Ozakia, a new genus of winged fruit shared between the Miocene of Japan and western North America
- First Online:
- Cite this article as:
- Manchester, S.R. & Uemura, K. J Plant Res (2014) 127: 187. doi:10.1007/s10265-013-0602-2
- 227 Views
A new genus is recognized based on winged fruits with a single species shared between the Miocene of southwestern Honshu, Japan, and the Miocene of Oregon and Idaho, USA. Calyces of Ozakia emryi gen. et sp. n. were formerly attributed to Heptacodium (Caprifoliaceae) and Amelanchier (Rosaceae); however, newly recovered specimens reveal additional characters that contradict these assignments. The pedicellate fruits are obovate, tapering basally and truncate apically, with about 10 longitudinal ribs, a prominent epigynous synsepalous calyx of five lobes, each with a midvein and a pair of weaker, ascending intramarginal primary veins. The single style has a capitate stigma. Ozakia is considered to represent an extinct eudicot genus, the familial affinities of which remain uncertain. The eastern Asian–western North American disjunction of Ozakia occurrences suggests that this plant traversed the Beringia land bridge during or prior to the Middle Miocene. Relatively few extinct angiosperm genera are known as late as the Miocene.
Miocene floras of eastern Asia and western North America share many genera in common, most of which are still living in modern north temperate forests. Most of the well-known examples belong to familiar extant arborescent genera that were widespread across the Northern Hemisphere during the Miocene, such as the deciduous eudicot trees Acer, Alnus, Carya, Fagus, Ostrya, Pterocarya, and Ulmus (Manchester, 1999). Extinct genera conforming to this biogeographic pattern are less well known.
Here we introduce a new genus of winged fruit, Ozakia, from the Miocene of western North America and Japan. Although specimens with the characteristic persistent epigynous calyx have been illustrated and discussed in the literature previously, their identifications to Amelanchier and Heptacodium are now seen to be incorrect. In North America, the specimens are rare components of the Sucker Creek, Trout Creek, and Moose Mountain floras of Oregon and the Latah flora of Emerald Creek, Idaho. In Japan, the fruits occur in the flora of Tatsumitoge in Tottori Prefecture, southwestern Honshu.
Materials and methods
Specimens were studied in collections of the National Science Museum, Tokyo (now National Museum of Nature and Science), the University of California Museum of Paleontology, Berkeley, and the Florida Museum of Natural History, Gainesville, USA. Features initially hidden within the sediment, including the pedicels, style, and stigma, were carefully exposed with a needle while observing with a dissecting microscope.
In North America, the fossils were obtained from the Middle Miocene Sucker Creek Formation of Sucker Creek, Malheur County, eastern Oregon (Molloy Ranch; 43° 13.45′ N, 117° 02.5′ W), the Middle Miocene Trout Creek Formation of Harney County, eastern Oregon (42° 11.959′ N, 118° 20.470′ W), the Miocene Moose Mountain flora (also referred to as the Menagerie Wilderness, or Cascadia flora), Linn County, western Oregon (44° 26.579′ N, 122° 18.957′ W), and the Middle Miocene Emerald Creek locality of the Latah Formation, Benewah County, Idaho (47°02.054′ N 116°20.254′ W) and the Haynes Creek Quarry, Lemhi County, Idaho (UF loc. 18088; ca 45° 03.23′ N 113° 41.42′ W). The latter site has been considered to be Oligocene (Axelrod 1998). In Japan, the specimens were collected from the Late Miocene Tochiwara Formation at Tatsumitoge in Tottori Prefecture in southwestern Honshu (35° 18′52″N, 134° 0′ll”E; Ozaki 1980). Another specimen is known from Ningyo-toge, ca. 8 km west of Tatsumi-toge (K. Uemura, pers. obs.). This flora was described by (Tanai and Onoe 1961 as the Ningyo-toge florule. Although they assigned it to the Pliocene, we now consider the age to be Late Miocene, nearly contemporanous with the Tatsumitoge flora.
Ozakia gen. n.
Diagnosis, same as for type species, below
Diagnosis Winged fruits with persistent actinomorphic, epigynous perianth. Pedicel persistent, dispersed with fruit, thin (0.2 mm), and 4–5 mm long. Fruit body obovate, 3.2–4.0 mm wide, 6.0–9.0 mm long, acute and tapering basally, truncate apically, with 8–10 straight to slightly sinuous, unbranched, longitudinal ribs that enter into the epigynous calyx. Calyx consisting of a circular disk or cup-shaped structure from which five oblanceolate lobes arise, each with a midvein bearing pinnate secondary veins and a pair of weaker, ascending intramarginal primary veins extending 1/3 or more of the distance from base to lobe apex. Intramarginal veins continuous across the sinus area between the bases of adjacent calyx lobes. Tertiary veins comprising a fine irregular meshwork and forming loops adjacent to the margin. Corolla not observed (absent, or not persistent). Style solitary, 2.5 mm long, with a capitate stigma.
