American Journal of Botany 91(10): 1666±1682. 2004.

FOSSIL EVIDENCE AND PHYLOGENY: THE AGE OF MAJOR ANGIOSPERM CLADES BASED ON MESOFOSSIL AND MACROFOSSIL EVIDENCE FROM DEPOSITS1

WILLIAM L. CREPET,2 KEVIN C. NIXON, AND MARIA A. GANDOLFO

228 Science Building, L. H. Bailey Hortorium, Department of Plant Biology, Cornell University, Ithaca, New York 14853-4301 USA

The record has played an important role in the history of evolutionary thought, has aided the determination of key relationships through mosaics, and has allowed an assessment of a number of ecological hypotheses. Nonetheless, expectations that it might accurately and precisely mirror the progression of taxa through time seem optimistic in light of the many factors potentially interfering with uniform preservation. In view of these limitations, attempts to use the fossil record to corroborate phylogenetic hypotheses based on extensive comparisons among extant taxa may be misplaced. Instead we suggest a methodÐminimum age node mappingÐfor combining reliable fossil evidence with hypotheses of phylogeny. We use this methodology in conjunction with a phylogeny for angiosperms to assess timing in the history of major angiosperm clades. This method places many clades both with and without fossil records in temporal perspective, reveals discrepancies among clades in propensities for preservation, and raises some interesting questions about angiosperm evolution. By providing a context for understanding the gaps in the angiosperm fossil record this technique lends credibility and support to the remainder of the angiosperm record and to its applications in understanding a variety of aspects of angiosperm history. In effect, this methodology empowers the fossil record.

Key words: angiosperms; fossil history; minimum age; node dating.

Recent developments have provided a better understanding morphology, biogeography, and ecology in a temporal context, of the history, evolution and relationships of (and within) the has both informative potential and the capacity to serve as a angiospermsÐa set of phenomena that together constitute corroborative mechanism for a variety of hypotheses about what Darwin considered to be the intractable ``abominable evolution, development/homology, relationships, biogeogra- mystery'' (e.g., Crepet, 2000). This special issue of the Amer- phy, and generally about the evolutionary play in the ecolog- ican Journal of Botany reveals that progress is being made in ical theater (sensu Hutchinson, 1965). In addition, the fossil understanding these problems and that there is imminent hope record, as representative of the history of life, holds the po- for more detailed and accurate understanding of these events. tential for clarifying relationships among extant taxa by re- It also illustrates that three broad areas in particular have con- vealing extinct mosaic taxa that link modern ones, in addition tributed to our improved understanding of angiosperm history to providing the general pattern of evolution of taxa through and relationships: new data and methods for molecular genet- time. Historically, and in the context of evolutionary biology, ics (Palmer and Zamir, 1982; Matthews and Donoghue, 1999), the fossil record has played both informative and corroborative renewed interest in paleontology (Crane et al., 1989; Crepet roles and continues to be called upon to do both. However, and Nixon, 1996; Gandolfo et al., 1998a; Zhou et al., 2001), the advent of modern methodologies for comparative studies and new developments in the analysis of phylogenetic rela- of extant taxa invites a reassessment of the primacy and scope tionships (Nixon, 1996, 1999; Goloboff, 1999). There is also of the fossil record in addressing questions of evolution and an invigorated interest in methods for dating of diversi®cation systematic relationships. events (Sanderson et al., 2004). What is distinctive about this moment and also raises the prospect that problems in under- The fossil record post evolutionary theory: utility and lim- standing angiosperm history and relationships will ultimately itations in phylogenetic studiesÐWhereas in a ``pre-evolu- be resolved is that there is increasing synthesis of these dis- tionary'' world scientists struggled to understand the meaning parate approaches and resultant heightening of insights into of the fossil record and were ultimately informed by it and the ``abominable mystery.'' This special issue illustrates the post evolutionary-theory scientists sought to con®rm momentum of this growing synthesis. Our contribution is in- and understand the process of evolutionary change and to clar- tended to provide a perspective on one aspect of the potential ify relationships among extant taxa, in this genomics era, sci- importance of paleontological data: timing. entists have an additional tendency to look to the fossil record The fossil record, as a body of knowledge that incorporates to con®rm increasingly accurate hypotheses of evolutionary relationships based on comparative studies of extant taxa. 1 Manuscript received 2 February 2004; revision accepted 17 June 2004. These relationships have become increasingly clear and ac- The authors thank Jennifer L. Svitko and Elizabeth Hermsen for their as- curate through analyses of multiple gene sequences made pos- sistance with literature searches. We also thank Jeff Doyle, L. H. Bailey Hor- sible principally through pioneering innovations by, for ex- torium and Plant Biology, Cornell, for helpful comments on the manuscript ample, Palmer and Zamir, 1982; Palmer et al., 1983, 1985; and Melissa Luckow, L. H. Bailey Hortorium and Plant Biology, Cornell for her insights into legume history and relationships. This research was sup- Downie and Palmer, 1992, who later, through multi gene anal- ported by NSF grant DEB 0108369 to WLC and KCN. ysis, suggested along with others that Amborella and Nym- 2 Author for correspondence ([email protected]). phaeales were among basal extant angiosperms (e.g., Parkin- 1666 October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1667 son et al., 1999; Matthews and Donoghue, 1999; Qiu et al., 2000; Soltis and Soltis, 2003; Sanderson et al., 2004) and what 1999; Soltis et al., 1999; Chaw et al., 2000; Graham and Olm- have been called deceptively ``errorless'' numbers (Graur and stead, 2000; Soltis et al., 2000; and Barkman et al., 2000, who Martin, 2004). These methods also usually suffer from the in contrast to the others, placed Amborella ϩ waterlilies as a need to simplify calculations by using a single divergence sister group to the remainder of angiosperms; also see Zanis point (i.e., often a single fossil or related fossils; see WikstroÈm et al., 2002, and Soltis and Soltis, 2004, for a detailed discus- et al., 2001; Davies et al., 2004; Yoo et al., 2004; see San- sion). Notable among these analyses is the most inclusive, derson et al., 2004, for relevant discussion) for calculating based on numbers of taxa simultaneously analyzed and made rates that are then applied throughout the tree, compounding possible via various powerful algorithms that allow the rapid the potential problem of incorrect calibration and problems analysis of large matrices representing numerous taxa (Golo- with error due to an incomplete fossil record (Graur and Mar- boff, 1999; Nixon, 1999) supporting the basal placement of tin, 2004). With all of these potential problems, given the pre- Amborella and Nymphaeales (Soltis et al., 2000). These anal- sent state of the art, even very careful and comprehensive anal- yses contribute to a growing consensus about the major rela- yses can be dramatically inaccurate and misleading. A relevant tionships in angiosperms (e.g., Graham and Olmstead, 2000; and striking example of a failure of molecular rate estimation Savolainen et al., 2000; Zanis et al., 2002, 2003; Borsch et al., even with an apparently careful analysis is provided by a re- 2003; Hilu et al., 2003; MagalloÂn and Sanderson, 2002; Chase, cent preliminary study of Nymphaeaceae using three of the 2004; Judd and Olmstead, 2004; Soltis and Soltis, 2004). Such most common estimation methods: rate smoothing (NPRS), comprehensive cladograms (e.g., Soltis et al., 2000) provide penalized likelihood (PL), and a ``Bayesian'' approach (Yoo additional opportunities for evaluating the fossil record or et al., 2004). Although their study estimated the age of modern vice-versa. Corroborative use of fossil evidence in supporting (``crown-group'') Nymphaeaceae to be ca. 40 million years such hypotheses can be as straightforward as observing broad old, extremely well-preserved fossils that nest within modern correlations between the fossil record and ordinal sequences Nymphaeaceae based on careful phylogenetic analysis are now implied by hypotheses of phylogeny (Friis et al., 2001) or known from ca. 90 million-year-old deposits (see Microvic- more complicated and sophisticated as in statistically checking toria discussed later; Gandolfo et al., 2004). In this case, the the ®t of the fossil record to model-based expectations derived failure of molecular substitution rate methods is probably at from phylogenies based on comparative analyses of extant least in part from poor calibration based on an incomplete taxa (e.g., Benton, 2001). These latter approaches and others understanding of the fossil record, emphasizing the need for (e.g., Norell and Novacek, 1992) have been somewhat contro- precisely identi®ed fossils in this context. Projecting such versial and variously have additional quali®cations (e.g., those models over larger trees based on single calibration points that limit the ®eld of the comparisons to make it more com- (e.g., WikstroÈm et al., 2001), therefore, is highly suspect and patible with available fossil data; Benton, 1998). raises questions about the ef®cacy of using calculated timing Another approach is to use fossil data to calibrate molecular of angiosperm radiation events in assessing ecological factors rate substitution models (often popularly called ``molecular in their radiation (e.g., Davies et al., 2004). clock'' modelsÐe.g., Soltis and Soltis, 2003; Hedges and Ku- Thus, attempts to use molecular substitution rate analyses mar, 2004; and see Sanderson et al., 2004) that are then used to assess timing in angiosperm evolution (Sanderson, 1997; to date evolutionary events with any one of several methods MagalloÂn et al., 1999; WikstroÈm et al., 2001; Davies et al., including most commonly ``rate smoothing'' or maximum 2004; Yoo et al., 2004), controversial in and of themselves likelihood. When estimates based on substitution models differ (e.g., Graur and Martin, 2004; see discussion in Sanderson et from observed fossil ages, the molecular estimates are typi- al., 2004), are, in any case, unavoidably compromised by the cally accepted as ``better'' (more accurate) due to the inherent inherently imperfect nature of the fossil record manifest in: underestimation and variability of an incomplete fossil record. This has resulted in extremely disparate and sometimes out- 1. Gaps in the record due to missing strata and to taxa or landish estimates of the ages of major clades (see Graur and organs of taxa that may not have been preserved because Martin, 2004; but also Hedges and Kumar, 2004, for a re- of inappropriate conditionsÐenvironmental or related to sponse). The purpose of this paper is not to review such meth- the ephemerality and durability of the organs in question. ods in detail (see Sanderson et al., 2004, for relevant discus- 2. Imperfect sampling due to stochastic elements including sion); the reader is also referred to Graur and Martin (2004; outcrop exposure and to the relatively low numbers of but also see the response to it by Hedges and Kumar, 2004) paleontologists/paleobotanists collecting and investigat- for cogent criticisms of both the methods and some interesting ing, even in aggregate over time. and celebrated speci®c applications and for the responses. 3. Skewing due to favorable preservational motifs that are Both of these publications point out two important factors that often ecologically correlated, such as charcoali®cation vs. have relevance to this paper and will be discussed further: compression in the fossil ¯ower record. weaknesses in the fossil record and the overriding importance 4. Dif®culties in identifying (i.e., phylogenetically placing) of accurate fossil identi®cations. Absent the alleged hyperbole fossil taxa and consequent use of imprecisely identi®ed of Graur and Martin (according to Hedges and Kumar, 2004), fossils in molecular substitution rate analyses. This prob- from our perspective, molecular rate substitution methods, al- lem of identi®cation occurs with almost every signi®cant though recently extremely popular (WikstroÈm et al., 2001; Da- fossil, but is particularly acute in the dispersed re- vies et al., 2004; Yoo et al., 2004), are complicated by nu- cord, due to a lack of discrete pollen characters, imperfect merous factors including but not limited to assumptions about understanding of pollen variation in modern taxa, and lack constant substitution rates among different lineages, problems of connection between pollen and meso- or macrofossils. with calibration due to incomplete fossil records, and/or the 5. Complications related to interpreting local changes in ¯ora complexity of models that attempt to compensate for substi- constituents vs. actual progression in global taxonomic di- tution variability (e.g., discussion in Sanderson, 1997; Bremer, versity due to evolution and extinction. 1668 AMERICAN JOURNAL OF BOTANY [Vol. 91

