• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of annbotAboutAuthor GuidelinesEditorial BoardAnnals of Botany
Ann Bot. Jul 2006; 98(1): 123–140.
PMCID: PMC2803531

Former Diversity of Ephedra (Gnetales): Evidence from Early Cretaceous Seeds from Portugal and North America

Abstract

Background and Aims The extant species of the seed plant group Gnetales (Ephedra, Gnetum and Welwitschia) have been considered a remnant of a much greater, now extinct, diversity due to the pronounced differences in form and ecology among the genera. Until recently, this hypothesis has not been supported by evidence from the fossil record. This paper adds to the expanding information on Gnetales from the Early Cretaceous and describes coalified seeds from Barremian-Albian localities in Portugal and USA.

Methods The fossils were extracted from sediment samples by sieving in water. Adhering mineral matrix was removed by chemical treatment. Seeds were investigated using light and scanning electron microscopy. Morphology and anatomy of the seeds were documented and compared with those of extant species.

Key Results The fossils share characters with extant Ephedra, for example papillae on the inner surface of the seed envelope and in situ polyplicate pollen grains that shed the exine during germination. They differ from extant Ephedra seeds in morphological and anatomical details as well as in their smaller size. Two new species of Ephedra are described together with one species assigned to a new genus of Gnetales. Other Ephedra-like seeds, for which pollen and critical morphological details are currently unknown, are also present in the samples.

Conclusions These Cretaceous seeds document that key reproductive characters and pollen germination processes have remained unchanged within Ephedra for about 120 million years or more. There is sufficient variety in details of morphology to suggest that a diversity of Ephedra and Ephedra-like species were present in the Early Cretaceous flora. Their presence in Portugal and eastern North America indicates that they were widespread on the Laurasian continent. The fossil seeds are similar to seeds of Erdtmanithecales and this supports the previously suggested relationship between Erdtmanithecales and Gnetales.

Keywords: Early Cretaceous, Ephedra, Erdtmanispermum, Erdtmanithecales, Eucommiidites, fossil pollen, fossils, Gnetales, Portugal, Potomac Group, seeds

INTRODUCTION

The three extant genera Ephedra, Gnetum and Welwitschia comprise the Gnetales, a small group of seed plants with a poorly known evolutionary history. The three genera are very different in overall form and ecology and the monophyly of the group has sometimes been questioned (Eames, 1952; Nixon et al., 1994). However, several molecular studies (Goremykin et al., 1996; Bowe et al., 2000; Chaw et al., 2000; Magallón and Sanderson, 2002; Rydin and Källersjö, 2002; Rydin et al., 2002) support monophyly of the Gnetales, as well as the position of Ephedra as sister group to a Gnetum-Welwitschia clade. Gnetalean monophyly and this pattern of relationships was hypothesized originally based on morphology (Crane, 1985; Doyle and Donoghue, 1986, 1992; Doyle, 1996).

Arber and Parkin (1908) interpreted the marked morphological differences among the three extant genera as an indication of a long evolutionary history and, even though no unambiguous fossil representatives of the group were known at that time, they suggested that extant Gnetales constitute merely a remnant of ‘former greatness’ (Arber and Parkin, 1908). Later studies of dispersed pollen lent support to this suggestion by documenting the extensive fossil record of polyplicate pollen grains often referred to as ‘ephedroids’ that in general morphology are very similar to the pollen of extant Ephedra and Welwitschia (Wilson, 1962; Crane and Lidgard, 1989; Osborn et al., 1993; Crane, 1996; Osborn, 2000). Discoveries of well-preserved and informative mega- and mesofossils from the Early Cretaceous have provided further undisputed support for a long and rich fossil history of the group. Cratonia cotyledon (Rydin et al., 2003) and Welwitschiostrobus murili (Dilcher et al., 2005) from Brazil, and possibly also Drewria potomacensis (Crane and Upchurch, 1987) from Virginia, USA, are Early Cretaceous members of the Gnetum-Welwitschia clade. Otherwise many of the Early Cretaceous gnetalean fossils in the literature appear more closely related to extant Ephedra (Wu et al., 1986; Cao et al., 1998; Krassilov et al., 1998; Guo and Wu, 2000; Wu et al., 2000; Sun et al., 2001; Tao and Yang, 2003; Rydin et al., 2004; Yang et al., 2005). There are also Mesozoic fossils, with affinity to the Gnetales, which are difficult to assign to any of the extant lineages (Krassilov, 1982, 1986; Krassilov and Bugdaeva, 1988; Duan, 1998; Sun et al., 2001).

Extant Ephedra comprises approx. 35–40 species (Kubitzki, 1990) inhabiting arid environments of the northern hemisphere and South America. They are shrubs with opposite to whorled branching and reduced leaves that are decussate or in whorls. Female cones have fleshy or dry bracts and seeds with an integument surrounded by an outer envelope. The apical part of the integument forms a micropylar tube that extends beyond the outer seed envelope and serves as the pollen-receiving area. Male cones consist of microsporangiophores with fused microsporangia, surrounded by paired bracts. The pollen grains are distinctive with aligned ridges and valleys (polyplicate) that form alternating thicker and thinner regions of the exine (Steeves and Barghoorn, 1959; El-Ghazaly and Rowley, 1997; Osborn, 2000). The exine is discarded during germination (El-Ghazaly et al., 1998).

All extant species of Ephedra are very similar in gross morphology and have few informative characters in gene regions investigated so far [18S, 26S, ITS, rbcL, trnL, rps4 (Rydin et al., 2004); rps4, ITS1 (Ickert-Bond and Wojciechowski, 2004); rbcL, matK, ITS1 (Huang et al., 2005)]. Phylogenetic analyses indicate that major groups of species within Ephedra correspond to geographic regions, but relationships among these groups are not well supported (Ickert-Bond and Wojciechowski, 2004; Rydin et al., 2004; Huang et al., 2005). Within the genus, the absence of obvious morphological synapomorphies for extant subgroups, and the apparently homoplasious nature of many characters, have hampered traditional classification as well as phylogenetic reconstruction of relationships and molecular dating analyses. Groups defined on morphological characters (Stapf, 1889; Mussayev, 1978; Freitag and Maier-Stolte, 1994) are not fully consistent with the phylogenetic history of the genus indicated by molecular data. Huang and Price (2003) suggested that extant species of Ephedra evolved recently (Oligocene to Miocene) based on the paucity of base substitutions in the rbcL gene and an assumption of a molecular clock calibrated by landmark events. The current lack of morphologically well-defined subgroups impedes a placement of the fossils in any of the subgroups and thus also makes it difficult to date subgroups by fossil calibration.

In this paper a set of fossils are described from the Early Cretaceous Buarcos and Torres Vedras localities, Portugal and from Drewry's Bluff, Virginia, USA. The material consists of coalified Ephedra-like seeds, which based on differences in detailed morphology are assigned to several new species. The fossil seeds add to knowledge of the evolutionary history of the Gnetales and provide clear evidence for the early occurrence of several unique characters of extant Ephedra.

MATERIALS AND METHODS

The fossils were collected from Early Cretaceous localities in Portugal and USA. Samples from Portugal were collected from fluviatile sands and clays at two localities: Buarcos and Torres Vedras, in the western Portuguese (Lusitanian) Basin. Both localities have now been obliterated by town development.

The Buarcos locality was a partly overgrown road cut in the town of Buarcos, situated north of Figueira da Foz (Beira Litoral region), 40 °09′54″N; 8 °52′11″W, Carta Geológica de Portugal 19C Figuera da Foz (Rocha et al., 1981). Samples were collected by E. M. Friis and K. R. Pedersen in 1992–1997. A small palynoflora was reported from the Buarcos locality by Pais and Reyre (1981). Subsequently, rich mesofossil assemblages have been extracted from the sediment sequence including many angiosperms (Friis et al., 1997, 1999, 2000a, b). The plant-bearing sequence from Buarcos is included in the lowermost member (Calvaria Member) of the Figueira da Foz Formation, thought to be of late Aptian to early Albian age (Dinis, 2001) or early Albian (Heimhofer et al., 2005).

The Torres Vedras locality was an open-cast clay pit on the north-eastern outskirts of Torres Vedras (Estremadura region), about 1 km north-east of Forte de Forca on the road to Sarge, 39 °06′13″N; 9 °14′47″W, Carta Geologica de Portugal Torres Vedras 30C, (Zbyszewski et al., 1955). The sample studied here (Torres Vedras 43) was collected in 1989 by E. M. Friis and K. R. Pedersen, in a sandy, lignitic horizon referred to the basal member of the Almargem Formation, at the same level as sample 44 described in Friis et al. (2004). According to Rey (1993), this part of the sequence is of late Barremian to early Aptian age. A late Barremian to early Aptian age of the Torres Vedras samples is also supported by mesofossil floras and in situ pollen grains (E. M. Friis, K. R. Pedersen and P. R. Crane, unpubl. res.).

Samples from the eastern USA were collected by P. R. Crane, A. N. Drinnan, E. M. Friis and K. R. Pedersen in 1988, from the Potomac Group sequence at the Drewry's Bluff locality along the James River, south-east of Richmond, Chesterfield County, VA (37 °25′30″N; 77 °25′15″W). All specimens examined here are from reworked clayballs sampled in the lower part of the exposed fluvial sequence, about 1–2 m above river level (Drewry's Bluff samples 163, 164 and 198). The Early Cretaceous sequence at the Drewry's Bluff locality belongs to the Patuxent Formation and the sediments are referred by Doyle and Hickey (1976) to the upper pollen Zone I of Brenner (1963), which according to Doyle (1992) may be of middle Aptian age. The clayballs are older than the sediments in which they occur and are referred by Brenner (1963) to the lower part of pollen Zone I. They were considered by Brenner (1963) to be of Barremian age, while Doyle (1992) indicated an early Aptian age for the lower part of pollen Zone I in the Potomac Group sequence.

