Chapter 44 
Compositae metatrees: the next generation 
Vicki A. Funk, Arne A. Anderberg, Bruce G. Baldwin, Randall J. Bayer, 
J. Mauricio Bonifacino, Use Breitwieser, Luc Brouillet, Rodrigo Carhajal, Raymund Chan, 
Antonio X.P. Coutinho, Daniel J. Crawford, Jorge V. Crisci, Michael O. Dillon, 
Susana E. Freire, Merce Galhany-Casals, Nuria Garcia-Jacas, Birgit Gemeinholzer, Michael 
Gruenstaeudl, Hans V. Hansen, Sven Himmelreich, Joachim W. Kadereit, Mari Kallersjo, 
Vesna Karaman-Castro, Per Ola Karis, Liliana Katinas, Sterling C. Keeley, Norhert Kilian, 
Rebecca T. Kimball, Timothy K. Lowrey, Johannes Lundberg, Robert J. McKenzie, 
Mesjin Tadesse, Mark E. Mort, Bertil Nordenstam, Christoph Oberprieler, Santiago Ortiz, 
Pieter B. Pelser, Christopher P. Randle, Harold Robinson, Nddia Roque, Gisela Sancho, 
John C. Semple, Miguel Serrano, Tod F. Stuessy, Alfonso Susanna, Matthew Unwin, 
Lowell Urbatsch, Estrella Urtubey, Joan Valles, Robert Vogt, Steve Wagstaff, Josephine Ward 
and Linda E. Watson 
INTRODUCTION 
Constructing a large combined tree of Compositae, a 
'metatree' (also called 'meta-supertree' by Funk and 
Specht 2007 and 'megatree' by R. Ree, pers. comm.) 
allows one to examine the overall phylogenetic and bio- 
geographic patterns of the family. The first modern at- 
tempts to understand the family were by the authors in 
Heywood et al. (1977) plus the paper by Cronquist (1977), 
which was initially intended to be in the Heywood 
publication. Literature prior to 1977 has been discussed 
in detail in other chapters (for the early literature, see 
Chapter 1). In Cronquist's 1977 paper he reaffirmed his 
agreement with Bentham's 13-tribe classification of the 
family and the concept that Heliantheae s.l. were the 
primitive members (Cronquist 1955; Bentham 1973a, b). 
Cronquist (1977) pointed out that the Heywood et al. 
volumes listed the tribes mostly in the order of Bentham 
1873a rather than beginning with Heliantheae, which 
Bentham thought was most primitive (Bentham 1873b). 
The papers in the 1977 volumes did accept some changes 
such as the recognition of Liabeae and the conclusion 
that Helenieae were not a 'good' group, both more or 
less accepted by Cronquist in 1977. However, most pro- 
posed changes such as the new tribe Coreopsideae, etc. 
were not accepted by the synantherological community. 
Cronquist (1977) believed that the primitive characters 
of the family were as follows (slightly modified): shrubby; 
leaves opposite; inflorescence cymose; heads few, each 
with many florets; involucre leafy, several-seriate; recep- 
tacle chaffy; ray florets present and fertile; disk florets 
perfect and fertile; lobes of the disk corollas with well 
developed mid-vein; pappus chaffy, of five members; and 
anthers  connate,  not  tailed.   Cronquist  stated  that the 
748 Funk et al. 
presence of ray florets may have predated the origin of 
Compositae, so that even discoid tribes might have had 
a radiate ancestry. 
The acceptance of the modified Bentham system was 
not universal. There were at least two papers in the 
Heywood et al. volumes (Jeffrey 1977; Skvarla 1977) 
and two individuals who published elsewhere (Carlquist 
1966, 1976; Robinson 1981) who had reservations about 
the concept of "13 tribes rooted in the Heliantheae". All 
of these dissenting authors observed that the data they 
were generating did not support all of the above-listed 
characteristics as primitive in the family. However, for 
the most part, the synantherological community contin- 
ued to use the Bentham classification. 
Not too long after 1977, opinions began to change with 
the advent of cladistic methodology and molecular data. 
Jansen and Bremer and their collaborators (Bremer 1987, 
1992, 1994; Jansen and Palmer 1987, 1988; Hansen 1991a, 
b; Jansen et al. 1991a, b; Bremer and Jansen 1992; Jansen 
and Kim 1996; Bremer and Gustafsson 1997) reordered 
Compositae by placing Barnadesiinae as the sister group of 
the family and placing Heliantheae (including Eupatorieae) 
highly nested in the phylogeny of the family. 
Bremer's cladistic analysis (1994) was the first revi- 
sion of the whole family based on morphology since 
Bentham, and he recognized many of the problem areas 
in the cladograms of the family and tribes, but the mor- 
phology did not generate enough data to resolve many of 
the issues. Over ten years later Kadereit and Jeffrey (2007) 
reordered the genera, tribes, and subfamilies within the 
family based on morphology and molecular results, and 
this work is now the standard reference for descriptions 
of the tribes and genera of the family. 
This chapter seeks to link the most recent molecular 
trees together in a metatree framework (Funk and Specht 
2007) and to use that tree to provide a basis for under- 
standing the systematics, evolution, and biogeography of 
the family. 
MATERIALS AND METHODS 
Construction of the metatree 
The metatree for Compositae was developed using a com- 
pilation of trees. The name metatree was adopted for this 
type of tree because it is a "tree of trees", one that is based 
on a fixed 'base tree' topology (Funk and Specht 2007). 
This type of tree has also been called a meta-supertree or 
megatree (R. Ree, pers. comm.), and some authors refer 
to it as a supertree. It is, however, neither a tree produced 
by a combined analysis of coded cladograms obtained 
from individual datasets (classic 'supertrees') nor is it the 
result of analyzing a dataset in which data from multiple 
datasets have been combined ('supermatrix' trees). There 
has been some discussion on the pros and cons of the 
'supertree' and 'supermatrix' methods (Steel et al. 2000; 
Gatesy et al. 2002; Bininda-Emonds et al. 2003), and 
both methods are compared with the metatree approach 
by Funk and Specht (2007). The metatree for this analysis 
was constructed in the following manner: 
1. A 'base tree' was formed from the phylogeny of 
Panero and Funk (2008) with a few alterations. The 
most important change was the addition of taxa 
from the Heliantheae Alliance. The Heliantheae 
Alliance section of the Panero and Funk tree 
(which had only a few taxa) was replaced with the 
branching pattern of the Heliantheae Alliance from 
Baldwin (Baldwin et al. 2002; Chapter 41). Also, 
some refinements were made using the work of 
Ortiz (Chapters 18 and 19) and Ortiz et al. (Chapter 
17) for Carduoideae, and Funk and Chan (Chapter 
23) for Cichorioideae. The base tree was reduced 
to a matrix using Brooks Parsimony Analysis 
(BPA; Brooks 1982; Brooks and McLennan 2002), 
wherein any branching diagram can be reduced to 
a series of zeros and ones in a data matrix. We used 
MacClade to generate the data matrix (Maddison 
and Maddison 2001). The data matrix was run in 
a tree program (PAUP 4.0M0; Swofford 2002) to 
check for errors. All trees have been "ladderized to 
the right" for consistency, although anyone familiar 
with cladistics will understand that the tree can be 
"rotated" at any node. This feature is amply dem- 
onstrated by comparing the rooted tree (Fig. 44.1) 
and the unrooted tree (Fig. 44.2). 
2. The most recent (and available) tree for each clade 
(see below) was reduced to a matrix (as above) and 
these matrices were added to the original matrix. 
Each time a new clade tree was added, the overall 
analysis was re-run to insure an accurate replication 
of the newly added tree, as well as to confirm that 
the addition did not result in topological changes 
elsewhere in the metatree. It should be noted that 
when a phylogeny for a tribe contained many taxa 
from the same area in a monophyletic group or a 
grade, these were often pruned to decrease the size 
of the tree without subtracting any biogeographi- 
cal information. For instance, the phylogeny of 
Gnaphalieae contained a clade of 58 terminal taxa 
all endemic to Australia; this clade was reduced to 
25 taxa. 
3. A summary tree (Fig. 44.1) was produced in which 
each major clade was reduced to a single branch. 
This tree also shows the phylogenetic position of 
critically placed taxa and is displayed as an unrooted 
tree in Fig. 44.2. 
See the section on optimization for an explanation of 
the biogeographic areas and how they were assigned. 
Chapter 44: Compositae metatrees: the next generation 749 
Sources of the trees 
General references for this study were Bremer (1994), 
Heywood (1993), Heywood et al., (1977), Hind (1996), 
and Kadereit and Jeffrey (2007). Below, the origin of each 
phylogeny on the metatree is discussed. 
Outgroups 
Lundberg (Chapter 10) examined the relationships among 
the families now contained in Asterales, including Com- 
positae. His work indicated that Calyceraceae were the 
sister group of Compositae (1st outgroup) and that Good- 
eniaceae (2 outgroup) were the sister group of the 
Calyceraceae + Compositae clade. The next most closely 
related family is Menyanthaceae, and it is followed by a 
clade containing Stylidiaceae, Alseuosmiaceae, Phellinac- 
eae, and Argophyllaceae. The distribution ofthese eight fam- 
ilies (Fig. 44.1) shows that the Compositae + Calyceraceae 
clade is nested in a grade of Australasian taxa (Australia, 
New Guinea, New Caledonia, and New Zealand). Each 
ofthese families is discussed below (listed in reverse order 
of relatedness to Compositae). 
Argophyllaceae. ? Two genera with ca. twenty spe- 
cies that are distributed on Australia, Lord Howe Island, 
New Caledonia, New Zealand, and Rapa Island. 
Phellinaceae. ? One genus with eleven species, all of 
which are found on New Caledonia. 
Alseuosmiaceae. ? Five genera and ten species all 
located on Australia, New Caledonia, New Guinea, and 
New Zealand. 
Stylidiaceae. ? Six genera with 245 species found in 
Australia and New Zealand with a few species in East 
Asia and South America. 
Menyanthaceae. ? Five genera with sixty species 
having an almost cosmopolitan distribution; however, 
four of the five genera are found in Australia, and because 
the closely related taxa are found in the Australia?New 
Zealand?New Guinea?New Caledonia area, this family is 
treated as having an Australasian distribution at its base. 
Goodeniaceae. ? The second outgroup of Compos- 
itae is a moderate-sized family of herbs and some shrubs: 
Goodeniaceae (fourteen genera, over 400 species). The 
family is largely confined to Australia, particularly west- 
ern Australia, with only a few species extending else- 
where, mostly in the Pacific area (Gustafsson et al. 1996, 
1997). A recent study (Howarth et al. 2003) has shown 
that the base of the phylogeny of Goodeniaceae is in 
Australia with dispersals by members of Scaevola into the 
Pacific area, coastal areas in southern Asia and Africa, and 
the east coast of the Americas. 
Calyceraceae. ? The first outgroup of Compositae, 
and therefore its sister group, is Calyceraceae, a small 
family (six genera, ca. sixty species) of annual and pe- 
rennial herbs. The family is entirely South American, 
being most abundant in the Andes south from Bolivia, 
extending eastwards through Paraguay to Uruguay and 
southern Brazil and down through Argentina to southern 
Patagonia (Heywood 1993). 
Cassini, in his famous 1816 diagram (Chapter 41: Fig. 
41.1), showed Calyceraceae and Campanulaceae to be 
closely related to Compositae. Even though he did not 
have it in the diagram, he also thought Goodeniaceae 
were close (see Chapter 1). 
Compositae 
The base tree. ? The basic structure of the tree was 
taken from Panero and Funk (2002, 2008) and Baldwin 
(Baldwin et al. 2002; Chapter 41); see above for details. 
The trees in Panero and Funk (2008) contained exten- 
sive sampling from the base of the tree, Mutisieae (sensu 
Cabrera), three to ten genera representing all other tribes 
(including the Heliantheae Alliance), and many taxa that 
had been "hard to place" in previous studies (includ- 
ing Hecastodeis, Gymnarrhena, and Corymbium). The Panero 
and Funk phylogeny was based on data from ten chlo- 
roplast gene regions (ndhF, tmL-tmF, matK, ndhD, rbcL, 
rpoB, rpoCl, exonl, 23S-t.rnI, and ndhl). Relationships 
within tribes of the Heliantheae Alliance were taken from 
Baldwin et al. (2002) and Chapter 41 and 'were based 
on data from the ITS region of rDNA. Modifications 
were made in Cichorioideae (based on Chapter 23) and 
in Carduoideae (based on Ortiz, Chapters 18 and 19; and 
Ortiz et al. (Chapter 17). 
Mutisieae s.l. sensu Cabrera (Chapter 12). ? The 
tribe Mutisieae (sensu Cabrera) has 84 genera and ca. 
900 species. The paraphyly of Mutisieae (sensu Cabrera) 
was suggested by morphological studies (Cabrera 1977; 
Hansen 1991b) as well as the first molecular studies of 
the family. The subtribe Barnadesiinae was recognized 
as being the sister group to the rest of the family (Jansen 
and Palmer 1987, 1988; Bremer 1994; Kim and Jansen 
1995). Kim et al. (2002) showed that the remainder of 
the tribe (sensu Cabrera) could not be supported as a 
monophyletic group. Most recently, Panero and Funk 
(2002, 2008) published phylogenies based on molecular 
data from ten chloroplast regions that (1) confirmed that 
Mutisieae (sensu Cabrera) were paraphyletic, (2) identi- 
fied additional clades, and (3) elevated several groups to 
tribal and subfamily levels. Except for Barnadesieae, the 
phylogeny of Panero and Funk (2008) formed the base 
tree for Mutisieae (sensu Cabrera) with a few additions 
from Kim et al. (2002) and Katinas et al. (2007). 
Barnadesieae (Chapter 13). ? The subfamily Barn- 
adesioideae (nine genera; 91 species) has one tribe, and 
it is the sister group for the rest of Compositae. This 
has been known since the seminal papers by Jansen and 
Palmer (1987, 1988) established the presence of a chlo- 
roplast DNA inversion shared by the rest of the family, 
but not by Barnadesieae or other flowering plants. The 
750 Funk et al. 
MUT. WUNDER.    CARDU. CICHORIOIDEAE 
Outgroups Stifftieae Wun. 
