Discussion

CM01 is the most abundant morphotype in the Soap Wash collection, represented by 43 well-preserved specimens. These specimens fit the generic description of Pandemophyllum provided by Upchurch and Dilcher (Reference Upchurch and Dilcher1990) in details of size, shape and venation, but they do not correspond to any previously described species. CM01 has an acute and asymmetrical leaf apex that is not mentioned or illustrated for any of the described species of Pandemophyllum. Upchurch and Dilcher (Reference Upchurch and Dilcher1990) noted other Dakota Formation forms, such as “Inga” cretacea Lesquereux Reference Lesquereux1891 (Upchurch and Dilcher, Reference Upchurch and Dilcher1990, pl. 14, fig. 1), that shared similarities with Pandemophyllum and also showed asymmetry in the leaf apex. CM01 shares morphological characteristics with “Ficus” inaequalis Lesquereux Reference Lesquereux1891 from the Dakota Formation of Kansas (Lesquereux, Reference Lesquereux1891 [1892], p. 82, pl. XLIX, figs. 6–8, pl. L, fig. 3). However Lesquereux described the species as craspedodromous, whereas CM01 has brachidodromous secondary venation. Berry (Reference Berry1919, p. 80, pl. XII, fig. 1) reported “Ficus” inaequalis from the Gulf Coastal Plain and illustrated brochidodromous specimens in which secondary vein loops are flattened at the margin in the apical half of the lamina. Gulf Coastal Plain representatives also display a rounded base (Berry, Reference Berry1919), which is circumscribed into the range of variation typical of CM01. The illustrated specimen of “Ficus” inaequalis (Berry, Reference Berry1919, pl. XII, fig. 1) is distinctly symmetrical although the description notes that the species is “inequilateral” (Berry, Reference Berry1919, p. 80). Berry (Reference Berry1919) also segregated two additional species—F. daphnogenoides (Heer) Berry Reference Berry1919 (p. 80, pl. XIII, figs. 6–7) and F. shirleyensis Berry Reference Berry1919 (p. 81, pl. XIII, fig. 8)—that share affinities with Pandemophyllum, but the bases of these species are distinctly and uniquely acute and likely fall within the range of variation circumscribed for Pandemophyllum kvacekii Upchurch and Dilcher Reference Upchurch and Dilcher1990. Therefore, we place CM01 in Pandemophyllum but do not assign it to a previously described species.

Morphotype CM02

Figure 3.7

Discussion

Preservation precludes determining the number and the arrangement of leaflets in the Cedar Mountain specimens.

Sapindopsis Fontaine Reference Fontaine1889 was originally recognized for compound leaves of lanceolate shape from the Potomac Group that had a generally lancet shape, prominent primary vein and untoothed margin. Leaflets occur in opposite pairs with the uppermost pairs of leaflets decurrent. Golovneva (Reference Golovneva2007) described a larger collection from the middle Albian to Cenomanian of Central Asia and western Siberia that circumscribed a wider range of variation, including toothed margins and a wider range in shape in proximal leaflets. In this collection, a single whole leaf (PB3012) with attached leaflets is preserved, but the leaflets are folded over one another so the decurrent attachment cannot be confirmed. Isolated leaflets comprise the remainder of the collection and are allied with specimen PB3012 by venation and shape, although shape, particularly L:W ratio, is variable. Fontaine (Reference Fontaine1889) described several species including Sapindopsis variabilis (p.298, pl. CLI, fig. 1, pl. CLII, figs. 1, 4, pl. CLIII, figs. 2, 3, pl. CLIV, figs. 2–4, pl. CLV, figs. 2–5) and S. magnifolia (p. 297, pl. CLI, figs. 2, 3, pl. CLII, figs. 2, 3, pl. CLIII, fig. 2, pl. CLIV, figs. 1, 5). Together these forms circumscribe the range of variation observed in the small Soap Wash collection. Peppe and colleagues (Reference Peppe, Hickey, Miller and Green2008) reported S. magnifolia (AS19) from the Aspen Shale (Albian) of Wyoming. Some of the specimens that they illustrated (e.g., YPM56073a, Peppe et al., Reference Peppe, Hickey, Miller and Green2008, fig. 19) have slightly obovate, rather than elliptical shape and more rounded apices than those recovered from the Cedar Mountain Formation. This variation may reflect position of leaflets on the leaf and so we place the Cedar Mountain specimens under this name.

