THE EARLY MIOCENE CAPE BLANCO FLORA OF COASTAL OREGON by LISA FRANCIS EMERSON A DISSERTATION Presented to the Department of Geological Sciences and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy September 2009 11 University of Oregon Graduate School Confirmation of Approval and Acceptance of Dissertation prepared by: Lisa Emerson Title: "The Early Miocene Cape Blanco Flora of coastal Oregon" This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Geological Sciences by: Gregory Retallack, Chairperson, Geological Sciences Rebecca Dorsey, Member, Geological Sciences Joshua Roering, Member, Geological Sciences Barbara Roy, Outside Member, Biology and Richard Linton, Vice President for Research and Graduate Studies/Dean of the Graduate School for the University of Oregon. September 5, 2009 Original approval signatures are on file with the Graduate School and the University of Oregon Libraries. © 2009 Lisa Francis Emerson 111 An Abstract ofthe Dissertation of Lisa Francis Emerson for the degree of in the Department of Geological Sciences to be taken IV Doctor ofPhilosophy September 2009 Title: THE EARLY MIOCENE CAPE BLANCO FLORA OF COASTAL OREGON Approved: Gregory 1. Retallack This dissertation establishes the age, depositional environment, composition, and climatic conditions for the Cape Blanco flora. The paleotemperature estimated by the Cape Blanco flora, the Temblor flora ofCalifornia, and the Seldovia flora ofAlaska are then compared with sea surface temperatures estimated from oxygen isotope analysis of benthic foraminifera. The unconformity-bound shallow marine sandstone ofFloras Lake includes a redeposited tuffbed which contains fossil leaves at Cape Blanco. An 40Ar/39Ar age of 18.26 ± 0.86 Ma is presented for the tuff as well as a paleomagnetic stratigraphy ofthe sandstone. Sedimentary structures ofthe tuffbed are evidence that the tuffwas deposited at or just above the strand line. The depth oftuffdeposition was shallower than the adjacent marine sands, and this short-lived shoaling may have been a result of increased sediment supply. vThe fossil flora was an oak forest with numerous species ofFagaceae. Additional components include lanceolate Salicaceae leaves, entire margined Lauraceae, fragmentary Betulaceae, and lobed Platanaceae. Coniferous debris, charcoal, Equisetales, and Typhaceae forms are also figured. Ten leaffonns could not be confidently assigned to established names but are described, figured, and called angiosperm forms 1-10. In total 44 unique fonns are identified. The size and margin type ofthe dicot specimens are quantified, and by comparison with known modem floras, a former mean annual precipitation of201 (+86, -61) cm and a former mean annual temperature of 18.26 ± 2.6°C are estimated. The paleotemperature ofthe ~17.5 Ma Seldovia Flora and the ~17.5 Ma Temblor Flora are estimated using the same method, establishing a ~0.7°C per degree of latitude temperature gradient for the northern hemisphere temperate zone. The leafbased . gradient is steeper than the sea surface temperature gradient, of~0.26°C per degree of latitude as estimated from oxygen isotopic composition of foraminifera collected from ocean sediment cores. Both fossil leaf and isotope methods suggest that the early Miocene was ~5°C warmer than today. This thesis includes unpublished co-authored material. CURRICULUM VITAE NAME OF AUTHOR: Lisa Francis Emerson PLACE OF BIRTH: Redmond, Oregon USA DATE OF BIRTH: October 16, 1979 GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene, OR Dartmouth College, Hanover, NH DEGREES AWARDED: Doctor of Philosophy, Geological Sciences, 2009, University of Oregon Bachelor of Arts, 2001, Earth Science, Dartmouth College AREAS OF SPECIAL INTEREST: Neogene Stratigraphy Paleobotany Paleoclimate PROFESSIONAL EXPERIENCE: Graduate Teaching Fellow, University of Oregon Department of Geological Sciences, September 2004 to June 2009 Geoscience Intern, ExxonMobil Exploration Company, Houston, TX, Spring 2008 Geoscience Intern, BP North America, Houston, TX, Summer 2007 Naturalist, Aspen Center for Environmental Studies, Aspen, CO, 2002-2004 vi vii GRANTS, AWARDS AND HONORS: Baldwin Fellowship, University of Oregon Department of Geological Sciences, $1,000.00, June 2009 Dunbar Scholarship, University of Oregon, College of Arts and Sciences, $5,000.00, June 2008 Baldwin Fellowship, University of Oregon Department of Geological Sciences, $1,200.00, June 2007 Condon Fellowship, University of Oregon Department of Geological Sciences, $600.00, June 2006 University of Oregon, Graduate School Research Award to attend the Advances in Paleobotany Conference, January 2006 Baldwin Fellowship, University of Oregon, Department of Geological Sciences, $300.00, June 2005 The Geological Society of America, Student Grant, "Climatic and Biogeographical Implications of the 12 Ma cape Blanco Flora, Cape Blanco, OR, $1,500.00, May 2005 Dartmouth College Earth Science Department, Student Grant, $1,500.00, September 2000 PUBLICATIONS: Raymond, Kristina R., Prothero, D., Emerson, L., Retallack, G., 2008, Magentostratigraphy of the lower Miocene Sandstone of Floras Lake and the C ape Blanco Flora, Oregon. Geological Society of America abstracts with programs,40 (6),477. Emerson, L. F., and Rempel, A. W., 2007, Thresholds in the sliding resistance of simulated basal ice. The Cryosphere, v. 1, p. 11-19. Emerson, L. F., and Retallack, G. J., 2007, Miocene coastal vegetation preserved by volcanic eruption at Cape Blanco, OR. Geological Society of America abstracts with programs, 39 (6), 401. ------------ viii Emerson, L. F., Soule, A, Belien, 1., Deardorff, N., Gottesfeld, E., Johnson, E., McKay, D., and Wisely, B. A, 2007, Implementation of ground-based LiDAR, total station, and GPS in an advanced geophysical study of basaltic cinder cone morphology, Four Craters Volcanic Field, Oregon. Geological Society of America abstracts with programs, 39 (6), 123. Emerson, Lisa F. and Rempel, Alan W., 2006, Laboratory Study of the Frictional Properties of Simulated Basal ice. Eos Trans. AGU, 87 (52) Fall Meeting SuppL, Abstract C31A-1245. Emerson, Lisa F. and Retallack, Gregory J., 2006, A nearest living relative comparison of the Cape Blanco Flora, Oregon. USA Advances in paleobotany conference; Gainesville, Florida, USA, March 2006. Emerson, Lisa F. and Retallack, Gregory J., 2006, Comparison of middle Miocene angiosperm leaf fossils from Cape Blanco with type specimens. Proceedings of the Oregon Academy of Sciences, v. 62, p. 35. Emerson, Lisa F. and Rempel, Alan W., 2005, Experimental studies of the influence of entrained sediment on ice friction. Eos Trans. AGU, 86(52) Fall Meeting SuppL, Abstract C51B-0287. Emerson, Lisa F. and Retallack, Gregory J., 2005, A new middle Miocene flora from Cape Blanco, OR. Abstracts with Programs- Geologic Society of America Fall Meeting 37(7), 362. --_._._._----------- ix ACKNOWLEDGMENTS My committee provided continuous support and guidance. I appreciated Greg Retallack's good humor, Josh Roering's insistence on clarity, Becky Dorsey's enthusiasm, and Bitty Roy's humanity. I am grateful to the Department of Geological Sciences under the direction of Department Heads, Dana Johnson, Kathy Cashman and Becky Dorsey, for fostering an environment with high standards and respectful approach to graduate education. The University of Oregon Department of Geological Sciences provided continuous financial support through teaching appointments. I appreciate their commitment, the commitment by the University of Oregon College of Arts and Sciences, and the taxpayers of the State of Oregon to higher education. My research required examination of materials from distant libraries. Thank you to all librarians, past and present, who maintain these resources. Many in the scientific community contributed to my education, and for their patience and attention I am grateful: Don Prothero of Occidental College, Scott Wing and Jonathan Wingerath at the Smithsonian, Diane Erwin and Howard Schorn of the University of California Museum of Paleontology, Jeff Myers of Western Oregon University, and Robert Duncan of Oregon State University. Two terms of Greg Retallack and his Paleobotany students broke and carried a huge amount of rock, as did Peter Almond, thank you. Brendan Toch, my husband, was a bottomless well of patience. My family and friends were always encouraging, and despite the long road, I never felt abandoned. x TABLE OF CONTENTS Chapter 1. INTRODUCTION .. II. RADIOMETRIC AND PALEOMAGNETIC DATING OF AN EARLY MIOCENE FOSSIL FLORA FROM CAPE BLANCO, SOUTHERN OREGON COAST . 1. Introduction . 2. Geologic Setting . 3. Sedimentary Environment of the Tuff.. .. 3.1. Methods . 3.2. Tuff Outcrop Results . 3.3. Tuff Outcrop Interpretation .. 40 39 . f h ff4. Arl Ar Datmg 0 t e Tu .. 40 394.1. Arl Ar Methods .. 4.2. 4°Ar/39Ar Results . 5. Provenance of the Tuff .. 5.1. XRF Methods .. 5.2. XRF Results .. 5.3. XRF Interpretation . 6. Paleomagnetic Stratigraphy . 6.1. Paleomagnetic Method .. 6.2. Paleomagnetic Results . xi Page 1 2 2 5 8 8 9 11 12 12 14 15 15 15 16 20 20 21 XlI Chapter Page 6.3. Paleomagnetic Interpretation........................................................... 26 6.4. Paleomagnetic Discussion... 26 7. Conclusions 27 III. SYSTEMATICS OF THE EARLY MIOCENE CAPE BLANCO FLORA OF COASTAL OREGON... 29 Introduction................................................................................................. 29 Geological Setting 31 Materials and Methods 33 Megafossil Remains 33 Systematics.................................................................................................. 35 Equisetales............................................................................................. 35 Pinales 36 Typhales 39 Salicales 39 Fagales 44 Juglandales 54 Urticales 55 Hamamelidales 56 Ranunculales 59 Laurales 60 xiii Chapter Page Rosales 61 Rhamnales 61 Lamiales 63 Dipsacales 64 Unknown Botanical Affiliation 64 Taxonomic Summary.................................................................................. 73 Floristic Interpretation................................................................................. 73 Physiognomic Interpretation........................................................................ 75 Paleotopography 77 Taphonomy............. 79 Conclusion 80 IV. EARLY MIOCENE TEMPERATURE GRADIENTS, A COMPARISON OF PALEOBOTANICAL AND ISOTOPIC PREDICTIONS................................ 81 Introduction................................................................................................. 81 Methods 82 Paleobotanical........................................................................................ 82 Geochemical.......................................................................................... 85 Modern Observed Temperatures .. 86 Analysis................................................................................................. 86 Results 86 Discussion.. 88 Chapter XIV Page Conclusion .,. 90 V. CONCLUSIONS..................................................................................................... 91 REFERENCES 92 Page xv LIST OF FIGURES Figure Chapter II: 1. Location map for the sandstone of Floras Lake outcrop..................................... 3 2. Map of Pacific Northwest showing tectonic terranes.......................................... 4 3. Stratigraphic section.......................................................................................... 7 4. Stratigraphic section of tuff 8 5. Photographs of tuff bed 10 6. Plagioclase cumulative Argon release plateau.................................................... 14 7. Total Alkali-Silica diagram 16 8. Orthogonal demagnetization plots 23 9. Stereonet of paleomagnetic results..................................................................... 25 Chapter III: 1. Location map of fossil locality.......................................................................... 31 2. Stratigraphic section of the tuff bed.. 32 3. Rarefaction curve 34 4. Equisetales and Pinales 37 5. Typhales, Salicales, and Fagales 41 6. Fagales and Juglandales 48 7. Urticales, Hamamelidales, Ranunculales, Laurales, Rosales.. 57 8. Laurales, Rhamnales, Lamiales, Dipsacales, and undetermined......................... 65 XVI Figure Page 9. Unknown botanical affiliation 69 Chapter IV: 1. Map of Flora Locations 84 2. Temperature data............................................................................................... 87 Page xvii LIST OF TABLES Table Chapter II: 1. Analytical data for radiometric dates................................................................. 14 2. XRF results 15 3. Possible volcanic sources 19 4. Paleomagnetic data............................................................................................ 24 Chapter III: 1. Miocene floras of Oregon.................................................................................. 30 2. Fossil quantities and similar modern plants........................................................ 74 3. Morphology of leaves........................................................................................ 78 Chapter IV: 1. Climate data 85 1CHAPTER I INTRODUCTION The Cape Blanco flora and its paleoclimate implications are the subject of this dissertation. The second chapter establishes the age and depositional environment of the flora. The third chapter describes the flora and includes systematic assignments, photographs of the fossil leaves, and an estimate ofthe mean annual temperature and mean annual precipitation during late Early Miocene time. The fourth chapter expands the scope of the study to include other Early Miocene floras and the floral data is compared to oxygen isotope based paleoclimate estimates. Together, the second, third, and fourth chapters provide the setting, composition, and importance of the Cape Blanco flora. The paleomagnetic laboratory analysis and technical figures of the second chapter were produced by my co-authors, Don Prothero and Kristina Raymond. Greg Retallack, a co-author for the second and third chapters assisted in all facets of the work, including idea development, data collection, data analysis, and editing. 