Nordenskioldia and Trochodendron (Trochodendraceae) from the Miocene of Northwestern
Author(s): Steven R. Manchester, Peter R. Crane, David L. Dilcher
Source: Botanical Gazette, Vol. 152, No. 3 (Sep., 1991), pp. 357-368
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/2995219
Accessed: 29/05/2010 15:12
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact email@example.com.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to
BOT. GAZ. 152(3):357-368. 1991.
t 1991 by The University of Chicago. All rights reserved.
0006-807 1/9 1/5203-0007$02.00
FROM THE MIOCENE OF NORTHWESTERN
AND TROCHODENDRON (TROCHODENDCEAE)
STEVEN R. MANCHESTER,' * PETER R. CRANE,t AND DAVID L. DILCHER' *
* Departments of Geology and Biology, Indiana University, Bloomington, Indiana 47405; and
t Department of Geology, Field Museum of Natural History, Chicago, Illinois 60605
The extinct trochodendraceous genus Nordenskioldia,
Hemisphere, is documented for the first time from the Neogene, based upon infructescences, fruits,
associated twigs, and foliage from the Miocene of Idaho, Washington, and southern British Columbia.
The infructescences and fruits, assigned to Nordenskioldia
similar to Paleocene N. borealis, but differ in ranging to a higher number of carpels per fruit and in being
less regularly dehiscent. The leaves, Zizyphoides auriculata (Heer) comb. nov., formerly attributed to
Populus and Cocculus,
are clearly congeneric with the leaves associated with Nordenskioldia
deposits. Zizyphoides auriculata leaves differ from Paleocene Z. flabellum in having generally more
prominent dentations along the margin and a broader divergence of the lateral primary veins. Excellent
preservation of the Miocene material reveals features not preserved in the Paleocene specimens, and in
particular, lignified fruitlets clearly show aborted ovules in addition to the single mature seed. Infruc-
tescences of Trochodendron
are also documented from the same Miocene localities at which N. intergla-
cialis occurs. The close similarities between Paleocene and Miocene species of Nordenskioldia,
between the Miocene and extant species of Trochodendron,
evolution of the Trochodendraceae over intervals of up to 45 million years.
Fossil plants preserved in Miocene lake de-
posits in eastern Washington and adjacent Idaho
1926; BERRY 1929a, 1934; BROWN
1 985a, 1985b) indi-
cate the presence of a temperate hardwood forest
broadly similar in structure and composition to
the present-day mixed mesophytic forests of east-
ern Asia and eastern North America. However,
in addition to genera that now survive in eastern
Asia (e.g., Cercidiphyllum,
sequoia) or elsewhere in the Northern Hemi-
sphere (e.g., Fagus, Hydrangea, Liquidambar,
Liriodendron, Tilia), these fossil floras also con-
tain taxa that are now extinct (e.g., Pseudofagus
[SMILEY and HUGGINS 1981]).
One of the most abundant leaf types in these
Miocene floras was originally identified as POP-
UIUS (Salicaceae [KNOWLTON
1926]), and LA-
MOTTE (1952) listed the many published reports
of these kinds of leaves under the name Cocculus
(Knowlton) Brown (Menisperma-
ceae). Studies of Paleocene floras, however, have
documented the same type of leaf in consistent
association with fizits assigned to the extinct
trochodendraceous genus Nordenskioldia Heer
(KRYSHTOFOVICH 1956; CRANE et al. 1991). The
well represented in the Paleocene of the Northern
(Hollick) comb. nov., are very
suggest relative stasis in the morphological
l Present address: Department of Natural Sciences, Florida
Museum of Natural History, University of Florida, Gaines-
ville, Florida 32611-2035.
Manuscript received June 1990; revised manuscript received
Address for correspondence and reprints: STEVEN R.
Department of Natural Sciences, Florida Mu-
seum of Natural History, University of Florida, Gainesville,
leaves are now assigned to the fossil genus Zi-
zyphoides Seward and Conway (CRANE et al.
1991). The associated shoots and fruiting axes of
have been shown to lack vessels,
and the combined evidence from vegetative and
reproductive characters clearly indicates that
is an extinct genus of Trochoden-
et al. 1990, 1991).
In this paper we review Miocene records of
emphasizing material from
northwestern North America. Nordenskioldia,
which is widespread in the Paleogene of the
Northern Hemisphere, is recognized for the first
time from the Neogene based on fruits, infruc-
tescences, shoots, and leaves. We also present the
first unequivocal evidence for the presence of
in the Tertiary of North America
based on infructescences
and fruits from the Mio-
cene of Idaho. Comparison with extant T. ara-
lioides and with Paleocene N. borealis provides
new information on the history of the Trocho-
dendraceae and documents the importance of the
family in the evolution of temperate vegetation
in the Northern Hemisphere.
Material and methods
The leaves, twigs, and fruits considered here
are preserved as compressions and impressions
in Miocene shales and clays from northern Idaho,
adjacent Washington, and southeastern British
Columbia (fig. 1; App. 1). Most of the localities
are lacustrine interbeds within the Columbia Riv-
er Basalt Group. Localities in eastern Washington
include railroad cuts and brickyard excavations
in the type area of the Latah Formation (PARDEE
and BRYAN 1926; GRIGGS
which produced the flora described by KNOWLTON
1976) near Spokane,
4 ] r
(1926) as well as a site at the north end of Grand
Coulee (BERRY 1931; App. 1). Similar deposits
in Idaho have also been attributed to the Latah
Formation (KIRKHAM and JOHNSON
calities in northern Idaho, including Clarkia (P-
33), Emerald Creek (P-37, 40), and Oviatt Creek
(P-35), have attracted considerable interest be-
cause of the exquisite preservation of their fossil
biota (BOYD 1985; SMILEY
The stratigraphic relationships of the localities
to radiometrically dated basalt flows are pre-
sented in Appendix 1. The Grande Ronde locality
is the oldest, occurring within the Grande Ronde
Basalt which is dated between 15.6 and 16.5 mil-
lion years (REIDEL and HOOPER
the other localities, including those near Spokane
and in northern Idaho, are interbedded with, or
lateral to, basalts of the Priest Rapids Member
of the Wanapum Basalt Formation. The Priest
Rapids Member is radiometrically dated at 14.5
million years (REIDEL and HOOPER
The most southerly localities occur in a slump
block oftuffaceous shales near White Bird, Idaho,
while the excellent specimens figured by HOLLICK
(1915, 1927) are from the Saint Eugene Silts,
Kootenay Valley, on the Saint Mary River in
southeastern British Columbia. Floristic similar-
ities suggest that the deposits in British Columbia
are similar in age to the Latah deposits of Wash-
ington and Idaho (BERRY 1929b). Miocene in-
fructescences and fruits of Trochodendron
observed from Iwate Prefecture, Japan, and from
the Clarkia and Emerald Creek localities in Ida-
Specimens and localities cited are from the fol-
lowing institutions: the United States Geological
Survey (USGS), the United States National Mu-
seumv Washington, D.C. (USNM), the Washing-
ton State Thomas Burke Memorial Museum, Se-
attle (UWBM, collection of E. E. ALEXANDER
formerly housed at the Spokane Museum), the
University of Idaho, Moscow (UIMM), Indiana
University (IU, now housed at University of
Florida, Gainesville), Field Museum of Natural
History, Chicago (FM-PP), Geological Survey of
Canada, Ottawa (GSC), National Science Mu-
seum, Tokyo (NSM-PP).