Biogeographic and stratigraphic analysis suggests that Sapindopsis evolved in Eurasia during the Albian and migrated to North America during the early–middle Albian climate optimum (Golovneva, Reference Golovneva2007). Hickey and Doyle (Reference Hickey and Doyle1977) reported that Sapindopsis with untoothed margins were among the most common elements of the Albian floras of the Potomac Group of the mid-Atlantic region of North America. Three species of Sapindopsis are reported from the Albian-age Aspen Shale of Wyoming, with two of these, S. magnifolia and S. belviderensis, among the most common elements of the flora (Peppe et al., Reference Peppe, Hickey, Miller and Green2008).

Morphotype CM03

Figure 3.8–3.10

Discussion

Our review of the literature has not revealed a previously described species that matches the range of variation present in this morphotype. Details of secondary venation, particularly the basal pair of secondary veins that branch at low angle, and tertiary venation, particularly tertiaries in the basal half of the lamina that are regularly spaced and more uniformly perpendicular to the midrib ally this form with the modern genus Cinnamomum Schäffer Reference Schäffer1762 in the Lauraceae. Seward (Reference Seward1925) erected the genus Cinnamomoides Seward Reference Seward1925 for Cretaceous age leaves that resembled living Cinnamomum but with dubious alliance to Lauraceae. Cinnamomoides ovalis (Dawson) Bell Reference Bell1956 was described from the Albian of Alberta and British Columbia. Crabtree (Reference Crabtree1987) reported this species from similar age rocks in Montana. A variety of other species have been ascribed to the modern genus including “Cinnamomum” scheuchzeri Heer Reference Heer1856 from the Dakota Formation (Lesquereux, Reference Lesquereux1891 [1892]), and “Cinnamomum” canadense Dawson Reference Dawson1885 from the Albian of Alberta (Crabtree, Reference Crabtree1987). While it seems clear that CM03 is allied with the “Cinnamomophylls” of Crabtree (Reference Crabtree1987), this may be a polyphyletic group with an as yet undetermined diversity in the mid-Cretaceous. In the Cedar Mountain flora, this form is variable in size, shape and the details of venation, particularly at the lamina base. Consequently, it does not fit easily into any previously described species concept.

Morphotype CM04

Figure 3.11–3.14

Discussion

The acropetiolar gland characteristic of this morphotype is obvious and distinct. We have found no other specimens reported in the literature from the Early to mid-Cretaceous that preserve this feature. Therefore, we do not ally CM04 with any previously described taxon.

Morphotype CM05

Figure 4.1–4.3

Figure 4

Morphotype CM05 (Magnoliaephyllum sp.): (1) PB2041, (2) PB2235, (3) PB2982. Morphotype CM06: (4) PB2989. Morphotype CM07 (cf. ”Ziziphus” fibrillosus): (5) PB2559, (6) PB2052, (7) PB2983. Morphotype CM18 (“Cinnamomum” newberryi): (8) PB2980, (9) PB2984, (10) PB2240.

Discussion

The shape of the lamina in combination with the characteristics of the secondary and tertiary venation set these specimens apart from others in the Soap Wash collection. The genus Magnoliaephyllum Seward Reference Seward1926 includes leaves that resembled living Magnolia in leaf shape but for which the details of venation were not preserved. In his discussion, Seward (Reference Seward1926) noted that many paleobotanical specimens from the Cretaceous and Tertiary had been ascribed to the modern genus Magnolia without preservation of the details of tertiary, or sometimes even secondary, venation needed to make the link to the extant genus. He therefore erected the genus Magnoliaphyllum for these Magnolia-like leaves. In the case of CM05, tertiary venation is preserved and resembles, but is not identical to, modern Magnolia in having a tertiary vein fabric that is generally opposite percurrent with generally sinuous course. Therefore, we have chosen Seward’s (Reference Seward1926) form genus for these specimens. CM05 is also smaller than most modern members of the genus. Two species of “Magnolia” have been described from the mid-Cretaceous of North America. “Magnolia” alternans Heer 1866 (Capellini and Heer, Reference Capellini and Heer1866) is reported from the Albian of Alberta and British Columbia (Crabtree, Reference Crabtree1987) and the Cenomanian of New Jersey (Newberry, Reference Newberry1895) and Nebraska (Capellini and Heer, Reference Capellini and Heer1866). Dawson (Reference Dawson1885 [1886]) reported “Magnolia” magnifica Dawson Reference Dawson1882 [1883] from the Albian of Alberta. CM05 is consistently smaller than the specimens ascribed to these species; therefore, we do not assign CM05 to them, particularly since the details of venation are not described in the original publications.