The fossils were extracted from the sediment samples by sieving in water. Adhering mineral matrix was removed by treatment with hydrofluoric acid (40 %) and hydrochloric acid (10 %), followed by rinsing in water, following standard methods (Friis et al., 1988). The fossils were investigated initially using a stereomicroscope. For scanning electron microscopy (SEM), specimens were mounted on cleaned aluminium stubs, coated with gold for 90 s in a sputter coater, and examined with a Hitachi field emission-scanning electron microscope at 2 kV. Selected seeds were macerated for light microscope studies, using nitric acid followed by ammonia or by using sodium hypochlorate. Specimens and preparations are housed in the Swedish Museum of Natural History, Stockholm (S) and in the Field Museum, Chicago (PP).

RESULTS

Formal descriptions

Spermatophyta

Order Gnetales

Genus Ephedra L.

Ephedra portugallica sp. nov. (Figs 1A–K and 2A–F).

Specific diagnosis

Fossil seeds and in situ pollen with their general organization and structure as in modern Ephedra (nucellus surrounded by integument and outer sclerenchymatous envelope, long micropylar tube extending beyond the envelope, papillae on inner apical part of envelope, pollen exine shed during germination), and with the following defining features:

Seeds small, 0·85–1·2 mm long, 0·7–0·8 mm wide, ovoid with rounded base and pointed apex. Attachment scar prominent, rounded to slightly squared, approx. 0·25 mm in diameter. Sclerenchymatic cells of outer seed envelope narrowly rectangular, approx. 10–30 µm long, longitudinally arranged. Pollen approx. 33–40 µm long, 20–25 µm wide, approx. 10–15 ridges on exposed side (inferred total of 20–30 ridges).

Etymology

Named from Portugal, where the fossils were found.

Holotype

S107680 (Buarcos sample 157, Figs 1A, 2A–C).

Fig. 1.
Ephedra portugallica sp. nov., Early Cretaceous seeds from Buarcos, Portugal. (A) Holotype with in-situ polyplicate pollen grains indicated by an arrow (S107680). (B–F) Paratypes showing the ovoid form, the rounded base, the pointed apex and the ...
Fig. 2.
Ephedra portugallica sp. nov. (A) apical view showing the four-angled seed envelope, the circular integument and in-situ polyplicate pollen (S107680); (B) in-situ pollen grain indicated by the arrow in Fig. 1A (S107680); (C) rugulate surface of pollen ...

Paratypes

S107678–S107679, S107681, S136808 (Buarcos sample 157); S107682, S107684–S107686, S148070–S148072, 148102 (Buarcos sample 209); S107687–S107689, S148077–S148078 (Buarcos sample 211); S107683, S134042, S148073, S148087–S148088, S148090, S148092, S148094–S148096 (Buarcos sample 243); S148060 (Buarcos sample 267).

Locality

Buarcos, North of Figueira da Foz, Portugal (40 °09′54″N; 8 °52′11″W).

Stratigraphic position and age

Calvaria Member, Figueira da Foz Formation. Early Cretaceous (late Aptian or early Albian).

Description and comments on Ephedra portugallica

This seed type is the most abundant and informative of the fossil ephedroid seeds from the Buarcos locality and is known from more than 50 specimens.

Shape and size of seeds

The seeds are small, about 0·85–1·2 mm long and 0·7–0·8 mm wide. They are often compressed to various degrees but have a basically ovoid shape with a rounded base and pointed apex (Fig. 1A–G). In cross-section they are almost circular, to faintly four-angled but they are more distinctly four-angled towards the apex (Fig. 2A). In some specimens the four-angled form is emphasized by distinct longitudinal ridges that probably mark the position of vascular bundles along the four edges (Figs 1A, B and E). The seeds sometimes split apically into four valves along the ridges (Fig. 1G). The basal attachment scar is distinct and circular to four-angled, approx. 0·25 mm in diameter (Fig. 1B, C and F).

Organization and structure of seeds

The seeds are composed of three distinct parts: an innermost membranous nucellus preserved as a delicate cuticle, a membranous integument and a thick sclerenchymatic outer seed envelope. The nucellus has been observed only in macerated specimens. It is ovate to elliptic in longitudinal outline with an irregular apical area. The integument is ovate to elliptic in longitudinal outline and extended at the apex into a long, narrow micropylar tube (Fig. 2E) that is circular in cross-section (Fig. 2A). The outer seed envelope consists of an outer epidermis, a middle sclerenchyma layer, and an inner epidermis. In one specimen the outer epidermis is well preserved (Fig. 1B), but otherwise this layer is usually abraded over most of the seed surface except for small areas in the apical region. The shape and size of the epidermal cells are unclear. The sclerenchyma layer constitutes the bulk of the envelope and the sclerenchyma cells are generally well preserved. They are narrowly rectangular, about 10–30 µm long, and longitudinally arranged (Fig. 1H and I).

The inner epidermis of the seed envelope has, in the apical region adjacent to the micropylar tube, a distinct and well-preserved zone of papillae extending from the top of the seed envelope and about 0·17 mm downwards. The papillae are observed in specimens where the envelope is split apically (Fig. 1G, J and K), and sometimes in well-preserved intact specimens where the micropylar area is not filled in with debris or collapsed cells. Each papilla is about 10 µm long, but they decrease in size and prominence towards the base of the papillate zone.

In-situ pollen grains

Pollen grains were found in situ in the micropylar tube of three specimens. In two specimens (Fig. 1A and D) pollen grains were observed at the top of the micropyle of seeds mounted for SEM (Fig. 2A–D). The pollen grains are ellipsoid, approx. 33–40 µm long, and approx. 20 µm wide, with 10–15 ridges on the exposed surface; the inferred number of ridges for the whole grain is thus 20–30. There is no visible aperture. The surface of the pollen wall is indistinctly rugulate (Fig. 2C). In the macerated specimen (Fig. 2E and F) pollen grains were observed inside the integument in the micropylar tube near the apex of the nucellus. These pollen grains show upcurled pollen exines (Fig. 2F).

Ephedra drewriensis sp. nov. (Figs 3A–K and 4A–D)

Fig. 3.
Ephedra drewriensis sp. nov., Early Cretaceous seeds from Drewry's Bluff, Virginia, USA. (A) Holotype, showing the ellipsoid form of the seeds and papillae on the seed envelope (PP44843). (B–F) Paratypes showing the faintly four-angled seed envelope ...
Fig. 4.
Macerated seed of Ephedra drewriensis sp. nov. (PP44875): (A) seed envelope and integument; (B) seed envelope, integument and nucellus, in-situ pollen in the upper part; (C) close up of polyplicate pollen grains with a non-germinated grain in the middle ...

Specific diagnosis

Fossil seeds and in situ pollen with their general organization and structure as in modern Ephedra (nucellus surrounded by integument and outer sclerenchymatous envelope, long micropylar tube extending beyond the envelope, papillae on inner apical part of envelope, pollen exine shed during germination), and with the following defining features:

Seeds small, 1·2 and 1·4 mm long, 0·8–0·9 mm wide, ellipsoid with rounded base and short pointed apex. Attachment scar indistinct, rounded, approx. 0·15 mm in diameter. Sclerenchymatic cells of outer seed envelope narrowly rectangular, up to 70 µm long, longitudinally arranged. Pollen approx. 50 µm long, 20 µm wide, narrowly ribbed, six or seven ridges on exposed side (inferred total of 12–14 ridges).

Etymology

Named from Drewry's Bluff, where the fossils were found.

Holotype

PP44843 (Drewry's Bluff sample 198), (Fig. 3A and J).

Paratypes

PP44844-PP44845, PP44878 (Drewry's Bluff sample 198); PP44854, PP44866–PP44867 (Drewry's Bluff sample 163); PP44874–PP44875 (Drewry's Bluff sample 164).

Locality

Drewry's Bluff on the James River, south of Richmond, Chesterfield County, VA. USA.

Stratigraphic position and age

Clayballs in Patuxent Formation, Potomac Group. Pollen Zone I, Early Cretaceous (early Aptian).

Description and comments on Ephedra drewriensis

About 10–15 seeds of this type were found in the samples from Drewry's Bluff.

Shape and size of seeds

The seeds are about 1·2–1·4 mm long and 0·8–0·9 mm wide. They are compressed to various degrees but have a basically ellipsoid shape with a rounded base and pointed apex. They are often flattened and faintly four-angled in cross-section, but they are more distinctly four-angled towards the apex (Fig. 3A–G). Weak longitudinal ridges correspond to the position of vascular bundles (Fig. 3B, C, E and H). The seed envelope sometimes split apically into four valves (Fig. 3G). The basal attachment scar is indistinct, circular, and approx. 0·15 mm in diameter.