2CHAPTER II RADIOMETRIC AND PALEOMAGNETIC DATING OF AN EARLY MIOCENE FOSSIL FLORA FROM CAPE BLANCO, SOUTHERN OREGON COAST The paleomagnetic laboratory analysis and paleomagnetic technical figures of the second chapter were produced by my co-authors, Don Prothero and Kristina Raymond. Greg Retallack, assisted in all facets of the work, including idea development, data collection, data analysis, and editing. Though they are not co-authors, Robert Duncan and John Huard at Oregon State University completed the mineral separation and laboratory analysis for the 40Ar/39Ar date. 1. Introduction Accurate reconstruction ofpast climates using terrestrial fossils requires knowledge of the age, provenance, and depositional environment for host rocks. The Cape Blanco Flora is a macrofossil assemblage used for paleoclimatic estimates (Emerson and Retallack 2005, Wolfe 1994) and this paper establishes the age and depositional environment ofthe rocks that contain the flora. Cape Blanco is the westernmost extent of the north/south trending Oregon coast, and the sea cliffs southeast of the cape expose a sequence ofEocene to Holocene sedimentary rocks on the south limb ofan east/west trending anticline (Armentrout 1980). The axis of the anticline is 1 3km north of the cape (Figure 1). Here, emphasis is placed on a 7.5 m bed of redeposited pumice-rich lapilli tuff that contains the fossil flora. Paleomagnetic stratigraphy, calibrated by a single-crysta1 4oArP9Ar date, constrains the age and tectonic rotation of 124°35' W 1 kll1 . Legend ~ Anticl!ne o/u Fault . Pacific Ocean Cape Blanco Cape Blanco 11 ?, 'iJ~ ~ " Oregon. USA Fignre 1: Location map for sandstone of Floras Lake outcrop, modified from Kelsey (1990) and Addicott (1980). For detailed bedrock geology refer to Dott (1962) and for detailed terrace maps refer to Kelsey (1990). 4the sediments. Whole rock XRF analysis of the tuff and the adjacent marine sandstones aid in determining provenance. Sedimentary structures refine the depositional environment for the tuffbed. Combined, these results serve as a guide to volcanic source terrane, tectonic setting, and ancient environment ofthe fossil flora. Cape Blanco lies near the intersection ofthree geologic provinces (Figure 2): the subducting Gorda plate, the accreted and metamorphosed Mesozoic Klamath OR l,-------rN1T7V- • _110 Gcrrelalive un:s ... t-lsed on K-Ar dales, Nurri:ers o::rrespClod m refl?"€nce nurJt,;;r.; "lii.;.1e 3. • t:ii1-)' Miooe"e i'lo',-y f<::nr-abcn • Early Mc<-"!!l'O RD,!? :::reek ',lErnter of 1hf- J::>"n D3)' Foml3:r.o lim l.le5az(),~ arr.- Pale::w:o ~I"letal"101Jhicr{)~k~ ' cascadia S'JbdLdion Zone -hrJS1 fa'Jlt ~ Strike'SID Fa.lt ;::::::::::::::;;:::::::::::::::l ,:i r::j 'C S O I 8st k ElI))i :::1 ~L-~ " DC ;;:::~, .,::~ ~I ::;:;:;:;::\; "Cascade ~ j·!!lilli!iiili ·~~( :;:. r.,. 'I '.: .. ,. ,) , ....... '" I1I.·<:~, CAt r.., t..,,) ~ '7 >- ",-..,.. KI th " L-'" L- ,>. 1arna . ~ .,"/': .,"."') Terrane. , ~ • .".~..- ..... ., ... ~:;: ~:.: r. I ~r"-~--..b/ Gorda Plate .... "' ..... 200km Pacific Ocean Figure 2: Map of the Pacific Northwest showing tectonic terranes relevant to the history ofCape Blanco, Continental province base map modified from Wells and Heller (1988) and oceanic plate boundaries modified from Dziak et al. (2001). Multiple samples are included in the correlation points because they are too close together to differentiate at this scale, SAF: San Andreas Fault, OR: Oregon, WA: Washington, CA: California, NY: Nevada, ID: Idaho 5Terrane and the Tertiary marine basins of the Coast Ranges. The tectonic setting is complex, and addressing uncertainty regarding the areas tectonic history also motivated this study. Cape Blanco has the shortest trench normal distance of the Oregon coast (Miller et al. 2001) and the age of the Gorda Plate at the trench west of Cape Blanco is 6 Ma (Wilson 1989). The accreted Mesozoic rocks of a broadly defined Klamath Terrane have their most northerly coast outcrop at Cape Blanco. The Siletzia terrane thought to underlie the Cenozoic sediments of the Oregon coast range and Willamette Valley has its most southward extent just north of Cape Blanco (Wells et al. 1998). Geological refinement of this section allows for increased precision in the volcanic and tectonic models of the region. 2. Geologic Setting The northernmost exposure of accreted Mesozoic rocks along the Oregon Coast is at Cape Blanco (Kelsey 1990), and the Jurassic Otter Point sandstones and mudstones (Dott 1962, 1971, Koch 1966) are the resistant rocks of the headland. Unconformably overlying the Otter Point Formation are Cenozoic deposits that have been offset by roughly north trending normal faults (Dott 1962) and folded by the Cape Blanco anticline (Diller 1902, Kelsey 1990). The southern limb of the east-west trending anticline exposes a sequence ofSE dipping sedimentary rocks; Eocene Roseburg formation shale (Armentrout et al. 1983, Bandy 1950), early Miocene sandstone of Floras Lake (Addicott 1980, 1983; Armentrout 1980), late Miocene Empire sandstone (Armentrout et al.,1983; Diller 1902), Pliocene Port Orford sandstone (Baldwin 1945), and a flight ofPleistocene 6marine terraces (Bockheim et al. 1992, 1996; Diller 1902; Janda 1969; Kelsey 1990). The dip of each unconformity-bound unit (Durham 1953) decreases with age, indicating that the structure has been active since the deposition of the sandstone ofFloras Lake (Kelsey 1990). The fault at the neck of the cape brings down late Miocene Empire Formation sandstone which fonus the buff colored cliffs that are the source of Cape Blanco's name (Dott 1962). This study focuses on the 158 m thick Burdigalian (Addioctt 1976, 1980, 1983) sandstone ofFloras Lake (Figure 3) which is punctuated by a 7.5 meter thick tuff bed (Figure 4). Previous Miocene age estimates were based on molluscan biostratigraphy (Addicott 1976, 1980, 1983; Allison and Addicott 1976; Moore 1963; Moore and Addicott 1987). The paleoecology of those assemblages indicated rapid deposition along a NW trending rocky coast (Leithold and Bourgeois 1983). Leithold and Bourgeois (1983) subdivided the Sandstone ofFlora Lake into four units; the conglomeratic sandstone unit, the pebbly sandstone unit, the tuff, and the sandstone unit (Figure 3). The fming-upward trend seen in these units records a transgressive sequence with a continuous increase in depositional depth from intertidal at the bottom to below stormweather wave base at the top (Addicott 1983, Leithold and Bourgeois 1983). The apparent deepening was interrupted by the deposition ofthe tuff and was attributed to vo lcanically induced uplift (Leithold and Bourgeois 1983). Here we closely examine the tuff bed in order to better interpret the fossil flora. 75C • • • • • In~inatc(" • • • • • • • • • • • • • • • • • • • • • )=:li"<~~n 1lJ "7J 19722 18524 13D55 18745 20 ,1\ u \ • 32 \ \ 036 \ 035 \ \ \ 0 37 - 039 040 \ \ 0 41 \ \ \ )!-~= ,.m~.: 031 ~6 M3gne10SI'at graph)': <;3m~je "and pclar~ 0 30 1:0 :X Tuff 1JI.:!e : 0.'6 M" )J 1!:O Pebbly l'~(' -li=:;;;~ Sandstone Sandstone Conglonleratic 5C Sandstone DepositionJ Units of Leorod,md Bourgecis: ~ 983;, Figure 3: Stratigraphic section of sandstone ofFloras Lake with sample locations for XRF and paleomagnetic stratigraphy. Black circles have normal polarization and white circles are reversed. Leaf symbol indicates the location of the fossil flora. Unexposed portions ofthe section are marked by crossing lines. Assigned polarity zones, inclination, and declination results are on the right. 8750 em 't; 'i :'~ ,>'f' '.: . Uppe: ::::'f,-::';'," +-; -- Middle (Fossil -..- f'i"'1>. E-=aring Un:) :u-u ~; LcW=' Upper 580cm··· ......,.,........~""" Middle _ :.._. ,.':.~:':~: 5 '1::. em --- ,,':':"::-::':'::'.'.' ::r:~;;; ~:i\\~:;;: :~::: ':.' ~.; ,".~.)"..\.- :.,: ...~-.~~':.:. ~. =_:.. Lower t~=~ Siltstone (Lagoon Facies) P:/;7":j Crossbedded coarse'~/<:); sandstone (Channel Facies) Parallel thin beds (Swash zone Facies) ~ Fossil Leaves (f) UJ tJ OJ ro ::l IT CL. IT it) Figure 4: Stratigraphic section of tuff with three zones; upper, middle,and lower. Leafsymbol indicates location of fossil bearing unit. 3. Sedimentary Environment of the Tuff 3.1. Methods The section is a sea cliff exposure intermittently interrupted by landslides (Figure 3). The section was measured using the eye-height method and grain size and meter scale observations of sedimentary structures confIrmed the descriptions ofLeithold and 9Bourgeois (1983). Grain size and sedimentary structures were recorded for the tuff bed with 0.1 m precision. 3.2. Tuff Outcrop Results The 7.5 m tuff is a thick, gray, erosion resistant unit composed of a series of even and parallel tabular beds that vary in thickness from a few millimeters to nearly a meter (Figure 4). Sand sized grains are sub-angular and granules are mainly pumice. W,e divide the tuff into three units; lower, middle, and upper (Figure 4, SA). These units are evaluated separately and then combined to construct a depositional model for the tuff bed. The lower tuff unit (0-515 cm) is composed ofparallel laminated thin (0.2 cm to 2 cm) beds (Figure 5B). These beds are massive with the exception of a small 3 cm trough cross bed and a 30 cm exposure ofripple cross lamination. There are two distinct bed types within the lower tuff unit based on grain size: uniform and bimodal. Uniform beds are sandy and very well to well sorted. Bimodal beds contain sand and fine pebble gravel (up to 2 cm) pumice grains. This rock fractures conchoidally and does not split evenly on bedding planes. The lower contact of the tuff is obscured. The middle tuff unit (515-580 cm) is well sorted, sub-angular, coarse grained, trough cross-bedded sandstone (Figure 5C). The bi-directional cross-bedding is truncated at a high angle and the contact with the underlying lower tuff is abrupt and erosionaL The upper tuff unit is subdivided into three sub-units, including a central fossil- bearing unit. The sub-units are: lower (580 to 615 cm), middle (fossil leafbearing) (615- 10 655 cm), and upper (655-750). The lower sub-unit of the upper tuff is 15 cm ofmassive siltstone overlain by 5 cm ofparallel laminated coarse grained thin beds similar to the basal tuff The lower contact is gradational and the siltstone has a few interbedded very fine sand beds with some lenses of coarser sand to granule material. The middle sub-unit is a massive siltstone and the contact between the lower and middle sub-units is abrupt. Figure 5: Photographs of tuff bed outcrop. A: outcrop with scale bar, B: lower unit with planar bedding, c: Lower/Middle contact with hammer on right for scale, D. Upper unit with hammer for scale. 11 The upper sub-unit is a medium grained paralle11aminated sandstone with beds varying from 0.5 cm to 15 cm thick. The contact between the upper and middle sub-units is abrupt and planar and the contact between the tuff and the overlying marine sands is loaded. 3.3. Tuff Outcrop Interpretation The sedimentary structures, rounded granules, predominantly sand sized sediments, and thin planar beds in the lower sub unit of the tuff unit indicate remobilization of sediment rather than direct pyroclastic base surge or airfall on land or into water. These sediments lack dune forms ofbase surge and graded beds typical of airfall into water (Cas and Wright 1987). The coarse sand is evidence that the sediment was remobilized through traction by water. Evidence ofbioturbation is absent indicating rapid deposition, and given the difficulty of fracturing the rock along bedding planes and thus the inferred absence ofparting lineation it is likely that the beds were deposited in the lower flow regime (Tucker 2003). Such criteria are met in the swash zone along the shore line (Clifton et aI1971). The trough cross bedding in the middle part of the tuff suggests complex dunes of channeled flow (Selley 2000). We interpret it to be the trough cross-bedded facies ofbarred nearshore system (Hunter et al. 1979) or a channel on a prograding delta (Reading and Collinson 1996). The massive siltstones are interpreted as lagoonal deposits that trapped fme-grained sediment and floating leaves. The association of swash zone, channeled flow, and lagoon indicate deposition in small barred estuary or a delta. 12 The abrupt shallowing that occurs at the base of the tuff as well as the thickness (7.5 m) of the tuff layer suggest that it was the product of at least a moderate size eruption. Grain size and uniformity suggest this tuffbed was in the distal zone of volcanic deposition (Vessel and Davies 1981, Scott 1988). The distal zone can extend from 40-120 km downstream from the volcanic source (Orton 1996, Smith 1988). Because of the abundant pumice, we infer that the tuffwas derived from a Plinian style volcanic eruption. Such an eruption would have provided a large amount of easily remobilized sediment to drainage networks supplying sediment to this shallow marine system. Such rapid inputs of sediment can quickly fill available accommodation space at the paleoshoreline and cause shallowing and progradation. Stream aggradation was observed at Mt. St. Helens fo llowing the 1980 eruption (Scott 1988) and in Guatemala following the 1902 eruption of Santa Maria. In Guatemala, volcanically derived sediment raised the river bed 10-15 m and an elongate deltaic platform prograded 7 km (Kuenzi et al. 1979). A similar process likely led to the shallow depositional environment that is recorded in the tuffbed in the sandstone ofFloras Lake. The aggraded primary stream may have dammed smaller drainages creating lakes. Such lakes may have assisted in the mechanical size sorting of the remobilized volcanic products by impounding the coarser fractions and leaving ash and leaves in suspension. When the lakes were subsequently breeched by headwall erosion or a large precipitation event, large plugs of ash and leaves could have been remobilized downstream. We call upon an increase sediment supply as the mechanism for shallowing as opposed to volcanically 13 induced uplift (Leithold and Bourgeois 1983) because the distance to a plausible source of the tuff is at least 100 km (see section 5.3). 4. 4°Ar/39Ar Dating of the Tuff 4.1. 40Ar/39Ar Methods Two fresh tuff samples were collected in the lower sub unit of the tuff and sent for analysis by Robert Duncan and John Huard at Oregon State University. A sample of plagioclase phenocrysts and hornblende phenocrysts each with a total mass of50-100 mg as well as FCT-3 biotite monitors (standard age of28.03 ± 0.18 Ma after Renne et a1., 1994) were wrapped in Cu-foil and stacked in evacuated quartz vials and irradiated with fast neutrons for 6 hr in the core of the lMW TRlGA reactor at Oregon State University. Incremental heating experiments were performed using the MAP2l5/50 mass spectrometer. The measured argon isotopes (40Ar, 39Ar, 38Ar, 37Ar, and 36Ar) were corrected for interfering Ca, K, and Cl, nuclear reactions (McDougall and Harrison, 1999) and for mass fractionation. Apparent ages for individual temperature steps were calculated using ArArCALC software (Koppers, 2002) assuming an initial atmospheric 4°Ar/39Ar value of295.5, and reported uncertainty (2a) includes error in regression of peak height measurement, in fitting the neutron flux measurements (J-values), and uncertainty in the age ofthe monitor. The 40Ar/36Ar normal isochron intercepts confirm that there was no excess Ar in the samples. Plateau ages are the average ofconcordant step ages, comprising most of the gas released, weighted by the inverse of their standard errors. The decay constants used were AE = 0.58lE-lO/yr and A~ = 4.963E-IO/yr. 14 4.2. 