Morphotype CM06

Figure 4.4

Discussion

Peppe and colleagues (Reference Peppe, Hickey, Miller and Green2008) reported a similar form from the middle to late Albian Aspen Shale from Lincoln County, Wyoming. Illustrated specimens of AS16 are similar in size to PB2989 from the Cedar Mountain Formation and smaller than the large specimen PB2075. Peppe and colleagues (Reference Peppe, Hickey, Miller and Green2008) assign the form to the Magnoliopsida but do not make a family or genus determination. Crabtree (Reference Crabtree1987) placed similar specimens including those described as “Platanus” (Crabtree Reference Crabtree1987, fig. 10) and Platanoplyllum (Crabtree, Reference Crabtree1987, fig. 11) in his broad Platanophylls category. Without more details of venation, even this assignment would be provisional for CM06. In shape, size and primary venation, CM06 also resembles Aralia eutawensis Berry 1914 (Berry Reference Berry1914b, p. 119, pl. XX, fig. 7) from the Late Cretaceous Eutaw Formation of Georgia and A. newberryi Berry Reference Berry1907 (p. 201, pl. XV, fig. 1) from the Late Cretaceous Raritan Formation of New York and New Jersey, the Magothy Formation of New Jersey, and the Black Creek Formation of North Carolina. Both A. eutawensis and A. newberryi are quinquelobate with the two basal lobes directed outward perpendicularly from the midvein and the other three directed laterally. As in CM06, Berry’s (Reference Berry1907, Reference Berry1914b) specimens do not preserve venation other than the palmately compound primaries, which make allying CM06 with these forms tentative. Most previous taxonomic assignments appear to have been based on overall shape of the leaf because the details of secondary and more detailed venation are poorly preserved. Since a diversity of described forms display quinquelobate shape, we are reluctant to ally CM06 to any previously described taxon based only on shape.

Morphotype CM07

Figure 4.5–4.7

Discussion

Simple, unlobed leaves with actinodromous venation are common in the latest Cretaceous and Tertiary floras of the mid and high latitudes of the Northern Hemisphere (Crane et al., Reference Crane, Manchester and Dilcher1991) but rare in Early and mid-Cretaceous. Our review of previously described floras yielded no convincing match to CM07 in collections from the Early Cretaceous through the Cenomanian. In his dissertation, Miller (Reference Miller2007) erected the new genus Paterosia Miller and Hickey in the Piperales to include simple, unlobed leaves with actinodromous venation. Miller (Reference Miller2007) described five species of Paterosia from the Albian-age Winthrop Formation of Washington State. As of this writing, this work, and the new taxa erected as part of it, has not been formally published and only text descriptions, not illustrations, are available for comparison. Verbal descriptions of Paterosia are consistent with CM07 in having 3–5 basal primary veins, a marginal petiole, simple agrophic veins and tertiary vein networks that may be opposite percurrent and oriented perpendicular to the primary veins. However, Miller (Reference Miller2007) defines secondary venation as brochidodromous and festooned brochidodromous rather than eucamptodromous as in CM07. Miller (Reference Miller2007) further noted that Paterosia species are commonly asymmetrical and have a deeply cordate or lobed base, both features absent in CM07. Therefore, we do not align CM07 with this Paterosia.

The closest match to CM07 that we were able to locate is “Zizyphus” fibrillosus Lesquereux Reference Lesquereux1878. Note that Lesquereux’s spelling of the genus name (Zizyphus versus Ziziphus, the spelling commonly used for the extant form) seems to be used interchangeably in the literature and many contemporary databases (e.g., Paleobiology Database, https://paleobiodb.org) with “Ziziphus” fibrillosus, which is favored by most 20^th Century authors publishing on this fossil form. “Ziziphus” appears to be preferred based on the extant genus Ziziphus Miller 1754 (Sprague, Reference Sprague1929), which defers to a spelling used before formal, Latinized names were standardized for species. Lesquereux (Reference Lesquereux1878) described Z. fibrillosus with both a rounded (p. 276, pl. LII, figs. 1, 3, 4) and cordate (p. 276, pl. LII, fig. 2) base and acute apex. Margin lacks teeth. Lesquereux (Reference Lesquereux1878) described three basally palmate primaries with a pair of lesser veins emerging from the point of divergence (p. 276, pl. LII, fig. 1). Simple agrophic veins are illustrated and opposite percurrent veins emerging at right angles to the primaries are described and illustrated. This form is described from the Late Cretaceous (Maastrichtian) Wyoming (Lesquereux, Reference Lesquereux1878), North and South Dakota (Johnson, Reference Johnson2002), the Paleocene of Montana and Colorado (Johnson and Ellis, Reference Johnson and Ellis2002; Wilf et al., Reference Wilf, Labandeira, Johnson and Ellis2006), and the Eocene of Wyoming (Knowlton, Reference Knowlton1899). This species has not previously been reported from the Early Cretaceous, therefore assigning CM07 to this form taxon would be a significant range extension. However, a broadly similar form, “Zizyphus” lamarensis Berry Reference Berry1912, is reported from the Woodbine and Tuscaloosa formations of mid-Cretaceous age. Given our few and incomplete specimens, assignment to “Ziziphus” fibrillosus should be considered provisional.

