The Botanical Review

, Volume 76, Issue 1, pp 83–135

Fossil Record and Age of the Asteridae

Authors

    • L. H. Bailey Hortorium, Department of Plant BiologyCornell University
Article

DOI: 10.1007/s12229-010-9040-1

Cite this article as:
Martínez-Millán, M. Bot. Rev. (2010) 76: 83. doi:10.1007/s12229-010-9040-1
  • 545 Views

Abstract

The Asteridae is a group of some 80,000 species of flowering plants characterized by their fused corollas and iridoid compounds. Recent phylogenetic analyses have helped delimit the group and have identified four main clades within it; Cornales, Ericales, Lamiids and Campanulids, with the last two collectively known as the Euasteridae. A search for the oldest fossils representing asterids yielded a total of 261 records. Each of these fossils was evaluated as to the reliability of its identification. The oldest accepted fossils for each clade were used to estimate minimum ages for the whole of the Asteridae. The results suggest that the Asteridae dates back to at least the Turonian, Late Cretaceous (89.3 mya) and that by the Late Santonian-Early Campanian (83.5 mya) its four main clades were already represented in the fossil record.

Keywords

AsteridaeCampanulidsEuasteridaeFossilLamiidsMinimum Age

Resumen

Las Asteridas son un grupo de unas 80,000 especies de plantas con flor caracterizadas por sus corolas fusionadas y compuestos iridoides. Análisis filogenéticos recientes han ayudado a delimitar al grupo y han identificado cuatro clados principales en él; Cornales, Ericales, Lamiidas y Campanulidas, con las últimas dos conocidas colectivamente como Euasteridas. Una búsqueda por los fósiles más antiguos que representan asteridas produjo un total de 261 registros. Cada uno de estos fósiles fue evaluado en cuanto a la confiabilidad de su identificación. Los fósiles aceptados más antiguos de cada clado se usaron para estimar edades mínimas para las Asteridas. Los resultados sugieren que las Asteridas datan al menos del Turoniano, Cretácico Tardío (89.3 ma) y que para el Santoniano Tardío-Campaniano Temprano (83.5 ma) sus cuatro clados principales ya estaban representados en el registro fósil.

Introduction

The Asteridae

The Asteridae is a group of flowering plants characterized by their fused corollas and iridoid compounds (Bremer et al., 2002). This group has been recognized by botanists since the eighteenth century, receiving names such as the Monopetalae, Gamopetalae or Sympetalae (Wagenitz, 1992) all of which allude to the characteristic connate corolla. In more recent times, classification systems based on morphology such as those of Cronquist (1981) and Takhtajan (1997), recognized relationships among several families displaying these characters and placed them in the similar subclass(es): Asteridae in the case of Cronquist (1981) and Asteridae, Lamiidae and Cornidae of Takhtajan (1997). With the advent of molecular systematics, the delimitation of the group has become clearer (Chase et al., 1993; Savolainen et al., 2000; Soltis et al., 2000; APG, 1998, 2003, 2009). Most of the taxa that Cronquist (1981) and Takhtajan (1997) placed in their Asteridae or separately in the Asteridae, Lamiidae and Cornidae are still accommodated in the current concept of Asteridae, but several other taxa traditionally placed in the Dillenidae and Rosidae have also been proven to be asterids (compare the three classification systems in “Appendix A”). This new, expanded and redefined Asteridae includes some 80,000 species in 102–106 families, that is, about 1/3–1/4 of all angiosperm species (APG, 1998, 2003, 2009; Bremer et al., 2002).

The Asteridae as defined today is a remarkable group in many respects; for example, two of its families, Asteraceae (=Compositae, sunflower family) and Rubiaceae (coffee family) are among the most biodiverse plant families in terms of number of species. From an ecological and evolutionary point of view, Asteraceae (the sunflower family), Campanulaceae / Lobeliaceae (the bell-flower family) and Apocynaceae / Asclepiadaceae (the milkweed family) have some of the most specialized pollen presentation mechanisms in the plant kingdom. And from an economical perspective, important crops and other widely cultivated plants are asterids: tomato, potato, chili pepper, eggplant, tobacco (Solanaceae), tea (Theaceae), carrot, caraway, celery, dill (Apiaceae), oregano, thyme, sage, mint (Lamiaceae), sunflower, lettuce, artichoke (Asteraceae), coffee (Rubiaceae), blueberries and cranberries (Ericaceae).

Phylogenetic works focusing on all or parts of the Asteridae have substantially increased over the last few years (i.e. Hufford, 1992; Olmstead et al., 1992, 1993; Albach et al., 2001; Bremer et al., 2001, 2002; Lundberg & Bremer, 2003; Zhang et al., 2003; Chandler & Plunkett, 2004; Albach et al., 1998, 2005; Oxelman et al., 2005; Geuten et al., 2004; Schönenberger et al., 2005, etc) and have provided us with a more robust and better supported hypothesis of relationships among the asterid taxa. The most comprehensive study to date is that of Bremer et al. (2002) who focused on the whole of the Asteridae and included 132 genera in their analysis of six chloroplast markers (Fig. 1). Their results, although consistent with previous studies that focused on all flowering plants (e.g. Chase et al., 1993; Savolainen et al., 2000; Soltis et al., 2000), are an important contribution towards resolving and understanding of the relationships among asterid lineages.
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig1_HTML.gif
Fig. 1

Phylogenetic relationships of Asteridae according to Bremer et al. (2002). Numbers in brackets indicate number of terminals used in the original analysis. Names in bold indicate Orders. Bold lines-Asterid clade, patterned lines-Lamiid clade (Euasterid I of Soltis et al., 2000 and APG, 1998), grey lines-Campanulid clade (Euasterid II of Soltis et al., 2000 and APG, 1998)

Bremer et al. (2002) identified four main clades (Fig. 1): Cornales, Ericales, Lamiids and Campanulids. Cornales is sister to the rest of the Asteridae; Ericales is sister to the largest clade, the Euasteridae, formed by the Lamiids (formerly Euasterid I) and the Campanulids (formerly Euasterid II). The Euasteridae is where most of the diversity of the group is found (42.91% of all extant eudicot species [Magallón et al., 1999]).

Fossil Record of the Asteridae

Although the importance of the Asteridae has attracted attention in many areas of research and nearly 1/3 of all angiosperm species are asterids, their fossil record is not as extensive or even reflective of their extant diversity, especially in the large euasterid clade (sensu Soltis et al., 2000, APG, 2003 and Bremer et al., 2002). Moreover, with the exception of selected families (e.g. Eucommiaceae by Call & Dilcher, 1997, Symplocaceae by Kirchheimer, 1949 and Mai & Martinetto, 2006 or Cornaceae by Manchester, 2002), the relatively scarce fossil record of the group has never undergone specialized systematic revision or comprehensive treatment. For the most part, reports of fossils identified as asterids are scattered in paleofloral treatments, preliminary reports and short communications. As an additional problem, most of the reports are old—19th or early 20th Century—and poorly documented.

Estimating Ages of Divergence

Ever since the development of the concept of the molecular clock (Zuckerkandl & Pauling, 1962, 1965; Langley & Fitch, 1974) and of stochastic changes in the genome not subject to natural selection (i.e. neutral theory of molecular evolution (Kimura, 1983)), estimation of ages of divergence have no longer been the exclusive province of paleontology. Fossils alone do not pinpoint the place and time of origin of natural groups any more; instead they are used in conjunction with methods that incorporate the current knowledge of molecular evolution and the vast reservoir of genomic data available. In recent years, with the increase in computational power, correlated new algorithms, and the better understanding of genome evolution and of phylogenetics, the interest in molecular dating has increased at an unprecedented rate (see Bromham & Penny, 2003; Sanderson et al., 2004; Welch & Bromham, 2005 for reviews).

Today, it is widely accepted that genes do not necessarily evolve in a clock-like manner—i.e. according to a strict molecular clock sensu Langley and Fitch (1974)—and that different rates of evolution can be found in different genes, in different partitions and/or on different lineages (Sanderson, 2002). This has led to the development of techniques or methods that, coupled with those that estimate phylogenies themselves, give an estimate of timing in the divergence of lineages (clades) not based on a fixed molecular evolution rate. For example, Sanderson (1997) proposed NPRS (NonParametric Rate Smoothing), a method based on the assumption that evolutionary rates are not clock-like and can change from lineage to lineage (the estimation of that change is highly dependent on the rates of the descendant lineages). An improvement over this method that allows control of the level of smoothing through the introduction of a parameter is PL (Penalized Likelihood) also called Semiparametric Rate Smoothing (Sanderson, 2002). These methods and others have been widely applied to estimate ages of divergence of large clades across the entire tree of life (summarized in Hedges & Kumar, 2003; Magallón, 2004 and Welch & Bromham, 2005).

One feature that all these methods have in common is their need for at least one (but often more) calibration or constraint point(s) which will help determine the rate(s) at which the genomic sequences change. In a very simplified way, the number of base differences / time = rate of molecular substitution. Since the fossil record is the main source of calibration point(s) for these analyses, it is of the utmost importance that the fossils used are reliable, both in terms of taxonomic identity and in terms of age. Unfortunately misidentifications of taxa are common in the fossil record, especially in the case of angiosperms. As Collinson et al. (1993) remarked, “These problems have been exacerbated in the past by a common tendency to include fossils in modern taxa based on superficial similarity rather than in-depth analysis. Although the latter is now the rule rather than the exception, many older determinations have not yet been revised.”

Other methods that do not use molecular sequences such as minimum-age-dating still depend on a reliable fossil record. Minimum-age-dating assigns ages to the different nodes in the phylogeny by choosing the oldest date among the daughter nodes descended from the node in question (Crepet et al., 2004). Progressing this way along the tree, it is possible to assign objective minimum ages to each node in the phylogeny. This method, although free from the pitfalls that plague rate estimation, is very susceptible to errors due to misidentifications and taxonomic misplacements of fossils. Due to its influence and central role in both kinds of methods, molecular-based and fossil-only-based, it is highly desirable that the fossil record be revised and that the reliability of fossils used to assign ages to phylogenies be assessed.