Given the limitations noted, the fossil record is likely to be MATERIALS AND METHODS biased, both ecologically and phylogenetically relative to mod- ern diversity. Thus, in addition to the potential dangers in at- There have been several signi®cant developments relevant to the angio- tempting to use the fossil record to calibrate rates of evolution sperm fossil record within the past 30 years. With respect to the nature of the based on molecular data [especially in cases where the choice fossil evidence itself, perhaps the addition of ¯owers and in¯orescences to of fossil(s) for calibration might be problematical], there is no the array of organs that have been fossilized and that are now routinely studied reasonable expectation that available fossils will accurately re- constitutes an important addition to angiosperm paleontology. And, because ¯ect the evolutionary history of every group, and certainly, of the high information content of such fossils, they are of primary interest the fossil records of some groups (e.g., vertebrates, shelled to us in the context of dating angiosperm history. Ironically, ¯owers, the mollusks) are likely to be much more reliable than others. organs that embody so many important characters of angiosperm taxa (and Therefore, we believe that efforts to ``test'' the congruence of indeed were once the primary characters used in angiosperm systematics and the fossil record with cladograms of modern and/or fossil taxa remain so in the array of angiosperm morphological features; Cronquist, 1981; (e.g., Norell and Novacek, 1992) are of ambiguous value. If Lawrence, 1951), had with occasional exceptions (Conwentz, 1886) been es- sentially unstudied from the perspective of fossils because they were generally such a test revealed that the fossil record was congruent (or regarded as too rare and/or poorly preserved to make substantial contributions consistent) with the cladograms of modern taxa, the possibility to our understanding of angiosperm history and relationships. Now they are would remain that such congruence was illusory due to sam- the principal foci of many research laboratories, and have been since the pling error. Alternatively, if the test reveals incongruence, it beginning of a concerted and continuing effort on investigations of fossil would be unwise to reject cladograms based on molecular and/ ¯owers (Crepet et al., 1974), which have been followed by a sustained and or morphological evidence of extant, observable taxa due to continuing line of investigations based on fossil ¯owers by the same authors incongruence with an admittedly incomplete and thus possibly and others (e.g., Crepet et al., 1975, 1980; Crepet and Dilcher, 1977; Tiffney, biased fossil record. Furthermore, using this ``congruence'' ap- 1977; Crepet, 1978, 1979a, b, 1984a, b, 1989; Crepet and Stuessy, 1978; proach to discriminate between alternative extant taxon-based Crepet and Daghlian, 1980, 1982a, b; Daghlian et al., 1980; Zavada and Cre- hypotheses of phylogeny might be useful in certain instances pet, 1981; Krassilov et al., 1983; Crepet and Taylor, 1985, 1986; Crepet and today (e.g., Benton, 1998), but one might speculate that, as Friis, 1987; Friis and Crepet, 1987; Taylor and Crepet, 1987; Stockey, 1987; analyses of modern taxa become more inclusive and complete, Crepet and Feldman, 1991; Stockey and Pigg, 1991; Pedersen et al., 1994; there will be fewer instances of con¯icting hypotheses until Stockey and Rothwell, 1997; HernaÂndez-Castillo and Cevallos-Ferris, 1999; there is general agreement, as seems to be the trend in esti- Boucher et al., 2003). mating angiosperm relationships (see papers in this volume A very signi®cant development within fossil ¯ower studies has been the and references therein e.g., Parkinson et al., 1999; Chaw et investigation of charcoali®ed fossils preserved in three dimensions. These al., 2000; Savolainlen et al., 2000; Soltis et al., 2000; Hilu et fossils have been of particular importance because their detailed three-dimen- al., 2003). sional preservation captures numerous characters that are often unavailable in We believe that a straightforward and informative approach compressed specimens and because their agesÐranging from Early to Late to combining fossil evidence with hypotheses of phylogeny CretaceousÐrepresent a critical interval in angiosperm radiation. Since their ®rst discovery on Martha's Vineyard, Massachusetts, USA by Tiffney (1977) based on comparative studies of extant taxa is available and later in southern Sweden by Friis and Skarby (1981), such fossils have through the simple exercise of minimum age node mapping. been the object of numerous signi®cant continuing investigations as have sim- We use this methodology herein to assess timing in the history ilar fossils from and Portugal (e.g., Friis, 1983, 1984, 1990; of the angiosperms and to explore some of the broad impli- Friis et al., 1986, 1988, 1992, 1994, 1997a, b; Crane et al., 1989; Drinnan et cations of the results. al., 1990, 1991; Crepet et al., 1991, 1992; Herendeen et al., 1993, 1994; Nixon and Crepet, 1993; Crepet, 1996, 1999, 2000, 2001; Crepet and Nixon, 1996, GoalsÐminimum age mappingÐa means for the fossil 1998a, b; Eklund et al., 1997; Gandolfo et al., 1998a, b, c, 2000, 2002, 2004; record of angiosperms to retain the potential to be both MagalloÂn-Puebla et al., 1997; Zhou et al., 2001; Eklund, 2003; Hermsen et informative and corroborativeÐBy mapping estimated min- al., 2003; and others). In aggregate, these fossils have made substantial con- imum ages for all possible clades on a large comprehensive tributions to our understanding of Cretaceous angiosperm history and they angiosperm cladogram, we intend to provide estimates for continue to be the objects of sustained inquiries. Yet, it is important to note some of the more important angiosperm clades based on the that compressed and even petri®ed ¯owers have also continued to make con- best available mesofossil (middle-sized fossils), macrofossil tributions to our understanding and constitute a signi®cant element of angio- and, in a few cases, palynological evidence. We have taken a sperm paleobotany to this day (e.g., Dilcher et al., 1976a, b, 1978; Doyle and selective approach, using only the fossils that have either been Hickey, 1976; Crane, 1981, 1989; Basinger and Dilcher, 1984 ; Crane and assigned through direct cladistic analysis or have clearly iden- Stockey, 1985; Dilcher and Kovach, 1986; Dilcher and Manchester, 1988; ti®able characters to place them ``unequivocally'' (i.e., with Dilcher and Basson, 1990; Call and Dilcher, 1992; Manchester, 1992; Mohr high con®dence) in an identi®able clade of modern angio- and Friis, 2000; Sun et al., 2002; Taylor and Hickey, 1990, 1992). sperms. As will be discussed in Materials and Methods, we Flower fossil studies would have failed to reach their full potential without have excluded some fossils from our analysis, following the the second major developmentÐthat of algorithms for evaluating phyloge- netic relationships based on character arrays. The general topic of phyloge- caveats raised by the original authors related to ambiguity, netics and fossils has been addressed by Nixon (1996) and is discussed by con¯icting placements, or low con®dence in their taxonomic Crane et al. (2004), but the need for precise and reliable identi®cation of assignments. This does not mean that these excluded fossils fossils that often represent extinct taxa requires that they be identi®ed through should be ignored in future analyses; indeed, many are merely a phylogenetic context. Such an approach makes the process of identifying a in need of careful analysis in a phylogenetic context in order fossil transparent and includes an explicit illumination of the characters attri- to place them in a broader phylogeny of angiosperms. buted to it. This process adds objectivity to identi®cations that may be inher- Because the evidence for both extant and fossil taxa is much ently dif®cult, as in cases where the fossils might be mosaics relative to extant more complete for certain clades, we devote sections to some taxa. Overall phylogenetic approaches to understanding plant relationships by of the more important ones (assessed subjectively) and deal means of modern cladistic methods were initiated by Parenti (1980) followed with these and the fossils that represent them in greater detail. by Young and Seigler (1981; angiosperms), Hill and Crane (1982), Crane October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1669