Morphotype CM18

Figure 4.8–4.10

Discussion

Placing this morphotype in a previously described taxon is complicated by the absence of preserved venation. Berry (Reference Berry1911a) proposed the name “Cinnamomum” newberryi for leaves previously described by Newberry (Reference Newberry1895) as Cinnamomum intermedium from the Early Cretaceous Amboy Clays. Specimens illustrated by Newberry (Reference Newberry1895) included actinodromous (p. 89, pl. XXIX, figs. 1–5) and acrodromous (p. 89, pl. XXIX, figs. 6, 10) primary venation. All specimens had acute basal and apical angles (Newberry, Reference Newberry1895). Berry (Reference Berry1911b) illustrated a similar specimen from the Raritan Formation that had an acute base, three actinodromous primary veins, and clearly developed agrophic veins (p. 150, pl. XVI, fig. 3). In 1925, Berry further expanded the concept of “Cinnamomum” newberryi by erecting several subspecies. Among these were specimens with obtuse base angles, acute apical angles and basally acrodromous primary venation (e.g., Berry, Reference Berry1925, p. 77, pl. XVI, fig. 6, “Cinnamomum” newberryi minimum). Although C. newberryi minimum is smaller in size (3.5 cm in length, 1.2 cm in width) than the average specimens of CM18 (approx. 6 cm in length, 3 cm in width, but note PB2984, Fig. 4.9, which matches Berry’s Reference Berry1925, p. 77 measurements), conforms to the description and illustration in most other respects. Uncertainty arises because the details of agrophic veins, secondaries and tertiary veins are not preserved in CM18.

“Cinnamomum” newberryi is described from the Cenomanian age Raritan Formation of New Jersey (Berry, Reference Berry1911b) and subsequently from the Dakota Formation of Kansas (Albian to Cenomanian), the Magothy Formation of New Jersey, Delaware, Maryland and New York (Santonian), and the Late Cretaceous Black Creek Formation of South Carolina and the Tuscaloosa Formation of Alabama (Knowlton, Reference Knowlton1919). The small form, C. newberryi minimum Berry Reference Berry1925, that best matches CM18, was reported as a single specimen from the Ripley Formation (Campanian-Maastrichtian) from Tennessee (Berry, Reference Berry1925).

Morphotype CM08

Figure 5.1–5.2

Figure 5

Morphotype CM08 (Protophyllum sp.): (1) PB2558, (2) PB2990. Morphotype CM09 (Leguminosites shirleyensis): (3) PB2995. Morphotype CM10 (Liriodendrites sp.): (4) PB2991, (5) PB2073, (6) arrow indicates location of spine that extends past the leaf margin in the center of the sinus, (7) PB2992. Morphotype CM11: (8) PB2986, (9) PB2987, (10) PB2985.

Discussion

The particular combination of characters that typify CM08 coupled with the lack of complete specimens makes assignment to previously described morphospecies difficult. CM08 most nearly fits Protophyllum Lesquereux Reference Lesquereux1874 described from the Dakota Formation (Cenomanian) of Kansas. Lesquereux (Reference Lesquereux1874) segregated this genus from Credneria because the leaves of Protophyllum lacked the horizontal secondary veins at the base of the leaf that characterize Credneria (Tschan et al., Reference Tschan, Denk and Von Balthazar2008; Zenker, Reference Zenker1833). Lesquereux (Reference Lesquereux1874) further distinguished Protophyllum from Pterospermites because the latter has peltate petiole attachment while the former has subpeltate or marginal attachment. Protophyllum is variable in many of the features characteristic of CM08. For example, P. rugosum (Lesquereux Reference Lesquereux1874, p. 105, pl. XVII, figs. 1–2) has a cordate base and subpeltate petiole attachment while P. haydenii (Lesquereux, Reference Lesquereux1874, p. 106, pl. XVII, fig. 3) has a concavo-convex base shape and marginal petiole attachment. All described species of Protophyllum have craspedodromous secondary venation in which veins terminate on the leaf margin and compound agrophic veins. Margins vary from untoothed and smooth to sinuous. When tertiary vein fabric is described and illustrated, it is mixed percurrent with sinuous course (Lesquereux, Reference Lesquereux1874). CM08 most nearly resembles Protophyllum quadratum (Lesquereux Reference Lesquereux1874, p. 104, pl. XIX, fig. 1), but CM08 differs from P. quadratum in appearing to have a marginal, rather than subpeltate petiole attachment. However, some specimens of CM08 have the deeply undulate margin that characterizes P. quadratum (Lesquereux, Reference Lesquereux1874). The difference in petiole attachment precludes assignment of these specimens to P. quadratum, however the variation observed in CM08 fits well within the range described for members of the genus from the Dakota Formation.