Estimated Age for the Asterids

One of the earliest attempts to date the angiosperm phylogenetic tree by incorporating fossil data into the angiosperm phylogeny was that of Magallón et al. (1999) who assigned dates to the nodes of the Eudicot clade of the Chase et al. (1993) cladogram by referencing the fossil record. While they did not explicitly date the asterid clade in that study, its “older” order, the Ericales was assigned a date of 89.5 my (Table 1). More recently, Crepet et al. (2004) assigned minimum ages to the early lineages of the angiosperm phylogeny of Soltis et al. (2000) using the fossil record. Although that work deals with the earliest nodes on the angiosperm phylogeny and not with the more derived groups, the minimum age for the Asterid clade (90 my; Table 1) is similar to that reported by Magallón et al. (1999). Age estimates of the angiosperms as a whole, based on molecular clock techniques (Wikström et al., 2001) give an older estimate for the origin of the Asterid clade (112–122 my, Table 1) while studies focusing on the Asteridae alone (Bremer et al., 2004) give an even older date, somewhere before 128 million years ago (Table 1). The results of these studies imply a gap of at least 22 million years for which, if accurate, the early asterids left no identifiable fossil record. Only through careful revision and critical study of the fossils upon which these age estimates are based would we be able to assess if this discrepancy is real or is an artifact of the methods. Did the asterids not leave a recognizable fossil record for 22 million years? Or, is the proposed 128 million years an inaccurate estimate?
Table 1

Estimated Times of Divergence of Relevant Angiosperm Groups Based on Fossil Estimates and Molecular Dating, Ages are Given in Millions of Years Before Present (MYBP)

Clade

Magallón et al., 1999

Wikström et al., 2001

Bremer et al., 2004

Crepet et al., 2004

Estimate

Fossil

Molecular

Molecular

Fossil

Angiosperms

158–179

113

Eudicots

131–147

100

Asterids

112–122

90

Cornales

69.5

106–114

128

Ericales

89.5

106–114

127

90

Euasterids

107–117

127

Campanulids

102–112

123

Aquifoliales

69.5

99–107

121

Apiales

69.5

85–90

113

Dipsacales

53.2

85–90

111

Asterales

29.3

101–94

112

Lamiids

102–112

123

Garryales

45.9

100–107

114

Gentianales

53.2

83–89

108

Solanales

53.2

82–86

106

Lamiales

37

71–74

106

The accurate dating of correctly identified fossils is critical in any age estimation, based either on fossil record alone or in molecular dating techniques with fossil calibration points. Even the best method for molecular dating will generate meaningless results if the original calibration point(s) on which the whole analysis is/are based, is not reliable (Crane et al., 2004; Graur & Martin, 2004; Benton & Donoghue, 2007). The two problem areas that need to be critically evaluated before a fossil taxon can be accepted as a reliable calibration point are: [1] the certainty of its taxonomic placement and [2] the correctness of the age assignment of the sediments in which the fossil was found, which implies an understanding and a correct interpretation of the geological time scale. This work is a step towards achieving a better understanding of the early Asterid fossil record; by evaluating the reliability of those fossils that could potentially represent the earliest members of asterid families, by producing a time scale for Asterid diversification based on the reliable fossils identified and applying minimum-age-dating, and by comparing this fossil-based time scale to molecular-based age estimates.

Materials and Methods

Literature Review

The evaluation of the fossil record started with a literature search of the fossils that have been published as having affinities with groups that today constitute the Asteridae (sensu Bremer et al., 2002). The search focused on the family level, that is, on the oldest fossils ever reported for each one of the 100–104 families that constitute the group. However, fossils unassigned to family but to higher taxonomic levels were also considered. In the assembling of this list, original descriptions and monographs were preferred.

Evaluation of Fossils

Each one of the fossils was evaluated with respect to the reliability of its identification by reviewing their protologues or monographs. Eight criteria were taken into consideration for each fossil, each one evaluated as provided/not provided by the authors. In order of decreasing reliability the criteria are: [1] inclusion of the fossil in a phylogenetic analysis, [2] discussion of key characters that place the fossil in a group, [3] list of key characters that place the fossil in the group, [4] full taxonomic description and diagnosis of the fossil, [5] photographs of the specimens, [6] drawings, diagrams or reconstructions of the fossils, [7] specimen information; housing institution, collection number, holotype designation, [8] collection information; locality, formation, age.

Once the list was compiled, it was subject to a filter designed to identify the reliable and well supported records by focusing on those fossils that fulfilled the first three criteria. These fossils were accepted as representing reliable records. The remainder of the list was subject to a second filter for which the criterion was the existence of a reliable older fossil belonging to the same family. That is, once a fossil was accepted as member of a family, any younger fossils assigned to that family were excluded from further analysis because they could no longer be considered evidence of the first appearance of that family. The fossils that were not removed by either filter are not only ambiguous and in need of revision but also potentially the earliest evidence for a family/order of Asteridae, that is, the putative oldest evidence of a lineage.

Age Determinations

The age assigned to the fossils follows the most recent accepted date for the sediments in which they are found, and not the age that was assigned to them when they were first described. This is important to consider, especially with regard to older reports in which boundaries for time periods were differently defined. For the purpose of assigning numerical dates to time periods, the upper bound (end) of that period as defined in the International Geologic Time Scale (Gradstein et al., 2004) was used.

Minimum Age Dating

The fossils accepted as reliable after applying criteria mentioned above were incorporated as minimum age indicators in a phylogeny of the Asteridae following the method of Crepet et al. (2004). The asterid cladogram used is based on the results of Bremer et al. (2002) modified by the substitution of particular clades that are now available and that represent more comprehensive and resolved cladograms for those particular groups: Zhang et al. (2003) for Dipsacales, Chandler and Plunkett (2004) for Apiales, Lundberg and Bremer (2003) for Asterales, Oxelman et al. (2005) for parts of Lamiales and Schönenberger et al. (2005) for Ericales.

Results

Fossil Record of the Asteridae

A total of 261 fossils once described as asterids were identified (Table 2). This list should not be considered exhaustive since many younger reports were not included in instances where older, reliable reports had been already listed. Also, reproductive structures were favored over vegetative structures because it is in the reproductive structures that synapomorphies and diagnostic characters of the groups are more likely to be found. Therefore, this list should not be considered a fair representation of the status of the uncritically assessed asterid fossil record; however, it does provide a more accurate assessment of asterid history and also represents a coarse approximation of the abundance of fossil reports for the different clades. From this listing, it can be seen that, although some families (e.g. Fouqueriaceae, Loasaceae) have no reported fossil record, in the end, all orders are represented in the fossil record.
Table 2

Summary of the Fossil Record of Asteridae by Orders. Numbers Indicate the Number of Fossil Occurrences Found During the Literature Review Before Evaluating Each Record. “Reproductive” Includes Macrofossil Remains of Flowers, Fruits and Seeds, “Vegetative” Mostly Includes Leaves and Wood. Six Fossils are Represented Twice Since They are Known From Organically Connected Reproductive and Vegetative Remains. “Unplaced Families” Include the Clades Escalloniaceae-Paracryphiaceae (Paracryphiales of APG, 2009), Icacinaceae-Oncothecaceae and Boraginaceae-Vahliaceae (see Fig. 1)

Order

Fossils

Reproductive

Vegetative

Pollen

Cornales

21

17

4

2

Ericales

81

49

24

10

Campanulids

56

40

6

10

Aquifoliales

5

3

0

2

Apiales

18

16

1

1

Dipsacales

8

5

2

1

Asterales

24

15

3

6

Lamiids

103

58

30

16

Garryales

11

7

2

2

Gentianales

24

9

7

8

Solanales

14

9

2

3

Lamiales

24

16

8

2

Unplaced families

31

18

12

1

TOTAL

261

164

65

38

The order of asterids with the best fossil record is the Ericales (Table 2), with 80 reports; however, this apparent abundance of ericalean taxa is misleading since a good portion of these records is based on reports of the genus Symplocos, monographed in 1949 by Kirchheimer. If the taxa described in that work were removed, only 48 records would remain, leaving the Ericales as the most frequently reported order of asterids, but with a more modest advantage.

Of the two clades of the Euasteridae, the Lamiids (Euasteridae I) present a more abundant fossil record than its counterpart, the Campanulids (Table 2). However, it is noticeable that despite the biodiversity these two groups display today, their fossil record combined is barely larger than the combined fossil record of the two early diverging orders, the Cornales and the Ericales.

Cornales

The Cornales is a well supported and well studied group, however, different authors treat the families Cornaceae, Nyssaceae and Mastixiaceae differently. Under some schemes, the families are treated as one broadly defined Cornaceae, while other authors prefer to treat them as separate, although closely related, families (see Xiang et al., 2002; Fan & Xiang, 2003). In this work, they will be referred to as different families, as that provides more information as to the inferred relationships of the fossils. The oldest reliable fossils for this clade are Hironoia fusiformis, a “cornalean” fruit from the Coniacian-Santonian of Japan and Tylerianthus crossmanensis, a fossil flower from the Turonian of New Jersey (Table 3; Fig. 2). Although initially Tylerianthus was described as having affinities with the Hydrangeaceae or the Saxifragaceae, the authors indicate that cladistic analyses placed it as sister to Hydrangeaceae. Later, Crepet et al. (2004) confirmed the placement of Tylerianthus in that family. This fossil places the Order in at least the Turonian (Fig. 2).
Table 3

Early Fossil Record of the Cornales. Acc = Accepted

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Hydrangeaceae

Tylerianthus crossmanensis Gandolfo, Nixon et Crepet 1998

Flower

Turonian

Raritan, NJ, USA

Gandolfo et al., 1998

Yes

Cornaceae

Cornus clarnensis Manchester 1994

Endocarp

Middle Eocene

Clarno, OR, USA

Manchester, 1994

Yes

Mastixioxylon symplocoides Meijer 2000

Wood

Middle-Late Santonian

Aachen, La Calamine, NE Belgium

Meijer, 2000

Yes

Cornaceae/Nyssaceae

Nyssoxylon sp.

Wood

Middle-Late Santonian

Aachen, La Calamine, NE Belgium

Meijer, 2000

Yes

Nyssaceae

Davidia antiqua (Newberry) Manchester 2002

Leaf/fruit

Early Paleocene

Bureya, SE Russia

Manchester, 2002

Yes

Davidia antiqua (Newberry) Manchester 2002

Leaf/fruit

Paleocene

Fort Union Gr, ND, MT, WY, USA

Manchester, 2002

Yes

Nyssa

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Nyssa sp.

Endocarp

Early Ypresian, Early Eocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

Tricolporopollenites kruschii Potonié 1934

Pollen

Paleocene

Wilcow flora, TX, USA

Elsik, 1968

No

Nyssaceae/Mastixiaceae

Hironoia fusiformis Takahashi, Crane et Manchester 2002

Fruit

Early Coniacian-Early Santonian

Ashizawa Fm, NE Honshu Japan

Takahashi et al., 2002

Yes

Mastixiaceae

Beckettia sp.

