RESEARCH ARTICLE A M E R I C A N J O U R N A L O F B O TA N Y
A mosaic Lauralean flower from the Early Cretaceous of Myanmar1 William L. Crepet2,6, Kevin C. Nixon3, David Grimaldi4, and Mark Riccio5
PREMISE OF THE STUDY: The floral history of early angiosperms is far from complete. The fossil discussed here has the potential to expand our knowledge of timing, reproductive biology, and paleobiogeography in early angiosperms. METHODS: Cutting-edge methodologies in CT scanning in conjunction with tomography software have opened new possibilities for discovering details in amber-preserved fossils that were inaccessible for meaningful study in the past. KEY RESULTS: The fossil is small and complex, cupulate, with numerous stamens and a suite of characters distributed in the modern families of Laurales. The most parsimonious placement of the fossil based on morphology is as a sister taxon of Atherospermataceae + Gomortega (Gomortegaceae). CONCLUSIONS: This fossil taxon, a Laurasian Lauralean from the mid-Cretaceous, is an important example of fossil Laurales with implications for biogeography and timing in the radiation and extinction in this group. KEY WORDS Atherospermataceae; Cretaceous; diversification; Gomortegaceae; insect pollination; Lauraceae; Laurales; Monimiaceae
A great deal of progress has been made in understanding angiosperm history since the rapid Cretaceous radiation of angiosperms inspired Darwin to refer to this phenomenon as the “abominable mystery” (e.g., Friedman, 2009). The coincidence in the development of algorithms generating objective phylogenies (e.g., TNT, Goloboff et al., 2008), with the emergence of techniques generating inexpensive DNA sequence data, has yielded a better understanding of relationships among extant angiosperms and has also provided important context for evaluating the affinities and evolutionary significance of fossils (e.g., Crepet et al., 2004). At the same time, there have been important advances in studies of fossil angiosperm leaves and pollen that have greatly improved our understanding of angiosperm radiation and paleobiogeography while lending new data useful to the understanding of past climates (Doyle, 1969; Walker and Doyle, 1975; Doyle and Hickey, 1976; 1
Manuscript received 1 September 2015; revision accepted 5 January 2016. 412 Mann Library, Plant Biology Section, School of Integrative Plant Science, Cornell University, Ithaca, New York 14853 USA; 3 L. H. Bailey Hortorium, Plant Biology, School of Integrative Plant Science, Cornell University, Ithaca, New York 14853 USA; 4 Division of Invertebrate Biology, American Museum of Natural History, New York, New York 10024-5192 USA; 5 Biotechnology Center, Cornell University, Ithaca, New York 14853 USA 6 Author for correspondence (e-mail: [email protected]) doi:10.3732/ajb.1500393 2
Wolfe, 1995; Crisp et al., 2009). Studies of fossil flowers, which came later than leaf and pollen studies (e.g., Crepet et al., 1974; Tiffney, 1977 ; Nixon and Crepet, 1989 , 1993 ; Friis et al., 2011 ), have provided remarkable insights into early angiosperms and their reproductive strategies (especially in conjunction with the burgeoning fossil record of insects; Grimaldi and Engel, 2005; Poinar and Danforth, 2006). Flowers also have the potential for generally providing more precise indications of minimum timing in the history of angiosperm taxa due to their diagnostic value. Hence, variously fossilized flowers have added important dimensions to the fossil record of angiosperms that have not usually been available from the records of other preserved angiosperm organs. Here we describe a mid-Cretaceous flower entombed in amber with a unique combination of lauralean features not found in any modern groups.
MATERIALS AND METHODS This study reports a fossil flower from Burmese amber of earliest Cenomanian to latest Albian age (ca. 98 million years ago [Ma]) that was acquired commercially. The amber-containing sediments are well known, and their geology and age have been exhaustively studied (Cruickshank and Ko, 2003; Shi et al., 2012), and a number of significant fossils have been reported from these sediments (e.g.,
290 • A M E R I C A N J O U R N A L O F B OTA N Y 103(2): 290–297, 2016; http://www.amjbot.org/ © 2016 Botanical Society of America
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
Grimaldi et al., 2002; Poinar and Danforth, 2006). The specimen studied herein is a tiny flower (2.1 mm diameter) that is complex in structure with numerous floral parts. The preserving medium (amber) is relatively dark and variably somewhat cloudy. Given the difficulties imposed by the nature of the preserving amber and the small size and structural complexity of this flower, it was difficult to determine floral structural details with light or with Zeiss 710 confocal microscopy. It was therefore necessary to resort to Micro-CT scan analysis to determine its detailed structural features. Micro-CT scans were performed at the Cornell University Institute of Biotechnology Imaging Facility using a Zeiss Xradia 520 Versa X-ray instrument. For each data set, 3600 X-ray projections were digitized at 0.1° intervals over 360° using 60 kV and 5 W with a 5000 ms exposure time. Reconstructions used a modified Feldkamp filtered back projection algorithm yielding isotropic x-y-z voxels. The sample was scanned at various resolutions, yielding 500 nm, 750 nm, and 1.8 μm voxels.
RESULTS Description—Flower small (ca. 2.1 mm diameter), bisexual. Floral
cup hemispheric, bearing imbricate ovate tepals, ca. 12, near the rim (Fig. 1A-C). Stamens ca. 12, arranged in an apparent tight spiral with the fertile stamens having longer filaments than the inner staminodes (Figs. 1D, 1E, 3A). Stamens are flattened and bear elongate unicellular hairs on adaxial surfaces. Similar hairs line the entire floral cup (Appendices S1, S5, see online Supplemental Data). The fertile stamens bear two paired basal appendages terminating in sagittate heads (one of a pair visible in Fig. 1F). Stamens are bilocular (presumed bisporangiate) introrse, and dehiscent by apical valvate flaps (Figs. 1G, 2A, 2B; online Appendix S2). Anther connectives extend slightly beyond the thecae, are flattened and terminally pointed (Figs. 1D, 1E, 2A). Several staminodes occur between the stamens and gynoecium (Fig. 2C), and these are filamentous with sagittate heads but lack lateral appendages (Fig. 2C, D). The exact number of these staminodes is uncertain, but is in the range of 10–12. Pollen is abundant in the amber matrix, concentrated in the vicinity of the anther locules and valvate flaps and averages 15 μm in diameter. Pollen morphology is difficult to interpret in nano-CT scan sections due to limits of resolution, apparent folding of the grains, and uneven wall preservation. There are four carpels (illustrated in cross section in Fig. 2G; longitudinal in Fig. 2H), each with an elongate, tapering style terminating in a slightly expanded stigma (Fig. 2E, F; online Appendix S3). Carpels are free from the floral cup, but basally surrounded by the cupule wall, and each includes a single ovule/seed with a developed embryo (Fig. 2G, H). Concentric, but not spirally arranged layers that represent the inner and outer epidermal boundaries of the carpel walls and seed coat surround the embryos. Taxonomic treatment—Jamesrosea gen. nov.
Etymology: Jamesrosea is named in honor of Professor James L. Reveal and his wife Rose. Type: Jamesrosea burmensis, new species Diagnosis: Flower bisexual, with hemispheric floral cup bearing imbricate tepals near the rim. Stamens ca. 12, each with two paired basal appendages. Anthers introrse, bilocular, with valvate dehiscence by apical flaps. Staminodes present between stamens and carpels. Gynoecium apocarpous, carpels superior or half-inferior, four, with elongate filiform styles. Ovules/seeds apparently one per carpel.
• 291
Jamesrosea burmensis sp. nov. Diagnosis: as for the genus Jamesrosea. Type: holotype: AMNH JZ Bu-124 Phylogenetic analysis—To place the fossil phylogenetically, a mor-
phological matrix was scored based on a significant modification of the matrix for Laurales published by Renner (1999). Focusing the analysis on Laurales is justified because of the clear synapomorphies and unique combination of lauralean features exhibited by the fossil (see Discussion). The original matrix had 15 characters, and the fossil could be unambiguously scored for nine of those characters. The new morphological matrix has 14 characters for which the fossil could be unambiguously scored (online Appendix S4). It was analyzed under a parsimony criterion using TNT (Goloboff et al., 2008). We also reduced the number of extant taxa in the matrix to 14 homogeneous terminals that are representative of all lauralean families. Thus, the final matrix had 15 taxa and 14 morphological characters. All analyses were constrained according to the molecular results for Laurales published by Renner (1999) and found in subsequent analyses. An alternative topology that differs in placing Hernandiaceae as sister to Lauraceae (Angiosperm Phylogeny Group, 2009) based on morphology was also tested as a constraint, with identical results to those reported below. The fossil was not constrained within this framework, and thus its placement was determined solely by the morphological characters, while retaining the scaffold tree determined by the molecular analyses. Phylogenetic results—The analysis of the Laurales morphological
matrix constrained by the molecular scaffold tree resulted in one most parsimonious tree of 32 steps placing Jamesrosea as the sister to Atherospermataceae + Gomortegaceae (Fig. 3). The fossil is similar to modern Gomortega in perianth features, stamen morphology including paired basal glands, anther dehiscence and number of locules, orientation of dehiscence in the fertile stamens (introrse), and the presence of staminodes between the fertile stamens and gynoecium (inconsistently seen in Gomortega). The fossil differs from Gomortega in having four free carpels as opposed to a syncarpous, inferior ovary in the latter. The fossil has introrse stamen dehiscence, while Atherospermataceae have extrorse stamen dehiscence and unisexual flowers. Moving outward in the tree, Jamesrosea differs from Siparunaceae in having valvate anther dehiscence as opposed to a variant (Endress and Hufford, 1989; longitudinal slits/valves) in the latter. Fossil pollen cannot be evaluated at adequate resolution for determination of pollen characters. Thus, the fossil is best considered an early offshoot of the AtherospermataceaeGomortegaceae lineage that predates the occurrence of syncarpy, a feature unique to Gomortega within Laurales. It is important to note that using the morphology alone, unconstrained by the molecular scaffold, results in some trees with Jamesrosea nested within the Monimiaceae-Lauraceae clade. However, the position presented here remains the best estimate based on all evidence, and the striking similarity of the fossil to Gomortega except for the apocarpous, superior gynoecium is also compelling.
