Phylogenetic Relationships and Taxonomic Status Of the Paleoendemic Fagaceae Of Western North America: Recognition Of A New Genus, Notholithocarpus

2008·8.3 MB·English

by  Paul S Manos

Madrono, Vol. 55, No. 3, pp. 181^190, 2008
PHYLOGENETIC RELATIONSHIPS AND TAXONOMIC STATUS OF THE
PALEOENDEMIC FAGACEAE OF WESTERN NORTH AMERICA:
RECOGNITION OF A NEW GENUS, NOTHOLITHOCARPUS
Paul Manos
S.
Department of Biology, Duke University, Box 90338, Durham, NC 27708
[email protected]
Charles H. Cannon
Department ofBiological Sciences, Texas Tech University, Box 43131, Lubbock, TX 79409;
Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun,
Mengla, Yunnan 666303, China
[email protected]
Sang-Hun Oh
L. H. Bailey Hortorium, 412 Mann Library, Cornell University, Ithaca, NY 14853
[email protected]
Abstract
We investigated the phylogenetic relationships and taxonomic status ofthe castaneoid component
{Lithocarpus and Chrysolepis) of the family Fagaceae that is endemic to the California Floristic
DNA
Province (CA-FP). Over 7800 basepairs ofnuclear and chloroplast were analyzed in 17 taxa
representing the breadth ofphylogenetic diversity in the family. The genus Lithocarpus^ as currently
defined, isclearlypolyphyleticduetotheinclusion ofL. densiJJorus. Here, wedesignatethistaxonasa
new genus, Notholithocarpus, which can be recognized morphologically by its relatively small,
subprolate pollen. Notholithocarpus is more closely related to Quercus, Castcinea, and Castcmopsis;
Chrysolepis was resolved as the sister group to Lithocarpus sensu stricto. These results indicate that
Notholithocarpus does not possess true 'flower cupules,' which define Lithocarpus sensu stricto, but
liketheoaks, thesinglefiowerpercupuleisderivedthroughtheabortionoflateralflowerswithineach
cupule. Further study is required to confirm this characteristic. A formal taxonomic treatment is
presented with new combinations.
Key Words: California Floristic Province, Chrysolepis, Fagaceae, Lithocarpus, Notholithocarpus,
phylogeny, sudden oak death, tanbark oak.
The Cahfornia Floristic Province (CA-FP) of florus var. densiflorus (Hook. & Arn.) Rehder, a
western North America is rich in paleoendemic tree that reaches 45 m, often occurs as a co-
plant species that highlight the botanical legacy dominant in the redwood and mixed evergreen
of the region and stimulate discussion on the forests ofthe north coast ranges; its shrub form,
affinities ofmany ofthese putatively isolated taxa L. densiflorus var. echinoides (R. Br. ter) Abrams
(Stebbins and Major 1965; Raven and Axelrod is more common in open conifer forests and dry
1978). Among these taxa are representatives of slopes of the northern interior CA-FP at higher
the largely tropical East Asian castaneoid sub- elevations. The giant chinquapin, Chrysolepis
family of the Fagaceae: Chrysolepis Hjelmq. and chrysophylla (Douglas ex Hook.) Hjelmq., a tree
Lithocarpus Blume. Both are morphologically potentially reaching 50 m, also occurs mostly in
similar to other broadleafevergreen Fagaceae of redwood and mixed evergreen forest. The bush
eastern Asia, but their phylogenetic position chinquapin, Chrysolepis sempervirens (Kell.)
within Fagaceae is either unresolved or has been Hjelmq., is more common at higher elevations
put into question with recently available data. In as a low-growing shrub on rocky slopes under
this study, we address the relationships of the coniferous forests ofthe Sierra Nevada and in the
castaneoids of the CA-FP and discuss the isolated interior ranges ofsouthern California.
implications of our findings in the context of Taxonomic delimitation of the castaneoid
taxonomy, cupule evolution, and biogeography. genera is based largely on characters of the
The two castaneoid taxa in question are well pistillate flower and cupule. The most important
represented in several CA-FP communities (Bar- morphological characteristic for defining Chryso-
bour and Minnich 2000), and both express lepis is a cupule with internal valves separating
sufficient differentiation in habit and leafmorphol- the fruits (Hjelmquist 1948). This condition is
ogyfortherecognition oftaxaatthesubspecific-to autapomorphic within Fagaceae, and the valves
species-level. The tanbark oak, Lithocarpus densi- have been interpreted as vestiges ofthe branches
MADRONO
182 [Vol. 55
ofhigher-order inflorescences (Nixon and Crepet Whilethecombination ofreproductivefeatures
1989). Both chinquapin species originally were observedinL. densijlorusfitscomfortablyintothe
treated in the chestnut genus Castcmea Miller, range ofvariation observed in Asian Lithocarpus,
and later transferred to the strictly Asian genus non-floral studieshave suggested otherwise. Most
Castanopsis(D. Don) Spach, consistent with their Asian Lithocarpus possess 2 to 4-rayed thick-
evergreen habit, leafvenation, and spiny cupule. walled trichomes, while L. densijlorus has distinct
Hjelmquist's (1948) segregation of the chinqua- multiradiate thin-walled trichomes (Jones 1986;
pins into Cluysolepis was supported by Jones' Cannon and Manos 2000). In addition, previous
(1986) foliarinvestigation ofthe family. He noted molecular phylogenetic studies showed that L.
that compared to Castanopsis, Chrysolepis has densijlorus does not form a clade with species of
smaller leaves and unique thick-walled peltate Asian Lithocarpus, one of the most strongly
trichomes, likely adaptations to xeric conditions. supported groups in all of Fagaceae (Manos et
Recent phylogenetic studies also have supported al. 2001); the placement oftanbark oak remained
this distinction, resolving Castanea and Casta- largely unresolved in that analysis.
nopsis as sister genera to the exclusion of In this study, we clarify these lingering !'