Fruit

Early Ypresian, Early Eocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

cf. Mastixia

Endocarp

Early Paleocene-Earliest Eocene

Fort Union, WY, USA

Tiffney & Haggard, 1996

Yes

cf. Mastixia

Fruit

Late Early Eocene-Early Middle Eocene

Sepulcher, MT-WY, USA

Tiffney & Haggard, 1996

Yes

Diplopanax eydei Stockey, LePage et Pigg 1998

Fruit

Middle Eocene

Princeton Chert, BC, Canada

Stockey et al., 1998

Yes

Langtonia bisulcata Reid et Chandler 1933

Endocarp

Early Eocene

London Clay, England

Reid & Chandler, 1933

Yes

Langtonia bisulcata Reid et Chandler 1933

Endocarp

Middle Eocene

Clarno, OR, USA

Manchester, 1994

Yes

Mastixia eydei Tiffney et Haggard 1996

Endocarp

Late Eocene

Auriferous Gravels, CA, USA

Tiffney & Haggard, 1996

Yes

Mastixia oregonense (Scott) Tiffney et Haggard 1996

Endocarp

Middle Eocene

Clarno, OR, USA

Tiffney & Haggard, 1996

Yes

Mastixicarpum occidentale Manchester 1994

Endocarp

Middle Eocene

Clarno, OR, USA

Tiffney & Haggard, 1996

Yes

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig2_HTML.gif
Fig. 2

Minimum age dating of the Cornales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Tylerianthus crossmanensis, Turonian (89.3 mya). B- Hironoia fusiformis, Santonian (83.5 mya)

Ericales

The ericalean clade in the both the Bremer et al. (2002) and Schönenberger et al. (2005) analyses is composed of two sister clades, one includes Marcgraviaceae, Balsaminaceae, Tetramerista and Pelliciera, the “balsaminoid” clade, and the other one includes the rest of the Ericales. From the balsaminoid clade, pollen grains representing Pellicieraceae (=Tetrameristaceae in the Schönenberger et al. (2005) analysis) and Marcgraviaceae have been reported from several localities around the neotropics (Table 4). However, none of these reports have provided a detailed account of the characters that identify these fossils as Pelliciera, Marcgravia or Norantea. For this reason these reports are not accepted in the present treatment. It is interesting to notice, however, that the clade has only been reported in the palynological fossil record, no macrofossils have been assigned to this group.
Table 4

Early Fossil Record of the Ericales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Marcgraviaceae

Marcgravia sp

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Norantea sp

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Tetrameristaceae (“Pellicieraceae”)

Pelliciera”-like

Pollen

Oligocene-Miocene

Simojovel, Chis, Mexico

Lagenheim et al., 1967

No

Pelliciera

Pollen

Early Middle Eocene

Chapelton, Jamaica

Graham, 1977

No

Pelliciera

Pollen

Eocene

Gatuncillo, Panamá

Graham, 1977

No

Pelliciera

Pollen

Oligocene-Miocene

La Boca, Panamá

Graham, 1977

No

Pelliciera

Pollen

Oligocene-Miocene

La Quinta, Chis, Mexico

Graham, 1999

No

Pelliciera

Pollen

Middle Oligocene

San Sebastián, Puerto Rico

Graham & Jarzen, 1969

No

Psilatricolporites crassus van der Hammen et Wigmstra 1964

Pollen

Tertiary

Guiana Basin, Guianas

Graham, 1977

No

Lecythidaceae

Barringtonioxylon arcotense Awasthi 1969

Wood

Tertiary

Cuddalore Series, India

Awasthi, 1969a

Yes

Barringtonioxylon eopterocarpum Prakash et Dayal 1964

Wood

Early Tertiary (Eocene?)

Deccan Intertrappean Beds, India

Prakash & Dayal, 1964

Yes

Careyoxylon pondicherriense Awasthi 1969

Wood

Tertiary

Cuddalore Series, India

Awasthi, 1969a

Yes

Sapotaceae

Chrysophyllum tertiarum Mehrotra 2000

Leaf

Late Paleocene

Nangwalbibra India

Mehrotra, 2000

Yes

Ebenaceae

Austrodiospyros cryptostoma Basinger et Christophel 1985

Flower/ leaf

Late Eocene

Anglesea, Victoria, Australia

Basinger & Christophel, 1985

Yes

Diospyros palaeoebenum Prasad 1994

Leaf

Middle Miocene-Pliocene

Siwalik, Nepal

Prasad & Pradhan, 1998

Yes

Ebenoxylon arcotense Awasthi 1969

Wood

Tertiary

Cuddalore Series, India

Awasthi, 1969b

Yes

Ebenoxylon kalagarhensis Prasad 1988

Wood

Middle Miocene

Siwalik, India

Prasad, 1988

No

Ebenoxylon kartikcherrense Prakash et Tripathi 1969

Wood

Late Miocene

Kartikcherra, India

Prakash & Tripathi, 1969

Yes

Ebenoxylon miocenicum Prakash 1978

Wood

Middle Miocene

Siwalik, India

Prasad, 1993

Yes

Ebenoxylon palaeocandoleana Prasad 1993

Wood

Middle Miocene

Siwalik, India

Prasad, 1993

Yes

Ebenoxylon siwalicus Prakash 1981

Wood

Middle Miocene

Siwalik, India

Prasad, 1993

Yes

Myrsinaceae

Ardisia palaeosimplicifolia Prasad 1994

Leaf

Middle Miocene-Pliocene

Siwalik, Nepal

Prasad & Pradhan, 1998

Yes

“Myrsinaceae”

Leaf

Early Miocene

Foulden Hills, New Zealand

Pole, 1996

Yes

Polemoniaceae

Gilisenium hueberi Lott, Manchester et Dilcher 1998

Plant

Middle Eocene

Green River, UT, USA

Lott et al., 1998

Yes

Theaceae

Andrewsiocarpon henryense Grote et Dilcher 1989

Seed/fruit

Middle Eocene

Claiborne Fm, KY, TN, USA

Grote & Dilcher, 1989

Yes

Gordonia lamkinensis Grote et Dilcher 1992

Fruit

Middle Eocene

Claiborne Fm, KY, USA

Grote & Dilcher, 1992

Yes

Gordonia warmanensis Grote et Dilcher 1992

Fruit

Middle Eocene

Claiborne Fm, TN, USA

Grote & Dilcher, 1992

Yes

Gordoniopsis polysperma Grote and Dilcher 1992

Fruit

Middle Eocene

Claiborne Fm, TN, USA

Grote & Dilcher, 1992

Yes

Pentaphylacaceae (“Ternstroemiaceae”)

Eurya crassitesta Knobloch 1975

Seed

Maastrichtian-Paleocene

Eisleben, Germany

Mai, 1987

Yes

Eurya microstigmosa Mai 1987

Seed

Early Paleocene

Gunna, Germany

Mai, 1987

Yes

Eurya stigmosa (Ludwig) Mai 1987

Seed

Paleocene

Eisleben, Germany

Mai, 1987

Yes

Pentaphylacaceae (“Sladeniaceae”)

Sladenioxylon africanum Giraud, Bussert et Schrank 1992

Wood

Albian-Cenomanian

Wadi Awatib, Sudan

Giraud et al., 1992

Yes

Pentaphylacaceae (“Pentaphylacaceae”)

Pentapetalum trifasciculan_ dricus Martínez-Millán, Crepet et Nixon, 2009

Flower

Turonian

Raritan, New Jersey

Martínez-Millán et al., 2009

Yes

Actinidiaceae/Theaceae

Paradinandra suecica Schönenberger et Friis 2001

Flower

Late Santonian-Early Campanian

Åsen, Scania, N Sweden

Schönenberger & Friis, 2001

Yes

Actinidiaceae

Actinidia argutaeformis Dorofeev 1963

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Actinidia faveolata Reid 1915

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Actinidia oregonensis Manchester 1994

Seed

Middle Eocene

Clarno, OR, USA

Manchester, 1994

Yes

Parasaurauia allonensis Keller, Herendeen et Crane 1996

Flower

Early Campanian

Gaillard Fm, Buffalo Creek GA, USA

Keller et al., 1996

Yes

Saurauia antiqua Knobloch et Mai 1986

Seed

Senonian-Santonian

Klikov-Schichtenfolge, Germany

Knobloch & Mai, 1986

Yes

Ericaceae

“Ericaceae”

Pollen

Oligocene-Miocene

La Quinta, Chis, Mexico

Graham, 1999

No

Paleoenkianthus sayrevillensis Nixon et Crepet 1993

Flower

Turonian

Raritan, New Jersey

Nixon & Crepet, 1993

Yes

Diapensiaceae

Actinocalyx bohrii Friis 1985

Flower

Late Santonian-Early Campanian

Åsen, Scania, S Sweden

Friis, 1985

Yes

Styracaceae

Rehderodendron stonei (Reid et Chandler) Mai 1970

Fruit

Eocene

Sabals d’ Anjou, France

Vaudois-Mieja, 1983

Yes

Styrax hradekense Schweigert 1992

Leaf

Oligocene

Hessenreuth, Germany

Schweigert, 1992

Yes

Symplocaceae

Durania ehrenbergii Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Palliopora symplocoides Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Sphenoteca gigantea Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Sphenoteca incurva Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Symplocos arecaeformis (Schlotheim) Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany & Poland

Kirchheimer, 1949

Yes

Symplocos braunii Kirchheimer 1949

Endocarp

Late Miocene

Hessen, Germany

Kirchheimer, 1949

Yes

Symplocos bureauana Saporta 1868

Leaf

Early Eocene

Marne, France

Kirchheimer, 1949

No

Symplocos casparyi Ludwig 1857

Endocarp

Early Miocene-Early Pliocene

Hessen, Germany

Kirchheimer, 1949

Yes

Symplocos cf. crataegoides Buchanan-Hamilton 1937

Endocarp

Early Pliocene

Hessen, Germany

Kirchheimer, 1949

No

Symplocos commutatifolia Berry 1938

Leaf

Eocene

Rio Pichilefu, Argentina

Kirchheimer, 1949

No

Symplocos detrita Velenovsky 1882

Leaf

Early Miocene

Vrsovice, Czech Rep.