40ArP9Ar Results Ofthe two samples, only plagioclase yielded a reliable plateau age (Figure 6, Table 1). The hornblende age was unreliable because ofa low amount oforiginal K. The weighted plateau age ofplagioclase is 18.24 ± 0.86 with a MWSD of0.82. The plagioclase age values for each step are within the error range ofall other steps indicating a valid age determination. ~O.,..... ...... 70 Plagioclase Plateau Age: 18.24 ± 0.86 Ma 20 10 I It- L- o~ ......-... ..... o 10 2D 31) 40 W 60 70 8D .~ 10'2· Cumulative 39Ar Released (%) Figure 6: Plagioclase cumulative Argon release plateau. Weighted Plateau age 18.24 ± 0.86 Ma; Total Fusion 17.88 ± 1.03; Normal Isochron 16.28 ± 2.55; Inverse Isochron 18.45 ± 1.23; MSWD 0.82.The weighted plateau age is favored because it has the smallest error. 15 Table 1: Analytical Data for Radiometric Dates - Sample Plateau 2(J Total 20' MSWD Normal 2(J 40Ar/36Ar 2(J J age error Fusion error Isochron error normal error (Ma) Age age isochron IMa) (Ma) Plagioclase 18.24 0.86 17.88 1.03 0.82 16.28 2.55 309.66 172,16 0.0021259 Hornblende 11,43 6.46 13,78 9.82 0.26 5.89 14,8::· 301,58 1462 0.0021169 5. Provenance of the Tuff 5.1. XRF Methods The purpose of X-ray florescence analysis was to determine chemical differences between the tuff and the surrounding sandstone, and to identify a possible volcanic source for the eruption. We sampled the least weathered rock available within the tuff and approximately 10 meters above and below (Figure 3). The whole-rock samples were then analyzed by ALS Chemex of Vancouver, BC, using CANMET standard SDMS2 with FeO by Pratt titration. 5.2. XRF Results The X-ray florescence results in weight percentages were normalized for material lost on ignition (LOI) and results are shown in Table 2. There was a higher loss on ignition (LOI) from the tuff samples (sample numbers: 2,3,4 and 5), compared with marine sandstones. This difference can be related to devitrification and hydration of volcanic glass in the interval between airfall and redeposition and/or a greater amount of organic matter in the tuff. 16 Table 2: XRF Results' Sampte# Si02 A1203 Fe203 CaO MgO Na20 K20 Cr203 Ti02 MoO P205 srO BaO FeO Total LOI 6 71.61 14.07 4.15 2.31 2.26 2.42 2.42 0.01 0.51 0.02 0.09 0.04 0.07 3.01 93.66 6.32 5 70.42 15.94 3.41 1.46 2.59 3.61 1.S'4 0 0.4 0.03 O.OB 0.02 0.07 1.16 9'3.45 10.3 4 69.91 16.1 3.61 2.25 2.98 2.89 '1.59 0 0.46 0.03 0.08 0.03 0.06 '1.37 98.7 12.1 3 71.51 15.46 3.29 1.72 1.96 3.38 2.05 0 0.42 O.O~, 0.09 0.02 0.07 1.16 98.85 9.99 2 70.47 16.46 3.34 1.68 2.77 2.79 1.81 a 0.47 0..03 O.OB 0.02 0.07 '1.51 99.86 12.25 1 71.83 13.25 4.47 2.29 2.44 2.56 2.1'1 0.05 0.72 0.04 0.09 0.05 0.09 1.99 99.74 4.26 A.n. Er. 2.71 0.83 0.4 0.22 0.18 0.11 0.13 - 0..06 0.03 0.04 - - - nla 0.35 Tuff Ave 70.58 15.99 3.41 '1.79 2.58 3.16 1.85 0 0.44 0.04 0.08 0.Q3 0.07 '1.3 nia nl8 Tuff S.E. 0.33 0.2t 0.07 0.17 0.22 0.2 0 ..1 0 0.02 0 0 0 0 O.OB nia nia . •Analysis results with analytical error (An. Er.) reported from 10 replicate samples, mean of the four tuff samples (Tuff Ave.) and standard error of tuff samples (Tuff S. E.). The stratigraphic location of samples is shown on figure 3. Dashes indicate that analytical error data was not available. 5.3. XRF Interpretation The chemical analyses shows that the tuff is chemically distinct from the adjacent marine sandstones and allow for the classification of the tuff as a dacite (Figure 7) using the total alkali and silica method of LeBas et aL (1992). Change in the sediment supply system is observed by comparing the average and standard error of the four tuff values with the measurements from the adjacent sandstones. Between the tuff and the sandstone FeG were outside the 95% confidence interval of an average of tuff values indicating that the tuff inundated the source area and the change in sediment supplied to the system was FeG oxide values for the sandstone above the tuff were also significantly different indicating that once the volcanic material was flushed through, the system generally returned to pre-eruptive conditions. However some dilution of the sediments by 17 -'c;(t!)'w ;;n~lI~ Figure 7: Total Alkali-Silica diagram showing classifications from Le Bas et al (1992). Squares are tuff data, error bars are the analytical error, for the Na20 + K20 the error was summed. Circles are from Verplanck (1985) and triangles are from Mertzman (2000) and correspond with Cascade samples that have Ar/Ar dates that overlap with the Cape Blanco specimens. Errors were not reported and the plotted values are normalized to account for reported totals (Verplanck 1985) and LOI (Mertzman 2000). controlled redeposition ofsmall amounts of tuff did occur because the SiOz, Ah03, FeZ03, CrZ03, TiOz, PzOs, SrO and BaO, values for the sandstone above the tuff fall between the values for the tuff and the sandstone below. Thus, most ofthe tuffwas deposited without significant mixing with local sediment. However, tuffaceous sediment continued to be eroded into shallow marine sandstone after eruption diluting the concentration of elements in subsequent sediments. 18 Given age constraints for the tuff, its source was a part of the 100 km wide predominantly andesitic volcanic belt that spanned the western Cordillera ofNorth America (Christiansen and Yeats 1992), referred to in Oregon as the Early Western Cascades (Priest et al. 1983). Although data on early Miocene Cascade volcanism is limited (Sherrod and Smith 2000), however, information that may permit correlation is of two types. The first type is geologic maps that combine sparse radiometric dating with stratigraphic position to determine age. Rocks of early Miocene age are present in the southern Washington Cascades (Hammond 1980; Swanson 1989), but the most likely correlation is with the compositionally diverse but generally siliceous Roxy formation of the Southern Oregon Cascades (Hladky 1994, 1996, 1999a, 1999b) (Figure 2). The radiometric dates that constrain the age ofthe Roxy formation are 2-3 million years older than the Floras Lake tuff, though younger potentially correlative units may have been removed by erosion. Geochemical and geochronological studies (du Bray 2006, Hammond 1980, Mertzman 2000, Verplanck 1985) allow for identification of additional correlative units. We identity 10 samples (Table 3) from these studies that may correlate with the tuffbed. The criteria used for identification were that the samples have radiometric dates with two sigma error ranges that overlap with the Floras Lake tuffdate and that the samples are classified as andesite or dacite (LeBas et al1992). The possible correlations are all located within the Cascade Arc (Figure 2). 19 Table 3: Possible Volcanic Sources a ::.rror {2 er I (1\'13) "~ge Calculation Me:hod 0.2 K-Ar 0.2 K-Ar 0.7 K-Ar 08 K-Ar Sample ID Reference Age (Ma) M3-36 1 17.8 M3-38 '1 19.0 £17~':'5 2 17.6 97-54Z 2 17.1 S/7-,:'3 2 We. J1\>197-16 2 17.6 FRL #481~; (3) 3 18.4 FRL #4816 n)) ::, 18.9 FRL #4817 (3) ., HI.4..> FRL #4817 (I)) ::, 20.0 06 0.6 0.3 03 10 10 K-Ar K-Ar K-Ar K-Ar K-Ar K-Ar T.A.S ClassIfication (Iacjle dacite rhyolite andesite andesite dacite Ilia nla nla Illa a Samples that may correlate with the tuff in the sandstone ofFloras Lake. Reference 1 is Verplanck (1985), reference 2 is Mertzman (2000), reference 3 is Hammond (1980). Chemical analysis was not provided by Hammond (1980) the rocks are described as dacitic pyroclastic flow deposits. For Verplanck (1985) and Mertzman (2000) samples we normalized the geochemical analysis to account for material LOI and then determined the TAS classification Quaternary Plinian eruptions ofthe Cascades produced tephra fallout patterns primarily to the east and deposits rarely reach the coast (Sarna-Wojcicki et al. 1991). Sedimentary units that may correlate with the tuffbed are therefore more likely to be found on the leeward side ofthe Cascade range. Ofcomparable age is the Rose Creek Member of the volcaniclastic Eocene to Miocene John Day formation in Central Oregon (Albright et aI2008). The Rose Creek Member contains a number of ash fall tephra beds (Hunt and Stepleton 2004) that may have had the same source eruption as the tuff at Cape Blanco. Also ofnote is the amount ofchromium in the sandstones (Table 2). The amount of chromium in the adjacent sandstones is much higher than in the tuff. The presence of detrital chromite in sandstone has been used as evidence ofonset and development of collision in India and Pakistan because the uplift exposed ultramafic source rocks (Garzanti et a11996, Warwick et al. 1998). Elevated concentrations ofCr in mudstones 20 have also been used to identify ultramafic provenance (Garver and Scott 1995). The Klamath Terrane contains ultramafic ophiolites (Harper 1984) and the high Chromium in the sandstones of Floras Lake, excluding the tuff, indicates that the catchment area for the sandstone of Floras Lake likely included the Klamath Terrane. 6. Paleomagnetic Stratigraphy 6.1. Paleomagnetic Method To better refine the age and tectonic history of the sandstone ofFloras Lake, the entire marine sandstone section, including the tuff, was sampled at approximately 15 meter intervals (Figure 3) for paleomagnetic analysis. Oriented block samples were collected at 13 sites spaced approximately IS-meter apart through the 158 meters of section. The surface ofthe outcrop was scraped to remove the oxidized outer layer, and we also avoided hematite coated fractures. At each site three large block samples were collected with hand tools. The samples were then subsampled into 2.5-cm diameter cores using a drill press or by grinding down the more friable samples with sandpaper and a grinding wheel, then hardening the sample with dilute sodium silicate. From the three large block samples, a minimum of six to eight sub-samples were turned into cores for analysis, except for sites 41 and 42, which were so crumbly that only four sub-samples could be recovered for each site. Samples were analyzed on a 2G cryogenic magnetometer with a Caltech-style automatic sample changer in the Occidental College paleomagnetics laboratry. After measurement ofnatural remanent magnetization (NMR), each sample was demagnetized 21 in alternating field (AF) of2.5, 5.0, 7.5, and 10.0 mT (millitesla). This enables determination of the coercivity behavior of each sample individually, and demagnetizes multi-domain grains before their magnetism is locked in by heating. After AF demagnetization, each sample was then thermally demagnetized at 50°C steps from 100 to 650°C. This removes overprinting by iron hydroxides such as goethite (which dehydrates to hematite at 200°C), and determines how much of the remanence is held by magnetite and hematite as the blocking temperature of magnetite (580°C) is exceeded. About 0.1 g ofpowdered samples of a number ofrepresentative lithologies were placed in Eppendorftubes and subjected to increased isothermal remanent magnetization (IRM) to determine their IRM saturation behavior. These same samples were also demagnetized twice, once after having acquired an IRM produced in a 100 MT peak field, and once after having acquired an anhysteretic remanent magnetization (ARM) in a 100 mT oscillating field. Such data are useful in conducting a modified Lowrie-Fuller test (Pluhar et al. 1991). Results were plotted on orthogonal demagnetization ("Zijderveld") plots, and the average direction of each sample was determined by the least-squares method of Kirschvink (1980). Mean directions for each sample were then analyzed using Fisher (1953) statistics, and classified according to the scheme of Opdyke et al. (1977). 6.2. Paleomagnetic Results Representative orthogonal demagnetization plots and are shown in Figure 8. Most of the samples showed a dip-corrected NRM direction either northeast and down (a normal 22 direction rotated clockwise) or southwest and up (a reversed direction rotated clockwise), and most ofthe samples showed only this single component as their vectors decayed to the origin. A small number of samples showed some overprinting that was typically removed by 300°C, and then a single component ofremanence that decayed to the origin. The fact that nearly every sample showed a significant drop in intensity during AF demagnetization indicates that most of the remanence is held by a low-coercivity mineral such as magnetite; this is consistent with the fact that nearly all the remanence was lost when the blocking temperature of magnetite (580°C) was exceeded. Thus, it appears that the samples show a primary or characteristic remanence that is rotated counterclockwise. The IRM acquisition experiments indicated that the samples contained significant amounts ofmagnetite, as the IRM saturated at 300 mT. The Lowrie-Fuller tests indicated that the grains in the sample were single-domain or pseudo-single-domain, because the ARM was more resistant to AF demagnetization than the IRM. The mean directions of all 13 sites were significantly removed from a random distribution at the 95% confidence level, or Class I sites in the system of Opdyke et al. (1977) (Table 4). The mean direction for all the normal samples (n = 34) was D = 49.7, I = 41.0, k = 14.7, (X,95 = 6.6; for the reversed samples, the mean (n = 37) was D = 208.9, I = -42.3, k = 6.0, (X,95 = 10.5 (Table 4). These mean directions are antipodal within the error estimates, giving a positive reversal test and showing that the remanence is primary with most of the overprinting removed (Figure 9). 23 '" • :-~~ .• "\. , /." ,.I \ " •~,/ '" .[ /" "E " ,. • ~. , I q. , ,l \ \,, f.... t~., t~-·-+-S I: '+1 " \ \ ~I~. ,\ " \ '"I " ,t,.-", , n ___ ... I ., --11 h \ \ \ \ \ "\ :'1 \,~ ~// ~-'. ·'w Figure 8: Orthogonal demagnetization ("Zijderveld") plots showing representative results. Solid circles = declination (compass direction); open circles = inclination (up-down component). First step is NRM, followed by 25,50, 75 and 100 mT AF, then thermal steps from 100° to 650°C in 50°C increments. Top two plots are normally magnetized with slight clockwise rotation; bottom two are reversely magnetized, also rotated. Most of the remanence is held in magnetite, based on low coercivity and loss ofmagnetization above the Curie point ofmagnetite (580°C). 24 Table 4: Paleomagnetic Data a SITE N D I k '2/.'?