early Lower Cretaceous of Southern Tuscany (Italy): Some aspects. 7th International Palynological Congress (Brisbane) Abstracts, 165.
  • Tucker SC. 1977. Foliar sclereids in the Magnoliaceae. Journal of the Linnean Society (Botany) 75: 325356.
  • Tucker SC. 1982. Inflorescence and flower development in the Piperaceae III. Floral ontogeny of Piper. American Journal of Botany 69: 13891401.
  • Tucker SC, Douglas AW. 1996. Floral structure, development, and relationships of paleoherbs: Saruma, Cabomba, Lactoris, and selected Piperales. In: TaylorDW, HickeyLJ eds. Flowering plant origin, evolution & phylogeny. New York : Chapman & Hall. 141175.
  • Upchurch GR. 1984. Cuticle evolution in Early Cretaceous angiosperms from the Potomac Group of Virginia and Maryland. Annals of the Missouri Botanical Garden 71: 522550.
  • Upchurch GR, Dilcher DL. 1990. Cenomanian angiosperm leaf megafossils, Dakota Formation, Rose Creek locality, Jefferson County, southeastern Nebraska. US Geological Survey Bulletin 1915: 155.
  • Upchurch GR, Crane PR, Drinnan AN. 1994. The megaflora from the Quantico locality (upper Albian), Lower Cretaceous Potomac Group of Virginia. Virginia Museum of Natural History Memoir 4: 157.
  • Vakhrameev VA. 1952. Stratigrafiya i iskopaemaya flora melovykh otlozheniy Zapadnogo Kazakhstana. Regional’naya Stratigrafiya SSSR 1: 1340.
  • Venkata Rao C. 1960. Studies in the Proteaceae I. Tribe Persoonieae. Proceedings of the National Institute of Science of India, Part B—Biological Sciences 26: 300337.
  • Venkata Rao C. 1961. Studies in the Proteaceae II. Tribes Placospermeae and Conospermeae. Proceedings of the National Institute of Science of India, Part B—Biological Sciences 27: 126151.
  • Venkata Rao C. 1971. Proteaceae. New Delhi : Council of Scientific and Industrial Research.
  • Viehofen A, Hartkopf-Fröder C, Friis EM. 2008. Inflorescences and flowers of Mauldinia angustiloba sp. nov. (Lauraceae) from Middle Cretaceous karst infillings in the Rhenish Massif, Germany. International Journal of Plant Sciences 169: 871889.
  • Vink W. 1970. The Winteraceae of the Old World I. Pseudowintera and Drimys—morphology and taxonomy. Blumea 18: 225354.
  • Von Balthazar M, Endress PK. 2002a. Development of inflorescences and flowers in Buxaceae and the problem of perianth interpretation. International Journal of Plant Sciences 163: 847876.
  • Von Balthazar M, Endress PK. 2002b. Reproductive structures and systematics of Buxaceae. Botanical Journal of the Linnean Society 140: 193228.
  • Von Balthazar M, Schönenberger J. 2009. Floral structure and organization in Platanaceae. International Journal of Plant Sciences 170: 210225.
  • Von Balthazar M, Pedersen KR, Crane PR, Friis EM. 2008. Carpestella lacunata gen. et sp. nov., a new basal angiosperm flower from the Early Cretaceous (Early to Middle Albian) of eastern North America. International Journal of Plant Sciences 169: 890898.
  • Von Balthazar M, Pedersen KR, Crane PR, Stampanoni M, Friis EM. 2007. Potomacanthus lobatus gen. et sp. nov., a new flower of probable Lauraceae from the Early Cretaceous (Early to Middle Albian) of eastern North America. American Journal of Botany 94: 20412053.
  • Von Balthazar M, Schatz GE, Endress PK. 2003. Female flowers and inflorescences of Didymelaceae. Plant Systematics and Evolution 237: 199208.
  • Walker JW, Walker AG. 1984. Ultrastructure of Lower Cretaceous angiosperm pollen and the origin and early evolution of flowering plants. Annals of the Missouri Botanical Garden 71: 464521.
  • Walker JW, Brenner GJ, Walker AG. 1983. Winteraceous pollen in the Lower Cretaceous of Israel: Early evidence of a magnolialean angiosperm family. Science 220: 12731275.
  • Wanke S, Vanderschaeve L, Mathieu G, Neinhuis C, Goetghebeur P, Samain MS. 2007. From forgotten taxon to a missing link? The position of the genus Verhuellia (Piperaceae) revealed by molecules. Annals of Botany 99: 12311238.
  • Ward LF. 1895. The Potomac Formation. US Geological Survey Fifteenth Annual Report, Washington , DC. 307397.
  • Ward JV, Doyle JA. 1994. Ultrastructure and relationships of mid-Cretaceous polyforates and triporates from Northern Gondwana. In: KurmannMH, DoyleJA eds. Ultrastructure of fossil spores and pollen. Kew : Royal Botanic Gardens. 161172.
  • Weberling F. 1985. Zur Infloreszenzmorphologie der Lauraceae. Botanische Jahrbücher für Systematik 107: 395414.
  • Weidlich WH. 1973. The organization of the vascular system in the stems of the Nymphaeaceae. Ph.D. Dissertation. Durham : Duke University .
  • Wiens JJ. 2004. The role of morphological data in phylogeny reconstruction. Systematic Biology 53: 653661.
  • Wolfe JA, Doyle JA, Page VM. 1975. The bases of angiosperm phylogeny: Paleobotany. Annals of the Missouri Botanical Garden 62: 801824.
  • Zanis MJ, Soltis DE, Soltis PS, Mathews S, Donoghue MJ. 2002. The root of the angiosperms revisited. Proceedings of the National Academy of Sciences USA 199: 68486853.
  • Zhang XH, Ren Y. 2008. Floral morphology and development in Sargentodoxa (Lardizabalaceae). International Journal of Plant Sciences 169: 11481158.