Chrvsolepis (e.g., Manos et al. 2001; Oh and questions about the phylogenetic relationships
Manos 2008). among the castaneoid taxa of the Fagaceae,
The taxonomic history of Lithocarpus densi- particularly in relation to the endemic CA-FP
florus is relatively straightforward. The species taxa. Our analysis involves three main goals: 1) |
was recognized first within Qiiercus, with the phylogenetic reconstruction of a broad sampling
following commentary: ''This remarkable plant of Fagaceae, representing all major lineages,
DNA
has very much the appearance ofa Castanea, the using multiple sequence data sets; 2)
fruit in the only specimen we possess being analysis of pollen morphology of L. densijlorus
situated at the base ofa male somewhat fascicled to further elucidate its relationship to Asian
catkin of the former year, while the numerous Lithocarpus; and 3) taxonomic revision of L.
male catkins of the present year present no densijlorus based upon these results.
appearancewhateveroffemale flowers." (Hooker
and Arnott 1840, p. 391). The acorn-like fruit of Materials and Methods
L. densiflorus develops biennially from a single
pistillate flower surrounded by a valveless cupule. Molecular phylogenetics
Miquel (1857) first introduced Qiiercus subg.
Pasania in a treatment of south Asian species of Seventeen taxa representing all ofthe currently
Fagaceae and Orsted (1866) later raised the recognized genera and major lineages of Faga-
subgenus Pasania to the generic rank, wherein ceae were chosen for study based on previous
Pasania c/ensiflora (Hook. & Arn.) Oerst. was phylogenetic analyses (Manos and Stanford 2001;
treated together with a mixture of largely Asian Manos et al. 2001; Oh and Manos 2008).
species under subg. Eupasania. Although no type Previous analyses addressed intra- and interspe-
has been associated with the name Pasania, the cific samplingofthecastaneoids oftheCA-FP, as
name is often misapplied to Asian species of well as Lithocarpus and Castanopsis. The current
Lithocarpus (Prantl 1888). Rehder (1917) treated sample is designed to focus on the broader
Pasaniadensiflorawithin Lithocarpus, and Camus relationships within Fagaceae using more data.
(1936-1954) placed L. densiflorus within Litho- The names, authorities, sources, geographic
carpus subg. Pasania, in the monotypic sect. origin, and GenBank accession numbers are
Androgyne. summarized in Appendix I. The genus Fagus is
Historically, there has been little doubt about consistently recognized as sister to all remaining
the relationship of L. densijlorus to the approx- Fagaceae; it is used here as the outgroup (Manos
imately 250 species of Asian Lithocarpus (stone and Steele 1997; Li et al. 2004).
oaks). The genus concept is supported by the We included two nuclear loci (ITS region and
flower-cupule described by Forman (1966a, b), CRABS CLAW or CRC) and three chloroplast
wherein each pistillate flower of a dichasium is regions {trnK-matKltrnK, atpB-rbcL and ndhF).
surrounded by a distinct valveless cupule. As a The CRC data were taken from Oh and Manos
whole, the genus Lithocarpus shows abundant (2008) and the nuclear ITS and chloroplast trnKJ
variation in inflorescence structure, sexual condi- matK and atpB-rbcL data were obtained from
tion of the flower spikes, numbers of staminate Manos and Stanford (2001) and Manos et al.
and pistillateflowersineachdichasium, fruittype, (2001). Some sequences of the trnKlmatK region
and cupule type and ornamentation (Kaul 1987; and most of atpB-rbcL sequences were newly
Cannon and Manos 2001). In L. densijlorus, the generated here, using the methods described in
inflorescences may be androgynous with pistillate Manos and Steele (1997) and Manos and
flowers at the base, orentirely staminate. There is Stanford (2001). The ndhF region was newly
a single pistillate flower per dichasium, and the sequenced in this study, using the PCR primers
cupules bear strongly reflexed scales. ndhF972 (5'- ATG TCT CAA TTG GGT TAT
2008] MANOS ET AL.: PHYLOGENY AND TAXONOMY OF PALEOENDEMIC FAGACEAE 183
Table 1. Specimens Usedinthe Pollen Study. ssiwmapplpeingseiqnuePnAcUeP*a.ddTithieonbesatnfdittTinBgRevoblurtainocnh-
ML
NotJwlithocarpiis deusifloi-us, California, Butte Co., ary model in the analysis was determined by
border ofponderosa pine, Douglas fir forest, 5 mi S the hierarchical likelihood ratio test using Mod-
ofSterling City, 12 Jul 1968, Terrells.n. (US); eltest 3.06 (Posada and Crandall 1998).