(1985), Doyle and Donoghue (1986), Loconte and Stevenson (1990), Nixon for not including some fossil taxa in an effort to clarify the actual sequence et al. (1994), Rothwell and Serbet (1994) and by subsequent reanalyses by of appearance of taxa in the fossil record of the angiosperms. Thus the in- the same authors. Most recently, there have been analyses of molecular-based formed and interested reader can draw their own conclusions regarding pu- or combined data sets, which are noted later (e.g., Soltis et al., 2000). Use of tative af®nities and the validity of our treatments of particular fossil taxa. phylogenetic analysis to identify fossil ¯owers has a separate, but more or Due to well-known problems with existing information on dispersed pollen less parallel history and issues speci®c to including fossil evidence in phy- grains (e.g., Zavada, 1984 ; Harley and Zavada, 2000), we have used such logenetic analysis have been discussed by Nixon (1996). Early applications evidence sparingly in this analysis. In cases where the pollen identi®cation of phylogenetic context for identi®cation of fossils included Crane and Man- seems relatively strong, based on unique combinations of characters that are chester (1982, fruits of Juglandaceae) and Crepet and Nixon (1989a, b) on found only in one modern clade, we have accepted determinations and used Tertiary Fagaceae. And, this methodology has now become standard in iden- these records for our estimates. The caveats and problems associated with tifying charcoali®ed ¯owers as in the Turonian taxa Tylerianthus and Dres- literature references to ostensibly identi®ed fossils outlined earlier apply to siantha (Gandolfo et al., 1998a, b) and others (e.g., Sims et al., 1999). There fossil pollen as well, due to fewer distinctive characters. A cogent example is additional and interesting potential for illuminating the fossil history of of this phenomenon is the early record of supposed monocot pollen, but dif- through the use of phylogenetic analysis as seen in Gandolfo et al. ®culties in pollen identi®cation extend to various modern dicotyledonous fam- (1997)Ðthat of assembling the dispersed organs of a taxon through their ilies. relative placements in phylogenetic context. One of the major problems with attempts to rationalize the fossil record Criteria for inclusion of fossils for determining the ages of modern clades with phylogenies based on modern taxa [either by various means of direct (i.e., through minimum age mapping)Ð(1) Fossil identi®cations are depen- comparison testing for congruence (e.g., Norell and Novacek, 1992; Benton, dent on either cladistic morphological analyses or the presence of well-doc- 2001)] or by using fossils as a basis for timing the rate of substitution in umented synapomorphies. (2) Accurate aging of fossil sites is required. We lineages and extrapolating that to other groups to assess the pattern on timing selected sites whose geology and ages are well known and are not contro- in angiosperms clades (Bremer, 2000) is the circumscription of the set of fossil versial and have interpreted the ages conservatively by assigning the youngest taxa that will be used in the analysis. Indeed, this selection dramatically alters of the suggested range of ages. (3) A high quality cladogram of modern taxa the outcome of any analysis, given that fossil reports of almost any existing is required. We used the widely accepted Soltis et al. (2000) three-gene 567- angiosperm taxon can be found in the paleobotanical literature. Many taxa taxon cladogram as the basis for mapping major clades of angiosperms. This have been misidenti®ed over the years as revealed by new analytical meth- cladogram was selected because it provides suf®cient detail and resolution to odologies and investigatory tools. In other cases, perfectly reasonable but too map the most important Cretaceous fossils of angiosperms that can be rea- general assessments of af®nities have been constrained by the nature of the sonably identi®ed. fossil evidence. Because selection of fossil taxa that are reliably identi®ed is critical to this analysis and to the utility of fossil evidence in related appli- Magnoliid groupsÐWe have found it most challenging to delimit the num- cations (e.g., Yoo et al., 2004), we have been conservative in selecting taxa ber of fossil taxa included in magnoliid groups (sensu Soltis et al., 2000, as to be included in determining the minimum ages of modern clades (Table 1). opposed to the more restrictive use of ``magnoliid'' in APG II, 2003). These Consequently, we restrict our analysis to taxa whose presence in the Creta- taxa are signi®cant because they are relevant to early branches in the angio- ceous is supported by the strongest fossil evidence. We de®ne the latter as sperms, thus the timing of their history rami®es signi®cantly throughout the having af®nities based upon the outcome of phylogenetic analysis that in- rest of the angiosperms. They are dif®cult because they include some of the cludes the fossils or alternatively, the presence of a suite of unequivocally earliest fossil evidence of the angiosperms, and these fossils are often imper- demonstrated fossil characters that uniquely matches the taxonomic unit to fectly or incompletely preserved, making them the most dif®cult group to which the fossil is assigned. identify and thus the most likely group to be inadequately identi®ed. The Fossils utilized in this paper are from two main sources: the Cornell Uni- monocots too are nested within the broadly de®ned magnoliid group (sensu versity Paleobotanical Collection (CUPC), which has been the source of nu- Soltis et al., 2000), and although many claims have been made regarding Early merous papers on Late Cretaceous ¯ower and fruit fossils from New Jersey Cretaceous monocot fossils, their early fossil record is generally problematical (Figs. 1±15), and the large number of macro- and mesofossils that have been (Gandolfo et al., 2000). Although the actual number of magnoliid/monocot described in the literature over the past 20±30 years. We have attempted a fossils that we include in this analysis based on our criteria is relatively low, thorough survey of published Cretaceous angiosperm ¯owers, and in each available evidence suggests that a large portion of the modern diversity within case where a putative identi®cation has been proposed, we have evaluated the the ``magnoliid'' families, including the basal monocots, had developed by nature of the evidence that supports that identi®cation. We have not made an about 90 million years ago (mya). There have been a remarkable number of effort to reanalyze these fossils in a cladistic context where none has been signi®cant fossil discoveries of mesofossils or compressions of ¯owers of offered in the original description, except in a few cases where they impinged Cretaceous age during the past 25 years (e.g., Dilcher and Crane, 1984; Crane upon our own work and demanded a transparent and objective analysis. We et al., 1993; Nixon and Crepet, 1993; Crepet and Nixon, 1994, 1998a, b; Friis have excluded a number of previously described Cretaceous fossils, because et al., 1997a, b). Even more recently, the number of reports of Early Creta- frequently, the original authors have themselves expressed doubts about the ceous angiosperm mesofossils of ¯owers has increased. We have been very possible af®nities of the fossils, and thus they were not identi®ed with cer- careful in accepting purported af®liations for the fossils using the criteria tainty in the original descriptions. Consistent with and possibly symptomatic delineated because it is our intention to make the dating as reliable as possible of these dif®culties, is that in most of these cases, the published fossils have (for example, all records of Early Cretaceous monocots have been excluded not been included in cladistic analyses, nor have there been concise presen- based on the limited characters available; Gandolfo et al., 2000). Several tations of synapomorphies or unique sets of features that would be needed reports of Early Cretaceous angiosperms that support fossil record congruence for identi®cation. Further complicating the literature is the fact that in some with phylogenies based on nucleic acid sequences in key aspects or that are cases, although particular fossils were originally presented as of ambiguous relevant in the timing of monocot history have not been included in our af®nity with possible af®nities to more than one taxon, subsequent authors, analysis and deserve additional discussion. without further analysis or explanation, later cited them as representative of a speci®c taxon, despite the reservations expressed by the original author(s). Included/excluded fossilsÐWe have included as many Cretaceous fossils This resulted in a trail of further citations that were based on little if any new as possible that conform with the criteria delineated (Table 1). Fossil reports evidence, and the ``identity'' of the tentatively identi®ed fossil then became have been excluded if, based on the available evidence (including caveats entrenched in the literature and a part of the conventional wisdom of the ®eld. raised by the original authors regarding their af®nities), their taxonomic place- Thus, it is sensible not to include such fossils in an analysis whose value ments are too ambiguous to provide minimum age estimates with suf®cient depends on the accuracy of the fossil identi®cations. We discuss our reasons con®dence. The sum of excluded ϩ included fossils is not exhaustive relative 1670 AMERICAN JOURNAL OF BOTANY [Vol. 91 Ènenberger et al., 2001 Cevallos-Ferriz, 1999 Sims et al., 1999 Drinnan et al., 1991 Friis et al., 1994 Gandolfo et al., 1998b Friis et al., 1986 Crepet and Nixon, 1998a Scho Nixon and Crepet, 1993 Nixon et al., 2001 Hernandez-Castillo and Zhou et al., 2001 Gandolfo et al., 1998a Hermsen et al., 2003 Friis, 1983 Friis, 1983 Knobloch and Mai, 1986 Drinnan et al., 1990 Herendeen et al., 1994 Herendeen and Crane, 1992 Crepet and Nixon, 1998b Crepet and Nixon, 1998b Friis et al., 1992 Friis et al., 2001 Gandolfo et al., inCrane press et al., 1993 Pedersen et al., 1994 Friis et al., 1988 Crane et al., 1993 Krassilov et al., 1983 Basinger and Dilcher, 1984 Crepet and Nixon, 1996 Stockey and Rothwell, 1997 Gandolfo et al., 2002 Gandolfo et al., 2002 ¯ower, fruit Flower, fruit Flower Flower Flower Stamen Flower Flower Flower Cupule In¯orescence In¯orescence, fruit Flower Flower, fruit Fruit Fruit Flower In¯orescence Flower Infructescence Flower Flower Flower Flower Flower In¯orescence In¯orescence, ¯ower In¯orescence In¯orescence In¯orescence Flower Flower Leaf, rhizome, root, Flower Flower USA USA USA USA USA USA Sweden USA USA Mexico USA USA USA Sweden Sweden Germany USA USA England USA USA Portugal Portugal USA USA USA USA USA Kazakhstan USA USA Canada USA USA 90 90 90 90 70 90 90 90 90 90 90 90 94 90 90 90 73±88 83±88 98±113 98±113 98±113 73±88 73±88 83±98 91±98 65±83 98±113 65±73 98±113 98±113 98±113 98±113 51±60 113±125 Santonian-Campanian Turonian Turonian Campanian Turonian Santonian Albian Albian Turonian Albian Turonian Turonian Turonian Santonian-Campanian Santonian-Campanian Cenomanian-Santonian Cenomanian Turonian Paleocene- Turonian Turonian Campanian-Maastrichtian Barremian-Aptian Turonian Albian Albian Albian Albian Albian Cenomanian Turonian Maastrichtian Turonian Turonian Platydiscus Paleoenkianthus Unnamed Tarahumara Microaltingia Bedellia Spanomera Virginianthus Dressiantha Unnamed Paleoclusia Tylerianthus Divisestylus Antiquocarya Caryanthus Budvaecarpus Mauldina Perseanthus Leguminocarpon Cronquisti¯ora Detrusandra Esqueiria Unnamed Microvictoria Auqia Hamatia Platananthus Platanocarpus Hyrcantha Unnamed Unnamed Trapago Mabelia Nuhliantha Order Genus Geologic Period Age (mya) Country Organ Reference Family or Cunoniaceae Ericaceae Fagaceae Haloragaceae Hamamelidaceae/Altingiaceae Betulaceae Buxaceae Calycanthaceae Capparales Chloranthaceae Clusiaceae Hydrangeaceae Iteaceae Juglandales Juglandales Juglandales Lauraceae Lauraceae Leguminosae/Fabaceae Magnoliales Magnoliales Myrtales Nymphaeales/Illiciaceales Nymphaeaceae Platanaceae Platanaceae Platanaceae Platanaceae Ranunculales Rhamnaceae Rosaceae Trapaceae Triuridaceae Triuridaceae 1. Fossils used to date the trees illustrated in Figs. 16±18. ABLE G H I J K A B C D E F L M N N N O O P Q Q R S T U U U U V W X Y Z Z T Node October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1671 to all known Cretaceous angiosperms, and in many cases where younger fos- Elsemaria and the possible polyphyletic nature of multilocular capsules make sils occur for a clade, we have neither accepted nor rejected these because, it dif®cult to attribute Elsemaria to a particular order or family'' (p. 133). in either case, they would not affect the estimated age of the nodes ancestral to them. We provide next a brief discussion of the most prominent of the Appomattoxia ancistrophora Friis et al. (1995)ÐEarly or Middle Albian. explicitly excluded fossils (predominantly fossils that have been assigned to Well-preserved fossilized fruits and pollen are known, but these have never the magnoliids/monocots). Note that it is quite possible that some of these been placed in a cladistic analysis. The pollen bears similarity to modern taxa do indeed have af®nities suggested for them in the literature (and in fact, Piperales, but the authors discuss other possible relationships and do not con- as will also be noted, our methodology has the potential of providing some clude with a de®nite taxonomic placement. The authors stated, ``Detailed eval- indirect support for these assignments), but the number of characters present uation of the phylogenetic position of Appomattoxia will require a much clear- in the fossils precludes de®nitive assignments by our criteria, thus it would er understanding of relationships among extant `basal' angiosperms than is be potentially misleading to include them in our analysis. Moreover, in some currently available, as well as additional information on the ¯oral and vege- cases, as will be noted, uncertainty about their af®nities is expressed in the tative structure in the fossil material'' (p. 294). original descriptions. Anacostia Friis et al. (1997b)ÐAlbian? Although the pollen found adhering Lesqueria elocata Crane and Dilcher (1984)ÐMid Cretaceous. This is a to fossilized unicarpellate fruits has some features usually associated with fruiting axis with magnoliid af®nities, but could represent a taxon that would monocotyledons (®ner reticulum at the ends of the grains), the combined be placed on a lower node of the cladogram. The authors do not place it in features of the fruits and pollen are not found in any extant monocotyledonous any particular family. They agreed that this fossil taxon has some characters family, and the authors leave open the possibility that this fossil is a magnoliid of the Magnoliidae; however, they pointed out that ``we know of no Recent (ϭdicotyledonous). Anacostia comprises four species, A. marylandensis, A. plant that has the combination of numerous helically arranged, dehiscent fol- virginiensis, A. portugallica, and A. teixeirae; all are new species. Explicitly, licles with bi®d tips borne in a swollen head at the apex of a long receptacle the authors remarked, ``The character combination in the fossil material, with bearing numerous, persistent, spirally arranged, laminar structures'' (p. 398). unicarpellate fruiting units containing a single anatropous seed and trichoto- mocolpate/monocolpate pollen, indicates a relationship between Anacostia Archaeanthus linnenbergeri Dilcher and Crane (1984)ÐUppermost Albian and extant Magnoliidae or monocotyledons'' (p. 241). They also pointed out, or lowermost Cenomanian. This also represents an infructescence of magno- ``The available characters for Anacostia are not suf®cient to assign the fossil liid af®nity but cannot be placed in any particular clade without further anal- material to any extant family. The weight of the evidence from the fruiting ysis. In the original paper, the authors concluded, ``Archaeanthus is clearly units and seeds, as well as the pollen indicates a relationship with the mag- most similar to extant Magnoliidae and share features with supposedly prim- noliids, but af®nity with extant monocotyledons cannot be ruled out'' (p. 242). itive members of the Hamamelidae and Dilleniidae. . . However, we do not Until the fossil is placed in a cladistic analysis, its taxonomic position cannot believe that Archaeanthus usefully can be assigned to any extant family; Ar- be evaluated. chaeanthus is a unique and extinct genus of fossil angiosperms'' (p. 378). Couperites Pedersen et al. (1991)ÐMid-Cretaceous. Although these fossils Prisca reynoldsii Retallack and Dilcher (1981)ÐMid Cretaceous. Another are likely chloranthaceous, they were characterized as incertae sedis by the fructi®cation is, Prisca reynoldsii, a raceme of unisexual multifollicles. The authors and need to be evaluated in a cladistic context. In the original paper, authors did not place these fossils within any modern family and suggested the authors presented doubts about where this fossil taxon should be placed the erection of the family Priscaceae in order to place the fossils (``Subclass taxonomically as indicated by the following quotation: ``The chloranthaceous and order incertae sedis, Family Priscaceae fam. nov. This family is intended af®nity of Couperites is also strongly supported by similarities in fruit and for fructi®cations like Prisca reynoldsii and other organs, such as leaves, seed structure. However, differences in seed organization preclude unequiv- pollen organs, staminate in¯orescences and woods, which appear related to ocal assignment of Couperites to the family (table 1)'' (p. 588). Couperites and distinctive of plants with such fructi®cations,'' p. 107). Later, Drinnan et is compared by the authors with selected families of the Laurales, Magnoliales al. (1990) tentatively referred Prisca to Lauraceae, based on gross similarities and Piperales (table 1, p. 587). to the fossil Mauldinia, but provided no evidence of diagnostic features of Lauraceae in Prisca, nor did they provide any synapomorphies or cladistic Silvanthemum Friis (1990)ÐLate Cretaceous, Santonian±Campanian. Sil- analysis. vanthemum was originally considered to have strong af®nities with Escallon- iaceae, but that placement was not veri®ed by subsequent analyses (Backlund, Protomonimia kasai-nakajhongii Nishida and Nishida (1988)ÐTuronian 1996). Because the reanalysis did not include all taxa included here and there (Late Cretaceous). This is a reproductive structure similar to Archaeanthus are signi®cant differences in the topologies, it is not possible to place this and Lesqueria. However, the authors did not place it within any family and fossil de®nitively. Evaluation in a broader, more complete analysis is neces- considered that Protomonimia represents a magnolioid ancestor from which sary. In the original paper, Friis (1990, p.15) pointed out, ``The great similarity some modern Magnoliales may have been derived. In the diagnosis, the au- between Silvianthemum and modern Escalloniaceae further suggests that the thors said that it is an ``Angiosperm fructi®cation consisting of a globose head woody taxa of the saxifragalean complex constitute an ancient stock,'' and and a woody peduncle; apocarpous and apistillate.'' They also pointed out, also mentioned, ``the close relationship to members of the Escalloniaceae, ``Judging from the polycarpelous ¯oral structure, presence of oil cells and particularly to species of Quintinia, is documented not only by morphological relatively primitive wood structure composed mainly of scalariform vessels and organizational similarities, but also by anatomical features including cel- with steep scalariform perforation plates, the specimen can be related to the lular details of the trichomes'' (p. 15). However, when Silvianthemum was Magnoliales (sensu Cronquist, 1968)'' (p. 419), and ``. . . the direct connec- included in a phylogenetic analysis for the Dipsacales, it did not appear close- tion between Protomonimia and Monimiaceae is still based more on specu- ly related to Quintinia (Backlund, 1996). Backlund (1996) clearly stated, lation than on unequivocal evidence'' (p. 422). This fossil requires a cladistic ``Analysis of the matrix shows that Silvianthemum occupies a stable but not placement prior to inclusion for dating nodes. strongly supported position as somewhat more advanced than the extant Quin- tinia (see ®g. 3), and just outside the basal node of the Dipsacales'' (p. 18). Elsemaria kokubunii H. Nishida (1994)ÐConiacian±Santonian. An angio- sperm permineralized fruit that has af®nities to some extant taxa formerly NymphaeaceaeÐThe putative key phylogenetic position within angio- placed in the Dilleniidae, but the lack of stamens and other ¯oral parts prevent sperms of the Nymphaeales and Amborella, as a sister group or sister groups placement within a family. It is in need of cladistic reevaluation. Nishida of the remainder of angiosperms in DNA-based hypotheses of angiosperm presented the following systematics: Subclass Dilleniidae, Order and Family phylogeny including the 567-taxon three-gene tree (Soltis et al., 2000) and incertae sedis, and indicated that ``However, the lack of other ¯oral parts in others (e.g., Parkinson et al., 1999; Chaw et al., 2000), makes the fossil record 1672 AMERICAN JOURNAL OF BOTANY [Vol. 91

Figs. 1±9. Charcoali®ed fossil ¯owers from Sayreville, Raritan Formation (Turonian, Upper Cretaceous), New Jersey, USA. 1. Divisestylus brevistamineus Hermsen, Gandolfo, Nixon and Crepet, based on cladistic analyses, this taxon has close af®nities to the modern family Iteaceae. 80ϫ. 2. Detrusandra mystagoga Crepet and Nixon, a magnoliid ¯ower has a unique combination of characters relative to the extant magnoliid ¯owers. 20ϫ. 3. Paleoclusia chevalieri Crepet and Nixon, placed parsimoniously within the family Clusiaceae, Paleoclusia is well nested within the subfamily Clusioideae. 64ϫ. 4. Pentamerous ¯ower with af®nities to the , 51ϫ. 5. Microvictoria svitkoana Gandolfo, Nixon and Crepet, a fossil member of the family Nymphaeaceae share characters with the modern genus Victoria, cladistic analysis placed the fossil within the family. 33ϫ. 6. Perseanthus crossmanensis Herendeen, Nixon and Crepet, a complete October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1673