DISCUSSION As indicated, assessment of the fossil’s structure using standard light microscopy or confocal microscopy was difficult due to limitations on viewing interior details of the flower in the darkened amber
292
•
A M E R I C A N J O U R N A L O F B OTA N Y
FIGURE 1 (A) Light photograph of Jamesrosea showing overall fossil. Bar = 1 mm. (B) Higher magnification photograph of the fossil illustrating the cloudy nature of the amber matrix. Bar = 500 μm. (C) Screenshot of three-dimensional (3D) movie of Jamesrosea generated by CT scan/tomography (online Appendix S1). Note stamen connectives (c) and recurved dehiscence valves (v). Bar = 120 μm. (D) Screenshot of another 3D movie generated by CT scan/tomography of a top view of the fossil embedded in amber, illustrating the distal portions of the outermost stamens, their flattened
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
and because the cup-like hypanthium masks many of the inner floral features. High-resolution CT scanning, on the other hand, provided information about floral structure that was not available with light microscopy alone (online Appendices S6–S8). Most importantly, the CT scans revealed an important set of characteristics not visible with light including an apocarpous gynoecium of four carpels with long, slender styles that were otherwise obscured by the surrounding stamens (Figs. 1E, 2C, 2D, 2F; Appendices S1, S3, S8). The scans also show paired appendages at the base of the filaments (Fig. 2F; Appendices S8, S9, S10), and bilocular anthers dehiscent by apical flaps (Figs. 1G, 2A, 2B, 2D; Appendices S1, S3, S8). These three characters were vital in determining the appropriate placement of Jamesrosea as a sister taxon of Atherospermataceae-Gomortegaceae based on the phylogenetic analysis of morphology. The CT-scans also revealed a single ovule/seed in each carpel with concentric layers (Fig. 2G, H). It is difficult to interpret these structures. Within Laurales, spirally enfolded cotyledons are found only within Calycanthaceae sensu stricto (Calycanthus and Sinocalycanthus), but in section, these appear to be clearly spiral with many layers (K. C. Nixon, W. L. Crepet, unpublished data, examination of live material), unlike our fossil. As mentioned, in Jamesrosea the concentric pattern may simply reflect different wall boundaries within the seed and embryo and be difficult to determine given the limitations of the scan. Previously described fossil Cretaceous flowers include a number of unequivocal calycanthoid (Jerseyanthus, Crepet et al., 2005), a possible calycanthoid, Virginianthus (Friis et al., 1994), and core lauralean flowers (Friis et al., 2011), e.g., the mid-Albian Potomacanthus (Von Balthazar et al., 2007), mid-Cretaceous Mauldinia angustifolia (Viehofen et al., 2008), and Turonian Perseanthus (Herendeen et al., 1994). These fossils show affinity to the crown group of Lauraceae sensu stricto, although Potomacanthus and Mauldinia have not been placed in a formal phylogenetic analysis. In combination, the assemblage of calycanthaceous and lauraceous fossil flowers from the Early and mid-Cretaceous indicates that differentiation of the Laurales clade was already underway in the Albian, and some modern Lauraceae had already differentiated by the Turonian (e.g., Herendeen et al., 1994). None of these previously described lauralean flowers show any significant similarity to Jamesrosea, and all belong either to the calycanthoid or LauraceaeMonimiaceae clade according to published analyses. With the exception of the slightly older fossil Potomacanthus (Von Balthazar et al., 2007), Jamesrosea is the only known Early Cretaceous fossil flower identifiable as Laurales. Jamesrosea also has the most convincing suite of lauralean features, including paired basal appendages associated with stamen bases (ambiguous in Potomacanthus) and valvate apical anther dehiscence. Within Laurales, Jamesrosea is clearly the oldest fossil record for the Siparunaceae-Gomortega-Atherospermataceae (SGA) clade (see Knight and Wilf [2013] for an excellent review of the fossil record of Atherospermataceae). The position of Jamesrosea within the SGA crown group places a minimum age on the Gomortegaceae-
• 293
Atherospermataceae clade, the SGA and its sister group, the Hernandiaceae-Monimiaceae-Lauraceae clade, and the ancestor of both, of ca. 98 Ma. The Atherospermataceae-Gomortegaceae clade is widely distributed in Patagonia (three genera), New Zealand (one genus), Australia (four genera), and New Guinea and New Caledonia (two genera; Knight and Wilf, 2013). There are no modern species of the crown clade (Gomortegaceae/Atherospermataceae) to which Jamesrosea is a sister taxon in mainland Asia. Thus, Jamesrosea, from the Burmese amber, represents a considerable paleo-range extension for the clade in addition to being the oldest record for the SGA clade, suggesting a more widespread distribution of the SGA clade in the Early Cretaceous. Pollination of Early Cretaceous angiosperms is interesting from the perspective of both plants and insects, and understanding the possible pollination syndrome of Jamesrosea has implications for understanding the evolution of pollination within the Laurales. Floral morphology often can predict potential pollinators, although with limited precision. Within extant Laurales, a wide range of pollination syndromes and pollinators have been reported (e.g., Grant, 1950; Feil, 1992; Momose et al., 1998; Kato and Kawakita, 2004; Peñalver et al., 2012). Documented pollinators include flies, thrips, beetles, lepidopterans, and hymenopterans. Different pollinators within the order are associated both with floral syndromes, and with particular clades reflected at the family or generic level. Because of the phylogenetic position of Jamesrosea, the modern pollinators of Siprunaceae, Atherospermataceae, and Gomortegaceae are of special interest in understanding possible pollination and pollinators in Jamesrosea. Additionally, because of similarities in floral morphology, the other members of Laurales, most importantly Monimiaceae, are also of interest in terms of pollination. Within the Laurales, Calycanthaceae is considered to be the outgroup to the remaining families based on all molecular analyses. In Calycanthus, pollination is primarily by beetles, and a single food body on the stamen connective acts as a floral reward in modern species (Daumann, 1930; Grant, 1950; Rickson, 1979). In the calycanthaceous Late Cretaceous fossil Jerseyanthus, a more complex anther connective occurs with several food bodies in a candelabralike structure (Crepet et al., 2005). In Idiospermum, pollination has been reported to be by small beetles or thrips (Worboys and Jackes, 2005). Modern Siparunaceae species are typically pollinated by gall midges (family Cecidomyiidae) (Renner et al., 1997). Gall midge pollination is associated with unisexual, relatively closed flowers that provide a host for larvae. Male flowers harbor large numbers of galls, partly because the flowers are more open and allow oviposition more easily. Female flowers are more tightly closed and typically do not harbor galls. Attempts to oviposit on female flowers results in pollination by pollen acquired from male flowers. According to Feil (1992, p.171), “The evolution of unisexual flowers in Siparuna can be explained as a result of the differential predation by larvae: unimportant in male flowers, destructive if occurring in
→
connective extensions (c), and pairs of recurved valves (arrows). Bar = 500 μm. (E) Another top view of the 3D model illustrated in (D) at an optical level revealing the stigmas in the center of the flower (arrow) and including views of some of the tepals. Bar = 500 μm. (F) Screen shot from CT-scangenerated 3D model illustrating one of the appendages attached to the base of a stamen filament. Bar = 300 μm. (G) Screen shot from CT scan illustrating longitudinal view through the flower showing stamens with two thecae (t) and recurved valves (arrowhead) and elongate hairs in floral cup. Bar = 167 μm.