Humboldt Co., Seely Ranch, ca. 2 mi E ofWillow ML bootstrap analyses were conducted with
(sClrDoeUpeeKkEo)ofn;ritMdhgeeenTSdroEicnoiiftnyRoeRCidov.eM,rt,.a,l3o8JnuJglulR1e917d917,0M,tS.Ctloranoueasd3e,n03N740- G50A0RLpIseudvoerrseipolnica0t.e9s51 by(Zwuisciknlg 2t0h0e6).prAogrsaixm-
91 (NY); Oregon, Curry Co., Babyfoot Lake on the substitution parameter model was employed to
upper drainage ofthe Chetco River, 22 Sep 1964, calculate a likelihood value. Empirical base
Chambers2277 (US) frequencies were used, and other parameters,
including the shape parameter of gamma distri-
Chrysolepis chrysophylla. California, Shasta Co., E-face
ofCrater Peak, 8200 ft, volcanic, subalpine, 16 Aug bution and proportion of invariable sites, were
1964, /. Olmsted259 (DUKE). estimated from the data. Default values were
LitRheohcdaerrp;usCdheianlah,atSuisch(uHaono,k.Yfon&g-JTihao:mMsaonnoesx1M2i8q8.) useWdefortegsetneedticthaelgmoorintohpmhyanldy otohferListethtoicnagrs.pus,
(DUKE) including L. densijlorus, using the Shimodaira-
Hasegawa test (Shimodaira and Hasegawa 1999;
CasYtuannnopasni,s Minednigcalu(mR:oxMb.aneoxsLiannddleZyh)oAu.1D42C6:(CDhiUnKaE,) SH teMstP), as implemented in PAUP*. For the SH
test, trees generated enforcing the topological
MP
constraint were evaluated against the original
ATG ATG; Olmstead and Sweere 1994) and tree based on 10,000 bootstrap pseudoreplicates
FndhF2110R (5'- CCT CCT ATA TAT TTG using the re-estimated log likelihood (RELL)
ATA CCC TCT CC) with Phusion High-Fidelity method.
DNA polymerase (New England Biolabs, Ips-
wich, Massachusetts, USA) under the following Palynology
conditions in 25 \xL reactions: initial denatur-
ation at 98 C for 1 min 30 s, 40 cycles of98 C for The pollen of four specimens of Lithocarpus
10 s, 60 C for 30 s, and 72=C for 1 min, followed densiflorus collected in California and Oregon
by final extension at 72 C for 7 min. PGR was examined, in addition to representative
products were purified with the QlAquick PGR species of other castaneoid genera (Table 1).
Purification kit (Qiagen, Valencia, California, Pollen grains were acetolyzed using the method
USA). All sequences were deteimined at the in Faegri and Iversen (1950). Specimens were
DNA sequencing facility at the Duke Institute for mounted on stubs, coated with gold palladium
DGeNnAomAenaSlcyizeenrce(sApapnldiedPoBliiocsyysttheatmsu,sFeosstaer37C3i0txy1, woibtsheraveHduamnmdeprho6t.2ogsrpuatptheerdcuosatienrg(aAnPhaitleicphs)X,La3n0d
California, USA). Sequences were edited in ESEM TMP (FEI Company) in the Department
Sequencher version 4.5 (Gene Codes Corpora- of Biology SEM facility of Duke University. We
tion, Ann Arbor, Michigan). also measured polar and equatorial diameter and
Sequences were manually aligned using Mac- calculated the P/E ratio for 30 pollen grains ofL.
Clade version 4.06 (Maddison and Maddison densiflorus with light microscopy (LM) using the
2000). Phylogenetic analysis of the combined following collections: Terrell s.n. (US), Clausen
data was conducted with maximum parsimony 70-91 (NY), and Chambers2277(US). The pollen
(MP) and maximum likelihood (ML) methods. was mounted and stained with lactophenol-
Eighteen sites in intron four of CRC were cotton blue and viewed under a Leitz microscope
excluded in the analyses because of alignment at lOOOX magnification.
ambiguity. All other characters were treated as
MP
unordered and equally weighted in the Results
analyses. Gaps resulting from multiple alignment
of indels were treated as missing data. In both The final alignment ofcombined data included
MP ML
and analyses, heuristic searches in 7866 characters of which 1022 were parsimony-
PAUP* (Swofford 2002) were used to find the informative. The alignment was deposited in
best scored tree with 100 replicates of random TreeBASE (URL: http://www.treebase.org/).
taxon addition and tree bisection-reconnection The MP analysis produced a single tree with a
(TBR) branch swapping, saving all of the best length of 1946 steps, consistency index of 0.573
trees at each step (MulTrees). Branches with a (excluding uninformative characters), and reten-
minimum length of zero were collapsed using tion index of 0.643. ModelTest selected the
"amb-" option during the searches in the MP general time-reversal model (GTR; Swofford et
analysis (Nixon and Carpenter 1996). Bootstrap al. 1996) with six rate parameters, the gamma
analysis (Felsenstein 1985) using the MPcriterion distribution parameter (F), a = 0.793, and
with 500 pseudoreplicates was conducted with proportion of invariable sites (Pinvar) = 0.4386
MADRONO
184 [Vol. 55
Castanopsis argyrophylla
54
56
100 Castanopsis fissa SEAsia
100
100 Castanopsis carlesii
100
62 Castanea mollissima NH
D
86
Quercus myrsinifolia
99
Old World
97
57 Quercus suber
Notholithocarpus densiflorus
100
100
Quercus robur
97
99 94 Quercus rubra New World*
100 100
Quercus tomentella
100
Lithocarpuspachylepis
100 SEAsia
100
93 Lithocarpus xylocarpus
81
Chrysolepis chrysophylla
Trigonobalanus verticillata
90
88
73 Formanodendron doichangensis
83
Colombobalanus excelsa
Fagus grandifolia
Fig. 1. Maximum likelihood tree from the analysis of the combined ITS, CRC, and cpDNA data. Bootstrap
proportions using MP are indicated above branches and those based on ML are below. Boldfaced taxa are
paleoendemics of the California Floristic Province; shaded boxes indicate subfamily Castaneoideae; area
designations: SE Asia = Southeast Asia, NH = Northern Hemisphere; *New World Quercus clade contains
approximately 15 species ofEurasian white oaks {Quercus sect. Quercus s. s.).
