THE AFFINITIES AND DISTRIBUTION OF THE LOWER

EOCENE FLORA OF SOUTHEASTERN

NORTH AMERICA

By EDWARD WILBER BERRY

Reprinted from

Proceedings of the American Philosophical Society

Vol. LIII, No. 214, June-July, 1914.

THE AFFINITIES AND DISTRIBUTION OF THE LOWER

EOCENE FLORA OF SOUTHEASTERN

NORTH AMERICA.

By EDWARD WILBER BERRY.
(Read April 25, 19 14.)

Introduction.

Three years ago I made a preliminary announcement before this
Society 1 concerning the fossil floras of southeastern North America.
I have, in the interim, completed a monograph of the extensive and
especially well preserved plants of the Lower Eocene, and it is some
of the results of this detailed study that are given in the present
communication. This work has been done under the auspices of the
United States Geological Survey, to the director of which organiza-
tion I am indebted for permission to publish the following prelim-
inary abstract. I also wish to express my great indebtedness to Dr.
T. Wayland Vaughan, who has had general charge of the Coastal
Plain investigations and to whom great credit is due for their com-
prehensive character.

Physical Conditions Indicated by the Flora.
There is no part of North America so favorably situated for the
study of the floras which preceded the present, extending backward
1 Proc. Amer. Philos. Soc, Vol. 50, No. 199, 1911.

Reprinted from Proceedings American Philosophical Society, Vol. liii., iqia

130 BERRY— LOWER EOCENE FLORA OF [April 25,

to a time which marks the first recorded appearance of angiosperms,
as that of the South Atlantic and Gulf states. Xo single part of
Ninth America contains so continuous a series of Tertiary deposits
carrying fossil plants. In this area are found abundant floras in the
lower and middle stages of the Eocene, a small flora in the Upper
Eocene, considerable floras in the Oligocene, some in the later Mio-
cene, and rather abundant fossil plants in the Pliocene, as well as
numerous Pleistocene deposits carrying fossil plants. The Rocky
Mountain region is rich in Eocene fossil plants and there are some
Miocene floras, but practically no Oligocene or Pliocene floras are
known. The Pacific coast region likewise furnishes Eocene and
Miocene fossil plants but none of Oligocene age.

The fossil floras of the Coastal Plain are found in an area where
it is possible to attain to some measure of accuracy in predicating the
general character and course of ocean currents and w r inds and other
physical features of the environment. On the other hand the west-
ern floras just mentioned grew in areas where vulcanism was great at
times ; in areas of great orogenic activity, where changes in topog-
raphy were numerous and elevations of several thousands of feet are
recorded ; areas in which climatic conditions not only varied from
place to place, but passed through a large cycle of secular changes.
All these factors greatly complicate the floral history.

The floras of the southern Coastal Plain are moreover checked for
the most part by very abundant marine fauna in intercalated beds, or
the plant-bearing beds which represent the coastal swamps and the
shallow water deposition of the old embayment merge laterally with the
contemporaneous limestones or marls which were forming in more
open waters along the coasts to the southward, so that there is a con-
siderable body of facts bearing on depth, character of the bottom, and
marine temperatures, with which to compare land temperatures.
These criteria have been admirably worked out for the Florida area
by Doctors Dall and Vaughan for the post-Eocene and their results
furnished a reliable datum plan for the deductions to be derived from
the study of the fossil floras of those times.

With the exception of fragments of the petrified stems of coni-
fers, palms and dicotyledons the plant-remains are in the form of
impressions, mostly of foliage, but with a goodly representation

I9I4-] SOUTHEASTERN NORTH AMERICA. 131

of fruits and seeds, and in some few cases even flowers are pre-
served.

While the oscillations of the embayment area have been numerous
they have been, as I have just mentioned, inconsiderable in amount,
only a few hundred feet at most, and the coastal region has uniformly
been one of slight relief. The various floras show an almost com-
plete absence of upland types. This is in striking contrast to the
European older Tertiary floras. The only large area of the globe
which has been thoroughly studied— Europe — was far less stable
than this region in Tertiary times and lying much farther toward the
pole was subsequently subjected to the rigors of Pleistocene condi-
tions whose influence never reached our southern states.