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results and Discussion
  5. Acknowledgments
  6. References
  7. Appendix

Appendix I: Taxa and characters

The data matrix is presented as Table 1.

Recent taxa
  • 1
    Amborella (= Amborellaceae).
  • 2
    Cabomba (Cabombaceae).
  • 3
    Brasenia (Cabombaceae).
  • 4
    Nuphar (Nymphaeaceae).
  • 5
    Barclaya (Nymphaeaceae).
  • 6
    Nymphaeoideae (Nymphaeaceae).
  • 7
  • 8
    Austrobaileya (= Austrobaileyaceae).
  • 9
    Trimenia (= Trimeniaceae).
  • 10
    Illicium (= Illiciaceae).
  • 11
  • 12
    Hedyosmum (Chloranthaceae).
  • 13
    Ascarina (Chloranthaceae).
  • 14
    Sarcandra (Chloranthaceae).
  • 15
    Chloranthus (Chloranthaceae).
  • 16
    Liriodendron (Magnoliaceae).
  • 17
    Magnolioideae (Magnoliaceae).
  • 18
    Degeneria (= Degeneriaceae).
  • 19
    Galbulimima (= Himantandraceae).
  • 20
    Eupomatia (= Eupomatiaceae).
  • 21
  • 22
  • 23
    Calycanthoideae (Calycanthaceae).
  • 24
    Idiospermum (Calycanthaceae).
  • 25
  • 26
  • 27
    Hortonia (Monimiaceae).
  • 28
    Monimioideae (Monimiaceae).
  • 29
    Mollinedioideae (Monimiaceae).
  • 30
    Gomortega (= Gomortegaceae).
  • 31
  • 32
    Hernandioideae (Hernandiaceae).
  • 33
    Gyrocarpoideae (Hernandiaceae).
  • 34
  • 35
  • 36
  • 37
  • 38
    Lactoris (= Lactoridaceae).
  • 39
    Asaroideae (Aristolochiaceae).
  • 40
    Aristolochioideae (Aristolochiaceae).
  • 41
    Euptelea (= Eupteleaceae).
  • 42
  • 43
  • 44
    Circaeaster (Circaeasteraceae).
  • 45
  • 46
  • 47
    Glaucidium (Ranunculaceae).
  • 48
    Hydrastis (Ranunculaceae).
  • 49
    Core Ranunculaceae.
  • 50
    Nelumbo (= Nelumbonaceae).
  • 51
    Platanus (= Platanaceae).
  • 52
  • 53
    Tetracentron (Trochodendraceae).
  • 54
    Trochodendron (Trochodendraceae).
  • 55
  • 56
    Acorus (= Acoraceae).
  • 57
  • 58
    Butomus (= Butomaceae).
  • 59
    Aponogeton (= Aponogetonaceae).
  • 60
    Scheuchzeria (= Scheuchzeriaceae).
  • 61
  • 62
  • 63
  • 64
  • 65
    Ceratophyllum (= Ceratophyllaceae).
Characters and sources of data

DE, character number in Doyle & Endress (2000); DEU, character number in Doyle et al. (2008a); ED, character number in Endress & Doyle (2009). Characters not included in Endress & Doyle (2009) are marked with an asterisk (*). When no comments are given for a character, see discussion in articles where the character was defined.

1 (ED 1, DE 1). Habit (0) tree or shrub, (1) rhizomatous, scandent, or acaulescent.

When not otherwise indicated, data on anatomical characters not included in Endress & Doyle (2009) for those taxa added or subdivided since Doyle & Endress (2000) are from general compendia and references on taxa cited in Doyle & Endress (2000), especially Metcalfe & Chalk (1950) and Metcalfe (1987); for Hydatellaceae: Cutler (1969); added monocots: Buxbaum (1922, 1927), Tomlinson (1982); Ceratophyllum: Ito (1987), Schneider & Carlquist (1996).

*2 (DE 2). Stele (0) eustele, (1) (pseudo) siphonostele, (2) monocot-type (atactostele). Winteraceae changed from (0/1) to (1) based on presence of a siphonostele in Takhtajania (Keating, 2000).

*3 (DE 3). Inverted cortical bundles (0) absent, (1) present.

4 (ED 2, DE 4). Protoxylem lacunae (0) absent, (1) present. Amborella changed from (0) to (?) following Eklund et al. (2004).

5 (ED 3, DE 14). Pith (0) uniform, (1) septate (plates of sclerenchyma).

6 (ED 4, DE 5). Cambium (0) present, (1) absent.