Measurements of fruits were taken from the
molds left in the clay or silt matrix rather than
directly from the carbonized remains, because
there may be up to 20% shrinkage in the com-
and REMBER 1985 a,
1989). Most of
MOSCOW < }
WASH I NGTON
FIG. 1.-Map showing localities for Nordenskioldia inter-
glacialis. 1, Kootenay Valley, B.C. (HOLLICK
1929b). 2, North end of Grand Coulee, Wash. (USGS 9068).
3, Vicinity of Spokane, Wash. (USGS 7559,7589,7872,7884).
4, Emerald Creek, Idaho (UIMM P-37, P-40).5, Clarkia Race
Track, Clarkia, Idaho (UIMM P-33). 6, Oviatt Creek, Idaho
(UIMM P-35, P-42).7, Juliaetta, Idaho (USGS 8443).8, Oro-
fino, Idaho (USGS 8512).9, White Bird, Idaho (USGS 8444).
See App. 1 for locality details.
pressions from these localities. It should also be
noted that the sediments may shrink 9%-l 5% on
dehydration after collection (SMILEY
1985a, p. 20). A leafof Zizyphoides from Clarkia,
Idaho, was lifted from the matrix, cleared, and
mounted between glass microscope slides by
WILLIAM REMBER to reveal fine venation and
glands of the leaf margin. Leaf impressions from
White Bird, Idaho, also reveal excellent details
of leaf architecture with very low, oblique light-
ing. Complete and variously fractured i-ruits were
cleaned in hydrofluoric acid, washed in water,
and studied with reflected light and scanning elec-
tron microscopy. The seeds and abortive ovules
were examined in situ by scanning electron mi-
croscopy after flaking away the carpel wall with
a needle. The seeds and ovules were then re-
moved, cleared, and mounted on glass slides with
Canada Balsam for light microscopy.
FIGS. 2-12.-Miocene infructescences and fruits of Nordenskioldia interglacialis (Hollick) comb. nov. Fig. 2, Incomplete
infructescence with eight attached fruits, showing scars on the axis where additional fruits were attached, SP&S. cut, Spokane
Wash., UWBM 57254; x 1. Fig. 3, Infructescence with paired fruits remaining attached in lower portion- fruits shed from
upper portion of axis, leaving characteristic umbrella-shaped structures with central columns, Spokane, Wash., UWBM 57253
X 1. Fig. 4, Infructescence previously figured as Equisetum underground stem (KNOWLTON 1926, pl. 9, fig. 1), showing three
nodes with attached fruits. Central node shows the attachment of three fruits, Spokane, Wash. (USGS 7559), USNM 36875
x 1. Fig. 5, Mold of a laterally compressed fruit showing rounded basal scar, segmented morphology, and a row of 10 persistent
styles, Spokane, Wash. (USGS 7589), USNM 435125; x 3.5. Fig. 6, Fruit compressed obliquely showing two rows of seven
or more styles (arrows), White Bird, Idaho (USGS 8444), USNM 435145; x 3.5. Fig. 7, Mold of the base of a fruit showing
elliptical central scar and radial arrangement of fruitlets, Spokane brickyard, Wash., UWBM 57276B x 3.5. Fig. 8, Apical
counterpart of the specimen in fig. 7, showing bilateral symmetry and 20 persistent styles, Spokane brickyard, Wash., UWBM
57276A; x 3.5. Fig. 9, Transversely compressed umbrella-like structure showing basal cup and central column around which
the fruitlets were attached, Spokane, Wash. (USGS 7884), USNM 435126, x 3.5. Fig. 10, Mold of a relatively large fruit with
28 persistent styles, Spokane brickyard, Wash., UWBM 57264; x 3.5. Fig. 11, Base of a fruit showing cycle of marks (not seen
in other specimens) that may represent stamen scars, north end of Grande Coulee, Wash. (USGS 9068), USNM 435142* x
3.5. Fig. 12, Carbonized fruit viewed apically, showing separation between fruitlets, Juliaetta, Idaho, UWBM 57395; x 3.5.
INFRUCTESCENCES AND FRUITS
p. 44, pls. 152, 153.
(Hollick) comb. nov.
1915, p. 44, pls. 152, 153, HOLLICK 1927, p.405,
pls. 34, 35, BERRY 1931, p. 37; Equisetum,
derground stem KNOWLTON 1926, p. 24, pl. 9,
fig. 1, pl. 26, fig. 5, pl. 29, fig. 8; Carpites men-
BERRY 1929a, p. 264; Phyllites amplexicaulis
KNOWLTON 1926, p. 47, pl. 29, fig. 3; Malva?
BERRY 1934, p. 120, pl. 24, fig. 2; Carpolithes
BROWN 1937, p. 185, pl. 63, fig. 2, 3, Carpites
KNOWLTON 1926, p. 49, pl. 26, fig. 4,
KNOWLTON 1926, p. 47, pl. 29, fig. 1 1,
LAMOTTE 1952, p. 95.