The paleobotanical record of the Atlantic and Gulf Coastal Plain
furnishes a history which extends back as I have just mentioned be-
yond the oldest known angiosperms to a time (Lower Cretaceous)
when the flora was made up almost entirely of tree-ferns, conifers
and those interesting cycadophytes (Cycadeoidea) whose trunks are
sometimes preserved with such marvelous perfection that the outlines
of the embryos in the ovules can often be made out in detail. Coming
a step nearer my present theme, a step of some millions of years
from the Lower into the Upper Cretaceous, wt find the first great
modernization of the floras of the world due to the seemingly sudden
evolution of the main types of angiosperms. These Upper Creta-
ceous floras are well represented in the Coastal Plain from Marthas
Vineyard to Texas. They extend northward to Greenland and south-
ward to Argentina in South America, and are found to indicate very
different physical conditions from those which prevail at the present
time. I do not intend, however, to dwell upon the Upper Cretaceous
floras in this connection but pass to a consideration of the succeeding
Eocene epoch of plant evolution.

The Eocene as defined by Lyell was marked by the dawn of the
recent species of marine mollusca. It is equally well marked by the
sudden expansion and evolution of modern types of plants after a
long antecedent Cretaceous development. The floras become thor-
oughly modernized as compared with those which preceded them,
although they are still very different in their general facies and dis-
tribution from those of the present.

132

BERRY— LOWER EOCENE FLORA OF

[April 25,

In the earliest epoch of the Eocene known as the Midway, the re-
lations of sea and land in the Gulf area differed in only minor par-
ticulars from that of the late Cretaceous. The waters of the Missis-
sippi Gulf were, however, deeper. This factor combined with a
much less influx of fresh water from the tributary streams, due in
some measure to the low relief of the land, enabled marine faunas to
reach well toward the head of the gulf. These faunas indicate sub-
tropical bottom temperatures northward as far as Paducah, Ky. The

Fig. 1. Sketch map showing the approximate position of the shore line,
A-A. at the beginning of the Wilcox transgression and B-B the area covered
by the Wilcox sea during its maximum transgression. C-C, the extreme
northern limit of the Willcox flora under existing climatic conditions.

known floras are very scanty and unsatisfactory and in the present
state of our knowledge do not merit an extended discussion. The
maximum transgression of the sea during the Midway epoch is
shown on the accompanying sketch map (Fig. 1).

The Midway Eocene was succeeded by a long interval during
which the sea is believed to have withdrawn southward at least as
far as the position indicated on the accompanying sketch map (Fig.
1. ./-. / ). since terrestrial conditions are known at the extreme base

i 9 i4.] SOUTHEASTERN NORTH AMERICA. . 133

of the Wilcox in the most southerly areas of their outcrop. This
interval of emergence of the embayment area was followed by an
equally long interval during which a great thickness of deposits was
laid down that are collectively known as the Wilcox group. The
character of these sediments and their faunas show that the Missis-
sippi gulf was somewhat restricted and much shallower than in the
preceding epoch, with true marine conditions prevalent only in its
lower portion. The shores were low and relatively flat. They were
flanked by current- or wave-built bars and separated from the main-
land by shallow inlets or lagoons. The lower courses of the streams
were transformed into shallow estuaries or broad swamps through
which the smaller streams meandered.

The maximum area covered or underlain by, Wilcox deposits is
also indicated on the accompanying sketch map (Fig. i, B-B) which
shows approximately the shore line along which the vegetation
migrated. As has been already remarked the Wilcox deposits have
yielded one of the most extensive of known fossil floras, an assem-
blage of extinct species which sheds considerable light on the phys-
ical conditions of the marginal lands of Wilcox time.

Before taking up in detail the evidence of the flora I wish to point
out certain general climatic conditions based on cosmic causes and de-
duced for the Wilcox from studies of recent climates.