294
•
A M E R I C A N J O U R N A L O F B OTA N Y
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
• 295
On the basis of these comparisons and overall morphology and size of the flower Jamesrosea, pollination was most likely by small flies, thrips, small hymenoptera, or small beetles as in some modern Monimiaceae. The relatively open, bisexual flowers and presence of paired nectar glands eliminates gall midges as likely pollinators, and no evidence of flower galls is present in the fossil. Also, the only gall midge taxa preserved in Cretaceous ambers (including Burmese amber) are of basal subfamilies that do not form galls (e.g., Lestremiinae, Porricondylinae); the gall-forming subfamily Cecidomyiinae appears in the Tertiary (D. Grimaldi, unpublished data). Thus, it is highly unlikely that Jamesrosea was pollinated by cecidomyiid midges. Thrips pollination is known in a variety of FIGURE 3 Single most-parsimonious cladogram based on parsimony analysis of the morphological “basal” angiosperm groups (e.g., Annonamatrix and constrained to the consensus phylogeny for Laurales (Renner, 1999). ceae: Momose et al., 1998), as well as cycads, and the earliest direct evidence of insect pollination may be a thrips female flowers.” Although Jamesrosea is phylogenetically close to preserved with Cycadopites-like pollen on its surface in Early CretaSiparunaceae, it has bisexual flowers, and it is unlikely that gall ceous amber (Peñalver et al., 2012). Thus, the possibility of thrips midges would be effective pollinators. pollination in Jamesrosea and other small flowers at this early phase Gomortega pollination has been minimally studied, and based of angiosperm divergence is reasonable. Further investigation of on very limited observations, may be pollinated by syrphid flies modern Gomortegaceae pollination with methods that could detect (Lander et al., 2009). Because of the limited data in this study, and thrips pollination is warranted. mixed pollen loads from other species on pollinators, it is premaRenner (2004) has considered the pattern of evolution in Lauture to exclude other pollinators as candidates for pollination in rales from a number of perspectives, focusing especially on disJamesrosea, and the primary difference between Jamesrosea and parities in relative species numbers among extant families Gormortega, the syncarpous, unistylous inferior ovary in the latter, (Calycanthaceae, 10 species; Siparunaceae, 70 spp.; Atherosperis likely important in presentation to pollinators. mataceae, 14 spp.; Gomortegaceae, 1 sp.; Lauraceae, 2500–3000 Modern Monimiaceae is noteworthy for having a high proporspp.; Hernandiaceae, 50 spp.; Monimiaceae, 195 spp.). Patterns of tion of thrips and gall midge pollination. Flowers pollinated by disparate clade sizes are common at almost any level (order, family thrips often, but not always, have floral rewards such as nectar in or genus) within angiosperms (for example, within Fagaceae, addition to pollen. Mollinedia has unisexual flowers pollinated by where monophyletic genera size ranges from 1–3 species to several thrips, without paired stamen glands (Renner, 2004), but thrips hundred). She observed (p. 444) that “The result that extant spepollination is not associated with unisexuality per se. However, cies numbers of Laurales families are significantly imbalanced, at some thrips flowers are also similar to those pollinated by small least against a Markov null model, clearly does not address differbeetles, and indeed in Monimiaceae both pollinators are seen ences in speciation and extinction rates through time.” The issue is within the same genus (Kato and Kawakita, 2004). most likely whether a Markov null model is a realistic expectation Within the Lauraceae/Hernandiaceae clade, flowers have a for diversification and/or extinction of clades within a complex single carpel and seed, most are bisexual, are usually more open landscape of climate, geography and biotic interactions. Nonethe(without a constricted orifice), and there has been a tendency to less, Renner speculated on possible environmental factors that reduce the number of stamens relative to Monimiaceae and offer might have affected speciation rates or taxon longevity in certain nectar rewards from the paired basal glands on the stamen filalauralean clades and suggested that a variable speciation–extinction ments and the glandular head of the staminodes (Rohwer, 2009). dynamic may explain differences in species diversity among differThis clade has a wide variety of pollinators, including mainly ent lauralean families (although it is difficult to imagine any other hymenopterans. →
FIGURE 2 (A) Screen shot from CT-scan movie (Appendix S1) illustrating the stamens with elongate filaments (f ) and open anther locules with recurved valves (arrowheads) as well as a sagittate stamen appendage (sa) and staminodes (st). Bar = 388 μm. (B) Top view in CT scan section showing three of the four styles in view (arrowhead) at this level and dehisced introrse locules of the stamens. Bar = 167 μm. (C) Transverse CT scan section through flower illustrating the basal portions of inner staminodes in section (arrowheads). Bar = 35 μm. (D) Screen shot of three-dimensional (3D) model movie in longitudinal aspect illustrating a style in the foreground (Sty) and four of the innermost staminodes (arrowheads), surrounding the carpels, and two microsporangia of a stamen in the background (OS). Bar = 35 μm. (E) Top view of 3D CT-scan-based model of the flower showing tepals and the four stigmas in the center (arrowhead). Bar = 140 μm. (F) Another transverse CT scan view of the flower illustrating staminodes surrounding all four styles (arrowhead). Bar = 28 μm. (G) CT scan section through the four carpels (arrowheads) illustrating layers connoting tissue boundaries. Bar = 110 μm. (H) Longitudinal CT scan through the center of the flower illustrating carpels, elongate styles, and tissue boundaries (arrowheads). Bar = 110 μm.
296
•
A M E R I C A N J O U R N A L O F B OTA N Y
explanation). One question in that context follows: Does the fossil record provide clues to modern day diversity differences among families in Laurales by revealing patterns of radiation and extinction with paleoenvironmental correlates that might reflect causative factors? While the fossil record of Laurales is good relative to the Cretaceous records of some other taxa, it is not complete enough to illuminate possible causes of the inequitable species distribution in that order. It has an inadequate pollen record—i.e., it is depauperate because, in Laurales, pollen typically is not preserved due to the (chemically and structurally), delicate nature of the exine (e.g., Herendeen et al., 1994), and pollen in some lauralean taxa could not reveal taxon level diversity because of the apparent uniformity of pollen in some clades. The leaf record, while relatively extensive in numbers of reports at least, may fail to reveal true past diversity because variation among modern leaves within the same individual or taxon or similarities among leaves in different taxa may have influenced interpretations of lauralean fossil leaves by obscuring the significance of variation in superficially similar leaves that might actually represent a number of different taxa, even at the generic level. For example, Herendeen et al. (1994, p. 29) pointed out that “Specialists in Lauraceae have noted that it is frequently difficult, if not impossible, to provide generic identifications for sterile or fruiting extant laurels and therefore they have questioned the generic identifications of fossil Lauraceae particularly sterile or fruiting material (Kostermans, 1957).” In contrast, the fossil flower record does provide interesting insights into a pattern of evolution of past lauralean diversity that is more or less congruent with phylogenetic analyses of modern lauralean taxa, and which may offer evidence of how at least one factor may be correlated with dramatically differential species diversity among lauralean families. As noted, the fossil record of the Laurales is sparse in the Early Cretaceous, making each precisely identified fossil potentially significant. The fossil taxon Jamesrosea suggests both a more diverse, widespread SGA clade in the past, spanning northern and southern hemisphere land masses in what is now both Asia and South America. The fossil record of lauralean flowers reveals an emergence of the crown groups of modern lauralean families during the Albian-Turonian transition. Known Early Cretaceous lauralean fossils do not fit into modern families (even though at least one generalized Cretaceous flower has been placed with Lauraceae [von Balthazar et al., 2007]). However, the latter fossil lacks structural features typically found in flowers of extant family crown groups, including features related to pollination syndromes such as stamen filament appendages characteristic of Lauraceae, while the Lower Cretaceous genus Virginianthus (Friis et al., 1994) may represent a higher-level calycanthoid taxon that is transitional between Calycanthaceae and the remaining Laurales (Crepet et al., 2005). By the Turonian, lauralean flower fossils can be placed within modern crown groups. Thus, Jerseyanthus (Crepet et al., 2004), while an extinct taxon represented by fossil flowers with rather dramatic differences in stamen connective morphology from those in taxa in modern Calycanthaceae, fits phylogenetically within that modern crown group. Other Turonian fossils such as Perseanthus fit neatly into modern Lauraceae based on floral characters alone. This progressive modernization of floral structure and the taxa they represent in Laurales implies a parallel change in pollination biology from flies, small beetles, and thrips to more specialized/ effective pollinators, especially Hymenoptera, and is consistent with
pollinators and phylogenetic relationships within modern lauralean crown groups. The shift to more modern pollination may in part explain the higher diversity of the Lauraceae/Hernandiaceae crown group relative to other lauralean clades. While all Laurales are insect-pollinated, not all forms of insect pollination are equal relative to their potential for catalyzing speciation (e.g., Grant, 1950; Crepet and Niklas, 2009). This pollinator transition implied by progressive changes in floral characters through time in Laurales is consistent with the Cretaceous radiation of more-derived insect pollinators, notably hymenopterans, especially Apinae (Grimaldi and Engel, 2005; Poinar and Danforth, 2006). Generally, there is a correlation between hymenopteran pollination and clade diversity in modern angiosperms (Crepet and Niklas, 2009). In summary, Jamesrosea is clearly an early representative of Laurales, with a rather typical laurad hypanthium, lauralean bilocular, valvately dehiscing pollen sacs, and the diagnostic presence of paired basal stamen appendages typical of the Atherospermataceae-Gomortegaceae clade. Pollination was likely by thrips or possibly small beetles. The presence of a core Lauralean in the Early Cretaceous that can be placed firmly within the SiparunaceaeAthersopermataceae clade, along with other previously published lauralean fossils, strongly suggests that by 100 Ma, the lauralean clade was well established, but modern groups that are now recognized as families had not become fully differentiated. Jamesrosea, although close to Gomortegaceae in the context of modern taxa, differs in having free carpels, as shared with both Siparunaceae and Atherospermataceae, but stamen morphology that is suggestive of Gomortega and may be plesiomorphic within the GomortegaAtherospermataceae clade.