ML
as the best fitting model for the data. The oaks, but this relationship was not supported in
MP
tree (Fig. 1) is identical to the tree, except for bootstrap analysis (results not shown). Overall
ML
the position ofLithocarpusdensiflorus. In the close relationship of L. densiflorus to Quercus
tree, L. densiflorus was sister to a weakly Castanea, and Castanopsis was strongly support-
MP ML i
supported clade that contains Old World oaks, ed (100% bootstrap value) in both and
MP i
Castanea, and Castanopsis. In tree, L. analyses (Fig. 1). Asian species of Lithocarpus
densiflorus was placed as sister to the New World formed a strongly supported clade with Chryso-l
2008] MANOS ET AL.: PHYLOGENY AND TAXONOMY OF PALEOENDEMIC FAGACEAE 185
lepis. Forcing the monophyly ofLithocarpus and (Crepet and Daghlian 1980; Wang and Chang
L. densijlorus required 48 additions steps, and the 1989; Fig. 2). In contrast, the pollen of Notho-
SH test indicated that the constrained topology is lithocarpus is distinctly smaller and rounder than
significantly worse than the original tree (P < the grains of Lithocarpus s.s. (mean P/E ratio of
0.0001). 1.68), and appears to be among the smallest and
The pollen of Lithocarpus densijlorus is tricol- most spheroidal grains (P/E ratio of 1.2) thus far
porate and subprolate (Fig. 2A) with an irregu- observed among extant castaneoid species
larly striate to smooth exine (Fig. 2B). On the (Fig. 2). In summary, we consider the distinct
basis of examining 30 pollen grains under LM, subprolate shape of the pollen grains of Notho-
the average measurement was 17.8 X 12.7 \xm lithocarpus combined with its unique multi-
(polar = 17.2 |Lim, minimum and 18.8 |Lim max- radiate leaf trichomes not found in Lithocarpus
imum; equatorial = 9.42 |am, minimum and s.s. (Jones 1986)anditsphylogeneticplacementto
14.1 |am maximum), and the average P/E ratio be strongevidence to recognize this paleoendemic
was 1.4. The pollen was smaller under the SEM ofthe CA-FP as a separate genus.
(12.5 X 10.4 |Lim), but the P/E ratio under both Our robust phylogenetic reconstruction of the
conditions was similar (1.4 in light vs. 1.2 in Fagaceae provides strong evidence of the para-
SEM). Apparently, the pollen grains increased phyly of subfamily Castaneoideae (Fig. 1) with
LM
slightly in size afterthe initial preparation for major implications for our understanding of
because of fluid uptake. The pollen grains of cupuleevolution (Fig. 3). Support for Chrysolepis
other representative castaneoids were prolate (Fig. as the sister genus to Lithocarpus s.s. combines
2C-H): Chrysolepis (P/E = 1.8), Asian Lithocar- two genera that each possess well-defined cupule/
pus (P/E = 1.6), and Castanopsis (P/E = 1.5). fruit apomorphies. One possible synapomorphy
for this relationship is the flower cupule {sensu
Discussion Forman 1966a), where each pistillate flower in a
dichasium is surrounded by cupular tissue. The
Phylogenetic analysis of over 7800 nucleotides placement of the new genus Notholithocarpus
resolves the placement of the castaneoids of the within the Quercus and Castanea + Castanopsis
CA-FP into two distinct lineages ofthe emerging clade suggests that its single-fruited cupule is
phylogeny of Fagaceae (Fig. 1). Chrysolepis is derived from reduced dichasial cupules via loss of
strongly supported as sister to Asian Lithocarpus; lateral flowers (Fig. 3). Developmental data from
while Lithocarpus densijlorus, hereafter referred Quercus indicates two initiation points or pri-
to as Notho/ithocarpus densijlorus (formal taxo- mordial cupule valves subtend the pistillate
nomic treatment is provided below) is placed flower (MacDonald 1979), which represents a
within the Quercus and Castanea + Castanopsis different ontogenetic process than the completely
clade, several nodes away from Lithocarpus s. s. valveless cupule of Lithocarpus s.s. (Okamoto
These results clearly indicate the polyphyly of 1989). Further studies are needed to confirm
Lithocarpus s. 1. A similar topology with high these evolutionary patterns of cupule develop-
bootstrap support resulted from analysis of ment inferred from our phylogenetic analysis of
nuclear sequences (CRABS CLAW; ITS) with a DNA sequence data.
larger sample oftaxa (Oh and Manos 2008) and a Delimitation of the castaneoid genera can be
previous analysis of only the ITS region across difficult, as few traits are truly fixed and exclusive
Fagaceae, including over 50 species of Lithocar- to a single genus. Forexample, while most species
pus s.s. also provided strong support for the of Castanea and Castanopsis possess three
monophyly Lithocarpus s.s., but no close rela- pistillate flowers per cupule, certain species in
tionship to Notholithocarpus (Manos et al. 2001). these genera have a single pistillate flower per
Pollen morphology and ultrastructure have cupule, a trait that normally defines Lithocarpus
provided important systematic data to classify s.s. (Fig. 3). A combination of traits is often
modern Fagaceae, aswellastoassesstheaffinities needed and Notholithocarpus is no exception. In
DNA
offossil taxa (Crepet and Daghlian 1980; Crepet addition to sequences and pollen morphol-
and Nixon 1989b; Nixon and Crepet 1989). The ogy, other potentially diagnostic features to
pollen of Quercus and the castaneoid genera are distinguish Notholithocarpus from Lithocarpus s.