*7 (DE 6). Storied structure (in tracheids and axial parenchyma, phloem) (0) absent, (1) present. Scored (?) when secondary growth is nearly or entirely lacking.

*8 (DE 7, modified). Tracheary elements (0) tracheids and elements with porose pit membranes, (1) vessel members with typical perforations. State (0) broadened to include porose membranes in Nymphaeales and Acorus, scored (?) in Doyle & Endress (2000), following Eklund et al. (2004).

*9 (DE 9). Vessel perforations (end-wall pits in vesselless taxa) (0) scalariform, (1) scalariform and simple in the same wood, (2) simple. Myristicaceae changed from (0/1) to (1) based on data and topology of Sauquet et al. (2003); Platanus from (0/1) to (1) with elimination of fossil outgroups; Butomus from (2) to (1) based on report of some scalariform pitting by Cheadle (1942). Nartheciaceae and Melanthiaceae: Cheadle & Kosakai (1971).

*10 (DE 10). Fiber pitting (lateral pitting of tracheids in vesselless taxa) (0) distinctly bordered, (1) minutely bordered or simple. Scored (?) when secondary xylem is absent or fibers are replaced by pervasive parenchyma. Hedyosmum changed from (1) to (0/1) based on data and topology of Eklund et al. (2004).

*11 (DE 8). Vessel grouping (0) predominantly solitary, (1) mostly pairs or multiples. Hedyosmum changed from (0/1) to (0) based on Eklund et al. (2004).

*12 (DE 11). Rays (0) narrow (generally not more than four cells wide), (1) wide.

*13 (DE 12). Paratracheal parenchyma (0) absent or scanty, (1) well developed. Taxa with pervasive parenchyma scored (?).

*14 (DE 13). Tangential apotracheal parenchyma bands (0) absent, (1) present. Taxa with pervasive parenchyma scored (?). Myristicaceae changed from (0/1) to (1) based on the nested position of Brochoneura (Sauquet et al., 2003).

*15 (DE 15). Secondary phloem (0) simple, (1) stratified (fibers in small tangential rows or bands several cells thick).

16 (ED 5, DE 16). Sieve element plastids (0) S-type (starch), (1) PI-type, (2) PII-type. Hydatellaceae changed from PII- to S-type based on Tratt et al. (2009).

17 (ED 6, DE 17 part). Fibers or sclerenchyma in pericyclic area (including modified protophloem) of vascular bundles (0) present, (1) absent.

*18 (DE 17 part, modified). Pericyclic ring (0) separate fiber bundles with no intervening fibers or sclerenchyma, (1) more or less continuous ring of fibers and non-U-shaped sclereids, (2) ring of fibers alternating with U-shaped (hippocrepiform) sclereids, (3) continuous homogeneous ring of fibers. Taxa with no fibers or sclerenchyma scored (?). State (1) of Doyle & Endress (2000) split into (1) and (3), with (3) found in monocots and some Aristolochiaceae, and former state (3), no sclerenchyma, expressed in character 17. Scoring of most taxa confirmed by JAD observations on anatomical collections at Harvard and Kew; changes are Lardizabalaceae from (0) to (0/1) and core Ranunculaceae from (1) to (0/1), based on variation among basal genera, and Platanus from (0) to (1). Amborella, which has a continuous ring of U-shaped sclereids, is scored as (2/3). Sarcandra, Chloranthus, and Saururaceae based on Eklund et al. (2004); Myristicaceae changed from (0/1) to (0) based on data and topology of Sauquet et al. (2003).

19 (ED 7, DE 18). Laticifers in stem (0) absent, (1) present.

20 (ED 8, DE 19). Raphide idioblasts (0) absent, (1) present. Scoring of monocots confirmed by Prychid & Rudall (1999).

21 (ED 9, DE 20 part). Phyllotaxis (0) alternate (spiral or distichous), (1) opposite or whorled.

22 (ED 10, DE 20 part). Distichous phyllotaxis (0) absent, (1) on some or all branches.

*23 (DE 21). Nodal anatomy (0) multilacunar, (1) unilacunar one-trace, (2) unilacunar two-trace, (3) trilacunar. Nuphar: Weidlich (1973); Circaeaster: Foster (1963); Glaucidium, Hydrastis, core Ranunculaceae: Tobe & Keating (1985); Tofieldiaceae, Nartheciaceae, and Melanthiaceae assumed to have typical multilacunar monocot nodes based on the steles described by Buxbaum (1922, 1927).

24 (ED 11, DE 22 modified). First appendage(s) on vegetative branch (0) paired lateral prophylls, (1) single distinct prophyll (adaxial, oblique, or lateral). Gomortega: JAD observations, Berkeley Botanical Garden.

25 (ED 12). Leaf base (0) non-sheathing, (1) sheathing (half or more of stem circumference).

26 (ED 13, DE 23 modified). Stipules (0) absent, (1) adaxial/axillary, (2) interpetiolar, (3) paired cap.

27 (ED 14, DE 24). Axillary squamules (0) absent, (1) present.

28 (ED 15, DE 25). Leaf blade (0) bifacial, (1) unifacial.