Figs. 10±12. Charcoali®ed fossil ¯owers from Sayreville, Raritan Formation (Turonian, Upper Cretaceous), New Jersey, USA. 10. Tylerianthus crossmanensis Gandolfo, Nixon and Crepet, fossil taxon with af®nities to the family Hydrangeaceae. 79ϫ. 11. Microaltingia apocarpela Zhou, Crepet and Nixon. This fossil taxon, with characters of the Hamamelidaceae, is placed within the subfamily Altingioideae. 28ϫ. 12. Paleoenkianthus sayrevillensis Nixon and Crepet share a mosaic of characters found in modern Ericaceae. 73ϫ. of Nymphaeales potentially important to understanding angiosperm origins taxon of modern angiosperms is thus ultimately reinforced, not refuted, by a and the nature of the earliest ¯owering plants. Thus, the occurrence of early careful examination of the Friis et al. (2003) analysis. While Friis et al. (2003) nymphaeaceous fossils is critical in dating the lowest nodes in the angiosperm appealed to the possibility that any morphologically simple taxon could owe phylogeny, so it is important for the date to be based on unequivocal fossil its structure to evolutionary reduction (and thus to the notion that relationships evidence. of structurally simple aquatics can at times be dif®cult to determine based on Archaefructus, an Early Cretaceous angiosperm from China, has been in- morphology alone), there is absolutely no evidence to support this possibility. terpreted both as sister group to the rest of the angiosperms (Sun et al., 1998, The fossil is indeed unusual relative to modern taxa and thus remains enig- 2002) or as an aberrant member of the Nymphaeales, morphologically distant matic, but even with available state-of-the-art techniques for identi®cation, from any modern member of that group (Friis et al., 2003). The latter inter- analysis only provides one result. Further insights into the af®nities or support pretation is based on a phylogenetic reanalysis of the original matrix of Sun for the phylogenetic position indicated by analysis of the morphological char- et al. (2002) that includes an additional taxon (Cabombaceae) and has several acters now available from the fossils will most likely come from additional changed and unconventional character codings. As in the Sun et al. analysis, fossil evidence and its analysis. In addition, placement within the Nymphae- the results of Friis et al. (2003) place Archaefructus as the sister group to the ales as suggested by Friis et al. requires a number of ad hoc hypotheses to remainder of the angiosperms in some of the most parsimonious trees, but explain the dramatic morphological contrast between Archaefructus and ¯ow- Archaefructus is placed with Cabomba in the Nymphaeales in other trees of ers of modern Nymphaeales, but such speculation is not necessary given the the same length. We are skeptical of the latter interpretation because it was results of the reanalysis. Thus, based on the best available evidence, we do executed with one empirically veri®able character miscoding (in their revised not accept the notion that Archaefructus is an extinct nymphaealean, consis- matrix, Friis et al. coded Cabomba as having only dichotomously veined tent with our criterion that it shares no identi®able synapomorphies with any dissected leaves, while in fact it also has entire ¯oating leaves with numerous extant member of that clade. clearly anastomosing veins). When the Friis et al. matrix is modi®ed to re¯ect Another report of fossil Nymphaeaceae (sensu Les et al., 1999) from the the actual variation in leaf form within Cabomba by changing only the single Early Cretaceous (approximately 100±125 mya) sediments in Portugal (Friis matrix cell from dichotomous to polymorphic, Archaefructus again becomes et al., 2001), while not de®nitive, is still signi®cant in dating angiosperm a sister group to the angiosperms in 100% of the shortest trees and is never history. The actual character suite demonstrated in the fossil is, upon phylo- placed with Cabomba or in the Nymphaeales, thus conforming with the results genetic reanalysis that includes additional relevant taxa (Gandolfo et al., of Sun et al. (2002). This restoration of the outcome of the original analysis 2004), equally compatible with Illiciaceae as well as with other angiosperm by Sun et al. (2002) has been achieved by making only that one change even families. The incomplete preservation of the fossil in combination with the though other matrix modi®cations by Friis et al. (2003), including a somewhat absence of de®nitive synapomorphies of Nymphaeaceae argue against its pre- unorthodox coding of a perianth as possibly present in Ginkgo (which has cise placement in any particular clade of angiosperms. We have re¯ected this naked stalked ovules intermixed with leaves), were left in the matrix. The ambiguity by including it as an estimator of the youngest node that subtends resultant stability of the phylogenetic position of Archaefructus as a sister both Nymphaeaceae and Illiciaceae in the three-gene tree, aged at 113 mya

←

Figs. 1±9. Continued. ¯ower of the fossil taxon found after the initial publication of the taxon. 25ϫ. 7. Mabelia archaia Gandolfo, Nixon and Crepet, the oldest monocot ¯ower in the fossil record; this taxon appear well nested within the subfamily Triurideae (Triuridaceae) based on cladistic analyses of all the modern and fossil genera of the family. 40ϫ. 8. Male fagoid ¯ower. 39ϫ. 9. Dressiantha bicarpellata Gandolfo, Nixon and Crepet, placed in a phylogenetic context within the capparalean clade as suggested by Rodman, it is the oldest record for the order Capparales sensu Cronquist. 60ϫ. 1674 AMERICAN JOURNAL OF BOTANY [Vol. 91

monocotyledonous (Gandolfo et al., 2000). There are relatively few reports of Upper Cretaceous monocots that conform to the criteria we employ for this study. There have been several reviews of the fossil record of the monocot- yledons (Horwood, 1912 a, b; Doyle, 1973; Daghlian, 1981; Muller, 1981; Herendeen and Crane, 1995; Gandolfo et al., 2000), and in each case, selected fossils have been helpfully discussed. However, as mentioned in all of these reviews, it is sometimes dif®cult to recognize fossils with monocotyledonous af®nities, particularly, we now realize, due to a frequent lack of monocot synapomorphies in supposed monocot fossils. Without such synapomorphies, it is impossible to distinguish some monocots from a number of magnoliid dicots, compromising the early fossil record of monocots and defeating the purpose and utility of a review of the early monocot record. Thus, as with other fossil evidence used in this analysis, we excluded reports of fossil taxa that could not be unequivocally related to monocotyledons based on the pres- ence of diagnostic synapomorphies. For example, in the latest review by Gan- dolfo et al. (2000), the genera Liliacidites Couper (pollen taxon) and Aca- ciaephyllum Fontaine (leaves) were included in a cladistic analysis because these two fossil taxa had previously been considered to be the oldest examples of monocots. However, both fossils actually lack monocot synapomorphies, and the most parsimonious placement of these purported early monocotyle- donous taxa was among dicot clades. Recently, Bremer (2000) drew attention to a subset of Cretaceous fossils with purported monocotyledonous af®nities by using them to date nodes on a plastid rbcL-based phylogeny of monocot taxa, providing the means for calculating the rate of change (base substitution rate) in particular lineages. The results of such a combined approach, while potentially informative, de- pend on the quality of the identi®cation of the fossils. While it was certainly appropriate for Bremer to choose relevant fossil evidence from the paleobo- tanical literature, we are using a different and more stringent set of standards for fossil identi®cation and do not include the same fossil taxa in this analysis for reasons discussed next. Dicolpopollis P¯anzl was used to provide the minimal age for the family To®eldiaceae (Chmura, 1973) and Milfordia Erdtman for the age of the Res- toniaceae-Flagellariaceae-Joinvillaceae complex. These assignments, however, were made in the absence of comparative ultrastructural data necessary to de®nitively identify these taxa on the basis of palynological evidence alone (Zavada, 1984; Harley and Zavada, 2000). As with many other fossil assignments that have not been included in the present analysis, lack of synapomorphic characters does not necessarily pre- clude the possibility that the suggested af®nities may be accurate (as in the suggested af®nity of Dicolpopollis to To®eldiaceae), but it would be prema- ture and potentially misleading to use such unsupported identi®cations for dating particular clades. Indeed, in the absence of additional characters, cred- ibility of these identi®cations can ultimately be assessed best in the context of a temporal framework provided by the methodology used herein. Another very important complex of monocots has been placed temporally on the basis of the fossil palynomorph Milfordia, but Linder (1986) indicated Figs. 13±15. Reconstructions of three fossil ¯owers. 13. Paleoclusia chev- alieri Crepet and Nixon. 14. Microvictoria svitkoana Gandolfo, Nixon and that there are three pollen morphs of Maastrichtian- age related to Crepet. 15. Mabelia connati®la Gandolfo, Nixon and Crepet. Reconstructions Restoniaceae and Poaceae. They are Monoporites annulatus Hammen, Mil- done by Michael Rothman. fordia hungarica (Kedves) W. Kr., and Milfordia incertae (Th. & Pf.) W. Kr. None are adequately characterized (they lack comparative ultrastructural fea- tures); thus it is dif®cult and complicated to determine their af®nities with (Fig. 16). Because these two clades are sister taxa of the remainder of angio- certainty. sperms (excluding Amborella), it provides the oldest record of the extant clade, or ``crown group'' of angiosperms. The most conclusive, earliest fossil Pistia corrugata LesquereuxÐThe oldest con®rmed fossil record of P. cor- evidence for Nymphaeaceae fossil ¯owers is from the Turonian (90 mya) of rugata dates from the Paleocene of southwestern Saskatchewan, Canada the USA, and these fossils have ¯oral structures almost identical to those of (McIver and Basinger, 1993); earlier fossils assigned to Pistia are suspect modern Victoria (Nymphaeaceae). These fossil ¯owers have been placed as (Hickey, 1991). unequivocally nested within modern Nymphaeaceae by explicit phylogenetic analysis (Gandolfo et al., 2004). Cymodocea KoenigÐThe record of Cymodocea presented by Bremer is MonocotyledonsÐOther Cretaceous taxa in the ``magnoliid'' clade that mo- based on the fossil Thalassocharis bosqueti Debey ex Miquel collected from lecular (and other) evidence indicates participated in the early radiation of Upper Maastrichtian sediments of the Kunrade region, Netherlands (Voigt and angiosperms are represented by the monocotyledons (Soltis et al., 2000). Domke, 1955). However, Voigt (1981, p. 283) notes that ``Thalassocharis Monocots have a traditionally sparse Cretaceous fossil record, and recent crit- seems to show some af®nities with the recent genus Cymodocea but because ical analysis of Lower Cretaceous fossils suggests that none are de®nitively ¯owers and fructi®cations are not known the exact phylogenetic and system- October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1675

Fig. 16. Minimum ages (in millions of years) for major ``basal'' lineages of angiosperms (113±90 million years ago) based on the best available me- sofossil evidence mapped onto the three-gene 567-angiosperm tree. The tree has been collapsed to show major clades of interest. Numbers in brackets following a name indicate the number of distinct sequences that were used in the original molecular analysis.

atic position of Thalassocharis is still uncertain,'' a reservation shared by Fig. 17. Minimum ages (in millions of years) for major lineages within Daghlian (1981). the tricolpate or ``Eudicot'' clade (some collateral groups not shown, such as Platanaceae, Buxaceae; see Fig. 1).

Typha L.ÐBremer (2000) used the fossil seeds T. ochreacea Knobloch and Mai and T. protogaea Knobloch and Mai, from the Maastrichtian of Eisleben, MethodsÐFossil data from both the literature and unpublished sources (Knobloch and Mai, 1986) to date the node for bur-weeds and cat-tails. How- were entered into a relational database structure, and the program Winclada ever, these fossils have no unique characters to unequivocally relate them to (Nixon, 2003) was utilized to automatically map minimum ages for each node the extant genus Typha and have not been evaluated in a cladistic analysis. on the consensus of the 567-taxon three-gene tree (Soltis et al., 2000). The As in the cases of the putatively monocotyledonous palynomorphs, the procedure for optimizing node ages was simple: each node was assigned the preceding assessments do not exclude the possibility that any or all of these greatest minimum age of any node that had been assigned above it. Ages for early fossils are monocots or even that they belong to the groups used to modern taxa with no fossil record were initially assigned an age of zero. support Bremer's analysis, but they have not been shown to have suf®cient Fossils were positioned on the 567-taxon tree according to their most likely characters to justify such speci®c taxonomic assignments. Until the relation- identi®cation based on assignment either by published cladistic analyses or ships are con®rmed, they cannot reliably be used as a basis for estimating the the occurrence of de®nite synapomorphies and/or of unique character com- ages of monocot clades. binations. Thus by default, the oldest unequivocal evidence for fossil monocots is The resulting cladogram with nodes assigned minimum ages was then re- embodied in charcoali®ed fossil ¯owers from the Turonian of New Jersey that duced by collapsing nodes to show a more tractable view, emphasizing major bear a combination of features found today only in the achlorophyllous Triur- clades for which fossil evidence is available. Estimated minimum ages are idaceae. These fossil ¯owers were described and placed phylogenetically in only mapped in these trees for nodes that have at least one fossil that can be extensive detail (but after the publication of Bremer, 2000), and their rela- associated with at least one descendant branch (0 age estimates are not tionships within the monocots have been well established through parsimony- shown). based phylogenetic analysis (Gandolfo et al., 2002). Based on these analyses, they are well nested within the Triuridaceae as the sister group to the clade RESULTS AND DISCUSSION that encompasses the tribe Triurideae. Although these ¯owers constitute the earliest unequivocal record for monocots, the group is well nested within the The reduced cladogram with minimum ages is presented in monocot clade, and there is no reason to assume that triurids are morpholog- Figs. 16±18. This cladogram has been reduced to only 28 ter- ically similar to the basal monocot condition. The lack of other veri®ed minals. Terminals have been assigned to family according to mesoÐand macrofossils of monocots from similarly aged or older Cretaceous the classi®cation of Takhtajan (1997). sediments is likely due to preservational bias, assuming that early monocots Improved understanding of angiosperm relationships made were herbaceous, as discussed later. The Late Cretaceous also includes several possible by expanded molecular data sets and cladograms have fossils of Maastrichtian (Upper Cretaceous) age with unmistakable monocot by themselves clari®ed our understanding of trends in angio- synapomorphies (Hickey and Petersen, 1978; Daghlian, 1981; Herendeen and sperm evolution and in the evolution of ¯ower morphology. Crane, 1995; Harley, 1996; Herendeen et al., 1999). These will be analyzed Assessing the implications of these relationships will occupy in greater detail in another paper (Gandolfo et al., unpublished manuscript). considerable effort in the future. When the dimension of time 1676 AMERICAN JOURNAL OF BOTANY [Vol. 91