SPECIMENS.- The holotype (HOLLICK 1915, pl.
152, 153; GSC 4958) is from the Kootenay Val-
ley, British Columbia. Additional specimens ex-
amined: USNM 36875, 36984, 37005, 37010,
37013, 435125-435135, UWBM 57253-57256,
from Spokane, Washington;
435137-435138, UWBM 57395-57397 from Ju-
liaetta, Idaho; USNM 435142-435143 from
Grand Coulee, Washington; USNM 435144-
from Orofino, Idaho; IU 9247 from Clarkia, Ida-
DESCRIPTION. Infmctescences consist of elon-
gate woody axes, 1.2-3.0 mm in diameter and up
to 95 mm long (longest axis incomplete), with
many sessile fruits (figs. 2, 3). Fruits are attached
singly, or in pairs or triplets (fig. 4), at intervals
of 12-35 mm along the axis. All of the infruc-
tescence specimens are broken at one or both
ends so that the total number of fruiting nodes is
unknown; the longest specimen (fig. 2) shows that
there were up to at least 11 fruiting nodes per
The fruits are subglobose, elliptical in lateral
view (fig. 5), elliptical to circular in transverse
view (figs. 7, 8, 10-12), 9-12.5 mm long (major
axis of symmetry), 9-12 mm broad (minor axis
of symmetry), and 7-8 mm high. Fruits are par-
tially schizocarpic, with 14-29, typically 18-24,
laterally concrescent single-seeded fruitlets ar-
ranged in a whorl around a parenchymatous
tral column (fig. 12). The central column tapers
apically and is bilaterally symmetrical, with reg-
ular longitudinal ridges developed between each
of the adjacent fruitlets. Remains of the recep-
tacle and exocarp persist on the infructescence
axis after dehiscence as an umbrella-like shallow
cup (figs. 3, 9) with a central column. At the base
of the fruit the arrangement
radially symmetrical (fig. 7), but apically the ar-
rangement is bilateral, with persistent abaxially
curved styles arranged in two rows (figs. 6, 8, 10).
In one specimen, the base of the fruit shows a
cycle of circular marks that may be staminal scars
Fruitlets are wedge-shaped in transverse sec-
tion, tapering toward the fruit axis, and are
D-shaped in lateral view. The ventral margin is
more or less straight and the dorsal margin is
convex, with a short abaxially recurved style near
the apex. Oblique bands of fibers are inconspic-
uous on the radial walls. Each fruitlet contains a
single mature seed (fig. 15) and up to six abortive
ovules (figs. 16, 17).
The seeds are anatropous and pendulous from
the apex at the ventral side of the locule (fig. 15).
The single seed within each fruitlet is ovate in
outline, 3.5 mm wide, 7 mm long, and rounded
at the chalazal end but tapered toward the mi-
cropyle and hilum. Seeds are flattened parallel to
the median plane of the fruitlet, and the testa is
extended to form a wing surrounding
seed body except at the hilar end. The raphe runs
directly from the hilum to the chalaza without
forming a hairpin loop. The seed surface is retic-
ulate (figs. 15, 16), resulting from a layer of large
cuboidal cells, 0.074.09 mm diameter, forming
the outer layer of the seed coat.
of fruitlets is virtually
correspond closely in structure and morphology
to those of the genus Nordenskioldia
from well-preserved Paleocene material (CRANE
et al. 1990, 1991). Diagnostic features ofthe ge-
nus include elongate fruiting axes with numerous
sessile fruits; fruits composed of many single-
seeded, wedge-shaped fruitlets arranged in a
whorl; fruitlets with an abaxially recurved per-
sistent style near the apex on the dorsal side; and
conspicuous umbrella-like structures that persist
on the infructescence axes after the fruits are shed.
from the Paleogene N. borealis
ber of carpels per fruit and apparent differences
in dehiscence. The range in carpel number (14-
29) greatly exceeds that observed in Paleocene
specimens (12-19). While isolated fruitlets and
seeds are common at the Paleocene and Eocene
localities and were evidently shed individually,
no isolated seeds and only a single isolated fruitlet
(UWBM 57424) have been found in the extensive
collections from Miocene localities. In contrast,
complete fruits with all of the fruitlets intact, but
detached from the infructescence axis, are rela-
tively common at the Miocene localities, sug-
gesting that infructescences of N. interglacialis
shed whole fruits rather than individual fruitlets.
fruiting axes and fruits
based on the num-
MANCHESTER ET AL.-MIOCENE TROCHODENDRACEAE
FIGS. 13-17.-Fruits and seed of Nordenskioldia
between adjacent carpels; most of the exocarp, except near the apex, has fractured away (note styles near the apex); reflected
light photography, x 10. Fig. 14, The same specimen split longitudinally, showing remains of central column and adjoining
fraitlets; SEM, x 10. Fig. 15, Same specimen, showing smooth lateral wall of fizitlet partially removed, revealing a single
winged seed with reticulate surface, x 12. Fig. 16, Same specimen, showing the apex of seed: note large cells comprising the
seed coat, and flattened elliptical aborted ovules, x 50. Fig. 17, Aborted ovule removed from the cluster in fig. 16; transmitted
light, x 85.
Clarkia, Idaho, IU 9297. Fig. 13, Lateral view offruit splitting
In addition, the absence of isolated seeds among
the Miocene collections suggests that the fruitlets
were not regularly dehiscent. Prominent oblique-
ly oriented fiber bands, which surround the vas-
cular bundles, occur on the lateral walls of each
fruitlet in N. borealis but have not been observed
in N. intergZaciaZis material. Although this may
partly reflect differences in the preservation ofthe
Paleocene and Miocene specimens, it may also
reflect a difference in dispersal system if these
bands functioned as part ofthe fruitlet dehiscence
Zizyphoides auriculata (Heer) new comb.
BASIONYM. Hedera auriculata HEER 1869, p.
36, pl.9, fig.6.