It is to be noted that the factors governing atmospheric circulation
are general and not local and the relatively slight changes in the rela-
tion of land to sea in Wilcox time as compared with the present are
entirely too small to have caused much modification of existing con-
ditions. Then as now there was a persistent area of high pressure
over the North Atlantic and a low over the continent. Consequently
the winds were prevailingly from the east. Cyclonic disturbances
like those that originate today in the Gulf of Mexico or those more
violent and widespread storms of the West Indian hurricane type
which today originate in the Caribbean Sea would traverse at least a
part of the Mississippi embayment. So large an area of shallow
more or less landlocked water would have a very appreciable effect
in raising total temperatures and in the prevention of widely sepa-
rated extremes. At the same time it would increase the rainfall and
increase the width of the marginal lands over which this augmented

134 BERRY— LOWER EOCENE FLORA OF [April 25,

rainfall would be effective. Whether or not this would be sufficient
to furnish the subtropical conditions that the flora seems to indicate
is doubtful. Speculation regarding the Eocene climate of the world
as a whole is perhaps out of place, nevertheless it remains true that
the sum total of paleontologic evidence indicates that the familiar
succession of seasons or of types of vegetation in passing from the
luxuriant tropics to the ice-capped poles did not hold good for the
Eocene. Paleobotanists have long maintained that the existing cli-
mate is essentially a Pleistocene climate of an interglacial character
and that for the great bulk of geologic time uniformity and not dif-
ferentiation has been the rule rather than the exception. While the
older paleobotanists were inclined to overestimate the conditions
of torridity, it remains true that from the Lower Cretaceous until
toward the close of the Oligocene, not to mention still older floras of
more remote botanical affinities, whenever fossil floras are found,
from beneath the Equator to within the Arctic circle, they show a
degree of uniformity that proves that former climates were secularly
unlike these of today and as is obvious this floral evidence would be
equally convincing if all the vast number of fossil plants were simply
called Phyllites as in Schlotheim's day and no attempt were made to
determine their botanical affinity.

The student of fossil floras is naturally more sanguine and en-
thusiastic in predicting former physical conditions than perhaps is
warranted by his facts. When however a common Upper Creta-
ceous flora can be traced from Texas to Greenland or when we find
in the Eocene such unmistakable forms as Artocarpus leaves, Engel-
hardtia fruits, and nuts of the Nipa palm associated with forms as
characteristic as ferns of the genus Acrostichum all extending almost
across the temperate zone in both the eastern and western hemi-
spheres it would seem that the burden of proof that climates were
not very different from those of today rests with the physicist and
not with the paleobotanist.

It may be noted that all of the Wilcox plants, almost without ex-
ception, are plants whose modern representatives inhabit the warmer
parts of the earth. There is not a single strictly temperate type in
the whole assemblage, the nearest approach to such types being in
the genera Juglans, Myrica, Magnolia, Cercis, Ilex, Nyssa, ami Frax-

I9I4-] SOUTHEASTERN NORTH AMERICA. 135

inus, and in all of these genera or closely related ones there are ex-
isting tropical forms. None extend beyond the warmer parts of
the temperate zone and some of these as in the case of Juglans and
Fraxinus indicate in their compound leaves their tropical ancestry, as
was first pointed out by Grisebach. The ferns are all tropical types
and their relative unimportance in the Wilcox flora furthermore in-
dicates that the major part of this flora is a strand flora. This is
shown more especially by forms like the Nipa palm which never
grows outside of tidal marshes, by Conocarpus, Laguncularia and
Avicennia which inhabit like situations; by coastal marsh or lagoon
plants like Canna, Trapa and Sabalites; and by the large number of
strand types that inhabit beaches or the jungle behind the beach
ridges or dunes. The more striking of these genera are Myrica,
Artocarpus, Ficus, Coccolobis, Pisonia, Anona, Capparis, Chryso-
balanus, various Lauraceae, Apocynacese, Sapotaceae and Legumino-
sse, Fagara, Dry petes, Metopium, Ilex, Celastrns, Sapindus, Dodoncea,
Reynosia, Rhamnus, Myrcia, Eugenia, Laguncularia, Combretum,
Tcrminalia, Cordia, Citharexylon, Exostema and Guettarda.