Kirchheimer, 1949

No

Symplocos elongata Ludwig 1857

Endocarp

Early Pliocene

Hessen, Germany

Kirchheimer, 1949

No

Symplocos globosa Ludwig 1857

Endocarp

Early Pliocene

Hessen, Germany

Kirchheimer, 1949

No

Symplocos gothani Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Symplocos grimsleyi Tiffney 1999

Endocarp

Early Ypresian, Early Eocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

Symplocos headonensis Chandler 1926

Fruit

Late Eocene

Hampshire, UK

Kirchheimer, 1949

No

Symplocos kirstei Kirchheimer 1939

Endocarp

Early-Middle Oligocene

Thüringen, Germany

Kirchheimer, 1949

No

Symplocos laurifolia Hofmann 1926

Leaf

Miocene

Kathrein, Austria

Kirchheimer, 1949

No

Symplocos lignitarum (Quenstedt) Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

Yes

Symplocos ludwigii Kirchheimer 1949

Endocarp

Early Pliocene

Hessen, Germany

Kirchheimer, 1949

Yes

Symplocos microcarpa Reid 1920

Endocarp

Early Pliocene

Pont-de-Gail, France

Kirchheimer, 1949

No

Symplocos minutula (Sternberg) Kirchheimer 1949

Endocarp

Early Oligocene-Late Miocene

Rheinland, Germany, Switzerland, Austria, France, Czech Rep, Poland

Kirchheimer, 1949

Yes

Symplocos n. sp. Reid 1923

Endocarp

Early Pliocene

Pont-de-Gail, France

Kirchheimer, 1949

No

Symplocos oleaceae Ludwig 1858

Endocarp

Late Miocene

Hessen, Germany

Kirchheimer, 1949

Yes

Symplocos oregona Chaney et Sanborn 1933

Leaf

Late Eocene

Goshen, Oregon, USA

Kirchheimer, 1949

No

Symplocos poppeana Kirchheimer 1940

Endocarp

Middle-Late Oligocene

Lausitz, Germany

Kirchheimer, 1949

No

Symplocos pseudogregaria Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

No

Symplocos quadrilocularis Reid et Chandler 1933

Fruit

Early Eocene

Minster, Kent, UK

Kirchheimer, 1949

No

Symplocos radobojana Unger 1866

Fruit

Late Oligocene-Early Miocene

Radoboj, Yugoslavia, Czech Rep

Kirchheimer, 1949

No

Symplocos salzhausenensis (Ludwig) Kirchheimer 1949

Endocarp

Late Miocene

Hessen, Germany

Kirchheimer, 1949

No

Symplocos schereri Kirchheimer 1935

Endocarp

Middle-Late Oligocene

Rheinland, Germany

Kirchheimer, 1949

No

Symplocos smithii Florin 1920

Leaf

Late Pliocene

Amakusa, Japan

Kirchheimer, 1949

No

Symplocos sp. Kirchheimer 1949

Endocarp

Late Eocene-Early Pliocene

Hessen, Germany, England, Netherlands

Kirchheimer, 1949

No

Symplocos subspicata Friedrich 1883

Leaf

Eocene

Eisleben, Germany

Kirchheimer, 1949

Yes

Symplocos trilocularis Reid et Chandler 1933

Fruit

Early Eocene

Minster, Kent, UK

Kirchheimer, 1949

No

Symplocos urceolata Reid 1920

Fruit

Early Pliocene

Pont-de-Gail, France

Kirchheimer, 1949

No

Symplocos wiesaensis Kirchheimer 1949

Endocarp

Middle-Late Oligocene

Lausitz, Germany

Kirchheimer, 1949

No

The second clade includes families with no known fossil record such as Fouqueriaceae and Sarraceniaceae, and families with relatively extensive fossil records, like Ebenaceae and Symplocaceae. This group includes many arborescent members whose fossil record is mostly wood (e.g. Ebenaceae, Theaceae, Lecythidaceae).

The most remarkable of ericalean fossils are the fusainized flowers found in Cretaceous sediments in different localities of Eastern North America (New Jersey and Georgia), Europe (Sweden) and Japan (Crepet, 1996; Crane & Herendeen, 1996; Herendeen et al., 1999; Takahashi et al., 1999; Friis et al., 2006). The preliminary surveys of these floras indicate that ericalean flowers are abundant and diverse in these localities. Unfortunately many of them have not been formally described and the ones that have been described frequently show an array of characters not found in modern genera and sometimes not completely conforming to the families to which they are believed to be related. The fossil record also indicates that many modern families were well established by the Eocene (Fig. 3) as evidenced by very complete fossils that include both, vegetative and reproductive structures (i.e. Christophel & Basinger, 1982; Basinger & Christophel, 1985; Lott et al., 1998).
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig3_HTML.gif
Fig. 3

Minimum age dating of the Ericales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Schönenberger et al. (2005). A- Gordonia lamkinensis, Middle Eocene (40.4 mya). B- Gilisenium hueberi, Lutetian-Bartonian (40.4 mya). C- Symplocos grimsleyi, Ypresian (48.6 mya). D- Actinocalyx bohrii, Late Santonian-Early Campanian (83.5 mya). E- Pentapetalum trifasciculandricus, Turonian (89.3 mya). F- Austrodiospyros cryptostoma, Late Eocene (33.9 mya). G- Parasaurauia allonensis, Early Campanian (70.6 mya). H- Paleoenkianthus sayrevillensis, Turonian (89.3 mya)

As of today, the oldest representatives of Ericales are Paleoenkianthus sayrevillensis (Nixon & Crepet, 1993) and Pentapetalum trifasciculandricus (Martínez-Millán et al., 2009), both from the Turonian of New Jersey (Table 4). These two fossils have been associated to clades that in the most recent phylogenetic hypothesis are not too closely related (Fig. 3), the Ericaceae and a part of the Theaceae s.l. that Schönenberger et al. (2005) call the Pentaphylacaceae and Bremer et al. (2002), the Ternstroemiaceae.

Aquifoliales

This order sensu Bremer et al. (2002) includes only four families (Fig. 4), of these, only Aquifoliaceae has a fossil record (Table 5), starting with Maastrichtian fruits mentioned by Knobloch and Mai (1986) and Paleocene fruits reported by Mai (1987). Brown (1962) reported some leaves from the Early Tertiary of Colorado, however, Collinson et al. (1993) have pointed out the need for a critical revision of the these leaves. Pollen belonging to Ilexpollenites has been reported from the Late Cretaceous of South Australia (Martin, 1977).
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig4_HTML.gif
Fig. 4

Minimum age dating of the Aquifoliales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Ilex hercynica, Early Paleocene (61.7 mya)

Table 5

Early Fossil Record of the Aquifoliales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Aquifoliaceae

Ilex antiqua Knobloch et Mai 1986

Fruit

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Ilex gonnensis Mai 1970

Seed

Late Paleocene

Gonna, Germany

Mai, 1987

Yes

Ilex hercynica Mai 1970

Seed

Early Paleocene

Gonna, Germany

Mai, 1987

Yes

Ilex

Pollen

Middle Oligocene

San Sebastián, Puerto Rico

Graham & Jarzen, 1969

No

Ilexpollenites

Pollen

Campanian

San Joaquín Valley, CA, USA

Martin, 1977

No

Apiales

Seven families compose this order in the Bremer et al. (2002) cladogram; Apiaceae, Araliaceae, Aralidiaceae, Torriceliaceae, Melanophyllaceae, Griseliniaceae, and Pittosporaceae, of which three have been reported from the fossil record (Table 6). The more comprehensive analysis of Chandler and Plunkett (2004) confirms these families as Apialean lineages and also segregates some members of Apiaceae and Araliaceae, the Mackinlaya and the Myodocarpus groups (Fig. 5).
Table 6

Early Fossil Record of the Apiales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Torricelliaceae

Torricellia bonesii (Manchester) Manchester 1999

Fruit

Early Miocene

Oberdorf, Austria

Manchester, 1999

Yes

Torricellia bonesii (Manchester)Manchester 1999

Fruit

Middle Eocene

Roslyn, Washington

Manchester, 1999

Yes

Torricellia bonesii (Manchester) Manchester 1999

Fruit

Middle Eocene

Clarno, OR, USA

Manchester, 1999

Yes

Torricellia bonesii (Manchester) Manchester 1999

Fruit

Middle Eocene

Messel, Germany

Manchester, 1999

Yes

Araliaceae (“Apiaceae”)

Hydrocotyle sp Łańcucka-Środoniowa 1979

Fruit

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Araliaceae

Aralia antiqua Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Aralia cf. ucrainica Dorofeev 1963

Endocarp

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Aralia rugosa Dorofeev 1963

Endocarp

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Aralia tertiaria Dorofeev 1963

Endocarp

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Acanthopanax fiedrichii Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Acanthopanax gigantocarpus Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Acanthopanax mansfeldensis Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Acanthopanax obliquocostatus Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

Dendropanax eocenensis Dilcher et Dolph 1970

Leaf

Middle Eocene

Claiborne, TN, USA

Dilcher & Dolph, 1970

Yes

Dendropanax

Pollen

Middle Oligocene

San Sebastián, Puerto Rico

Graham & Jarzen, 1969

No

Oreopanax dakotensis Melchior 1976

Fruit

Paleocene

Wannagan Creek Flora, ND, USA

Melchior, 1976

No

Paleopanax oregonensis Manchester 1994

Endocarp

Middle Eocene

Clarno, Oregon

Manchester, 1994

Yes

Schefflera dorofeevii Łańcucka-Środoniowa 1975

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

No

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig5_HTML.gif
Fig. 5

Minimum age dating of the Apiales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Chandler and Plunkett (2004). A- Torricellia bonesii, Middle Eocene (40.4 mya). B- Dendropanax eocenensis, Middle Eocene (40.4 mya)

Torricelliaceae has representatives in the Eocene of Washington, Oregon and Germany and the Miocene of Austria (Table 6). Araliaceae has an extensive fossil record that goes back to the Late Cretaceous but that is in need of revision; few of the araliaceous fossils can be considered reliable and revising them would be of extreme importance. As of now, the oldest reliable record are the leaves of Dendropanax described by Dilcher and Dolph (1970) from the Eocene of Tennessee and fruits of Paleopanax Manchester (1994) from the Eocene of Oregon (Table 6; Fig. 5).

Escalloniaceae-Paracryphiaceae

In the Bremer et al. (2002) analysis, Escalloniaceae is resolved as polyphyletic, with Escallonia, Tribeles and Polyosma in one clade along with the Eremosynaceae (the newly recognized Order Escalloniales of APG, 2009), and Paracryphiaceae and Quintinia in a second clade, sister to the Dipsacales (the Order Paracryphiales of APG, 2009). The uncertainty about the relationships within the family makes it difficult to place fossils in proper context. However, fossil pollen from the Upper Eocene of New Zealand has been related to the genus Quintinia (Mildenhall, 1980), as have fusainized fossil flowers from the Late Santonian-Early Campanian of Southern Sweden (Friis, 1990).