Organization and structure of seeds

The seeds are composed of three distinct tissues: an innermost nucellus surrounded by a membranous integument, and a thick sclerenchymatic outer seed envelope. The nucellus has been observed only in macerated specimens. It is ovate to elliptic in shape and with an irregular apical area. The integument is elliptic in shape and apically extended into a micropylar tube that is circular in cross-section (Fig. 4A, B and D). The outer seed envelope consists of an outer epidermis, a middle sclerenchyma layer and an inner epidermis. The outer epidermis is partly preserved in some specimens (Fig. 3B and D), and consists of more or less isodiametric cells approx. 9–13 µm across. The sclerenchyma cells are generally well preserved, narrowly rectangular up to 70 µm long, and longitudinally arranged (Fig. 3H and I).

Papillae occur on the inner surface of the outer seed envelope in an area extending from the apex to approx. 0·17 mm downwards. The papillae are 10 µm wide and 10–20 µm long, but the lowermost papillae are less well developed. Papillae can be observed in split open seed envelopes (Fig. 3G and K), in some intact seeds (Fig. 3A, F and J) and in macerated specimens (Fig. 4D).

In-situ pollen grains

Specimen PP44875 (Figs 3C and 4A–D) was macerated and at least five polyplicate pollen grains were found in situ inside the integument (Fig. 4C). They are narrowly elliptic, nearly 50 µm long and 20 µm wide, ribbed with six or seven ridges on one side. The total number of ridges inferred for pollen grains of Ephedra drewriensis is 12–14. There is no visible aperture in the pollen grains. Some grains are intact, but there are also several upcurled pollen exines (Fig. 4C).

Ephedrispermum gen. nov. (Fig. 5A–N)

Fig. 5.
Ephedrispermum lusitanicum gen. et sp. nov., Early Cretaceous seeds from Buarcos and Torres Vedras, Portugal. (A) Holotype: in-situ pollen grains in the micropyle (S148062). (B–K) Paratypes showing the rounded shape of the seed envelope, with ...

Generic diagnosis

Fossil seeds with their general organization and structure similar to modern Ephedra. Seeds ovoid to almost spherical with a short pointed apex. Nucellus and integument surrounded by an outer sclerenchymatous seed envelope, seed envelope distinctly four-angled towards apex. Sclerenchyma cells rectangular and in two layers, outer layer cells transversely arranged, inner layer longitudinally arranged. Papillae on seed envelope not observed. Pollen grains narrowly ribbed.

Etymology

These Ephedra-like seeds have been named from Ephedra and spermum meaning seed.

Type species

Ephedrispermum lusitanicum gen. et sp. nov.

Comments on the new genus Ephedrispermum

The seeds are similar in organization and structure to seeds of Ephedra, but discarded pollen exines and papillae on the seed envelope, characters that are diagnostic of extant Ephedra, have not been detected. Seeds of Ephedrispermum further differ from Ephedra in the cross-wise pattern of sclerenchyma cells on the seed envelope. As far as is known, this pattern is not present in Ephedra and, therefore, a new genus within the Gnetales has been established for these seeds.

Ephedrispermum lusitanicum gen. et sp. nov. (Fig. 5A–N)

Specific diagnosis

As for the genus with the following additions. Seeds minute, 0·6–0·9 mm long, 0·6–0·8 mm wide, ovoid to spherical with short pointed apex. Attachment scar prominent, rounded, approx. 0·1 mm in diameter. Outer layer of sclerenchyma cells rectangular, up to 30 µm long, transversely arranged. Inner layer cells up to 30 µm long, longitudinally arranged. Pollen grains approx. 23 µm long, 10 µm wide, approx. six ridges on the exposed side (inferred total of up to 12 ridges).

Etymology

From Lusitania, name of Roman province covering central Portugal where the fossils were collected.

Holotype

S148062 (Buarcos sample 244; Fig. 5A and L).

Paratypes

S148062–S148064 (Buarcos sample 244); S148079–S148083 (Buarcos sample 157); S148097, S148099 (Buarcos sample 209); S148076 (Buarcos sample 211); S148074, S148085–S148086, S148091 (Buarcos sample 243); and S148106–S148109, S148111–S148112, S148114–S148120, S148125, S149286–S149289 (Torres Vedras sample 43).

Locality

Buarcos, North of Figueira da Foz, Portugal (40 °09′54″N; 8 °52′11″W).

Stratigraphic position

Calvaria Member, Figueira da Foz Formation. Early Cretaceous (late Aptian or early Albian).

Description and comments on Ephedrispermum lusitanicum

About 35 seeds of this type have been found in samples from Torres Vedras and Buarcos. They are typically less well preserved than seeds of Ephedra portugallica and Ephedra drewriensis. This is particularly true for seeds from the Buarcos locality.

Shape and size of seeds

Seeds are minute, 0·6–0·9 mm long and 0·6–0·8 mm wide, broadly ovoid to almost spherical in shape (Fig. 5A–K) and faintly four-angled in cross-section except at the apex where they are distinctly four-angled (Fig. 5A–K). The seeds are rounded at the base and have a circular attachment scar, about 0·1 mm in diameter, and a short, pointed apex (Fig. 5E and F).

Organization and structure of seeds

The seeds are apparently identical to other ephedroid seeds from Buarcos in organization and structure, although they are smaller, nearly spherical and not as well preserved. Maceration of seeds failed to provide information on the nucellus, the shape of micropylar tube, and the inner epidermis of the outer seed envelope. The outer epidermis of the outer seed envelope is poorly preserved in most specimens. The sclerenchymatic part of the seed envelope consists of at least two cell layers. Cells in the outer layer are rectangular, up to 30 µm long and arranged in transverse rows (Fig. 5 M and N). Cells of the inner layer are more narrowly rectangular, up to 30 µm long and longitudinally aligned. The micropylar region is generally poorly preserved and papillae have not been observed.

In-situ pollen grains

Two specimens (Fig. 5A and B) have pollen grains in the micropyle. The pollen grains are polyplicate and small, only approx. 23 µm long and 10 µm wide. The grains in the holotype (Fig. 5A and L) have up to six exposed ridges (the total number of ridges is probably less than 12). The poorly preserved pollen grain in the micropyle of the paratype S148064 (Fig. 5B, grain not shown in close up) is only 15 µm long and has two or three ridges. This is likely to be a foreign grain.

DISCUSSION

Systematic and nomenclatural comments

The fossil seeds described here share many characters with each other and with seeds of extant Ephedra. These include a sclerenchymatous outer seed envelope that surrounds the nucellus and integument, a long micropylar tube that extends beyond the outer envelope, and in situ polyplicate pollen grains. Based on differences in pollen features, in the shape and size of the seeds and in the cellular details of the seed wall (Table 1), they have been described as three new species. Two of the species have papillae on the inner, apical part of the seed envelope and pollen that shed the exine during germination. This character combination is only known for seeds of Ephedra and they have been assigned to two fossil species of the extant genus: Ephedra portugallica and Ephedra drewriensis.

Table 1.
Morphological variation in seed and pollen characters of Erdtmanithecales, Ephedrispermum and Ephedra

The third species also has general characters of Ephedra including a sclerenchymatic seed envelope and polyplicate pollen in situ in the micropyle. However, two diagnostic characters of Ephedra (discarded pollen exines and papillae on the seed envelope) have not been observed in any specimen of this species. Further, these seeds differ from Ephedra in the cross-wise pattern of sclerenchyma cells on the seed envelope, a pattern that as far as is known is not present in Ephedra and, therefore, a new genus has been established, Ephedrispermum, to accommodate these seeds.

Interpretations of the new fossils and comparison with extant species

Seed characters

Seeds of the three new fossil species are all small, ranging in length between 0·6 and 1·6 mm, and all possess a slightly four-angled sclerenchymatic outer seed envelope with a rounded base and a pointed apex (Figs 1–5). The shape of the fossil seeds varies from ovoid in Ephedra portugallica, ellipsoid in Ephedra drewriensis to more globose in Ephedrispermum lusitanicum. Many extant species have ovoid seeds, but more ellipsoid and globose seeds also occur. The apex can be distinctly or less prominently pointed, in the fossils as well as in extant seeds. In the fossils, four longitudinal ridges extend from the apex and the seed envelope is sometimes split from the apex into four valves. Xylem cells with a spiral to helical wall pattern have been observed on the longitudinal ridge of one fossil (Figs 3C and H) and the ridges probably correspond to the position of vascular bundles. The four-angled nature is evident in all the fossils, whereas the shape of extant seeds varies from two to three to four angles (four-angled seed of Ephedra californica shown in Fig. 6A). The shape (the number of angles) of extant seeds is partly related to the number of seeds in the cone, a character that varies among and within species.

Fig. 6.
Seeds and pollen of extant Ephedra: (A) Ephedra californica, seed (S-C598); (B) Ephedra alata, sclerenchymatic cells of the seed envelope (S-C311); (C) Ephedra likiangensis, micropylar tube supported by papillae on the seed envelope (S-03926); (D) Ephedra ...

The outer seed envelope in Ephedra seeds consists of an outer epidermis, a middle sclerenchyma layer, and an inner epidermis. In the fossils, the outer epidermis is typically abraded over most of the seed surface except for small areas in the apical region, but the sclerenchyma cells that comprise the bulk of the envelope are generally well preserved in all three species. In Ephedra drewriensis the sclerenchyma cells are very long and narrow, up to 70 µm long, while those of the Portuguese species Ephedra portugallica and Ephedrispermum lusitanicum are 30 µm or shorter. In Ephedra portugallica and Ephedra drewriensis, these cells are longitudinally arranged. The seeds of Ephedrispermum lusitanicum are often preserved slightly differently and the shape and size of the sclerenchyma cells are not always clear. Species with well-preserved cell patterns reveal an outer layer of transversely arranged sclerenchymatic cells covering an inner layer of longitudinally arranged cells, which together give the seeds a chequered appearance (Fig. 5G–J and M–N), not known from seeds of extant Ephedra. This pattern is less clear in some specimens (Fig. 5A–C and E), but is still obvious on parts of these seeds upon careful examination (Fig. 5 M).