It needs but a slight acquaintance with the existing Antillean
flora or that of the Florida keys, or in lieu of actual acquaintance a
perusal of the as yet too few ecological discussions of the flora of
the American tropics or even of Schimper's classic Indo- Malayan
strand flora to see at once that the general facies of the Wilcox flora
is overwhelmingly that of a strand flora of which some of the
elements indicate that they grew on the sandy beaches, others in
muddy tidal flats, others between or behind dunes or beach ridges,
and others in estuary bayous or marshes. None of the forms can
certainly be considered as inland or upland types. Even genera like
Banksia which is not usually considered a coastal type in the existing
flora furnishes Banksia marginata Cav. to the coastal sand dunes of
South Australia (Tepper) and several others species of the genus
occur on the dunes of Queensland, Victoria and western Australia.

Little has been written of the plant associations of the American
tropics and collectors notes almost always fail to adequately describe
habitats. While the marginal Wilcox lands were low there was such
a large area of continent to the northward to draw from, and the
long coast furnishes such varied edaphic conditions, that the flora

136 BERRY— LOWER EOCENE FLORA OF [April 25,

was far richer than floras of small insular areas of the American
tropics of the present, as for example, that of the Bahamas which are
relatively close to mainland, where in addition to difficulties of intro-
duction there is relatively great uniformity of edaphic factors and
directly ail verse factors such as winds, which limit the floral display.

Without pursuing the subject in greater detail it may be assumed
to be proven that the Wilcox flora is a typical coastal flora. Com-
pared with recent coastal floras it is at once apparent that its affin-
ities are entirely with those of tropical and subtropical America. It
has much in common with the Bahaman flora and that of the Florida
keys, but is far richer in arborescent forms. Comparisons with the
larger islands of the West Indies show more elements in common,
such differences as are apparent being due to the prevalence of por-
ous coral rock along these recent shores while the Wilcox shores were
not of this character. The most complete agreement is furnished by
the floras along the Caribbean coast from Central America to north-
ern Brazil. A considerable number of genera found in the Wilcox
flora do not range through the West Indies at the present time and
the explanation seems to be that the Wilcox flora more closely re-
sembles the original flora of the whole American equatorial region
which became restricted during the epeirogenetic and climatic changes
of the Miocene or Pleistocene and the elements now lacking in the
West Indies never regained all of the area of distribution lost at that
time.

It may seem improper to say that a flora with abundant forms
of Artocarpus, Nipa, Cinnamomum, Banksia, etc., is entirely Amer-
ican in character but if the brief sketches in the botanical discussion
which follows are read it will be obvious that these genera, oriental
in the existing flora, were cosmopolitan in the early Tertiary, so that
it would be entirely misleading to draw conclusions from existing dis-
tribution alone.

The Wilcox waters of the upper embayment were always shal-
low ; there were fringing bars and lagoons as well as deltas, estuaries
and swampy bayous. The deposits in places show river action and
streams shifting about over sand flats. Regarding actual tempera-
tures so little is known after all of the relations of modern plants to
their climatic environment that results can only be qualitative and
IV it quantitative.

■9I4-] SOUTHEASTERN NORTH AMERICA. 137

It is obvious that the flora could not have existed if the region was
ever visited by frost, and temperatures appear to have been like those
found today on the Florida keys. Aside from the meteorological
certainty that there was a wide coastal belt of abundant precipitation,
there is the confirmation furnished by the flora itself. It would seem
to me proper to compare the Wilcox flora with those of the regions
to which the somewhat loosely used term subtropical rain forest is
applied by plant geographers. Too little is known of the -Midway
flora for accurate comparisons. Compared with the Upper Cre-
taceous flora of the embayment area, in which however 40 percent
of the genera are extinct, the Wilcox would seem to have become
more tropical, a progression from what might be termed a warm
temperate to a subtropical rain forest. On the other hand the floras
as well as the faunas show a gradual increase of tropical conditions
in the later Eocene which culminate in the Oligocene, the flora of
which in southeastern North America is strictly tropical.