?    o 
'--Si 9   <>> 
o o 
CD TT 
03 ?' 
CO CD 
CD CO 
O CD 
CO O 
?5 .g 
c c 
CO CD 
>--o 
c O 
.?? UJ   <~   d' CD O 
tt> < a. < 2 CD 
o 
3 
^ CD 
>>  CO 
i- CD 
??' CO 
CD CD 
CO O 
CD CO 
U  = It 
co O) 
^co 
CN  %T 
lO CO 
?S 
f~~ CD 
CD   CO 
Hyg.? 
Cichorieae 
CO       ?2,      ~_ CO ? 
CM 
</) 
CD 
"O 
CO 
>> = 
CO CO 
O CO 
CD CD 
CO CO 
CD CD 
5 > 
i- 3 
CD CO 
W CO 
O CO 
= CO O Z 
$? 
J5 co 
?'5.-S.<o 
CD  o  0  ~ 3a8^ I 
???15=53 
CD ^ 
- ,T3  O 
LO.J0  N 
S3 ? 
Z: co "K 
CD  o. 2 
T3   O   F C   C    o 5 .ffi o 
scoo 
o 
o 
c > 
o 
LO LD 
N CM 
CD CD 
CO m 
CD CD (- 3 
o u 
CJ ro 
G O 
CO 
CD ~? 
CO CD 
CD CO 
g CD 
D) 
CD S2 
CO rn 
t? o 
CD   t   CO 
S   CO   c 
03   ><^ 
+ + 
o o 
o o 
TT if3 
o o 
03 03 
CO CO 
O3 ,03 
6 6 
u u b b 
CO + 
? CD 
03  N 
CO ?' 
03 03 
C    CO 
E .5 
CO  "O 
11 
+ 
CO c? 
?22 c 
I" I p t 
03 V I  < 
03 ^ 
CO TT 
03 N 
#^ J0 CO 
D__l 
+ 
^     o 
o      o 
&w!2 
CO   CD II CO   o 
c3 s 
I, J W L_J 
i     -r 
?"r-J 
L.T_J 
?.____. 
b3 
Fig. 44.1. A summary tree based on the metatree (Figs. 44.3?44.7). The tribes or clades have been represented by one to four- 
branches. The branches and internodes were colored according to the distribution of the taxon or the optimization of those 
distributions. The numbers by the terminal taxa reflect the number of species in that clade. Note that some areas have been 
combined (e.g., Mexico and North America) and that the red color in Vernonieae represents Tropical America. Subfamilies 
that have more than one tribe are indicated on the summary tree in capital letters (see Chapter 11 for details). A = Arctotideae; 
CARDU. = Carduoideae; Hya. = Hyalideae; MUT. = Mutisioideae; S = Senecioneae; Wun. = Wunderlichieae; WUNDER. 
= Wunderlichioideae. 
Chapter 44: Compositae metatrees: the next generation 751 
ASTEROIDEAE 
Senecioneae (3500) 
CD 
CD 
O  ? 
r-   CO 
.a   3 
-Q .o 
E 5 &e 
o o 
O Q 
O o 
03 ^ 
co C 
.E a) 
5 cc 
O CD 
u "5 0 -c 
CD 
CO 
.E   03 
Eg 
CI3-- 
M = 
CO   O 
CO  J= 
o 
CM ^ 
i- o 
? CO 
CD O 
CC CO 
03 
03  . 
g=0 
-= =  CD !2 ^ 
CD 
CD 
CO (/) (0 U 
Q. c 
CO   03 
=  "CO o < 
t   03 
CD   CO 
O CO 
^^         m m 
in      ,-^m T- 
m      o ? ? 
? ^CM 03 03 
CD i- i- CO co 
CO ??'??'  CD 03 
CD 03 CD "O CZ 
P CO CO w 03 
W 03 .03 Q. JO 
= CD = O O 
1?  Si   C 3 
5   03   03   O 03 
<  LL  %  U Z 
a> 
CM CO 
(26
7) 
id
ea
e 
83
) 
^^^r 
co J3. 
CD   03 
a) o w 03   03 
ge
te
a 
a
e
n
a
 
hi
ea
e 
1.1 
|2 6 CO O   03 0_ X 
o 
o 
^^ CM 
O 'T-CN -~- 
CO a ?CO 
CO JO 03 O 
? CO c\i 
03 03 03 """ C0%   c   03 
03 .92 o co 
?=>>tt.2 
???? S.-o 
=   CD   3  JO 
S 0_ UJ S 
F 
South America 
Brazil 
| Guiana Shield 
North & central Andes 
Southern Andes, southern South America 
| General South America 
North America 
North America, Mexico 
Central America, Caribbean 
Eurasia 
Eurasia, Europe 
Eastern & central Asia 
Southern Africa 
Madagascar, tropical Africa 
Northern Africa, Mediterranean, southern Europe 
General Africa 
Australia and the Pacific 
| Australia, N. Guinea, N. Caledonia, N. Zealand 
Widespread or ambiguous 
Africa 
first phylogeny of this tribe was done by Gustafsson et al. 
(2001), but it was not completely resolved. The phylogeny 
for the tribe was taken from Gruenstaeudl et al. (2009). It 
was based on DNA sequence data of nine chloroplast gene 
regions (atpI-atpH IGS, matK, psbA-tmH IGS, rbcL, partial 
rpoCl gene + intron, rpsl6-trnK IGS, partial trnK intron, 
trnL intron, trnL-trnF IGS), the nuclear ribosomal ITS 
region (ITS1, 5.8S, ITS2), recoded DNA insertions/dele- 
tions, and selected morphological characters from previous 
investigations. In their analysis all genera were monophyl- 
etic except for Dasyphyllum, which fell into two groups 
reflecting the subgenera and their respective distributions 
"east of the Andes" and "west of the Andes". There are 
two possible positions for Schlechtendalia, one of which is 
basal for the tribe, and the other is more highly nested. 
The ambiguity of the position of Schlechtendalia does not 
affect the biogeographic hypothesis for this tribe. 
African Mutisieae (Chapters 17-19). ? With the ex- 
ception of Gerbera and the closely related and sometimes 
congeneric Perdicium, which are found in Africa and 
to a lesser extent in Asia, all Mutisieae (sensu Cabrera) 
from Africa are no longer part of Mutisioideae (sensu 
Panero and Funk) and are now in Carduoideae. Using 
ITS and ndhF sequence data, Ortiz and his collaborators 
(Chapters 17?19) have shown that these segregate African 
Mutisieae form three (or four) distinct groups that are 
separated by striking morphological as well as molecular 
differences. Currently, there are three tribes: Dicomeae, 
Oldenburgieae, and Tarchonantheae. However, it is pos- 
sible, but not yet certain, that the tribe Dicomeae may fall 
into two distinct groups that are not sister taxa. In addi- 
tion, there is still some ambiguity as to the relationships 
among some of the tribes. 
The tribe Dicomeae contains seven African genera (ca. 
75?100 species) occurring in tropical and southern Africa 
and Madagascar with a minor presence in the Arabian 
Peninsula, India, and Pakistan. The tribe Tarchonantheae 
contains   two   African   genera   (13   species)   occurring 
752 Funk et al. 
1. Cichorieae Eupatorieae 
2. Eremothamneae 
3. Moquineae 
4. Calenduleae 
Madieae 
Heliantheae 
Bahieae 
Chaenactideae 
Neurolaeneae 
Tageteae 
Feddeeae 
Inuleae 
Senecioneae 
Abrotanella 
Doronicum 
A-Gorteriinae 
Heterolepis 
A-Arctotidineae 
3 
Li. 
3 Platycarpheae Liabeae 
Distephanus 
Vernonieae   ^ 
Hecastocleideae 
Stenopadus clade 
Wunderlichia ? 
Leucomeris clade <? 
Hyalis clade 
Gongylolepis clade o 
Dinoseris clade I 
Jl 
? 
= 3 
Perityleae 
Millerieae 
Polymnieae 
4 Coreopsideae 
Helenieae 
Athroismeae 
4    ^A Asterc t reae 
Anthemideae 
Gnaphalieae 
> Corymbieae 
Gymnarrheneae 
Pertyeae 
f Dicomeae 
o Oldenburgieae 
Tarchonantheae 
Cardueae 
Gochnatieae 
i Onoserideae 
Stifftia 
Barnadesieae 
I -0 Mutisieae Nassauvieae 
Calyceraceae 
Fig. 44.2. An unrooted representation of the summary tree. 
The size of the circle indicates the number of species found 
in that clade. Colors are the same as in Fie;. 44.1. 
mainly in tropical and southern Africa, and Madagascar, 
but it is also present on the Arabian Peninsula. The tribe 
Oldenburgieae has only the genus Oldenburgia (4 species), 
which is endemic to the Cape Floristic Region of South 
Africa. 
Cardueae (Chapter 20). ? Cardueae (thistles; 73 gen- 
era, ca. 2500 species) are now known to be nested within 
a paraphyletic Mutisieae (sensu Cabrera). This tribe is 
the sister group of the African Mutisieae clades. The 
tribes Cardueae, Tarchonantheae, Oldenburgieae, and 
Dicomeae form a monophyletic group that is now the 
subfamily Carduoideae. The Cardueae tree used for the 
metatree is based on matK, trnL-F, and ITS sequence data 
(Susanna et al. 2006). 
Cichorieae (Lactuceae; Chapter 24). ? The phy- 
logeny of the mainly north temperate dandelion tribe 
Cichorieae (Lactuceae) has long been problematic. It has 
93 genera arranged in eleven subtribes, but the number 
of species varies depending on one's species concept. If 
one excludes the problematic genera Hieracium, Pilosella, 
and Taraxacum, there are about 1400 species (Kilian et al., 
Chapter 24). The Cichorieae tree used in this study was 
provided by Gemeinholzer and her collaborators based 
on recent molecular analyses of a large ITS dataset (428 
taxa of 83 genera; Gemeinholzer and Bachmann 2003; 
Kilian et al., Chapter 24; Gemeinholzer et al., unpub.). 
The analyses revealed the existence of five major clades, 
with a total of eleven subclades, within the tribe. 
The position of Gundelia (Gundelieae) as basal within 
Cichorieae was suggested by Karis et al. (2001) based 
on ndhF data, and this was supported by Panero and 
Funk (2008), who also found Warionia to be at the base. 
However, the current studies of Gemeinholzer and her 
collaborators comprising more basally branching taxa 
place the Northern African genus Warionia at the base 
of Cichorieae with the Mediterranean Gundelia slightly 
more highly nested. Since both are from the same bio- 
geographic area, the two different placements of Gundelia 
and Warionia do not affect the biogeographic analysis. 
Arctotideae, Eremothamneae, Platycarpheae, and 
Heterolepis (Chapters 25, 26, 29, 31). ? The tribe Arcto- 
tideae (African Daisies) is a diverse and interesting group 
(18 genera, 215 species). Recent molecular studies are am- 
biguous as to the monophyly of this tribe, and some former 
members have been moved out of the tribe based on mor- 
phology and/or molecular data. The positions of Heterolepis 
(Funk and Karis, Chapter 31) and the tribe Eremothamneae 
(2 genera, 3 species; Robinson and Funk, Chapter 26) vary 
depending on the data used in the analysis, and the new 
tribe Platycarpheae (2 genera and 3 species) is most likely 
closely related to the Liabeae + Vernonieae clade (Funk et 
al., Chapter 29). Although Arctotideae cannot be unam- 
biguously diagnosed, the two core subtribes are distinctive 
based on morphology as well as molecular data (Funk et 
Chapter 44: Compositae metatrees: the next generation 753 
al. 2004; Karis et al., Chapter 25). Recently published 
phylogenies using both chloroplast and nuclear DNA and 
representing all of the genera (some with many species) 
provided the structure for the trees (Funk and Chan 2008; 
McKenzie and Barker 2008) and the relationships among 
the clades was taken from Funk et al. 2004 and Funk and 
Chan, Chapter 23). 
Liabeae (Chapter 27). ? Liabeae are a monophyletic 
Neotropical tribe containing approximately 174 species 
distributed in 17 genera and occupying a wide variety of 
habitats throughout Mexico, Central America, the West 
Indies, and the Andes. The greatest diversity in the tribe 
is found in Peru, where no fewer than 14 genera and over 
70 species are represented. After a long history of moving 
from tribe to tribe, the current members were brought 
together by Robinson (1983). A previous morphological 
analysis resolved a northwestern Andean origin (Funk et 
al. 1996). The tree for our study was based on Dillon et 
al. (Chapter 27) and contains all the genera of the tribe 
except the monotypic Bishopanthus, which is only known 
from the type. Although the type was relatively recently 
collected, it is just a small piece of the original collection, 
most of which was destroyed by one of the collectors. 
Vernonieae and Moquinieae (Chapters 28, 30). ? 
The tribe Vernonieae with 126 genera and 1500 spe- 
cies has until recently had most of its species placed 
in the large and complicated genus Vernonia (ca. 1000; 
Jones 1977; Keeley and Robinson, Chapter 28). The tribe 
is widely distributed with centers of diversity in tropi- 
cal Africa and Madagascar, Brazil, and North America. 
However, it has been the subject of recent revisions that 
concentrated on recognizing monophyletic genera from 
within the 1000 species of the core genus Vernonia s.l. 
(e.g., Robinson 1999), first in the Americas and more re- 
cently in Africa and Asia. Vernonieae have recently been 
examined by Keeley et al. (2007; Keeley and Robinson, 
Chapter 28) based on ndhF, trnL-trnF, and ITS sequence 
data. Their work supports the monophyly of the tribe and 
the non-monophyly of Vernonia. However, in the analysis 
of the subfamily Cichorioideae (Funk and Chan, Chapter 
23), Distephanus had alternative placements: as the sister 
group to the rest of Vernonieae, or unresolved at the base 
with the tribe Moquinieae (Robinson, Chapter 30). 