The shape and arrangement of sclerenchymatic cells of the seed envelope were investigated for nine extant species from all major subgroups of Ephedra (see Table 1). Removal of the epidermis revealed a layer of longitudinally arranged sclerenchymatic cells that range in length from approx. 50 µm (Fig. 6B) to 200 µm or more. These cells gradually decrease in length towards the apex. From this limited sample it seems as if Old World species on average have shorter cells than New World species. The same pattern is observed in the fossils.

Maceration of the seeds to isolate acid resistant cuticles and pollen walls, shows both the cuticles of the outer seed envelope and the thin cuticles of the integument and nucellus (Figs 2E and 4A and B). The main body of the integument is ovate to elliptic in lateral view and tightly adheres to the nucellus. Apically it is extended into a long micropylar tube that protrudes beyond the outer seed envelope (Figs 2E and 4A and D). The protruding part of the tube breaks easily, which is obvious in the fossils as well as in extant seeds. Inside the envelope, the micropylar tube is supported by papillae (Fig. 6C and D) located on the inside of the upper part of the outer seed envelope (Thoday and Berridge, 1912; Rydin et al., 2004). Occurrence of papillae on the outer seed envelope was investigated for 25 extant Ephedra species representing major subgroups (Rydin et al., 2004); all possessed this character.

The fossils Ephedra drewriensis and Ephedra portugallica also have papillae, which may be observed in specimens where the envelope is split apically (Figs 1G, J and K and 3G and K), in some well-preserved intact specimens (e.g. Fig. 3J), and also in macerated specimens (Fig. 4D). In extant and fossil Ephedra, the papillae adhere to the micropylar tube (Fig. 6C) and are gradually transformed downwards into ordinary cells of the inner epidermis of the envelope (Fig. 3G and K). The papillae of the fossil seeds are identical in shape and position to those of extant Ephedra. Papillae were not found in Ephedrispermum lusitanicum, but no seed of this type had split open prior to preservation, and papillae can be difficult to see in intact seeds.

Pollen characters

Polyplicate pollen grains (Fig. 6E and F) occur in all extant species of Ephedra, and have also been found in situ inside the fossil seeds. The surface of the pollen wall is indistinctly rugulate in the holotype of Ephedra portugallica (Fig. 2C), but this feature is not seen in other specimens and it is unclear if this is a result of the fossilization process. The pollen grains are narrowly ribbed and about 33–40 µm long in Ephedra portugallica, about 50 µm long in Ephedra drewriensis, and only about 23 µm long in Ephedrispermum lusitanicum. Pollen of extant Ephedra ranges between 30 and 70 µm in length, 40–50 µm being common in many species. The pollen grains of Ephedrispermum lusitanicum are thus smaller than what is usually seen in extant taxa.

The number of ridges on pollen grains of extant Ephedra differs among species. According to Kubitzki (1990), the range is from 6 to 18, but Steeves and Barghoorn (1959) report up to 20 ridges for Ephedra alata. Intraspecific variation has been reported for grains of extant Ephedra (Steeves and Barghoorn, 1959; El-Ghazaly and Rowley, 1997; Ickert-Bond et al., 2003). In the fossils, pollen grains are always preserved inside the micropyles and are often folded or overlapping. It is not possible to get a three-dimensional view of these grains that exposes all sides, and it is therefore difficult to assess the exact number of ridges. Therefore, the number of visible ridges (generally the number of ridges on one side) and the total number of ridges (inferred to be about twice as many) are reported. Grains of Ephedra portugallica have 10–15 exposed ridges (inferred total of 20–30) (Figs 2A, B and D). Pollen grains of Ephedra drewriensis have fewer ridges, only six or seven on the exposed side (inferred total of 12–14) (Fig. 4C). The numbers of ridges on the fossil pollen grains of Ephedra drewriensis are well within the range of variation seen in modern species, whereas the inferred total of 20–30 ridges in Ephedra portugallica (Fig. 2A, B and D) is higher than what is generally seen in extant taxa. In Ephedrispermum lusitanicum, one of the grains in the holotype has six visible ridges (total number probably <12) (Fig. 5L). The other grain conceivably has fewer ridges, but it is partly covered and the number of ridges is difficult to assess. A single pollen grain observed in paratype S148064 (Fig. 5B, not shown in close up), is only 15 µm and seems to have two or three ridges. This grain is covered in a gluey substance and is considered to be foreign. Thus, the variation in size and number of ridges seen in grains of Ephedrispermum lusitanicum is probably an artefact owing to one partly covered and one foreign grain.

The exine of extant Ephedra pollen grains splits up during germination and is shed, leaving the male gametophyte naked in a characteristic way (El-Ghazaly et al., 1998). Pollen grains were found inside the micropyles at the top of the megasporangium in specimens of Ephedra drewriensis and in Ephedra portugallica. In both species these included grains with curled up, shed exines (Figs 2E and F and 4B and C), which completely match the shed, upcurled, transversely striated exines described for extant Ephedra (El-Ghazaly et al., 1998). Pollen grains of Welwitschia are similar to those of Ephedra in size and shape but Welwitschia grains possess a sulcus through which germination occurs and the exine remains on the gametophyte (Rydin and Friis, 2005). The unusual habit of shedding the pollen exine prior to male gametophyte growth and pollen tube formation is evidently unique to Ephedra among the polyplicate pollen-producing genera of the Gnetales (Rydin and Friis, 2005).

Comparison with other fossil seeds and pollen

Fossil Gnetales

The fossils described here are currently the only coalified and three-dimensionally preserved ephedroid seeds reported from the fossil record. In addition to the three newly described species, a diversity of other Ephedra-like seeds has been found in samples of Early Cretaceous age from Portugal and the Potomac Group sequences (Fig. 7A–L). However, these seeds differ from those described here in key details and lack diagnostic features such as in situ pollen grains and papillae on the seed envelope. This material requires further study to establish their exact systematic position, but they provide additional evidence that a diversity of ephedroid species were present in the Early Cretaceous vegetation.

Fig. 7.
Additional diversity of Early Cretaceous putatively ephedroid seeds. (A–D) Ovoid and prominently four-angled seeds, with folded seed envelope from Drewry's Bluff (A, PP44869; B; PP44848; C, PP44868; D, PP44870). (E) Obovoid seed with transverse ...

The only other Ephedra fossil in which details of seeds are preserved is Ephedra archaeorhytidosperma, from the Early Cretaceous Yixian Formation of north-eastern China (Yang et al., 2005). It is a compression fossil of a fertile shoot that has strong, vegetative similarities to Ephedra. Female cones are terminal on the branches. Seeds are partly enclosed by remnants of bracts, and there are weak impressions of an extended micropylar tube. The seeds of Ephedra archaeorhytidosperma resemble the fossils presented here in their size, ovoid shape with a pointed apex and the presence of an extended micropyle, but are clearly different in having a prominent, longitudinal, dorsal ridge and a seed surface characterized by wavy transverse ridges.

A variety of dispersed polyplicate pollen types have previously been recorded from the Drewry's Bluff locality (Brenner, 1963). Many are clearly ephedroid but others are more similar to Welwitschia, e.g. those found in association with Drewria potomacensis (Crane and Upchurch, 1987). The pollen grains associated with Drewria are clearly distinct from those found in situ in Ephedra drewriensis. Grains of Ephedra drewriensis appear inaperturate as in pollen of extant Ephedra, and they shed the pollen exine during germination.

In addition, there are dispersed and in situ grains of Eucommiidites (Erdtmanithecales) at the same locality (further discussed below).

Fossil Erdtmanithecales

Among seeds of other gymnosperms only those of the extinct Erdtmanithecales (Fig. 8A–D) are comparable in gross morphology to the fossil Ephedra seeds described here. Erdtmanithecales is an order of seed plants established by Friis and Pedersen (1996) to encompass dispersed pollen organs and seeds with Eucommiidites pollen in situ in the pollen sacs and inside the micropyles of the seeds.

Fig. 8.
Erdtmanithecales from the earliest Cretaceous of Bornholm, Denmark (Erdtmanispermum seeds and Eucommiidites pollen). (A, B) Seeds showing the three-angled nature of the seed envelope and the pointed apex (A, S148043; B, S148050). (C) Isodiametric sclerenchyma ...

The pollen genus Eucommiidites (Fig. 8D) was established by Erdtman (1948) for pollen grains from the Early Jurassic of Scania, Sweden, characterized by three elongated slits. Erdtman described the pollen as tricolpate and compared it with pollen of the angiosperm genus Eucommia. It was later shown that the colpi are unevenly distributed and the grains were re-interpreted as non-angiospermous pollen grains with one distal colpus and two subsidiary, lateral colpi (Couper, 1956, 1958). A non-angiospermous affinity of the grains was later demonstrated by the discovery of pollen grains in situ in the micropyle of small seeds (Hughes, 1961; Brenner, 1963, 1967) (Fig. 8D).