Silvianthemum suecicum, the flower-based taxon from Sweden, was not put in phylogenetic context when described but it is still considered by its authors to be related to Quintinia (Friis et al., 2006). In one attempt to assess its phylogenetic relationships, Backlund (1996) added it to the Dipsacales matrix analyzed by Backlund and Donoghue (1996) concluding that Silvianthemum occupied “a stable but not strongly supported position … just outside the basal node of the Dipsacales”, although it is shown as sister to a Dipsacales-Apiales clade that also includes Tribeles, Polyosma and Bruniaceae, a result not fully compatible with current hypotheses of Campanuliid (Euasterid II) relationships. For this reason, a new analysis was performed using a fixed backbone based on the analyses by Bremer et al. (2002), Zhang et al. (2003) and Donoghue et al. (2003); the 58 taxa and characters 32–60 of the Backlund and Donoghue (1996) morphology matrix; and the Friis (1990) description of Silvianthemum to code its characters. Details of this analysis can be found in “Appendix B”. The strict consensus places Silvianthemum as sister to Quintinia with dorsifixed anther attachment as synapomorphy (Fig. 6). Although under the current phylogenetic framework, Silvianthemum is better placed with Quintinia, this might change when the phylogenetic relationships among members of the apparently polyphyletic Escalloniaceae are better understood.
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig6_HTML.gif
Fig. 6

Strict consensus of 312 trees (L = 204, CI = 32, RI = 68) showing the position of the fossil Silvianthemum as sister to Quintinia with dorsifixed anther attachment (ch 52) as synapomorphy (see “Appendix B” for details). Numbers in brackets indicate number of descendant terminals

Dipsacales

Several fossils have been described as leaves of Viburnum, however many are dubious and many have been shown to represent different taxa; for example, those described by Brown (1962) were transferred to Davidia of the Cornales (Manchester, 2002). The macrofossil record of the Dipsacales was reviewed by Manchester and Donoghue (1995) and several reports that were rejected by those authors have not been included here (i.e. fossils formerly referred to Abelia). Bell and Donoghue (2005) have also evaluated the available fossil record of the Dipsacales when searching for suitable calibration points for their molecular age estimation analysis and found that Diplodipelta is the oldest most reliable fossil of this group. Diplodipelta places the Dipsacales in the Late Eocene (Table 7; Fig. 7) and although it was not placed in phylogenetic analysis as a terminal, enough synapomorphies were found to confidently place it as sister of Dipelta (Manchester & Donoghue, 1995).
Table 7

Early Fossil Record of the Dipsacales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Adoxaceae

Sambucus

Seed/fruit

Middle Eocene

Messel, Germany

Collinson, 1988

No

Sambucus

Leaf

Late Eocene

Florissant, CO, USA

Manchester, 2001

No

Caprifoliaceae (“Diervillaceae”)

Diervilla echinata Piel 1971

Pollen

Oligocene

Fraser River, BC, Canada

Piel, 1971

Yes

Caprifoliaceae

“Caprifoliaceae”

Seed/fruit

Middle Eocene

Clarno, OR, USA

Bones, 1979

No

Symphoricarpos elegans (Lesquereux) Smith 1937

Leaf

Eocene

Ruby River Basin, MT, USA

Becker, 1961

No

Linnaeaceae

Dipelta europaea Reid et Chandler 1926

Fruit

Late Eocene-Early Oligocene

Bembridge, UK

Reid & Chandler, 1926

Yes

Diplodipelta miocenica (Berry) Manchester et Donoghue 1995

Fruit

Miocene

Succor Creek, WA, ID, OR, USA

Manchester & Donoghue, 1995

Yes

Diplodipelta reniptera (Becker) Manchester et Donoghue 1995

Fruit

Late Eocene-Oligocene

Florissant, Mormon Cr, Ruby, CO, MT, USA

Manchester & Donoghue, 1995

Yes

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig7_HTML.gif
Fig. 7

Minimum age dating of the Dipsacales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Zhang et al. (2003). A- Diplodipelta reniptera, Late Eocene-Oligocene (33.9 mya)

Asterales

The most recent phylogenetic study of the Asterales is that of Lundberg and Bremer (2003) where they include 12 families in the order. Few of these are represented in the fossil record; macrofossil records include seeds assigned to Campanulaceae and Menyanthaceae (Table 8), a stem assigned to Donatia (Donatiaceae or Stylidaceae) and fruits assigned to Asteraceae (Table 8). In this group, the pollen record is more extensive than the macrofossil one, with families like Stylidaceae s.s. and Goodeniaceae known only from fossil pollen.
Table 8

Early Fossil Record of the Asterales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Campanulaceae

Campanula palaeopyrami_ dalis Łańcucka-Środoniowa 1977

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1977

Yes

Campanula sp. Łańcucka-Środoniowa 1979

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Stylidiaceae

Tricolpites stylidioides Macphail et Hill 1994

Pollen

Early Oligocene

Lemonthyme Creek, NW Tasmania

Macphail & Hill, 1994

No

Donatia novae-zelandiae Hook f. 1853

Stem

Pleistocene

Comstock, King River Valley, Tasmania

Gibson et al., 1987

Yes

Menyanthaceae

Menyanthes cf. trifoliata L 1753

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Striasyncolpites laxus Mildenhall et Pocknall 1989

Pollen

Late Oligocene-Middle Miocene

Cullen, Tierra del Fuego, Argentina

Zamaloa, 2000

No

Goodeniaceae

Poluspissusites digitatus Salard-Cheboldaeff 1978

Pollen

Oligocene

Kwa-Kwa, Cameroon

Salard-Cheboldaeff, 1978

Yes

Asteraceae

Cypselites aquensis Saporta 1889

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Cypselites fractus Saporta 1889

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Cypselites gypsorum Saporta 1861

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1862

No

Cypselites philiberti Saporta 1872

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1873

No

Cypselites spoliatus Saporta 1889

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Cypselites stenocarpus Saporta 1872

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1873

No

Cypselites tenuirostratus Saporta 1889

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Cypselites trisulcatus Saporta 1889

Fruit

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Hieracites nudatus Saporta 1889

Head

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Hieracites salyorum Saporta 1861

Leaf

Oligocene

Aix-en-Provence, France

Saporta, 1862

No

Hieracites stellatus Saporta 1889

Head

Oligocene

Aix-en-Provence, France

Saporta, 1889

No

Mutisiapollis patersonii Macphail et Hill 1994

Pollen

Early Oligocene

Lemonthyme Creek, NW Tasmania

Macphail & Hill, 1994

Yes

Parthenites priscus Saporta 1861

Leaf

Oligocene

Aix-en-Provence, France

Saporta, 1862

No

Tubulifloridites antipodica Cookson ex Potonie 1960

Pollen

Late Paleocene-Eocene

Koingnaas, South Africa

Zavada & de Villiers, 2000

Yes

Tubulifloridites viteauensis Barreda 1993

Pollen

Eocene

Shearwater Bay, South Africa

Zavada & de Villiers, 2000

Yes

Viguiera cronquistii Becker 1969

Head

Late Oligocene-Early Miocene

Beaverhead Basin, sw MT, USA

Crepet & Stuessy, 1978

No

“Asteraceae”

Fruit

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

No

The fossil record of the Asteraceae, one of the most species-rich families of flowering plants, has been elusive. Graham, in 1996, reviewed the available fossil record up to that point, however, that information is now outdated. In 2000, Zavada and de Villiers reported pollen grains of the tribe Mutisiae from South Africa and assigned them the name Tubulifloridites antipodica (Table 8). These pollen grains were assigned a Late Paleocene-Eocene age and became the oldest fossils for the family and for the order. But Scott et al. (2006) cast doubts on their identity suggesting that the South African T. antipodica is probably conspecific with T. viteauensis, a second taxon described by Zavada and de Villiers (2000) from Middle Eocene (Bartonian) sediments (Scott et al., 2006) offshore the coast of Namibia (Table 8; Fig. 8).
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig8_HTML.gif
Fig. 8

Minimum age dating of the Asterales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Lundberg and Bremer (2003). A- Campanula palaeopyramidalis, Miocene (5.33 mya). B- Donatia novae-zelandiae, Pleistocene (0.01 mya). C- Tubulifloridites viteauensis (37.2 mya)

Many reports of asteraceous macrofossils have been discredited: Cypselites has been reinterpreted as representing seeds of Apocynaceae instead of achenes of Asteraceae (Reid & Chandler, 1926; Manchester, 2001), Viguiera cronquistii had no characters linking it definitely to the Asteraceae (Crepet & Stuessy, 1978) and Parthenites priscus is not even a real fossil (pers. obs).

Oncothecaceae-Icacinaceae

In the Bremer et al. (2002) analysis, the Icacinaceae turned out as polyphyletic, with Apodytes as sister to Oncotheca in one clade and Cassinopsis, Icacina and Pyrenacantha in a different clade, more closely related to the rest of the Lamiid groups than to the Apodytes clade (Fig. 1). This partly agrees with the results of Kårehed (2001) who recovered a Garryales-Apodytes group and a separate Icacina group. This condition makes it difficult to assign some fossils to particular clades, especially when the fossil is of a genus not represented in the phylogenetic analysis. As of today, the position and relationships of the members of Icacinaceae are still unresolved and fluctuating (APG, 2009). For example, in the analysis of Soltis et al. (2007), the only member of Icacinaceae included, Icacina, was recovered as sister to a clade that includes all Lamiid groups except a clade Garryales-Oncotheca. This contrasts with the analysis of Bremer et al. (2002) where the Icacina would be sister to all other Lamiids including Garryales but not Oncotheca.