diagnostic in shape and exine sculpture. The s. are foliar, namely the lack ofan acuminate tip,
pollen grains of Quercus are relatively large (27.5 sclerophyllous leaves bearing multiradiate tri-
X 25.4 |Lim), subspheroidal, with a granularexine; chomes (Jones 1986, see Type 10), unbranched
the grains of the castaneoid genera are smaller secondary veins that extend to the margin, and
(18.0 X 10 |am), prolate, with a smooth to striate the potential for serrate margins in Notholitho-
exine (Crepet and Daghlian 1980). Pollen varia- carpus, although some species ofLithocarpus s. s.
tion among and within the castaneoid genera is also bear slightly toothed margins. Our studies
mostly limited to size, but with a trend from more also included a comprehensive examination of
prolate grains in Chrysolepis, Castanea, and literature-based reports on inflorescence structure,
Lithocarpus s. s., to less prolate in Castanopsis morphology of staminate and pistillate flowers.
Fig. 2. Pollen of select castaneoid taxa. A & B. Notholithocarpus densijlorus {Stone 3034); C & D. Chysolepis
chrysophylla (I. Olmsted259)\ E & F. Lithoearpiis dealhatus {Manos 1288); G & H. Castcmopsis iiidica {Memosand
Zhou 1426); Scales: A, C, E, and G = 5 ^im; B, D, F, and H = 2 [im.
2008] MANOS ET AL.: PHYLOGENY AND TAXONOMY OF PALEOENDEMIC FAGACEAE 187
Notholithocarpus (#)
E
®
Quercus
'Castanea
Castanopsis
Lithocarpus
Chrysolepis
<Trigonobalanus
Formanodendron
Colombobalanus
Fagus
Fig. 3. Summary ofthe phylogeny ofFagaceae with a set ofdiagrams and images ofrepresentative cupule/fruit
types. Cupule valvesare indicated with straight lines and valvelesscupules with largecircles. Fruits aredrawn with
discs or triangles, and aborted flower position with small circles. For example, the unifloral cupule/fruit in
Notholithocarpus is shown with a disc in a large circle. Arrows indicate observed transformations within that clade
orwithintheinflorescence. Photographsaretakenfromthefollowingspecies: Notholithocarpusdensiflorus: Quercus
coccinea\ Castanea crenata; Castanopsis tribuloides (left) and Castanopsisfissa (right); Chysolepis chrysophylla:
Lithocarpusfenestratus(left)and Lithocarpusrotunclatus(right); Colombobalanusexcelsa(left)andFornumoclenclron
doichangensis (right); Fagus grcmdifolia.
wood anatomy, and cupule ornamentation among evergreen taxa known to have dispersed through
castaneoids (Carlquist 1984; Kaul and Abbe 1984; the North Hemisphere via Beringian and North
Kaul 1987; also see http://insidewood.lib.ncsu.edu/ Atlantic Land Bridges (Tiffney and Manchester
search/), but no clear differences were detected in 2001).
these features.
The phylogenetic position of CA-FP casta- Taxonomic Treatment
neoids as highly divergent members of two
different major Fagaceae lineages reinforces the Tanbark oak is a well-known element of the
concept that these paleoendemic taxa are rem- CA-FP and is a major focus ofresearch into the
nants of an ancient and formerly widespread spread of sudden oak death (SOD; Rizzo et al.
broadleafevergreen flora, which persists today in 2005). Ourresultsclarify its phylogenetic position
the Indochinese tropics. Additional evidence for in the Fagaceae and its affinities with temperate
this is based on the appreciable level of fossil oaks and tropical chestnuts. The fact that it is
diversity among likely thermophihc castaneoids more closely related to the victims of the
in North America (Crepet 1989; Crepet and devastating chestnut blight, Castanea spp., than
Daghlian 1980; Crepet and Nixon 1989a; Man- toitsmistakencongenerics, thestoneoaks, should
chester 1994; Mindell et al. 2007). The macrofos- provide insight into the evolutionary dynamics of
sil record also indicates the minimum divergence fungal infection and resistance. Here, we place
time between Chrysolepis and Lithocarpus s.s. to tanbark oak into a newly recognized genus,
be at least 40 mya based on fossils unambigu- Notholithocarpus, which reflects its false or
ously assigned to Lithocarpus s.s. (Kvacek and convergent similarity to Asian Lithocarpus s.s.
Walther 1989). This scenario suggests that the We make this new combination without subfa-
CA-FP castaneoids were established in the early milial designation, as further work is required to
Tertiary, which agrees with other broadleaf produce a valid nomenclature at that level.
)
MADRONO
188 [Vol. 55
NothoUthocarpus Manos, Cannon & S. Oh, gen. Cannon, C. H. and P. S. Manos. 2000. The Bornean
nov. TYPE: N. densiflorus (Hook. & Arn.) LitlwcarpusBl. section Synaedrys(Lindley) Barnett
Manos, Cannon & S. Oh. (Fagaceae): discussion of its circumscription and
A Lithocarpo foliis sclerophyllibus trichomatibus description of a new species. Botanical Journal of
the Linnaean Society 133:343-357.
multiradiatis et polline subprolato (sphaeroideo AND 2001. Combining and comparing
differt. morphometric s.hape descriptors with a molecular
Notholithocarpus densiflorus (Hook. & Arn.) phylogeny: the case of fruit type evolution in
Manos, Cannon & S. Oh, comb. nov. Basionym: Bornean Litliocarpus (Fagaceae). Systematic Biol-
Quercus densiflora Hook. & Arn., Bot. Voy. ogy 50:1-21.