Dispersed Eucommiidites pollen grains are widespread in Mesozoic palynofloras from the northern hemisphere, ranging in age from the Early Jurassic to the early Late Cretaceous, but all pollen organs and seeds discovered so far are from the Cretaceous. Pollen organs that can be assigned to the order include Erdtmanitheca (Pedersen et al., 1989), Eucommiitheca (Friis and Pedersen, 1996) and Bayeritheca (Kvacek and Pacltová, 2001). Seeds of Erdtmanithecales include fossils assigned to Erdtmanispermum (Pedersen et al., 1989) as well as two unnamed fossils (Hughes, 1961; Brenner, 1963, 1967). Erdtmanitheca is from the early Late Cretaceous (early Cenomanian) of North America (Texas), while all the other taxa are from the Early Cretaceous.

Seeds of Erdtmanispermum balticum (Erdtmanithecales) are coalified and comparable in preservation to the fossil Ephedra seeds described here. They were described from the earliest Cretaceous (Berriasian-Valanginian) sediments of Bornholm, Denmark (Pedersen et al., 1989) (see also Fig. 8A–D). They are ovoid in outline with a rounded base and a long, distinctly pointed micropyle (Fig. 8A and B). The seed wall is composed of a membranous integument and an outer sclerenchymatic seed envelope. The sclerenchyma cells are small and isodiametric (approx. 10 µm; Fig. 8C). The envelope is three-parted and gives the seed a weakly triangular cross-section. The envelope sometimes splits from the apex into three valves. In apical view the micropylar tube is circular and surrounded by the triangular envelope. Pollen grains of Eucommiidites-type were found inside the seeds by maceration (Pedersen et al., 1989) and were also observed by SEM in the top of the micropylar tube (Fig. 8D). The two unnamed seed types with Eucommiidites pollen in situ are preserved only by their seed cuticles and the nature of the seed wall is unknown, but the specimen from Drewry's Bluff figured by Brenner (1963) clearly shows a triangular form.

Seeds of Erdtmanispermum balticum strongly resemble the seeds described here, in size, in their ovoid shape with a prominently pointed apical region, in the angled envelope, probably derived from bracts, and in having a long micropylar tube. Erdtmanispermum (Erdtmanithecales) differs from fossil Ephedra seeds in the three-angled rather than four-angled nature of the envelope, the absence of papillae on the envelope, and in the much smaller, iso-diametric sclerenchymatic cells under the epidermis. They also differ in having a distinct megaspore membrane, which is unknown for Ephedra. Eucommiidites pollen (Erdtmanithecales) is smaller than Ephedra grains and distinct with only three ridges.

The microsporangiate units of Erdtmanitheca are described from the Early Cretaceous Vila Verde locality (Pedersen et al., 1989), not far from Buarcos, in sediments that may be of approximately the same age as the Buarcos locality, and plants of Gnetales and Erdmanithecales occur together, e.g. in the clayballs at the Drewry's Bluff locality. The possibility that small angular seeds from Buarcos and Drewry's Bluff could belong to Erdtmanithecales cannot be excluded. However, seeds of Ephedra portugallica and Ephedra drewriensis have clear affinity to Ephedra based on their relatively large pollen grains with many ridges and the presence of discarded exines in the micropyle. Ephedrispermum lusitanicum is more problematic as it has small pollen grains with few ridges. Ephedrispermum is, however, clearly distinct from Erdmanithecales in the four-angled nature of the seeds, the shape of the sclerenchyma cells and in the presence of pollen grains with more than three ridges.

The evolutionary history of Ephedra

Fossil evidence

Polyplicate pollen grains that may resemble those of extant Ephedra and Welwitschia are common in the fossil record and extend back to the Permian (Wilson, 1962). They become more common in the Triassic (Traverse, 1988), decline again in the Jurassic and expand rapidly in both diversity and abundance in the Early Cretaceous (Crane and Lidgard, 1989; Osborn et al., 1993; Crane, 1996). Many of the Cretaceous dispersed polyplicate pollen types most likely represent the Gnetales (e.g. Osborn et al., 1993), but the systematic position of pre-Cretaceous polyplicate pollen is more uncertain. Some of the Permian pollen grains, such as Vittatina, are known to be produced by extinct seed plants (peltasperms), while others are of unknown origin (Meyen, 1987).

The earliest megafossil possibly representing the Gnetales, is the Permian cone Palaeognetaleana auspicia that has in situ polyplicate pollen (Wang, 2004), but details of the organization and structure of this plant are not fully established. Pre-Cretaceous megafossils associated with the Gnetales are often morphologically distinct from extant groups and difficult to interpret. The Late Triassic plant Dechellyia gormani (Ash, 1972) and the early Jurassic Piroconites kuespertii (van Konjinenburg-van Cittert, 1992) are other examples of fossil plants that have been compared with Gnetales but are of uncertain relationship.

A study of Cretaceous floristic diversity based on palynological assemblages suggest that there was a striking increase in gnetalean diversity simultaneously with the initial angiosperm radiation in the Early Cretaceous, but that the Gnetales underwent a significant decline again through the Late Cretaceous (Crane and Lidgard, 1989). Unequivocal gnetalean megafossils are now known from Early Cretaceous strata. The Crato flora from Brazil (Mohr and Friis, 2000; Mohr and Eklund, 2003; Mohr and Bernardes-de-Oliveira, 2004; Mohr et al., 2004; Dilcher et al., 2005) contains several ephedroid plants. The Yixian Formation in western Liaoning China, also famous for its feathered dinosaurs and other exquisitely preserved vertebrates, has proven particularly rich in gnetalean fossils. The lower part of the formation has been radiometrically dated to approx. 125 Myr, corresponding to the Barremian (Early Cretaceous), and the fossil-bearing sediments extend into the Aptian (for a review of the geology and biology of the Yixian Formation, see Zhou et al., 2003).

Several Ephedra-like impression fossils have been discovered from the Yixian Formation including currently unpublished material (C. Rydin, S. Q. Wu and E. M. Friis in progress). Eragrosites changii and Liaoxia chenii (Cao et al., 1998) from the Yixian Formation both have opposite branching and small cone-like reproductive structures. They were originally described as angiosperms but are clearly ephedroid and Liaoxia and Eragrosites were later combined and re-described jointly as Ephedrites chenii (Guo and Wu, 2000). Ephedra archaeorhytidosperma (Yang et al., 2005), also from the Yixian Formation, has seeds with transverse laminar protuberances, comparable to those of extant Ephedra rhytidosperma. Ephedrites? elegans (Sun et al., 2001) from the Early Cretaceous of Liaoning, China, Ephedrites sinensis and Ephedra exhibens (Wu et al., 1986) from the Jurassic Xiaomeigou Formation, Qinghai, China, and Alloephedra xingxuei (Tao and Yang, 2003) from the Early Cretaceous Dalazi Formation in Jinlin, China, also belong to the Gnetales and are probably most closely related to Ephedra.

The fossils presented here provide details of the reproductive parts of Cretaceous Ephedra-like plants that have not been available in previously described material. They document that plants with characters such as apical papillae on the outer sclerenchymatic envelope, and polyplicate pollen grains that discard the exine during germination were already established in the Early Cretaceous. Such plants were widespread across Laurasia and were present in what is now the Old World, as well as in the New World. Together with other material that is not preserved at the same level of detail, these fossils suggest that a diversity of Ephedra and Ephedra-like plants were present in the Early Cretaceous. The only difference in general appearance between the fossil seeds and those of extant Ephedra is size. Seeds of modern Ephedra are much larger than their Cretaceous relatives. Whereas the fossils are generally <1·5 mm long, the length of extant Ephedra seeds is often up to 6–7 mm or more. Increase in seed size through time has been observed also for several angiosperm taxa, e.g. Spirematospermum (extinct Zingiberales) (Friis, 1988) and Stratiotes (Hydrocharitales) (Mai, 1985).

Phylogenetic implications

The information currently available on Cretaceous ephedroids is insufficient for detailed phylogenetic analysis, and the phylogenetically isolated nature of extant Gnetales further complicates outgroup comparison. It is interesting, however, that seeds and pollen of Erdtmanithecales are very similar to those of Ephedra. Based mainly on the granular structure of the exine, and the possible homology between the two lateral colpi in Eucommiidites with the furrows of Ephedra and Welwitschia grains, it has been suggested that Eucommiidites may have been produced by plants related to the Gnetales (Doyle et al., 1975; Trevisan, 1980; Crane, 1985; Pedersen et al., 1989). The decussate arrangement of the microsporangiate units of Eucommiitheca is in agreement with the opposite, decussate or whorled phyllotaxis of the Gnetales, but the peltate microsporangia-bearing units of Erdtmanithecales are very different from those in extant Gnetales. Peltate structures are, however, present in Aegianthus sibiricus, supposedly gnetalean male cones from the mid-Jurassic of East Siberia (Krassilov and Bugdaeva, 1988).

Seeds of Erdtmanithecales and Gnetales are also similar in overall structure. An origin of the outer seed envelope from bracts has previously been suggested for extant Gnetales, e.g. by Crane (1985) and by Pedersen et al. (1989). This is consistent with ontogenetic studies of extant Ephedra, which show that the inner integument is initiated as a circular protrusion, while the seed envelope initiates as two, three or four protrusions and is apparently of foliar nature (Pearson, 1929; Takaso and Owens, 1994; Yang, 2001, 2004). The angled nature of the seed envelope is prominent in the fossil forms and it is very likely that the seed envelopes of Ephedra, Ephedrispermum (Gnetales) and Erdtmanispermum (Erdtmanithecales) are derived from paired or whorled bracts. It seems probable that the Cretaceous forms more evidently reflect this evolutionary origin. In conclusion, some but not all features of the Gnetales are established in the extinct plant group Erdtmanithecales, and whether this indicates a position of Erdtmanithecales as sister group to Gnetales needs to be tested as more information accumulates on Gnetales and their fossil relatives.