timing in angiosperm diversi®cation. Thus for example, al- though our analysis illustrates that the node that precedes the divergence of Nymphaeaceae and Illiciaceae is at a minimum 113 million years old, this is not based on a model of gene substitution, nor does it suppose an unbiased sample. This es- timate is dependent only on the identi®cation of the fossil that is the basis for the estimate and of course the veracity of the phylogeny. The estimates provided here are consistent with a major diversi®cation of several important angiosperm clades by the Turonian (ϳ90 mya) that included numerous magnoliids, the monocots, and several tricolpate clades. Extensive Early Cre- taceous monocot diversi®cation is suggested by the presence of Triuridaceae ¯oral evidence in the Turonian, but this diver- si®cation is not re¯ected in the mesofossil or megafossil re- cords, most likely due to preservational bias. Monocot fossil pollen reports from the pre-Turonian are all based on charac- ters that are mostly not exclusive to monocots or on the single feature of gradation in the size of the reticulum of the pollen wall (also found on some fossil that may best be placed outside the monocots based on attached organs see Couperites; Pedersen et al., 1991). While it is likely that some of these early Cretaceous palynomorphs do represent monocots, veri- ®cation must be deferred until diagnostic leaf and ¯oral ma- terials are available. We must face the possibility that for cer- tain herbaceous groups (probably including early monocots), the likelihood of preservation is so low that we may never have unequivocal early examples of these clades, and our di- rect estimates will remain conservative. The occurrence of platanoid fossils (leaves, ¯owers, fruits, and pollen), among the oldest representatives of the ``tricol- pate'' clade (or ``'' of some authors) at about 100± 108 mya (we always use the youngest age of a range in this paper) may seem puzzling at ®rst, given the supposed ``de- rived'' nature of this group. However, as can be seen in Fig. Fig. 18. Minimum ages (in millions of years) of clades within the ``Rosid 16, the age of the platanoids is actually consistent with the I'' (sensu APG II, 2003) clade. As for Fig. 1. minimum age estimate for its sister group (the remainder of tricolpates at 98 mya). Because of the highly resolved structure of the phylogenetic is added through careful dating using rigorously identi®ed fos- tree used in this analysis, minimum ages for the ``stem sil evidence, the fossil record is more informative, allowing groups'' of some families that lack a fossil record can still be major events in the evolution and radiation of the ¯owering calculated, and the minimum ages for the stem lineages of plants to be placed in a temporal perspective. Precise assess- some families are calculated as greater than the ages of the ment of timing also permits evaluation of various ecological oldest known fossils for the particular families. This is the evolutionary hypotheses. The method we have employed pre- case, for instance, with Sabiaceae, which has a calculated min- sents the minimal time of divergence of major angiosperm imum age of 98 mya for its divergence from other extant an- clades, allowing a more accurate appraisal of evolutionary giosperms, based on fossils assignable to other families (Fig. rates within and among lineages than do similar estimates 16), even though the oldest putative fossils known for Sabi- based on inadequate fossil identi®cation and avoids the dif®- aceae, seeds from the Late Cretaceous, are at least 27 million culties and complexities attendant to the ``molecular clock'' or years younger (Knobloch and Mai, 1986; used by MagalloÂn- substitution-rate modeling approaches and more complex at- Puebla et al., 1999). It should be noted that the estimated age tempts to use molecular data to estimate timing in lineage of the minimal clade that includes Sabiaceae does not imply history (e.g., Sanderson, 1997; Bremer, 2000). It is not a meth- that the ``crown group'' or modern Sabiaceae had diverged by od per se for evaluating congruence of the hypothesis of phy- this time, but that the divergence of any Sabiaceae from any logeny of angiosperms and the fossil record; nonetheless, it other extant angiosperm not in the Sabiaceae (which includes does place the phylogeny in a temporal framework and does all other extant angiosperms) occurred at least 98 mya. The have the potential to reveal inconsistencies between the fossil modern family may have diverged much more recently. Al- record and the best estimates of phylogenetic relationships though MagalloÂn-Puebla et al. (1999) aged modern Sabiaceae based on comparative studies of nucleic acid sequences of liv- at 67.5 mya, this conclusion is not supported by cladistic anal- ing taxa. It also allows comparison of the temporal framework ysis of the fossil nor by synapomorphies, and we have not implied by the technique with the background fossil record of included it in our tree. If correct, the ®rst known occurrence organs that are inconclusively assignable to particular taxa but of sabiaceous fossils more than 20 million years after the es- that may give broader indications of the general pattern of timated divergence of the group is neither consistent nor in- October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1677 consistent with the proposed molecular phylogeny (Soltis et angiosperm species not in the same family (based on this tree). al., 2000), because the perceived hiatus may be explained by Likewise, a minimum age of 90 mya for the divergence of the numerous factors, including lack of suf®cient diversity in the stem group of Malpighiaceae can be estimated based on the clade for the ®rst 27 million years, lack of identi®able features extremely well-known Clusiaceae fossil from the Turonian of of modern Sabiaceae in the early members of the clade, pres- New Jersey (Crepet and Nixon, 1998b). ervational bias due to morphology, mode of pollination or eco- Another important clade within the ``Rosid I'' group in- logical distribution, or random events that might affect the cludes Rosaceae, Fabaceae, Urticales, Cucurbitales, and the preservation of fossils. We see no way to tease apart such families formerly included in the ``Higher Hamamelididae'' explanations and thus view with some skepticism the idea that (e.g., Fagaceae, Nothofagaceae, Juglandaceae, Myricaceae, any realistic or useful models can be applied to the supposed Betulaceae, Casuarinaceae). This clade is of great interest in ``®t'' of fossil data to phylogenetic trees. part because there are no obvious morphological features that The MagalloÂn-Puebla et al. (1999) paper is distinguished by unite this clade, and it includes a diverse assemblage of eco- its inclusion of numerous fossils in addition to fossil Sabiaceae nomically important taxa with an exceptional fossil record. and as observable before, it illustrates how selection of fossils The minimum age of the stem group of the clade that includes and criteria used in fossil identi®cation can affect the outcomes Rosaceae and Fagales but excludes Fabaceae and Polygalaceae of such analyses. The results of their analysis differ from our is estimated at 94 mya, while the ®rst identi®able fossil of own in other ways partly and perhaps most substantively due Fabaceae is from the Tertiary, ca. 51 mya (Herendeen and to different fossil selection. Employing criteria for including Crane, 1992), or at least 40 million years after the minimum fossil evidence that we apply would eliminate some of the estimated divergence for the clade. It should be noted in this fossils utilized in the MagalloÂn-Puebla et al. analysis. For ex- context that strongly zygomorphic ¯owers are not present in ample, the age of the Trochodendrales is based on leaves as- the best-known Late Cretaceous deposits, and the radiation of signed to the modern genus Tetracentron (Tetracentraceae); both Polygalaceae and zygomorphic Fabaceae (e.g., papilion- however, Collinson et al. (1993, the reference used by Ma- oid and caesalpinoid ¯owers) may have been a Tertiary event, galloÂn-Puebla et al.) explicitly stated that ``Con®rmation of a possibly in response to the increased availability of specialized fossil record for the Tetracentraceae must await complete re- hymenopteran pollinators (apparently available at some level vision of all the leaf fossils and recovery of appropriate re- since Turonian times; Crepet, 2000). It is likely that members productive material'' (p. 835), and we would not have included of the ``stem group'' subtending Polygalaceae-Fabaceae were it based on the criteria discussed previously. There are other actinomorphic and closely resembled modern Rosaceae, and leaves assigned to the family Trochodendraceae, Trochoden- thus identi®cation of such taxa would be problematic without droides rhomboideus (Lesquereux) E. W. Berry 1922, from extensive characters for evaluation, typically unavailable for sediments of upper Cretaceous age; however, these leaves are fossils. This provides yet another example of how disparity poorly preserved and lack diagnostic characters. However, (incongruence) in the fossil record should not be the basis for such differences in approaches to fossil selection are not the acceptance or rejection of particular phylogenetic trees and of only reasons for different results. We use more conservative how gaps in the fossil record, once characterized through this dating of clade ages in angiosperms (Fig. 17). For example, methodology, do not necessarily detract from the overall in- Esgueiria adenocarpa Friis, Pedersen and Crane from the Late formative value of fossil evidence that does exist. In a sense, Cretaceous (Campanian-Maastrichtian, 85±65 mya) was used the utility of the fossil record is enhanced by a methodology in both analyses; however, we selected the youngest date (as that points to clear gaps in a context that helps to understand we have throughout), while MagalloÂn-Puebla et al. used the them and thus, to understand the rest of the fossil record and older one (85 mya). In addition, when possible, we use the its imperfections and to validate its utility in assessing various exact ages as proposed by the original author(s) for each spe- aspects of angiosperm evolution. ci®c taxon. For example, the fossil ¯ower Spanomera was The extensive record of the Normapolles palynomorphs in used in both analyses for dating the Buxaceae; however, Ma- the Late Cretaceous (e.g., Christopher, 1979; SchoÈnenberger galloÂn-Puebla et al. dated it as mid-Albian (104.5 mya), while et al., 2001) seems to re¯ect diversi®cation of the Juglandales we followed the original authors suggestion of a Late Albian lineage, which is well nested within the group that has recently age (ϳ99 mya; Drinnan et al., 1991). There has been only one been dubbed ``Fagales'' (sensu lato) by the APG II (2003). additional effort to assess clade ages over an entire angiosperm The complexity of assigning phylogenetic positions to the ma- phylogeny (Davies et al., 2004), and we will compare their jority of Normapolles types is seen in the various taxonomic results and those of MagalloÂn-Puebla et al. to our ®ndings in placements that have been proposed, including various ``High- more depth in a subsequent manuscript. er Hamaelididae'' families with pororate apertures and gran- ular wall structure, as well as in some cases to Urticales. How- ``Rosid I'' cladeÐThe rosid clade that includes Rosaceae, ever, there seems little doubt that at least some of the Nor- Malpighiaceae and a host of other families traditionally placed mapolles palynomorphs are correctly placed in the broad clade in either the Rosidae or the ``Higher Hamamelididae'' is often that includes Juglandaceae, Betulaceae, and Myricaceae (Table referred to as ``Rosid I'' and is shown in some detail in Fig. 1), and we have assigned the minimum age of that clade ac- 18. The existence of only two ®rmly identi®ed fossils in the cordingly at 83 mya. Mesofossils and megafossils identi®able lower part of this subtree (Cunoniaceae, Clusiaceae), never- to particular families do not occur until later, with the excep- theless, allows estimating the minimum divergence times for tion of cupulate Fagaceae from the Turonian of New Jersey the stem lineages of several families, due to the pectinate to- (Crepet et al., 2001; Nixon et al., 2001). pology. The divergence of the stem groups for Oxalidaceae, Notably absent are fossils of the Cucurbitales clade, which Elaeocarpaceae s. s., Eucryphiaceae, and Cunoniaceae can be is mostly herbaceous, and for which the ``stem group'' should estimated at 73 mya. Thus any modern species in those fam- have been present at least by 90 mya. The pattern within these ilies has a minimum divergence of 73 mya from any other clades nicely illustrates the probable effect of depositional, 1678 AMERICAN JOURNAL OF BOTANY [Vol. 91 preservational, and ecological bias in the fossil record. The morphic and/or bilabiate ¯owers at this time suggests that ¯o- largely herbaceous and insect-pollinated Cucurbitales is essen- ral zygomorphy is an ultimate re®nement in pollination syn- tially without a fossil record [with the exception of Tertiary drome that followed the development of other features. It is fossils of the woody genus Tetrameles (Datiscaceae); Lakhan- worth noting that the groups that later developed zygomorphy pal and Verma, 1966], while the concomitant ``Fagales'' (probably almost entirely in the Tertiary, based on current ev- group, which is entirely woody and mostly wind-pollinated idence) have become some of the most diverse and successful has one of the most extensive fossil records for angiosperms, clades of modern angiosperms, including a large portion of the in terms of leaf, pollen, and reproductive structures, beginning ``asterids'' (I and II) as well as the papilionoid and caesalpi- in the Late Cretaceous and extending throughout the Tertiary noid legumes. Just as striking, however, is the success and (Crepet and Nixon, 1989a, b). Another factor that might con- persistence of numerous clades in which zygomorphic ¯oral tribute to the abundance of fossil Fagaceae and other members presentation is absent or rare and which generally have less of the woody ``Fagales'' clade is their tendency to form ex- stringent pollinator speci®city (or passive pollination syn- tensive dominant forests (or riparian forests), while the Cu- dromes), such as the majority of rosid and ranunculid lineages, curbitales (e.g., Cucurbitaceae and Begoniaceae in particular) including the mimosoid legumes. are generally scattered individuals with far less cumulative biomass. Overall pattern implicationsÐThe overall pattern seen in the estimated times of divergence of major angiosperm clades Ericanae (``Asterid III'') and ``Asterid IV''ÐThe Ericales in the Cretaceous suggests either a rapid diversi®cation of are well represented in the Turonian of New Jersey (Nixon groups between 113 and 80 mya, or alternatively, an earlier and Crepet, 1993; Crepet, 2000), providing an estimate of 90 diversi®cation with an extremely poor fossil record for these mya for the minimum divergence of that group, as well as the groups. In some cases, such as that of the platanoids, we be- broader group known as ``Asterid III'' in recent works. Be- lieve that it is likely that the fossil record is ``good'' (i.e., cause both Asterid III and Asterid IV are made up of families re¯ective of reality) because the modern equivalents of these mostly not originally placed in the Asteridae, and mostly lack- taxa occupy prime habitats for fossil deposition, they are large ing fused corollas, we prefer the name ``ERICANAE'' for the deciduous trees with extensive populations, they are wind-pol- former group, which is much more descriptive. The ``Asterid linated and produce numerous ¯owers, and their leaves are IV'' group includes Hydrangeales, for which a well-estab- large and coriaceous. Given that the minimum ages provided lished fossil occurs in the Turonian of New Jersey, again pro- by platanoids are relatively consistent with those of the sister viding an estimate of ca. 90 mya for the divergence of this group of the platanoids (which includes numerous other tri- group. This is consistent with MagalloÂn-Puebla et al. (1999), colpates), we see no reason to assume a much greater age for who estimated the minimum age at 89.5 mya based on the the divergences at that level in the phylogeny. The greater same fossils. question that arises is whether identi®able fossils that might be placed on the ``stem lineage'' below angiosperms but above Asterids (``Asterid I'' and ``Asterid II'')ÐThe traditional their divergence with other gymnosperms might be identi®ed Asteridae of Cronquist (1981), with typically gamopetalous, from existing fossil materials and while controversial, based often tubular and/or zygomorphic ¯owers, are mainly found on all available evidence and no doubt subject to validation in what have recently been called the ``Asterid I'' and ``Asterid from future fossil evidence, it appears that Archaefructus may II'' clades and lack a fossil record in the Cretaceous. This be the ®rst such fossil angiosperm. includes Asteraceae, Solanaceae, Lamiaceae, Rubiaceae, and a host of other families that together contain an enormous spe- CONCLUSIONS cies diversity, particularly in the tropics. As is the case with the zygomorphic legumes, the diversi®cation of these asterids, This paper represents an attempt to comprehensively age the with the associated features of tubular ¯owers, extensive zy- major Cretaceous radiations of angiosperms. The method pro- gomorphy, and various often highly specialized pollination vides a minimum age estimation without models, and thus is syndromes, very likely took place in the Tertiary. Alternative- a conservative estimate of divergence times. Given that there ly, if the early asterids were herbaceous, in low population are clear and inherent biases in the fossil record of ¯owering densities, and insect-pollinated, they may have had an earlier plants, the consistency of estimation in many cases is actually diversi®cation but have been relatively ``invisible'' in the fos- striking (e.g., independent estimates of 98 vs. 100 mya for the sil record. Nonetheless, the apparently parallel radiation of le- common ancestor of the magnoliids/monocots and tricolpates gumes, a group with generally similarly specialized pollina- in Fig. 16). While it is likely that most of these estimates will tors, is consistent with a radiation of asterids in the Tertiary probably be extended further back in time as new fossils are (e.g., Herendeen and Crane, 1992), possibly for similar rea- discovered, there are several groups (such as the ``platanoids'') sons. that are probably well represented in the fossil record. The As pointed out in previous papers, various complex polli- estimated ages for these clades are unlikely to increase signif- nation syndromes likely were present by 90 mya as evidenced icantly. Groups that have inherently low probabilities of being by the diversity of Turonian ¯oral morphologies (Nixon and preservedÐsuch as herbaceous, insect pollinated cladesÐare Crepet, 1993; Crepet and Nixon, 1998b; Gandolfo et al., the most inconsistent (i.e., providing signi®cantly younger es- 1998a, b, 2004; Crepet, 2000). These include features that im- timates of divergence than their woody, wind-pollinated sister ply high pollinator speci®city, such as viscin threads (Nixon groups; e.g., Cucurbitales vs. Fagales sensu lato). Unfortu- and Crepet, 1993), enclosed ¯oral chambers that imply en- nately, if the earliest angiosperms were frequently herbaceous, trapment pollination (Gandolfo et al., 2004), non-nectar ¯oral and/or aquatic, as may be suggested by both molecular evi- rewards (Crepet and Nixon, 1998b), and at least some sym- dence and some fossil evidence (and, there is no credible early petaly (Nixon and Crepet, 1993). The lack of strongly zygo- fossil evidence of Amborella), it is unlikely that we will ever October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1679 have a satisfying fossil record for the early history of the CRANE, P. R., E. M. FRIIS, AND K. R. PEDERSEN. 1989. Reproductive structure group. The next phase in any such analysis will be to combine and function in Cretaceous Chloranthaceae. Plant Systematics and Evo- existing and future molecular data with sound morphological lution 165: 211±226. CRANE,P.R.,AND S. R. MANCHESTER. 1982. An extinct juglandaceous fruit analyses of potential gymnospermous groups, with the goal of from the Upper Paleocene of Southern England. Botanical Journal of ®nding transitional forms between woody gymnosperms (or the Linnean Society 85: 89±101. ``progymnosperms'') and herbaceous angiosperms. Only after CRANE, P. R., P. HERENDEEN, AND E. M. FRIIS. 2004. Fossils and plant phy- such a careful analysis can we accurately estimate minimum logeny. American Journal of Botany 91: 1683±1699. divergence times for the entire angiosperm clade relative to CRANE, P. R., K. R. PEDERSEN,E.M.FRIIS, AND A. N. DRINNAN. 1993. other seed plants. Early Cretaceous (early to middle Albian) platanoid in¯orescences as- sociated with Sapindopsis leaves from the Potomac Group of eastern LITERATURE CITED North America. Systematic Botany 18: 328±344. CRANE,P.R.,AND R. A. STOCKEY. 1985. Growth and reproductive biology APG II. 2003. An update of the angiosperm phylogeny group classi®cation of Joffrea speirsii gen. et sp. nov., a Cercidiphyllum-like plant from the for the orders and families of ¯owering plants: APG II. Botanical Journal Late Paleocene of Alberta, Canada. Canadian Journal of Botany 63: of the Linnean Society 141: 399±436. 340±364. BACKLUND, A. 1996. Phylogeny of Dipsacales. Acta Universitatis Upsaliensis CREPET, W. L. 1978. Investigations of angiosperms from the Eocene of North 243: 1±34. America: an aroid spadix. Review of Paleobotany and Palynology 25: BARKMAN, T. J., G. CHENERY,J.R.MCNEAL,J.LYONS-WEILER,W.J.EL- 241±252. LISENS,G.MOORE,A.D.WOLFE, AND C. W. DE PAMPHILIS. 2000. In- CREPET, W. L. 1979a. Insect pollination: a paleontological perspective. dependent and combined analyses of sequences from all three genomic BioScience 29: 102±108. compartments converge on the root of ¯owering plant phylogeny. Pro- CREPET, W. L. 1979b. Some aspects of the pollination biology of Middle ceedings of the National Academy of Sciences, USA 97: 13166±13171. Eocene angiosperms. Review of Paleobotany and Palynology 27: 213± BASINGER,J.F.,AND D. L. DILCHER. Ancient bisexual ¯owers. Science 224: 238. 511±513. CREPET, W. L. 1984a. Advanced (constant) insect pollination mechanisms: BENTON, M. J. 1998. Molecular and morphological phylogenies of mammals: patterns of diversi®cation and implications vis a vis angiosperm diversity. congruence with stratigraphic data. Molecular Phylogenetics and Evo- Annals Missouri Botanical Garden 71: 607±630. lution 9: 398±407. CREPET, W. L. 1984b. Ancient ¯owers for the faithful. Natural History 93: BENTON, M. J. 2001. Finding the tree of life: matching phylogenetic trees to 39±45. the fossil record through the 20th century. Proceedings of the Royal CREPET, W. L. 1989. The history and implications of the early North Amer- Society of London, B, Biological Sciences 268: 2123±2130. ican Fossil record of Fagaceae. In P. R. Crane [ed.], The Hamamelidae, BORSCH, T., K. W. HILU,D.QUANDT,V.WILDE,C.NEINHUIS, AND W. 45±66. Oxford University Press, Oxford, UK. BARTHLOTT. 2003. Noncoding plastid trnT-trnF sequences reveal a well CREPET, W. L. 1996. Timing in the evolution of derived ¯oral characters: resolved phylogeny of basal angiosperms. Journal of Evolutionary Bi- Upper Cretaceous (Turonian) taxa with tricolpate and tricolpate-derived ology 16: 558±576. pollen. Review of Palaeobotany and Palynology 90: 339±360. BOUCHER, L. D., S. R. MANCHESTER, AND W. S. JUDD. 2003. An extinct CREPET, W. L. 1999. Early bloomers. Natural History 108: 40±41. genus of Salicaceae based on twigs with attached ¯owers, fruits, and CREPET, W. L. 2000. Progress in understanding angiosperm history, success, foliage from the Eocene Green River Formation of Utah and Colorado, and relationships: Darwin's abominably ``perplexing phenomenon.'' Pro- USA. American Journal of Botany 90: 1389±1399. ceedings of the National Academy of Sciences, USA 97: 12939±12941. BREMER, K. 2000. Early Cretaceous lineages of monocots ¯owering plants. CREPET, W. L. 2001. Plant±animal interactions: insect pollination. In D. E. Proceedings of the National Academy of Sciences, USA 97: 4707±4711. G. Briggs and P. R. Crowther [eds.], Palaeobiology II, 426±429. Black- CALL, V. B., AND D. L. DILCHER. 1992. Investigations of Angiosperms from well Scienti®c, Oxford, UK. the Eocene of southeastern North America: Samaras of Fraxinus wiil- CREPET,W.L.,AND C. P. DAGHLIAN. 1980. Castaneoid in¯orescences from coxiana Berry. Review of Palaeobotany and Palynology 74: 249±266. the Middle Eocene of Tennessee and the diagnostic value of pollen (at CHASE, M. W. 2004. Monocot relationships: an overview of the best char- the subfamily level) in the Fagaceae. American Journal of Botany 67: acterized major angiosperm clade. American Journal of Botany 91: 739±757. 1645±1665. CREPET,W.L.,AND C. P. DAGHLIAN. 1982a. Euphorbioid in¯orescences from CHAW, S.-M., C. L. PARKINSON,Y.CHENG,T.M.VINCENT, AND J. D. PALM- the Middle Eocene Claiborne Formation. American Journal of Botany ER. 2000. Seed plant phylogeny inferred from all three plant genomes: 69: 258±266. monophyly of extant gymnosperms and origin of Gnetales from conifers. CREPET,W.L.,AND C. P. DAGHLIAN. 1982b. Lower Eocene and Paleocene Proceedings of the National Academy of Sciences, USA 97: 4086±4091. Gentianaceae: ¯oral and palynological evidence. Science 214: 75±77. CHMURA, C. A. 1973. Upper Cretaceous (Campanian-Maastrichtian) angio- sperm pollen from the western San Joaquin Valley, California, USA. CREPET, W. L., C. P. DAGHLIAN, AND M. ZAVADA. 1980. Investigations of Paleaontographica B 141: 89±171. angiosperms from the Eocene of North America: a new juglandaceous catkin. Review of Paleobotany and Palynology 30: 361±370. CHRISTOPHER, R. A. 1979. Normapolles and triporate pollen assemblages from the Raritan and Magothy Formations (Upper Cretaceous) of New CREPET,W.L.,AND D. L. DILCHER. 1977. Investigations of angiosperms Jersey. Palynology 3: 73±121. from the Eocene of North America: a mimosoid in¯orescence. American COLLINSON, M. E., M. C. BOULTER, AND P. L. HOLMES. 1993. Magnoliophyta Journal of Botany 64: 714±725. (`Angiospermae'). In M. J. Benton [ed.], The fossil record 2, 809±849. CREPET, W. L., D. L. DILCHER, AND F. W. P OTTER. 1974. Eocene angiosperm Chapman and Hall, London, UK. ¯owers. Science 185: 781±782. CONWENTZ, H. 1886. Die Flora des Bernsteins, 2. Die Angiospermen des CREPET, W. L., D. L. DILCHER, AND F. W. P OTTER. 1975. Investigations of Bernsteins. Wilhelm Engelman, Leipzig, Germany. angiosperms from the Eocene of North America: a catkin with juglan- CRANE, P. R. 1981. Betulaceous leaves and fruits from the British Upper daceous af®nities. American Journal of Botany 62: 813±823. Palaeocene. Botanical Journal of the Linnean Society 83: 103±136. CREPET,W.L.,AND G. D. FELDMAN. 1991. Earliest fossil evidence of grasses CRANE, P. R. 1985. Phylogenetic analysis of seed plants and the origin of in the fossil record. American Journal of Botany 78: 1010±1014. angiosperms. Annals of the Missouri Botanical Garden 72: 716±793. CREPET, W. L., AND E. M. FRIIS. 1987. The evolution of insect pollination CRANE, P. R. 1989. Early fossil history and evolution of the Betulaceae. In in the angiosperms. In E. M. Friis, W. G. Chaloner, and P. R. Crane P. R. Crane and S. Blackmore [eds.], Evolution, systematics and fossil [eds.], Angiosperm origins and biological consequences, 181±201. Cam- history of the Hamamelidae, vol. 2, 87±116. Systematics Association bridge University Press, Cambridge, UK. Special Volume 40. Clarendon Press, Oxford, UK. CREPET,W.L.,AND K. C. NIXON. 1989a. Earliest megafossil evidence of CRANE,P.R.,AND D. L. DILCHER. 1984. Lesqueria: and early angiosperm Fagaceae: phylogenetic and biogeographic implications. American Jour- fruiting axis from the mid-Cretaceous. Annals of the Missouri Botanical nal of Botany 76: 842±855. Garden 71: 384±402. CREPET,W.L.,AND K. C. NIXON. 1989b. Extinct transitional Fagaceae from 1680 AMERICAN JOURNAL OF BOTANY [Vol. 91