ACKNOWLEDGEMENTS The authors thank Jennifer L. Svitko for expertise in CT scan tomography and for technical assistance throughout this research project and thank anonymous reviewers for helpful comments and suggestions. The authors also acknowledge Adrienne Roeder, Cornell SIPS, Plant Biology, and Weill Institute for Cell and Molecular Biology for help with confocal microscopy and use of her Zeiss 710 and financial support from Cornell CALS.
LITERATURE CITED Angiosperm Phylogeny Group. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161: 105–121. Crepet, W. L., D. L. Dilcher, and F. W. Potter. 1974. Eocene angiosperm flowers. Science 185: 781–782. Crepet, W. L., and K. J. Niklas. 2009. Darwin’s second “abominable mystery”: Why are there so many angiosperm species? American Journal of Botany 96: 366–381. Crepet, W. L., K. C. Nixon, and M. A. Gandolfo. 2004. Fossil evidence and phylogeny: The age of major angiosperm clades based on mesofossil and macrofossil evidence from Cretaceous deposits. American Journal of Botany 91: 1666–1682. Crepet, W. L., K. C. Nixon, and M. A. Gandolfo. 2005. An extinct calycanthoid taxon, Jerseyanthus calycanthoides, from the Late Cretaceous of New Jersey. American Journal of Botany 92: 1475–1485. Crisp, M. D., M. T. K. Arroyo, L. G. Cook, M. A. Gandolfo, G. J. Jordan, M. S. McGlone, P. H. Weston, et al. 2009. Phylogenetic biome conservatism on a global scale. Nature 458: 754–756.
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
Cruickshank, R. D., and K. Ko. 2003. Geology of an amber locality in the Hukawng Valley, Northern Myanmar. Journal of Asian Earth Sciences 21: 441–455. Daumann, E. 1930. Das Blütennektarium von Magnolia und die Futterkörper in der Blüte von Calycanthus. Planta 11: 108–116. Doyle, J. A. 1969. Cretaceous angiosperm pollen of the Atlantic Coastal Plain and its evolutionary significance. Journal of the Arnold Arboretum 50: 1–35. Doyle, J. A., and L. J. Hickey. 1976. Pollen and leaves from the Potomac Group and their bearing on early angiosperm evolution. In C. B. Beck [ed.], Origin and early evolution of angiosperms, 139–206. Columbia University Press, New York, New York, USA. Endress, P. K., and L. D. Hufford. 1989. The diversity of stamen structures and dehiscence patterns among Magnoliidae. Botanical Journal of the Linnean Society 100: 45–85. Feil, J. 1992. Reproductive ecology of dioecious Siparuna (Monimiaceae) in Ecuador—A case of gall midge pollination. Botanical Journal of the Linnean Society 110: 171–203. Friedman, W. E. 2009. The meaning of Darwin’s “abominable mystery”. American Journal of Botany 96: 5–21. Friis, E. M., P. R. Crane, and K. R. Pedersen. 2011. Early flowers and angiosperm evolution. Cambridge University Press, Cambridge, UK. Friis, E. M., H. Eklund, K. R. Pedersen, and P. R. Crane. 1994. Virgininanthus calycanthoides gen. et sp. nov.: A calycanthaceous flower from the Potomac Group (Early Cretaceous) of eastern North America. International Journal of Plant Sciences 155: 772–785. Goloboff, P. A., J. S. Farris, and K. C. Nixon. 2008. TNT, a free program for phylogenetic analysis. Cladistics 24: 774–786. Grant, V. 1950. The pollination of Calycanthus occidentalis. American Journal of Botany 37: 294–297. Grimaldi, D., and M. S. Engel. 2005. Evolution of the insects. Cambridge University Press, Cambridge, UK. Grimaldi, D., M. S. Engel, and P. Nascimbene. 2002. Fossiliferous Cretaceous amber from Myanmar (Burma): Its rediscovery, biotic diversity, and paleontological significance. American Museum Novitates 3361: 1–72. Herendeen, P. H., W. L. Crepet, and K. C. Nixon. 1994. Fossil flowers and pollen of Lauraceae from the Upper Cretaceous of New Jersey. Plant Systematics and Evolution 189: 29–40. Kato, M., and A. Kawakita. 2004. Plant–pollinator interactions in New Caledonia influenced by introduced honey bees. American Journal of Botany 91: 1814–1827. Kostermans, A. G. J. H. 1957. Lauraceae. Reinwardia 4: 193–256. Knight, C. L., and P. Wilf. 2013. Rare leaf fossils of Monimiaceae and Atherospermataceae (Laurales) from Eocene Patagonian rainforests and their biogeographic significance. Palaeontologia Electronica 16(3): 26A; palaeo-electronica.org/content/2013/546-eocene-laurales-from-patagonia. Lander, T. A. A., S. A. Harris, and D. H. Boshier. 2009. Flower and fruit production and insect pollination of the endangered Chilean tree, Gomortega keule in native forest, exotic pine plantation and agricultural environments. Revista Chilena de Historia Natural 82: 403–412.
• 297
Momose, K., T. Nagamitsu, and T. Inoue. 1998. Thrips cross-pollination of Popowia pisocarpa (Annonaceae) in a lowland dipterocarp forest in Sarawak. Biotropica 30: 444–448. Nixon, K. C., and W. L. Crepet. 1989. Trigonobalanus: Taxonomic status and phylogenetic relationships. American Journal of Botany 76: 828–841. Nixon, K. C., and W. L. Crepet. 1993. Late Cretaceous fossil flowers of Ericalean affinity. American Journal of Botany 80: 616–623. Peñalver, E., C. Labandeira, E. Barrón, X. Delclòs, P. Nel, A. Nel, P. Tafforeau, and C. Soriano. 2012. Thrips pollination of Mesozoic gymnsoperms. Proceedings of the National Academy of Sciences, USA 109: 8623–8628. Poinar, G. O., and B. N. Danforth. 2006. A fossil bee from the Early Cretaceous Burmese Amber. Science 314: 614. Renner, S. S. 1999. Circumscription and phylogeny of the Laurales: Evidence from molecular and morphological data. American Journal of Botany 86: 1301–1315. Renner, S. S. 2004. Monimiaceae. In N. Smith, S. Mori, A. Henderson, D. Stevenson, and S. Heald [eds.], Flowering plants of the neotropics, 252–253. Princeton University Press, Princeton, New Jersey, USA. Renner, S. S. 2005. Variation in diversity among Laurales, Early Cretaceous to Present. Kongelige Danske Videnskabernes Selskab Biologiske Skrifter 55: 441–458. Renner, S. S, A. E. Schwarzbach, and L. Lohmann. 1997. Phylogenetic position and floral function of Siparuna (Siparunaceae: Laurales). International Journal of Plant Sciences 158: S89–S98. Rickson, F. 1979. Ultrastructural development of the beetle food tissue of Calycanthus flowers. American Journal of Botany 66: 80–86. Rohwer, J. G. 2009. The timing of nectar secretion in staminal and staminodial glands in Lauraceae. Plant Biology 11: 490–492. Shi, G. D. A. Grimaldi, G. E. Harlow, J. Wang, J. Wang, M. Yang, W. Lei, Q. Li, and X. Li. 2012. Age constraint on Burmese amber based on UePb dating of zircons. Cretaceous Research 37: 155–163. Tiffney, B. H. 1977. Fruits and seeds of the Brandon Lignite: Magnoliaceae. Botanical Journal of the Linnean Society 75: 299–323. Viehofen, A., C. Hartkopf-Fröder, and E. M. Friis. 2008. Inflorescences and flowers of Mauldinia angustiloba sp. nov. (Lauraceae) from Middle Cretaceous karst infillings in the Rhenish Massif, Germany. International Journal of Plant Sciences 169: 871–889. Von Balthazar, M., K. Pedersen, P. Crane, M. Stampanoni, and E. Friis. 2007. Potomacanthus lobatus gen. et. sp. nov., a new flower of probable Lauraceae from the Early Cretaceous (Early to Middle Albian) of eastern North America. American Journal of Botany 94: 2041–2053. Walker, J. W., and J. A. Doyle. 1975. The bases of angiosperm phylogeny: Palynology. Annals of the Missouri Botanical Garden 62: 664–723. Wolfe, J. A. 1995. Paleoclimate estimates from Tertiary leaf assemblages. Annual Review of Earth and Planetary Sciences 23: 119–142. Worboys, S. J., and B. R. Jackes. 2005. Pollination processes in Idiospermum australiense (Calycanthaceae), an arborescent basal angiosperm of Australia’s tropical rain forests. Plant Systematics and Evolution 251: 107–117.