-5 30 7 233.0 -46.9 9.4 207 31 6 216.5 -41.2 7.9 25.4 ~·1 S 45.2 45.6 l' "l 158)~ 5.", 33 6 419 43.0 114 20.7 34 S 381 4::·.5 53.3 7.7 35 7 239.0 -44.8 7.9 22.9 36 S 1919 -58-4 3.8 32.6 37 S 192.1 26.9 16 14.3 38 6 59.2 33.2 15.8 17.4 39 6 205.9 -35 6.5 28.3 40 4 206.1 -322 11.2 28.8 41 4 235.0 -501 9.2 32 42 6 64.9 31.1 15.4 17.6 a Paleomagnetic data and Fisher statistics. N = number of samples; D = declination; I = inclination; k = precision parameter; U95 = ellipse of95% confidence around mean. Plotting the magnetic results on the stratigraphic section (Figure 3) yields a magnetostratigraphy with five polarity zones. The basal 40 m of section has reversed polarity, followed by the next 40 m ofsection consisting of rocks wish normal polarity. From 85 to 115 m in the section, the rocks have reversed polarity. Between 115 and140 m in the section (including the dated tuft), the rocks have normal polarity. The uppermost 15 m of section have reversed polarity. Thus we have a R-N-R-N-R sequence of five roughly equal-length polarity zones spanning 158 m ofsection, and calibrated by the dated tuffnear the top. 25 Figure 9: Stereonet ofpaleomagnetic results showing mean and ellipse of95% confidence ofnormal samples (solid dot and circle) and reversed samples (open circle and dashed line). The latter is inverted through the center of the stereonet (solid square and circle), showing that the directions are antipodal, and that the overprinting has been removed. Inverting the reversed directions and averaging all the vectors yields a formational mean direction ofD = 38.0, I = 43.6, k = 6.7, U95 = 7.1 (n = 71). This mean is rotated clockwise by about 37.6° ± 7°, degrees from the expected early Miocene cratonic pole (Diehl et aI., 1983). This is much less than the 105 ± 5° reported from the middle Eocene Coaledo Formation in nearby Coos Bay, Oregon (Prothero and Donohoo, 2001 a), and 106 ± 18° for the lower Oligocene Tunnel Point Formation overlying the Coaledo Formation in Coos Bay (Prothero and Donohoo, 2001b). However, it is much more than the 9.0 ± 8° of clockwise rotation reported from the upper Miocene (6-8 Ma) Empire Formation in Coos Bay (Prothero et aI., 2001). Thus, the early Miocene rotation of about 38° clockwise at Cape Blanco is intermediate between the large (greater than 90°) 26 clockwise Eocene-Oligocene rotations and the small late Miocene rotations reported from the region. 6.3. Paleomagnetic Interpretation The plagioclase 40Ar;39Ar crystal yielding the 18.24 Ma ± 0.86 Ma came from the fourth polarity zone and may correspond with normal polarity chrons C5Dn, C5Dr.ln, C5En, or C6n (Ogg and Smith. 2004). To determine the preferred age assignment the deposition rate was estimated by dividing the thickness (m) ofpolarity zones 2,3, and 4, 42.5m, 32.5m, and 26.5m, respectively, by the duration (Ma) of the possibly correlating chrons. The 23,000 year chron, C5Dr.ln, produced deposition rates an order of magnitude greater than the adjacent zones and therefore the reversal was likely not captured by the coarseness ofour sampling. The other three chrons all yielded reasonable deposition rates and our preferred correlation for zone four is with polarity chron C5En because this chron overlaps with the 40Ar/39Ar date. 6.4. Paleomagnetic Discussion Cenozoic rotation of the Oregon coastal block has been recognized for some time (Simpson and Cox, 1977), and generally the amount ofrotation decreases through the Tertiary (Beck and Plumley, 1980; Magill and Cox 1980; Prothero and Donohoo, 2001 a,b; Prothero et aI., 2001). The tectonic mechanism driving deformation has been attributed to oblique subduction of the Juan de Fuca and associated microplates and expansion of the Basin and Range Province (Wells 1990; Wells and Heller, 1988; Wells 27 and Simpson 2001; Wells et aI, 1998). The rotations observed in our paleomagnetic data are consistent with previously documented clockwise Oregon coastal block rotations. But given the unlikelihood that Cape Blanco is underlain by Siletzia, the post Miocene motion observed at Cape Blanco may reflect dextral shearing (Wells 1990) as opposed to rigid block rotation. Prothero and Donohoo (200la) also suggested dextral shearing as the rotational mechanism for Eocene deposits 60 km north of Cape Blanco at Coos Bay, Oregon. These long-term geologic observations are supported by modern GPS studies that parse the observed North American plate motion (McCaffrey et al. 2000, Murray and Lisowski 2000, Savage et aI, 2000) into interseismic strain caused by a locked zone in the subducting slab and a residual velocity which is consistent with clockwise rotation (Miller et al. 2001) about a Euler pole located along the eastern portion of the Washington-Oregon border (Lewis et al. 2003). The depositional rates inferred by sediment thickness and the assigned polarity zones vary between 50 and 150 meters per million years. This rate is consistent with that predicted by Leithold and Bourgeois (1983). 7. Conclusions Our analysis ofthe sandstone ofFloras Lake and the tuffbed within it shows that the shallow marine depositional environment for the sandstone beds (Leithold and Bourgeois 1983) was interrupted by a transient deltaic progradation ofredeposited volcanic material which contains the Cape Blanco flora. Isotopic data constrain the age of the plant-bearing tuffto 18.24 ± 0.86 Ma, and we interpret this age as an age of 28 eruption with near instantaneous redeposition. The eruption took place in the Cascade are, but a precise source remains undetermined. Paleomagnetic data confIrm that the area was rotated clockwise after deposition, likely by dextral shear. The relation between the early Miocene Cascade volcanic arc and the Klamath Terrane was fixed since the early Miocene (Magill and Cox 1977; Wells and Heller 1988) and given the Cascade source and the high Cr203 in the sandstones it is likely that the source area for the sandstone of Floras Lake extended across the Klamath Terrane which may have been much lower relief at that time (Aalto 2006). The>100 km distance between the likely eruptive center and the site ofdeposition means that the leaves ofthe Cape Blanco flora capture the floral diversity and paleoclimate on the coast west ofthe early Miocene Cascade arc. The fo Howing chapter describes the fossils recovered from the tuff bed. 29 CHAPTER III SYSTEMATICS OF THE EARLY MIOCENE CAPE BLANCO FLORA OF COASTAL OREGON Greg Retallack assisted in all facets ofthe work, including idea development, data collection, data analysis, and editing. INTRODUCTION Coastal location and radiometric age of 18.26 ± 0.86 Ma (Chapter II) make the Cape Blanco flora important for paleoclimatic and floristic assessment, because previously documented Miocene floras of Oregon, USA, (Table 1) are inland. Coastal records are important because their paleoelevation is constrained and they provide a link between marine records ofglobal climate (Zachos et aI., 2001) and inland climate records (Wolfe, 1994; Retallack, 2007). Both records of climate indicate a time of global cooling from late early to middle Miocene peak temperatures. Regional reconstruction of paleoclimate is important to the testing of climate models (Wing and Greenwood, 1993; Wing et aI., 2005, Sewall and Sloan, 2006) as well as tectonic reconstructions that call upon climate as a potential driver (Ruddiman, 1997). Furthermore increasing temperature resolution at sea level allows for more accurate interpretation of terrestrial 30 lapse rate based studies ofpaleoelevations (Axelrod, 1968, Forest et al. 1999) and better reconstructions of paleo latitudinal temperature gradients (Miller et aI., 2006). Table 1: Miocene Floras of Oregon ab Flora Reference Age: From listed reference Status Lola Pass 17 Latest Miocene b Alvord Creek' 4 Late Miocene a,b,d Unity '13 Late Miocene a,b,d Vibbert 1 Late Miocene a,b,d Mascall • 5,9 Mid-Late Miocene a,b,d Troutdale' 4,16 '10 a,b,c Faraday 16,17 11 1012 b,c Liberal 16 n b,c Molalla 16 13 b,c Vinegar Creek 17 Late middle to early late Miocene Weyerhauser 16, 17 14-18 b,c Blue Mountain' 5,12 Mid Miocene a,b,d Cape Blanco 6,16 Mid Miocene a,b,c Jamison 17 Early to Mid Miocene b Baker 17 Early to Middle Miocene Hidden Lake 17, Mentioned in 7 Late Early the Middle Late b Little Butte Creek 17 Late early to early middle Miocene b Skull Spnng 17 Late early to early middle Miocene Sparta 17, Mentioned by 7 Late early to early middle Miocene b Beulah 2 Miocene (UCMP database) b,d Pelton 1 Miocene a,b,d Stinking Water' 5 Miocene a,b,d Succor Creek • 7,8,11,14 Miocene a,b,d Trout Creek' 5,8,10 Miocene a,b,d Eagle Creek' 16 18-22 b,c Collawash • 16 18-22 b,c Forman Point 1 Early Miocene b FIsh Creek Road 7 ? Cascadia 7 ? Sandstone Creek Mentioned in 7, 15 ? • Multiple localities a References: I (AshwillI983); 2 (Axelrod 1964); 3 (Chaney 1938); 4 (Chaney et a1. 1944); 5 (Chaney and Axelrod 1959); 6 (Emerson and Retallack 2005); 7 (Fields 1996); 8 (Graham 1963); 9 (Knowlton 1902); 10 (MacGinitie 1933); II (Niklas et a1. 1978); 12 (Oliver 1936); 13 (Retallack 2004); 14 (Smith 1932); IS (Wolfe 1981); 16 (Wolfe 1994); 17 (Wolfe and Tanai 1987); b Status: a) location described; b) location shown on map; c) published climatic information; d) published taxonomy 31 GEOLOGICAL SETTING Cape Blanco is located along the southern Oregon coast (Figure 1) and sea cliffs south of the cape expose a southeast dipping section of Cenozoic sedimentary rocks 124"35' 'N Pacific Ocean cape Blanco Oregon. USA 123"35' ",IV Legend 'I- .~,n1icline C/U F..ul: • • 1 k'll Figure 1: Location map showing location offossillocality (star). Modified from Kelsey (1990) and Addicott (1980). For detailed bedrock geology refer to Dott (1962) and for detailed terrace maps refer to Kelsey (1990). (Addicott, 1980). The unconformity bound, early Miocene, sandstone of Floras Lake contains a 7.5 m thick rhyodacite tuff bed with leaf bearing beds (Figure 2). The 32 pumiceous tuff is comparable with others from a Plinian eruptions in the Western Cascade Arc. Planar bedding and <1 meter scale trough cross bedding within the tuff are evidence that it was deposited at sea level (Emerson and Retallack 2007). The source 750 em 580 em -t.~.:~~;~~:"?:""'::\""'\:;C7i;;~!J 515 em k=-:-=-J Silt~tone (Lagoon Facies) [;.:.:o;.,:~:,:~. Crossbedded coarseIrolr"""··';';:.i:i:;·:~ sandstone I;Channei Facies} !::i::~'~:::':~ Para:lel1hin beds{:':/+::: {Swash zone Facies) ~ Fossil Leaves Figure 2: Stratigraphic section of tuff bed with three zones; upper, middle,and lower. Leaf symbol indicates location of fossil bearing unit. eruption may have dammed local catchments. When these lakes and washes were drained, the plug of ash and leaves was flushed to the shoreline where the fme grained, 1eafladen material settled in a small oxbow lakes on a delta or in lagoons adjacent to 33 barrier bars. The Cape Blanco flora was discovered and briefly characterized by Wolfe (1994) and noted by Leithold and Bourgeois (1983). MATERIALS AND METHODS Megafossil Remains Abundant new material (687 specimens) was collected for this study from a single quarry in six trips totaling 102 man hours. Rock was removed from the quarry in ~ 40 cm x 40 cm x 40 cm blocks and then broken down into fist sized pieces. The tuffwas difficult to break and the leaves were randomly oriented. The majority of leaf fossils were within a single massive, 40 cm thick grey shale and an underlying 10 cm of coarse sand to granule pumiceous tuffbed. This precluded stratigraphic investigations finer than the bed thickness. Coniferous needle litter was predominantly found in the coarse grained, basal, pumice conglomerate. Fossils that had promise of identifiable features were collected and brought to the laboratory for further analysis. In the laboratory, morphotypes were created to document diversity in the flora. The morphotypes were based on the part of the leafpreserved (eg. apex, base, margin), level ofpreservation, venation, size, and margin state. As the fossils were considered in the laboratory each was assigned to a morphotype and a total of 115 morphotypes were established. To ensure that the full diversity ofthe assemblage was captured we created a rarefaction curve (Figure 3) to show the rate ofnew morphotype establishment. When additional material failed to warrant the establishment ofnew morphotypes, we were satisfied that we had captured the diversity present in the flora. Well preserved 34 morphotypes were subsequently assigned fossil names, and in many cases morphotypes were merged where transitional examples warranted. Rarefaction Curve "140 11l ~ Q. "120~ 0 100£. E- o 80~ '0 60 ~ ~ .0 40E :J 20z 0 I J ..-r r, I I, a '100 200 300 400 500 600 700 800 Number of Specimens Figure 3: Rarefaction curve documenting the decrease in new morphotype discovery with increasing specimens. The flat portion ofthe graph (arrow) demonstrates that the morphological diversity present in the flora was captured by our collections. Photographs for figs 4-10 were taken using a Canon EOS 10D camera with a macro lens. Images were converted to grey scale and brightness and contrast were adjusted using Adobe Photoshop 6.0. No new species were described as a part of this study. Leaf forms that could not be assigned to a fossil name, but were unique, were assigned numbers and are described 35 at the end of the section. Specimens included in this study are from two collections; the majority were collected by the authors, as described above, and are housed in the Condon Collection of the Museum ofNatural and Cultural History, University ofOregon (UO specimen numbers). The Condon Collection specimens were supplemented by 73 specimens from the Wolfe collections on loan from the Smithsonian, Washington D. C. (W specimen numbers) and are noted when figured. We retain the usage of fossil names in this study and do not attempt to unambiguously associate the fossil species with a modern equivalent. However, we recognize the similarity of fossil species with extant plants and rely on the nearest living relatives assigned by previous workers to aid in the floristic interpretation of the flora. The systematic assignments are organized by order and the families are assigned to orders following USDA (2009). SYSTEMATICS Order- Equisetales Family-Equisetaceae Michx. ex DC. Genus-Equisetum sp. Referred specimens- UO F 38064 (FigAA); UO F 38088A (Fig. 4B) Description- Equisetum is recognized by longitudinally ridged axes, separated by nodes. Only stems were found. Stems were 1-2 em wide with nodes every 2.5 em. 36 Order- Pinales Family-Pinaceae Genus-Pinus Species-Pinus tiptoniana Chaney and Axelrod, 1959, p. 142; pI. 13, figs. 6, 7. Referred specimens- UO F 38089A (FigAC). Description-Pinus tiptoniana is represented by a single specimen which we identifY as a two needle bundle because of thickening at the base that we interpret as a fascicle. The needle is 15 cm long and each needle is 2 mm wide. Longitudinal ridges run the length of the needle. This is similar to Pinus ponderosoides Axelrod because both are long needled pines, however Pinus ponderosoides has three needles per bundle and a more prominent fascicle. Uncertain Affinity- Cone unit Referred specimens- UO F 38090A (FigAL). Description- The shape of this specimen is similar to a cone unit ofa conifer such as Pinus but is too abraded and frayed for secure identification. Uncertain Affinity- Charcoal Referred specimens- UO F 38091 (Fig. 4D), UO F 38093 (Fig. 4G), UO F 38094 (Fig. 4M), UO F 38095 (Fig. 4F). Description- Many rounded charcoal pieces up to a few centimeters in the long axis direction were found in the coarse grained layer. Vessels could not be identified in .