TON 1926, p.30, pl.12, figs.8 - 10, pl.13, figs. l -
7, pl. 14, figs. 1-3, pl. 15, figs. 3-5; P. fairfi
KNOWLTON 1926, pl. l5, fig.2, pl.16, figs.1 - 3;
Cocculus heteromorpha aKnowlton)
Populus heteromorpha BOWL-
p. 352, BOYD 1985, p. 94, pl. 9, figs. 6-8, pl. 10,
figs. 1, 2, LAMoTTE 1952, p. 129, TANAI 1961,
p.324, pl.21, fig.7; Cebatha multiformis
1927, p. 406, pl. 38, figs. 1-6, pl. 39, figs. 1-3;
(Knowlton) BERRY 1931,
p.37; probaby also Cissampelos dubiosa HOLLICK
fig. 6) is from English Bay, Cook Inlet, Alaska.
Additional specimens include: USNM 36903-
36914, 36920-36926 from Spokane, Washing-
ton; UIMM 0150-0152 from Oviatt Creek, Ida-
ho; EM-PP43584 from White Bird, Idaho; IU
9592-9594 from Clarkia, Idaho.
holotype (HEER 1869, pl. 9,
DEscRIPrIoN.-The leaves are simple and long-
petiolate, with the petiole nearly as long as the
lamina (fig. 20). The lamina is highly variable in
shape, ranging from ovate to ovate-elliptical to,
less commonly, obovate; frequently it is more or
less deltoid (figs. 18-22). The lamina is typically
20-85 mm long, 22-100 mm wide (BOYD 1985),
with a length-to-width ratio of 0.75-1.25 (average
1.00). The apex of the lamina is more or less
obtuse to rounded, with a slight retuse glandular
incision at the very tip (fig. 23). The leaf base
varies from rounded or very obtusely wedge-
shaped to nearly flat or cordate. The margin is
irregularly toothed, with large, often glandular,
crenations, typically four on each side of the lam-
ina, less commonly, faintly undulate or entire.
The lowest crenations are typically in the lower
third ofthe lamina. The crenations have a distinct
central vein running to the tip and a pair of lateral
accessory veins that converge into the apical gland
(figs. 24, 25).
The venation is actinodromous, with relatively
thin primary veins that are somewhat wavering
in course and are frequently deflected at the points
of origin ofthe secondary veins (figs.18,22). The
midvein runs directly to the leaf apex, flanked by
one or two pairs of successively thinner lateral
primary veins. The inner pair of lateral primary
veins typically are not curved inward toward the
leaf apex and usually join with a secondary vein
originating from the midvein to form a margina]
loop. The outer pair of primary veins are dis-
tinctly thinner than the inner pair, more or less
follow the curvature of the lateral leaf margins,
and usually terminate at the margin in the lower
third of the lamina. The midvein produces two
or three pairs of alternate secondary veins, with
the lowest pairs initiated at or slightly below the
midpoint of the lamina; weaker, intersecondary
veins originate at broader angles than the sec-
ondaries. Lateral primaries produce secondary
veins only abmedially, and together with second-
ary veins produced from the midrib, these form
distinct angular brochidodromous loops well
within the margin. Outer lateral primaries pro-
duce tertiary veins that form smaller angular
Tertiary veins between adjacent primaries are
percurrent, often forming arched or inverted
V-shaped connections. Abmedial tertiary veins
produced by secondary veins and lateral primar-
ies in the apical third of the lamina form small
loops just within the margin. Quaternary veins
arise orthogonally, forming a rectangular to po-
lygonal network. Quinternary veins also arise or-
thogonally and delimit rectangular to polygonal
areoles up to 0.34.6 mm diameter. Areoles have
straight to curved freely ending veinlets that
branch one to three times (fig. 26).
DISCUSSION.-Zizyphoides auriculata leaves are
common at most of the Miocene localities from
which Nordenskioldia interglacialis is known
(App. 1) and have also been recovered from the
Miocene of southern Alaska (WOLFE and TANAI
1980) and Japan (TANAI 196 1; WOLFE and TANAI
1980). They resemble the leaves of Z. flabellum
associated with N. borealis from the Paleocene of
North America, &reenland, Spitsbergen,
the Paleocene and Eocene of eastern Asia in the
long narrow petiole, the irregularly crenate mar-
gin, the retuse, glandular leaf apex, the actino-
dromous venation consisting of a midvein and two
pairs of thinner lateral primaries, and the fine
areolation with branched, freely ending veinlets.
However, Z. auriculata leaves can generally be
distinguished from those of Z. flabellum by fea-
tures of the margin and venation. The marginal
FIGS. 18-28.-Leaves of Zizyphoides
and crenate margin, White Bird, Idaho, FM-PP43584- x 1. Fig. 19, Compression previously figured by KNOWLTON 1926, pl.
13, fig.1, Spokane, Wash. (USGS 7559), USNM 36905; x 1. Fig.20, Compression showing typical long petiole, original figured
by KNOWLTON 1926, pl. 13, fig. 4, Spokane (USGS 7559), USNM 36908; x 1. Fig. 21, Leaf with prominent crenations Clarkia
Idaho (UIMM P-33), UIMM 350/IU 9592; x 1. Fig. 22, Leaf with well-preserved venation, Clarkia (UIMM P-33j, UIMM
351/IU 9593; x 1. Fig. 23, Retuse apex of leaf in fig. 22; x 2. Fig. 24, Isolated cleared leaf showing details of higher-order
venation and glandular crenation; note lateral veins converging into the glandular apex of tooth, Clarkia (UIMM P-33), UIMM
353/IU 9594; x 10. Fig. 25, Same leaf as in fig. 18, magnified to show details of tooth and marginal venation- x 8. Fig. 26
Same leaf as in fig. 18, magnified to show five orders of venation and branched freely ending veinlets; x 10. Fig. 27, Twig
associated with Zizyphoides
Impression of short shoot from fig. 27 (protruding into the sediment), showing a pair of lateral leaf scars (lower left and right),
one clearly showing the three vascular bundle scars (arrows); x 6.
auriculata (Heer) comb. nov. Fig. 18, Impression showing actinodromous venation
auriculata, showing opposite short shoots, Spokane (USGS 7559), USNM 435147; x 1. Fig. 28,
* ire'';,S A
8 f e . v # st
;^:e v :;.
/ : . s jt
18 Si. ,,,j. - \v-efe)
* 22¢. ;,,r,.X:-
- w _. i.