Ephedrispermum lusitanicum, described here, has rather small pollen (23 µm) and seems to lack papillae on the seed envelope, as do seeds of Erdtmanithecales, but Ephedrispermum is clearly distinct from Erdtmanithecales with its polyplicate pollen and four-angled seed envelope with rectangular sclerenchyma cells. The similarities between Erdtmanispermum (Erdtmanithecales) and Ephedrispermum lusitanicum (Gnetales) are probably plesiomorphies and not indications of a sister relationship between the two.

Seeds of Ephedra portugallica, Ephedra drewriensis and extant Ephedra are distinguished from those of Erdtmanispermum and Ephedrispermum by the presence of apical papillae on the seed envelope and pollen grains that discard the exine during germination. Resolving relationships among these fossil and extant species is difficult, because the extant genus comprise a homogenous group of species in which morphological characters are often more or less identical. Moreover, some characters are evidently homoplasious while others are sometimes not uniform within species or even within an individual (El-Ghazaly and Rowley, 1997; Huang, 2000; Ickert-Bond et al., 2003; Ickert-Bond and Wojciechowski, 2004; Rydin et al., 2004; Huang et al., 2005). Currently, the variation in anatomical details and reproductive characters among extant species of Ephedra is evidently complex and poorly understood [also see Ickert-Bond and Wojciechowski (2004) and Huang et al. (2005) for a discussion of extant morphology].

Evidence from molecular data provides limited information on relationships among extant species of Ephedra. A previous study (Rydin et al., 2004), investigated six gene regions from the chloroplast and the nuclear DNA, but all contained very few informative characters (Table 2). Most of the information came from the internal transcribed spacer (ITS) of the nuclear ribosomal DNA. As previously reported (Maggini et al., 1998), the ITS1 region of Ephedra is very long compared with ITS1 in angiosperms, and has three short subrepeats. However, only 170 of the total 1687 characters in ITS1 and ITS2 were informative. Phylogenetic analyses resulted in well-supported subgroups of Ephedra that correspond to geographical regions (Huang, 2000; Ickert-Bond and Wojciechowski, 2004; Rydin et al., 2004; Huang et al., 2005) with African–Mediterranean species in a basal grade or clade and Asian species forming two well-supported clades (Ickert-Bond and Wojciechowski, 2004; Rydin et al., 2004). New World species are monophyletic and comprise a South American clade (Ickert-Bond and Wojciechowski, 2004; Rydin et al., 2004) and a non-monophyletic grade of North American species (Ickert-Bond and Wojciechowski, 2004). Relationships among these subgroups are not well supported, and unfortunately they are not characterized by obvious morphological synapomorphies. It is consequently difficult to assign the fossils to any particular subgroup within Ephedra.

Table 2.
Further information on gene regions investigated in Rydin et al. (2004)

The age of Ephedra

A recent radiation of crown group Ephedra (Huang and Price, 2003), might be inferred from molecular data alone, but this requires that Cretaceous Ephedra (e.g. the fossil plants described here) belonged to extensively distributed but now extinct stem group(s), and that the modern diversity in the genus is the result of a second radiation in the Oligocene or Miocene. However, this supposedly second radiation does not appear to be associated with any morphological novelties. The consequence of this scenario is that despite an assumed high rate of molecular evolution in the Ephedra stem lineage (see below), there is no obvious morphological divergence between the Cretaceous and the modern species. An analogous pattern occurs in the lycopod genus Isoetes. As in Ephedra, modern species of Isoetes are morphologically similar and molecular divergence is very low. But phylogenetic analysis results in a tree topology that implies an ancient speciation pattern with multiple Gondwana distributions (Rydin and Wikström, 2002). This interpretation is in line with fossil evidence (possible crown group members are documented at least from the Early Cretaceous) and indicates that, perhaps as in Ephedra, low molecular and morphological divergence does not necessarily mean recent speciation.

A Cretaceous origin of crown group Ephedra raises, however, other questions that need to be answered. An old age of Ephedra requires extremely low substitution rates within the crown group, because the branches are very short (see, for example, the phylograms in Rydin et al., 2002). In contrast, the branch leading to extant Ephedra is very long and this has been attributed, at least in part, to a high substitution rate (Sanderson, 2002; Magallón and Sanderson, 2005). Studies have shown that the branches leading to all the extant genera of the Gnetales have the highest estimated substitution rates among seed plants (Sanderson, 2002; Magallón and Sanderson, 2005). If the crown group originated in the Early Cretaceous, there must accordingly have been a shift from a very high substitution rate in ephedran stem lineages to a very low substitution rate in the crown group.

Frequently used molecular dating methods often smooth rate variation among adjacent branches on a phylogeny (for a review of methods for estimating divergence times see, for example, Sanderson et al., 2004), and the concept of temporal autocorrelation (see Gillespie, 1991) implies that substitution rate is a heritable trait that thus is likely to be similar among closely related lineages. A sudden and substantial change in this trait is of course possible, as in any trait, [remarkable differences in branch lengths among sister clades, thus differences in substitution rates, have been proposed, for instance, for the lycopods Selaginella–Isoetes (Rydin and Wikström, 2002; Korall and Kenrick, 2004), and filmy ferns Hymenopyllum s.l.-Trichomanes s.l. (Pryer et al., 2001)], but its basis remains to be explained.

Further, currently used methods for molecular dating would not be able to manage such significant rate changes. For instance, a dramatic slowdown in rate of molecular evolution has been proposed for the eusporangiate ferns Marattia-Angiopteris and the tree ferns (Soltis et al., 2002). These groups have comparatively short branch lengths, and despite fossil documentation at least from mid-Jurassic, probably further back, their estimated ages were substantially younger. The authors refer to Marattiales and tree ferns as ‘molecular living fossils’, consistent with their relative morphological stasis for the past 165–200 million years (Soltis et al., 2002), and they argue that even dating methods that allow rate heterogeneity among lineages could not handle the dramatic decrease in rates inferred for these groups.

Conclusions

The Early Cretaceous (late Barremian to early Albian) seeds described here are assigned to three new species, of which two have a character combination that excludes a relationship to any other group but Ephedra. These fossils document that key reproductive features characteristic of extant Ephedra (i.e. an extra-integumentary seed envelope with apical papillae, and polyplicate pollen grains that shed the exine during germination), were established and widespread in the northern hemisphere in the early Aptian and have thus been retained for at least 120 million years. The third new species possesses some, but not all, of these characters. Further, a diversity of other seeds, which may belong to the ephedran lineage, have been presented briefly. The seeds have been found in sediments from what is now the Old World (Portugal) as well as in the New World (Virginia, USA) and the results demonstrate that a diversity of Ephedra and ephedroid species were present in the Early Cretaceous flora. The fossil seeds share some characters with seeds of Erdtmanithecales and this supports the previously suggested relationship between Erdtmanithecales and Gnetales.

Even though these seeds document the presence of unique Ephedra characters in the Early Cretaceous, secure evidence on the age of Ephedra is not available. The homoplasious nature of many morphological characters in extant Ephedra, intraspecific morphological variation, and sparse information in investigated gene regions, make it difficult to resolve phylogenetic relationships within the genus and therefore to assign fossils to extant subgroups. Two alternative hypotheses on the age of Ephedra: an origin of extant Ephedra in the Oligocene-Miocene based on molecular dating analysis (Huang and Price, 2003) and a Cretaceous origin proposed here, both leave key questions unanswered. Nevertheless, because of the identical reproductive characters in fossil and extant Ephedra presented here, it is probable that crown group Ephedra is of Mesozoic origin, an idea further supported by the Early Cretaceous fossil Ephedra archaeorhytidosperma (Yang et al., 2005), that has characters only present in some extant species.

Acknowledgments

The authors thank Peter Endress, Petra Korall and Susana Magallón for valuable comments on the text and Yvonne Arremo for technical assistance. The project was supported by grants from the Swedish Research Council to Else Marie Friis, from the Carlsberg Foundation to Kaj Raunsgaard Pedersen, and from the Royal Swedish Academy of Sciences to Catarina Rydin.