The earliest reproductive structures assigned to the Icacinaceae s.l. are the endocarps of Iodes germanica from the Maastrichtian of Germany (Table 9) while the earliest vegetative structures seem to go back to the Late Albian with the fossil wood Icacinoxylon (Table 9). Collinson et al. (1993), however, cast doubts on the identity of these woods and suggest the need for a revision. A review of fossil endocarps of the Icacinaceae is found in Pigg et al. (2008).
Table 9

Early Fossil Record of the Clades Oncotheca-Icacinaceae and Icacinaceae

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Icacinaceae

Calatoloides eocenicum Berry 1922

Fruit

Eocene

Wilcox, TX, USA

Berry, 1922

No

Goweria bibaiensis Tanai 1990

Leaf

Middle Eocene

Hokkaido, Japan

Tanai, 1990

Yes

Hosiea marchiaca Mai 1987

Endocarp

Middle Paleocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

Hosiea pterojugata Mai 1987

Endocarp

Late Paleocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

Huziokaea eoutilus (Endo) Tanai 1990

Leaf

Late Eocene

Hokkaido, Japan

Tanai, 1990

Yes

Icacinicarya dictyota Pigg, Manchester et DeVore

Endocarp

Late Paleocene

Beicegel Creek, ND, USA

Pigg et al., 2008

Yes

Icacinicarya collinsonae Pigg, Manchester et DeVore

Endocarp

Late Paleocene

Almont, ND, USA

Pigg et al., 2008

Yes

Icacinicaryites corruga (Brown) Pigg, Manchester et DeVore

Endocarp

Late Paleocene

USGS 9492, CO, USA

Pigg et al., 2008

Yes

Icacinicaryites linchensis Pigg, Manchester et DeVore

Endocarp

Late Paleocene

Linch WY, USA

Pigg et al., 2008

Yes

Icacinoxylon alternipunctata Wheeler, Lee et Matten 1987

Wood

Maastrichtian

McNairy Fm, IL, USA

Wheeler et al., 1987

Yes

Icacinoxylon pittiense Thayn, Tidwell et Stokes 1985

Wood

Late Albian

Cedar Mountain, UT, USA

Thayn et al., 1985

No

Iodes germanica Knobloch et Mai 1986

Endocarp

Maastrichtian

Eisleben, Germany

Knobloch & Mai, 1986

Yes

Iodes multireticulata Reid et Chandler 1933

Endocarp

Early Ypresian, Early Eocene

Nanjemoy, VA, USA

Tiffney, 1999

Yes

Iodes multireticulata Reid et Chandler 1933

Endocarp

Middle Eocene

Clarno, Oregon, USA

Manchester, 1994

Yes

Iodes multireticulata Reid et Chandler 1933

Fruit

Early Eocene

London Clay, England

Reid & Chandler, 1933

Yes

Merrilliodendron ezoanum Tanai 1990

Leaf

Late Eocene

Hokkaido, Japan

Tanai, 1990

Yes

Phytocrene microcarpa Scott et Barghoorn 1957

Fruit

Early Late Cretaceous

Raritan, NY, USA

Scott & Barghoorn, 1957

Yes

Phytocrene ozakii Tanai 1990

Leaf

Late Eocene

Hokkaido, Japan

Tanai, 1990

Yes

Pyrenacantha sp

Leaf

Late Eocene

Hokkaido, Japan

Tanai, 1990

Yes

Garryales

Despite the small size of this clade—only of two families, Garryaceae (including Aucuba) and Eucommiaceae (Bremer et al., 2002)—, the fossil record is somewhat extensive. Garryaceae has been reported from the Miocene (Table 10). Eucommiaceae, on the other hand, is well documented from the fossil record starting from the Late Early Eocene (Table 10; Fig. 9) which underwent revision by Call and Dilcher (1997), Manchester (1999) and Manchester et al. (2009).
Table 10

Early Fossil Record of the Garryales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Garryaceae

Garrya axelrodi Wolfe 1964

Leaf

Late Miocene

Stewart Spr, NV, USA

Wolfe, 1964

Yes

Eucommiaceae

Eucommia cf. E. ulmoides Leopold and Clay-Poole 2001

Pollen

Late Eocene

Florissant, CO, USA

Leopold and Clay-Poole, 2001

Yes

Eucommia constans Magallón-Puebla et Cevallos-Ferriz 1994

Fruit

Early Oligocene

Pie de Vaca, Pue, Mexico

Magallón-Puebla and Cevallos-Ferriz, 1994

Yes

Eucommia eocenica (Berry) Brown 1940

Fruit

Middle Eocene

Claiborne, Tenn, USA

Call & Dilcher, 1997

Yes

Eucommia jeffersonensis Call et Dilcher 1997

Fruit

Late Eocene

Lower John Day OR USA

Call & Dilcher, 1997

Yes

Eucommia kobayashii Huzioka 1961

Fruit

Eocene

Yubari, Hokkaido, Japan

Huzioka, 1961

Yes

Eucommia montana R. W. Brown 1940

Fruit

Late Eocene

Florissant, CO, USA

Manchester, 2001

No

Eucommia montana R. W. Brown 1940

Fruit

Late Early Eocene

Republic, WA, USA

Call & Dilcher, 1997

Yes

Eucommia rolandii Call et Dilcher 1997

Leaf

Middle Eocene

Talahatta, Holly Spr, MS, USA

Call & Dilcher, 1997

Yes

Eucommia ulmoides Oliv 1890

Fruit

Late Miocene (Tortonian)

Poland

Szafer, 1961

Yes

Tricolpites sp. cf. Eucommia

Pollen

Late Paleocene

Powder River Basin, WY, MT, USA

Pocknall, 1987

No

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig9_HTML.gif
Fig. 9

Minimum age dating of the Garryales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Eucommia montana, Late Early Eocene (48.6 mya)

Gentianales

The fossil record of this order, which contains two of the most species-rich families of angiosperms, dates back to the Early Tertiary, probably Paleocene but most likely Eocene (Table 11; Fig. 10). In a preliminary survey of the Black Peaks Formation from the Paleocene of Texas, Abbott (1986) mentioned the presence of wood of Rubiaceae, however it was never described. Graham (2009), who recently reviewed the fossil record of the Rubiaceae also accepts Emmenopterys as one of the its oldest members (Table 11). Fossils identified as Apocynaceae are relatively common in the Early Oligocene of England (Table 11) and probably elsewhere in Europe if the reports of Cypselites are proven to be apocynaceous (Reid & Chandler, 1926; Manchester, 2001). Woods from the Maastrichtian with affinities to the Apocynaceae were described by Wheeler et al. (1987), however, formal assignation to the family was never made. Gentianaceae has been reported from the Eocene (Table 11) based on a preserved flower and the pollen contained in it, however, in spite of the very distinctive pollen the paucity of other floral characters casts some doubt on this identification (Crepet, pers. comm.).
Table 11

Early Fossil Record of the Gentianales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Rubiaceae

Emmenopterys dilcheri Manchester 1994

Fruit

Middle Eocene

Clarno, OR, USA

Manchester, 1994

Yes

Faramea

Pollen

Middle Oligocene

San Sebastián, Puerto Rico

Graham & Jarzen, 1969

No

Remijia tenuiflorifolia Berry 1938

Leaf

Lutetian (Middle Eocene)

Laguna del Hunco, Argentina

Rodriguez de Sarmiento and Durango de Cabrera, 1995

No

Retitricolporites annulatus Salard-Cheboldaeff 1978

Pollen

Oligocene-Early Miocene

Kwa-Kwa, Cameroon

Salard-Cheboldaeff, 1978

Yes

“Rubiaceae”

Wood

Paleocene

Black Peaks Fm, TX, USA

Abbott, 1986

No

Loganiaceae

“Loganiaceae”

Pollen

Pliocene

Cerro la Popa, Colombia

Sole de Porta, 1960

No

Gentianaceae

Pistillipollenites mcgregorii Rouse

Pollen

Early Eocene

Wasatch Fm, WY, USA

Pocknall, 1987

No

Voyrioseminites magnus Trivedi and Chaturvedi 1972

Seed

Eocene

Kuala Lumpur, Malaysia

Trivedi and Chaturvedi, 1972

Yes

“Gentianaceae”

Flower

Early Eocene

Wilcox, TX, USA

Crepet & Daghlian, 1981

Yes

Apocynaceae

Apocynophyllum helveticum Heer 1859

Leaf

Middle Eocene

Messel, Germany

Wilde, 1989

Yes

Apocynospermum dubium Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Apocynospermum elegans Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Apocynospermum rostratum Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Apocynospermum striatum Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Apocynospermum

Seed

Late Eocene

Florissant, CO, USA

Manchester, 2001

No

Brevicolporites molinae (Schuler et Doubinger) Salard-Cheboldaeff 1978

Pollen

Oligocene-Early Miocene

Kwa-Kwa, Cameroon

Salard-Cheboldaeff, 1978

Yes

Echitonium ashczisaicum Vassiljev 1976

Leaf

Early Tertiary

Aktyubinsk, Kazakhstan

Vassiljev, 1976

Yes

Echitonium sophiae O. Web 1852

Leaf

Early Tertiary

Aktyubinsk, Kazakhstan

Vassiljev, 1976

Yes

Euholarrhenoxylon aisnense Gros 1993

Wood

Lutetian

Aisne valley, France

Gros, 1993

Yes

Paraapocynaceoxylon barghoorni Wheeler, Lee et Matten 1987

Wood

Maastrichtian

McNairy Fm, Illinois

Wheeler et al., 1987

Yes

Phyllantera vectensis Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Rauwolfia

Pollen

Middle Oligocene

San Sebastián, Puerto Rico

Graham & Jarzen, 1969

No

Tabernaemontana cf. T. coronaria Leopold and Clay-Poole 2001

Pollen

Late Eocene

Florissant, CO, USA

Leopold and Clay-Poole, 2001

Yes

Apocynaceae (“Asclepiadaceae”)

Polyporotetradites laevigatus Salard-Cheboldaeff 1978

Pollen

Oligocene-Early Miocene

Kwa-Kwa, Cameroon

Salard-Cheboldaeff, 1978

Yes

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig10_HTML.gif
Fig. 10

Minimum age dating of the Gentianales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Emmenopterys dilcheri, Middle Eocene (40.4 mya). B- Apocynospermum rostratum, Early-Middle Oligocene (28.4 mya)

Vahliaceae-Boraginaceae

In the analysis of Bremer et al. (2002), Vahliaceae and Boraginaceae form a clade (Fig. 1), however this arrangement is different in other analyses (e.g. Soltis et al., 2007). The fossil record of the Boraginaceae is restricted to the Tertiary starting in the Early Eocene (Table 12; Fig. 11). Boraginaceous seeds were reported from India (Table 12) and assigned a Paleocene age based on the presence of these same seeds, this age assignment was arbitrary and cannot be considered reliable.
Table 12

Early Fossil Record of the Vahliaceae-Boraginaceae Clade

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Vahliaceae

Scandianthus costatus Friis et Skarby 1982

Flower

Late Santonian-Early Campanian

Åsen, Scania, S Sweden

Friis & Skarby, 1982

Yes

Scandianthus major Friis et Skarby 1982

Flower

Late Santonian-Early Campanian

Åsen, Scania, S Sweden

Friis & Skarby, 1982

Yes

Boraginaceae

Cordia amurensis (Kryshtofovich et Baikovskaya) Chelebayeva 1984

Leaf

Paleogene

Kamchatka, Russia

Chelebayeva, 1984

Yes

Cordia congerminalis (Hollick) Chelebayeva 1984

Leaf

Paleogene

Kamchatka, Russia

Chelebayeva, 1984

Yes

Cordia kamtschatica Chelebayeva 1984

Leaf

Paleogene

Kamchatka, Russia

Chelebayeva, 1984

Yes

Cordia ochotensis Chelebajeva 1984

Leaf

Paleogene

Kamchatka, Russia

Chelebayeva, 1984

Yes

Cordia platanifolia (Ward) Chelebayeva 1984

Leaf

Paleogene

Kamchatka, Russia

Chelebayeva, 1984

Yes

Ehretia clausentia Chandler 1961

Fruit

Early Eocene

London clay, England

Chandler, 1964

Yes

Lithospermum dakotense Gabel 1987

Fruit

Late Miocene

Ash Hollow, Bennett, SD, USA

Gabel, 1987

Yes

Tournefortia

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

“Boraginaceae”

Seed

Paleocene?