DB.eeDchoeuyg,laps.s3.9n1.,(1h8o4l0o.tyTpeY:PEK!:, iUsSoAty,pe:CaGliHf!o)rnia, Carlwtqrouaoicdsh:eta,sriygSn.eilfe1im9ce8an4nt.cse.VaeAnslsdieslroelg1ar0to:iu5op0ni5sh-ni5gp25it.no diimcpoetryfloerdaotne
For intraspecific taxonomy, we follow other Crepet, W. L. 1989. History and implications of the
treatments (Tucker 1993; Nixon 1997). early North American fossil record of Fagaceae.
Notholithocarpus densiflorus var. densiflorus Pp. 45-66 in P. R. Crane and S. Blackmore (eds.).
Evolution, systematics and fossil history of the
Notholithocarpus densiflorus var. echinoides (R. Hamamelidae.Clarendon,Oxford,UnitedKingdom.
Br. ter) Manos, Cannon & S. Oh, comb. nov. AND C. P. Daghlian. 1980. Castaneoid
Basionym: Quercus echinoides R. Br. ter, Ann. inflorescences from the Middle Eocene ofTennes-
Mag. Nat. Hist., 4. 7:251. 1871. TYPE: USA, see and the diagnostic value of pollen (at the
Oregon, Canon Creek, Siskiyou Mountains, subfamily level) in Fagaceae. American Journal of
Botany 67:739-757.
BrTohwen g2e5n0us(HnoalomteypPeasKan,iai,soatylpoec:alK!n).ame for one evidAeNncDe Ko.fCF.agNaicxeoaen:. p1h9y8l9oag.enEeatrliicesatnmdegbaifoogsesiol-
of the species in Java (Rehder 1916), also was graphic implications. American Journal ofBotany
applied to tanbark oak byOrsted (1866). In order 76:842-855.
to preserve its usage within the synonomy of AND 1989b. Extinct transitional Faga-
.
Lithocarpus, we designate a type for the genus ceae from the Oligocene and their phylogenetic
name Pasania by using one of the species implications. American Journal of Botany
originally listed by both Miquel (1857: Quercus 76:1493-1505.
subgenus Pasania) and Orsted (1866: Pasania Faegri, K. and J. Iversen. 1950. Text-book of
subgenus Eupasania). modern pollen analysis. Ejnar Munksgaard, Co-
penhagen, Denmark.
Pasania (Miq.) Oerst., Vidensk. Meddel. Dansk Felsenstein, J. 1985. Confidence limits on phyloge-
Naturhist. Foren. Kjobenhavn, 8:81, 1866. Ba- nies: an approach using the bootstrap. Evolution
sionym: Quercussubgenus Pasania Miq., Fl. Ned. 39:783-791.
Ind. Vol. 1(1): 848. 1856. TYPE: Pasania sun- FoRMAN, L. L. 1966a. On the evolution ofcupules in
daica (Blume) Oerst., {Quercus sundaica Blume) the Fagaceae. Kew Bulletin 18:385-^19.
(lectotype designated here). . 1966b. Generic delimitation in the Castaneoi-
deae (Fagaceae). Kew Bulletin 18:421-426.
Acknowledgments Hjelmquist, H. 1948. Studies on the floral morphol-
ogy and phylogeny of the Amentiferae. Botanical
We are grateful to Don Stone and Leslie Eibest for Notiser 2(Suppl.):l-171.
advice on acetolysis and assistance with the SEM Hooker, W. J. and G. A. W. Arnott. 1840. The
images and to Elizabeth Wheeler for her insights on botany of Captain Beechey's voyage. Henry G.
woodanatomy. Bill Buck,JonShaw,AlanWhittemore, Bohn, London, United Kingdom.
Erin Tripp, and Robert Wilbur helped with the InsideWood. 2004-onwards. Published on the Inter-
taxonomy, and Bill Buck added the Latin diagnosis. net. Available at: http://insidewood/lib.ncsu.edu/
Thanks to Erin Tripp, John L. Strother, and Bruce search
Baldwin for improving the manuscript. This research Jones, J. H. 1986. Evolution of the Fagaceae: the
was supported by National Science Foundation grants implications of foliar features. Annals of the
DEB-9707945 to P. S. Manos and DEB-0445193 to P. Missouri Botanical Garden 73:228-275.
S. Manos and S. Oh, and the National Evolutionary Kaul, R. B. 1987. Reproductive structure of Litho-
Synthesis Center Northern Hemisphere Phytogeogra- carpus sensu lato (Fagaceae): cymules and fruits.
phy Working Group. Journal ofthe Arnold Arboretum 68:73-104.
and E. C. Abbe. 1984. Inflorescence architec-
Literature Cited
ture and evolution in the Fagaceae. Journal ofthe
Barbour, M. G. and R. A. Minnich. 2000. Califor- Arnold Arboretum 65:375^01.
nian uplandsandwoodlands. Pp. 161-202 in M. G. KvACEK, Z. AND H. Walther. 1989. Paleobotanical
Barbourand W. D. Billings(eds.).North American studiesin FagaceaeoftheEuropeanTertiary. Plant
terrestrial vegetation, 2nd ed. Cambridge Univer- Systematics and Evolution 162:213-229.
sity Press, New York, NY. Li, R.-Q., Z.-D. Chen, A.-M. Lu, D. E. Soltis, P. S.