Lianas were apparently not as common in the Wilcox as they
are in the existing floras with which it has been compared. No
traces of the Bignoniaceas, so common in the American tropics, have
been detected, the scandent types being represented by Lyg odium,
Aristolochia, Malpighiaceae, Canavalia, Pisonia (?), and Zizyphus,
I am inclined to think that the great uniformity of climatic condi-
tions together with the abundant rainfall have combined to make the
Wilcox flora seem more tropical in character than was actually the
case. That reef corals are not found in the Wilcox is, I believe, en-
tirely due to physical conditions other than those of temperature
as Yaughan- has shown to be the case so often in such a striking
manner in recent seas.

1 have indicated upon the sketch map (Fig. 1 C-C) what I con-
ceive would be the northern limit of range of the Wilcox flora under
existing climatic conditions in southeastern North America.

It would seem to be probable that most of the generic types of
the Wilcox were differentiated by the close of the Cretaceous. If the
equatorial region of America was the place of origin of a majority of
those types which have not as yet been recorded from the Cretaceous
as I believe to be the case, they must have spread northward along

2 Yaughan, T. \V., Jouni. Wash. Acad. Sci., Vol. 4, pp. 26-34, IQM-

l:;s BERRY— LOWER EOCENE FLORA OF [April 25,

the Mississippi Gulf either during the Cretaceous-Eocene interval,
during the Midway or during the Midway-Wilcox interval. While
the time available for this northward dispersal was thus sufficiently
long to account for the migration of even the most slowly spreading
forms a short statement on the adaptations and agencies of this
dispersal is not without interest. The Wilcox genera with winged
fruits or seeds are Engelhardtia, Paraeiu/ellwrdtia, Dodoncea, Paliit-
rus, Fraxinus and the Proteaceae and Malpighiaceae. None of these
are capable of long flights except those of the last two families and
these during high winds might readily be carried for miles along
coasts, although it is doubtful if they could have crossed great
stretches of open water, even through the agency of a West Indian
hurricane. The heavier winged fruits such as those of Engelhardtia,
Paraengelhardtia, Paliurus, Dodoncca and Fraxinus float readily,
although as far as I know there is no experimental data to show
how long they float in oceanic waters without losing their vitality.
Certainly Dodoncca has reached the Bermudas in recent times
through the agency of the Gulf Stream. Among the Wilcox forms
more or less adapted for floating the following genera may be enu-
merated : Nipadites, Canna, Taxodium, Pisonia, Sapindus, Sterculio-
carpus, Trapa, Aviccnnia, Solanites, Exostema and the Combretaceae.
Among the foregoing Canna, Taxodium, Trapa and Exostema are
scarcely adapted for sea voyages while on the other hand, Nipadites,
Sapindus, Stercnliocarpits, Avicennia and the Combretaceae are sin-
gularly adapted for dispersal by ocean currents and would be in the
van of forms colonizing the shores of the transgressing Wilcox sea.

A large number of the Wilcox genera had fleshy or drupaceous
fruits and formed the dietary of both mammals and birds. Among
these the following may be mentioned : Myrica, Ficus, Coccolobis,
Magnolia, Anona, Asiniina, CJirysobalanus, Simaruba, all the Laura-
ceae, Ilicaceae, Celastraceae, Myrtaceae, Ebenaceae, Sapotaceae, Melia-
ceae, Euphorbiaceae, Anacardiaceae, Zizyplius, Guettarda, Citharcxy-
lon, Cordia, Osmantlnis, Icacorea, Rhamnus and Rcynosia. Many
of these have crustaceous stones that pass uninjured through birds
or are voided with their vitality unimpaired and these could undoubt-
edly be carried long distances over seas. Even in the case of soft
seeds like those of a large number of the Leguminosse it has been

I9I4-] SOUTHEASTERN NORTH AMERICA. 139

found that birds that have eaten greedily often void them uninjured
and others meet with fatalities before the seeds are digested and
these constitute by no means unimportant factors in distribution.
Clement Reid in his discussion of the origin of the British flora gives
an instance of a dead wood-pigeon with beans sprouting from its
crop, and when it is remembered what a great percentage of birds
meet an untimely end it is conceivable that a single hurricane might
readily be the means of introducing new forms from the Antilles
upon the Wilcox coast. Other Leguminosse, although more rarely,
are dispersed by ocean currents, as is the case in an eminent degree
with the modern Entada or snuffbox seabean.