Specimens Holotype: UF18240-44561 from Sucker Creek, Molloy Ranch, Oregon. Paratypes: UF44560 from Sucker Creek, Molloy Ranch, Oregon; UF18998-35555; UCMP 594 from Trout Creek, Oregon; UF18110-55057, 55058, 55376 from Moose Mountain, Oregon; UF18596-59771, from Emerald Creek, Idaho; UF18088-14768 from Haynes Creek, Idaho; NSM PP16115, 16227 from Tatsumitoge flora, Tottori Pref., Honshu, Japan.
Etymology The generic name recognizes Dr. Kimihiko Ozaki (1941–1989) in memory of his contributions to Neogene Paleobotany of Japan. The specific epithet recognizes Mr. Howard Emry, who collected the Sucker Creek specimens that provided the initial impetus for this investigation.
Discussion Although at first glance the fossils might be interpreted as flowers with a fused dish-like calyx and protruding free elongate petals (Fig. 1a, j), closer observation indicates that the dish-like flange and protruding lobes all represent one perianth whorl, with interconnected venation (Fig. 2b–d). Thus it represents a calyx only, with-out petals or stamens; apparently the latter organs did not persist to fruiting stage. The circular dish-like (disk) portion of the calyx arises from the top of the ovary (epigynous), with venation that is continuous with the longitudinal ovary ribs. The dish like portion is thicker in texture than the lobes (darker in color, Figs. 1e, 2b, e), possibly nectariferous. Prominent longitudinal ribs on the fruit body, regularly arranged around the fruit (Figs. 1d–f, j, k, 2d–f) provided vasculature to the calyx, and possibly other epigynous organs. Collectively, the epigynous, actinomorphic five-lobed calyx, elongate, longitudinally ribbed fruit, single style and capitate stigma are helpful in considerations of the systematic affinities of Ozakia.
MacGinitie (1933) illustrated a specimen of this species (p. 59, pl. 9, Fig. 5; reillustrated here as Figs. 1j, k) from the Trout Creek flora of Oregon as “Amelanchier grayi (?) Chaney,” along with leaves that he assigned to the same species (but without the question mark). He considered the flower to have a pentamerous perianth, perigynous stamens, and carpels adnate to the surface of a deep cup-like axis and concluded that it “is almost certainly a member of the Rosaceae and is similar to mature flowers of the genus Amelanchier.” However, evidence for the presence and number of stamens is not seen in his specimens, nor in any of the specimens now available, and it is now clear that there is a single unbranched style with one capitate stigma (Fig. 1h). The style of Amelanchier flowers is branched with four or five stigmas, the calyx lobes are sharply pointed, and the fruit is subglobose without longitudinal ribbing. Because the attachment of petals and stamens are not preserved in these fossils, we are unable to discern whether the flower possessed a hypanthium.
Ozaki (1980) documented the same kind of fruits from the Tatsumitoge flora of Japan (e.g., Fig. 1a–d). He assigned them, along with leaves that he believed to be conspecific, to the caprifoliaceous genus Heptacodium as H. hokianum. However, the leaves and fruits were not found in attachment and the interpretation that they represent a single taxon remains open to question. The leaf, which was designated as holotype, retains the epithet hokianum, whereas we now transfer the fruits to a new genus named in Dr. Ozaki’s memory. The fruit does indeed resemble those of extant Heptacodium in the persistent epigynous calyx and in the venation of the perianth lobes (Figs. 2h, i). However, extant Heptacodium fruits are sessile and shed without pedicels (contrasting with the long narrow pedicels of the fossil, Fig. 1a, e, f), lack the epigynous circular dish-like flange, and possess only the prominent, deeply lobed calyx. The ephemeral gamopetalous corolla and style of Heptacodium are shed prior to fruiting stage. In view of these differences, the assignment to this genus cannot be maintained. Abelia, also in the Caprifoliaceae, bears epigynous persistent pentamerous calyces similar in configuration to those of Ozakia. However, in Abelia, as in Heptacodium, the style does not persist to fruiting stage, because it is torn away as the gamopetalous corolla detaches. The persistent style seen in the fossil fruits (e.g., Figs. 1a, 2f), indicates that these fossils probably do not belong to Caprifoliaceae, and more likely belong to a family with absent, or in which the corolla was unfused, such that the petals could fall away without dislodging the style.
Similarities with Calycites ardtunensis Crane from the Paleocene of Scotland and USA, and the Eocene of western North America (Crane 1987) were also considered. Although C. ardtunensis fruits have a similar general appearance, they bear six, rather than five wings, they have fewer ribs on the ovary, lack the epigynous dish like structure, and their stylar condition is unknown.
To assess the likely systematic affinities of Ozakia with extant taxa, we compiled a listing of all extant genera of angiosperms, of which we are aware, that possess fruits with persistent actinomorphic perianth. Among these 35 genera in 26 families, most can be excluded from consideration by their hypogynous perianth. Those with epigynous 5-merous perianth, are Ancistrocladus (Ancistrocladaceae), Abelia and Heptacodium (Caprifoliaceae), Calycopteris (Combretaceae), Petraeovitex (Lamiaceae), Kissenia (Loasaceae), Buckleya (Santalaceae), and Korupodendron (Vochysiaceae). However, under detailed comparison, each of these is readily distinguished from the fossil at hand, by the features of ovary ribbing, epigynous disk, and the single style with capitate stigma.