Camus, A. 1936-1954. Les chenes monographic du SoLTis, AND P. S. Manos. 2004. Phylogenetic
DNA
genre Quercus (et LitJwcarpus). Encyclopedic econ- relationships in Fagales based on sequences
omique de sylviculture. Vol. 6-8. Academic des from threegenomes. InternationalJournalofPlant
Sciences, Paris, France. Sciences 165:311-324.
2008] MANOS ET AL.: PHYLOGENY AND TAXONOMY OF PALEOENDEMIC FAGACEAE 189
Maddison, D. R. and W. P. Maddison. 2000. Rehder, a. 1916. Additional species. Pp. 2479 in L.
MacClade 4: Analysis ofphylogeny and character H. Bailey (ed.). The standard cyclopedia of
evolution. Sinauer, Sunderland, MA. horticulture. Vol. V. The MacMillan Company,
Manos, p. S. and a. M. Stanford. 2001. The London, United Kingdom.
biogeography of Fagaceae: tracking the Tertiary . Additional species. Pp. 3569 in L. H. Bailey
history oftemperate and subtropical forests ofthe (ed.). The Standard cyclopedia of horticulture.
Northern Hemisphere. International Journal of Vol. VI. The MacMillan Company, London, Unit-
Plant Sciences 162:S77-S93. ed Kingdom.
and K. p. Steele. 1997. Phylogenetic analyses Rizzo, D. M., M. Garbelotto, and E. M. Hansen.
of 'higher' Hamamelididae based on plastid se- 2005. Phytophthora ramorum: integrative research
quence data. American Journal of Botany and management of an emerging pathogen in
84:1407-1419. California and Oregon forests. Annual Review of
Z. Zhou, and C. H. Cannon. 2001. System- Phytopathology 43:309-335.
,
atics of Fagaceae: phylogenetic tests of reproduc- Shimodaira, H. and M. Hasegawa. 1999. Multiple
tive trait evolution. International Journal ofPlant comparisonsoflog-likelihoods with applicationsto
Sciences 162:1361-1379. phylogenetic inference. Molecular Biology and
MacDonald, a. D. 1979. Inception of the cupule of Evolution 16:1114-1116.
Quercus macrocarpa and Fagus grandifolia. Cana- Stebbins, G. L. and J. Major. 1965. Endemism and
dian Journal ofBotany 57:1777-1782.^ speciation in the California Flora. Ecological
Manchester, S. R. 1994. Fruits and seeds of the Monograph 35:1-35.
middleEoceneNut Beds Flora, Clarno Formation, Swofford, D. L. 2002. PAUP* Phylogenetic Analysis
Oregon. Palaeontographica Americana 58:1-205. Using Parsimony (* and other methods), version
MiQUEL, F. A. W. 1857. Cupuliferae. Pp. 843-871 in 4.0. Sinauer Associates, Sunderland, MA.
Flora van nederlandsch Indie. Van de Post, , G. K. Olsen, p. J. Waddell, and D. M.
Amsterdam, the Netherlands. HILLIS. 1996. Phylogenetic inference. Pp. 407-514
MiNDELL, R. A., R. A. Stockey, and G. Beard. in D. M. Hillis, C. Moritz, and B. K. Mable (eds.).
2007. Cascadiacarpa spinosa gen. et sp. nov. Molecular systematics. SinauerAssociates, Sunder-
(Fagaceae): castaneoid fruits from the Eocene of land, MA.
Vancouver Island, Canada. American Journal of TiFFNEY, B. H. AND S. R. MANCHESTER. 2001. Theusc
Botany 94:351-361. of geological and paleontological evidence in
Nixon, K. C. 1997. Fagaceae. Pp. 436-506 in Flora of evaluating plant phylogenetic hypotheses in the
North America Editorial Committee (eds.). Flora Northern HemisphereTertiary. International Jour-
of North America north of Mexico. Vol. 3. nal ofPlant Sciences 162(Suppl.):S3-S17.
Magnoliophyta: Magnoliidae and Hamamelidae. Tucker, J. M. 1993. Fagaceae. Pp. 657-665 in J. C.
Oxford University Press, New York, NY. Hickman (ed.). The Jepson manual. University of
AND J. M. Carpenter. 1996. On consensus, California Press, Berkeley, CA.
collapsibility, andcladeconcordance. Cladistics 12: Wang, P-L. and K-L. Chang. 1989. A study on
305-321. pohen morphology and ultrastructure ofsubfamily
AND W. L. Crepet. 1989. Trigonobalanus Castaneoideae (Fagaceae) in China. Acta Phyto-
(Fagaceae): taxonomic status and phylogenetic taxonomica Sinica 3:205-214.
relationships. American Journal of Botany 6: ZwiCKL, D. J. 2006. Genetic algorithm approaches for
828-841. the phylogenetic analysis of large biological se-
Oh, S. and p. S. Manos. 2008. Molecular phyloge- quence datasets under the maximum likelihood
netics and cupuleevolution in Fagaceae as inferred criterion. Ph.D. dissertation. The University of
from nuclear CRABS CLAW sequences. Taxon Texas, Austin, TX. Available at: http://www.bio.