A mosaic Lauralean flower from the Early Cretaceous of Myanmar1 William L. Crepet2,6, Kevin C. Nixon3, David Grimaldi4, and Mark Riccio5
PREMISE OF THE STUDY: The floral history of early angiosperms is far from complete. The fossil discussed here has the potential to expand our knowledge of timing, reproductive biology, and paleobiogeography in early angiosperms. METHODS: Cutting-edge methodologies in CT scanning in conjunction with tomography software have opened new possibilities for discovering details in amber-preserved fossils that were inaccessible for meaningful study in the past. KEY RESULTS: The fossil is small and complex, cupulate, with numerous stamens and a suite of characters distributed in the modern families of Laurales. The most parsimonious placement of the fossil based on morphology is as a sister taxon of Atherospermataceae + Gomortega (Gomortegaceae). CONCLUSIONS: This fossil taxon, a Laurasian Lauralean from the mid-Cretaceous, is an important example of fossil Laurales with implications for biogeography and timing in the radiation and extinction in this group. KEY WORDS Atherospermataceae; Cretaceous; diversification; Gomortegaceae; insect pollination; Lauraceae; Laurales; Monimiaceae
A great deal of progress has been made in understanding angiosperm history since the rapid Cretaceous radiation of angiosperms inspired Darwin to refer to this phenomenon as the “abominable mystery” (e.g., Friedman, 2009). The coincidence in the development of algorithms generating objective phylogenies (e.g., TNT, Goloboff et al., 2008), with the emergence of techniques generating inexpensive DNA sequence data, has yielded a better understanding of relationships among extant angiosperms and has also provided important context for evaluating the affinities and evolutionary significance of fossils (e.g., Crepet et al., 2004). At the same time, there have been important advances in studies of fossil angiosperm leaves and pollen that have greatly improved our understanding of angiosperm radiation and paleobiogeography while lending new data useful to the understanding of past climates (Doyle, 1969; Walker and Doyle, 1975; Doyle and Hickey, 1976; 1
Manuscript received 1 September 2015; revision accepted 5 January 2016. 412 Mann Library, Plant Biology Section, School of Integrative Plant Science, Cornell University, Ithaca, New York 14853 USA; 3 L. H. Bailey Hortorium, Plant Biology, School of Integrative Plant Science, Cornell University, Ithaca, New York 14853 USA; 4 Division of Invertebrate Biology, American Museum of Natural History, New York, New York 10024-5192 USA; 5 Biotechnology Center, Cornell University, Ithaca, New York 14853 USA 6 Author for correspondence (e-mail: [email protected]) doi:10.3732/ajb.1500393 2
Wolfe, 1995; Crisp et al., 2009). Studies of fossil flowers, which came later than leaf and pollen studies (e.g., Crepet et al., 1974; Tiffney, 1977 ; Nixon and Crepet, 1989 , 1993 ; Friis et al., 2011 ), have provided remarkable insights into early angiosperms and their reproductive strategies (especially in conjunction with the burgeoning fossil record of insects; Grimaldi and Engel, 2005; Poinar and Danforth, 2006). Flowers also have the potential for generally providing more precise indications of minimum timing in the history of angiosperm taxa due to their diagnostic value. Hence, variously fossilized flowers have added important dimensions to the fossil record of angiosperms that have not usually been available from the records of other preserved angiosperm organs. Here we describe a mid-Cretaceous flower entombed in amber with a unique combination of lauralean features not found in any modern groups.
MATERIALS AND METHODS This study reports a fossil flower from Burmese amber of earliest Cenomanian to latest Albian age (ca. 98 million years ago [Ma]) that was acquired commercially. The amber-containing sediments are well known, and their geology and age have been exhaustively studied (Cruickshank and Ko, 2003; Shi et al., 2012), and a number of significant fossils have been reported from these sediments (e.g.,
290 • A M E R I C A N J O U R N A L O F B OTA N Y 103(2): 290–297, 2016; http://www.amjbot.org/ © 2016 Botanical Society of America
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
Grimaldi et al., 2002; Poinar and Danforth, 2006). The specimen studied herein is a tiny flower (2.1 mm diameter) that is complex in structure with numerous floral parts. The preserving medium (amber) is relatively dark and variably somewhat cloudy. Given the difficulties imposed by the nature of the preserving amber and the small size and structural complexity of this flower, it was difficult to determine floral structural details with light or with Zeiss 710 confocal microscopy. It was therefore necessary to resort to Micro-CT scan analysis to determine its detailed structural features. Micro-CT scans were performed at the Cornell University Institute of Biotechnology Imaging Facility using a Zeiss Xradia 520 Versa X-ray instrument. For each data set, 3600 X-ray projections were digitized at 0.1° intervals over 360° using 60 kV and 5 W with a 5000 ms exposure time. Reconstructions used a modified Feldkamp filtered back projection algorithm yielding isotropic x-y-z voxels. The sample was scanned at various resolutions, yielding 500 nm, 750 nm, and 1.8 μm voxels.
RESULTS Description—Flower small (ca. 2.1 mm diameter), bisexual. Floral
cup hemispheric, bearing imbricate ovate tepals, ca. 12, near the rim (Fig. 1A-C). Stamens ca. 12, arranged in an apparent tight spiral with the fertile stamens having longer filaments than the inner staminodes (Figs. 1D, 1E, 3A). Stamens are flattened and bear elongate unicellular hairs on adaxial surfaces. Similar hairs line the entire floral cup (Appendices S1, S5, see online Supplemental Data). The fertile stamens bear two paired basal appendages terminating in sagittate heads (one of a pair visible in Fig. 1F). Stamens are bilocular (presumed bisporangiate) introrse, and dehiscent by apical valvate flaps (Figs. 1G, 2A, 2B; online Appendix S2). Anther connectives extend slightly beyond the thecae, are flattened and terminally pointed (Figs. 1D, 1E, 2A). Several staminodes occur between the stamens and gynoecium (Fig. 2C), and these are filamentous with sagittate heads but lack lateral appendages (Fig. 2C, D). The exact number of these staminodes is uncertain, but is in the range of 10–12. Pollen is abundant in the amber matrix, concentrated in the vicinity of the anther locules and valvate flaps and averages 15 μm in diameter. Pollen morphology is difficult to interpret in nano-CT scan sections due to limits of resolution, apparent folding of the grains, and uneven wall preservation. There are four carpels (illustrated in cross section in Fig. 2G; longitudinal in Fig. 2H), each with an elongate, tapering style terminating in a slightly expanded stigma (Fig. 2E, F; online Appendix S3). Carpels are free from the floral cup, but basally surrounded by the cupule wall, and each includes a single ovule/seed with a developed embryo (Fig. 2G, H). Concentric, but not spirally arranged layers that represent the inner and outer epidermal boundaries of the carpel walls and seed coat surround the embryos. Taxonomic treatment—Jamesrosea gen. nov.
Etymology: Jamesrosea is named in honor of Professor James L. Reveal and his wife Rose. Type: Jamesrosea burmensis, new species Diagnosis: Flower bisexual, with hemispheric floral cup bearing imbricate tepals near the rim. Stamens ca. 12, each with two paired basal appendages. Anthers introrse, bilocular, with valvate dehiscence by apical flaps. Staminodes present between stamens and carpels. Gynoecium apocarpous, carpels superior or half-inferior, four, with elongate filiform styles. Ovules/seeds apparently one per carpel.
• 291
Jamesrosea burmensis sp. nov. Diagnosis: as for the genus Jamesrosea. Type: holotype: AMNH JZ Bu-124 Phylogenetic analysis—To place the fossil phylogenetically, a mor-
phological matrix was scored based on a significant modification of the matrix for Laurales published by Renner (1999). Focusing the analysis on Laurales is justified because of the clear synapomorphies and unique combination of lauralean features exhibited by the fossil (see Discussion). The original matrix had 15 characters, and the fossil could be unambiguously scored for nine of those characters. The new morphological matrix has 14 characters for which the fossil could be unambiguously scored (online Appendix S4). It was analyzed under a parsimony criterion using TNT (Goloboff et al., 2008). We also reduced the number of extant taxa in the matrix to 14 homogeneous terminals that are representative of all lauralean families. Thus, the final matrix had 15 taxa and 14 morphological characters. All analyses were constrained according to the molecular results for Laurales published by Renner (1999) and found in subsequent analyses. An alternative topology that differs in placing Hernandiaceae as sister to Lauraceae (Angiosperm Phylogeny Group, 2009) based on morphology was also tested as a constraint, with identical results to those reported below. The fossil was not constrained within this framework, and thus its placement was determined solely by the morphological characters, while retaining the scaffold tree determined by the molecular analyses. Phylogenetic results—The analysis of the Laurales morphological
matrix constrained by the molecular scaffold tree resulted in one most parsimonious tree of 32 steps placing Jamesrosea as the sister to Atherospermataceae + Gomortegaceae (Fig. 3). The fossil is similar to modern Gomortega in perianth features, stamen morphology including paired basal glands, anther dehiscence and number of locules, orientation of dehiscence in the fertile stamens (introrse), and the presence of staminodes between the fertile stamens and gynoecium (inconsistently seen in Gomortega). The fossil differs from Gomortega in having four free carpels as opposed to a syncarpous, inferior ovary in the latter. The fossil has introrse stamen dehiscence, while Atherospermataceae have extrorse stamen dehiscence and unisexual flowers. Moving outward in the tree, Jamesrosea differs from Siparunaceae in having valvate anther dehiscence as opposed to a variant (Endress and Hufford, 1989; longitudinal slits/valves) in the latter. Fossil pollen cannot be evaluated at adequate resolution for determination of pollen characters. Thus, the fossil is best considered an early offshoot of the AtherospermataceaeGomortegaceae lineage that predates the occurrence of syncarpy, a feature unique to Gomortega within Laurales. It is important to note that using the morphology alone, unconstrained by the molecular scaffold, results in some trees with Jamesrosea nested within the Monimiaceae-Lauraceae clade. However, the position presented here remains the best estimate based on all evidence, and the striking similarity of the fossil to Gomortega except for the apocarpous, superior gynoecium is also compelling.