. :.J :
£ B: ;
- w:s ;X
* tN i.,.,4 q
. ; !s' ,4
20 - ' j
-*sl;j ' ,F,<e
crenations of Z. auriculata
nent, with glandular tips, and begin in the lower
one-third of the lamina. In contrast, Z. flabellum
leaves are frequently entire margined and, when
toothed, have more poorly developed crenations
that are often nonglandular and typically con-
fined to the upper half of the lamina. Also, in Z.
flabellum the inner pair of lateral primary veins
frequently curve admedially as they approach the
apex. In Z. auriculata
these veins typically follow
a more abmedial course, and the pinnate sec-
ondary veins arising from the midvein are more
With regard to leaves of Z. auriculata,
KNOWLTON (1926, p. 30) commented that this
form is represented by approximately 200 spec-
imens, and it is without exception one ofthe most
polymorphous species that I have ever studied."
His illustrations document this range of varia-
tion, including narrow to broad leaves and a wide
range in the number and spacing of crenations
on the margin. This polymorphisin may reflect
differences between long and short shoot leaves,
as in extant Cercidiphyllum.
Isolated twigs showing differentiation
and short shoots occur together with Zizyphoides
leaves and Nordenskioldia
from the Spokane brickyard (e.g., fig. 27), but
because they are preserved as molds, with the
carbonaceous contents missing, it is not possible
to examine internal anatomy. However, their
morphology is similar to that of long and short
shoots associated with Paleogene Nordenskioldia
in which anatomical information is preserved
(CRANE et al. 1 990, 1 99 1). Although Zizyphoides
leaves have not been found attached to these
shoots, the anatomical similarity between per-
mineralized twigs and the infructescence axes of
from the Paleocene of North Da-
kota supports the inference that they were pro-
duced by the same biological entity. Like the Pa-
leocene shoots, the Miocene short shoots have
diamond-shaped leaf scars, some of which show
a row of three circular marks that may indicate
the position of three vascular bundles (fig. 28).
infructescences, and the distinctive
shoots described above have not been found in
organic connection. However, their consistent co-
occurrence at more than 10 Miocene localities
(App. 1) strengthens the association evidence
based on congeneric Paleocene material that these
different organs are produced by the same plant
species (CRANE et al. 1 99 1). We predict that future
studies also will reveal Nordenskioldia
calities in Alaska and Japan from which Z. au-
are typically promi-
fruits in the collections
at the lo-
riculata has been recovered. During the Paleo-
gene Zizyphoides and Nordenskioldia occur
together at numerous localities, but the interpre-
tation of such patterns of co-occurrence is com-
plicated by the presence of extinct Cercidiphyl-
laceae with leaves very similar to those of
(CRANE et al. 1991). The apparent
absence of such extinct cercidiphyllaceous
from Neogene floras makes the association evi-
dence from Miocene localities particularly com-
pelling. The lignitic preservation of many of the
Miocene specimens also provides new structural
data complementary to those obtained from Pa-
leocene impression, cast, and silicified material.
Most significantly, the presence of several abor-
tive ovules in addition to the single enlarged seed
per locule is clearly seen in the Miocene speci-
mens although it has not yet been observed in
the Paleocene material.
Although reconstructions of the Nordenskiol-
dia plants are now relatively complete with regard
to infructescences, fruits, seeds, leaves, and shoot
systems, neither the Paleogene nor the Miocene
material has provided direct information on the
androecium and pollen, although a single speci-
men (fig. 1 1) suggests that the flowers may have
had numerous stamens. Detailed morphological
comparison of Nordenskioldia
of Trochodendrales based on the evidence cur-
rently available is presented by CRANE et al.
Comparison with Nordenskioldia
While Nordenskioldia borealis from the Paleo-
cene and Eocene and N. interglacialis
Miocene are sufficiently similar to warrant con-
generic status, morphological differences in both
fruits and foliage justify the recognition of two
distinct but variable species. The probable foliage
of N. interglacialis (Zizyphoides
more numerous and more distinct crenations on
the leaf margin, and the lateral primary veins are
generally straighter and not prominently curved
toward the leafapex. In addition, N. interglacialis
typically has more fruitlets per fruit than the Pa-
leocene and Eocene species. The rarity of dis-
persed fruitlets and seeds at Miocene localities,
combined with the common occurrence of dis-
persed fruits with all the fruitlets intact, suggests
that the whole fruit may have been the primary
unit of dispersal in the Miocene species.
The morphological variability in both the Pa-
leogene and Miocene species is such that there is
some overlap between them. Occasional Paleo-
cene leaves have prominent crenations typical of
the Miocene material, while some Miocene leaves
have more subtle crenations as are common in
the Paleocene species. Similarly, there is overlap
in the number of fruitlets per fruit, so that a fruit
with extant genera
MANCHESTER ET AL.-MIOCENE TROCHODENDRACEAE
latest Cretaceous and early Tertiary by a variable
complex of closely related plants (Nyssidium/Jof-
frea complex) that include Nyssidium (CRANE
(CRANE and STOCKEY
Although most closely related to extant Cercidi-
phyllum, these extinct genera are distinguished
by differences in the structure of infructescences,
fruits, and seeds (CRANE and STOCKEY 1985,
1986). Tertiary leaves and fruits corresponding
closely to the extant genus have been assigned to
(Heer) Jahnichen, Mai, and Walther,
respectively, and are known from the Lower Oli-
gocene of Oregon (MANCHESTER
1987), Miocene of Idaho (SMILEY
1985b), and Miocene to Pliocene of Europe (JAH-
NICHEN et al. 1980). The presence of Cercidi-
phyllum in the Miocene of Idaho is clearly dem-
onstrated by a fossil twig with attached leaves
and a cluster of fruits closely similar to extant
illustrated by SMILEY
sER (1985b, pl. 4, fig 8)
Taken as a whole, the fossil record of the
Trochodendraceae and Cercidiphyllaceae is sur-
prisingly extensive and, unlike that of many an-
giosperm groups, includes several fossil plants
that are well understood from both vegetative
and reproductive structures. These data contrib-
ute to more reliable estimates of rates of change
in the angiosperm fossil record than is provided
by studies of single organs, and in the history of
both the Trochodendraceae and Cercidiphylla-
ceae the striking pattern is one of stasis. Based
on the material presented in this article, it is clear
that while minor (possibly anagenetic) morpho-
logical change may have occurred during the in-
terval of about 45 million years between Paleo-
cene Nordenskioldia borealis and Miocene N.
there is no evidence of any grad-
ualistic trend toward the morphology of the ap-
parent sister genus Trochodendron.