29 (ED 16, DE 26). Leaf shape (0) obovate to elliptical to oblong, (1) ovate, (2) linear. See Doyle (2007) for discussion of characters 29–35. Calycanthoideae changed from (0/1) to (1) because Chimonanthus, which is basal (Li et al., 2004), is ovate, Calycanthus varies between species, and Sinocalycanthus is slightly ovate; Trochodendron varies around the limit between obovate and ovate (0/1).

30 (ED 17, DE 27 modified). Major venation (0) pinnate with secondaries at more or less constant angle, (1) palmate (actinodromous or acrodromous) or crowded (pinnate with crowded basal secondaries, upward decreasing angle), (2) parallel (lateral veins departing at low angles from the midrib and converging and fusing apically).

31 (ED 18). Fine venation (0) reticulate, (1) open dichotomous in some or all leaves.

32 (ED 19, DE 28 modified). Base of blade (0) not peltate, (1) peltate in some or all leaves.

*33 (new). Apex of blade (0) simple, (1) bilobed (Liriodendron).

34 (ED 20, DE 29 modified). Leaf dissection (0) simple, (1) some or all leaves lobed or compound.

*35 (DE 30). Marginal teeth (0) absent, (1) chloranthoid, (2) monimioid, (3) platanoid. Changes in scoring of Trimenia and Illicium are discussed in Eklund et al. (2004) and Doyle (2007).

*36 (DE 31 modified). Stomata (predominant type on leaf) (0) paracytic, (1) laterocytic, (2) anomocytic, (3) stephanocytic (including tetracytic). State (3) broadened from tetracytic, following Eklund et al. (2004); in Chloranthaceae (all scored as laterocytic in Doyle & Endress, 2000), Hedyosmum is scored as (3), Ascarina (encyclocytic) as (?), Sarcandra as (1), and Chloranthus as (0), based primarily on Kong (2001). Proteaceae changed from (?) to (0/1) based on laterocytic in Bellendena and predominance of paracytic in others (Carpenter et al., 2005); Araceae from (0/2) to (0) based on Keating (2002).

*37 (DE 32). Midrib vasculature (0) simple arc, (1) arc with adaxial plate, (2) ring.

*38 (DE 33). Palisade parenchyma (0) absent (mesophyll homogeneous), (1) present (mesophyll dorsiventral). Cabomba and Brasenia confirmed as (1), Barclaya changed from (?) to (1), Illicium from (0) to (1), Schisandraceae from (?) to (0), and Ascarina, Sarcandra, and Chloranthus from (?) to (0) based on Feild et al. (2004), as discussed in Eklund et al. (2004); Araceae from (0) to (0/1) based on Keating (2002). Hydatellaceae: Rudall et al. (2007); Ceratophyllum: Jones (1931).

39 (ED 21, DE 34). Asterosclerids in mesophyll (0) absent, (1) present. Liriodendron scored (0) based on Tucker (1977).

*40 (DE 35). Oil cells in mesophyll (0) absent, (1) present. Hydatellaceae and Ceratophyllum scored as unknown because of the possibility that oil cells would be lost for functional reasons in submerged aquatics and the presence of tanniniferous cells in Ceratophyllum (Metcalfe & Chalk, 1950) that might be modified oil cells. In Tetracentron, Endress (1986) reported cells resembling oil cells in flowers, but because their contents are unknown we score it as (?).

*41 (DE 36). Mucilage cells in mesophyll (0) absent, (1) present. Hedyosmum changed from (?) to (0) based on Eklund et al. (2004). Hydatellaceae and Ceratophyllum scored as unknown for reasons given for character 40.

42 (ED 22, DE 37 part). Inflorescence (0) solitary flower (or occasionally with 1–2 lateral flowers), (1) botryoid, panicle, or thyrsoid (monotelic), (2) raceme, spike, or thyrse (polytelic).

43 (ED 23). Inflorescence partial units (0) single flowers, (1) cymes. Lauraceae were scored as (0) in Endress & Doyle (2009), but based on Weberling (1985), Endress & Lorence (2004), PKE observations, and the phylogeny of Rohwer & Rudolph (2005), cymes are more likely ancestral, as Hypodaphnis has thyrsoids, Cryptocaryeae have some simple and compound botryoids but thyrsoids are particularly common, and most of the remaining groups have thyrsoids or botryoids.

*44 (new). Inflorescence (or partial inflorescence) (0) not modified, (1) modified into globular head (Platanus). In Myristicaceae, either umbel-shaped fascicles or heads (Brochoneura, Cephalosphaera, Pycnanthus, Scyphocephalium, Staudtia) may be ancestral (Sauquet et al., 2003). Menispermaceae with heads (Stephania, Penianthus species: Kessler, 1993) are well nested in the phylogeny of Hoot et al. (2009).

45 (ED 24). Pedicel (0) present in some or all flowers, (1) absent or highly reduced (flower sessile or subsessile). Menispermaceae changed from (0/1) to (0) (see character 44).

46 (ED 25). Floral subtending bracts (0) present, (1) present in female, absent in male flowers, (2) absent in all flowers.

47 (ED 26, DE 38 modified). Sex of flowers (0) bisexual, (1) unisexual. Taxa with both bisexual and unisexual flowers scored as (0/1).

48 (ED 27, DE 39 modified). Floral base (0) hypanthium absent, superior ovary, (1) hypanthium present, superior ovary, (2) partially or completely inferior ovary.