DISCUSSION As indicated, assessment of the fossil’s structure using standard light microscopy or confocal microscopy was difficult due to limitations on viewing interior details of the flower in the darkened amber
292
•
A M E R I C A N J O U R N A L O F B OTA N Y
FIGURE 1 (A) Light photograph of Jamesrosea showing overall fossil. Bar = 1 mm. (B) Higher magnification photograph of the fossil illustrating the cloudy nature of the amber matrix. Bar = 500 μm. (C) Screenshot of three-dimensional (3D) movie of Jamesrosea generated by CT scan/tomography (online Appendix S1). Note stamen connectives (c) and recurved dehiscence valves (v). Bar = 120 μm. (D) Screenshot of another 3D movie generated by CT scan/tomography of a top view of the fossil embedded in amber, illustrating the distal portions of the outermost stamens, their flattened
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
and because the cup-like hypanthium masks many of the inner floral features. High-resolution CT scanning, on the other hand, provided information about floral structure that was not available with light microscopy alone (online Appendices S6–S8). Most importantly, the CT scans revealed an important set of characteristics not visible with light including an apocarpous gynoecium of four carpels with long, slender styles that were otherwise obscured by the surrounding stamens (Figs. 1E, 2C, 2D, 2F; Appendices S1, S3, S8). The scans also show paired appendages at the base of the filaments (Fig. 2F; Appendices S8, S9, S10), and bilocular anthers dehiscent by apical flaps (Figs. 1G, 2A, 2B, 2D; Appendices S1, S3, S8). These three characters were vital in determining the appropriate placement of Jamesrosea as a sister taxon of Atherospermataceae-Gomortegaceae based on the phylogenetic analysis of morphology. The CT-scans also revealed a single ovule/seed in each carpel with concentric layers (Fig. 2G, H). It is difficult to interpret these structures. Within Laurales, spirally enfolded cotyledons are found only within Calycanthaceae sensu stricto (Calycanthus and Sinocalycanthus), but in section, these appear to be clearly spiral with many layers (K. C. Nixon, W. L. Crepet, unpublished data, examination of live material), unlike our fossil. As mentioned, in Jamesrosea the concentric pattern may simply reflect different wall boundaries within the seed and embryo and be difficult to determine given the limitations of the scan. Previously described fossil Cretaceous flowers include a number of unequivocal calycanthoid (Jerseyanthus, Crepet et al., 2005), a possible calycanthoid, Virginianthus (Friis et al., 1994), and core lauralean flowers (Friis et al., 2011), e.g., the mid-Albian Potomacanthus (Von Balthazar et al., 2007), mid-Cretaceous Mauldinia angustifolia (Viehofen et al., 2008), and Turonian Perseanthus (Herendeen et al., 1994). These fossils show affinity to the crown group of Lauraceae sensu stricto, although Potomacanthus and Mauldinia have not been placed in a formal phylogenetic analysis. In combination, the assemblage of calycanthaceous and lauraceous fossil flowers from the Early and mid-Cretaceous indicates that differentiation of the Laurales clade was already underway in the Albian, and some modern Lauraceae had already differentiated by the Turonian (e.g., Herendeen et al., 1994). None of these previously described lauralean flowers show any significant similarity to Jamesrosea, and all belong either to the calycanthoid or LauraceaeMonimiaceae clade according to published analyses. With the exception of the slightly older fossil Potomacanthus (Von Balthazar et al., 2007), Jamesrosea is the only known Early Cretaceous fossil flower identifiable as Laurales. Jamesrosea also has the most convincing suite of lauralean features, including paired basal appendages associated with stamen bases (ambiguous in Potomacanthus) and valvate apical anther dehiscence. Within Laurales, Jamesrosea is clearly the oldest fossil record for the Siparunaceae-Gomortega-Atherospermataceae (SGA) clade (see Knight and Wilf [2013] for an excellent review of the fossil record of Atherospermataceae). The position of Jamesrosea within the SGA crown group places a minimum age on the Gomortegaceae-
• 293
Atherospermataceae clade, the SGA and its sister group, the Hernandiaceae-Monimiaceae-Lauraceae clade, and the ancestor of both, of ca. 98 Ma. The Atherospermataceae-Gomortegaceae clade is widely distributed in Patagonia (three genera), New Zealand (one genus), Australia (four genera), and New Guinea and New Caledonia (two genera; Knight and Wilf, 2013). There are no modern species of the crown clade (Gomortegaceae/Atherospermataceae) to which Jamesrosea is a sister taxon in mainland Asia. Thus, Jamesrosea, from the Burmese amber, represents a considerable paleo-range extension for the clade in addition to being the oldest record for the SGA clade, suggesting a more widespread distribution of the SGA clade in the Early Cretaceous. Pollination of Early Cretaceous angiosperms is interesting from the perspective of both plants and insects, and understanding the possible pollination syndrome of Jamesrosea has implications for understanding the evolution of pollination within the Laurales. Floral morphology often can predict potential pollinators, although with limited precision. Within extant Laurales, a wide range of pollination syndromes and pollinators have been reported (e.g., Grant, 1950; Feil, 1992; Momose et al., 1998; Kato and Kawakita, 2004; Peñalver et al., 2012). Documented pollinators include flies, thrips, beetles, lepidopterans, and hymenopterans. Different pollinators within the order are associated both with floral syndromes, and with particular clades reflected at the family or generic level. Because of the phylogenetic position of Jamesrosea, the modern pollinators of Siprunaceae, Atherospermataceae, and Gomortegaceae are of special interest in understanding possible pollination and pollinators in Jamesrosea. Additionally, because of similarities in floral morphology, the other members of Laurales, most importantly Monimiaceae, are also of interest in terms of pollination. Within the Laurales, Calycanthaceae is considered to be the outgroup to the remaining families based on all molecular analyses. In Calycanthus, pollination is primarily by beetles, and a single food body on the stamen connective acts as a floral reward in modern species (Daumann, 1930; Grant, 1950; Rickson, 1979). In the calycanthaceous Late Cretaceous fossil Jerseyanthus, a more complex anther connective occurs with several food bodies in a candelabralike structure (Crepet et al., 2005). In Idiospermum, pollination has been reported to be by small beetles or thrips (Worboys and Jackes, 2005). Modern Siparunaceae species are typically pollinated by gall midges (family Cecidomyiidae) (Renner et al., 1997). Gall midge pollination is associated with unisexual, relatively closed flowers that provide a host for larvae. Male flowers harbor large numbers of galls, partly because the flowers are more open and allow oviposition more easily. Female flowers are more tightly closed and typically do not harbor galls. Attempts to oviposit on female flowers results in pollination by pollen acquired from male flowers. According to Feil (1992, p.171), “The evolution of unisexual flowers in Siparuna can be explained as a result of the differential predation by larvae: unimportant in male flowers, destructive if occurring in
→
connective extensions (c), and pairs of recurved valves (arrows). Bar = 500 μm. (E) Another top view of the 3D model illustrated in (D) at an optical level revealing the stigmas in the center of the flower (arrow) and including views of some of the tepals. Bar = 500 μm. (F) Screen shot from CT-scangenerated 3D model illustrating one of the appendages attached to the base of a stamen filament. Bar = 300 μm. (G) Screen shot from CT scan illustrating longitudinal view through the flower showing stamens with two thecae (t) and recurved valves (arrowhead) and elongate hairs in floral cup. Bar = 167 μm.