Senecioneae (Chapter 34). ? Senecioneae are the 
largest tribe with over 150 genera containing 3500 species 
(Nordenstam 2007a) and they have a global distribution. 
Pelser et al. (2007), recently published a phylogenetic anal- 
ysis of the tribe based on ITS data that, while unresolved 
at the base, showed several well supported clades. The 
genus Senecio, which contained the majority of the species 
of the tribe, was shown to be non-monophyletic, and the 
authors indicated that revisions of the generic boundaries 
that are needed to achieve monophyletic groups are com- 
pleted or in progress. 
The relationship of Senecioneae to other clades is un- 
certain. The tribe is variously positioned as (1) the sister 
group to the rest of Asteroideae, (2) the sister group 
to the Calenduleae + Gnaphalieae + Astereae + Anthemi 
deae clade, or, in the least likely scenario, (3) the sis- 
ter group to the Inuleae + Athroismeae + Heliantheae 
Alliance clade. The support for its inclusion is stron- 
gest for option 1, but the relatively short branches make 
its placement there tentative (see Pelser and Watson, 
Chapter 33). This ambiguity will not be resolved until 
more taxa and characters from both plastid and nuclear 
markers are included in a tribal-level study of the sub- 
family. At this time we are following the resolution fa- 
vored by Panero and Funk (2008), which shows the 
Senecioneae in a polytomy with the clade formed by 
the Inuleae + Athroismeae + Heliantheae Alliance and the 
clade containing Calenduleae + Gnaphalieae + Astereae + 
Anthemideae. Doronicum and Abrotanella, the two addi- 
tional taxa in this polytomy, are Senecioneae genera that 
have been hard to place and may have to be excluded 
from the tribe (Pelser et al. 2007). 
Calenduleae (Chapter 35). ? The placement of Calen- 
duleae as the sister taxon to the Gnaphalieae + Anthem- 
ideae + Astereae clade is based on the Panero and Funk 
(2008) analysis as well as those by Kim and Jansen (1995) 
and Eldenas et al. (1999). The sister group relation- 
ship of Calenduleae to the other three tribes is strongly 
supported in the Panero and Funk (2008) study, even 
though the number of taxa sampled is small. The tribe 
Calenduleae has 12 genera with 120 species (Nordenstam 
2007b), and most genera have distinct centers of distribu- 
tion in southern Africa; most of the species occur in the 
Cape Floristic Region. However, one genus, Calendula, is 
found in northern Africa and the Mediterranean north to 
Central Europe and east into Turkey, Iraq, and Iran; but 
it is nested in the higher portion of the tree and so does 
not affect the biogeographic pattern. 
Gnaphalieae (Chapter 36). ? Gnaphalieae are a mod- 
erately large tribe whose members were traditionally in- 
cluded in the tribe Inuleae. It has only been recently 
that the tribe has been shown to be isolated from the 
remainder of "old" Inuleae (Anderberg 1989, 1991). The 
approximately 180?190 genera and ca. 1240 species of 
Gnaphalieae are most numerous in the southern hemi- 
sphere, with strong centers of diversity in southern Africa, 
Australia, and South America (Anderberg 1991; Bayer et 
al. 2007). The tree for this study was provided by Bayer 
and his collaborators (Ward et al., Chapter 36) and it is 
based on chloroplast DNA sequences for matK, the trnL 
intron, and the trnL-trnF intergenic spacer. The principal 
improvement of this tree over previously published DNA 
sequence phylogenies for Gnaphalieae is that it includes 
a broad sampling of genera from Africa and Australasia 
together with some from other continents. 
754 Funk et al. 
Astereae (Chapter 37). ? With 170 genera, ca. 3000 
species, and a worldwide distribution, Astereae are the 
second largest tribe after Senecioneae. It has centers of 
diversity in southwestern North America, the Andes, 
South Africa, Australia, and New Zealand. The tree 
presented in this book (Brouillet et al., Chapter 37) is the 
first global, molecular phylogenetic analysis of the tribe. 
It is based on ITS sequence data and shows that interrela- 
tionships among genera are better reflected by geographic 
origin than by the current classification. 
Anthemideae (Chapter 38). ? The tribe Anthemideae 
is composed of 111 genera and ca. 1800 species with 
main concentrations of species in southern Africa, the 
Mediterranean region, and Central Asia. The phytog- 
eny for the metatree was generated using data from 
two recent publications that used ndhF (Watson et al. 
2000; Himmelreich et al. 2008) and one that used ITS 
(Oberprieler et al. 2007). 
Inuleae and Plucheeae (Chapter 39). ? Plucheeae are 
now known to be nested within Inuleae, and so they are 
recognized as a single tribe with about 66 genera and ca. 
700 species (Anderberg and Eldenas 2007). The tree for 
this study was provided by Anderberg and his collabora- 
tors (Anderberg et al. 2005) based on ndhF data. Inuleae 
are a mainly Eurasian and east and southern African 
tribe, but some genera (e.g., Plucked) have a worldwide 
distribution. 
Athroismeae (Chapter 40). ? The tribe Athroismeae 
is the sister group to the rest of the large and diverse 
clade that is the Heliantheae Alliance. The five genera 
(only two were included in Panero and Funk 2008) and 
55 species in Athroismeae are centered in eastern tropical 
Africa and were in Inuleae until moved to Heliantheae 
s.l. (Eriksson 1991). 
Heliantheae Alliance (including Eupatorieae) (Chap- 
ters 41?43). ? The tribe Eupatorieae is nested in the 
Heliantheae Alliance, and former Heliantheae s.l. have 
been reorganized into twelve tribes (Baldwin et al. 2002; 
Panero and Funk 2002; Cariaga et al. 2008). Bremer 
(1994) divided this part of the family into three groups, 
Helenieae (including Athroismeae), Heliantheae, and 
Eupatorieae, but recognized that the groups would need 
to be re-arranged once additional information was avail- 
able. The studies of both Baldwin et al. (2002) and Panero 
and Funk (2002) showed Helenieae and Heliantheae of 
Bremer to be non-monophyletic, and they described ad- 
ditional tribes where needed. More recently, Cariaga et 
al. (2008) published a treatment of the problem genus 
Feddea based on ndhF sequence data. As part of their study 
the genus was placed in a new tribe by itself, Feddeeae, 
located as the sister group of the "rest" of the Heliantheae 
Alliance (minus Athroismeae). The inclusion of the tribes 
Feddeeae and Eupatorieae in the Heliantheae Alliance 
brings the total number of tribes in the Alliance to 13. 
The tree for this clade in the metatree was formed by 
using the Baldwin treatment of the Heliantheae Alliance 
(Chapter 41), the Coreopsideae treatment of Crawford et 
al. (Chapter 42), and the Funk et al. paper (2005). The 
branching within Eupatorieae was taken from Robinson 
et al. (Chapter 43). 
The tree for the Heliantheae Alliance section of the 
family contains 160 out of ca. 460 genera and so rep- 
resents about 35% of the generic diversity of this clade. 
This is the lowest percentage for any clade on the meta- 
tree, however the poor representation is found primar- 
ily in three tribes, Eupatorieae (the tree has 25 genera 
represented out of a total of 182; there are 2200 species), 
Heliantheae (6 out of 113 genera were represented; there 
are 1461 species), and Millerieae (3 genera out of 36 
were represented; there are 380 species). When totaled 
together, these three tribes are represented by only about 
10% of the generic diversity within them. The other ten 
tribes in the Alliance are much better represented, some 
at or close to 100% (see below). Because the members of 
former Helenieae form the basal grade, the under-repre- 
sentation of three of the more highly nested groups does 
not present an obstacle to the biogeographic analysis, al- 
though it does give an under-estimate of the importance 
of the northern and central Andes. 
The tribe Heliantheae s.l. was broken up by Baldwin 
et al. (2002) and by Panero and Funk (2002) when 
Eupatorieae were found to be nested within what is now 
referred to as the Heliantheae Alliance (Fig. 44.1). Most 
of the new tribes, however, were actually not new and 
had been described previously by others but not picked 
up by the synantherological community. In fact, only 
three of the tribes recognized by Baldwin needed to be 
described as new (Baldwin et al. 2002): Bahieae (17 out 
of 20 genera were represented in the analysis; there are 
83 species), Chaenactideae (all 3 genera were represented; 
29 species), and Perityleae (4 out of 7 genera were rep- 
resented; 84 species) (see Funk et al., Chapter 11). Other 
tribes in the Heliantheae Alliance (not mentioned above) 
include: Coreopsideae (21 genera out of 30 were repre- 
sented; 550 species), Helenieae (all 13 genera were repre- 
sented; 120 species), Madieae (35 genera were represented 
out of 36; 203 species), Neurolaeneae (1 out of 5 genera 
was represented; 153 species), Polymnieae (the only genus 
was represented; 3 species), and Tageteae (17 out of 32 
genera were represented; 267 species). 
Area optimization analysis using parsimony 
The terminal branches of the metatree were colored 
based on the distribution of each terminal taxon; taxa 
that span more than one area have multiple colors (Figs. 
44.1?44.7). The internode distributions were mapped 
onto the metatree using the Farris double pass method 
(1970). The results of the mapping were checked using 
Chapter 44: Compositae metatrees: the next generation 755 
the PAUP 'Acctran' option (Swofford 2002). These tech- 
niques provided the hypothesized distributions at deep 
branches and nodes. 
Following the theory that bold hypotheses are better 
than weak ones (courtesy of Popper), equivocal situa- 
tions were resolved when possible to present the most 
predictive estimate of the biogeographic history. In a 
few instances there were equivocal resolutions which 
were left black, or if the two areas were contained in a 
single continent, they were coded for that continent (e.g., 
general Africa). In essence, we created an 'area metatree' 
as opposed to an 'area cladogram'. In the summary tree 
and unrooted tree (Figs. 44.1, 44.2), some of the biogeo- 
graphic areas were combined (e.g., North America was 
combined with Mexico). 
RESULTS AND DISCUSSION 
The first supertree (= metatree) for Compositae was pub- 
lished in 2005 (Funk et al.), and since then there has been 
considerable progress in the reconstruction of evolution- 
ary relationships in many clades. In fact, we now have 
robust phylogenies for most of the clades in the family. 
Descriptions and diagnostic characters for all of the tribes 
and critical clades are found in Chapter 11. Without a 
doubt the most substantial progress has been made in the 
large and complicated Astereae, Cichorieae, Senecioneae, 
and Vernonieae tribes, all of which were problematic in 
the 2005 publication (Funk et al. 2005) but now have 
their first comprehensive molecular phylogenetic hypoth- 
eses (Keeley et al. 2007; Pelser et al. 2007; Brouillet et al., 
Chapter 37; Kilian et al., Chapter 24; and other references 
in the corresponding chapters). For the first time within 
these tribes we have a fairly good idea of what the basal 
groups are and where different clades are found, and we 
know that the large genera with global distributions are 
not monophyletic. 
Considering the entire metatree, most of the tradi- 
tional thirteen tribes were found to be monophyletic 
or could easily be made monophyletic with only a few 
rearrangements. The big exceptions to this are Mutisieae 
(sensu Cabrera) and the Heliantheae Alliance, both were 
broken up into many groups. The genera that were once 
placed in Mutisieae by Cabrera or others are now in four- 
teen tribes, Helenieae are in seven, and Heliantheae are 
in six (including Feddeeae). 
For such a large and interesting family, relatively little 
has been published on its geographic origin and diversi- 
fication since Bentham (1873b). Bentham (1873b), Small 
(1919), Raven and Axelrod (1974), and Turner (1977) all 
believed that Compositae had their origin in the north- 
west portion of South America, in the Andes. Rzedowski 
(1972) and Hu (1958) pointed out the high diversity of the 
family in montane areas. More recently, Bremer (1992, 
1994) developed a method he called Ancestral Areas 
Analysis' and came to the conclusion that the family 
originated in "South America and the Pacific". DeVore 
and Stuessy (1995) suggested that the family originated 
in southern South America, which was re-emphasized 
by Bremer and Gustafsson (1997). Graham (1996) sum- 
marized the fossils for the family but had wide estimates 
of the age of some of the pollen. Other than these efforts, 
little attention has been paid to this topic. Perhaps the 
size of the family, its global distribution, the lack of mac- 
rofossils and paucity of discriminating characters in fossil 
pollen, and the lack of an agreed upon phylogeny have 
restricted attempts to understand its history. 
The meta showing the overall phylogeny of Compositae 
allows us to use information from the most recent avail- 
able molecular phylogenies to look at the family as a 
whole and to try to discern its origin and history. It is also 
an excellent method for determining critical areas of the 
tree for future work (Funk and Specht 2007). 
The metatree and its sections 
In order to more easily discuss the tree it has been bro- 
ken into sections. Section 1 (Fig. 44.3) covers the Basal 
Grade, from the outgroups through monotypic Gym- 
narrheneae. Section 2 (Fig. 44.4) covers the large subfamily 
Cichorioideae. Section 3 (Fig. 44.5) covers Corymbieae, 
Senecioneae, Calenduleae, and Gnaphalieae; Section 4 
(Fig. 44.6) Anthemideae and Astereae; and finally, Section 
5 (Fig. 44.7) Inuleae, Athroismeae, and the Heliantheae 
Alliance (including the Eupatorieae). Figure 44.8 has 
some of the proposed ages of the clades and Figs. 44.9 
and 44.10 show some of the morphological variation. 
Since we have no macrofossil data, the following dis- 
cussion is based on extant taxa. 
Section 1, Basal Grade (Figs. 44.3, 44.9A-D). ? 
Except for Calyceraceae (the sister group of Compositae), 
the most closely related families to Compositae are 
found in Australia, New Zealand, New Guinea, and 
New Caledonia (purple lines; Fig. 44.3). The members of 
Calyceraceae are from southern South America. 