&1.,,0 , JIIIII ----------~ 37 Figure 4: Cape Blanco fossils, orders Equisetales and Pinales, scale bar is I em. A. Equisetum sp. UO F 38064 B. Equisetum sp. UO F 38088 A, arrow at node. C. Pinus tiptoniana UO 38089A. D. Charcoal, UO F 38091. E. Shredded wood, UO F 38104. F. Charcoal, UO F 38095. G. Charcoal, UO F 38093. H. Needle, UO F 38114. I. Shredded wood, UO F 38101B. J. Shredded wood, UO F 38103. K. Needle, UO F 42033A. L. Cone Unit, UO F 38090A. M. Charcoal, UO F 38094. N. Splayed needles, UO F 38105B. 38 the samples suggesting it is conifer wood. There was a range in sizes from a few cm, DO F 38093, to a piece that was larger than 10 cm, DO F 38092. Uncertain Affinity- Needles Referred specimens- DO F 38114 (Fig. 4H), DO F 42033A (Fig. 4K). Description- Many disaggregated pieces of coniferous debris were found in the coarse grained, basal, pumiceous layer. A few 1-2 cm solitary needles with a midrib, DO F 38114 and DO F 42033, resemble Abies needles although limited material precluded specific assignment. Other, larger needle pieces may be fragments ofPinus tiptoniana. Uncertain Affinity- Shredded wood Referred specimens- DO F 38101B (Fig. 41), UO F 38103 (Fig. 4J), DO F 38104 (Fig. 4E). Description- Compressed wood is dark and thick, but lacks the equant rounded shape of the charcoal morphotype. The ends of the pieces are frayed giving the wood a shredded appearance. Modem wood of cedar (Thuja), and pine (Pinus), shreds in a comparable way as it weathers. Uncertain Affinity- Splayed needles Referred specimens-DO F 38105B (Fig. 4N). Description- These specimens have needles attached to a stem and are likely part of a conifer short shoot, similar to Douglas fir (Pseudotsuga menziesii), and spruce (Picea). 39 Order-Typhales Family-Typhaceae Uncertain Afjinity- Typha sp. Referred specimens- DO F 38108A (Fig. 5A), DO F 38110 (Fig. 5B). Description- These specimens showed longitudinal parallel venation and did not have relief No connected stems, nodes, or sheaths, characteristics ofPoaceae, were found. The laminae were 2-5 mm wide. Order- Salicales Family- Salicaceae Genus-Populus Species- Populus eotremuloides Knowlton. Populus eotremuloides Knowlton, 1898, . 725; pI. 100, figs. 1, 2; pI. 101, figs. 1, 2. Populus eotremuloides Chaney and Axelrod 1959, p 151; pI. 17, fig 4. Populus eotremuloides Axelrod 1991, P 41; pI. 6, figs. 1, 3, 8. Populus eotremuloides Buechler et a~ 2007, p 322; Fig. 3C. Referred specimens- DO F 38118A (Fig.5C), DO F 38119A (Fig. 5D), DO F 38121 (Fig. 5E), DO F 38120 (Fig. 5F). Description- Suprabasal actinodromous primary venation is a diagnostic character ofPopulus. The two Populus species observed at Cape Blanco are distinguished by the margin with P. eotremuloides tending toward an entire to very fme1y serrate margin and P. lindgreni having a crenate margin with pronounced teeth. This 40 corresponds with the serration observed on the often cited nearest living relative of this species, Populus trichocarpa, which has since been revised to, Populus balsamifera L. ssp. trichocarpa (Torr. & A. Gray ex Hook.)(USDA 2009). Many leaf fragments were also assigned to this species because of the unique margin, apex angle, and venation pattern. Family- Salicaceae Genus-Populus Species- Populus lindgreni Knowlton. Populus lindgreni Knowlton, 1898, p. 725; pI. 100, fig. 3. Populus lindgreni Chaney and Axelrod, 1959, p. 151; pI. 17, fig. 1-3. Populus lindgreni Buechler et aI, 2007, p. 322; Fig. 3E. Referred specimens- W 50 (Fig. 5G), W 51 (Fig. 5H). Description- Buechler et aI. (2007) included Populus voyana Chaney and Axelrod (1959) in P. lindgreni because the evidence used by Chaney and Axelrod (1959) to split P. voyana from P. lindgreni, petiole and midrib thickness, could be explained by natural variation within the species. Fields (1996) did not document P. lindgreni at Succor Creek, but in his discussion ofP. voyana he recommended the separation. We assigned our specimen to the priority name, P. lindgreni because our specimen, particularly the teeth, most closely resembles P. lindgreni Chaney and Axelrod (1959 pI. 17, fig 2). We follow Chaney and Axelrod (1959) who suggested that the nearest living 41 Figure 5: Cape Blanco fossils, orders Typhales, Salicales, and Fagales, scale bar is 1 em. A. Typha sp., DO F 38108A. B. Typha sp., DO F 38110. C. Populus eotremuloides, DO F 38118A. D. Populus eotremuloides, DO F 38119A. E. Populus eotremuloides, DO F 38121. F. Populus eotremuloides, DO F 38120. G. Populus lindgreni, W 50. H. Populus lindgreni ,W 51. I. Salix laevigatoides DO F 38136A. J. Salix laevigatoides W 7. K. Salix succorensis W 29. L. Salix succorensis W 68. M. Alnus harneyana DO F 42024B. N. Alnus harneyana DO F 42025A. 42 relative to P. lindgreni may be P. heterophylla Linaeus, and also concur with Fields (1996) that P. voyana is more like the extant aspen, P. tremuloides. Family- Salicaceae Genus-Salix Species- Salix laevigatoides Axelrod Salix coalingensis Dorf, 1930, p. 78-79; pI. 7, fig 6 only. Salix laevigatoides Axelrod, 1950, p. 55; pI. 2, fig. 10. Salix laevigatoides Axelrod, 1991, p. 43; pI. 8, figs. 3, 4. Salix laevigatoides Axelrod, 1995, p 45; pI. 11, figs. 6, 7. Referred specimens- DO F 38l63A (Fig. 51), W 7 (Fig. 5J) Description- The diagnostic characters ofSalix are lanceolate leaves with semicraspedodromous to brochidodromous secondary venation and reticulate tertiary venation. Many Salix species present in Miocene floras have a serrated margin. Axelrod (1995, p. 44) provides a key to a few ofthe Miocene Purple Mountain species. We follow that key and assign the entire and erose margined lanceolate Salix specimens from Cape Blanco to S. laevigatoides with the nearest living form Salix laevigata Bebb of western North America. When originally described, Dorf(1930, pg. 78) illustrated five leaves ofS. coalingensis. These fossils were split by Axelrod (1944a, pg 133-134) into Persia coalingensis (Dorf) Axelrod (Dorf 1930; pI. 7, fig. 7; pI. 8, figs. 1,2), Salix hesperia (Knowlton) Condit (Dorf 1930, pI . 7, fig. 6) and Salix wildcatensis Axelrod (Dorf 1930, 43 pI. 7, fig. 5). Later, Axelrod (1950, pg 55) reassigned one ofDorf's specimens (1930, pI. 7, fig. 6) to Salix laevigatoides Axelrod when he established the name Salix laevigatoides. No fonnal description was provided for the new name and therefore the only published fonnal description for this species is the one by Dorf (1930). The Dorf description applies to more material than is currently included in Salix laevigatoides and so we provide a supplemental diagnosis below, relying heavily on the original description ofSalix coalingensis Dorf (1930) and the discussion ofAxelrod (1991, pg. 43). Supplemental Diagnosis- Leaves rather firm in texture; margin chiefly entire, may be fmely serrate; leaf narrowly lanceolate; 3.5 to 11.4 cm long; 1.2 to 2.6 cm wide; elliptic to ovate; base cuneate to acute; apex acute to attenuate; midrib strong; secondaries 10 to 18 pairs, thin, alternate, arising from midrib at high angle, secondary veins brochidodromous to semicraspedodromous; tertiary veins orthogonal reticulate. Family-Salicaceae Genus-Salix Species- Salix succorensis Chaney and Axelrod Salix succorensis Chaney and Axelrod, 1959, p. 154; PI. 16, fig. 8. Salix succorensis Buechler et aI., 2007, p. 323; Fig. 4B, G. Referred specimens- W 29 (Fig. 5K), W 68 (Fig. 5L). Description-The large length to width ratio, size, entire margin, and 45° angle of secondary departure are characters unique to Salix succorensis. The specimens from Cape Blanco are a very good match for the leaf figured by Buechler et aI. (2007), which 44 has an entire margin. This echoes the comments of Fields (1996, p 411) who describes the margin as "faintly or non-toothed". However the figured specimen of Chaney and Axelrod (1959) shows teeth. The Cape Blanco specimens are entire. Modem species most closely resembling S. succorensis are S. nigra Marshall, and S. longipes Anderson (Chaney and Axelrod 1959), which was later reassigned to S. caroliniana (USDA 2009) Order- Fagales Family-Betu1aceae Leaves of the Birch family have compound serrate margins and straight secondary vems. The primary obstacle to proper name assignment in the Cape Blanco specimens is the lack of complete specimens. Here we differentiate large leaves with widely spaced secondaries, and three-ordered serrate margin into Alnus harneyana Chaney and Axelrod and the small leaves into Betula thor Knowlton. The genera Alnus and Betula were distinguished by Chaney and Axelrod (1959) as follows, "(a) the leaves ofAlnus are characterized by strong subsecondaries which diverge from the abaxial side of the secondaries in the outer part ofthe blade, whereas if subsecondaries are present in Betula they are only weakly developed; (b) in leaves ofcomparable size, Betula has more numerous and more closely spaced secondaries than Alnus; (c) in shape Alnus is generally wider in the lower halfof the blade, not near the middle of the blade like Betula; (d) the marginal outline is more even in Betula than in Alnus."(page 158). Axelrod (1985, pg. 138-140) further stresses the need to consider the whole leaf to make assessments, but laments that complete Betulaceae material is rarely found. Fossil 45 Betulaceae have been much debated in the literature (Fields 1996, pg. 380; Axelrod 1985, pg 138; Wolfe 1966, pg. B15) although a clear review ofthe family has yet to be presented. Our fossils leaves are membranaceous and many are tom, indicating less resistance to transport than associated leaves. Genus-Alnus Species-Alnus harneyana Chaney and Axelrod Alnus harneyana Chaney and Axelrod, 1959, p. 158-159: pi. 21, figs. 3-9 Alnus harneyana Axelrod 1985, p. 140-141; pl. 7, fig. 5 Referred specimens- UO F 42024B (Fig. 5M), UO F 42025A (Fig. 5N). Description- A subcoriaceous to membranaceous texture, straight and uniform secondaries, and toothed margin all mark this species. The fine details ofvenation were preserved in the Cape Blanco specimens, but no complete leaves were recovered. Many leafbase fragments were found that match the description ofChaney and Axelrod (1959, pg. 158). The extant species suggested by Chaney and Axelrod (1959) and Axelrod (1985) as closely resembling Alnus harneyana is Alnus tenuifolia. This name is now associated with the modem species, Alnus incana (L.) Moench ssp. tenuifolia (Nuttall) Breitung (USDA 2009) and its range is western North America. The other subspecies, Alnus incana (L.) Moench ssp. rugosa (Du Roi) R. T. Clausen, has single serrate margins and an eastern North America range. The morphological diversity observed in the sub- 46 species level underscores the difficulty in assigning fossils to species in the same way that we differentiate modem plants. There is a discrepancy in the description of Chaney and Axelrod (1959, pg. 158) and modem tenninology (Hickey, 1973) regarding the definition oflobed. Chaney and Axelrod (1959, pg. 158) describe Alnus harneyana as, "margin sinuate with the lobes mostly rounded but varying to subacute, the marginal lobes with forward-pointing serrate teeth varying from 2 to 4 (averaging 3) on lower side oflobe and usually only I distally." However the correct description ofthe marginal feature is compound serrate (following Hickey, 1973). To the original description we can add the following observations ofleaf margin: toothed; serrate; with three orders convex ofteeth, but teeth may have two or four orders; apical angle obtuse. Genus-Betula Species-Betula thor Knowlton Betula thor Knowlton 1926, p. 35; pI. XVII, fig 3 Betula thor Chaney and Axelrod 1959, p. 160-161; pI. 23, figs. 2-6 Betula thor Wolfe 1964, p N21; pI. 1, fig. 14 Betula thor Axelrod 1985, p 141-142; pI. 23, figs. 3, 6 Betula thor Axelrod 1991, p 47-48; pI. 9, figs. 1-3,5, 7 Betula thor Axelrod 1992, p 35-36; pI. 8, fig. 3; pI. 12, figs. 1,2 Betula thor Axelrod 1995, p 47; pI. 14, fig. 5, pI. 15, fig. I Referred specimens- DO F 42027B (Fig. 6A) 47 Description-Small size and serrate margins characterize this species. The living plants most similar to Betula thor are Betula papyr(fera Marshall and Betula occidentalis Hooker. Chaney and Axelrod (1959) also suggest a resemblance with Betula nigra Linnaeus. Combined the range of these three modem species spans North America. The modem species grow in streamside environments that are moist year round (Axelrod 1991). Family- Fagaceae Oak and beech leaves from Cape Blanco are diverse and distinct, without intermediate forms. In some cases two or more of the distinct forms have been included within the same fossil species as documented by Fields (1996) in his discussion of the Quercus simulata morphoplex. Genus- Chrysolepis Species- Chrysolepis sonomensis (Axelrod) Axelrod Castanopsis sonomensis Axelrod 1944b, p. 196 Chrysolepis sonomensis Axelrod 1985, p. 144, pI. 10, figs. 1-3 Chrysolepis sonomensis Axelrod 1991, p. 49, pI. 14, figs 1,2 Chrysolepis sonomensis Axelrod 1992, p. 37, pI. 10, figs 7, 8 Referred specimens- UO F 42054A (Fig. 6C), UO F 42053 (Fig. 6D) Description- This entire margined, medium sized leaf with a thick midrib and clear tertiary venation is common in the Cape Blanco flora. The most similar modem species suggested by Axelrod (1985) for this fossil species is Chrysolepis chrysophylla 48 Figure 6: Cape Blanco fossils, order Fagales and Juglandales, scale bar is 1 Cill. A. Betula thor, VO F 42027B. B. Carya bendirei, VO F 42029B. C. Chrysolepis sonomensis, VO F 42054A. D. Chlysolepis sonomensis, VO F 42053. E. Fagus washoensis VO F 42077A. F. Fagus washoensis VO F 42078. G. Lithocarpus nevadensis VO F 42088. H. Lithocarpus nevadensis VO F 42087. I. Quercus dayana VO F 42105A. J. Quercus hannibali VO F 42122. K. Quercus hannibali VO F 42115. L. Quercus simulata W 66. M. Quercus prelobata VO F 42l38A. N. Castanopsis perplexa VO F 38056. O. Quercus simulata VO F 42142A. 49 (Douglas ex Hook.) Hjelmquist var. chrysophylla (USDA 2009) which ranges from California to Oregon and is associated with redwood forests. Genus- Fagus Species- Fagus washoensis LaMotte Fagus washoensis Chaney and Axelrod, 1959, p. 164; pI. 25, figs. 7-10 Fagus idahoensis Chaney