49 (ED 28). Floral receptacle (female portion) (0) short, (1) elongate.

*50 (new). Pits in receptacle bearing individual carpels (0) absent, (1) present (Nelumbo).

51 (ED 29). Cortical vascular system (0) absent or supplying perianth only, (1) supplying androecium, (2) supplying androecium plus gynoecium.

52 (ED 30). Floral apex (0) used up after production of carpels, (1) protruding in mature flower. Unicarpellate taxa scored as unknown.

53 (ED 31, DE 41 part). Perianth (0) present, (1) absent.

54 (ED 32, DE 40). Perianth phyllotaxis (0) spiral, (1) whorled. Monimioideae changed from (?) in Endress & Doyle (2009) to (0), as in Peumus and Palmeria (Staedler & Endress, 2009); Monimia has dimerous whorls (Lorence, 1985), but because of its nested position we assume this condition is derived.

55 (ED 33, DE 42 modified). Perianth merism (0) trimerous, (1) dimerous, (2) polymerous. Spiral taxa scored as unknown. In Platanus, von Balthazar & Schönenberger (2009) show variation between trimerous and tetramerous, confirming our previous scoring as (0/2).

56 (ED 34, DE 41 modified). Perianth whorls (series when phyllotaxis is spiral) (0) one, (1) two, (2) more than two. Includes petals (character 58); taxa with no perianth scored as unknown. Von Balthazar & Schönenberger (2009) confirm our previous scoring of Platanus as having two whorls (misstated as being 1/2 in the text of Endress & Doyle, 2009). Gomortega changed from (2) in Endress & Doyle (2009) to (1) based on Staedler & Endress (2009).

57 (ED 35, DE 43 modified). Tepal differentiation (0) all more or less sepaloid; (1) outer sepaloid, inner distinctly petaloid; (2) all distinctly petaloid. Does not include petals (36). Single sepaloid cycle scored as (0/1).

58 (ED 36). Petals (0) absent, (1) present. Taxa with no perianth or only one cycle scored as (?).

59 (ED 37, DE 45 modified). Nectaries on inner perianth parts (0) absent, (1) present.

60 (ED 38, DE 44 part). Outermost perianth parts (0) free, (1) at least basally fused. Idiospermum: Staedler et al. (2007).

61 (ED 39, DE 44 part). Calyptra derived from last one or two bracteate organs below the flower (0) absent, (1) present.

62 (ED 40). Stamen number (0) more than one, (1) one.

63 (ED 41, DE 46). Androecium phyllotaxis (0) spiral, (1) whorled. Laurales: Staedler et al. (2007), Staedler & Endress (2009). Monimioideae changed from (?) in Endress & Doyle (2009) to (0) based on Staedler & Endress (2009); Mollinedioideae changed from (1) to (0/1) because of spiral phyllotaxis in Xymalos (Staedler & Endress, 2009).

64 (ED 42, DE 47 modified). Androecium merism (0) trimerous, (1) dimerous, (2) polymerous. Spiral taxa scored as unknown. According to von Balthazar & Schönenberger (2009), Platanus varies between trimerous and tetramerous, so we have changed its scoring from (0/1/2) to (0/2).

65 (ED 43). Number of stamen whorls (series when phyllotaxis is spiral; includes inner staminodes) (0) one, (1) two, (2) more than two. Single stamens scored as unknown. Platanus changed from (0/1) to (0) in Endress & Doyle (2009) based on von Balthazar & Schönenberger (2009); Siparunaceae changed from (?) to (1/2) based on Renner & Hausner (2005) and Staedler & Endress (2009); Gomortega changed from (?) to (2) based on Staedler & Endress (2009).

66 (ED 44). Stamen positions (0) single, (1) double (at least in outer whorl). Taxa with no perianth and/or single stamens scored as unknown. Platanus: von Balthazar & Schönenberger (2009). Monimioideae changed from (?) in Endress & Doyle (2009) to (0) as in other taxa with spiral phyllotaxis.

We have reevaluated characters 64–66 in Saururaceae and Piperaceae in light of phylogenetic results of Wanke et al. (2007) and observations on the order of initiation of stamen primordia and their position relative to each other and to the carpels, as described by Tucker (1982), Liang & Tucker (1995), Tucker & Douglas (1996), and Hufford (1996). In Saururaceae, we changed stamen merism from trimerous to trimerous or dimerous (0/1), number of whorls from two to one or two (0/1), and positions from unknown to double, on the assumption that Saururus and Gymnotheca have two dimerous whorls with lateral stamens in double positions and Anemopsis and Houttuynia have one trimerous whorl, with (Anemopsis) or without (Houttuynia) double positions. In Piperaceae, Wanke et al. (2007) inferred that the ancestral stamen number was two, based on recognition of the basal position of Verhuellia and a Bayesian analysis with stamen numbers simplified to 2, 3, 4, and 6, but this is not easily compared with the more detailed system of three characters used here. We have retained our scoring of merism as either trimerous or dimerous and number of whorls as one or two, but we have rescored stamen position as (0/1) to recognize the possibility that 6-staminate Piper species (Tucker, 1982) and Zippelia (Liang & Tucker, 1995) are basically dimerous with lateral stamens in double positions.