The first split within Compositae is between the sub- 
family Barnadesioideae and the remainder of the family 
(Fig. 44.3). Gustafsson et al. (2001) and Stuessy et al. 
(Chapter 13) examined the biogeography and concluded 
that the Barnadesioideae clade has its origin in southern 
South America; this is confirmed by our analysis. In the 
sister group of Barnadesioideae the relationships among 
the basal groups are largely unresolved and are shown 
as a trichotomy (Fig. 44.3). However, this part of the 
tree could have been shown as a polytomy containing 
four or even five clades because support for monophyly 
of the subfamily Wunderlichioideae is not consistently 
strong, nor is its phylogenetic position; this ambiguity is 
756 Funk et al. 
Outgroups Bamadesieae Hyalideae Gochnatieae 
.?Oc-?cocDajOt 
?o =3? R>--o 0.9- - 
^?J5 ^z c^Onio - . ^ - ^ -? >v j -. -k. - - --,-- - >, ^-^ - --. ^,_-vwi ui . Ul- v. 
T?> 
Fig. 44.3. The metatree of Compositae has been broken up into five figures with two to three parts for each figure. The original 
trees are from the various chapters in this volume, but some taxa with redundant distributions have been pruned from the tree to 
save space. Figure 44.3 covers the Basal Grade of the family and includes the outgroups through Gymnarrheneae, including this- 
tles (some of the internodes have been compressed). All outgroups except for the sister group are Australasian. The extant taxa 
from the sister group of the family, Calyceraceae, along with those from the basal grade of Compositae have a southern South 
American origin. For subfamily groups see Chapter 11, for color chart see Fig. 44.7. Gy. = Gymnarrheneae; H. = Hecastocleideae; 
O. = Oldenburgieae; Onoser. = Onoserideae; Perty. = Pertyeae; Tar. = Tarchonantheae; Wund. = Wunderlichieae. 
Chapter 44: Compositae metatrees: the next generation 757 
Dicomeae 
Tar. 
SS m 5 baa coKmujcLS 
Cardueae 
OQrJcj 
?2 co     S 2 to | 
Perty.  ^ 
CM C 2 Q.C 
P?      ."CO CD'S 
==i?    o 0) qj J= JBsgssg 
SScB X=.S S, 
p. 758 
General Africa 
indicated in Fig. 44.3 by a dotted line. However, many of 
the main clades basal to the clade formed by Hecastodeis 
and its sister group are consistently resolved as having a 
southern South American origin, with the exception of 
the tribe Wunderlichieae whose members are found in 
the Guiana Shield and Brazil. The large Mutisioideae 
clade (composed of the tribes Mutisieae, Nassauvieae, and 
Onoserideae) contains mostly southern South American 
taxa, but it also contains Gerbera from tropical and south- 
ern Africa and Asia, North America taxa (e.g., Acourtia), 
and Leibnitzia from Asia and Mexico. Hyalideae have 
two clades, one from Asia and one from southern South 
America. Gochnatieae contain genera mainly from south- 
ern South America and Brazil, but there is also a radiation 
in Cuba. It is clear from the optimization that the extant 
taxa at the base of the Compositae metatree have their 
origin in southern South America. 
The internode between the southern South American 
grade and the beginning of the African radiation (labeled 
"General Africa" in Fig. 44.3) is left unresolved as to 
origin because there are no areas shared among the three 
(South American base, African radiation, and the North 
American genus Hecastodeis). A species level analysis of 
the tribe Gochnatieae (4?5 genera) is underway (Sancho 
et al., pers. comm.) and its relationships to Hecastodeis 
may provide some insight into the problem, because one 
of the genera (Gochnatia) is found in South America, the 
West Indies, and North America. 
758 Funk et al. 
Cichorieae Arctotideae-Arct 
Arctotidieae-Gort 
ill 
a 8 E Hi 
?S 5 S S I W&S 
, 8fea S5S* 
? ?OU.KC5( 
frT>J.S 
gin a ?!roo 
jOOJ535?:n.a. 
s3sr? - KsaA^A, 
U 
il 
II 
u 
ff 
I 
y 
F Lr 
-:?  p. 760 
p. 757 <? 
Fig. 44.4. Monotypic Cichorioideae (internodes have been compressed). For subfamily groups see Chapter 11, for color chart see 
Fig. 44.7. Di. = Distcphanus; Er. = Eremothamneae; He. = Hccastodeis; Mo. = Moquinieae; PL = Platycarpheae. 
The largest clade of the basal grade contains the sub- 
family Carduoideae (Tarchonantheae, Oldenburgieae, 
Dicomeae, Cardueae; Fig. 44.3); this is followed on the 
metatree by the Pertyeae (Asia) and Gymnarrheneae 
(northern Africa). At the base of Carduoideae are several 
former members of Mutisieae from southern and tropi- 
cal Africa (African Mutisieae). The relationships of these 
clades to one another are unresolved at this time, except 
for the sister group relationship between Oldenburgieae 
and Tarchonantheae. The thistles (Cardueae) are mono- 
phyletic and show a Mediterranean?northern African ra- 
diation with numerous incursions into Eurasia and Asia. 
The combination of the Mediterranean?northern African 
base of the thistles and the tropical and southern African 
Tarchonantheae, Oldenburgieae, and Dicomeae give a 
'general Africa' base to this clade. 
Chapter 44: Compositae metatrees: the next generation 759 
Liabeae 
.Mo. 
E.  E 
E . 
= a 
,E c 
s-s-sa-sjGssss 
no 
cooooooo 
&ESSSSSS 
U 
ff ffu 
t 
u 
U 
a 
U 
\ 
Tropical, 
America 
The sister group of Carduoideae is the remainder of 
the family (Pertyeae, Gymnarrheneae, Cichorioideae, 
Corymbieae, and Asteroideae) all of which, except for 
Pertyeae (Asia), presumably originated in Africa. The first 
group to split off is Pertyeae followed by Gymnarrheneae 
(Northern Africa) followed by Cichorioideae. 
Section 2, subfamily Cichorioideae (Figs. 44.4, 44.9E, 
F). ? This large clade contains six tribes: Cichorieae 
(Fig. 44.4; also referred to as Lactuceae) is the sister 
group to the remainder. This tribe has a Mediterranean- 
northern African base with independent radiations in 
North America and Asia. Interestingly, the main North 
American clade of Cichorieae is not nested within the 
Asian radiation as was predicted (Funk et al. 2005). In 
that paper, it was thought that the biogeographic pathways 
of Cichorioideae would lead from the Mediterranean via 
Eurasia to Asia and across to North America but it seems 
that the Asian and North American taxa are separately 
derived from Mediterranean clades. 
At the base of the rest of the subfamily Cichorioideae 
there are five clades containing members of the former 
Arctotideae: two are subtribes of that tribe (Arctotidinae 
and Gorteriinae), two are now recognized at the tribal 
level (Eremothamneae and Platycarpheae), and one is an 
unplaced genus (Heterolepis). All are from southern Africa 
(Fig. 44.4) and are prominent members of the Cape Floral 
Region, which is the subject of intense conservation in- 
terest. Because all of the basal taxa in each subtribe are in 
760 Funk et al. 
Senecioneae 
E eg 3 S| 
OWO< 
5-2-2 ^ ? 2 
- 5 c 8 =c to qj qj en .2 o> (0=3 0)^  O-C 0) UBSQOto 
-a 
#3 
>QWUj3Uj?tO 
Australasia 
& Pacific 
Sub-Saharan 
p. 758  ? 
Fig. 44.5. Corymbieae, Senecioneae, Calenduleae, and Gnaphalieae. For subfamily groups see Chapter 11, for color chart see 
Fig. 44.7. Co. = Corymbieae. 
southern Africa, the lack of evidence for the monophyly 
of Arctotideae does not affect the biogeographic hypoth- 
eses produced in this study. 
The tribe Platycarpheae (southern Africa) is the sis- 
ter taxon of the Liabeae + Vernonieae clade (including 
Distephanus and Moquinieae) but without strong support. 
This clade is nested in a grade formed by the southern 
African clades (Fig. 44.4). Liabeae are predominantly 
central Andean and the tribe is believed to have origi- 
nated in northern Peru and southern Ecuador with small 
incursions into Central America and radiations in Mexico 
(Sindairid) and the Caribbean (Liabum). The basal branches 
of Vernonieae are from the area we have designated as 
'tropical Africa and Madagascar'. New to the analysis is 
Chapter 44: Compositae metatrees: the next generation 761 
Calenduleae Gnaphalieae 
SB 
IE 
is.i @es 
El     . 
wiES-SiE^T 
soooof 
S 2 5E- 
??{?.8 Sg 
1       E 
3^ P- 764 
the small Brazilian tribe Moquinieae (Pseudostifftia and 
Moquinid). In some of the analyses the inclusion of this 
tribe results in Distephanus changing position from being 
the sister group of the rest of the tribe to being ambiguous 
at the base of the Vernonieae-Moquinieae clade (Keeley 
and Robinson, Chapter 28; Funk and Chan, Chapter 23). 
More highly nested members of Vernonieae are from 
Brazil and North America. In Vernonieae, the unusual 
North American genera Stokesia and Elephantopus are not 
in the main North American clade but rather represent 
two independent lineages (Fig. 44.4). With the excep- 
tion of Liabeae, every tribe or subtribe in Cichorioideae 
s.str. has its origins in Africa, either north, tropical or 
southern, in effect covering the whole continent. As a 
result the final biogeographic resolution of the subfamily 
is listed as 'General Africa'. 
Sections 3?5 (Figs. 44.5?44.7) cover Corymbieae 
(Corymbioideae) and its sister group Asteroideae. 
Section 3, tribes Corymbieae, Senedoneae, Calen- 
duleae and Gnaphalieae (Figs. 44.5, 44.10). ? The tribe 
Corymbieae (Corymbioideae) consists of only one genus, 
Corymbium, and this distinctive group is restricted to 
southern South Africa (Nordenstam 2007c; Fig. 44.5). 
Asteroideae encompass the remainder of the family 
phylogeny, and it is the largest subfamily. It was recog- 
nized by Cassini (1816) and Bentham (1873a) due to the 
762 Funk et al. 
Anthemideae 
Fig. 44.6. Anthemideae and Astereae. All taxa are in the subfamily Asteroideae; see Fig. 44.7 for the color chart. 
Chapter 44: Compositae metatrees: the next generation 763 
Astereae 
RE 
?'
S
-IJi 
.2.2 ? !2 c 
11 | I'i 
CLCLSOLL 
i111 si 
D ^"> COO 
???0 c" 53 53 
^-co o E co 
CO O 5 CO  C; CO o 
3E ?s.?E 
Jl SI 
1 Q.CD"S       'C-C CD'C^ % "9-    ^-coSS?
tt 
E I-Is 
g8#a o.S|Si> 
S^c O 0) t: 
CD ?-E 
'c?S 
si K 
m a 
combination of its capillary pappus, true rays, and a recep- 
tive area in two lines on the inside of the style branches. 
Senecioneae have long been one of the largest and 
most difficult groups to understand; they are truly a 
global tribe with major radiations in sub-Saharan Africa, 
West and East Asia, Andean South America, and Mexico. 
Because of uncertainty about the phylogenetic positions 
of the core of Senecioneae and two of the genera usually 
assigned to this tribe, relationships among these taxa and 
the clade formed by the other Asteroideae tribes are pres- 
ently unresolved. 
Two genera of Senecioneae, Doronicum (Eurasia and 
northern Africa) and Abrotanella (Australasia and southern 
South America), "fall out" of monophyletic Senecioneae. 
Since there is a general problem of tribal relationships 
among Senecioneae and its potential sister groups, 
there are not enough data to determine whether or not 
these two genera should stay in the tribe as subtribes 
or be moved to tribes of their own. The authors of the 
Senecioneae Chapter (Nordenstam et al., Chapter 34) 
have reserved final judgment on this matter until they 
have more information. 
Within the core Senecioneae clade there are four clades 
that form a polytomy. One represents the bulk of the 
species, which are found in two monophyletic subtribes 
that are sister taxa (Othonninae + Senecioninae). This 
major clade has a sub-Saharan African base with highly 
nested groups of species from South America and Central 
America?Caribbean basin. The second clade, the core of 
subtribe Tussilagininae, has clades in Asia, Eurasia, North 
764 Funk et al. 
Inuleae Coreopsideae 
% .Jo      o ^.5 
I Hill I'll u u 
? 
a 
SW USA & NW Mexico 
P. 761 -C1C- 
Fig. 44.7. Inuleae, Athroismeae, and the Heliantheae Alliance. All taxa are in the subfamily Asteroideae. See p. 766 for color chart. 
Chaen. = Chaenactideae; F. = Feddeeae; Miller. = Millerieae; N. = Neurolaeneae; P. = Polymnieae; Ferity. = Perityleae. 
Chapter 44: Compositae metatrees: the next generation 765 
Tageteae Bahieae Eupatorieae 
America, Mexico, and South America, but the relation- 
ships among these clades are not resolved. Two small 
clades complete Senecioneae. One is the South American 
genus Chersodoma, and the other is a clade composed of 
two groups: (1) an Australasian-Pacific subclade of nine 
genera, and (2) a subclade composed of a few succulent 
species from sub-Saharan Africa that were historically 
included in Senecio (Senecio m-w group, Fig. 44.5). 
On the basis of the biogeographic patterns observed 
in the metatree, one can certainly propose a sub-Saharan 
origin for the Othonninae + Senecioninae clade and pos- 
sibly a southern African origin for the tribe with ra- 
diations into other areas.  The southern African origin 
is reinforced by the fact that the most closely related 
clades in more basal and derived positions in the metatree 
have a southern African origin (i.e., Corymbieae and the 
clade made up by Calenduleae + Gnaphalieae + Astereae 
+ Anthemideae). 
The small tribe Calenduleae is found almost exclusively 
in Africa with its greatest diversity in southern Africa; 
the tropical and northern African groups are nested 
high in the tree. It is the sister taxon of Gnaphalieae + 
Anthemideae + Astereae clade (Figs. 44.5, 44.6). 