and Axelrod, 1959, p. 163; pI. 25, figs. 1-6 Referred specimens- UO F 42077A (Fig. 6E), UO F 42078 (Fig. 6F) Description- Fields (1996) synonymized two Fagus species based on the existence ofa continual size gradation between them, which contradicted separation of the smaller Fagus idahoensis by Chaney and Axelrod (1959). Most of the specimens from Cape Blanco are of the small variety, but we follow Fields (1996) for clarity. The Cape Blanco specimens are signally serrated with regular tooth spacing; apical and basal sides of the teeth convex; sinus rounded; apex simple. Secondaries diverge from primary from 30-60° and are craspedodromous straight to tooth apex. The most similar modem plant is Fagus grandifolia Ehrhardt. Genus- Lithocarpus Species- Lithocarpus nevadensis Axelrod Lithocarpus nevadensis Axelrod, 1985, p. 145; pI. 8 figs. 8-10; pI. 24, figs. 1-7 Lithocarpus nevadensis Axelrod, 1995, p. 49; pI. 15, fig. 6 Lithocarpus nevadensis Axelrod, 1992, p. 38; pI. 9, figs. 4, 5 50 Referred ftpecimens-UOF42087(Fig.6H),UOF42088(Fig.60) Description-Lithocarpus nevadensis was separated from similar Quercus species by Axelrod (1985). The primary differences between L. nevadensis and Quercus simulata are the spacing of the teeth, which are further apart in L. nevadensis, and the shape of the secondary veins which are curved in L. nevadensis, and straight on Q. simulata. These observations are in addition to the differences outlined by Axelrod (1992, p 38) ofL. nevadensis having a heavier texture and more blunt teeth. The most similar modem species is Lithocarpus denstflorus (Hook. and Am.) Rehder. Genus- Quercus Species- Quercus dayana Knowlton Quercus dayana Knowlton, 1902. p. 51; pI. 6, fig. 1 Quercus dayana Chaney and Axelrod, 1959, p. 165-166; pI. 24, figs. 3-7 Referred specimens- UO F 521 05A (Fig. 61) Description- The original description of Quercus dayana by Knowlton is very apt for Cape Blanco specimens, particularly the emphasis on the thickened mid rib, weak secondaries, entire margin, and the mesh tertiary pattern that makes individual tertiary veins difficult to distinguish. The prominence of tertiary venation is the primary difference between Q. dayana and Q. hannibali, with Q. hannibali having visible and percurrent tertiary veins. The modem plant most similar to Q. dayana is Quercus virginiana. 51 Fields (1996, pg. 342-345) suggested that Q. dayana of Chaney and Axelrod (1959) should be reassigned to Chrysolepis and is also ajunior synonym ofDiospyros elliptica Knowlton (Knowlton 1902, p. 83-84, pI. 16, fig. 5). Diospyros elliptica is described and figured with only four to five secondaries. This is different from the 12 or more pairs common to the leaves from Cape Blanco and the specimens listed above. The argument for placing this leaf within Chrysolepis is the mesh pattern ofthe leaf which is uncommon among the oaks. Most oak tertiary veins tend to have high relief, with exception ofthe modem Q. virginiana which resembles the Cape Blanco fossils retained in Q. dayana. Genus- Quercus Species- Quercus hannibali Dorf Quercus hannibali Dorf, 1930, p 86; pI. 8,figs. 8-11 Quercus hannibali Chaney and Axelrod, 1959, p. 168; pI. 24, fig. 2; pI. 25, figs. 11-13 Quercus hannibali Axelrod, 1991, P 50; pI. 12, figs. 1-7 Quercus hannibali Buechler et aI., 2007, p. 328; fig. 8H Referred specimens- VO F 42122 (Fig. 6J), VO F 42115 (Fig 6K) Description-The discussion in Chaney and Axelrod (1959, pg 166-167) ofthe characters of Q. hannibali was helpful in differentiating this species from Q. dayana Knowlton. In Q. hannibali there is a slight curve where the 2nd order veins meet the primary vein, and the 3rd order veins are usually percurrent, as they are in Cape Blanco specimens. The margin of Q. pollardiana can vary from entire to serrate, but the 52 specimens from Cape Blanco are entire. The modem plant most similar to Q. hannibali is Q. chrysolepis Liebmann. Genus- Quercus Species- Quercus prelobata Condit Quercus prelobata Condit, 1944, p. 43; pI. 7, figs. 3,4 Quercus prelobata Chaney and Axelrod, 1959, p. 169; pI. 26, figs 8, 9; plate 27, figs 1, 2 Quercus prelobata Buechler et aI., 2007, p. 332; Fig. 8A, B Referred specimens- VO F 42138A (Fig. 6M) Description- Quercus prelobata is a fossil oak that has rounded lobes. A whole leafwas not recovered from Cape Blanco, but complete lobes were collected which contribute to our confidence in this assignment. The size range ofthis species was expanded by Buechler et al (2007) to include leaves up to 14.5 cm long. The fungal fruiting bodies discussed by Buechler et al (2007) were not observed on the Cape Blanco specimens. This oak has limited occurrences in Miocene floras, and it most resembles the modem Quercus garryana Hooker. Genus- Quercus Species- Quercus simulata Knowlton Quercus simulata Knowlton, 1898, p. 728-729; pI. 101, figs. 3,4; pI. 102, figs. 1,2. Quercus simulata Chaney and Axelrod, 1959, p. 171: pI. 30, fig. 2,3,5-8; pI. 31, fig. 1-4. Quercus simulata Buechler et aI, 2007, p. 332; Fig. 8C, D. 53 Referred specimens- W 66 (Fig. 6L), DO F 42142A (Fig. 60). Description- Quercus simulata is a common component of Miocene floras and has been the subject ofmultiple taxonomic revisions and reassignments. A thorough discussion is provided by Fields (1996, pages 363-379), who develops the concept of a "Quercus simulata morphoplex", which is helpful in documenting convergent morphology within Fagaceae. The primary contention regarding this species results from the substantial morphological variation within the species, most notably a full transition from toothed to entire margined specimens. The Cape Blanco specimens of Q. simulata are toothed and have straight secondaries. Stringently defining the morphology ofthe fossil, Quercus simulata, creates confusion when trying to identifY the extant plant that is most closely resembles. Fields (1996, pg. 375) demonstrates this difficulty by presenting 17 modem species that various authors had suggested resembled Quercus simulata. Given this confusion we are unable to identifY a closest modem relative of Quercus simulata. Genus- Castanopsis Species- Castanopsis perplexa (Knowlton) Brown Castanopsis perplexa Brown, 1940, p. 348-349. Castanopsis convexa Brown, 1936, p. 171; pI. 49, figs 8-11. Referred specimens- DO F 38056 (Fig 6N). Description- This species was reported by Brown (1936; p. 171) as resembling the modem Castanopsis sempervirens Dudley, which has since been revised to 54 Chrysolepis sempervirens (Kellogg) Hjelmqvist (USDA 2009). The fossil Castanopsis perplexa was included in the Quercus dayana synonomy of Chaney and Axelrod (1959, pg. 165). We disagree with this assessment because the tertiary venation in Quercus dayana is a fme mesh of~1 mm2 areoles with secondary veins 2-3 mm apart, while in the Cape Blanco specimens of Castanopsis perplexa the tertiary venation encloses a large areola of~4 mm2 and lower angle, more widely spaced secondary veins. The Cape Blanco leaves are larger than the figured examples of Castanopsis convexa (Brown, 1936), but are similar in size to modem Chrysolepis sempervirens. Order- Juglandales Family- Juglandaceae Genus- Carya Species- Carya bendirei (Lesquereux) Chaney and Axelrod Rhus bendirei Lesquereux 1888, p. 15; pI. 9, fig 2. Carya bendirei Chaney and Axelrod, 1959, p 155; pI. 19, fig 1-5. Carya bendirei Axelrod, 1991. p 48; pI. 11, fig 5-7. Referred specimens- UO F 42029B (Fig. 6B) Description- The original description ofthis species (Lesquereux 1888) indicates that the leaf tapers toward the base, however significant variation in leaf shape is shown by Chaney and Axelrod (1959) and Axelrod (1991). The more recently figures fossils compare well with our specimens. The primary recognition criteria is the curved secondary veins which diverge from the primary at a ~90° angle. Fine serration and a 55 transition from brochidodromous to craspedodromous are also diagnostic. The closest extant species is Carya ovata (Miller) Koch, a water loving species in eastern North America. Order- Urticales Family- Ulmaceae Genus-Ulmus Species-Ulmus speciosa Ulmus speciosa Newberry, 1898, p. 80; pI. 45, figs. 3, 4. Ulmus speciosa Tanai and Wolfe 1977, p. 8; pI. 3C, F. Ulmus speciosa Axelrod, 1991, p. 51; pI. 12, figs. 8-10; pl. 13, fig. 10. Ulmus speciosa Meyer and Manchester, 1997, p. 82-83; pI. 18, figs. 1-6. Referred specimens-W 61 (Fig. 7A). Description- The key ofTanai and Wolfe (1977) was followed in assigning our specimens to this species, however there remains confusion about the teeth that characterize this species. In the descriptions leaves are described as doubly serrate, however in the plate both double and singular serration is documented. The tertiary venation, and asymmetrical bases observed in our specimens are identical to Ulmus speciosa even though the Cape Blanco specimens are overwhelmingly singly serrate. Axelrod (1991) was not satisfied with the revisions ofTanai and Wolfe (1977) regarding species level differentiation within Ulmus. Meyer and Manchester (1997) suggest the most similar living species is Ulmus Americana, but Axelrod (1991) reports it as extinct. 56 Genus-Zelkova Species- Zelkova browni Tanai and Wolfe Zelkova oregoniana Chaney and Axelrod, 1959 [part] p.174; p1.3l, figs. 6, 7. Zelkova browni Tanai and Wolfe, 1977, p. 8; pI. 4A, C-G. Zelkova browni Buechler et aI, 2007, p. 333; fig. 81. Referred specimens- DO F 42155 (Fig. 7D), DO F 38053 (Fig. 7F). Description- The unique characteristic that separates Zelkova from Ulmus is alternate percurrent to regular polygonal tertiary venation (Tanai and Wolfe 1977). This character is well preserved in one ofour specimens and moderate to poorly preserved in the other. The subsidiary teeth also characteristic of this species are observed in one specimen. The modem plant most similar may be Zelkova carpin!folia, currently restricted to Asia. Order- Hamamelidales Family- Platanaceae Genus- Platanus Species- Platanus bendirei (Lesquereux) Wolfe Platanus dissecta Lesquereux, 1878, p. 13; pI. 7, fig. 12; pI. 10, figs. 4, 5. Platanus dissecta Berry, 1929, p. 248; pI, 53, figs. 1, 2; pI 61. 57 Figure 7: Cape Blanco fossils, orders; UrticaIes, Hamamelidales, RanunculaIes, LauraIes, and Rosales, scale bar is 1 em. A. Ulmus speciosa, W 61. B. Hamamelis merriami, UO F 42181. C. Angiosperm 8, UO F 42196. D. Zelkova browni, UO F 42155. E. Cercocarpus nevadensis, UO F 42208. F. Zelkova browni, UO F 38053. Platanus bendirei, UO F 42180. G. Mahonia macginitiei, UO F 42191. H. Mahonia macginitiei, UO F 42189. I. Platanus bendirei, W 60. J. Platanus bendirei, UO F 42172. 58 Platanus dissecta Chaney and Axelrod, 1959, p. 182; pI. 36, fig 3. Platanus bendirei (Lesquereux) Wolfe, 1964, p. N24; pI 4, figs. 1, 2, 4. Platanus bendirei (Lesquereux) Wolfe and Tanai, 1980, p. 28; pI. 3, fig. 4; pl. 4, figs. 1, 2,4. Referred specimens- W 60 (Fig. 71), DO F 42172 (Fig. 7J) Description- We follow Wolfe and Tanai (1980) in assigning our specimens to his Platanus bendirei. Wolfe and Tanai (1980) cite expansion of the original concept of P. dissecta to include both 3- and 5-lobed leaves by workers in the middle twentieth century as the reason to separate P. bendirei as the 3-lobed species that is more common in the Middle and Early Miocene. We do not have complete leaves from Cape Blanco, however the leaf bases that we recovered have three primary veins, indicating a three lobed leaf, and the margin clearly shows the Platanus axial and basal concave teeth. Platanus bendirei was widespread in the Miocene and all Succor Creek sycamores were assigned this name (Fields 1996). Platanus is commonly confused with members of Acer, because both families tend toward large, lobed, toothed leaves; however the large concave teeth observed in the Cape Blanco specimens are more like Platanus. There is no modem Platanus that has the marginal characteristics, lobation, and size ofPlatanus bendirei, but the characteristic shape of the Platanus teeth can be seen in the modem Platanus occidentalis. Family- Hamamelidaceae Genus- Hamamelis 59 Species- Hamamelis merriami Chaney and Axelrod Hamamelis merriami Chaney and Axelrod, 1959, p. 180; pI. 35, figs. 1-3. Referred specimens- DO F 42181 (Fig. 7B). Description- This species has been identified at few localities in the Miocene, but the strong secondaries, coarsely crenate-sinuate leaves and firm texture confirm the presence of this species at Cape Blanco. The modem Hamamelis virginiana Linnaeus has a strong resemblance with the fossil specimens. Order- Ranunculales Family- Berberidaceae Genus- Mahonia Species-Mahonia macginitiei Axelrod Odostemon hollicki Dorf, 1930. p. 93; pI. 10, figs. 7, 8. Mahonia hollicki Arnold, 1936, p. 61; pI. II, figs. 3-8; pI. III, figs. 5, 7, 9. Mahonia macginitiei Axelrod, 1985, p. 150; pI. 11, figs. 2, 4,9; pI. 27, fig. 7. Mahonia macginitiei Axelrod, 1991, p. 52; pI. 14, figs, 3, 4, 7, 9. Mahonia macginitiei Axelrod, 1995, p. 50; pI. 16, figs. 4, 6-9. Mahonia macginitiei Buechler et aI. 2007, p. 334; fig. 8G. Referred specimens- DO F 42191 (Fig 7G), DO F 42189 (Fig. 7H). Description- The secondary veins on Mahonia are festooned brochidodromous with the first loop occurring roughly half way between the primary vein and the margin. Tertiary veins form an irregular polygonal reticulate mesh with rectangular areoles. The 60 long axis of the rectangular areoles is at a high angle to the primary vein. These identifying characteristics were sufficient to determine a generic assignment for these specimens. The assignment to Mahonia macginitiei Axelrod is preferred over M. simplex (Newberry) Arnold because our specimens lack the pair ofprominent basil veins that are unique to M simplex (Arnold 1936, p. 61). The modem equivalent suggested by Dorf (1930, p 93) was Odostemon aquifolium Persh, which has been synonymized into Mahonia aquifolium (Persh) Nutt (USDA 2009), and this modem equivalent has been supported by subsequent authors. Order- Laurales Family- Lauraceae Genus- Persea Species- Persea pseudocarolinensis Lesquereux Persea pseudo-Carolinensis Lesquereux, 1878, p 19-20; pI. 7, figs. 1,2. Persea pseudocarolinensis Chaney and Axelrod, 1959, p. 177-179; pI. 34, figs. 1, 3, 4, 6. Referred specimens- UO P 15013 (Fig. 8A). Description- Long arching secondary veins are the unique identifier for this species. This species had a complicated taxonomy in the early part of the twentieth century, but we follow Chaney and Axelrod (1959) acknowledging the generalized Lauraceous appearance. The modem equivalent is likely Persea americana Miller, or Persea carolinensis Nees. 61 Order- Rosales Family- Rosaceae Genus- Cercocarpus Species- Cercocarpus nevadensis Axelrod Cercocarpus nevadensis Axelrod, 1991, p. 51; pI. 17, figs. 6-13. Referred specimens- UO F 42208 (Fig 