67 (ED 45, DE 48). Stamen fusion (0) free, (1) connate. Taxa with one stamen scored as unknown. Menispermaceae changed from (0/1) to (0): genera with synandria (Kessler, 1993) are well nested in the phylogeny of Hoot et al. (2009). Stamens of Platanus are shortly fused with the neighboring tepals but not among themselves (von Balthazar & Schönenberger, 2009).

68 (ED 46, DE 70). Inner staminodes (0) absent, (1) present. Taxa with one stamen or one whorl of stamens scored as unknown.

69 (ED 47). Glandular food bodies on stamens or staminodes (0) absent, (1) present.

70 (ED 48, DE 49 modified). Stamen base (0) short (2/3 or less the length of anther), (1) long (>2/3 length of anther) and wide (>1/2 width of anther), (2) long (2/3 or more length of anther) and narrow (<1/2 width of anther) (typical filament).

71 (ED 49, DE 50). Paired basal stamen glands (0) absent, (1) present.

72 (ED 50, DE 51 modified). Connective apex (0) extended, (1) truncated or smoothly rounded, (2) peltate.

73 (ED 51, DE 53). Pollen sacs (0) protruding, (1) embedded.

74 (ED 52, DE 52). Microsporangia (0) four, (1) two.

75 (ED 53, DE 54 modified). Orientation of dehiscence (0) distinctly introrse, (1) latrorse to slightly introrse, (2) extrorse.

76 (ED 54, DE 55). Mode of dehiscence (0) longitudinal slit, (1) H-valvate, (2) valvate with upward-opening flaps.

77 (ED 55, DE 56). Connective hypodermis (0) unspecialized, (1) endothecial or sclerenchymatous.

78 (ED 56, DE 57). Tapetum (0) secretory, (1) amoeboid.

79 (ED 57, DE 58). Microsporogenesis (0) simultaneous, (1) successive.

80 (ED 58). Pollen nuclei (0) binucleate, (1) trinucleate.

81 (ED 59, DE 59). Pollen unit (0) monads, (1) tetrads.

82 (ED 60, DE 62). Pollen size (average) (0) large (>50 μm), (1) medium (20–50 μm), (2) small (<20 μm); ordered.

83 (ED 61, DE 60 modified). Pollen shape (0) boat-shaped, (1) globose, (2) triangular, angulaperturate (Proteaceae).

84 (ED 62, DE 61 modified). Aperture type (0) polar (including sulcate, ulcerate, and disulcate), (1) inaperturate, (2) sulculate, (3) (syn)tricolpate with colpi arranged according to Garside's law, with or without alternating colpi, (4) tricolpate.

85 (ED 63). Distal aperture shape (0) elongate, (1) round.

86 (ED 64). Distal aperture branching (0) unbranched, (1) with several branches (Hedyosmum).

87 (ED 65, DE 63). Infratectum (0) granular (including “atectate”), (1) intermediate, (2) columellar; ordered.

88 (ED 66, DE 64). Tectum (0) continuous or microperforate, (1) perforate (foveolate) to semitectate (reticulate), (2) reduced (not distinguishable from underlying granules).

89 (ED 67). Grading of reticulum (0) uniform, (1) finer at ends of sulcus (liliaceous), (2) finer at poles (rouseoid). Scored only in taxa with state (1) in character 88.

90 (ED 68, DE 65). Striate muri (0) absent, (1) present.

91 (ED 69, DE 66). Supratectal spinules (smaller than the width of tectal muri in foveolate-reticulate taxa) (0) absent, (1) present.

92 (ED 70, DE 67). Prominent spines (larger than spinules, easily visible with light microscopy) (0) absent, (1) present.

93 (ED 71, DE 68). Aperture membrane (0) smooth, (1) sculptured.

94 (ED 72, DE 69 modified). Extra-apertural nexine stratification (0) foot layer, not consistently foliated, no distinctly staining endexine or only problematic traces, (1) foot layer and distinctly staining endexine, or endexine only, (2) all or in part foliated, not distinctly staining. Brasenia and Cabomba changed from (0) to (1) based on Taylor & Osborn (2006) and Taylor et al. (2008).

95 (ED 73). Nexine thickness (0) absent or discontinuous traces, (1) thin but continuous, (2) thick (1/3 or more of exine); ordered.

96 (ED 74, DE 71), modified. Carpel number (0) one, (1) 2–5 in one whorl (series when phyllotaxis is spiral), (2) more than 5 in one whorl or series, (3) more than one whorl or series. State (2) corresponds to the “star-shaped” carpel arrangement of von Balthazar et al. (2008); see also Endress (2006). Scoring based on references cited in Endress & Doyle (2009) and the following: Nymphaeales and other ANITA taxa: Ito (1986), Schneider & Williamson (1993), Endress (2001); Laurales: Staedler & Endress (2009); Winteraceae: Vink (1970); Asaroideae: Leins & Erbar (1995); Lardizabalaceae: Zhang & Ren (2008); Platanus: von Balthazar & Schönenberger (2009); Tetracentron: Chen et al. (2007). Estimates of ancestral states in variable taxa based on the following: Illicium: Oh et al. (2003); Annonaceae: Doyle & Le Thomas (1996); Menispermaceae: Kessler (1993), Hoot et al. (2009).