The extant members of the Gnaphalieae had a major 
radiation in southern Africa early in their history 
with large radiations into Australia and New Zealand. 
766 Funk et al. 
S-S.g 
la ? ^wm^ i&g #mmm.## 
e (5 erg ?3 ff saS'E?S-aS'S.S5 cSS*? ? g-Ega - , Cn3C^,(OCQ)t03CO-9; CD O m(0T3 0)0)O(0W0)^)a)^C(0^ 
ISO 
I' 
North America 
Eurasia 
p. 765 <? 
South America 
Brazil 
Guiana Shield 
North & central Andes 
Southern Andes, southern South America 
General South America 
North America 
Mexico 
Central America, Caribbean 
Europe 
Eurasia 
Eastern & central Asia 
Southern & southeastern Asia 
Southern Africa 
Madagascar, tropical Africa 
Northern Africa, Mediterranean, southern Europe 
General Africa 
Australia and the Pacific 
Australia, New Guinea, New Caledonia 
New Zealand 
Pacific Islands 
Hawaii 
Widespread or ambiguous 
Africa 
Although not reflected in this figure, this tribe also has 
large highly nested groups of taxa in South America and 
Asia indicating dispersal to these regions as well (Ward 
et al., Chapter 36). 
Section 4, tribes Anthemideae and Astereae (Fig. 
44.6, pp. 762, 763; Fig. 44.10C). ? Sister to Gnaphalieae 
is the clade consisting of Anthemideae + Astereae (Fig. 
44.6). The tribe Anthemideae has a southern African grade 
at the base followed by a Mediterranean?northern African 
clade as well as one or two Asian clades (Oberprieler et 
al. 2009, Chapter 38). 
The phylogeny of the tribe Astereae (Fig. 44.6) is not as 
clearly based in southern Africa as are the other tribes in 
this clade: Calenduleae, Gnaphalieae, and Anthemideae. 
Nevertheless, this origin is the most parsimonious expla- 
nation for the basal grade of this tree. Although the tribe 
is nested among clades with a southern African origin, 
there are several taxa from other regions that are found 
in basal positions in the Astereae clade (e.g., Nannoglottis 
from south-central China, a clade from South America, 
and one from New Zealand). More highly nested in the 
tribe are some tropical African and Asian groups as well 
a clade with representatives in South America, North 
America, and Australia, although their relationships to 
one another are somewhat unresolved. 
The extant members of the large clade consisting of 
Calenduleae + Gnaphalieae + Anthemideae + Astereae, 
has an African origin, most likely sub-Saharan or south- 
ern Africa. As mentioned earlier it is possible that the 
Senecioneae are the sister group of this clade. 
Section 5, tribes Inuleae, Athroismeae and the Heli- 
antheae alliance (Fig. 44.7, pp. 764-766; Fig. 44.10D- 
F). ? The next clade on the metatree (Fig. 44.7) contains 
Inuleae (including Plucheeae). The tribe is divided into 
two subtribes, Plucheinae and Inulinae (Anderberg et 
al., Chapter 39). The Inulinae clade has a split between 
a Mediterranean?northern African clade and an Asian 
clade. The Plucheinae clade has a southern African basal 
polytomy (except for Stenachaeniam) with a pantropical 
clade nested within (including southern Africa, tropical 
Chapter 44: Compositae metatrees: the next generation 767 
Africa, and northern Africa). Given that one subtribe 
has the potential for being rooted in the Mediterranean- 
northern African area and the other in southern Africa 
and that the clades basal to Inuleae as well as Athroismeae 
are most likely rooted in sub-Saharan Africa or southern 
Africa, it seems likely that Inuleae have an African origin, 
and it is shown as 'General Africa' in origin in Fig. 44.7. 
The tribe Athroismeae is the sister group of the 
Heliantheae Alliance and includes five genera from 
Africa, mostly from the tropical eastern region (Fig. 44.7). 
This clade marks the end of the African influence on 
the family and signals a dramatic shift to the Americas, 
most notably southwestern United States (SW USA) and 
northwestern Mexico (NW Mexico). 
The recently described tribe Feddeeae is endemic to 
Cuba and is supported as being part of the Heliantheae 
Alliance (Cariaga et al. 2008). However, it may be the 
sister group to the rest of the Alliance, grouped near the 
base, or related to Athroismeae. For now it sits with some 
ambiguity at the base (Fig. 44.7). 
The core Heliantheae Alliance begins with the tribe 
Helenieae and its sister group (Fig. 44.7). This clade 
has strong support. Many of the clades within the core 
Heliantheae Alliance are ambiguous as to whether 
they are rooted in Mexico or North America (north 
of Mexico). This is the result of the somewhat artificial 
political categories selected for the biogeographic portion 
of this analysis. Some of the clades of the Heliantheae 
Alliance are from both SW USA and NW Mexico and 
frequently switch from one location to the other or in- 
habit both. Other clades are more firmly affiliated with 
either Mexico or North America (north of Mexico). For 
instance, the tribe Madieae (Fig. 44.7) is almost totally 
in North America (north of Mexico) while Helenieae 
(Fig. 44.7A), Coreopsideae (Fig. 44.7), Tageteae (Fig. 
44.7), and Bahieae (Fig. 44.7), are frequently found in 
both areas. For tribes such as Heliantheae (Fig. 44.7) and 
Millerieae (Fig. 44.7), there are too few taxa sampled to 
make a decision on the origin of these clades. These sam- 
pling concerns are minor since the root of the entire ra- 
diation is clearly in NW Mexico and the SW USA, with 
repeated incursions into Central America, the Andes, 
and back to North America. This agrees with Baldwin 
et al. (2002) who said, "the most recent common ances- 
tor of taxa referable to Helenieae s.l. (and to Heliantheae 
s.l. + Eupatorieae) ... probably occurred in southwestern 
North America (including northern Mexico)." Baldwin et 
al. (2002) also pointed out that the endemic Californian 
diversity in the Heliantheae Alliance is mostly confined 
to one clade, Madieae. 
Nested within the Heliantheae Alliance is the large 
and distinctive tribe Eupatorieae (Fig. 44.7), a large New 
World tribe with its base in Mexico and repeated disper- 
sals to Brazil, South America, and North America. 
What happened in the history of Compositae between 
the radiations in Africa and the Heliantheae Alliance in 
North America? Previously, Funk et al. (2005) specu- 
lated that since the base of the Heliantheae Alliance was 
in the SW USA and NW Mexico, the path from Africa 
to North America and Mexico might have been via 
Asia. However, if Feddea (Cuba) is the sister taxon of the 
core Heliantheae Alliance, then that proposition seems 
less likely. One possibility might be something like a 
peri-Tethyan dispersal, but these dates (late Triassic 6?2 
Ma) would make the clade much younger than previ- 
ously thought. Much depends on whether or not Feddea 
is ultimately supported as the sister group of the core 
Alliance. 
The summary trees for the family (Figs. 44.1, 44.2) 
show the results of the parsimony mapping of the dis- 
tributions. In this condensed tree it is even more evi- 
dent that extant Compositae had a South American base 
with an African diversification and radiation into Asia, 
Eurasia, Europe, Australia, etc. followed by the burst of 
diversification in North America. The unrooted diagram 
provided greater clarity as to the biogeographic patterns 
of the phylogeny (Fig. 44.2). 
Odd genera 
Throughout the history of the classification of Compositae 
there have been a number of difficult-to-place genera. 
Funk et al. (2005) and Panero and Funk (2008) discussed 
how important these genera were to resolving biogeo- 
graphic hypotheses for the family. These problem gen- 
era were traditionally grouped with taxa that they were 
"less different from" rather than groups with which they 
shared characters. It is interesting to note that many of 
these taxa have secondary or tertiary heads, with primary 
heads reduced to one or a few florets and then re-aggre- 
gated onto a common receptacle. As a result they usually 
lack ray florets and do not have the common involucral 
and receptacular characters, adding to the difficulty of 
assigning them to tribe. 
The advent of molecular data has allowed us to de- 
termine the relationships of many of these odd genera. 
Some that have relevance to the biogeography of the 
family are discussed here. Their positions have turned 
out to be among the more interesting aspects of this 
study because they are frequently relatively species-poor 
sister groups of large radiations: Cratystylis, in Plucheinae, 
Athroismeae, or even Feddea, as the sister group to the 
Heliantheae Alliance, Corymbium as the sister group to 
Asteroideae, Platycarpha as the sister group to the Ver- 
nonieae + Liabeae clade, Gymnarrhena as the sister group of 
Cichorioideae + Asteroideae, African Mutisieae at the base 
of the thistles, and Hecastodeis as the sister group to the 
major radiation of the family. All of these have important 
phylogenetic positions for the biogeographic analysis and 
768 Funk et al. 
illustrate the fact that odd taxa should always be included 
in analyses at all levels (Funk et al. 2005; Funk and Chan, 
2008; Panero and Funk 2008). Although, some of these 
taxa are on long branches and their position may be af- 
fected by 'long branch attraction'. 
Age of origin 
Considering the size and importance of Compositae, sur- 
prisingly little has been published about the possible area 
of its origin or its age since Bentham (1873a, b). As men- 
tioned before, one reason may be because of the absence 
of any reliable macrofossils from the early diversificaiton 
of the family. A few individuals have guessed at a possible 
age. Turner (1977) thought that the family originated in 
the mid-late Cretaceous (ca. 100 Ma), possibly near the 
time of the first upheaval of the Andes (ca. 90 Ma). Other 
recent estimates include 60 Ma (Zavada and de Villiers 
2000), 53-43 Ma from DeVore and Stuessy (1995), and 
38 Ma from Bremer and Gustafsson (1997). In the 2005 
supertree paper (Funk et al.), an examination of the re- 
lationship of Compositae to its two most closely related 
families was used to suggest an age of around 50 Myr for 
the separation of Compositae + Calyceraceae (southern 
South America) from Goodeniaceae (Australia). 
Lundberg's study (Chapter 10) included the whole of 
the order Asterales. In addition to Goodeniaceae, the other 
families of the order that are closely related to Compositae 
are all found in Australia, New Guinea, New Caledonia, 
and/or New Zealand (Fig. 44.3). As a result of these dis- 
tribution patterns, one can hypothesize that the ances- 
tor of these eight families of Asterales had a Gondwanan 
distribution, and that the split between the ancestor 
of Goodeniaceae and the ancestor of Calyceraceae + 
Compositae took place with the formation of the Drake 
Passage that separated South America and Australia from 
Antarctica. Estimations of when that passage was formed 
range from middle Eocene to Oligocene to early Miocene 
but recent evidence narrows it to 50?41 Ma (Ghiglione 
et al. 2008 and references cited therein). The earlier date 
reflects a spreading with low incursions of water and the 
younger time period reflects a deeper water passage. The 
question then becomes how deep and wide did the Drake 
Passage have to be to prevent easy dispersal of pollen and 
seeds? Other factors to consider include the fact that the 
oldest part of the Andes Mountains is the southern section, 
and the uplift of this area began ca. 90 Ma and lasted until 
ca. 50 Ma. The mountains were high enough to cause a 
drying effect only late in this time period; in fact, pollen 
records show that 53 Ma southern South America was 
forested. So, the earliest time of separation between the 
continents coincides with the final uplift of the southern 
mountains. Geological, climatic, and ecological consider- 
ations, therefore, can be used to suggest an origin of the 
Calyceraceae-Compositae clade at some time after 50 Ma 
(perhaps as recent as 41 Ma), with the base of Compositae 
radiating as the Andes developed. Since Africa drifted 
away from Gondwana some time before South America 
and Australia each drifted away from Antarctica, it appears 
unlikely that the movement of the African continent had 
any influence on the base of the cladogram. 
Within the family, most authorities agree that, based 
on pollen data (Germeraad et al. 1968; Muller 1970), 
most of the current tribes were in existence by the end of 
the Oligocene (25-22 Ma; Muller 1981). An older date 
is given by Graham (1996) who dates the earliest pollen 
from Mutisieae as Eocene to middle Oligocene (50?25 
Ma), pollen from the Astereae-Heliantheae-Helenieae 
group as Eocene (50?35 Ma), and pollen of the Ambrosia- 
type (Heliantheae) from latest Eocene/early Oligocene 
(35?25 Ma). Given the phylogenetic position of taxa with 
the Ambrosia-type pollen, we can use the date of 35?25 
Ma for the base of the Heliantheae Alliance (Fig. 44.8). 
There are four Hawaiian taxa estimated to have diverged 
7?5 Ma nested high in the metatree. A radiation in 
the northern Andes (Espeletiinae), with an age of ap- 
proximately 2 Myr, is in line with the occurrence of the 
sub-paramo habitat. The tribe Liabeae is a north-central 
Andean clade that can be dated 15?5 Ma when the central 
Andes were uplifted. Finally, there are taxa from the basal 
grade that are found on the Guiana and Brazilian Shields; 
these plants inhabit areas where the rock is older than the 
family. For instance, in the Guiana Shield area, the final 
uplift was probably in the Cretaceous (Gibbs and Barron 
1993), and so predates the origin of Compositae and is of 
no help in determining the ages of those clades. 
The authors of some of the chapters in this book have 
speculated as to the age of origin of their clades. The 
Barnadesieae clade, which is the sister group to the rest of 
the family, is estimated to be at least 23 Myr old (Stuessy 
et al., Chapter 13). A minimum age of 23?28 Myr (Late 
Oligocene) for fossil pollen related to the extant genera 
of Gochnatieae, and a minimum of 20?23 Myr (Early 
Miocene) for fossil pollen ofNassauvieae and Barnadesieae 
were reported (Katinas et al. 2007). On the basis of ITS 
divergence, Wang et al. (2007) suggested a date of 29?24 
Ma for the separation of Cardueae from the African 
(former Mutisieae) tribes; and in Chapter 20, Susanna 
and Garcia-Jacas stated that Cardueae originated as part 
of the Tertiary flora and benefited extensively from the 
new habitats that were open during the deep climatic and 
geological changes during the Miocene (24?5 Ma), based 
on data from Cox and Moore (2004). In Cichorieae, 
Kilian et al. (Chapter 24) point out that the fossil record 
shows three different types of echinolophate pollen, i.e., 
the Cichorium intybus L. type (age 22?28.4 Myr; Hochuli 
1978), the Scorzonera hispanica L. type (minimum age 3.4 
Myr; Blackmore et al. 1986), and the Sonchus oleraceus 
type  (minimum age 5.4 Myr; Blackmore et al.   1986), 
Chapter 44: Compositae metatrees: the next generation 769 
^ 
? 23 ma 
20-23 ma ?? 