Morphotype CM09

Figure 5.3

Discussion

Bowerbank (Reference Bowerbank1840) erected the genus Leguminosites for seeds from the London Clay (Early Eocene) that resembled those of the “Leguminosae” in the modern Fabaceae. However, since Bowerbank (Reference Bowerbank1840) could not confidently ascribe the fossil specimens to any living genus, he chose to create a name for fossils that he felt confident were members of the legume clade. Heer (Reference Heer1874) first applied Leguminosites to small, rounded leaflets from the Green River Formation (Eocene) and Lesquereux (Reference Lesquereux1891 [1892]) applied the name to specimens from the Dakota Formation (Cenomanian). Lesquereux (Reference Lesquereux1891 [1892]) also noted “Colutea” primordialis Heer Reference Heer1882 from the Dakota Formation and used this name for legume-like leaflets with rounded bases. “Colutea” primordialis was also illustrated from the Raritan, Magothy (Berry, Reference Berry1911b) and the Woodbine (Berry, Reference Berry1922) formations of mid-Cretaceous age. Our single specimen is identical to Leguminosites shirleyensis Berry Reference Berry1919 from the Tuscaloosa Formation of Alabama, which Berry deemed identical to Diospyros rotundifolia Lesquereux Reference Lesquereux1891 from the Dakota Formation. Both Lesquereux (Reference Lesquereux1891 [1892]) and Berry (Reference Berry1919) allied such small, rounded leaflets with the modern Fabaceae.

Morphotype CM10

Figure 5.4–5.7

Discussion

A variety of generic names have been applied to bilobed fossil leaves that resemble modern Liriodendron in shape. Liriophyllum Lesquereux Reference Lesquereux1878 was restricted to leaves in which the midvein dichotomizes with one strand entering each of the two lobes (Dilcher and Crane, Reference Dilcher and Crane1984). Liriodendropsis Newberry Reference Newberry1895 had secondary veins that were regularly and closely spaced, while “Liriodendron” has been applied to some fossil leaves with more than two lobes (Newberry, Reference Newberry1895). From the Cenomanian floras of the Dakota Formation, many species of bilobed “Liriodendron” have been described, including L. primaevum Newberry Reference Newberry1895, L. semialatum Lesquereux Reference Lesquereux1891, L. giganteum Lesquereux Reference Lesquereux1891, and L. intermedium Lesquereux Reference Lesquereux1891. L. intermedium (Lesquereux, Reference Lesquereux1891 [1892], p. 207, pl. XXV, fig. 5) differs from CM10 in that it has multiple oblong, lanceolate lobes rather than two broad lobes. CM10 differs from L. giganteum (Lesquereux, Reference Lesquereux1891 [1892], p. 206, pl. XXV, fig. 1, pl. XXVI, fig. 5, pl. XXVII, fig. 1) in that the sinus of L. giganteum is broader and the lobes more oblique than observed in the Soap Wash specimens. Similarly, L. semialatum (Lesquereux, Reference Lesquereux1891 [1892], p. 204–205, pl. XXV, figs. 2–4, pl. XXIX, fig. 3) is not bilobed like those of the Soap Wash flora and the sinus of L. primaevum (Lesquereux, Reference Lesquereux1891 [1892], p. 203–204, pl. XXIV, fig. 4, pl. XXVI, figs. 1–4) is more shallow than those observed from Soap Wash. Newberry (Reference Newberry1895) described two “Liriodendron” species from the Amboy Clays of New Jersey. L. quercifolium Newberry Reference Newberry1887 (Newberry, Reference Newberry1895, p. 81, pl. LI, figs. 1–6) and L. oblongifolium Newberry Reference Newberry1887 (Newberry, Reference Newberry1895, p. 81–82, pl. LII, figs. 1–5) differ from CM10 in that the New Jersey forms have more than two lobes with acute, pointed apices and shallow sinuses. Liriodendropsis simplex Newberry Reference Newberry1895 (p. 83, pl. LIII, figs. 1–4, 7) has shallower lobes—some that approximated an emarginated apex—than CM10, an acute base and distinctly different venation. L. angustifolia Newberry Reference Newberry1895 (p. 84, pl. LIII, fig. 8) also has an elongate lamina, acute base and emarginated apex with closely spaced secondaries. Therefore, none of these previously defined species concepts can be applied to CM10.