Lameta Beds of Gujarat, India

Mathur & Mathur, 1985

No

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig11_HTML.gif
Fig. 11

Minimum age dating of the Vahliaceae-Boraginaceae clade. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Scandianthus major, Late Santonian-Early Campanian (83.5 mya). B- Ehretia clausentia, Early Eocene (48.6 mya)

Two fusainized flowers from the Late Cretaceous of Sweden, Scandianthus major and S. costatus, were originally described as belonging to the Saxifragalean complex and compared to Hydrangeaceae, Vahliaceae, Escalloniaceae and Saxifragaceae (Friis & Skarby, 1982; Friis, 1984). Subsequent phylogenetic analyses spanning all of the angiosperms have shown that those families are not closely related. Nevertheless, the authors have maintained the fossils close to the Vahliaceae (Friis et al., 2006) despite the fact that they had not been subject to phylogenetic analysis or had their characters reviewed in light of more recent phylogenetic hypotheses (Hermsen et al., 2006). For this reason, the comparison table included in the protologue of Scandianthus was adapted for phylogenetic analysis using a fixed backbone compatible with recent hypotheses of angiosperm relationships—Soltis et al. (2000), Bremer et al. (2002) and APG (2003). The final matrix has 12 morphological characters, 28 families and the fossil genus Scandianthus. In this analysis Scandianthus was resolved as sister taxa to Vahliaceae with one locule and pendant placenta as synapomorphies (Fig. 12). Details of this analysis can be found in the “Appendix C”.
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig12_HTML.gif
Fig. 12

Part of the strict consensus of 32 trees (L = 62, CI = 22, RI = 28) showing the position of the fossil genus Scandianthus as sister to Vahliaceae with one locule (ch 7) and pendant placentae (ch 10) as synapomorphies (see “Appendix C” for details and full tree)

Solanales

This clade is composed of five families in the cladogram of Bremer et al. (2002) and all of them have scarce or nonexistent fossil records based on existing literature. The Convolvulaceae fossil record is mostly represented by pollen; however there is a leaf from the Late Eocene (Table 13) that could be assigned to the group.
Table 13

Early Fossil Record of the Solanales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Convolvulaceae

Convulvulites orichitus MacGinitie 1953

Leaf

Late Eocene

Florissant, CO, USA

MacGinitie, 1953

Yes

Merremia

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Tricolpites trioblatus Mildenhall et Pocknall 1989

Pollen

Early-Middle Miocene

Etadunna, Lake Hydra, Australia

Martin, 2000

Yes

Solanaceae

Cantisolanum daturoides Reid et Chandler 1933

Fruit

Early Eocene

London clay, England

Reid & Chandler, 1933

No

Datura cf. D. discolor Leopold and Clay-Poole 2001

Pollen

Late Eocene

Florissant, CO, USA

Leopold and Clay-Poole, 2001

Yes

Physalis pliocaenica Szafer 1947

Fruit

Late Miocene (Tortonian)

Stare Gliwice, Poland

Szafer, 1961

No

Solanites brongniartii Saporta 1862

Flower

Oligocene

Aix-en-Provence, France

Saporta, 1862

No

Solanites crassus Berry 1930

Flower

Early Eocene

Claiborne, TN, USA

Berry, 1930

No

Solanites pusillus Berry 1930

Flower

Early Eocene

Claiborne, TN, USA

Berry, 1930

No

Solanites saportanus Berry 1916

Flower

Early Eocene

Claiborne, TN, USA

Berry, 1916

No

Solanites sarachaformis Berry 1930

Flower

Early Eocene

Claiborne, TN, USA

Berry, 1930

No

Solanispermum reniforme Chandler 1957

Seed

Eocene

Lower Bagshot, UK

Chandler, 1962

Yes

Solanum arnense Chandler 1962

Seed

Eocene

Lower Bagshot, UK

Chandler, 1962

Yes

Solandra haeliadum Massalongo 1851

Leaf

Eocene

Salcedo, Italy

Massalongo, 1851

No

The Solanaceae, however, has had a few fossils assigned to it, although most of these reports have been unreliable due to poor preservation, poor descriptions, or poor comparative work (Table 13). Cantisolanum daturoides has frequently been cited as the oldest evidence for Solanaceae (i.e. Knapp, 2002), however, the taxon is only known from the type specimen, a seed which does not show enough characters to support its assignment to Solanaceae or to any other family (Collinson, 1983; pers. obs.). Several flowers assigned to Solanaceae from the Eocene of Eastern North America (Table 13) by Berry (1914, 1916, 1930) clearly do not show characters of this family and are therefore, also rejected (Martínez-Millán, unpubl.). The flower-based taxon Solanites brongniartii from the Oligocene of France and the seed-based taxon Solanispermum reniforme from the Eocene of England (Table 13; Fig. 13) show solanaceous characters and could potentially belong in this family. The oldest pollen record is probably that of Datura from the Late Eocene, although details about the structure of these grains were not provided (Table 13).
https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig13_HTML.gif
Fig. 13

Minimum age dating of the Solanales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002). A- Solanispermum reniforme, Eocene (33.9 mya)

Lamiales

Despite the size of this clade in terms of number of families (Bremer et al., 2002), the fossil record is sparse, with few examples from the Tertiary. Some isolated Eocene fossils had been reported in the past, however Collinson et al. (1993) reported most of these as unconfirmed or rejected them based on the poor preservation. Only one record of Oleaceae, Fraxinus excelsior, based on both reproductive and vegetative organs can be considered reliably supported by available evidence (Table 14; Fig. 14). In the Plantaginaceae, Acanthaceae and Pedaliaceae, all of them with very scarce fossil records, there are only one or two reports which seem to be reliable (Table 14; Fig. 14), while in the Bignoniaceae, a family a more extensive fossil record, around half of its reports should be considered unreliable. The Lamiaceae has been elusive in the fossil record; two fossils described by Cockerell (1926, 1927) were later rejected by MacGinitie (1953, 1969) leaving the fruits from the Bembridge flora in England, as the oldest fossils of this family (Table 14; Fig. 14).
Table 14

Early Fossil Record of the Lamiales

Fossil taxon

Organ

Age

Locality

Reference

Acc.

Oleaceae

Fraxinus cf. rupinarum Becker 1961

Fruit

Middle Eocene

Quilchena, BC, Canada

Mathewes & Brooke, 1971

No

Fraxinus excelsior L

Leaf/fruit

Late Miocene

Depresión Ceretana, Spain

Barron, 1992

Yes

Fraxinus leii Berry 1934

Leaf

Maastrichtian

Lance Flora, SD, USA

Berry, 1934

No

Fraxinus rupinarum Becker 1961

Fruit

Oligocene

Ruby River Basin, MT, USA

Becker, 1961

No

Plantaginaceae (“Scrophulariaceae”)

Gratiola tertiaria Łańcucka-Środoniowa 1977

Seed

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1977

Yes

Acanthaceae

Acanthus rugatus Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Bignoniaceae

Catalpa coloradensis (Axelrod) Wolfe et Schorn 1990

Leaf

Oligocene

Creede Flora, CO, USA

Wolfe & Schorn, 1990

No

Catalpa rugosa Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Catalpa sp

Leaf

Oligocene

Creede Flora, CO, USA

Wolfe & Schorn, 1989

No

Incarvillea pristina Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Jacaranda

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Radermachera pulchra Reid et Chandler 1926

Seed

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Tecoma

Pollen

Middle Oligocene

San Sebastián Puerto Rico

Graham & Jarzen, 1969

No

Verbenaceae

Gmelina tertiara Bande 1986

Wood

Paleogene

Deccan Intertrappean Beds, India

Bande, 1986

Yes

Holmskioldia quilchenensis Mathewes et Brooke 1971

Calyx

Middle Eocene

Quilchena, BC, Canada

Mathewes & Brooke, 1971

No

Holmskioldia speiri (Lesquereux) MacGinitie 1953

Leaf/fruit

Oligocene

Ruby River Basin, MT, USA

Becker, 1961

No

“Verbenaceae”

Wood

Paleocene

Black Peaks Fm, TX, USA

Abbott, 1986

No

Pedaliaceae

Trapella cf. antennifera (Léveillé) Glück

Fruit

Pliocene

Swisterberg, Germany

Tralau, 1965

Yes

Trapella weylandi (Thomson et Grebe) Tralau 1964

Fruit

Pliocene

Swisterberg/Weilerswist, Germany

Tralau, 1964

Yes

Lamiaceae

Ajuginucula smithii Reid et Chandler 1926

Fruit

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Lycopus cf. antiquus Reid 1926

Fruit

Miocene

Nowy Sacz Basin, Poland

Łańcucka-Środoniowa, 1979

Yes

Melissa parva Reid et Chandler 1926

Fruit

Early-Middle Oligocene

Bembridge, England

Reid & Chandler, 1926

Yes

Menthites eocenicus Cockerell 1926

Calyx

Eocene

Green River, CO, USA

Cockerell, 1926

No

Nepeta? pseudaeluri Cockerell 1927

Leaf

Miocene

Florissant, CO, USA

Cockerell, 1927

No

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig14_HTML.gif
Fig. 14

Minimum age dating of the Lamiales. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Oxelman et al. (2005). A- Fraxinus excelsior, Late Miocene (5.33 mya). B- Gratiola tertiaria, Miocene (5.33 mya). C- Acanthus rugatus, Early-Middle Oligocene (28.4 mya). D- Radermachera pulchra, Early-Middle Oligocene (28.4 mya). E- Trapella weylandi, Pliocene (1.8 mya). F- Melissa parva, Early-Middle Oligocene (28.4 mya)

Dating of the Asterid Tree

The estimated minimum age estimated for the whole of the Asteridae is the Turonian (Late Cretaceous), some 89.3 my ago (Table 15; Fig. 15) with the oldest fossils appearing simultaneously in the Cornales and the Ericales (Table 15). The Euasteridae makes its appearance shortly after, in the Late Santonian-Early Campanian, some 83.5 my ago (Table 15; Fig. 15) when the oldest fossils of its two clades, the lamiids and the campanulids, make their first appearances simultaneously (Table 15; Fig. 15). Diversification within these two clades took place shortly after. By the Eocene, most orders were present in the fossil record. Only Lamiales diversified later, in the Oligocene (Table 15; Fig. 15).
Table 15