-23-28 ma tE 
24-29 ma 
25-36 ma % 
tE 
29-30 ma 
26-29 ma 
35-39 ma 
17 ma ^ 
25-35 ma 
35-50 ma 
Stylidiaceae 
Alseuosmiaceae 
Phellinaceae 
Argophyllaceae 
Menyanthaceae 
Goodeniaceae 
Calyceraceae 
Barnadesieae 
Stifftieae 
Hyalideae 
Wunderlichieae 
Onoserideae 
Nassauvieae 
Mutisieae 
Gochnatieae 
Hecastocleideae 
Tarchonantheae 
Oldenburgieae 
Dicomeae 
Cardueae 
Pertyeae 
Gymnarrheneae 
l/Var/on/aC1-3 
Cichorieae C4 
Cichorieae C5 
Platycarpheae 
Liabeae 
Moquineae 
Vernonieae 
Heterolepis 
Eremothamneae 
Arctotideae-A 
Arctotideae-G 
Corymbieae 
Senecioneae 
Calenduleae 
Gnaphalieae 
Astereae 
Anthemideae 
Inuleae 
Athroismeae 
Feddea 
Helenieae 
Coreopsideae 
Neurolaeneae 
Tageteae 
Chaenactideae 
Bahieae 
Polymnieae 
Heliantheae 
Millerieae 
Perityleae 
Eupatorieae 
Madieae 
Fig. 44.8. There are few dates that can be placed on the phylogeny of Compositae with any certainty. The separation of the 
outgroup lineages from that of Calyceraceae-Compositae may be placed at a time when Australia separated from Antarctica- 
South America (the flora is believed to have separated about 50-41 Ma), and the radiation at the base of Compositae may be 
linked to uplift of the southern Andes. 
770 Funk et al. 
that were used to calibrate the phylogenetic tree; and es- 
timates were calculated by using an uncorrelated relaxed 
molecular clock approach (Drummond et al. 2006). The 
results indicated a most probable origin of the tribe in the 
Late Eocene or Oligocene (36.2?25.8 Ma; Tremetsberger 
et al., unpub. data) in North Africa. 
Pelser and Watson (Chapter 33) discussed the age of 
the subfamily Asteroideae on the basis of age estimates 
in the recent literature. Hershkovitz et al. (2006) es- 
timated the crown age of Asteroideae to be ca. 29?30 
Myr. Kim et al. (2005) used nonparametric rate smooth- 
ing in their molecular dating study of ndhF data and 
Cornus as an internal calibration point and arrived at an 
estimate for the subfamily of 26?29 Myr. Their age esti- 
mate for Asteroideae derived from average synonymous 
nucleotide substitutions using the same dataset and sub- 
stitution rates for Poaceae and Oleaceae was 35?39 Myr 
(Kim et al. 2005). These studies and unpublished data for 
Senecioneae (Pelser et al., in prep.) further indicate that 
the Heliantheae Alliance and all Asteroideae tribes out- 
side of it are 17 Myr old or older and were the result of a 
family-wide, rapid Oligocene?Early Miocene diversifica- 
tion. These results are roughly in line with other molecu- 
lar dating studies in Compositae (e.g., Wikstrom et al. 
2001; Wagstaff et al. 2006) and with paleo-palynological 
data (e.g., Katinas et al. 2008), although the latter source 
of data generally results in somewhat lower age estimates 
for Asteraceae lineages. 
Most of the dates discussed above are displayed on 
Fig. 44.8 and if we eliminate some of the outliers, we 
find that all of the tribes are proposed to have, more 
or less, the same age, around 25?35 Myr, and the age 
of the family seems to be 41?50 Myr. Initiation of all 
of the known major radiations of Compositae 35?25 
Ma places their origins within the Oligocene, which is 
often considered an important time of transition, a link 
between "[the] archaic world of the tropical Eocene and 
the more modern ecosystems of the Miocene" (Scotese 
2008). It makes ecological sense that a rapid expansion of 
the number oftaxa in many groups of Compositae would 
have coincided with the regression of tropical broad-leaf 
forests to the equatorial belt and the expansion of open, 
drier areas. 
The comparatively recent origin and great diversity of 
Compositae are likely indicative of the ecological success 
and evolutionary lability of the family (as is evidenced 
by their diverse appearance in Figs. 44.9 and 44.10), 
especially in drier environments. Turner (1977) felt that 
the family's "rich secondary metabolite chemistry, often 
short life cycle, facultative pollination, and freedom from 
many co-evolutionary restraints may be responsible for 
this success." It seems likely that the high seed set, dis- 
persal ability, and ability to radiate into new habitats have 
helped as well. 
Barker et al. (2008) examined gene duplication and re- 
tention in Compositae and found that there were at least 
three ancient whole genome duplications in the family 
resulting from paleopolyploidization events: at the base 
of the family just prior to its radiation, and near the base 
of tribes Mutisieae and the large Heliantheae Alliance. As 
one explanation for Compositae's evolutionary success, 
they suggest that retention of the resulting duplicates of 
CYCLOIDEA genes, which code for transcription factors 
associated with floral symmetry and branching patterns, 
were likely significant in the evolution of Compositae, 
because Chapman et al. (2008) observed that some copies 
have experienced positive selection and that the expres- 
sion of CYC genes is subfunctionalized among the disk 
and ray florets of the composite inflorescence. Thus, an- 
cient polyploidization may be, in part, responsible for the 
evolutionary success of the family. 
CONCLUSIONS 
The Calyceraceae-Compositae clade (as we know it 
today) may have originated in southern South America 
ca. 50?41 Ma, and the diversification of the family started 
in the same area. The diversification of Calyceraceae was 
modest by comparison with that of Compositae, which 
have traveled the globe. In Compositae, following the 
southern South American radiation, there was an African 
explosion. Of the 1600?1700 genera in Compositae today, 
about two-thirds are in clades with the basal branches in 
Africa, many in southern Africa. In fact, with the excep- 
tion of the Mutisieae (sensu Cabrera) grade at the base 
and the highly nested Heliantheae Alliance, all of the 
major clades in the family appear to have an African 
origin or a major African presence near the base of their 
phylogenies. From this African origin came numerous 
movements into Asia, Eurasia, Europe, Australia, etc., 
many of which have spawned substantial radiations (e.g., 
Cardueae, Vernonieae, Anthemideae). The clade formed 
by the core Heliantheae Alliance has a North American 
(including NW Mexico) origin beginning by 35?22 Ma, 
which coincides with a land bridge connection from Asia. 
Previously (Funk et al. 2005) it was suggested that, be- 
cause the sister clade to the Heliantheae Alliance is found 
in tropical eastern Africa, the ancestor of the Heliantheae 
Alliance could have come over the land bridge from Asia 
into western North America and down into Mexico. 
However, the position of the Cuban Feddea at the base of 
the American clade of the Heliantheae Alliance does not 
reinforce a land-bridge hypothesis. Given the success of 
a diversity of young lineages in the Heliantheae Alliance 
and long-distance dispersal to remote oceanic islands and 
between continents (see Baldwin 2009, Chapter 41), the 
possibility of a direct Old World to New World dispersal 
Chapter 44: Compositae metatrees: the next generation 771 
Fig. 44.9. Members of Compositae, subfamilies Barnadesioideae through Cichorioideae. A Schlechtendalia luzulacjolia Less. 
(Barnadesieae from Uruguay: Maldonado, Piriapolis, Cerro San Antonio); B Mutisia clematis L. (Mutisieae from Colombia: 
Cundinamarca, Finca "El Cerro"); C Wunderlichia mirabilis Riedel (Wunderlichieae from Brazil: Minas Gerais, Cerra do 
Cipo; Roque 1622); D Centaurea stoebe L. (Cardueae from USA: Virginia, Shenandoah National Park); E Cichorium intybus L. 
(Cichorieae from Uruguay: Montevideo); F Didclta spinosa Ait. (Arctotideae from South Africa, Northern Cape: Funk and 
Kockcmocr 12641). [Photographs: A, B, D, E, J.M. Bonifacino; C, N. Roque; F, V.A. Funk] 
772 Funk et al. 
Fig. 44.10. Members of Compositae subfamily Asteroideae. A Scnccio ceratophylloides Griseb. (Senecioneae from Uruguay: 
Canelones); B Dimorphotheca sinuata DC. (Namibia); C Bcllis pcrctmis L. (Astereae from Argentina: Tierra del Fuego, close to 
Paso Garibaldi); D Stcnachaenium megapotamicum (Spreng.) Baker in Mart. (Inuleae from Uruguay: Maldonado, Sierra de las 
Animas); E Heiianthus annuus L. (Heliantheae s.str. from Uruguay: Rio Negro, close to Fray Bentos); F Gyptis pinnatifida Cass. 
(Eupatorieae from Uruguay: Rivera, Arroyo Lunarejo). [Photographs: B, C.A. Mannheimer; A, C?F, J.M. Bonifacino.] 
Chapter 44: Compositae metatrees: the next generation 773 
of the ancestor of the American clade of the Heliantheae 
Alliance must be taken seriously. 
What about Hecastodeis? This monotypic North Amer- 
ican genus from the mountains of Nevada and the Death 
Valley area sits between the southern South American 
basal radiation and the African diversification. In Funk 
et al. (2005) two possible scenarios were proposed (apart 
from errors and misidentifications). First, there could 
have been two events of long distance dispersal, one 
from South America to North America and one from 
North America to Africa. The second possibility is that 
Compositae moved into North America from South 
America, then over to Europe and down into northern 
Africa followed by extensive extinction in the north- 
ern hemisphere (Panero and Funk 2008). There are, no 
doubt, other explanations; however, we do not have suf- 
ficient data to favor one hypothesis over another. One 
key group, Gochnatieae, is located just below Hecastodeis 
on the metatree, and it is being studied at the species 
level using both molecular and morphological data in 
the hope of providing a better estimation of the early 
biogeographic history of Compositae (Sancho et al., pers. 
comm.). 
Prior to the development of molecular techniques, 
most workers in the family followed the traditional con- 
cept of the family laid down by Bentham and elaborated 
upon in Cronquist (1955, 1977). Cronquist had detailed 
ideas about the characteristics of ancestral Compositae. 
He believed that the tribe Heliantheae, and more spe- 
cifically core Heliantheae, were the cauldron out of 
which the rest of the tribes evolved. He thought that 
the ancestor might have been something like Viguiera 
(Heliantheae s.L), but he pointed out that it was still "not 
exact" because the genus has neutral ray florets and only 
two principal pappus members (Cronquist 1977). Several 
scientists disagreed with Cronquist. Skvarla (1977) and 
Jeffrey (1977) pointed out that the characters were not 
consistent with the position that Heliantheae s.L was the 
primitive group of the family. In publications outside the 
1977 Heywood et al. volumes, Carlquist (1966, 1976) and 
Robinson (1981) tried to add additional tribes and to point 
out that the proposed direction of evolution did not make 
sense. These synantherologists thought that the pollen, 
anatomy, and morphology of Mutisieae were more like 
that of the related families, and that Heliantheae and 
other tribes had derived characters. 
Acknowledging that extant lineages of Barnadesieae 
have been around for as long as the most highly nested 
branches of the family, it is wise to not put too much 
emphasis on the characteristics that are found in this ba- 
sally diverging group but rather on characteristics that are 
shared by all early diverging branches and the outgroups. 
Many characters of Barnadesieae and Mutisioideae are 
variable (e.g., corolla morphology), but a few common 
characteristics can probably be determined: the pol- 
len was probably psilate (Skvarla 1977; Zao et al. 2006; 
Blackmore et al., Chapter 7); the basal chromosome 
number for the closely related families is x = 9 and that 
number has been proposed for Compositae with x = 10 
as the apparent basal number for tribes of South African 
origin (Semple and Watanabe, Chapter 4); and second- 
ary chemical compounds have developed from a small 
number of relatively simple flavonoids, polyacetylenes, 
coumarins, and triterpenes to a large number of com- 
plex compounds from many different chemical classes 
(Calabria et al., Chapter 5). 
Bremer (1994) started the process of updating the char- 
acters attributed to a hypothetical ancestor and Lundberg 
(Chapter 11) has added to the list. Here we have refined 
some of the characters and added a few more. Here 
we offer a list of potential plesiomorphic characters for 
the extant members of Compositae (* indicates that the 
character defines a larger clade than the Calyceraceae + 
Compositae clade; bold indicates a potential character 
unique to the ancestral members of Compositae): 
O *Shrubs or subshrubs; *no internal secretory systems. 
O Inflorescence cyme-like. 
O *Leaves alternate and spirally inserted. 
O Heads indeterminate; few heads per plant, each 
with many flowers. 
O Involucral bracts in several series, imbricate 
without hyaline or scabrous margins. 
O Receptacle naked. 
O Florets perfect and fertile, arranged in a head; *parts 
in 5's; mostly one type of flower, some differentia- 
tion in floral morphology in peripheral florets 
possible but without true rays. 
O Corolla white or possibly pink, yellow or blue; 
*probably 5-lobed, lobes deeply divided and with 
much variation. 
O *Stamens alternate with the corolla lobes; anthers 
fully connate at the margins with the fila- 
ments free with upper part of filaments forming 
a filament collar; thecae spurred (calcerate) and 
possibly tailed (caudate); possibly without apical 
appendage; *dehisce by longitudinal slits; pollen kit 
present. 