57:434-451. utexas.edu/faculty/antisense/garli/Garli.html [Ac-
Okamoto, M. 1989. A comparative study of the cessed 10 May 2007].
ontogenetic development of the cupules in Casta-
nea and Lithocarpus (Fagaceae). Plant Systematics
Appendix
OlmsatnedaEdv,olRu.tiGo.na1n6d8:J7.-1A8..Sweere. 1994. Combining nDumNbAervsoufocrhetraxianfuosremdatiniotnheanmdoI.lGeceunlBaarnpkhyalcocgeesnseitoinc
data in phylogenetic systematics: an empirical
study. Each entry includes species, locality, voucher
approach using three molecular data sets in the specimen, and GenBank accessionnumbers(ITS; CRC\
Solanaceae. Systematic Biology 43:467^81.
Orsted, a. S. 1866. Bidrag til Egelaegtens systematik. trnKlmatK\ atpB-rbcL\ ncI/iFaccession).
Videnskabelige Meddelelser fra Dansk Naturhis- Castanea mollissima Blume; USA, New York,
torisk Forening i Kjobenhavn 8:11-88. Tompkins Co., Cornell University Plantations; Manos
Posada, D. and K. A. Crandall. 1998. MODELT- 1038 (BH); ITS-AY040396; C7?C EU189752; tniKJ
EST: testing the model of DNA substitution. matK-¥iU5050\ nd/iF-FJ1^5073; Connecticut, Con-
Bioinformatics 14:817-818. necticut Agricultural Research Station; StanfordRlT15
Prantl, K. 1888. Fagaceae. Pp. 47-58 inA. Englerand (UNC-CH); atpB-rbcL-AY0A2A53.
K. Prantl(eds.). Dienatiiralichen Pflanzenfamilien. Castanopsis argyrophylla King ex Hook.f.; China,
Division III, Vol. 1, Wilhelm Engelmann, Leipzig, Yunnan, Menglian; Manos and Zhou 1402 (DUKE);
Germany. AY040376; EU189758; FJ185051; FJ185059; FJ185074.
Raven, P. H. and D. I. Axelrod. 1978. Origin and Castanopsis carlesii (Hemsl.) Hayata; China, Yun-
relationships of the Cahfornia flora. University of nan,Jingu; ManosandZhou 1382 (DUKE); AY040372;
CaHfornia Publications in Botany 72:1-134. EU189760; FJ185052; FJ185060; FJ185075.
MADRONO
190 [Vol. 55
Castcmopsisfissa(Champ, exBenth.) Rehder& E. H. echinoides (R. Br.) Abrams); USA; California, Ne-
Wilson; China, Yunnan, Simao; Memos andZhou 1396 vada Co., Washington; Manos and Tucker 922
(DUKE); AY040392; EU189763; FJ185053; FJ185061; (BH); AY040370, EU189783; FJ185047; FJ185067;
FJ185076. FJ185081.
Chrysolepis chrysophylla (Douglas ex Hook.) Quercus myrsinifoHa Blume; USA, Georgia, USDA
Hjelmq.; USA, Oregon, Benton Co., St. Mary's Peak; Coastal Research Station. Savannah; Memos s.n.
Memoss.n. (DUKE); AF389087; EU189770; FJ185045; (BH); ITS-AF098414; California, Yolo Co., Shields
FJ185062; FJ185077. Grove Arboretum; no voucher; C7?C-EU189824; trnKI
Colombohahmus excelsa (Lozano, Hdz-C. & Henao) matK- FJ185048; atpB-rbcL- FJ185068; ndhF^
Nixon & Crepet; Colombia, Virolin; Nixon 4655 (BH); FJ185084.
AF098412; EU189772; AY042456+AY040492; Quercus rohur L.; USA, New York, Tompkins Co.,
FJ185063; FJ185078. Cornell Univ. Campus; Manos s.n. (BH); AF098424;
Fagusgrcmdifolia Ehrh.; USA, NewYork, Tompkins EU189832; FJ185056; AY042505; FJ185085.
Co.; Manos 114 (BH); AY040509; EU189851; Quercus rubra L.; USA, New York, Tompkins Co.,
U92861+AY042400; AY042412; FJ185079. Cornell Univ. Campus; Manos s.n. (BH); AF098418;
Fornumodendron doicJumgensis (A. Camus) Nixon & EU189836; FJ185057; FJ185069; not determined.
Crepet; China, Yunnan, Menglian; Manos and Zhou Quercus suber L.; USA, California, Santa Barbara
1400 (DUKE); AY040452; EU189773; FJ185046; Co., Orella St.; Manos 423 (BH); AF098434;
FJ185064; FJ185080. EU189843; FJ185049; FJ185070; FJ185086.
Lithocarpus pachvlepis A. Camus; China, Yunnan, Quercus tonientella Engelm. USA, California.
Da Wei Shan; Manos and Zhou 1451 (DUKE); Santa Barbara Co. Santa Cruz Island; Memos 983
AY040441; EU189792; FJ185054; FJ185065; FJ185082. (BH); AF098437; EU189845; FJ185058; FJ185071;
Lithocarpusxylocarpus (Kurz) Markgr.; China, Yun- FJ185087.
nan, Da Wei Shan; Manos and Zhou 1463 (DUKE); Trigemobalanus verticil/ata Forman; U.K., Scotland,
AY040426; EU189801; FJ185055; FJ185066; FJ185083. Edinburgh, Royal Botanical Gardens; RBG 1967-421;
NotJioUtJwcarpus densifloras var. echinoides (R. Br. AF098413; EU189847; U92866+AY042455; FJ185072;
ter) Manos & S. Oh {Lithocarpus densiflorus var. FJ185088.