Estimated Times of Divergence of Relevant Angiosperm Groups Based on Fossil Estimates and Molecular Dating, Ages are Given in Millions of Years Before Present (MYBP)

Clade

Magallón et al., 1999

Wikström et al., 2001

Bremer et al., 2004

Crepet et al., 2004

This work

Estimate

Fossil

Molecular

Molecular

Fossil

Fossil

Angiosperms

158–179

113

Eudicots

131–147

100

Asterids

112–122

90

89.3

Cornales

69.5

106–114

128

89.3

Ericales

89.5

106–114

127

90

89.3

Euasterids

107–117

127

83.5

Campanulids

102–112

123

83.5

Aquifoliales

69.5

99–107

121

61.7

Apiales

69.5

85–90

113

40.4

Dipsacales

53.2

85–90

111

33.9

Asterales

29.3

101–94

112

37.2

Lamiids

102–112

123

83.5

Garryales

45.9

100–107

114

48.6

Gentianales

53.2

83–89

108

40.4

Solanales

53.2

82–86

106

33.9

Lamiales

37

71–74

106

28.4

https://static-content.springer.com/image/art%3A10.1007%2Fs12229-010-9040-1/MediaObjects/12229_2010_9040_Fig15_HTML.gif
Fig. 15

Minimum age dating of the Asteridae. Numbers on branches indicate age of the node they precede in millions of years. Cladogram based on Bremer et al. (2002), see Fig. 1 for explanation of color codes. See Figs. 214 for sources of minimum ages for each Order and the clade Boraginaceae-Vahliaceae

Discussion

Interest in establishing ages of origin and diversification of existing taxonomic groups has increased in recent years. Traditionally it has been up to the paleontologists to establish those dates, based on the first appearances of a taxon in the fossil record. Today, with the increase in use in one form or other of molecular clocks, the demand for reliable calibration points has increased accordingly. Now, it is demanded of paleontology that it delivers taxonomically and stratigraphically reliable fossil identifications that can withstand the test of phylogenetic methods (Benton & Donoghue, 2007; Donoghue & Benton, 2007). Phylogenetic methods have provided a means to more stringently test the placement of fossils by identifying synapomorphies that define those groups. Only fossils whose characters have been properly described and compared can be considered confidently identified. Thus only such fossils can be reasonably used in various methods of dating first appearances of taxa.

The survey and evaluation of the early fossil record of the Asteridae carried out in this work (Tables 314) attempts to provide a list of those fossil taxa that have been described as asterids and their degree of reliability. Those fossils that have been included in phylogenetic analyses offer the highest degree of confidence as their characters have been objectively tested against those of assumed related taxa. Unfortunately, very few fossils putatively representing asterids meet this criterion. The fusainized flowers from New Jersey (Nixon & Crepet, 1993; Gandolfo et al., 1998; Martínez-Millán et al., 2009) and Georgia (Keller et al., 1996), the mastixioids reviewed by Tiffney and Haggard (1996), and the fusainized flowers from Sweden analyzed in this work are among the few that have been put to, and passed the test of the phylogenetic analysis. One of the reasons for the paucity of reliably identified asterids is the lack of available matrices of morphological characters. Since most phylogenetic studies are based on genes, the morphological matrix in which a fossil could be included is rarely compiled.

Two alternative methods for the inclusion of fossil taxa have been proposed; the total evidence analysis and the molecular scaffold approach (Hermsen & Hendricks, 2008). In the total evidence approach, the molecular data and the morphology data are analyzed simultaneously and the fossil taxon is treated as no different as any other terminal in the analysis; the fossil is part of the process of formulating the phylogeny itself. This approach provides the most rigorous test of the relationships of the fossil to the rest of the taxa. The molecular scaffold approach involves finding the most suitable place for the fossil taxa given a pre-defined phylogeny of extant taxa. The fossil is not included in the original analysis that produced the phylogeny, but in a subsequent analysis whose objective is to find the best place for said fossil in that particular phylogeny. This was the approach used in this work to place the fossils Silvianthemum and Scandianthus. Of the two alternatives, however, the total evidence approach is certainly superior; it increases taxon sampling, increases the amount of information used to create the phylogeny, allows the fossil to impact the phylogenetic hypothesis and allows the discovery of secondary signals (see Hermsen & Hendricks, 2008 for a detailed discussion).

A less preferable but still acceptable alternative to the inclusion of the fossil in a phylogenetic analysis, is a description of the fossil with a thorough discussion of diagnostic characters including potential synapomorphies that relate the fossil to a particular clade. About two-thirds of the asterid fossils listed here (175 out of 261) include such a discussion, which allows the placement of the fossils in the most “suitable” position in the tree according to their characters. The rest of the fossils need to be reevaluated as their descriptions are not detailed and specific enough to be convincing.

It was by following these criteria that the minimum age dating of the Asteridae depicted in Fig. 15 has been obtained. According to these results, it seems that there have been three diversification “explosions” in the history of the Asteridae marked by the simultaneous first appearances of clades in the fossil record. The first one was in the Late Cretaceous when the four main groups of the Asteridae enter the fossil record; two in the Turonian (89.3 mya) the Cornales and the Ericales, and two in the Santonian (83.5 mya) the Lamiids and the Campanulids. The second one was in the Early Tertiary, around the Eocene (55–33.9 mya) involving most of the euasterid orders. And the third one taking place around the Oligocene when the last order, the Lamiales, diversified (Table 15; Fig. 15).

The fact that more than one fossil places a clade in a particular time frame increases confidence in the reliability of the minimum age of that clade. For example, the minimum age of Late Cretaceous for the Cornales is given by Tylerianthus crossmanensis from the Turonian of New Jersey (Table 3; Fig. 2), but if this fossil were to be removed, Hironoia fusiformis from Early Coniacian-Early Santonian of Japan would still place the Cornales in the Late Cretaceous (Table 3). Even more dramatic is the case of the Ericales as this clade has numerous fossils in the Late Cretaceous (Table 4; Fig. 3). The fact that the fossils come from different geographical locations adds another layer of confidence as the independence across data points (fossil identifications) can be more objectively assessed.

The diversification of the Euasteridae in the Late Cretaceous is, in principle, more difficult to support since there are only three fossils that place this huge clade in the Late Cretaceous: Scandianthus major, Scandianthus costatus and Silvianthemum suecicum (Table 12; Figs. 6, 12), and all of them come from the same locality. For this reason, assessing the phylogenetic relationships of these fossils is imperative. In the analyses performed in this work, both fossil taxa were ultimately placed as sisters to the same extant taxa that their authors had suggested based on direct observations: Silvianthemum with Quintinia and Scandianthus with the Vahliaceae (Friis, 1990; Friis & Skarby, 1982). However, these results should be taken with caution. For example, the matrix for the analysis of Scandianthus was derived from the same table that the authors built to support their conclusions, therefore, it is not surprising that Scandianthus was placed with the Vahliaceae. Independent confirmation of the placement of these fossils, or discovery of other euasterid fossils in the Late Cretaceous would certainly increase confidence in and robustness of these results.

In contrast to the minimum ages obtained from looking at the fossil record, the estimates based on molecular evidence suggest that the diversification of the Asteridae happened during the Early Cretaceous instead of the Late Cretaceous (Table 15). However, different molecular dating studies disagree with each other as much as they disagree with the fossil record (Table 15).

One of the most frequently cited molecular dating works is that of Wikström et al. (2001) who used non-parametric rate smoothing (NPRS; Sanderson, 1997) and a single calibration point—in the Rosid clade—to date the cladogram of Soltis et al. (2000). In their results, the Asteridae was estimated to have originated 112–122 mya and its diversification to have started some 106–114 mya (Table 15). Bremer et al. (2004), on the other hand, used the cladogram of Bremer et al. (2002) to explicitly estimate the time of origin and divergence of different groups of asterids based on molecular dating (Table 15). Three methods were applied: strict molecular clock of Langley and Fitch (1974), non-parametric rate smoothing (NPRS) of Sanderson (1997) and penalized likelihood (PL) of Sanderson (2002), although the authors only report the ages obtained with PL.

A comparison between these ages and the ones estimated by Wikström et al. (2001) indicates that Bremer et al. (2004) consistently got older age estimates than Wikström et al. (2001), between 11 and 30 my older (i.e. Campanulids and Lamiales, Table 15). This discrepancy could be due to a number of factors including different methodological tools used for estimating ages (NPRS vs. PL), different phylogenetic hypothesis used (Soltis et al., 2000 vs. Bremer et al., 2002) and different placement calibration points (one fixed calibration point in the Rosid clade vs. one fixed calibration point at the base of the Asterid clade). However, both molecular-based estimates agree in that they give significantly older estimates than those based on the fossil record alone (Table 15).

Advances and improvements in the methods to estimate molecular ages and phylogenies are constantly being produced (see Magallón, 2004; Pulquerio & Nichols, 2007; Soltis et al., 2007). This will undoubtedly improve our estimates of divergence events, and with it our understanding of evolutionary events in the history of clades. However, even the most precise of methods will deliver unreliable results if the data analyzed are not of good quality, including the fossils used as calibration points. One important step towards the improvement of quality of this calibration points is distinguishing those fossils that can be trusted in their taxonomic assignment from those that need to be restudied. This work provides that first step for the early fossil record of the Asteridae.

Acknowledgments

The author would like to thank the following colleagues for helping locate type specimens and for providing access to the collections in their institutions: Magali Volpes (AIX, Aix-en-Provence), Dr. Dario De Franceschi, Dr. Jean Dejax and Mme. Christiane Gallet (MNHN, Paris), Dr. Paul Kenrick, Dr. Peta Hayes and Mr. Cedric Shutte (NHM, London), and Dr. Dena Smith and Dr. Jonathan Marcot (CU Museum, Boulder CO). Special thanks to the librarians at the Hortorium, Mann, Kroch and Engineering Libraries, the Library Annex and Interlibrary Loan for providing prompt access to obscure materials that would have been otherwise impossible to obtain. The author would also like to thank Dr. William Crepet for comments and suggestions that greatly improved this paper, and Dr. Steven Manchester for this thorough review of the manuscript. This research was supported by grants from the following sources: Sigma Xi Cornell Chapter, the Mario Einaudi Center for International Studies, Grants-in-Aid of Research (Sigma Xi), The Paleontological Society, the Mid-America Paleontology Society (MAPS), the American Society of Plant Taxonomists (ASPT), the Botanical Society of America (BSA), the Harold E. Moore, Jr Memorial and Endowment Funds, the Department of Plant Biology, Cornell University and NSF grant DEB 01-08369.

Copyright information

© The New York Botanical Garden 2010