Johnson (Reference Johnson1996) distinguished Liriodendrites for leaves from the upper Maastrichtian with a bilobed shape and deep sinus similar to those of living Liriodendron. Liriodendrites bradacii Johnson Reference Johnson1996 from the Hell Creek Formation of North Dakota has a wide range of morphological variation and compares favorably with many features of CM10. L. bradacii (Johnson, Reference Johnson1996, p. 3–4, fig. 1–13) is bilobed, has a deep sinus, rounded apices of lobes, and brochidodromous and eucamptodromous secondary venation. Among those specimens illustrated by Johnson (Reference Johnson1996), base shape ranges from convex (Johnson, Reference Johnson1996, fig. 2) to truncate (Johnson, Reference Johnson1996, figs. 5–7) to cordate (Johnson, Reference Johnson1996, figs. 1, 4). The base of CM10 is uniformly acute. Due to morphological differences and the significant age difference between the Soap Wash flora and North Dakota specimens used to define L. bradacii, we hesitate to assign the Soap Wash morphotype to this species. Moreover, CM10 of the Soap Wash retains a spine that extends past the leaf margin in the center of the sinus (Fig. 5.6) where the primary vein of L. bradacii is terminates at the margin. The sinus spine is distinct, present in all specimens of CM10 and differs from all described species of Liriodendron, both living and fossil, all species of Liriophyllum and Liriodendropsis, and from Liriodendrites bradacii.

CM10 preserves evidence of insect damage.

Morphotype CM11

Figure 5.8–5.10

Discussion

We did not find a convincing match to this morphotype among the literature that we surveyed. The simple lamina with untoothed margin and weakly developed brochidodromous venation suggests affinities with the subclass Magnoliidae as defined by Hickey and Wolfe (Reference Hickey and Wolfe1975). Their pre-phylogenetic classification would include the modern magnoliid clade (Cantino et al., Reference Cantino, Doyle, Graham, Judd, Olmstead, Soltis, Soltis and Donoghue2007, Angiosperm Phylogenetics Group, 2009) and basal clades such as Amborella and Austrobaileyales because Hickey and Wolfe (Reference Hickey and Wolfe1975) specifically included Chloranthus, Illicium and Austrobaileyaceae within the Magnoliidae in their classification of leaf morphology. CM11 resembles a number of species of “Magnolia” that have been described from the Upper Cretaceous sediments of the Atlantic Coastal Plain and Gulf Coast. In particular, “Magnolia” capellinii Heer 1866 (Capellini and Heer, Reference Capellini and Heer1866, p. 21, pl. III, figs. 5–6), which was originally described from the Dakota Formation of Nebraska, shares broad similarity with CM11 but differs in some details of shape (e.g., acute rather than rounded base) and venation, which precludes assignment to this taxon. Given that the genus “Magnolia” in the Cretaceous fossil record has become a wastebasket for entire-margin leaves with brochidodromous venation and a generally “primitive” (sensu Hickey and Wolfe, Reference Hickey and Wolfe1975) venation, provisional assignment of CM11 to this genus as a new form does not clarify its identity or relationships. Therefore, we do not assign this morphotype to a previously described taxon.

This morphotype preserves evidence of extensive insect damage.

Morphotype CM12

Figure 6.1

Figure 6

Morphotype CM12: (1) PB3037. Morphotype CM13 (cf. Kadsura sp.): (2) PB3006, (3) PB19866, (4) Kadsura oblongifolia Merr. (extant) from Hainan, China UCH612215 (http://clearedleavesdb.org/), (5) PB3004. Morphotype CM14 (Rogersia parlatorii): (6) PB2054, (7) PB2053, (8) vein detail of this morphotype.