Miocene remains of Trochodendron,
million years old, are sufficiently similar to extant
to raise the question of whether these
fossils warrant a distinct specific epithet. In the
from the Oligocene and Miocene
(JAHNICHEN et al. 1980) is apparently indistin-
guishable in both foliage and fruits from the ex-
tant species. The basic structure and organization
of plants in the Nyssidium/JoWrea
shows little change over an interval of about 30
million years between their first appearance in
the latest Cretaceous-earliest Paleocene and their
most recent records in the late Eocene (BECKER
1961; CRANE 1984).
An intriguing question raised by the pattern of
stasis in these two angiosperm families is how
(Heer) Brown and C.
m crenatu m/ C.
with 14 fruitlets would be normal for the Paleo-
cene species but would be at the low end of the
range for the Miocene species. It is only in the
context of large collections from several localities
that we have been able to distinguish the Paleo-
gene and Miocene species.
Fossil history of the Trochodendrales
The order Trochodendrales, comprising the
Trochodendraceae, Cercidiphyllaceae, Euptele-
aceae, and Myrothamnaceae (ENDRESS
cupies a critical position in angiosperm phylog-
eny as a probable basal, or near basal, grade within
the clade of nonmagnoliid (triaperturate) dicot-
yledons (CRANE 1989). The group has an ancient
fossil record, broadly consistent with its phylo-
genetic position. "Trochodendrophyll" leaves first
appear in the late Albian, and Spanomera
dinensis, of probable buxaceous affinities from
the early Cenomanian (late Cretaceous), also
shows many features that are characteristic of the
Trochodendrales (DRINNAN et al. 1991).
Nordenskioldia borealis provides the earliest
unequivocal evidence of the Trochodendraceae
and is interpreted as closely related to extant
based on the semi-inferior gy-
noecium consisting of numerous whorled carpels
(CRANE et al. 1991). Fruits assigned to Trocho-
from the Lower Eocene
floras of the London Clay (REID and CHANDLER
1933) are of uncertain relationship although leaves
combining features of Trochodendron
centron have been recorded from the Eocene of
Washington (WOLFE 1989). Vesselless trocho-
dendralean woods assigned to the genus Trocho-
closely resemble that of Trocho-
and are known from the Eocene (SCOTT
and WHEELER 1982) and Upper Oligocene (HER-
GERT and PHINNEY 1954) of Oregon. Records of
in the fossil record (e.g., SUZUKI
1967; OZAKI 1987) are based on leaves without
corroborative fruiting remains.
Fossil infructescences and fruits of Trochoden-
dron, illustrated here for the first time, are known
from the Miocene of Clarkia and Emerald Creek,
Idaho (figs. 29, 31), and from the late Miocene
of Yonosawa, Shizukuishi Basin, Iwate Prefec-
ture, Japan (figs.30,33; locality in UEMURA
pp. 70-71), and appear to be indistinguishable
from those of extant T. aralioides
Trochodendron is represented by only a few in-
fructescences at Clarkia and Emerald Creek and
by a single partial leaf from Clarkia (C. J. SMILEY,
personal communication.) and has not been ob-
served at other localities in Idaho or Washington.
Leaves from Japan associated with the Trocho-
infructescences have been assigned to T.
Murai (UEMURA 1988).
The Cercidiphyllaceae are represented in the
(figs. 32, 34).
FIGS. 29-34.-Infructescences and fruits of Miocene and extant Trochodendron.
fruits on long pedicels, Clarkia, Idaho (UIMM P-33), UIMM 354/IU 9575; x 1.5. Fig. 30, Infructescence with at least six fruits
on long pedicels, late Miocene of Yonosawa, Shizukuishi Basin, Iwate Prefecture, Japan, NSM-PP 15854; x 1.5. Fig. 31, Fruit
viewed laterally, showing acute base, truncate apex, persistent styles; note elliptical dark area representing remains of nectar-
iferous tissue (arrow) comparable to that in extant fruits, Clarkia (UIMM P-33), UIMM 354/IU 9596; x 3. Fig. 32, Fruit of
extant T. aralioides viewed laterally, showing acute base, truncate apex, adaxially recurved styles and deltoid-elliptical pads
indicating the remains of nectariferous tissue; x 3.5. Fig. 33, Fossil fruit associated with infructescence in fig. 30, Yonosawa,
Iwate Prefecture, Japan, NSP-PP 15854; x 3. Fig. 34, Same specimen as in fig. 32, tilted to show radial arrangement of carpels;
Fig. 29, Infructescence showing at least eight
taxa such as Nordenskioldia, Trochodendron,
Nyssidium/Joffirea, and Cercidiphyllum have
managed to persist, apparently with only minor
morphological change, in the face of major cli-
matic and vegetational changes over long periods
of geologic time. In part, the long-term survival
of these plants may be due to their ability to
disperse uridely and thus to establish themselves
opportunistically in an appropriate habitat, but
it is also clear from the history of Nordenskioldia
that these taxa are capable
of surviving in various ecological situations. Both
are wldespread in low-diversity Paleocene vege-
tation but also persist in the more diverse vege-
tation ofthe late Eocene (Nyssidium/Joffrea)
The Trochodendraceae and Cercidiphyllaceae
appear to have achieved their zenith, in terms of
abundance, during the Paleogene with the max-
imum spread of Nordenskioldia
JoXrea. The Nyssidium/Joffrea
ently became extinct at around the beginning of
the Oligocene but the Miocene mixed mesophytic
forests bordering the North Pacific, including
northern Idaho, Washington, British Columbia,
southern Alaska, and Japan, provided the last
known refuge for"Nordenskioldia plants." In
North America these Miocene forests in Idaho
also provided the last known refuge of Trocho-
dendron and Cercidiphyllum.