97 (ED 75, DE 72). Carpel form (0) ascidiate up to stigma, (1) intermediate (both plicate and ascidiate zones present below the stigma) with ovule(s) on the ascidiate zone, (2) completely plicate, or intermediate with some or all ovule(s) on the plicate zone.

98 (ED 76, DE 73 part). Postgenital sealing of carpel (0) none, (1) partial, (2) complete.

99 (ED 77, DE 73 part). Secretion in area of carpel sealing (0) present, (1) absent.

100 (ED 78, DE 74). Pollen tube transmitting tissue (0) not prominently differentiated, (1) one layer prominently differentiated, (2) more than one layer prominently differentiated.

101 (ED 79, DE 75). Style (0) absent (stigma sessile or capitate), (1) present (elongated, distinctly constricted apical portion of carpel).

102 (ED 80, DE 76). Stigma (0) extended (half or more of the style-stigma zone), (1) restricted (above slit or around its upper part).

103 (ED 81, DE 77 part). Multicellular stigmatic protuberances or undulations (0) absent, (1) present.

104 (ED 82, DE 77 part, modified). Stigmatic papillae (most elaborate type) (0) absent, (1) unicellular or with a single emergent cell and one or more small basal cells, (2) uniseriate pluricellular with emergent portion consisting of two or more cells. Idiospermum scored (1) based on Staedler et al. (2009).

105 (ED 83, DE 78). Extragynoecial compitum (0) absent, (1) present.

106 (ED 84, DE 79). Carpel fusion (0) apocarpous, (1) parasyncarpous, (2) eusyncarpous (at least basally). Taxa with one carpel scored as unknown. Eupomatia, Calycanthoideae, and Gomortega show irregular and presumably autapomorphic types of fusion (Doyle & Endress, 2000) that we previously scored as (?), but because Calycanthoideae are only postgenitally coherent at the styles (Staedler et al., 2009), we have rescored them as (0).

107 (ED 85, DE 80). Oil cells in carpels (0) absent or internal, (1) intrusive. Taxa with no oil cells in any tissue scored as unknown.

108 (ED 86). Long unicellular hairs on and/or between carpels (0) absent, (1) present.

109 (ED 87). Short curved appressed unlignified hairs with up to two short basal cells and one long apical cell on carpels (0) absent, (1) present.

110 (ED 88). Nectary on dorsal or lateral sides of carpel or pistillode (0) absent, (1) present.

111 (ED 89, DE 81). Septal nectaries or potentially homologous basal intercarpellary nectaries (0) absent, (1) present.

112 (ED 90, DE 82 modified). Number of ovules per carpel (0) one, (1) two or varying between one and two, (2) more than two.

113 (ED 91, DE 83 modified). Placentation (0) ventral, (1) laminar-diffuse or “dorsal”.

114 (ED 92, DE 84). Ovule direction (0) pendent, (1) horizontal, (2) ascendent.

115 (ED 93, DE 85). Ovule curvature (0) anatropous (or nearly so), (1) orthotropous (including hemitropous).

Data on integument characters not included in Endress & Doyle (2009) for taxa added or subdivided since Doyle & Endress (2000) are from Corner (1976), Takhtajan (1988), Taylor (1991), Endress & Igersheim (1997, 1999), Igersheim & Endress (1997, 1998), and Igersheim et al. (2001); for Glaucidium, Hydrastis, and core Ranunculaceae: Tobe & Keating (1985); Nartheciaceae, Melanthiaceae: Oganezova (1984).

116 (ED 94, DE 86). Integuments (0) two, (1) one.

*117 (DE 87). Outer integument shape (0) semiannular, (1) annular. Orthotropous taxa scored as unknown. Myristicaceae changed from (1) in Doyle & Endress (2000) to (0/1) based on Sauquet et al. (2003), Calycanthoideae from (0/1) to (0) based on Staedler et al. (2009). Atherospermataceae changed from (0/1) to (1) because the one semiannular group, Doryphora (Endress, 1972), is linked with Daphnandra, which is annular (Endress & Igersheim, 1997; Renner et al., 2000).

*118 (DE 88). Outer integument lobation (0) unlobed, (1) lobed. Calycanthoideae changed from (0/1) to (1), Idiospermum from (0) to (1) based on Staedler et al. (2009).

*119 (DE 89). Outer integument thickness (at middle of integument length) (0) two cells, (1) two and three to four, (2) four and five, or more; ordered.

*120 (DE 90). Inner integument thickness (0) two cells, (1) two and three, or three, (2) three and more; ordered. Proteaceae changed from (1) to (1/2) based on Venkata Rao (1960, 1961).

121 (ED 95, DE 91). Chalaza (0) unextended, (1) pachychalazal, (2) perichalazal. Orthotropous taxa scored as unknown.

122 (ED 96, DE 92 modified). Nucellus (0) crassinucellar (including weakly so), (1) tenuinucellar or pseudocrassinucellar. Idiospermum: Staedler et al. (2009).

123 (ED 97, DE 93 part). Fruit wall (0) wholly or partly fleshy, (1) dry.