Although an exact match to a modern genus has not been possible, and some important floral morphological characters like number and position of seeds, stamens, pollen, and corolla characters are unknown, we may approximate the general systematic position of Ozakia from available characters. The features of floral parts in five, synsepaly, epigyny allow us to deduce that Ozakia is a eudicot. The single style and stigma indicates either a single carpel or fully fused multiple carpels. Whether it represents Rosids or Asterids is more difficult to deduce. Among Asterids, there are suites of similar characters in Araliaceae, Lissocarpaceae, Loasaceae, and Caprifoliaceae. In the Rosids, the Myrtales—e.g., Onagraceae, Myrtaceae, Lythraceae seem possible. However, in the absence of more characters for Ozakia, e.g. corolla configuration, number and arrangement of stamens, pollen morphology, a more precise systematic placement for this extinct genus eludes us.
Biogeographic and ecological considerations
The presence of Ozakia in the Miocene of both Japan and western North America, together with its apparent absence from deposits of similar age in Europe, suggests that it spread directly between Asia and North America. The land connection across Beringia was apparently open through much of the Tertiary (Tiffney and Manchester 2001). Although Ozakia has not been collected from intervening fossil assemblages, the Miocene Seldovia Point flora of the Cook Inlet region of Alaska records a temperate fossil flora with many taxa shared with the midlatitude Miocene floras of Asia and North America (Wolfe and Tanai 1980). The occurrences in North America are older, mostly Middle Miocene, and extending back to the Oligocene based on the single occurrence from the Haynes Creek flora, while the Japanese occurrences are younger, Late Miocene. This might indicate a refugial situation in eastern Asia, or could be a sampling artifact due to fewer Late Miocene floras studied in North America.
The North American Miocene floras from which Ozakia has been recovered represent temperate mixed mesophytic forest. The Trout Creek flora includes Abies, Picea, Pinus, Alnus, Quercus, Nymphaea, Acer and some other angiosperms (MacGinitie 1933). The Sucker Creek flora includes Abies, Chamaecyparis, Picea, Pinus, Thuja, Tsuga, Smilax, Typha, Acer spp., Mahonia, Trochodendron, Alnus, Betula, Ostrya, Quercus spp., Carya, Juglans, Pterocarya, Lindera, Leguminosae (pods), Nymphaea, Nyssa, Fraxinus, Platanus, Amelanchier, Crataegus, Rosa, Ptelea, Populus, Salix, Hydrangea, Ailanthus, Tilia, Ulmus (Fields 1996; Graham 1963, 1965).
The plants with which Ozakia is associated in the Tatsumitoge flora of Tottori Prefecture, southwestern Honshu, Japan, indicates a flora also dominated by deciduous elements. The flora includes Abies, Picea, Pseudotsuga, Pseudolarix, Tsuga, Cunninghamia, Metasequoia, Sciadopitys, Cyclocarya, Juglans, Pterocarya, Populus, Salix, Alnus, Betula, Carpinus, Castanea, Fagus, Quercus, Celtis, Ulmus, Zelkova, Eucommia, Liriodendron, Magnolia, Lindera, Sassafras, Hamamelis, Philadelphus, many rosaceous and leguminous genera, Clematis, Ceratophyllum, Stewartia, Sapium, Koelreuteria, Acer, Meliosma, Ilex, Euonymous, Buxus, Berchemia, Hovenia, Rhamnus, Sageretia, Parthenocissus, Tilia, Rotala, Cornus, Acanthopanax, Aralia, Hedera, Clethra, Enkianthus, Rhododendron, Diospyros, Styrax, Chionanthus, Fraxinus, Paulownia, Patrinia, Potamogeton and Bambusium (Ozaki, 1980, 1981).
Many of the genera that cohabitated with Ozakia in the Miocene are still living today in modern forest of the Northern Hemisphere. However, Ozakia became extinct—possibly in response to post-Middle Miocene climatic cooling. Other examples of extinct genera from the Miocene include Pseudofagus (Smiley and Huggins 1981) Nordenskioeldia (Manchester et al. 1991), Diplodipelta (Manchester and Donoghue 1995), Buzekia (Manchester 1999) and Shirleya (Pigg and DeVore 2005) from the Middle Miocene of the western United States.
We thank Robert Rosé and Greg Retallack for contributing important specimens and data concerning the collections made from Moose Mountain, Howard and Darlene Emry for specimens from Sucker Creek, William C. Rember for the specimen from Emerald Creek, Allen Marquette for the specimen from Haynes Creek, and Diane Erwin for providing access to specimens at the University of California Museum of Paleontology. Hongshan Wang provided assistance with specimens cited from Florida Museum of Natural History. Walter S. Judd provided helpful advice concerning possible botanical affinities and Terry Lott assisted with the investigation of extant taxa for comparison. Patrick F. Fields and two anonymous reviewers provided helpful comments for improvement of the original manuscript.