We thank J. P. FERRIGNO,
SMILEY, K. UEMURA, W. WEHR, S. L. WING, and
J. A. WOLFE
for providing access to fossils studied
in this investigation. W. REMBER
in interpreting the stratigraphy and age of the
localities. A. N. DRINNAN provided assistance
with scanning electron microscopy. H. E. SCHORN
provided helpful review comments. This project
was supported in part by National Science Foun-
dation grants EAR 8707523 and EAR 8904234
to S. R. MANCHES1ER, BSR 8314592 and BSR
8708460 to P. R. CRANE, and DEB 7910720 and
BSR 8516657 to D. L. DILCHER.
W. REMBER, C. J.
Miocene localities for Nordenskioldia
on co-occurrence of Zizyphoides
shoots. General collection refers to uncataloged spec-
imens in floristic collections of the U.S. Geological
Survey housed in the Department of Paleobiology,
Smithsonian Institution, Washington, D.C. Comments
on the stratigraphic position of the localities in Wash-
ington and Idaho, based on recent fieldwork, are cour-
fruits with notes
foliage and short
MANCHESTER ET AL.-MIOCENE TROCHODENDRACEAE
tesy of WILLIAM
1991). The basalt stratigraphy and ages are presented
by BOND (1963), HOOPER et al. (1984), REIDEL
and REIDEL and HOOPER (1989).
REMBER (personal communication
IJSGS 7559: Cut on Spokane, Portland & Seattle
railway; fruit and leaf figured by KNOWLTON
29, fig. 3. Leaves in general collection. Overlain by
Priest Rapids Member of the Wanapum Basalt.
USGS 7589: SPtS railway cut no. 1, Spokane, coll.
H. FAIR; fruit figured by KNOWLTON
11. Leaf, fruits in general collection. Same stratigraphy
as USGS 7559.
USGS 7872: Brickyard, Spokane, about 1 mi SW of
the high Latah Creek bridge between tracks of OWR
and SP&S railway, coll. KNOWLTON
Leaves, fruits, twigs with short shoots in general col-
lection. Same stratigraphy as USGS 7559.
USGS 7884: Brickyard, Spokane, Latah Formation,
coll. E. E. ALEXANDER 1927, 1929. IEruits in general
USGS 9068: North end of Grand Coulee, coll. BONSER
et al. 1927. Leaf, fruits in general collection. Between
the R2 and N1 units of the Grande Ronde Basalt.
1926, pl. 29, fig.
and BONSER 1926.
USGS 8443: 2.5 mi NNE of Arrow Junction, on road
to Julieatta, along Potlach Creek. Exposure between
basalt flows of Latah, coll. BROWN and McGLosHEN
1934. Leaves, fruit in general collection. Between the
uppermost Imnaha and the Grande Ronde R, unit.
USGS 8444: 2.5 mi NE of White Bird, on road to
Grangeville, coll. R. W. BROWN 1934. Leaves, fruits
in general collection. This locality is interpreted as a
slump block with shales probably between the upper-
most Grande Ronde in the area (R2) and the Priest
Rapids Member of the Wanapum Basalt (Basalt of
Lolo) (J. C. BoND, R. REID, and W. C. REMBER,
USGS 8512: Latah in vicinity of Orofino, coll. BOYD
OLSON 1936. Leaves, fruits in general collection. Be-
tween the uppermost Imnaha and Grande Ronde R,
UIMM-P33: Clarkia Race Track locality (SMILEY
REMBER 1985 a). Result of damming by flows of Priest
Rapids Member of the Wanapum Basalt Formation.
UIMM-P37, P40: Emerald Creek (SMILEY
sER 1985a). Result of damming by flows of Priest Rap-
ids Member of the Wanapum Basalt Formation.
UIMM-P35, P42: Oviatt Creek (SMILEY
1985a). Between flows of the Priest Rapids Member
of the Wanapum Basalt.
1927) St. Eugene Silts Kootenay Valley (HOLLICK
(see also BERRY 1929b).
River Basin, southwestern Montana. Geol. Soc. Am. Mem.
BERRY, E. W. 1929a. A revision of the flora of the Latah
Formation. U.S. Geol. Surv. Prof. Pap. 154-H:225-265.
1929b. The age ofthe St. Eugene silt in the Kootenay
Valley, British Columbia. Trans. R. Soc. Canada, 3d Ser.,
1931. A Miocene flora from Grand Coulee, Wash-
ington. U.S. Geol. Surv. Prof. Pap. 170-C:31-42.
1934. Miocene plants from Idaho. U.S. Geol. Surv.
Prof. Pap. 185-E:97-125.
BoND, J. G. 1963. Geology of the Clearwater Embayment,
Idaho. Idaho Bureau Mines Geol. Pamphlet 128, pp. 1-83.
BoYD, A. E. 1985. A Miocene flora from the Oviatt Creek
Basin, Clearwater County, Idaho. Master's thesis. Univer-
sity of Idaho, Moscow.
BROWN, R. W. 1937. Additions to some fossil floras of the
western United States. U.S. Geol. Surv. Prof. Pap. 186-J:
1946. Alterations in some fossil and living floras.
J. Wash. Acad. Sci. 36:344-355.
CRANE, P. R. 1984. A re-evaluation of Cercidiphyllum-like
plant fossils from the British early Tertiary. Bot. J. Linn.
1989. Paleobotanical evidence on the early radiation
of nonmagnoliid dicotyledons. Plant Syst. Evol. 162: 165-
CRANE, P. R., and R. A. STOCKEY. 1985. Growth and re-
productive biology of JoXrea speirsii gen. et sp. nov., a
Cercidiphyllum-like plant from the late Paleocene of A1-
berta. Can. J. Bot. 63:340-364.
H. F. 1961. Oligocene plants from the upper Ruby
1986. Morphology and development of pistillate
inflorescences in extant and fossil Cercidiphyllaceae. Ann.
Mo. Bot. Gard. 73:382-393.
CRANE, P. R., S. R. MANCHESTER,
A preliminary survey of fossil leaves and well-preserved
reproductive structures from the Sentinel Butte Formation
(Paleocene) near Almont, North Dakota. Fieldiana, Geol.,
1991. Reproductive and vegetative structure of Nor-
denskioldia (Trochodendraceae), a vesselless dicotyledon
from the early Tertiary of the Northern Hemisphere. Am.