294
•
A M E R I C A N J O U R N A L O F B OTA N Y
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
• 295
On the basis of these comparisons and overall morphology and size of the flower Jamesrosea, pollination was most likely by small flies, thrips, small hymenoptera, or small beetles as in some modern Monimiaceae. The relatively open, bisexual flowers and presence of paired nectar glands eliminates gall midges as likely pollinators, and no evidence of flower galls is present in the fossil. Also, the only gall midge taxa preserved in Cretaceous ambers (including Burmese amber) are of basal subfamilies that do not form galls (e.g., Lestremiinae, Porricondylinae); the gall-forming subfamily Cecidomyiinae appears in the Tertiary (D. Grimaldi, unpublished data). Thus, it is highly unlikely that Jamesrosea was pollinated by cecidomyiid midges. Thrips pollination is known in a variety of FIGURE 3 Single most-parsimonious cladogram based on parsimony analysis of the morphological “basal” angiosperm groups (e.g., Annonamatrix and constrained to the consensus phylogeny for Laurales (Renner, 1999). ceae: Momose et al., 1998), as well as cycads, and the earliest direct evidence of insect pollination may be a thrips female flowers.” Although Jamesrosea is phylogenetically close to preserved with Cycadopites-like pollen on its surface in Early CretaSiparunaceae, it has bisexual flowers, and it is unlikely that gall ceous amber (Peñalver et al., 2012). Thus, the possibility of thrips midges would be effective pollinators. pollination in Jamesrosea and other small flowers at this early phase Gomortega pollination has been minimally studied, and based of angiosperm divergence is reasonable. Further investigation of on very limited observations, may be pollinated by syrphid flies modern Gomortegaceae pollination with methods that could detect (Lander et al., 2009). Because of the limited data in this study, and thrips pollination is warranted. mixed pollen loads from other species on pollinators, it is premaRenner (2004) has considered the pattern of evolution in Lauture to exclude other pollinators as candidates for pollination in rales from a number of perspectives, focusing especially on disJamesrosea, and the primary difference between Jamesrosea and parities in relative species numbers among extant families Gormortega, the syncarpous, unistylous inferior ovary in the latter, (Calycanthaceae, 10 species; Siparunaceae, 70 spp.; Atherosperis likely important in presentation to pollinators. mataceae, 14 spp.; Gomortegaceae, 1 sp.; Lauraceae, 2500–3000 Modern Monimiaceae is noteworthy for having a high proporspp.; Hernandiaceae, 50 spp.; Monimiaceae, 195 spp.). Patterns of tion of thrips and gall midge pollination. Flowers pollinated by disparate clade sizes are common at almost any level (order, family thrips often, but not always, have floral rewards such as nectar in or genus) within angiosperms (for example, within Fagaceae, addition to pollen. Mollinedia has unisexual flowers pollinated by where monophyletic genera size ranges from 1–3 species to several thrips, without paired stamen glands (Renner, 2004), but thrips hundred). She observed (p. 444) that “The result that extant spepollination is not associated with unisexuality per se. However, cies numbers of Laurales families are significantly imbalanced, at some thrips flowers are also similar to those pollinated by small least against a Markov null model, clearly does not address differbeetles, and indeed in Monimiaceae both pollinators are seen ences in speciation and extinction rates through time.” The issue is within the same genus (Kato and Kawakita, 2004). most likely whether a Markov null model is a realistic expectation Within the Lauraceae/Hernandiaceae clade, flowers have a for diversification and/or extinction of clades within a complex single carpel and seed, most are bisexual, are usually more open landscape of climate, geography and biotic interactions. Nonethe(without a constricted orifice), and there has been a tendency to less, Renner speculated on possible environmental factors that reduce the number of stamens relative to Monimiaceae and offer might have affected speciation rates or taxon longevity in certain nectar rewards from the paired basal glands on the stamen filalauralean clades and suggested that a variable speciation–extinction ments and the glandular head of the staminodes (Rohwer, 2009). dynamic may explain differences in species diversity among differThis clade has a wide variety of pollinators, including mainly ent lauralean families (although it is difficult to imagine any other hymenopterans. →
FIGURE 2 (A) Screen shot from CT-scan movie (Appendix S1) illustrating the stamens with elongate filaments (f ) and open anther locules with recurved valves (arrowheads) as well as a sagittate stamen appendage (sa) and staminodes (st). Bar = 388 μm. (B) Top view in CT scan section showing three of the four styles in view (arrowhead) at this level and dehisced introrse locules of the stamens. Bar = 167 μm. (C) Transverse CT scan section through flower illustrating the basal portions of inner staminodes in section (arrowheads). Bar = 35 μm. (D) Screen shot of three-dimensional (3D) model movie in longitudinal aspect illustrating a style in the foreground (Sty) and four of the innermost staminodes (arrowheads), surrounding the carpels, and two microsporangia of a stamen in the background (OS). Bar = 35 μm. (E) Top view of 3D CT-scan-based model of the flower showing tepals and the four stigmas in the center (arrowhead). Bar = 140 μm. (F) Another transverse CT scan view of the flower illustrating staminodes surrounding all four styles (arrowhead). Bar = 28 μm. (G) CT scan section through the four carpels (arrowheads) illustrating layers connoting tissue boundaries. Bar = 110 μm. (H) Longitudinal CT scan through the center of the flower illustrating carpels, elongate styles, and tissue boundaries (arrowheads). Bar = 110 μm.
296
•
A M E R I C A N J O U R N A L O F B OTA N Y
explanation). One question in that context follows: Does the fossil record provide clues to modern day diversity differences among families in Laurales by revealing patterns of radiation and extinction with paleoenvironmental correlates that might reflect causative factors? While the fossil record of Laurales is good relative to the Cretaceous records of some other taxa, it is not complete enough to illuminate possible causes of the inequitable species distribution in that order. It has an inadequate pollen record—i.e., it is depauperate because, in Laurales, pollen typically is not preserved due to the (chemically and structurally), delicate nature of the exine (e.g., Herendeen et al., 1994), and pollen in some lauralean taxa could not reveal taxon level diversity because of the apparent uniformity of pollen in some clades. The leaf record, while relatively extensive in numbers of reports at least, may fail to reveal true past diversity because variation among modern leaves within the same individual or taxon or similarities among leaves in different taxa may have influenced interpretations of lauralean fossil leaves by obscuring the significance of variation in superficially similar leaves that might actually represent a number of different taxa, even at the generic level. For example, Herendeen et al. (1994, p. 29) pointed out that “Specialists in Lauraceae have noted that it is frequently difficult, if not impossible, to provide generic identifications for sterile or fruiting extant laurels and therefore they have questioned the generic identifications of fossil Lauraceae particularly sterile or fruiting material (Kostermans, 1957).” In contrast, the fossil flower record does provide interesting insights into a pattern of evolution of past lauralean diversity that is more or less congruent with phylogenetic analyses of modern lauralean taxa, and which may offer evidence of how at least one factor may be correlated with dramatically differential species diversity among lauralean families. As noted, the fossil record of the Laurales is sparse in the Early Cretaceous, making each precisely identified fossil potentially significant. The fossil taxon Jamesrosea suggests both a more diverse, widespread SGA clade in the past, spanning northern and southern hemisphere land masses in what is now both Asia and South America. The fossil record of lauralean flowers reveals an emergence of the crown groups of modern lauralean families during the Albian-Turonian transition. Known Early Cretaceous lauralean fossils do not fit into modern families (even though at least one generalized Cretaceous flower has been placed with Lauraceae [von Balthazar et al., 2007]). However, the latter fossil lacks structural features typically found in flowers of extant family crown groups, including features related to pollination syndromes such as stamen filament appendages characteristic of Lauraceae, while the Lower Cretaceous genus Virginianthus (Friis et al., 1994) may represent a higher-level calycanthoid taxon that is transitional between Calycanthaceae and the remaining Laurales (Crepet et al., 2005). By the Turonian, lauralean flower fossils can be placed within modern crown groups. Thus, Jerseyanthus (Crepet et al., 2004), while an extinct taxon represented by fossil flowers with rather dramatic differences in stamen connective morphology from those in taxa in modern Calycanthaceae, fits phylogenetically within that modern crown group. Other Turonian fossils such as Perseanthus fit neatly into modern Lauraceae based on floral characters alone. This progressive modernization of floral structure and the taxa they represent in Laurales implies a parallel change in pollination biology from flies, small beetles, and thrips to more specialized/ effective pollinators, especially Hymenoptera, and is consistent with
pollinators and phylogenetic relationships within modern lauralean crown groups. The shift to more modern pollination may in part explain the higher diversity of the Lauraceae/Hernandiaceae crown group relative to other lauralean clades. While all Laurales are insect-pollinated, not all forms of insect pollination are equal relative to their potential for catalyzing speciation (e.g., Grant, 1950; Crepet and Niklas, 2009). This pollinator transition implied by progressive changes in floral characters through time in Laurales is consistent with the Cretaceous radiation of more-derived insect pollinators, notably hymenopterans, especially Apinae (Grimaldi and Engel, 2005; Poinar and Danforth, 2006). Generally, there is a correlation between hymenopteran pollination and clade diversity in modern angiosperms (Crepet and Niklas, 2009). In summary, Jamesrosea is clearly an early representative of Laurales, with a rather typical laurad hypanthium, lauralean bilocular, valvately dehiscing pollen sacs, and the diagnostic presence of paired basal stamen appendages typical of the Atherospermataceae-Gomortegaceae clade. Pollination was likely by thrips or possibly small beetles. The presence of a core Lauralean in the Early Cretaceous that can be placed firmly within the SiparunaceaeAthersopermataceae clade, along with other previously published lauralean fossils, strongly suggests that by 100 Ma, the lauralean clade was well established, but modern groups that are now recognized as families had not become fully differentiated. Jamesrosea, although close to Gomortegaceae in the context of modern taxa, differs in having free carpels, as shared with both Siparunaceae and Atherospermataceae, but stamen morphology that is suggestive of Gomortega and may be plesiomorphic within the GomortegaAtherospermataceae clade.