the Oligocene and their phylogenetic implications. American Journal of Potomac Group (mid-Cretaceous) of eastern North America. American Botany 76: 1493±1505. Journal of Botany 78: 153±176. CREPET,W.L.,AND K. C. NIXON. 1994. Flowers of Turonian Magnoliidae DOWNIE,S.R.,AND J. D. PALMER. 1992. Use of chloroplast DNA rearrange- and their implications. In P. K. Endress and E. M. Friis [eds.], Early ments in reconstructing plant phylogeny. In P. S. Soltis, D. E. Soltis, and evolution of ¯owers, Plant Systematics and Evolution (Supplement) 8: J. J. Doyle [eds.], Molecular systematics of plants, 14±35. Chapman and 73±92. Hall, New York, New York, USA. CREPET,W.L.,AND K. C. NIXON. 1996. The fossil history of stamens. In DOYLE, J. A. 1973. The monocotyledons: their evolution and comparative W. G. D'Arcy and R. C. Keating [eds.], The anther: form, function and biology. V. Fossil evidence on early evolution of the monocotyledons. phylogeny, 25±57. Cambridge University Press, Cambridge, UK. The Quarterly Review of Biology 48: 399±413. CREPET,W.L.,AND K. C. NIXON. 1998a. Two new fossil ¯owers of mag- DOYLE,J.A.,AND M. J. DONOGHUE. 1986. Seed plant phylogeny and the noliid af®nity from the Late Cretaceous of New Jersey. American Journal origin of angiosperms: an experimental cladistic approach. Botanical Re- of Botany 85: 1273±1288. view 52: 321±431. CREPET,W.L.,AND K. C. NIXON. 1998b. Fossil Clusiaceae from the Late DOYLE,J.A.,AND L. J. HICKEY. 1976. Pollen and leaves from the mid- Cretaceous (Turonian) of New Jersey and implications regarding the his- Cretaceous Potomac Group and their bearing on early angiosperm evo- tory of bee pollination. American Journal of Botany 85: 1122±1133. lution. In C. B. Beck [ed.], Origin and early evolution of angiosperms, CREPET, W. L., K. C. NIXON, AND E. M. FRIIS. 1991. Fossil evidence for the 139±206. Columbia University Press, New York, New York, USA. evolution of biotic pollination. Transactions of the Royal Society of Lon- EKLUND, H. 2003. First Cretaceous ¯owers from Antarctica. Review of Pa- don, B, Biological Sciences 333: 187±195. laeobotany and Palynology 127: 187±217. CREPET, W. L., K. C. NIXON,E.M.FRIIS, AND J. FREUDENSTEIN. 1992. EKLUND, H., E. M. FRIIS, AND K. R. PEDERSEN. 1997. Chloranthaceous ¯oral Oldest fossil ¯owers of hamamelidaceous af®nity, from the late Creta- structures from the Late Cretaceous of Sweden. Plant Systematics and ceous of New Jersey. Proceedings National Academy of Sciences, USA Evolution 207: 13±42. 89: 8969±8989. FRIIS, E. M. 1983. Upper Cretaceous (Senonian) ¯oral structures of juglan- CREPET, W. L., K. C. NIXON, AND M. A. GANDOLFO. 2001. Turonian ¯ora dalean af®nity containing Normapolles pollen. Review of Palaeobotany of New Jersey, USA. VII International Symposium on Mesozoic Terres- and Palynology 39: 161±188. trial Ecosystems, AsociacioÂn PaleontoloÂgica Argentina, PublicacioÂn Es- FRIIS, E. M. 1984. Preliminary report of Upper Cretaceous angiosperm re- pecial 7: 61±69. productive organs from Sweden and their level of organization. Annals CREPET,W.L.,AND T. F. S TUESSY. 1978. A reinvestigation of the fossil Missouri Botanical Garden 71: 403±418. Viguiera cronquistii (Compositae). Brittonia 30: 483±491. FRIIS, E. M. 1990. Silvianthemum suecicum gen. et. sp. nov., a new saxifra- CREPET,W.L.,AND D. W. TAYLOR. 1985. The diversi®cation of the Legu- galean ¯ower from the Late Cretaceous of Sweden. Biologiske Skrifter minosae: ®rst fossil evidence of the Mimosoideae and Papilionoideae. 36: 5±21. Science 288: 1087±1089. FRIIS, E. M., P. R. CRANE, AND K. R. PEDERSEN. 1986. Floral evidence for CREPET,W.L.,AND D. W. TAYLOR. 1986. Primitive mimosoid ¯owers from Cretaceous chlorantoid angiosperms. Nature 320: 163±164. the Paleocene-Eocene and their systematic and evolutionary implications. FRIIS, E. M., P. R. CRANE, AND K. R. PEDERSEN. 1988. Reproductive struc- American Journal of Botany 73: 548±563. tures of Cretaceous Platanaceae. Biologiske Skrifter 31: 1±55. CRONQUIST, A. 1981. An integrated system of classi®cation of ¯owering FRIIS, E. M., P. R. CRANE, AND K. R. PEDERSEN. 1997a. Fossil history of plants. Columbia University Press, New York, New York, USA. magnoliid angiosperms. In K. Iwatsuki and P. H. Raven [eds.], Evolution DAGHLIAN, C. P. 1981. A review of the fossil record of monocotyledons. The and diversi®cation of land plants, 121±156. Springer Verlag, New York, Botanical Review 47: 517±555. New York, USA. DAGHLIAN,C.P.,AND W. L. CREPET. 1982. The evolutionary signi®cance of FRIIS, E. M., P. R. CRANE, AND K. R. PEDERSEN. 1997b. Anacostia, a new Quercus sect. Erythrobalanus from the Oligocene of East Texas. Amer- basal angiosperm from the Early Cretaceous of North America and Por- ican Journal of Botany 70: 639±649. tugal with trichotomocolpate/monocolpate pollen. Grana 36: 225±244. DAGHLIAN, C. P., W. L. CREPET, AND T. D ELEVORYAS. 1980. Investigations FRIIS,E.M.,AND W. L. CREPET. 1987. Time of appearance of ¯oral features. of Tertiary angiosperms: Eomimosoidea plumosa from the Lower Oli- In E. M. Friis, W. G. Chaloner, and P. R. Crane [eds.], Angiosperm gocene of eastern Texas. American Journal of Botany 67: 309±320. origins and biological consequences, 145±179. Cambridge University DAVIES, T. J., T. G. BARRACLOUGH,M.W.CHASE,P.S.SOLTIS,D.E.SOLTIS, Press, Cambridge, UK. AND V. S AVOLAINEN. 2004. Darwin's abominable mystery: insights from FRIIS, E. M., J. A. DOYLE,P.K.ENDRESS, AND Q. LENG. 2003. Archaefruc- a supertree of the angiosperms. Proceedings of the National Academy of tusÐangiosperm precursor or specialized early angiosperm? Trends in Sciences, USA 101: 1904±1909. Plant Science 8: 369±373. DILCHER,D.L.,AND P. B ASSON. 1990. Mid-Cretaceous angiosperm leaves FRIIS, E. M., H. EKLUND,K.R.PEDERSEN, AND P. R. CRANE. 1994. Virgi- from a new fossil locality in Lebanon. Botanical Gazette 151: 538±547. nianthus calycanthoides gen. et sp. nov.Ða calycanthaceous ¯ower from DILCHER,D.L.,AND P. R. CRANE. 1984. Archaeanthus: an early angiosperm the Potomac Group (Early Cretaceous) of eastern North America. Inter- from the Cenomanian of the western interior of North America. Annals national Journal of Plant Sciences 155: 772±785. of the Missouri Botanical Garden 71: 351±383. FRIIS, E. M., K. R. PEDERSEN, AND P. R. CRANE. 1992. Esqueiria gen. nov., DILCHER, D. L., W. L. CREPET,C.D.BEEKER, AND H. C. REYNOLDS. 1976. fossil ¯owers with combretaceous features from the Late Cretaceous of Reproductive and vegetative morphology of a Cretaceous angiosperm. Portugal. Biologiske Skrifter 41: 1±45. Science 191: 854±856. FRIIS, E. M., K. R. PEDERSEN, AND P. R. CRANE. 1995. Appomattoxia ancis- DILCHER,D.L.,AND W. L. KOVACH. 1986. Early angiosperm reproduction: trophora gen. et sp. nov., a new Early Cretaceous plant with similarities a new fruiti®cation from the Dakota Formation (Cenomanian) of Kansas. to Circaeaster and extant Magnoliidae. American Journal of Botany 82: American Journal of Botany 73: 1228±235. 933±943. DILCHER,D.L.,AND S. R. MANCHESTER. 1988. Investigations of the angio- FRIIS, E. M., K. R. PEDERSEN, AND P. R. CRANE. 2001. Fossil evidence of sperms from the Eocene of North America: a fruit belonging to Eu- water lilies (Nymphaeales) in the early Cretaceous. Nature 410: 357± phorbiaceae. Tertiary Research 9: 45±68. 360. DILCHER, D. L., F. W. POTTER, AND W. L. CREPET. 1976. Investigations of FRIIS,E.M.,AND A. SKARBY. 1981. Structurally preserved angiosperm ¯ow- angiosperms from the Eocene of North America: Juglandaceous winged ers from the Upper Cretaceous of Southern Sweden. Nature 291: 485± fruits. American Journal of Botany 63: 532±544. 486. DILCHER, D. L., F. POTTER, AND H. REYNOLDS. 1978. Preliminary account GANDOLFO, M. A., K. C. NIXON, AND W. L. CREPET. 1998a. Tylerianthus of middle Cretaceous angiosperm remains from interior of North Amer- crossmanensis gen. et sp. nov. (aff. Hydrangeaceae) from the Upper Cre- ica. Courier Forschungs-Institut Senckenberg 30: 9±15. taceous of New Jersey. American Journal of Botany 85: 376±386. DRINNAN, A. N., P. R. CRANE,E.M.FRIIS, AND K. R. PEDERSEN. 1990. GANDOLFO, M. A., K. C. NIXON, AND W. L. CREPET. 1998b. A new fossil Lauraceous ¯owers from the Potomac Group (mid-Cretaceous) of eastern ¯ower from the Turonian of New Jersey: Dressanthia bicarpellata gen. North America. Botanical Gazette 151: 370±384. et sp. nov. (Capparales). American Journal of Botany 85: 964±974. DRINNAN, A. N., P. R. CRANE,E.M.FRIIS, AND K. R. PEDERSEN. 1991. GANDOLFO, M. A., K. C. NIXON, AND W. L. CREPET. 2000. Monocotyledons: Angiosperm ¯owers and tricolpate pollen of buxaceous af®nity from the a review of their Early Cretaceous record. In K. Wilson and D. Morrison October 2004] CREPETETAL.ÐFOSSIL EVIDENCE AND PHYLOGENY 1681