O Pollen grains 3-celled, *pollen prolate and psilate. 
774 Funk et al. 
O Styles slender, shortly bifid, without hairs; solid band 
of stigmatic surface on inside of style branches; 
ovary consistently inferior with ovule in a basal 
position. 
O Pappus of capillary bristles. 
O Fruit an indehiscent achene; ribbed. 
O *Base chromosome number: x = 9. 
O Secondary chemistry simple and characterized by 
a small number of flavonoids, polyacetylenes, cou- 
marins, and triterpenes. 
O Southern South American in distribution; probably 
growing in open dry habitats. 
Finally, looking to the future, advances in genomics 
are changing the way we do research in systematics. 
Phylogenomics, the use of whole genomes for phylo- 
genetic studies, is already occurring in many plant and 
animal groups and at ever increasing speeds (see brief 
overview in Pennisi 2008) and will no doubt become the 
standard of the future in Compositae systematics as costs 
decrease and technology becomes more widely avail- 
able. Whole chloroplast genomes have already been se- 
quenced for many plant groups and used in phylogenetic 
studies, particularly for establishing the position of basal 
angiosperms (Goremykin et al. 2004; Soltis et al. 2004). 
Phylogenomic studies in Compositae lag considerably be- 
hind those of the Angiosperm Phylogeny Group (APG) 
and that of many animals groups as well. Although the 
genomes of a number of Cichorioideae taxa are currently 
under study (Rieseberg, pers. comm.) only two economi- 
cally important taxa, Helianthus annuus L. and Lactuca sa- 
tiva L., are the subject of a coordinated, large scale effort. 
The Compositae Genome Project (CGP), headquartered 
at the UC Davis Genome Center, has a wide range of 
objectives for its studies of lettuce and sunflower (and 
presumably others in the future). The goal as given on 
the home page (http://compgenomics.ucdavis.edu/index. 
php) is to "integrate information at the genetic, physi- 
ological and population/evolutionary levels for a broad 
range of genes involved in evolution of cultivated plants 
and weeds, evaluate the relative importance of changes 
in gene sequence versus gene expression in phenotypic 
evolution, determine the genotypic consequences of par- 
allel phenotypic evolution, and provide a basis for future 
functional analyses." For most systematists, however, the 
focus of whole genome sequencing will be on more ac- 
curately reconstructing the evolutionary history of a par- 
ticular group of plants, most of which are not cultivated 
and for which the vast funding required to map genes and 
determine their functions will likely never be available. 
As in all molecular studies, a cautionary note has been 
sounded relative to the resolving power of genomics for 
phylogenetic study (Soltis et al. 2004; Pennisi 2008). Data 
analysis of huge numbers of sequences is daunting and 
will probably still require collaboration with mathemati- 
cians and bioinformaticists. Another issue is lack of con- 
gruence, particularly with existing trees. Herve Philippe 
(University of Montreal; cited in Pennisi 2008) stresses 
that datasets will have to be reanalyzed with different 
methods in order to determine the best tree. The latter 
is not necessarily guaranteed by more data. Additionally, 
taxon sampling will remain an issue. Lots of informa- 
tion from only a few taxa does not guarantee a sound 
phylogeny no matter how cutting-edge the sequencing 
or the analyses. Still, we can expect that genomes will be 
increasingly common tools in future phylogenetic stud- 
ies. Hopefully, as the data accumulate there will be better 
resolution of taxonomic placements, particularly in the 
location of Senecioneae and at the base of the Compositae 
family tree where the position of some mutisioids and 
some enigmatic genera remain unclear. 
Acknowledgements. We appreciate the funding that was pro- 
vided by all the funding bodies that supported the research in the 
chapters of this book. The order of authorship on this paper is 
alphabetical by family name after the first author. 
Literature cited 
Anderberg, A.A.   1989. Phylogeny and reclassification of the 
tribe   Inuleae   (Asteraceae).   Canadian Journal  of Botany.   67: 
2277-2296. 
Anderberg, A. A.  1991. Taxonomy and phylogeny of the tribe 
Inuleae (Asteraceae). Plant Systematics and Evolution 176: 75? 
123. 
Anderberg, A.A. & Eldenas, P. 2007 [2006]. Tribe Inuleae. 
Pp. 374-391 in: Kadereit, J.W. & Jeffrey, C. (eds.), The Families 
and Genera of Vascular Plants, vol. 8, Flowering Plants. Eudicots. 
Asterales. Springer, Berlin. 
Anderberg,   A.A.,   Eldenas,   P.,   Bayer,   R.J.   &  England, 
M. 2005. Evolutionary relationships in the Asteraceae tribe 
Inuleae (inch Plucheeae) evidenced by DNA sequences of 
ndhF; with notes on the systematic positions of some aberrant 
genera. Organisms, Diversity and Evolution 5: 135?146. 
Baldwin, B.C., Wessa, B.L. & Panero, J.L. 2002. Nuclear 
rDNA evidence for major lineages of helenioid Heliantheae 
(Compositae). Systematic Botany 27: 161?198. 
Barker, M.S., Kane, N.C., Matvienko, M., Kozik, A., 
Michelmore,   R.W.,   Knapp,   S.J.   &   Rieseberg,   L.H. 
Chapter 44: Compositae metatrees: the next generation 775 
2008. Multiple paleopolyploidizations during the evolution 
of the Compositae reveal parallel patterns of duplicate gene 
retention after millions of years. Molecular Biology and Evolution 
25: 2445-2455. 
Bayer, R.J., Breitwieser, I., Ward, J.M. & Puttock, C.F. 
2007 [2006]. Gnaphalieae. Pp. 246-284 in: Kadereit, J.W. & 
Jeffrey, C. (eds.), The Families and Genera of Vascular Plants, vol. 
8, Flowering Plants. Eudicots. Asterales. Springer, Berlin. 
Bentham, G. 1873a. Compositae. Pp. 163?533 in: Bentham, 
G. & Hooker, J.D. (eds.), Genera Plantarum, vol. 2(1). Reeve, 
London. 
Bentham, G. 1873b. Notes on the classification, history and 
geographical distribution of Compositae. Journal of the Linnean 
Society, Botany 13: 335?577. 
Bininda-Emonds, OR.P.,Jones, K.E., Price, S.A., Grenyer, 
R., Cardillo, M., Habib, M., Purvis, A. & Gittleman, 
J.L. 2003. Supertrees are a necessary not-so-evil: a comment 
on Gatesy & al. Systematic Biology 52: 724?729. 
Blackmore, S., Van Campo, E. & Crane, PR. 1986. 
Lophate Compositae pollen from the Miocene and Pliocene 
of the Mediterranean region. Pollen et Spores 28: 391?402. 
Bremer, K. 1987. Tribal interrelationships of the Asteraceae. 
CWkfi'cs 3: 210-253. 
Bremer, K. 1992. Ancestral areas?a cladistic reinterpreta- 
tion of the center of origin concept. Systematic Biology 41: 
436-445. 
Bremer, K. 1994. Asteraceae: Cladistics & Classification. Timber 
Press, Portland. 
Bremer, K. 1996. Major clades and grades of the Asteraceae. Pp. 
1-7 in: Hind, D.J.N. & Beentje, HJ. (eds.), Proceedings of the 
International Compositae Conference, Kew, 1994, vol. 1., Compos- 
itae: Systematics. Royal Botanic Gardens, Kew. 
Bremer, K. & Gustafsson, M.H.G. 1997. East Gondwana an- 
cestry of the sunflower alliance of families. Proceedings of the 
National Academy of Sciences of the United States of America 94: 
9188-9190. 
Bremer, K. & Jansen, R.K. 1992. A new subfamily of the 
Asteraceae. Annals of the Missouri Botanical Garden 79: 414?415. 
Brooks, DR. 1982. Hennig's parasitological method: a pro- 
posed solution. Systematic Zoology 30: 229?249. 
Brooks, D.R. & McLennan, D.A. 2002. The Nature of 
Diversity: An Evolutionary Voyage of Discovery. University of 
Chicago Press, Chicago and London. 
Cabrera, AL. 1977. Mutisieae?systematic review. Pp. 1039? 
1066 in: Heywood, V.H., Harborne, J.B. & Turner, B.L. 
(eds.), The Biology and Chemistry of the Compositae, vol. 2. 
Academic Press, London. 
Cariaga, K.A., Pruski, J.F., Oviedo, R., Anderberg, A.A., 
Lewis, C.E. & Francisco-Ortega, J. 2008. Phylogeny and 
systematic positon of Feddea (Asteraceae: Feddeeae): a taxo- 
nomically enigmatic and critically endangered genus endemic 
to Cuba. Systematic Botany 33: 193?202. 
Carlquist, S. 1966. Wood anatomy of Compositae. A sum- 
mary, with comments on factors controlling wood evolution. 
/4ko 6: 25-44. 
Carlquist, S. 1976. Tribal interrelationships and phylogeny of 
the Asteraceae. Aliso 8: 465-492. 
Cassini, H. 1816. Troisieme memoire sur les Synantherees. 
Journal de Physique, de Chimie, d'Histoire Naturelle et des Arts 
82: 116-146. 
Chapman, M.A., Leebens-Mack, J. & Burke, J. 2008. 
Positive selection and expression divergence following gene 
duplication in the sunflower CYCLOIDEA gene family. 
Molecular Biology and Evolution 25: 1260?1273. 
Cox, C.B. & Moore, P.O. 2004. Biogeography. An Ecological and 
Evolutionary Approach. Blackwell, Oxford. 
Cronquist, A. 1955. Phylogeny and taxonomy of the Compositae. 
American Midland Naturalist 53: 478?511. 
Cronquist, A. 1977. The Compositae revisited. Brittonia 29: 
137-153. 
DeVore, M.L. & Stuessy, T.F. 1995. The place and time 
of origin of the Asteraceae, with additional comments on 
the Calyceraceae and Goodeniaceae. Pp. 23?40 in: Hind, 
D.J.N., Jeffrey, C. & Pope, G.V. (eds.), Advances in Compositae 
Systematics. Royal Botanic Gardens, Kew. 
Drummond, A.J., Ho, S.Y.W., Phillips, M.J. & Rambaut, 
A. 2006. Relaxed phylogenetics and dating with confidence. 
PLoS Biology 4(5): e88 doi:10.1371/journal.pbio.0040088. 
Eldenas, P., Kallersjo, M. & Anderberg, A.A. 1999. Phylo- 
genetic placement and circumscription of tribes Inuleae s.str. 
and Plucheeae (Asteraceae): evidence from sequences of chlo- 
roplast gene ndhF. Molecular Phylogenetics and Evolution 13: 
50-58. 
Eriksson, T. 1991. The systematic position of the Blepharispermum 
group (Asteraceae, Heliantheae). Taxon 40: 33?39. 
Farris, J.S. 1970. Methods of computing Wagner trees. Systematic 
Zoo/ogy 19: 83-92. 
Funk, V.A., Bayer, R.J., Keeley, S., Chan, R., Watson, L., 
Gemeinholzer, B., Schilling, E., Panero, J.L., Baldwin, 
B.G., Garcia-Jacas, N., Susanna, A. & Jansen, R.K. 
2005. Everywhere but Antarctica: using a supertree to un- 
derstand the diversity and distribution of the Compositae. 
Biologiske Skrifter 55: 343-374. 
Funk, V.A. & Chan, R. 2008. Phylogeny of the spiny African 
daisies (Compositae, tribe Arctotideae, subtribe Gorteriinae) 
based on trnL-F, ndhF, and ITS sequence data. Molecular 
Phylogenetics and Evolution 48: 47?60. 
Funk, V.A., Chan, R. & Keeley, S.C. 2004. Insights into 
the evolution of the tribe Arctoteae (Compositae: subfam- 
ily Cichorioideae s.s.) using trnL-F, ndhF, and ITS. Taxon 53: 
637-655. 
Funk, V.A., Robinson, H. & Dillon, MO. 1996. Liabeae: 
taxonomy, phylogeny and biogeography. Pp. 545?567 
in: Hind, D.J.N. & Beentje, HJ. (eds.), Proceedings of the 
International Compositae Conference, Kew, 1994, vol. 1, Compos- 
itae: Systematics. Royal Botanic Gardens, Kew. 
Funk, V.A. & Specht, C. 2007. Meta-trees: grafting for a global 
perspective. Proceedings of the Biological Society of Washington 
120: 232-240. 
Gatesy, J., Matthee, C, DeSalle, R. & Hayashi, C. 2002. 
Resolution of a supertree/supermatrix paradox. Systematic 
Biology 51: 652-664. 
Gemeinholzer, B. & Bachmann, K. 2003. Reconstruction of 
the phylogeny of the Lactuceae (Asteraceae) using the Internal 
Transcribed Spacer regions ITS 1+2. Compositae Newsletter 40: 
15-16. 
Germeraad, J.H., Hopping, C.A. & Muller, J. 1968. 
Palynology of Tertiary sediments from tropical areas. Review 
of Palaeoboany and Palynology 6: 189?348. 
Ghiglione, M.C., Yagupsky, D., Ghidella, M. & Ramos, 
V.A. 2008. Continental stretching preceding the opening of 
the Drake Passage: evidence from Tierra del Fuego. Geology 
36: 643-646. 
Gibbs, A.K. & Barron, C.N. 1993. The Geology of the Guiana 
Shield. Oxford University Press, New York. 
Goremykin, V.V., Hirsch-Ernst, K.I., Wolfl, S. & Hell- 
wig, F.H. 2004. The chloroplast genome of Nymphaea alba: 
whole-genome analyses and the problem of identifying the 
776 Funk et al. 
most basal angiosperm. Molecular Biology and Evolution 21: 
1445-1454. 
Graham, A. 1996. A contribution to the geologic history of the 
Compositae. Pp. 123-140 in: Hind, D.J.N. & Beentje, HJ. 
(eds.), Proceedings of the International Compositae Conference, Kew, 
1994, vol. 1, Compositae: Systematics. Royal Botanic Gardens, 
Kew. 