J. Bot., vol. 78.
DRINNAN, A. N., P. R. CRANE, E. M. FRIIS, and K. R. PE-
DERSEN. 1991. Angiosperm flowers and tricolpate pollen
of buxaceous affinity from the Potomac Group (mid-Cre-
taceous) of eastern North America. Am. J. Bot. 78:153-
ENDRESS, P. 1989. The systematic position of the Myro-
thamnaceae. Pages 193-200 in P. R. CRANE and S. BLACK-
MORE, eds. Evolution, systematics, and fossil history of the
Hamamelidae. Vol. 1. Systematics Assoc. Spec. Vol. No.
40B. Clarendon, Oxford.
GRIGGS, A. B. 1976. The Columbia River Basalt Group in
the Spokane Quadrangle, Washington, Idaho and Montana.
U.S. Geol. Surv. Bull. 1413:1-39.
HEER, O. 1869. Flora Fossilis Alaskana. Kongl. Svensk.
Vetens-Akad. Hand.8(4): 14 1. (Reprinted as Flora fossilis
Arctica. 2(2):1-192 [F. Schulthess, Zurich].)
HERGERT, H. L., and H. K. PHINNEY. 1954. Trochoden-
droxylon beckii gen. et sp. nov. from the Tertiary of Oregon.
Bull. Torrey Bot. Club 81: 118-122.
HOLLICK, A. 1915. A new fossil species of Ficus and its
and D. L. DILCHER. 1990.
climatic significance. J. N.Y. Bot. Gard.16:4347, pls.152,
1927. The flora of the Saint Eugene Silts, Kootenay
Valley, British Columbia. Mem. N.Y. Bot. Gard. 7:389-
HOOPER, P. R., W. D. KLECK, C. R. KNOWELS,
and R. L. THIESSEN. 1984. Imnaha Basalt, Columbia Riv-
er Basalt Group. J. Petrology 25:473-500.
JAHNICHEN, H., D. H. MAI, and H. WALTHER.
und Fruchte von Cercidiphyllum
mitteleuropaischen Tertiar. Schriften. Geol. Wiss. 16:357-
KIRKHAM, V. D., and M. M. JOHNSON. 1929. The Latah
Formation in Idaho. J. Geol. 37:483-504.
KNOWLTON, F. H. 1926. Flora of the Latah Formation of
Spokane, Washington, and Coeur d'Alene, Idaho. U.S. Geol.
Survey Prof. Pap. 140-A: 17-82.
KRASSILOV, V. A. 1973. Mesozoic plants and the problem
of angiosperm ancestry. Lethaia 6: 163-178.
1977. The origin of angiosperms. Bot. Rev. (Lan-
KRYSHTOFOVICH, A. N. 1956. Oligocene flora of Mount
Ashutus in Kazahkstan. Paleobotanica Fasciculus 1. Insti-
tutum Botanicum nomine V. L. Komarovii Academiae
Scientarium URSS. ACTA Series 8. Moscow.
LAMoTTE, R. S. 1952. Catalogue of the Cenozoic plants of
North America through 1950. Geol. Soc. Am. Mem. 51:1-
MANCHESTER, S. R., and H. W. MEYER. 1987. Oligocene
fossil plants of the John Day Formation, Fossil, Oregon.
Oregon Geol. 49: 115-127.
OZAKI, K. 1987. Tetracentron
Japan. Trans. Proc. Palaeont. Soc. Japan, N.S. 146:77-87.
PARDEE, J. T., and K. BRYAN. 1926. Geology of the Latah
Formation in relation to the lavas of the Columbia Plateau
near Spokane, Washington. U.S. Geol. Surv. Prof. Pap.140-
S. P. REIDEL,
Siebold & Zuccarini im
leaves from the Neogene of
REID, E. M., and M. E. J. CHANDLER.
Clay flora. British Museum (Natural History), London.
REIDEL, S. P. 1978. Stratigraphy and petrogenesis of the
Grande Ronde Basalt in the Lower Salmon and adjacent
Snake River Canyons. RHO-SA-62. Rockwell Hanford Op-
erations, Richland, Wash.
REIDEL, S. P., and P. R. HOOPER,
tectonism in the Columbia River flood basalt province.
Geol. Soc. Am. Spec. Pap. 239. Boulder, Colo.
SCOTT, R. A., and E. A. WHEELER.
the Eocene Clarno Formation of Oregon. IAWA Bull. 3:
SMILEY, C. J., and L. M. HUGGINS. 1981. Pseudofagus ida-
hoensis n. gen. et sp. (Fagaceae) from the Miocene Clarkia
flora of Idaho. Am. J. Bot. 68:741-761.
SMILEY, C. J., and W. C. REMBER. 1985a. Physical setting
of the Miocene Clarkia fossil beds, northern Idaho. Pages
11-31 in C. J. SMILEY, ed. Late Cenozoic history of the
Pacific Northwest. Pac. Div., Am. Assoc. Adv. Sci., San
1985b. Composition of the Miocene Clarkia flora.
Pages 95-112 in C. J. SMILEY,
the Pacific Northwest. Pac. Div., Am. Assoc. Adv. Sci., San
SUZUKI, K. 1967. Discovery of Tetracentron leaves from the
Neogene in Japan. Proc. Jpn. Acad. 43:526-530.
TANAI, T. 1961. Neogene floral change in Japan. J. Hok-
kaido Univ. Fac. Sci. Ser. 4(11): 119-298.
UEMURA, K. 1988. Late Miocene floras of Northeast Hon-
shu, Japan. Bull. Natl. Sci. Mus., Tokyo Ser. C, 14:1-197.
WOLFE, J. A. 1989. Leaf-architectural analysis ofthe Hama-
melididae. Pages 75-104 in P. R. CRANE
eds. Evolution, systematics, and fossil history of the Ha-
mamelidae. Vol. l. Systematics Assoc. Spec. Vol. No.40B.
WOLFE, J. A., and T. TANAI. 1980. The Miocene Seldovia
Point flora from the Kenai Group, Alaska. U.S. Geol. Surv.
Prof. Pap. 1105:1-47.
1933. The London
eds. 1989. Volcanism and
1982. Fossil woods from
ed. Late Cenozoic history of
and S. BLACKMORE,