7E). Description- Represented by a single nearly complete specimen from Cape Blanco, the fossil C. nevadensis was introduced by Axelrod (1991). However, other Cercocarpus species are present in the Pacific Northwest, often C. antiquus Lesquereux, which has more closely spaced secondary veins than C. nevadensis. The modem species most similar to C. nevadensis was reported as Cercocarpus betuloides Nuttall by Axelrod (1991), which has since been reassigned to Cercocarpus montanus (USDA 2009). Family- Rosaceae Genus- Sorbus Species- Sorbus idahoensis Axelrod Sorbus idahoensis Axelrod, 1985, p. 169; pI. 29, figs. 2, 6-7,10. Referred specimens- W 40 (Fig. 8K). Description-Though based on a single specimen this leaf is unique because of its strongly single serrate margin, and narrow lanceolate shape. Higher venation details are not preserved in the Cape Blanco specimens. Axelrod (1985) suggests Sorbus 62 acuparia Linnaeus may be a modem equivalent. This may have been a typographic error for Sorbus aucuparia, European Ash Order- Rhamnales Family- Rhamnaceae Genus- Paliurus Species- Paliurus blakei (Chaney) Meyer and Manchester Ceanothus blakei Chaney, 1927, p. 128-129; pI. 16, figs. 4, 8. Paliurus blakei Meyer and Manchester, 1997, p. 145-146; pI. 61, figs. 5, 7, 8. Referred specimens- VO F 42224 (Fig. 8D), VO 42216 (Fig. 8F). Description- Rhamnaceae are most readily identified by oblong to ovate leaves with actinodromous primary venation. Originally described as a Ceanothus (Chaney 1927), Meyer and Manchester (1997) reassigned the smaller specimens with smaller teeth to Paliurus and the size range includes leaves as small as those recovered from Cape Blanco. The recovery ofPaliurus fruits from other Oligocene localities supported the change ofname, however they acknowledge the similarity in leaves ofPaliurus and Ceanothus. The Cape Blanco specimens are smaller than Ceanothus chaneyi Dorf (Dorf 1930; page 103, pI. 13, figs 3, 4), but resemble the Ceanothus sp. fragment figured by Axelrod (1944b; p. 203; pI. 38; fig. 7). The Cape Blanco specimen is differentiated from Ceanothus edenensis Axelrod by its serrate margin. The venation and margin have a clear resemblance to modem Ceanothus thyrsiflorus Eschscholtz of the California and Oregon coast. Axelrod (l944b) suggests that a more close resemblance is seen between 63 Ceanothus blakei and modem Ceanothus velutinus Douglas however, we favor the smaller and more prominently toothed C. thyrsiflorus. Since Meyer and Manchester reassigned the genera of these specimens they suggest that the most similar modem plant is within Paliurus which is currently found in southern Europe and Asia. Order- Lamiales Family- Boraginaceae Genus- Cordia Species- Cordia oregona Chaney and Sanborn Cordia oregona Chaney and Sanborn, 1933, p. 94; pI. 38, figs 4, 5. Referred specimens- UO F 42221A (Fig. 8C), UO F 42224 (Fig. 8G) Description- These leaves, subcoriaceous and well preserved at Cape Blanco, have lengths greater than 10 cm and widths greater than 7 cm. Those figured and described by Chaney and Sanborn (1933), had a maximum width of4.6 cm. Despite this discrepancy we assign the collected specimens to this name because of the similarity of all other leaf features. Unique to Cordia oregona are basal secondaries parallel to the margin for up to 1/3 ofthe leaflength. The modem leaf that most resembles this fossil fonn is Cordia collococca. Images of modem herbarium specimens for C. collacocca available online from the New York Botanical Garden, C. V. Starr Virtual Herbarium (http://sciweb.nybg.org/science2/VirtualHerbarium.asp, accessed June 2, 2009). Order- Dipsacales Family- Caprifoliaceae Genus- Viburnum Species- Viburnum lantanafolium Berry Viburnum lantanafolium Berry, 1929, p 264; pI. 60, fig. 6. Viburnum lantanafolium Chaney and Axelrod, 1959, p. 202; pI. 44, fig. 4. Referred specimens- UO F 42251 (Fig. 8H). Description- These specimens have strong third order veins that are nearly perpendicular to the primary vein. This is the character that places them in Viburnum. The margins of the specimens that were collected at Cape Blanco are not sufficiently preserved to conclusively say that all margin characters in the description are present, however, there is some evidence of teeth. The lack ofprominent teeth precludes their assignment to V. grahamii (Fields 1996 p. 515). The arching character of the secondaries and their tendency to travel greater that half ofthe 1eaflength also supports their assignment to Viburnum. A similar modem species is Viburnum lantana Linne (Berry 1929). Unknown Botanical Affiliation Species- Angiosperm Form 1 Referred specimens- UO F 42253 (Fig. 81). 64 A.. . '.'~.. "". . , ~ - ·K- 65 Figure 8: Cape Blanco fossils, orders; Laurales, Rhamnales, Lamiales, Dipsacales, and undetermined; scale bar is 1 em. A. Persea pseudocarolinensis P 15013. B. Angiosperm 9, DO F 42206. C. Cordia oregona, DO F 42221A. D. Paliurus blakei, DO F 42224. E. Angiosperm 2, DO F 42253. F. Paliurus blakei, DO F 42216A. G. Cordia oregona, DO F 42224. H. Viburnum lantanfolium, DO F 42251. I. Angiosperm 1, DO F 42253. J. Angiosperm 3, W 35. K. Sorbus idahoensis, W 40. 66 Diagnosis- Unlobed; elliptic; entire; length> 5 cm; width 1-2 cm; apex attenuate and straight; base unknown; length to width ratio> 4: I; secondary veins diverge from midrib at an angle near 90° and loop back to form semicircular area that encloses a third order mesh; tertiary veins regularly polygonal reticulate; third order veins are highest observed; tertiary vein course sinuous; areoles well developed.Description- The high length to width ratio and tertiary venation is suggestive ofSalicaceae, but these specimen have a unique tertiary vein pattern. No whole specimen was recovered further hindering their identification. The specimen is reminiscent of TWa pedunculata Chaney (Meyer and Manchester, 1997; page 111, pl. 40, fig. 6). Species- Angiosperm Form 2 Referred specimens- UO F 42253 (Fig. 8E). Diagnosis- Unlobed; elliptic; entire; length < 4 cm; width I cm; apex acute and straight; base acute; length to width ratio 3-4: 1; brochidodromous secondary veins diverge from midrib at a moderate angle near 60°; tertiary veins alternate percurrent and sinuous; quaternary veins regular polygonal reticulate; areoles well developed.- Description- The small size ofthis leaf form is what sets it apart. The shape and venation are similar to Quercus simulata however there are no transitional sized leaves present and so this form is set apart. Other documented leaves of this size are Cercocarpus however the entire margin and acute apex show that such an assignment is incorrect. 67 Species- Angiosperm Form 3 Referred specimens- W 35 (Fig. 8J). Diagnosis- Unlobed; oblong; entire; length> 3 cm; width 0.5 cm; apex acute and straight; base unknown; length to width ratio> 4: 1; brochidodromous secondary veins diverge from stout midrib at high angle; third degree veins regular polygonal reticulate and sinuous; areoles well developed. Description- Unlike Angiosperm 2, which it resembles in small size and high length to width ratio, Angiosperm 3 lacks clear quaternary venation on Angiosperm Form 3. While the size and shape ofthis leafis reminiscent ofSalix churchillensis Axelrod (1991) the venation is not similar. S. churchillensis secondaries loop well up into the blade and simulate a marginal vein. The secondaries on Angiosperm Form 3 are cladodromous to brochidodromous. Finer venation is not observed in S. churchillensis and is present on the Cape Blanco specimen, which was collected by Jack Wolfe. Species- Angiosperm Form 4 Referred specimens- UO P 15015 (Fig. 9D). Diagnosis- Unlobed; entire; coriaceous; elliptic; length> 5 cm, likely 8 cm; width 5 cm; apex not preserved; base rounded to straight, obtuse; length to width ratio <2: 1; primary veins 3 basal actinodromous with possible marginal veins also; secondary veins brochidodromous with spacing that increases toward the base; tertiary veins alternate and sinuous; quaternary and quinternary veins regular polygonal reticulate. 68 De!J'cription- Basal actinodromous primary venation is limited to Populus, Platanus and Paliurus in the Cape Blanco Flora. However the coriaceous texture, entire margin, and lack oflobes prevent assignment to Populus or Platanus. And there is not sufficient preservation to allow assignment to the Paliurus. There is some resemblance in form and secondary vein shape to Persea however the spacing of secondaries in Angiosperm Form 4 is not consistent with Persea. Species- Angiosperm Form S Referred !J'pecimens- UO F 42261 (Fig. 9A) Diagno!J'is- Unlobed; entire at base, full margin not preserved; subcoriaceous; oblong to elliptic; length> 10 cm, likely ~12 cm; width> 3 cm, likely ~4 cm; apex not preserved; base acute to cuneate; length to width ratio 2: 1 to 3:1; primary veins pinnate, secondary veins straight, depart margin at 30°, opposite, > 6 pairs, appear eucamptodromous, although margin is poorly preserved; tertiary veins not visible. Description- The secondary veins suggest an affinity with Lauraceae or Comus however without a complete margin a confident assignment is not possible. Species- Angiosperm Form 6 Referred !J'pecimens- UO F 380S4B (Fig. 9E). Diagno!J'is- Degree oflobation unknown, only base was recovered; entire at base; leaf shape and size unknown, length> Scm, width >7 cm; apex not preserved; base. ~E t 69 Figure 9: Cape Blanco fossils, unknown botanical affiliation, scale bar is 1 cm. A. Angiosperm form 5, UO F 42261. B. Angiosperm form 10, UO F 42207B. C. Angiosperm form 7, UO F 42263. D. Angiosperm form 4, UO P 15015. E. Angiosperm form 6, UO F 38054B. rounded and obtuse; length to width ratio unknown, but likely low; primary veins, 5 palmate; tertiary veins alternate 70 Description- The palmate venation is unique but based on a basal fragment only. The leaf shows some resemblance to Cercidiphyllum and the actinodromous venation ofAcer. Species- Angiosperm Form 7 Referred specimens- UO F 42263 (Fig. 9C). Diagnosis- Unlobed; entire; coriaceous; obovate; length> 10 cm, likely 14 cm; width > 6 cm; apex attenuate; base not preserved; length to width ratio ~2: 1; primary veins pinnate; secondary veins diverge at a moderate angle, are brochidodromous with spacing that increases toward the base; tertiary veins percurrent and sinuous; quaternary and quinternary veins regular polygonal reticulate. Description- These leaves resemble Parrotia brevipetiolata Meyer and Manchester (1997) as well as the modem Gaultheria shallon Pursh. The fossil record of Gaultheria is limited to a two specimens from the Latah flora that were originally described as Arbutus matthesii (Brown 1936; pg. 184-185, pI. 59, fig. 5, 6) and then later transferred to Gaultheria pacifica Brown (Brown 1946; pg. 351-352). The problem with assigning the Cape Blanco specimens to Gaultheria is the margin state which is fmely serrate in the fossil and modem Gaultheria, but entire in the Cape Blanco specimens. The similarity of leaf morphology in Hamamelidaceae and Ericaceae highlights the difficulty in assigning proper botanical affmity to fossil specimens that lack reproductive structures. Gaultheria is widespread along the west coast ofNorth America today, while extant Parrotia species that resemble the fossils are limited to Asia. Because of the 71 ambiguity in assignment and the ramifications to the floristic interpretation, we do not assign this form to fossil species Species- Angiosperm Form 8 Referred specimens- UO F 42196 (Fig. 7C). Diagnosis- Unlobed; entire; obovate; length> 10 cm, likely 14 cm; width > 5 cm; apex nor preserved; base not preserved; length to width ratio ~2-3: 1; primary veins pinnate; secondary veins diverge at a high angle, are brochidodromous with uniform spacing; tertiary veins percurrent and sinuous; quaternary regular polygonal reticulate. Description-This large, brochidodromous, leafis example of the mesophyll component of the Cape Blanco flora. Assigning this specimen to a fossil species was not possible because neither apex nor base was preserved. The high divergence angle and uniform secondary spacing separate this form from Angiosperm 7. The wide leaf is not within Lauraceae because the brochidodromous secondaries form tight loops close to the margin and the secondary veins are parallel for their whole length. Large leaves within the modem Anacardiaceae family have a strong resemblance to these fossils and are now found in tropical regions. Species- Angiosperm form 9 Referred specimens- UO F 42206 (Fig. 8B). Diagnosis- Unlobed; entire; oblong; length> 10 cm, but complete leafnot recovered; width ~ 4 cm; apex nor preserved; base acute; length to width ratio ~4-5:l; primary veins pinnate; secondary veins diverge at a moderate to high angle, are brochidodromous with uniform spacing, secondary veins dichotomize 2/3 of the way to the margin; tertiary veins weak; thick midrib at base. Description- This leafis well preserved and larger than most ofthe Cape Blanco fossils. It has a resemblance to the modem Rhododendron macrophyllum D. Don ex G. Don, ofwestern North America., and also the fossil Magnolia lanceolata Lesquereux, but without a complete specimen identification was not possible. Species- Angiosperm form 10 Referred specimens- UO F 42207B (Fig. 9B). Diagnosis- Unlobed; entire; oblong; length> 10 em, but complete leafnot recovered; width ~ 6 em; apex nor preserved; base acute; length to width ratio ~4-5: 1; primary veins pinnate; secondary veins diverge at a moderate, are brochidodromous to c1adodromous with uniform spacing; tertiary veins percurrent; quaternary venation not preserved. Description- The gently looping secondary veins of this specimen are nearly parallel to the margin before looping into the secondary vein above them. This gives the leaves a Lauraceae appearance, but without a completely preserved specimen identification was not possible. 