124 (ED 98, DE 93 part). Lignified endocarp (0) absent, (1) present. Taxa with dry fruit wall scored as unknown.

125 (ED 99, DE 94 modified). Fruit dehiscence (0) indehiscent or dehiscing irregularly, dorsally only, or laterally, (1) dehiscent ventrally or both ventrally and dorsally, (2) horizontally dehiscent with vertical extensions.

*126 (new). Hooked hairs on fruit (0) absent, (1) present (Circaeaster).

Data on seed characters not included in Endress & Doyle (2009) for taxa added or subdivided since Doyle & Endress (2000) are from Corner (1976) and Takhtajan (1985, 1988, 1991); for Hydatellaceae: Hamann (1975), Hamann et al. (1979); Atherospermataceae, Gomortega: Doweld (2001); Proteaceae: Bellendena and Persoonioideae as described by Venkata Rao (1960, 1961, 1971); Trochodendron, Tetracentron: Doweld (1998); Tofieldiaceae, Nartheciaceae, Melanthiaceae: Oganezova (1984).

127 (ED 100, DE 95). Testa (0) slightly or non-multiplicative, (1) multiplicative. Tofieldiaceae changed from (?) to (0) in Endress & Doyle (2009) based on Takhtajan (1985) and Oganezova (1984).

128 (ED 101, DE 96). Exotesta (0) unspecialized, (1) palisade or shorter sclerotic cells, (2) tabular, (3) longitudinally elongated, more or less lignified cells.

*129 (DEU 60, DE 97, in part). Mesotesta lignification (0) unlignified, (1) with sclerotic layer, (2) with fibrous layer. Splitting character 97 of Doyle & Endress (2000) into this and the following character (Doyle et al., 2008a) allows recognition of the presence of both sclerotic and fleshy layers in the mesotesta of Austrobaileya (Endress, 1980).

*130 (DEU 61, DE 97, in part, modified). Mesotesta fleshiness (0) not juicy, (1) wholly or partly modified into a juicy sarcotesta. We previously recognized a third state (2), spongy, in Hernandioideae, Gyrocarpoideae, and Aponogeton. However, the mesotesta is thin in Gyrocarpus jatrophifolius (Takhtajan, 1988), and according to Kimoto & Tobe (2008) it is parenchymatous in Illigera (Hernandioideae) and disappears during development in Sparattanthelium (Gyrocarpoideae). This would necessitate scoring both subfamilies as uncertain, which would make the spongy state uninformative, so we have eliminated this state and rescored the three taxa as (0).

*131 (DEU 62, DE 98). Endotesta (0) unspecialized, (1) single layer of thin-walled cells with fibrous endoreticulum, (2) multiple layer of thin-walled cells with fibrous endoreticulum, (3) tracheidal, (4) palisade of thick-walled cells.

*132 (DEU 63, DE 99). Tegmen (0) unspecialized, (1) both ecto- and endotegmen thick-walled, (2) exotegmen fibrous to sclerotic. In Doyle & Endress (2000) we scored Sarcandra/Chloranthus as having a fibrous exotegmen (2), modified in Eklund et al. (2004) to (0) for Sarcandra, based on Corner (1976) and Endress (1987), and (2) for Chloranthus. However, although Corner (1976) described exotegminal cells in Chloranthus erectus (as Chloranthus elatior) as lignified and fibriform, in Chloranthus spicatus the thick-walled tegminal cells are small and cuboidal rather than elongate, and restricted to micropylar end of the seed (Endress, 1987; original observations), and we have therefore rescored Chloranthus as (0/2). Trochodendron and Tetracentron changed from (0) to (2) based on Doweld (1998); Acorus from (2) to (1) because both layers are thick-walled (Takhtajan, 1985; Rudall & Furness, 1997).

133 (ED 102, DE 100). Ruminations (0) absent, (1) testal, (2) tegminal and/or chalazal.

134 (ED 103, DE 101). Operculum (0) absent, (1) present.

135 (ED 104, DE 102). Aril (0) absent, (1) present.

136 (ED 105). Female gametophyte (0) four-nucleate, (1) eight- or nine-nucleate. Tetrasporic types in Piperaceae scored as unknown. The four-nucleate state in Hydatellaceae has been confirmed by Friedman (2008).

137 (ED 106, DE 103). Endosperm development (0) cellular, (1) nuclear, (2) helobial.

138 (ED 107, DE 104). Endosperm in mature seed (0) present, (1) absent.

139 (ED 108, DE 105 modified). Perisperm (0) absent, (1) from nucellar ground tissue, (2) from nucellar epidermis (Acorus).

140 (ED 109, DE 106). Embryo (0) minute (less than 1/2 length of seed interior), (1) large.

141 (ED 110, DE 107). Cotyledons (0) two, (1) one. Hydatellaceae have been described as having one cotyledon by Tillich et al. (2007) but two cotyledons fused into a bilobed structure, as in Nymphaeales, by Sokoloff et al. (2008); because these structures are reduced and difficult to interpret we continue to score Hydatellaceae as (?).

*142 (DE 108). Germination (0) epigeal, (1) hypogeal. Hydatellaceae: Cooke (1983). Hedyosmum: Edwards (1920); Sarcandra: PKE observations; see Eklund et al. (2004).