[eds.], Proceedings of the Second International Conference on the Com- LAINEN,M.W.CHASE,M.P.POWELL,L.A.ALICE,R.EVANS,H.SA- parative Biology of the Monocotyledons, CSIRO, Sydney, Australia, 44± QUET,C.NEINHUIS,T.A.B.SLOTTA,J.G.ROHWER,C.S.CAMPBELL, 52. AND L. W. CHATROU. 2003. Angiosperm phylogeny based on matK se- GANDOLFO, M. A., K. C. NIXON, AND W. L. CREPET. 2002. Triuridaceae quence information. American Journal of Botany 90: 1758±1776. fossil ¯owers from the Upper Cretaceous of New Jersey. American Jour- HORWOOD, A. R. 1912a. The past history of the monocotyledons, with re- nal of Botany 89: 1940±1957. marks on their origin. The Scottish Botanical Review 1: 164±180. GANDOLFO, M. A., K. C. NIXON, AND W. L. CREPET. 2004. The oldest com- HORWOOD, A. R. 1912b. The past history of the monocotyledons, with re- plete fossil ¯owers of Nymphaeaceae and implications for the complex marks on their origin. The Scottish Botanical Review 1: 216±234. insect entrapment pollination mechanisms in early angiosperms. Pro- HUTCHINSON, G. E. 1965. The ecological theater and the evolutionary play. ceedings of the National Academy of Sciences, USA 101: 8056±8060. Yale University Press, New Haven, Connecticut, USA. GANDOLFO, M. A., K. C. NIXON,W.L.CREPET, AND G. E. RATCLIFFE. 1997. JUDD,W.S.,AND R. G. OLMSTEAD. 2004. A survey of tricolpate (eudicot) A new fossil fern assignable to Gleicheniaceae from late cretaceous sed- phylogenetic relationships. American Journal of Botany 91: 1627±1644. iments of New Jersey. American Journal of Botany 84: 483±493. KRASSILOV, V. A., P. V. SHILIN, AND V. A. VACHRAMEEV. 1983. Cretaceous GANDOLFO, M. A., K. C. NIXON,W.L.CREPET,D.W.STEVENSON, AND E. ¯owers from Kazakstan. Review of Palaeobotany and Palynology 40: M. FRIIS. 1998c. Oldest known fossil ¯owers of monocotyledons. Nature 91±113. 394: 532±533. KNOBLOCH, E., AND D. H. MAI. 1986. Monographie der fruÈchte und samen GOLOBOFF, P. A. 1999. NONA (Parsimony program for IBM compatibles and in der Kreide von Mitteleuropa. Edice Rozpravy uÂstrÏedniho uÂstavu Geo- SUN workstations). Distributed by the author, website: www.cladistics. logickeÂho 47: 1±219. com. LAKHANPAL,R.N.,AND J. K. VERMA. 1966. Fossil wood of Tetrameles from GRAHAM,S.W.,AND R. G. OLMSTEAD. 2000. Utility of 17 chloroplast genes the Deccan Intertrappean beds of Mohgaonkalan, Madhya Pradesh. The for inferring the phylogeny of the basal angiosperms. American Journal Paleobotanist 14: 209±213. of Botany 87: 1712±1730. LAWRENCE, G. H. M. 1951. Taxonomy of vascular plants. The Macmillan GRAUR, D., AND W. M ARTIN. 2004. Reading the entrails of chickens: molec- Company, New York, New York, USA. ular timescales of evolution and the illusion of precision. Trends in Ge- LES, D. H., E. L. SCHNEIDER,D.J.PADGETT,P.S.SOLTIS,D.E.SOLTIS, netics 20: 80±86. AND M. ZANIS. 1999. Phylogeny, classi®cation and ¯oral evolution of HARLEY, M. M. 1996. Palm pollen and the fossil record, vols. 1 and 2. Ph.D. water lilies (Nymphaeaceae; Nymphaeales): a synthesis of non-molecu- dissertation, University of East London and The Royal Botanic Garden, lar, rbcL, matK, and 18S rDNA data. Systematic Botany 24: 28±46. Kew, UK. LINDER, H. P. 1986. The evolutionary history of the Poales/RestonialesÐa HARLEY,M.M.,AND M. S. ZAVADA. 2000. Pollen of monocotyledons: se- hypothesis. Kew Bulletin 42: 297±318. lecting characters for cladistic analysis. In K. Wilson and D. Morrison LOCONTE, H., AND D. W. STEVENSON. 1990. Cladistics of the Spermophyta. [eds.], Proceedings of the Second International Conference on the Com- Brittonia 42: 197±211. parative Biology of the Monocotyledons. CSIRO, Sydney, Australia, MAGALLOÂ N-PUEBLA, S., AND M. J. SANDERSON. 2002. Relationships among 194±213. seed plants inferred from highly conserved genes: sorting con¯icting HEDGES,S.B.,AND S. KUMAR. 2004. Precision of molecular time estimates. phylogenetic signals among ancient lineages. American Journal of Bot- Trends in Genetics 20: 242±247. any 89: 1991±2006. HERENDEEN,P.S.,AND P. R. CRANE. 1992. Early caesalpiniod fruits from MAGALLOÂ N-PUEBLA, S., P. R. CRANE, AND P. S. HERENDEEN. 1999. Phylo- the Palaeogene of southern England. In P. S. Herendeen and D. L. Dilcher genetic pattern, diversity, and diversi®cation of eudicots. Annals of the [eds.], Advances in legume systematics, vol. 4, The fossil record, 57± Missouri Botanical Garden 86: 297±372. 68. MAGALLOÂ N-PUEBLA, S., P. S. HERENDEEN, AND P. R. CRANE. 1997. Quad- HERENDEEN,P.S.,AND P. R. CRANE. 1995. The fossil history of the mono- riplatanus georgianus gen. et sp. nov.: staminate and pistillate platana- cotyledons. In P. J. Rudall, P. J. Cribb, D. F. Cutler, and C. J. Humphries ceous ¯owers from the Late Cretaceous (Coniacian-Santonian) of Geor- [eds.], Monocotyledons: systematics and evolution, 1±21. Royal Botanic gia, USA. International Journal of Plant Sciences 158: 373±394. Gardens, Kew, UK. MANCHESTER, S. R. 1992. Flowers fruits and pollen of Florissantia an extinct HERENDEEN, P. H., W. L. CREPET, AND K. C. NIXON. 1993. Chloranthus-like malvalean genus from the Eocene and Oligocene of western North Amer- stamens from the Upper Cretaceous of New Jersey. American Journal of ica. American Journal of Botany 79: 996±1008. Botany 80: 865±871. MATTHEWS, S., AND M. J. DONOGHUE. 1999. The root of angiosperm phy- HERENDEEN, P. S., W. L. CREPET, AND K. C. NIXON. 1994. Fossil ¯owers logeny inferred from duplicate phytochrome genes. Science 286: 947± and pollen of Lauraceae from the Upper Cretaceous of New Jersey. Plant 950. Systematics and Evolution 189: 29±40. MCIVER,E.E.,AND J. F. BASINGER. 1993. Flora of the Ravenscrag Formation HERENDEEN, P. S., S. MAGALLOÂ N-PUEBLA,R.LUPIA,P.R.CRANE, AND J. (Paleocene), southwestern Saskatchewan, Canada. Palaeontographica KOBYLINSKA. 1999. A preliminary conspectus of the Allon ¯ora from Canadiana 10: 1±167. the Late Cretaceous (late Santonian) of central Georgia, U.S.A. Annals MOHR,B.A.R.,AND E. M. FRIIS. 2000. Early angiosperms from the Lower of the Missouri Botanical Garden 86: 407±471. Cretaceous Aptian Crato Formation (Brazil), a preliminary report. Inter- HERMSEN, E., M. A. GANDOLFO,K.C.NIXON, AND W. L. CREPET. 2003. national Journal of Plant Sciences 161(6 Supplement): 155±167. Divisestylus gen. nov. (aff. Iteaceae), a fossil saxifrage from the Late MULLER, J. 1981. Fossil pollen records of extant angiosperms. The Botanical Cretaceous of New Jersey, USA. American Journal of Botany 90: 1373± Review 47: 1±141. 1388. NISHIDA, H. 1994. Elsemaria, a Late Cretaceous angiosperm fructi®cation HERNANDEZ-CASTILLO,G.R.,AND S. R. S. CEVALLOS-FERRIZ. 1999. Re- from Hokkaido, Japan. Plant Systematic and Evolution (Supplement) 8: productive and vegetative organs with af®nities to Haloragaceae from 123±135. the Upper Cretaceous Huepac Chert locality of Sonora, Mexico. Amer- NISHIDA, H., AND M. NISHIDA. 1988. Protomonimia kasai-nakajhongii gen ican Journal of Botany 86: 1717±1734. et sp. nov.: a permineralized magnolialean fructi®cation from the mid- HICKEY, L. J. 1991. Preliminary report on the ¯ora of the Meeteetse For- Cretaceous of Japan. The Botanical Magazine 101: 397±426. mation (late Cretaceous) of Wyoming and Montana. American Journal NIXON, K. C. 1996. Paleobotany in cladistics and cladistics in paleobotany: of Botany 78: 115± (Abstract). enlightenment and uncertainty. Review of Paleobotany and Palynology HICKEY,L.J.,AND R. K. PETERSON. 1978. Zingiberopsis, a fossil genus of 90: 361±373. the ginger family from the Late Cretaceous to Early Eocene sediments NIXON, K. C. 1999. The parsimony ratchet, a new method for rapid parsi- of western interior North America. Canadian Journal of Botany 56: mony analysis. Cladistics 15: 407±414. 1136±1152. NIXON, K. C. 2003. Winclada. Computer program for phylogenetic analysis HILL,C.R.,AND P. R. CRANE. 1982. Evolutionary cladistics and the origin distributed by the author, Windows 95/98/NT version, website: of angiosperms. In K. A. Joysey and A. E. Friday [eds.], Problems of www.cladistics.com. phylogenetic reconstructions, 269±361. Academic Press, New York, New NIXON,K.C.,AND W. L. CREPET. 1993. Late Cretaceous fossil ¯owers of York, USA. ericalean af®nity. American Journal of Botany 80: 616±623. HILU,K.W.,T.BORSCH,K.MUÈ LLER,D.E.SOLTIS,P.S.SOLTIS,V.SAVO- NIXON, K. C., M. A. GANDOLFO, AND W. L. CREPET. 2001. Origins of Fa- 1682 AMERICAN JOURNAL OF BOTANY [Vol. 91