72 73 TAXONOMIC SUMMARY The Cape Blanco flora is composed of at least 36 unique angiosperm leaf forms, most of them identifiable to the species level. Table 2 shows the quantities of each form, as well as the most similar extant plant, where identification was possible. FLORISTIC INTERPRETATION The assemblage ofleaves observed in the Cape Blanco flora is a combination of forms that are well documented and commonly observed in the Miocene of the greater Oregon region (Chaney and Axelrod 1959) and few exotic, and geologically older, forms. The dominance ofoaks, for example, is observed in many Miocene localities and have been used as evidence for the onset ofdrying and cooling climate in western North America (Axelrod, 1995). Also common in Miocene floras are the Betulaceae, Juglandaceae, and Rosaceae, which are all present at Cape Blanco. However, some of the larger leaves that were found at the Cape Blanco locality, Cordia oregona and Angiosperms 7, 8, and 9, are not reminiscent ofleaves documented at other Miocene localities. The Cape Blanco flora is a unique Miocene assemblage with large leaved heat and water loving taxa, as well as more temperate Miocene forms. Such tropical hold over taxa are not widely known from Miocene floras of eastern Oregon and adjacent states (e.g. Chaney and Axelrod, 1959; Fields, 1996; Beuchler et al. 2007), where paleosols, plants, and mammals are evidence of cold, dry paleoclimate and Table 2: ~ossil Quantities and Similar Modern Plan~s Fossil Name Quantity Similar modern plant Equisetum sp. 4 Equisetum sp. Pinus tiptoniana 1 Long needle pine Cone unit 1 Pinus? Picea? Charcoal '14 Conifer Needle 26 Conifer Shredded wood 8 Thuja? Pinus? Splayed needles 2 Picea? Typha sp? 7 Typha sp? Populus eotremuloides 20 Populus balsamifera Populus lindgreni 6 Populus heterophyfia Salix laevigatoides 30 Salix laevigata Salix succorensis 6 SaJix nigra Alnus harneyana 3 Alnus tenuifolia Betula thor 3 Betula papyrifera Chrysolepis sonomensis 46 Chrysolepis chrysophylla Fagus washoensis 6 Fagus grandifolia Lithocarpus nevadensis 19 Lithocarpus densiflorus Quercus dayana 20 Quercus virginiana Quercus hannibali 38 Quercus chrysolepis Qurcus pre/abata 3 Quercus garrj"ana Quercus simulata '16? Castanopsis perplexa 3 Chryso.lepis sepervirens Carya bendirei 5 Carya ov'ata Ulmus speciosa 12 UlmLJS americana Zelkova browni 3 Zelkova carpinifolia PlatamJs bendera! '16? Hamamelis merriam! 8 Hamamelis virgli?iana Mahonia macginit!ei 'I 0 MahonIa aquifolium Persea pseudocaroHnensis 4 Persea carolinensis Cercocarpus ne~'adensis 1 Cercocarpus montanus Sorbus idahoensis 1 Sorbus aucupan'a Paliurus blake! 9 Ceanothus thyrsif.lorus Cordia oregona 30 Cordia collococca Viburnum fantanafolium 6 Viburnum lantana Angiosperm form 1 5 Tilia? Angiosperm form 2 5? Angiosperm form 3 1 Salix? Angiosperm form 4 1? Angiosperm form52? Angiosperm form 6 1 Cercidiphyllum? Angiosperm form 7 "0 Gaultheria? Angiosperm form 8 7 Anacardiaceae? Angiosperm form 9 1 Rhododendron? Angiosperm fom1 '10 2 Lauraceae? 74 75 open grassy vegetation (Retallack 2004,2008; Wolfe, 1994). The persistence of large leaved species at Cape Blanco may be evidence of coastal refugia, protected from extremes of frost and heat endured by Miocene plants of eastern Oregon. The presence ofrefugia assemblages has been documented through the Tertiary in the Amazon basin of Brazil (Hooghiemstra and van der Hammen, 1998) and their presence is used explain modern Amazonian diversity. Similarly in Oregon, the Miocene spread ofoak may not have entirely displaced geologically older tropical species in coastal refugia. Many of the larger leaves from Cape Blanco have prominent midribs and a coriaceous texture, suggesting an evergreen habit as in the live oaks. Broadleaf evergreens along with the large number ofoaks suggests that the Cape Blanco flora may represent the notophyllous broad-leaved evergreen forest ofWolfe (1979), suggesting a mean annual temperature (MAT) between 13 0 C and 20 0 C. Though the presence of conifers suggests a lower MAT, they are occasionally present in the notophyllous broad- leaved evergreen, or 'oak-laurel' forest ofWolfe (1979). PHYSIOGNOMIC INTERPRETATION The basic premise of morphologically based paleoclimate analysis is that morphological changes in modern floras can be correlated with climatic parameters. The shape of modem leaves has been shown to correlate with temperature and precipitation (Bailey and Sinnott, 1916; Wi1fet al. 1998; Wolfe 1993). The relationships are based on the whole composition ofthe flora, with the proportion of species with serrated margins deceasing with warmer temperatures (Bailey and Sinnott, 1916) and leaf size increasing 76 with increase precipitation (Wilf et al. 1998). One advantage ofthe physiognomic method is that it is ataxonomic. Complications arise in the morphological descriptions because there is not a documented theoretical understanding ofwhy leaves are shaped the way they are. Such data would inform the collection ofmorphological features. The characters considered important by Wolfe (1993), are shown for each species, including unassigned forms, in Table 3. Morphological description was based on material collected from Cape Blanco, and published descriptions. These data emphasize overall leaf shape and margin character without addressing venation patterns. However, the venation patterns are essential in name assignment and the establishment of the unnamed angiosperm forms. Two specific techniques of physiognomic interpretation are multivariate analysis of a suite of leaf morphologic characters are associated with many climate parameters (Wolfe 1993), and univariate analysis of a single morphological feature correlated with a single measured climate variable. Both methods have advantages and disadvantages, the multivariate method can yield more data, yet suffers from lack of transparency, variables that co-vary, and difficulty in defming the error ofresults. The univariate method is straight forward and transparent, but perhaps does not extract the full measure ofdata from the flora. Both methods suffer from a lack of theoretical underpinning and thus solely rely on empirical observations. Here we apply the two most robust univariate analysis, those that estimate mean annual temperature and mean annual precipitation, to the Cape Blanco Flora. 77 There are 36 unique angiosperm forms in the Cape Blanco flora, 34 of them have a single margin state, and 55.5% ofthe single margin state species or species equivalents are entire. Using the equation ofWing and Greenwood (1993) LMAT = 30.6P + 1.14, where LMAT is leaf estimated mean annual temperature, and P is the proportion ofspecies that have entire margins, the LMAT ofthe Cape Blanco Flora is, 18.24 0 C ± 2.6 0 C, the error is calculated following Wilf(1997). This result corresponds well with the floristically derived MAT estimate. The current mean annual temperature at Cape Blanco is ~11 0 C (Wolfe 1993). The paleoprecipitation at Cape Blanco can be estimated in a similar fashion. Following the equations of Wilf et al.(1998), the mean annual precipitation at Cape Blanco is 201 cm (+ 86 cm, -61 cm). This is slightly more than the195 cm per observed today at Cape Blanco year (Wolfe 1993). PALEOTOPOGRPAHY Paleoelevation ofthe Cape Blanco flora is based on the sedimentological evidence of sea level deposition. Comparable modem species are evidence that both streamside and hill-slope vegetation types were present. This supports the sedimentological evidence for disturbance of an entire local catchment area by volcanic ash fall. l link 3: \1 nqlh nln(:! ,. lit' I.t'li w, r,fl Species FcmTl Teeth Leal SILl! Regul"r CI""" Round Acute :1'2S J 25 .~12~, :t 5 ""I :) ~. ) S j 2S :' ~ Cmfk-l. N"nc. Le~c. 1 J J :) ':' _, J :1 ~I J ..I .:' J ';' " "' ;1 :) j ") :,) ~ J ::1 :--:. ) :) :' ::. :' :1 :) -' j J :' j :' J :) J :J :1 ~l J :' :1 :-:i J J :1 ? :. :1 ~l .) :i J J ~) J "' :' J ~t :) oJ :1 :' ~i :1 -' :' :' .:! J ~l :J ) ~, :J ';' ) :' ::J J J ::) :. J :1 :' :' ) :) ::> :' 5 :i,5 :'. ~l :-:' ) S J :. :' 5 ) :1 ~l :,5 :J 3::' :' :. :; ~I J ? ) 5 :. ::. )S ~I :i :) ).5 :' ':i :' 5 :':1 :' ) :. :1.5 ::. ::-. .) :. :'.5 ) S ) :' 5 :1 ~. ) :. ~! :1.5 21 :. )5 :) ::.::. :1 32. .:. :. ':') ~I )S :1 :' ? )S j J :i5 J5 J5 J e· ~1 ) J5 :. :i S :' ~j ~. :IS 1 :' 5 ,:1 ~. :'~ J :;.~) J5 :1 ~;2; Mlc::ro.:2 Micr M,>.~n. 1 M,,-'d .\i3.r:'D.r:'/:3 ,r,"j fj=~~:r:'';!,;f!i F'9tS9.;; .1''3 ~l..l·:f"':'·~1",,-·i.":'9,":":;,"_~ c~ r::v :;"""J fP.~.I~ r\!:,,,,,:.y.J~..1t·, ~~i s~.),.'tl~.l ~ :.:}:,~~·!.:;I(.~t·.~:,,~ C:3st:J.F)~P~,;S .OBrpr,9.,'.r:a C:3.')03'~.ry..~;."g.' i.lm'~f!.i S,C'-€,,:;,',o.-:;.~ Zt~ ....~)Il:.j ~.v:h\·r\, ;;~9!'ial~!..IS .t\~J~:)!:1'r~' ,~,r';;;)spiZ'rlTl brIT; 1 Haf;1 at:! £t.1,'$ ,~~.r·.,~gr:1,' p~J:!..I.'~LI:' .D,r.8l,~·Sli' Cd,!""J.'e! ':J.rEl}:J.~ ~·~t'~·.~~lm :a'l·~a'":'36N'J.~ ,4!'-';;;Jl':) ~;P~f:il r:)r~i'· '2 .:"f,;.-~~p:-rf"!'f:)fifl ::, !::'1.'.~ !':'J~ ,r..v.;.~r.:.Jf) :;J~) Q!}er::-~.}s s.'.':"'j ~','a.t~ ..;r.'~I)sp=-rm T::'ff':"1 If .~r:~;·~ ;p-a-!ITI f:I~:;: 5 ''='':':i:)~p;z-rr:-' r:;otTTi t:i ,.!&.r.;:::- sp ~rm brrn "7 Ar'~;'\:J ~;p-e;~f:~~ r:.)rf~·! ~; .1\f;';'::'5D=-~1"It)frr: :,:l .~r:::~;D=-rmfDr;r) 1) F'Jp~.\·,~.~t ~)". ~P~l ~.\I·~)r·~.t...,~ ;~:),L""".IJ..I S .1".':a:J}"'~.IY Ba'i.f.' .'a8'':''~ ::\;::,9 S S~JN.i: S.~l %o.rS!.'":l.:Ji:i .Aln '...'/5 ,""!3.rp9 ~var,1;J B~~ :1,';.1 ~'.! ~x c.~.;"ys'-:Jl~.o''5. S,:l'\).';·}f.iI.':'S,'S ;:g.:;} !"'£;' \o,':=:tSJ:'-::>oe.1:-S'S ~'.'.!'! xa.'pus :':"eV:30~.~s.'s Q.~\9.r::-,~,~ ~ya.':l.~ CI.~;I.":.:'.• ~; t-I;.A.t'.r~':l.'J.c).r,' ::11 t·.\"rrh.;l\('.'",\· ,·,ll,r,,;il k:l':,C.; '.'. illl e:t1'~)):lri,~·,; 1~:111';1';·' in.~ "S "Il~ : . '.1;:(, I \h hr~ ',j :tli,'n ..: Cmr>J. Ci1mp,:1I.DlJ: >hn.:,. n:lIl:·'l'hdl: 1dK 1(1. mi,~l(lrtr:11: ~" k ':'. m':"'rh:.-II· .~.t1,:n 111·:mIJl·~ --.l 00 0'1 r-- '- ,- (, .. ( (. ,: (. l r: C '.' t. (. t:. r: L··~ 'C} I...,, C , C C , c e c , c c c , c C ll~ JCJ ',.Ld.?Csc~5u"y" C , C , C C C C i- i- i- I- C C e , '." ,(:J I,'U;:; ctl'r,,f:U'i C l- e c e <;;e <;; e e c <;; c <;;c c , c e ;H I"";,':~;)I; t)1'; ;1~T;,' "~.l,,"-;I~·.~·'-; (j",/"1 £( C rc c ££C C C <;; e <;; e c <;;e <;;c c £e £ c C C r: l...~-;h...... ~~; c',r .~-:':I S I~ sc r: c ~:I.: l: ~t ( ~:( c c c sc sc· c C l C 1~.lf~.E.:lti .:;j(I.,IIl'£..::' C l e I. e c c c C I. I: C ·'=:!~L, ;:~: l.t,G I .$ r iG., ~..:-~~' C C l C C l C C l C C c C l C ::;l~l! i;::(:f:l, ;:;l~ .;:;r, ::(1'8.:; X.I~~) e l c e s c ~ c e c l C C e l c e ;:,'~tJ6\.lII:::""E.::~.c.r;~;;:cr' :::;i.!"6C e l c e c c l C C l C C l- e e J51.;.-l....i[;;:~L.u E"1·lJ;::Li=;,~.' SC ~ C c e c c l c e L- e e c l- e 1·~U=ii..~,..":..i ~.II:3I.U·E::':'u::~ C £c SC r C· e £c £e c so £c £ C £ C r C ...;;.':;;Cw',:;cr S''";t..>'E;EC• £C ~c C C ITC ITC ::::: C c C ~c S'C c , c C ,·l..··\1C~'G E"'::"'1',;,:! n I; C e c l~ c <;;c c e c <;; c I; C e l- e c r;::;:~;,"::... :=t.)'"~ .'- "~,".~'l:",! :2£"C £cc (£ C C c <;;c <;; c c , c e c ~~ ( £O( C I':, :'J~ I~~~ ('; r~': ,1':.:) C , e 1;'12 ~ C C C C ~ C C C l- e c r: :.:~~,' d,~-:,:)' .:;,; .:.; :)':'; l.r::;.;;r:.::.1 C I. e l~(: C u-: C l); C I:: r: SC c r: C q ~~ C C E:-+~:lrJLl,.II'S ::;r"':::.I~r<:, r. l C I: C r: I. e r: l- e e s c S c C ~:~~"a~:,o:r.i::i -=il·I~:.I(I~) C I- e c , c (:; r:: l: C I- e e c l C C "E.: Gill lit' ~.I -S ri'.:",,-~ !'I~~':' e l c C G C S C C C sc- Si C C ~ C ~ C C C E=u"E=/;e:.c ':;;l~ ..i:'::(j::') C l- e c s c sc c C l C e e GC S C C :i;;;u·a.r::~:, cou ;;r;::::i...::.:;..':': l.i~:t! C l- e e e ~ c ~ c e c ~c ~ C r.. C L- e :>,";1..l'5Cu"S:·\' 'sr,c..r:::: C , c n ~C e c c ~ c C IT"C f€ C ITC C ::. ;:'I..l·;;; •.•I.:.;:':Li·::·':; :;,,"d&1"':.:,.~~"l..t·:J £C £e c c C £C £C C C £C <;; c C l C ..·::;I~·; E","!"';; f~ £C SC C C C C L C C SG £C C £ C n c r: LiE/~ ,~w"~·El..t" ];,"'";\.;,'"1 C , C ~ e c c e , e c <;; c Ge c C ':-~;.::'l)"i: ;...~..~..~. '-..~ /,', l~~' <;;c l; C C I;'e I~ c C c C ~ C c <;;c ~'C e c .:";~';.I-:;:-.;r:t.:"l"';''-''' )"',",l':S <;;c £c c e c <;; C <;; c c c , c c <;; c £C C I l~~).;(r.I"·1 ';;"';1 r;:J':~\..~ S e S r: C r: C S r:· se e [( c n:: c ~~t (: C L- [- e ~;';'.l:,:::,rMl;:;'j~~:;:; .~rilr,~.i::""1 olt!·~O :)!ld!113 OI'?MX1O ~: t'<' ~:H: ~ :c-z ~ :z- ~ ~: ~:> ()~1:)1f pu OOl::! oll!pJo::J .U(lll'it . 1!- ~~ SekJovia Flora 10:1" 4"1' 13C. rd 84 Figure 1: Map of Flora Locations 85 deposition along with the formation thickness, areal extent, and duration ofdeposition, all suggest that the unit was deposited near sea-level (Wolfe and Tanai 1980). The margin type for fossil plant species from the Temblor and Cape Blanco floras was recorded, and the percent of entire margins was calculated. The determination of character state was from figures and descriptions ofRenny (1972) and Emerson (this Dissertation). Wolfe and Tanai (1980) reported the percentage of entire species for the Seldovia Point flora. Based on the number of entire leaves, leaf mean annual temperature (LMAT ofWing and Greenwood, 1993) was calculated (Table 1). Table 1: Climate data estimated from fossil florasa Temblor Cape Blanco Seldovia Lat. Tolal Forms Percent Entire LMAT error 36 43 59 31 34 54 21 19 9 21.818.2 6.2 2.6 2.6 1.5 a Latitude is in degrees N. 'Total Forms' is the number ofunique leaftypes that had a single margin state. 'Percent entire' is the number of forms that had a smooth margin. 'LMAT is calculated using the equation of Wing and Greenwood (1993) LMAT = 30.6P + 1.14, where LMAT is leaf estimated mean annual temperature, and P is the proportion of species that have entire margins. 'Error' is one standard deviation and is calculated following Wilf (1997, eq. 4). Geochemical: The early Middle Miocene (15.5 to 16.5 Ma, N-8) benthic foraminifera compilation (n==83) of Savin et al. (1985) was used to estimate late Early Miocene sea surface temperature (SST). To follow the calcite/water 0180 temperature equation of Erez and Luz (1983), it was necessary to estimate the isotopic composition ofthe early Miocene ocean at the sample latitudes ocean by using equation one ofZachos et al. 86 (1994) and adding a -0.5 %0 ice volume correction. Zachos et al. (1994) used the same ice volume correction value for their early Oligocene study. The SST estimates are the circle symbols on the plot of absolute latitude verses temperature (Figure 2). The entire global dataset is used because the northeastern Pacific region Miocene has been little sampled by deep sea drilling (e.g., Woodruff and Savin, 1989). The Holocene temperature measurements also used data compiled by Savin et al. (1985) and fo llowed the same method of SST estimation, however no ice volume correction was used. Modern Observed Temperatures: Modem weather station data for the west coast ofNorth America (Bryson and Hare, 1974) and South America (Schwerdtfeger, 1976) are also plotted on Figure 2. Analysis: The four global data sets are plotted with absolute latitude on the x-axis and temperature CO C) on the y axis, Figure 2A. A subset of the global data sets was also analyzed. The northern hemisphere, temperate latitude (23.3 0 N to 66.6 0 N), portions of the data sets, and their least squares linear regressions are shown in Figures 2B and 2C. Results: Linear regressions of the four global data sets showed no coincident slopes, and student t-test confirmed with greater than 99.9% confidence that the slopes were unique. &7 Fit ure 2A Temperature Estimates for Four Global Datsets x 35 o 0 30 25 ~ 20 111 .. 15::l m .. 10111 l;l. E 5111 I- 0 -5 -10 0 20 40 60 x &FoSSII Lea'./E!s • Miocene Isotope Deri'ved SST x Modern Coastal Temper:. + 47.17 xx" ~ W=0.9937)('"~ y = -O.618x + 38.6£1 "'!'~ R2= 0.9535 "", ;;-- x 2J 4IJ Ell North.rn H