Gruenstaeudl, M., Urtubey, E., Jansen, R.K., Samuel, 
R., Barfuss, M.H.J. & Stuessy, T.F. 2009. Phylogeny of 
Barnadesioideae (Asteraceae) inferred from DNA sequence 
data and morphology. Molecular Phylogenetics and Evolution 51: 
572-587. 
Gustafsson, M.H.G., Backlund, A. & Bremer, B. 1996. 
Phylogeny of the Asterales sensu lato based on rbcE sequences 
with particular reference to the Goodeniaceae. Plant Systematics 
<W EWwk'oM 199: 217-242. 
Gustafsson, M.H.G., Grafstrom, E. & Nilsson, S. 1997. 
Pollen morphology of the Goodeniaceae and comparisons 
with related families. Grana 36: 185?207. 
Gustafsson, M.H.G., Pepper, A.S.-R., Albert, V.A. & 
Kallersjo, M. 2001. MolecularphylogenyoffheBarnadesioideae 
(Asteraceae). Nordic Journal of Botany 21: 149?160. 
Hansen, H.V. 1991a. Phylogenetic studies in Compositae tribe 
Mutisieae. Opera Botanica 109: 1?50. 
Hansen, H.V. 1991b. SEM-studies and general comments on 
pollen in tribe Mutisieae (Compositae) sensu Cabrera. Nordic 
Journal of Botany 10: 607?623. 
Hershkovitz, M.A., Arroyo, M.T.K., Bell, C. & Hinojosa, 
L.F. 2006. Phylogeny of Chaetanthera (Asteraceae; Mutisieae) 
reveals both ancient and recent origins of the high elevation 
lineages. Molecular Phylogenetics and Evolution 41: 594?605. 
Hey wood, V.H. (ed.). 1993. Flowering Plants of the World. 
Oxford University Press, New York. 
Hey wood, V.H., Harborne, J.B. & Turner, B.L. (eds.). 
1977. The Biology and Chemistry of the Compositae, 2 vols. 
Academic Press, London. 
Himmelreich, S., Kallersjo, M., Eldenas, P. & Oberprieler, 
C. 2008. Phylogeny of southern hemisphere Compositae- 
Anthemideae based on nrDNA ITS and cpDNA ndhF se- 
quence information. Plant Systematics and Evolution 272: 
131-153. 
Hind, D.J.N. (ed.). 1996. Proceedings of the International Compositae 
Conference, Kew, 1994, 2 vols. Royal Botanic Gardens, Kew. 
Hochuli, PA. 1978. Palynologische Untersuchungen im Oli- 
gozan und Untermiozan der zentralen und westlichen Para- 
tethys. Beitrdge zur Paldontologie von Osterreich 4: 1?132. 
Howarth, D.G., Gustafsson, M.H.G., Baum, D. & Motley, 
T.J. 2003. Phylogenetics of the genus Scaevola (Goodeniaceae): 
implications for dispersal patterns across the Pacific basin 
and colonization of the Hawaiian Islands. American Journal 
of Botany 90: 915?923 and Supplementary Data from the 
American Journal of Botany website: Pp. 1?3. 
Hu, S.Y. 1958. Statistics of Compositae in relation to the flora of 
China. Journal of the Arnold Arboretum 39: 347?419. 
Jansen, R.K. & Kim, K.-J. 1996. Implications of chloro- 
plast DNA data for the classification and phylogeny of the 
Asteraceae. Pp. 317-339 in: Hind, D.J.N. & Beentje, HJ. 
(eds.), Proceedings of the International Compositae Conference, Kew, 
1994, vol. 1, Compositae: Systematics. Royal Botanic Gardens, 
Kew. 
Jansen, R.K., Michaels, HJ. & Palmer, J.D. 1991a. Phylog- 
eny and character evolution in the Asteraceae based on chlo- 
roplast DNA restriction site mapping. Systematic Botany 16: 
98-115. 
Jansen, R.K., Michaels, H.J., Wallace, R., Kim, K.-J., 
Keeley, S.C., Watson, L.E. & Palmer, J.D. 1991b. Chlo- 
roplast DNA variation in the Asteraceae: phylogenetic and 
evolutionary implications. Pp. 252?279 in: Soltis, D.E., 
Soltis, P.S. & Doyle, JJ. (eds.), Molecular Systematics of Plants. 
Chapman & Hall, New York. 
Jansen, R.K. & Palmer, J.D. 1987. A chloroplast DNA inver- 
sion marks an ancient evolutionary split in the sunflower fam- 
ily (Asteraceae). Proceedings of the National Academy of Sciences of 
(Ae UmW SM? of/lmen'fa 84: 5818-5822. 
Jansen, R.K. & Palmer, J.D. 1988. Phylogenetic implications 
of chloroplast DNA restriction site variation in the Mutisieae 
(Asteraceae). American Journal of Botany 75: 753?766. 
Jeffrey, C. 1977. Corolla forms in Compositae?some evolution- 
ary and taxonomic speculations. Pp. 111?118 in: Heywood, 
V.H., Harborne, J.B. & Turner, B.L. (eds.), The Biology and 
Chemistry of the Compositae, vol. 1. Academic Press, London. 
Jones, S. 1977. Vernonieae. Pp. 503?521 in: Heywood, V.H., 
Harborne, J.B. & Turner, B.L. (eds.), The Biology and Chemistry 
of the Compositae, vol. 1. Academic Press, London. 
Kadereit, J.W. & Jeffrey, C. (eds.). 2007 [2006]. The Families 
and Genera of Vascular Plants, vol. 8, Flowering Plants. Eudicots. 
Asterales. Springer, Berlin. 
Katinas, L., Crisci,J.V.,Jabaily, R.S., Williams, C., Walker, 
J., Drew, B., Bonifacino, J.M. & Sytsma, K. 2008. 
Evolution of secondary heads in Nassauviinae (Asteraceae, 
Mutisieae). American Journal of Botany 95: 229?240. 
Katinas, L., Crisci, J.V., Telleria, M.C., Barreda, V. & 
Palazzesi, L. 2007. Early history of Asteraceae in Patagonia: 
evidence from fossil pollen grains. New Zealand Journal of 
Botany 45: 605-610. 
Karis, P.O., Eldenas, P. & Kallersjo, M. 2001. New evi- 
dence for the systematic position of Gundelia L. with notes on 
delimitation of Arctoteae (Asteraceae). Taxon 50: 105?114. 
Keeley, S.C., Forsman, A.H. & Chan, R. 2007. A phylog- 
eny of the "evil tribe" (Vernonieae: Compositae) reveals Old/ 
New World long distance dispersal: support from separate and 
combined congruent datasets (trnE-F, ndhF, ITS). Molecular 
Phylogenetics and Evolution 44: 89?103. 
Kim, H.-G., Loockerman, D.J. & Jansen, R.K. 2002. 
Systematic implications of ndhF sequence variation in the 
Mutisieae (Asteraceae). Systematic Botany 27: 598?609. 
Kim, K.-J., Choi, K.-S. & Jansen, R.K. 2005. Two chloro- 
plast DNA inversions originated simultaneously during the 
early evolution of the sunflower family (Asteraceae). Molecular 
Biology and Evolution 22: 1783?1792. 
Kim, K.-J. & Jansen, R.K. 1995. ndhF sequence evolution 
and the major clades in the sunflower family. Proceedings of the 
National Academy of Sciences of the United States of America 92: 
10379-10383. 
Maddison, DR. & Maddison, W.P. 2001. MacClade, vers. 
4.03. Sinauer, Sunderland. 
McKenzie, R.J. & Barker, N.P. 2008. Radiation of south- 
ern African daisies: biogeographic inferences for subtribe 
Arctotidinae (Asteraceae: Arctotideae). Molecular Phylogenetics 
and Evolution 49: 1?16. 
Muller, J. 1970. Palynological evidence on early differentiation 
of angiosperms. Biological Review 45: 417?450. 
Muller, J. 1981. Fossil pollen records of extant angiosperms. 
Botanical Review 47: 1?142. 
Nordenstam, B. 2007a [2006]. Senecioneae. Pp. 208-241 in: 
Kadereit, J.W. & Jeffrey, C. (eds.), The Families and Genera 
of Vascular Plants, vol. 8, Flowering Plants. Eudicots. Asterales. 
Springer, Berlin. 
Chapter 44: Compositae metatrees: the next generation 777 
Nordenstam, B. 2007b [2006]. Calenduleae. Pp. 241-245 in: 
Kadereit, J.W. & Jeffrey, C. (eds.), The Families and Genera 
of Vascular Plants, vol. 8, Flowering Plants. Eudicots. Asterales. 
Springer, Berlin. 
Nordenstam, B. 2007c [2006]. Corymbieae. Pp. 207-208 in: 
Kadereit, J.W. & Jeffrey, C. (eds.), The Families and Genera 
of Vascular Plants, vol. 8, Flowering Plants. Eudicots. Asterales. 
Springer, Berlin. 
Oberprieler, C, Himmelreich, S. & Vogt, R. 2007. A new 
subtribal classification of the tribe Anthemideae (Compositae). 
WiYWeMowi'a 37: 89-114. 
Panero, J.L. & Funk, V.A. 2002. Toward a phylogenetic 
subfamilial classification for the Compositae (Asteraceae). 
Proceedings of the Biological Society of Washington 115: 909?922. 
Panero, J.L. & Funk, V.A. 2008. The value of sampling anoma- 
lous taxa in phylogenetic studies: major clades of the Asteraceae 
revealed. Molecular Phylogenetics and Evolution 47: 757?782. 
Pelser, P.B., Nordenstam, B., Kadereit, J.W. & Watson, 
L.E. 2007. An ITS phylogeny of tribe Senecioneae (Asteraceae) 
and a new delimitation of Senecio L. Taxon 56: 1077?1104. 
Pennisi, E. 2008. Building the tree of life, genome by genome. 
Sconce 320: 1716-1717. 
Raven, P.H. & Axelrod, D.I. 1974. Angiosperm biogeogra- 
phy and past continental movements. Annals of the Missouri 
Botanical Garden 61: 539?673. 
Robinson, H. 1981. A revision of the tribal and subtribal lim- 
its of the Heliantheae (Asteraceae). Smithsonian Contributions to 
Botany 51: 1-102. 
Robinson, H. 1983. A generic review of the tribe Liabeae 
(Asteraceae). Smithsonian Contributions to Botany 54: 1?69. 
Robinson, H. 1999. Generic and subtribal classification of 
American Vernonieae. Smithsonian Contributions to Botany 89: 
1-116. 
Rzedowski, J. 1972. Contribuciones a la fitogeografia floris- 
tica e historica de Mexico. Ill Algunas tendencias en la dis- 
tribucion geografica y ecologica de las Compositae mexica- 
nas. Ciencia (Mexico) 27: 123?132. 
Scotese, C. R. 2008. PALEOMAP website (http://www.sco 
tese.com). 
Skvarla, J.J. 1977. Pollen morphology. Pp. 141-248 in: Hey wood, 
V.H., Harborne, J.B. & Turner, B.L. (eds.), The Biology and 
Chemistry of the Compositae, vol. 1. Academic Press, London. 
Soltis, D.E., Albert, V.A., Salvolainen, V., Hilu, K., Qiu, 
Y.-L., Chase, M.W., Farris, J.S., Stfanovic, S., Rice, 
D.W., Palmer, J.D. & Soltis, P.S. 2004. Genome-scale 
data, angiosperm relationships, and 'ending incongruence': 
a cautionary tale in phylogenetics. Trends in Plant Science 19: 
477-483. 
Small, J. 1919. The Origin and Development of the Compositae. 
New Phytologist, Reprint No. 11. Wesley & Son, London. 
Steel, M., Dress, A.W.M. & Bocker, S. 2000. Simple but 
fundamental limitations on supertree and consensus tree 
methods. Systematic Biology 49: 363?368. 
Susanna, A., Garcia-Jacas, N., Hidalgo, O., Vilatersana, 
R. & Garnatje, T. 2006. The Cardueae (Compositae) revis- 
ited: insights from ITS, trnL-trnF, and mafiC nuclear and chlo- 
roplast DNA analysis. Annals of the Missouri Botanical Garden 
93: 150-171. 
Swofford, D.L. 2002. PAUP*, Phylogenetic Analysis Using Parsi- 
mony (*and Other Methods), vers. 4.0M0. Sinauer, Sunderland. 
Turner, B.L. 1977. Fossil history and geography. Pp. 21?39 
in: Heywood, V.H., Harborne, J.B. & Turner, B.L. (eds.), 
The Biology and Chemistry of the Compositae, vol. 1. Academic 
Press, London. 
Wagstaff, S.J., Breitwieser, I. & Swenson, U. 2006. Origin 
and relationships of the austral genus Abrotanella (Asteraceae) 
inferred from DNA sequences. Taxon 55: 95?106. 
Wang, Y.-J., Liu, J.-Q. & Miehe, G. 2007. Phylogenetic or- 
igins of the Himalayan endemic Dolomiaea, Diplazoptilon and 
Xanthopappus (Asteraceae: Cardueae) based on three DNA re- 
gions. Annals of Botany (London) 99: 311?322. 
Watson, L.E., Evans, T.M. & Boluarte, T. 2000. Molecular 
phylogeny and biogeography of tribe Anthemideae (Aster- 
aceae), based on chloroplast gene ndhF. Molecular Phylogenetics 
and Evolution 15: 59?69. 
Wikstrom, N., Savolainen, V. & Chase, M.W. 2001. Evo- 
lution of the angiosperms: calibrating the family tree. Proceed- 
ings of the Royal Society of London, Series B, Biological Sciences 
268:2211-2220. 
Zao, Z., Skvarla, J.J. & Jansen, R.K. 2006. Mutisieae (Aster- 
aceae) pollen ultrastructure atlas. Lundellia 9: 51?76. 
Zavada, M.S. & De Villiers, S.E. 2000. Pollen of the Aster- 
aceae from the Paleocene-Eocene of South Africa. Grana 39: 
39-45.