gaceae: a review of relevant Turonian fossil material from New Jersey. TELL,S.M.SWENSEN,L.M.PRINCE,W.J.KRESS,K.C.NIXON, AND American Journal of Botany 88: 68 (Abstract). J. S. FARRIS. 2000. Angiosperm phylogeny inferred from 18S rDNA, NIXON, K. C., W. L. CREPET,D.M.STEVENSON, AND E. M. FRIIS. 1994. A rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133: reevaluation of seed plant phylogeny. Annals Missouri Botanical Garden 381±461. 81: 484±533. SOLTIS, P. S., D. E. SOLTIS, AND M. W. CHASE. 1999. Angiosperm phylogeny NORELL,M.A.,AND M. J. NOVACEK. 1992. The fossil record and evolution: inferred from multiple genes as a tool for comparative biology. Nature comparing cladistic and paleontologic evidence for vertebrate history. 402: 402±403. Science 255: 1690±1693. STOCKEY, R. A. 1987. A permineralized ¯ower from the Middle Eocene of PALMER,J.D.,AND D. ZAMIR. 1982. Chloroplast DNA evolution and phy- British Columbia. American Journal of Botany 74: 1878±1887. logenetic relationships in Lycopersicon. Proceedings of the National STOCKEY,R.A.,AND K. B. PIGG. 1991. Flowers and fruits of Princetonia Academy of Sciences, USA 79: 5006±5010. allenbyensis from the Middle Eocene chert of British Columbia. Review PALMER, J. D., C. R. SHIELDS,D.B.COHEN, AND T. J. ORTEN. 1983. Chlo- of Palaeobotany and Palynology 70: 163±172. roplast DNA evolution and the origin of Brassica species. Theoretical STOCKEY,R.A.,AND G. W. ROTHWELL. 1997. The aquatic angiosperm Tra- and Applied Genetics 65: 181±189. pago angulata from the Upper Cretaceous (Maastrichtian) St. Mary Riv- PALMER, J. D., R. A. JORGENSEN, AND W. F. T HOMPSON. 1985. Chloroplast er Formation of southern Alberta. International Journal of Plant Sciences DNA variation and evolution in Pisum: patterns of changes and phylo- 158: 83±94. genetic analysis. Genetics 109: 195±213. SUN, G., D. L. DILCHER,S.ZHENG, AND Z. ZHOU. 1998. In search of the PARENTI, L. R. 1980. A phylogenetic analysis of the land plants. Biological ®rst ¯ower: a angiosperm, Archaefructus, from northeast china. Journal of the Linnean Society 13: 225±242. Science 282: 1692±1695. PARKINSON, C. L., K. L. ADAMS, AND J. D. PALMER. 1999. Multigene anal- SUN, G., Q. JI,D.L.DILCHER,S.L.ZHENG,K.NIXON, AND X. F. WANG. yses identify the three earliest lineages of extant ¯owering plants. Cur- 2002. Archaefructaceae, a new basal angiosperm family. Science 296: rent Biology 9: 1485±1488. 899±904. PEDERSEN, K. R., P. R. CRANE,A.N.DRINNAN, AND E. M. FRIIS. 1991. TAKHTAJAN, A. 1997. Diversity and classi®cation of ¯owering plants. Co- Fruits from the mid-Cretaceous of North America with pollen grains of lumbia University Press, New York, New York, USA. the Clavatipollenites type. Grana 30: 577±590. TAYLOR,D.W.,AND W. L. CREPET. 1987. Fossil ¯oral evidence of Mal- PEDERSEN, K. R., E. M. FRIIS,P.R.CRANE, AND A. N. DRINNAN. 1994. pighiaceae and an early plant±pollinator relationship. American Journal Reproductive structures of an extinct platanoid from the Early Cretaceous of Botany 74: 274±286. (Latest Albian) of eastern North America. Review of Palaeobotany and TAYLOR,D.W.,AND L. J. HICKEY. 1990. An Aptian plant with attached Palynology 80: 291±303. leaves and ¯owers: implications for angiosperm origin. Science 247: QIU, Y.-L., J. LEE,F.BERNASCONI-QUADRONI,D.E.SOLTIS,P.S.SOLTIS,M. 702±704. ZANIS,E.A.ZIMMER,Z.CHEN,V.SAVOLAINEN, AND M. W. CHASE. TAYLOR,D.W.,AND L. J. HICKEY. 1992. Phylogenetic evidence for the 1999. The earliest angiosperms: evidence from mitochondrial, plastid herbaceous origin of angiosperms. Plant Systematics and Evolution 180: and nuclear genomes. Nature 402: 404±407. 137±156. RETALLACK,G.J.,AND D. L. DILCHER. 1981. Early angiosperm reproduction: TIFFNEY, B. H. 1977. Upper Cretaceous angiosperm ¯ower from Martha's Prisca reynoldsii, gen. et sp. nov. from the mid-Cretaceous coastal de- Vineyard, Massachusetts. Nature 265: 136±137. posits in Kansas, USA. Palaeontographica B 179: 103±137. VOIGT, E. 1981. Upper Cretaceous bryozoan ±seagrass association in the Maastrichtian of The Netherlands. In G. P. Larwood and C. Nielsen ROTHWELL,G.W.,AND R. SERBET. 1994. Lignophyte phylogeny and the evolution of spermatophytes: a numerical cladistic analysis. Systematic [eds.], Recent and fossil bryozoa, 281±298. Olsen and Olsen, Fredens- Botany 19: 443±482. burg, Denmark. VOIGT, E., AND W. D OMKE. 1955. Thalassocaris bosqueti Debey ex Miquel, SANDERSON, M. J. 1997. A nonparametric approach to estimating divergence ein strukturell erhaltenes Seegras aus der hollaÈndischen. Mitteilungen aus times in the absence of rate constancy. Molecular Biology and Evolution dem Geologischen Staatsinstitut in Hamburg 24: 87±102. 14: 1218±1232. WIKSTROÈ M, N., V. SAVOLAINEN, AND M. W. CHASE. 2001. Evolution of the SANDERSON, M. J., J. L. THORNE,N.WILK, AND K. G. BREMER. 2004. Mo- angiosperms: calibrating the family tree. Proceedings of the Royal So- lecular evidence on plant divergence times. American Journal of Botany ciety of London, B, Biological Sciences 268: 2211±2220. 91: 1656±1665. YOO, M.-J., C. D. BELL,P.S.SOLTIS, AND D. E. SOLTIS. 2004. Estimation SAVOLAINEN, V. M., F. FAY,D.C.ALBACH,A.BACKLUND,M.VAN DER of divergence times and biogeography of major clades of Nymphaeales. BANK,K.M.CAMERON,S.A.JOHNSON,M.D.LLEDO,J.C.PINTAUD, American Journal of Botany 90 (Abstract). OWEL HEAHAN OLTIS OLTIS ESTON M. P ,M.C.S ,D.E.S ,P.S.S ,P.W ,W. YOUNG,D.A.,AND D. S. SEIGLER. 1981. Phytochemistry and angiosperm M. WHITTEN,K.J.WURDACK, AND M. W. CHASE. 2000. Phylogeny of phylogeny. Praeger Publishers, New York, New York, USA. the eudicots: a nearly complete familial analysis based on rbcL gene ZANIS, M. J., D. E. SOLTIS,P.S.SOLTIS,S.MATHEWS, AND M. J. DONOGHUE. sequences. Kew Bulletin 55: 257±309. 2002. The root of the angiosperms revisited. Proceedings of the National SCHOÈ NENBERGER, J., K. R. PEDERSEN, AND E. M. FRIIS. 2001. Normapolles Academy of Sciences, USA 99: 6848±6853. ¯owers of fagalean af®nity from the Late Cretaceous of Portugal. Plant ZANIS, M. J., P. S. SOLTIS, Y.-L. QIU,E.ZIMMER, AND D. E. SOLTIS. 2003. Systematics and Evolution 226: 205±230. Phylogenetic analyses and perianth evolution in basal angiosperms. An- SIMS, H. J., P. S. HERENDEEN,R.LUPIA,R.A.CHRISTOPHER, AND P. R . nals of the Missouri Botanical Garden 90: 129±150. CRANE. 1999. Fossil ¯owers with Normapolles pollen from the Upper ZAVADA, M. S. 1984. Angiosperm origins and evolution based on dispersed Cretaceous of southeastern North America. Review of Palaeobotany and fossil pollen ultrastructure. Annals of the Missouri Botanical Garden 71: Palynology 106: 131±151. 444±463. SOLTIS,D.E.,AND P. S. SOLTIS. 2003. The role of phylogenetics in com- ZAVADA, M., AND W. L. CREPET. 1981. Investigations of angiosperms from parative genetics. Plant Physiology 132: 1790±1800. the Eocene of North America: ¯owers of the Celtidoideae. American SOLTIS,P.S.,AND D. E. SOLTIS. 2004. The origin and diversi®cation of Journal of Botany 68: 924±933. angiosperms. The American Journal of Botany 91: 1614±1626. ZHOU, Z., W. L. CREPET, AND K. C. NIXON. 2001. The earliest fossil evidence SOLTIS, D. E., P. S. SOLTIS,M.W.CHASE,M.E.MORT,D.C.ALBACH,M. of the Hamamelidaceae: Late Cretaceous (Turonian) in¯orescences and ZANIS,V.SAVOLAINEN,W.H.HAUN,S.B.HOOT,M.F.FAY,M.AX- fruits of Altingioidea. American Journal of Botany 88: 753±766.