Discussion

Berry (Reference Berry1919) described a species of fungus, Sphaerites alabamensis Berry Reference Berry1919 (p. 49, pl. V, fig. 1) infecting a leaf from the Tuscaloosa Formation of Late Cretaceous age from Alabama. The figured, but unidentified, leaf is identical to the single specimen of CM12 found at Soap Wash. Unfortunately, Berry (Reference Berry1919) did not identify the figured leaf, saying only that the fungus commonly infects specimens of Platanus shirleyensis Berry 1914 (Berry, Reference Berry1914a, p. 5, fig. 4), which does not correspond to CM12 in shape, margin or venation. In addition to shape, margin and venation, the specimen figured in Berry’s (Reference Berry1919) description of the fungal infection shows a pattern of damage that is identical to that observed in the Soap Wash specimen (Fig. 6.1). Thus, we are confident that CM12 corresponds to the form reported by Berry (Reference Berry1919) from Alabama, but our review of the literature and museum records could not identify the leaf taxon into which Berry (Reference Berry1919) intended to place the illustrated leaf fossil. Therefore, we leave the leaf morphotype CM12 unidentified. The damage observed to this specimen is the topic of ongoing research.

Morphotype CM13

Figure 6.2–6.5

Discussion

These specimens differ from the other specimens of the Soap Wash in their lamina and base shape, and venation. Like CM11, CM13 has a simple lamina with an untoothed margin and weakly developed brochidodromous venation that suggests affinities with the subclass Magnoliidae as defined by Hickey and Wolfe (Reference Hickey and Wolfe1975), which includes the modern concept of the magnoliid clade plus the basal angiosperm clade, Austrobaileyales (Cantino et al., Reference Cantino, Doyle, Graham, Judd, Olmstead, Soltis, Soltis and Donoghue2007, Angiosperm Phylogenetics Group, 2009). CM13 is provisionally assigned to the living genus Kadsura based on the acute base, thick, apparently fleshy petiole, and similarity with the secondary, tertiary and quaternary vein fabrics (see Fig. 6.4 for comparison to the venation of modern Kadsura oblongifolia Merr.), which is best observed in the partial specimen PB19866 (Fig. 6.3). Although Kadsura oblongifolia illustrated in Figure 6.4 has glandular teeth, several species of living Kadsura (e.g., K. acsmithii, K. celebica and K. borneensis) have untoothed margins and others (e.g., K. philippinensis) have both toothed and untoothed margins. Therefore, the absence of marginal teeth does not exclude CM13 from the genus.

Morphotype CM14

Figures 6.6–6.8

Discussion

Fontaine (Reference Fontaine1889) erected the fossil genus Rogersia for leaves from the Potomac Group that were long and narrow with untoothed margins. Rogersia angustifolia Fontaine Reference Fontaine1889, with which CM14 is aligned, is the most common constituent of the angiosperm flora from some localities of the Patuxent Formation of Virginia, particularly at Fredericksburg (Fontaine, Reference Fontaine1889, p. 288, pl. CXLIII, fig. 2; pl. CXLIX, figs. 4, 8; pl. CL, figs. 2–7). The fossil form also appears in the Cenomanian-age Dakota Formation of Nebraska and Kansas (Wang and Dilcher, Reference Wang and Dilcher2006). Wang and Dilcher (Reference Wang and Dilcher2006) synonymized Rogersia angustifolia Fontaine Reference Fontaine1889 with “Andromeda” parlatorii Heer 1866 (Capellini and Heer, Reference Capellini and Heer1866) to produce the new combination Rogersia parlatorii, and placed it in the Laurales clade. We apply Rogersia parlatorii to CM14 because it matches Wang and Dilcher’s (Reference Wang and Dilcher2006) description and illustrations.

Morphotype CM15

Figure 7.1–7.4

Figure 7

Morphotype CM15 (Dewalquea smithi): (1) PB2997, (2) PB3014, (3) PB3015, (4) tooth detail of this morphotype.

Discussion

This morphotype appears identical to Dewalquea smithi Berry 1910 (Berry, Reference Berry1910a, p. 36, fig. 1). This species is abundant in the Late Cretaceous of the Middendorf arkose member of the Black Creek Formation in South Carolina (Berry, Reference Berry1914b) and the Late Cretaceous Tuscaloosa Formation of western Alabama (Berry, Reference Berry1919). Other members of the genus are described from Dakota Formation sediments (Lesquereux, Reference Lesquereux1891 [1892]). The divergence of the petiole into three branches, the lateral two of which bifurcate into two leaflets and the central branch leading to a single leaflet, is consistent with the original descriptions of D. smithi (Berry, Reference Berry1910a, p. 36, fig. 1). Additionally, the presence of serrate margins on the apical three-quarters of each lamina and craspedodromous secondary venation is consistent with later descriptions and illustrations (Berry, Reference Berry1914b, Reference Berry1919).