LITERATURE CITED

  • Arber EAN, Parkin J. 1908. Studies on the evolution of angiosperms: the relationship of the angiosperms to the Gnetales. Annals of Botany 22: 489–515.
  • Ash SR. 1972. Late Triassic plants from the Chinle Formation in north-eastern Arizona. Palaeontology 15: 598–618.
  • Bowe LM, Coat G, de Pamphilis CW. 2000. Phylogeny of seed plants based on all three genomic compartments: extant gymnosperms are monophyletic and Gnetales' closest relatives are conifers. Proceedings of the National Academy of Sciences of the USA 97: 4092–4097. [PMC free article] [PubMed]
  • Brenner GJ. 1963. The spores and pollen of the Potomac Group of Maryland. Maryland Department of Geology, Mines and Water Resources, Bulletin 27: 1–215.
  • Brenner GJ. 1967. The gymnospermous affinity of Eucommiidites Erdtman, 1948. Review of Palaeobotany and Palynology 5: 123–127.
  • Cao Z, Wu S, Zhang P, Li J. 1998. Discovery of fossil monocotyledons from Yixian Formation, western Liaoning. Chinese Science Bulletin 43: 230–233.
  • Chaw S-M, Parkinson CL, Cheng Y, Vincent TM, Palmer JD. 2000. Seed plant phylogeny inferred from all three plant genomes: monophyly of extant gymnosperms and origin of Gnetales from conifers. Proceedings of the National Academy of Sciences of the USA 97: 4086–4091. [PMC free article] [PubMed]
  • Couper RA. 1956. Evidence for a possible gymnospermous affinity for Tricolpites troedssonii Erdtman. New Phytologist 55: 280–285.
  • Couper RA. 1958. British Mesozoic microspores and pollen grains: a systematic and stratigraphic study. Palaeontographica B 103: 75–179.
  • Crane PR. 1985. Phylogenetic analysis of seed plants and the origin of angiosperms. Annals of the Missouri Botanical Garden 72: 716–793.
  • Crane PR. 1996. The fossil history of the Gnetales. International Journal of Plant Sciences 157 (Suppl. 6): S50–S57.
  • Crane PR, Lidgard S. 1989. Angiosperm diversification and palaeolatitudinal gradients in Cretaceous floristic diversity. Science 246: 675–678. [PubMed]
  • Crane PR, Upchurch GR. 1987. Drewria potomacensis gen. et sp. nov., an Early Cretaceous member of Gnetales from the Potomac Group of Virginia. American Journal of Botany 74: 1722–1736.
  • Dilcher DL, Bernardes-de-Oliveira ME, Pons D, Lott TA. 2005. Welwitschiaceae from the Lower Cretaceous of Northeastern Brazil. American Journal of Botany 92: 1294–1310. [PubMed]
  • Dinis JL. 2001. Definicão da Formacão da Figueira da Foz—Aptiano a Cenomaniano do sector central da margem oeste ibérica /Definition of the Figueira da Foz Formation—Aptian to Cenomanian of the central sector of the western Iberian margin. Comunicações Instituto Geológico e Mineiro 88: 127–160.
  • Doyle JA. 1992. Revised palynological correlations of the lower Potomac Group (USA) and the Cocobeach sequence of Gabon (Barremian-Aptian). Cretaceous Research 13: 337–349.
  • Doyle JA. 1996. Seed plant phylogeny and the relationships of Gnetales. International Journal of Plant Sciences 157 (Suppl. 6): S3–S39.
  • Doyle JA, Donoghue MJ. 1986. Seed plant phylogeny and the origin of angiosperms: an experimental cladistic approach. Botanical Review 52: 321–431.
  • Doyle JA, Donoghue MJ. 1992. Fossils and seed plant phylogeny reanalyzed. Brittonia 44: 89–106.
  • Doyle JA, Hickey LJ. 1976. Pollen and leaves from the mid-Cretaceous Potomac Group and their bearing on early angiosperm evolution. In: Beck CB, ed. Origin and early evolution of angiosperms. New York, NY: Columbia University Press, 139–206.
  • Doyle JA, Van Campo M, Lugardon B. 1975. Observations on exine structure of Eucommiidites and Lower Cretaceous angiosperm pollen. Pollen et Spores 17: 429–486.
  • Duan S. 1998. The oldest angiosperm—a tricarpous female reproductive fossil from western Liaoning Province, NE China. Science in China 41: 14–20.
  • Eames AJ. 1952. Relationships of Ephedrales. Phytomorphology 2: 79–100.
  • El-Ghazaly G, Rowley JR. 1997. Pollen wall of Ephedra foliata. Palynology 21: 7–18.
  • El-Ghazaly G, Rowley JR, Hesse H. 1998. Polarity, aperture condition and germination in pollen grains of Ephedra (Gnetales). Plant Systematics and Evolution 213: 217–231.
  • Erdtman G. 1948. Did dicotyledonous plants exist in early Jurassic time? Geologisk Förening i Stockholm Förhandlingar 70: 265–271.
  • Freitag H, Maier-Stolte M. 1994. Characterization of areas, Ephedraceae. In: Browicz K, ed. Chorology of trees and shrubs in southwest Asia and adjacent regions. Kornik: Polish Academy of Sciences, Institute of Dendrology, 5–16.
  • Friis EM. 1988. Spirematospermum chandlerae sp. nov., an extinct species of Zingiberaceae from the North American Cretaceous. Tertiary Research 9: 7–12.
  • Friis EM, Pedersen KR. 1996. Eucommiitheca hirsuta, a new pollen organ with Eucommiidites pollen from the Early Cretaceous of Portugal. Grana 35: 104–112.
  • Friis EM, Crane PR, Pedersen KR. 1988. Reproductive structure of Cretaceous Platanaceae. Biologiske Skrifter, Det Kongelige Danske Videnskabernes Selskab 31: 1–55.
  • Friis EM, Crane PR, Pedersen KR. 1997. Anacostia, a new basal angiosperm from the Early Cretaceous of North America and Portugal with monocolpate/trichotomocolpate pollen. Grana 36: 225–244.
  • Friis EM, Pedersen KR, Crane PR. 1999. Early angiosperm diversification: the diversity of pollen associated with angiosperm reproductive structures in Early Cretaceous floras from Portugal. Annals of the Missouri Botanical Garden 86: 259–296.
  • Friis EM, Pedersen KR, Crane PR. 2000. a Fossil floral structures of a basal angiosperm with monocolpate, reticulate-acolumellate pollen from the Early Cretaceous of Portugal. Grana 39: 226–245.
  • Friis EM, Pedersen KR, Crane PR. 2000. b Reproductive structure and organization of basal angiosperms from the Early Cretaceous (Barremian or Aptian) of Western Portugal. International Journal of Plant Sciences 161 (Suppl. 6): S169–S182.
  • Friis EM, Pedersen KR, Crane PR. 2004. Araceae from the Early Cretaceous of Portugal: evidence on the emergence of monocotyledons. Proceedings of the National Academy of Sciences of the USA 101: 16565–16570. [PMC free article] [PubMed]
  • Gillespie JH. 1991. The causes of molecular evolution. Oxford, UK: Oxford University Press.
  • Goremykin V, Bobrova VK, Pahnke J, Troitsky AV, Antonov A, Martin W. 1996. Noncoding sequences from the slowly evolving chloroplast inverted repeat in addition to rbcL data do not support gnetalean affinities with angiosperms. Molecular Biology and Evolution 13: 383–396. [PubMed]
  • Guo S, Wu X. 2000. Ephedrites from latest Jurassic Yixian Formation in western Liaoning, Northeast China. Acta Palaeontologica Sinica 39: 81–91.
  • Heimhofer U, Hochuli PA, Burla S, Dinis JL, Weissert H. 2005. Timing of Early Cretaceous angiosperm diversification and possible links to major paleoenvironmental change. Geology 33: 141–144.
  • Huang J. 2000. Molecular systematics and evolution of the genus Ephedra. PhD Dissertation, University of Georgia, Athens.
  • Huang J, Price RA. 2003. Estimation of the age of extant Ephedra using chloroplast rbcL sequence data. Molecular Biology and Evolution 20: 435–440. [PubMed]
  • Huang J, Giannasi DE, Price RA. 2005. Phylogenetic relationships in Ephedra (Ephedraceae) inferred from chloroplast and nuclear DNA sequences. Molecular Phylogenetics and Evolution 35: 48–59. [PubMed]
  • Hughes NF. 1961. Further interpretation of Eucommiidites Erdtman 1948. Palaeontology 4: 292–299.
  • Ickert-Bond SM, Wojciechowski MF. 2004. Phylogenetic relationships in Ephedra (Gnetales): evidence from nuclear and chloroplast DNA sequence data. Systematic Botany 29: 834–849.
  • Ickert-Bond SM, Skvarla JJ, Chissoe WF. 2003. Pollen dimorphism in Ephedra L. (Ephedraceae). Review of Palaeobotany and Palynology 124: 325–334.
  • van Konjinenburg-van Cittert JHA. 1992. An enigmatic Liassic microsporophyll, yielding Ephedripites pollen. Review of Palaeobotany and Palynology 71: 239–254.
  • Korall P, Kenrick P. 2004. The phylogenetic history of Selaginellaceae based on DNA sequences from the plastid and nucleus: extreme substitution rates and rate heterogeneity. Molecular Phylogenetics and Evolution 31: 852–864. [PubMed]
  • Krassilov VA. 1982. Early Cretaceous flora of Mongolia. Palaeontographica B 181: 1–43.
  • Krassilov VA. 1986. New floral structures from the Lower Cretaceous of Lake Baikal area. Review of Palaeobotany and Palynology 47: 9–16.