ACKNOWLEDGEMENTS The authors thank Jennifer L. Svitko for expertise in CT scan tomography and for technical assistance throughout this research project and thank anonymous reviewers for helpful comments and suggestions. The authors also acknowledge Adrienne Roeder, Cornell SIPS, Plant Biology, and Weill Institute for Cell and Molecular Biology for help with confocal microscopy and use of her Zeiss 710 and financial support from Cornell CALS.
LITERATURE CITED Angiosperm Phylogeny Group. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161: 105–121. Crepet, W. L., D. L. Dilcher, and F. W. Potter. 1974. Eocene angiosperm flowers. Science 185: 781–782. Crepet, W. L., and K. J. Niklas. 2009. Darwin’s second “abominable mystery”: Why are there so many angiosperm species? American Journal of Botany 96: 366–381. Crepet, W. L., K. C. Nixon, and M. A. Gandolfo. 2004. Fossil evidence and phylogeny: The age of major angiosperm clades based on mesofossil and macrofossil evidence from Cretaceous deposits. American Journal of Botany 91: 1666–1682. Crepet, W. L., K. C. Nixon, and M. A. Gandolfo. 2005. An extinct calycanthoid taxon, Jerseyanthus calycanthoides, from the Late Cretaceous of New Jersey. American Journal of Botany 92: 1475–1485. Crisp, M. D., M. T. K. Arroyo, L. G. Cook, M. A. Gandolfo, G. J. Jordan, M. S. McGlone, P. H. Weston, et al. 2009. Phylogenetic biome conservatism on a global scale. Nature 458: 754–756.
F E B R UA RY 2016 , V O LU M E 103
• C R E P E T E T A L. — L AU R A L E A N F LO W E R F R O M T H E E A R LY C R E TAC E O U S O F M YA N MA R
Cruickshank, R. D., and K. Ko. 2003. Geology of an amber locality in the Hukawng Valley, Northern Myanmar. Journal of Asian Earth Sciences 21: 441–455. Daumann, E. 1930. Das Blütennektarium von Magnolia und die Futterkörper in der Blüte von Calycanthus. Planta 11: 108–116. Doyle, J. A. 1969. Cretaceous angiosperm pollen of the Atlantic Coastal Plain and its evolutionary significance. Journal of the Arnold Arboretum 50: 1–35. Doyle, J. A., and L. J. Hickey. 1976. Pollen and leaves from the Potomac Group and their bearing on early angiosperm evolution. In C. B. Beck [ed.], Origin and early evolution of angiosperms, 139–206. Columbia University Press, New York, New York, USA. Endress, P. K., and L. D. Hufford. 1989. The diversity of stamen structures and dehiscence patterns among Magnoliidae. Botanical Journal of the Linnean Society 100: 45–85. Feil, J. 1992. Reproductive ecology of dioecious Siparuna (Monimiaceae) in Ecuador—A case of gall midge pollination. Botanical Journal of the Linnean Society 110: 171–203. Friedman, W. E. 2009. The meaning of Darwin’s “abominable mystery”. American Journal of Botany 96: 5–21. Friis, E. M., P. R. Crane, and K. R. Pedersen. 2011. Early flowers and angiosperm evolution. Cambridge University Press, Cambridge, UK. Friis, E. M., H. Eklund, K. R. Pedersen, and P. R. Crane. 1994. Virgininanthus calycanthoides gen. et sp. nov.: A calycanthaceous flower from the Potomac Group (Early Cretaceous) of eastern North America. International Journal of Plant Sciences 155: 772–785. Goloboff, P. A., J. S. Farris, and K. C. Nixon. 2008. TNT, a free program for phylogenetic analysis. Cladistics 24: 774–786. Grant, V. 1950. The pollination of Calycanthus occidentalis. American Journal of Botany 37: 294–297. Grimaldi, D., and M. S. Engel. 2005. Evolution of the insects. Cambridge University Press, Cambridge, UK. Grimaldi, D., M. S. Engel, and P. Nascimbene. 2002. Fossiliferous Cretaceous amber from Myanmar (Burma): Its rediscovery, biotic diversity, and paleontological significance. American Museum Novitates 3361: 1–72. Herendeen, P. H., W. L. Crepet, and K. C. Nixon. 1994. Fossil flowers and pollen of Lauraceae from the Upper Cretaceous of New Jersey. Plant Systematics and Evolution 189: 29–40. Kato, M., and A. Kawakita. 2004. Plant–pollinator interactions in New Caledonia influenced by introduced honey bees. American Journal of Botany 91: 1814–1827. Kostermans, A. G. J. H. 1957. Lauraceae. Reinwardia 4: 193–256. Knight, C. L., and P. Wilf. 2013. Rare leaf fossils of Monimiaceae and Atherospermataceae (Laurales) from Eocene Patagonian rainforests and their biogeographic significance. Palaeontologia Electronica 16(3): 26A; palaeo-electronica.org/content/2013/546-eocene-laurales-from-patagonia. Lander, T. A. A., S. A. Harris, and D. H. Boshier. 2009. Flower and fruit production and insect pollination of the endangered Chilean tree, Gomortega keule in native forest, exotic pine plantation and agricultural environments. Revista Chilena de Historia Natural 82: 403–412.
• 297
Momose, K., T. Nagamitsu, and T. Inoue. 1998. Thrips cross-pollination of Popowia pisocarpa (Annonaceae) in a lowland dipterocarp forest in Sarawak. Biotropica 30: 444–448. Nixon, K. C., and W. L. Crepet. 1989. Trigonobalanus: Taxonomic status and phylogenetic relationships. American Journal of Botany 76: 828–841. Nixon, K. C., and W. L. Crepet. 1993. Late Cretaceous fossil flowers of Ericalean affinity. American Journal of Botany 80: 616–623. Peñalver, E., C. Labandeira, E. Barrón, X. Delclòs, P. Nel, A. Nel, P. Tafforeau, and C. Soriano. 2012. Thrips pollination of Mesozoic gymnsoperms. Proceedings of the National Academy of Sciences, USA 109: 8623–8628. Poinar, G. O., and B. N. Danforth. 2006. A fossil bee from the Early Cretaceous Burmese Amber. Science 314: 614. Renner, S. S. 1999. Circumscription and phylogeny of the Laurales: Evidence from molecular and morphological data. American Journal of Botany 86: 1301–1315. Renner, S. S. 2004. Monimiaceae. In N. Smith, S. Mori, A. Henderson, D. Stevenson, and S. Heald [eds.], Flowering plants of the neotropics, 252–253. Princeton University Press, Princeton, New Jersey, USA. Renner, S. S. 2005. Variation in diversity among Laurales, Early Cretaceous to Present. Kongelige Danske Videnskabernes Selskab Biologiske Skrifter 55: 441–458. Renner, S. S, A. E. Schwarzbach, and L. Lohmann. 1997. Phylogenetic position and floral function of Siparuna (Siparunaceae: Laurales). International Journal of Plant Sciences 158: S89–S98. Rickson, F. 1979. Ultrastructural development of the beetle food tissue of Calycanthus flowers. American Journal of Botany 66: 80–86. Rohwer, J. G. 2009. The timing of nectar secretion in staminal and staminodial glands in Lauraceae. Plant Biology 11: 490–492. Shi, G. D. A. Grimaldi, G. E. Harlow, J. Wang, J. Wang, M. Yang, W. Lei, Q. Li, and X. Li. 2012. Age constraint on Burmese amber based on UePb dating of zircons. Cretaceous Research 37: 155–163. Tiffney, B. H. 1977. Fruits and seeds of the Brandon Lignite: Magnoliaceae. Botanical Journal of the Linnean Society 75: 299–323. Viehofen, A., C. Hartkopf-Fröder, and E. M. Friis. 2008. Inflorescences and flowers of Mauldinia angustiloba sp. nov. (Lauraceae) from Middle Cretaceous karst infillings in the Rhenish Massif, Germany. International Journal of Plant Sciences 169: 871–889. Von Balthazar, M., K. Pedersen, P. Crane, M. Stampanoni, and E. Friis. 2007. Potomacanthus lobatus gen. et. sp. nov., a new flower of probable Lauraceae from the Early Cretaceous (Early to Middle Albian) of eastern North America. American Journal of Botany 94: 2041–2053. Walker, J. W., and J. A. Doyle. 1975. The bases of angiosperm phylogeny: Palynology. Annals of the Missouri Botanical Garden 62: 664–723. Wolfe, J. A. 1995. Paleoclimate estimates from Tertiary leaf assemblages. Annual Review of Earth and Planetary Sciences 23: 119–142. Worboys, S. J., and B. R. Jackes. 2005. Pollination processes in Idiospermum australiense (Calycanthaceae), an arborescent basal angiosperm of Australia’s tropical rain forests. Plant Systematics and Evolution 251: 107–117.