Dewalquea smithi is distinct from another common compound leaf in Early and mid-Cretaceous floras, Sapindopsis (e.g., CM02) in being palmately compound, where Sapindopsis is pinnately compound with leaflets in opposite to subopposite arrangement (Fontaine, Reference Fontaine1889, p. 296). In specimens where only the three top leaflets of Sapindopsis are preserved, the form may appear palmately compound. However, specimen PB2997 clearly shows the petiole arrangement of CM15 (Fig. 7.1) in which the lateral petioles bifurcate. This arrangement is a good match for the original illustration of the species (Berry, Reference Berry1910a, fig. 1) and distinct from described species of Sapindopsis that we identified in our review of the literature.

Morphotype CM16

Figure 9.1–9.5

Figure 9

Morphotype CM16 (Saliciphyllum sp.): (1) PB2082, (2) PB3023, (3) PB3025, (4) tooth detail of this morphotype, (5) PB3036. Morphotype CM17 (“Celastrophyllum” shirleyensis): (6) PB2065, (7) PB2066.

Discussion

This linear to ovate lanceolate morphotype was one of two dominant forms in the Soap Wash flora, with 38 well-preserved specimens. CM16 is identical to morphotype AS28 reported by Peppe and colleagues (Reference Peppe, Hickey, Miller and Green2008) from the Albian-age Aspen Shale of Wyoming. AS28 was not assigned to any previously described taxon (Peppe et al., Reference Peppe, Hickey, Miller and Green2008) and our independent survey of the literature did not yield any convincing matches with described species. In leaf shape and venation, CM16 resembled any of several described species of fossil “Salix” from the mid-Cretaceous, for example “Salix” flexuosa Newberry Reference Newberry1868. Newberry’s (Reference Newberry1868) written description matches CM16 precisely except that margin type is not mentioned; Newberry (Reference Newberry1868) provided no illustrations. Illustrations associated with subsequent reports of “Salix” flexuosa show an untoothed margin (Newberry, Reference Newberry1878, p. 21, pl. I, fig. 4). Subsequent reports follow this diagnosis (e.g., Berry, Reference Berry1914b, p. 32, pl. VII, figs. 14–16 from the Atlantic coastal plain). Other species, such as “Salix” lesquereuxii Berry Reference Berry1909 (p. 252) and illustrated subsequently (Berry, Reference Berry1914b, p. 33, pl. VII, figs. 11–13), also lack marginal teeth. Thus, we cannot place CM16 in these previously described species.

The teeth of CM16 strongly resemble the marginal teeth of some living members of Salix (e.g., Salix fragilis L. and S. alba L., which are native to Europe and western Asia, and S. petiolaris J.E. Smith and S. sericea Marsh from North America). These modern species are also very similar to CM16 in shape, L:W ratio, base, apex and some aspects of primary and secondary venation, suggesting either evolutionary affinities with the willow clade or extensive convergence. CM16 also lacks stipules, which further allies it with a subsection of extant European and North American willows. However, modern Salix species generally share percurrent tertiary vein fabric; tertiary venation is not preserved in CM16. Therefore, we do not assign this form to the modern genus.

Fontaine (Reference Fontaine1889) erected the genus Saliciphyllum to include abundant but poorly preserved willow-like forms from the Potomac Group. We have adopted this name for CM16 because of CM16’s similarities with some regional variants of modern willow. Fontaine’s (Reference Fontaine1889) original publication included a description of the venation of S. longifolium (p. 302, pl. CL, fig. 12) that indicates mixed to brochidodromous secondary venation. CM16 does not conform to this species. Margin form is also not mentioned in either the genus or species description (Fontaine, Reference Fontaine1889), nor was it mentioned in subsequent use of the genus (Ward, Reference Ward1905).

Morphotype CM17

Figure 9.6–9.7

Discussion

Göppert (Reference Göppert1853) published the generic name Celastrophyllum without description or illustration and followed with a formal description (Göppert, Reference Göppert1854, p. 52). The genus was created for fossil forms that resembled modern Celastrus L. but were both ancient and did not conform to the modern concept in all details of higher-order venation. A variety of fossil species have been erected for forms from the Cretaceous and Paleogene. CM17 conforms closely to the single specimen of Celastrophyllum shirleyensis Berry Reference Berry1919 (p. 107, pl. XXVI, fig. 8) described and illustrated from the Tuscaloosa Formation of Alabama. CM17 shares similar shape and base, primary and secondary venation and the diagnostic tooth form. No other previously described species matches CM17 in both venation and tooth form.