  • Krassilov V, Bugdaeva EV. 1988. Gnetalean plants from the Jurassic of Ust-Balej, East Siberia. Review of Palaeobotany and Palynology 53: 359–374.
  • Krassilov VA, Dilcher DL, Douglas JG. 1998. New ephedroid plant from the Lower Cretaceous Koonwarra Fossil Bed, Victoria, Australia. Alcheringa 22: 123–133.
  • Kubitzki K. 1990. Ephedraceae. In: Kubitzki K, ed. The families and genera of vascular plants. I. Pteridophytes and gymnosperms. Berlin/Heidelberg: Springer-Verlag, 379–382.
  • Kvacek J, Pacltová B. 2001. Bayeritheca hughesii gen. et sp. nov., a new Eucommiidites-bearing pollen organ from the Cenomanian of Bohemia. Cretaceous Research 22: 695–704.
  • Magallón S, Sanderson MJ. 2002. Relationships among seed plants inferred from highly conserved genes: sorting conflicting phylogenetic signals among ancient lineages. American Journal of Botany 89: 1991–2006. [PubMed]
  • Magallón S, Sanderson MJ. 2005. Angiosperm divergence times: the effect of genes, codon positions, and time constraints. Evolution 59: 1653–1670. [PubMed]
  • Maggini F, Marrocco R, Gelati MT, de Dominicis RI. 1998. Lengths and nucleotide sequences of the internal spacers of nuclear ribosomal DNA in gymnosperms and pteridophytes. Plant Systematics and Evolution 213: 199–205.
  • Mai DH. 1985. Entwicklung der Wasser- und Sumpfpflanzen-Gesellschaften Europas von der Kreide bis ins Quartär. Flora 176: 449–511.
  • Meyen SV. 1987. Fundamentals of palaeobotany. London, New York: Chapman and Hall.
  • Mohr B, Eklund H. 2003. Araripia florifera, a magnoliid angiosperm from the Lower Cretaceous Crato Formation (Brazil). Review of Palaeobotany and Palynology 126: 279–292.
  • Mohr B, Friis EM. 2000. Early angiosperms from the Aptian Crato Formation (Brazil), a preliminary report. International Journal of Plant Sciences 161 (Suppl. 6): S155–S167.
  • Mohr BAR, Bernardes-de-Oliveira MEC. 2004. Endressinia brasiliana, a Magnolialean angiosperm from the Lower Cretaceous Crato Formation (Brazil). International Journal of Plant Sciences 165: 1121–1133.
  • Mohr BAR, Bernardes-de-Oliveira ME, Barreto AMF, Castro-Fernandes MC. 2004. Gnetophyte preservation and diversity in the early Cretaceous Crato Formation (Brazil). The Seventh International Organisation of Palaeobotany Conference, Bariloche, Argentina. Abstract: 81.
  • Mussayev IF. 1978. On geography and phylogeny of some representatives of the genus Ephedra L. Botanicheskij Zhurnal 63: 523–543.
  • Nixon KC, Crepet WL, Stevenson D, Friis EM. 1994. A reevaluation of seed plant phylogeny. Annals of the Missouri Botanical Garden 81: 484–533.
  • Osborn JM. 2000. Pollen morphology and ultrastructure of gymnospermous anthophytes. In: Harley MM, Morton CM, Blackmore S, eds. Pollen and spores: morphology and biology. London: Royal Botanical Gardens, Kew, 163–185.
  • Osborn JM, Taylor TN, Lima MRd. 1993. The ultrastructure of fossil ephedroid pollen with gnetalean affinities from the Lower Cretaceous of Brazil. Review of Palaeobotany and Palynology 77: 171–184.
  • Pais J, Reyre Y. 1981. Problèmes posés par la population sporo-pollinique d′un niveau à plantes de la série de Buarcos (Portugal). Sociedade Geológica de Portugal Boletim 22: 35–40.
  • Pearson HHW. 1929. Gnetales. Cambridge: Cambridge University Press.
  • Pedersen KR, Crane PR, Friis EM. 1989. Pollen organs and seeds with Eucommiidites pollen. Grana 28: 279–294.
  • Pryer KM, Smith AR, Hunt JS, Dubuisson J-Y. 2001. rbcL data reveal two monophyletic groups of filmy ferns (Filicopsida: Hymenophyllaceae). American Journal of Botany 88: 1118–1130. [PubMed]
  • Rey J. 1993. Comunicações Instituto Geológico e Mineiro 79: 75–85.
  • Rocha R, Manuppella G, Mouteride R, Ruget C, Zbyszewski G. 1981. Carta geológica de Portugal na escala de 1/50,000. Notícia explicativa da folha 19-C Figueira da Foz. Lisbon: Serviços Geológicos de Portugal.
  • Rydin C, Friis EM. 2005. Pollen germination in Welwitschia mirabilis Hook. f.: differences between the polyplicate pollen producing genera of the Gnetales. Grana 44: 137–141.
  • Rydin C, Källersjö M. 2002. Taxon sampling and seed plant phylogeny. Cladistics 18: 485–513.
  • Rydin C, Wikström N. 2002. Phylogeny of Isoetes (Lycopsida): resolving basal relationships using rbcL sequences. Taxon 51: 83–89.
  • Rydin C, Källersjö M, Friis EM. 2002. Seed plant relationships and the systematic position of Gnetales based on nuclear and chloroplast DNA: conflicting data, rooting problems and the monophyly of conifers. International Journal of Plant Sciences 163: 197–214.
  • Rydin C, Mohr B, Friis EM. 2003. Cratonia cotyledon gen. et sp. nov.: a unique Cretaceous seedling related to Welwitschia. Biology Letters. The Royal Society of London 270: 29–32. [PMC free article] [PubMed]
  • Rydin C, Pedersen KR, Friis EM. 2004. On the evolutionary history of Ephedra: Cretaceous fossils and extant molecules. Proceedings of the National Academy of Sciences of the USA 101: 16571–16576. [PMC free article] [PubMed]
  • Sanderson MJ. 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Molecular Biology and Evolution 19: 101–109. [PubMed]
  • Sanderson MJ, Thorne JL, Wikström N, Bremer K. 2004. Molecular evidence on plant divergence times. American Journal of Botany 91: 1656–1665. [PubMed]
  • Soltis PS, Soltis DE, Savolainen V, Crane PR, Barraclough TG. 2002. Rate heterogeneity among lineages of tracheophytes: integration of molecular and fossil data and evidence for molecular living fossils. Proceedings of the National Academy of Sciences of the USA 99: 4430–4435. [PMC free article] [PubMed]
  • Stapf O. 1889. Die Arten der Gattung Ephedra. Denkschriften der Mathematisch-Naturwissenschaftlichen Classe der Kaiserlichen Akademie der Wissenschaften Wien 56: 1–112.
  • Steeves MW, Barghoorn ES. 1959. The pollen of Ephedra. Journal of the Arnold Arboretum 40: 221–255.
  • Sun G, Zheng S, Dilcher DL, Wang Y, Mei S. 2001. Early angiosperms and their associated plants from western Liaoning, China. Shanghai: Shanghai Scientific and Technological Education Publishing House.
  • Takaso T, Owens JN. 1994. Effects of ovular secretions on pollen in Pseudotsuga menziesii (Pinaceae). American Journal of Botany 81: 504–513.
  • Tao J, Yang Y. 2003. Alloephedra xingxuei gen. et sp. nov., an Early Cretaceous member of Ephedraceae from Dalazi Formation in Yanji Basin, Jilin Province of China. Acta Palaeontologica Sinica 42: 208–215.
  • Thoday MG, Berridge EM. 1912. The anatomy and morphology of the inflorescences and flowers of Ephedra. Annals of Botany 26: 953–985.
  • Traverse A. 1988. Palaeopalynology. Boston, MA: Unwin Hyman.
  • Trevisan L. 1980. Ultrastructural notes and considerations on Ephedrites, Eucommiidites and Monosulcites pollen grains from Lower Cretaceous sediments of southern Tuscany (Italy). Pollen et Spores 22: 85–132.
  • Wang Z. 2004. A new Permian gnetalean cone as fossil evidence for supporting current molecular phylogeny. Annals of Botany 94: 281–288. [PubMed]
  • Wilson LR. 1962. Permian plant microfossils from the Flowerpot Formation, Oklahoma. Oklahoma Geological Survey Bulletin 49: 5–50.
  • Wu S, Duan S, Mohr B, Pedersen KR, Friis EM. 2000. Early Cretaceous diversity of Gnetales: macro- and mesofossil evidence from China, Brazil, and Portugal. The Sixth Conference of the International Organization of Palaeobotany, Qinhuangdao, China. Abstract volume: 37–38.
  • Wu X, He Y, Mei S. 1986. Discovery of Ephedrites from the Lower Jurassic Xiaomeigou Formation, Qinghai. Acta Palaeobotanica et Palynologica Sinica 8: 13–21.
  • Yang Y. 2001. Ontogeny and metamorphic patterns of female reproductive organs of Ephedra sinica Stapf (Ephedraceae). Acta Botanica Sinica 43: 1011–1017.
  • Yang Y. 2004. Ontogeny of triovulate cones of Ephedra intermedia and origin of the outer envelope of ovules of Ephedraceae. American Journal of Botany 91: 361–368. [PubMed]
  • Yang Y, Geng B, Dilcher DL, Chen Z, Lott TA. 2005. Morphology and affinities of an Early Cretaceous Ephedra (Ephedraceae) from China. American Journal of Botany 92: 231–241. [PubMed]
  • Zbyszewski G, Moitinho d'Almeida F, Torre de Assunção C. 1955. Carta geológica de Portugal na escala de 1/50,000. Notícia explicativa da folha 30-C Torres Vedras. Lisbon: Serviços Geológicos de Portugal.
  • Zhou Z, Barrett PM, Hilton J. 2003. An exceptionally preserved Lower Cretaceous ecosystem. Nature 421: 807–814. [PubMed]

Articles from Annals of Botany are provided here courtesy of Oxford University Press
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...