All of the storms moved from the equator northward, the main
ocean currents had the same general direction, while the prevailing
winds were easterly so that all of these important factors combined
in causing a relatively rapid introduction and spreading of forms
along the Wilcox coasts, so that given favorable climatic conditions
and many of the forms need not have taken the time to spread from
Central or South America along continuous coasts.

Botanical Character of the Flora.

That the method by which the bulk of the determinations in the
present study are made rests upon real and not fanciful affinities is
of vital importance, since the resulting climatic and other physical
data are largely controlled by these facts. The case is not as intri-
cate or as hopeless as it might seem to the student who remembers the
thousand of living and extinct genera. De Candolle estimated the
total number of flowering plants to be about 250,000 species. This
figure is swollen by the great multiplication of herbaceous species in
recent geologic times. The ratio of arborescent to herbaceous types
was much greater in the Tertiary than it is at the present time and
it seems probable that trees were actually more abundant and
varied than in the existing flora. This was certainly true for all
Tertiary floras outside the torrid zone and may be readily proved by
a consideration of the Eocene floras of North America, the Miocene
floras of Europe or to cite an extreme case the Tertiary floras of
the Arctic and Antarctic regions.

While the arborescent flora of the temperate zone is relatively

140 BERRY— LOWER EOCENE FLORA OF [April 25,

meager, trees increase in passing toward the equator. For example
the state of Maryland which is in latitude 38 to 39 30' is in effect
a cross section of the Coastal Plain, Piedmont Plateau and the Alle-
ghany Mountains with great differences of climate, topography and
soils. It is moreover the meeting ground for plants of northern and
southern range. In spite of these facts there are only about 150
arborescent forms in its flora. On the other hand in Small's recently
published Trees of Florida (1913) there are 366 native and natu-
ralized arborescent forms, and if Florida furnished much altitudinal
variation the number would be much larger. For example the ar-
borescent flora of the Philippine Islands includes 6C5 native species
and many additional introduced forms. Even remote oceanic islands
if of sufficient size and topographic variety to overcome the adverse
action of winds have a large arborescent flora. Thus the Sandwich
Islands have 225 native species of trees distributed among 45 families,
the larger being the Rutacese (32 species), Rubiacese (31 species),
Campanulaceae (15 species), Araliacese (14 species), Pittosporacere
(12 species), Palmaceas (11 species), Myrsinaceae (11 species), and
Malvaceae (10 species). 3

The general physical conditions of a remote geological epoch may
be more or less completely deduced in advance from the character of
the sediments : the run-off from the land can be approximated and
consequently the altitude of the land and the probable rainfall as well
as any periodicity in these factors. These are all reflected in the
sediments. Work like that of Vaughan 4 on the deposits of the
Florida Keys or that of Drew 5 on the part played by denitrifying
bacteria in the formation of limestones, enables a careful paleobota-
nist to in a measure predict the character of the flora that clothed the
marginal lands. In deposits that teem with the remains of marine
life as do many of the Tertiary formations of southeastern North
America it is possible to arrive at very close approximations of the
actual temperatures of the coastal waters. It may be safely assumed
that boreal or temperate floras did not flourish in proximity to trop-

3 Rock, J. F., "The Indigenous Trees of the Hawaiian Islands,'' Hono-
lulu, 1913.

1 Vaughan, T. \Y., Carnegie Institution. Publication 133, 1910.
■'Drew. < i. II., Vi arnegie Institution, No, 10, 1911.

i9i4-] SOUTHEASTERN NORTH AMERICA. 141

ical marine faunas : that plants reflected their environment in the
past as in the present.

A considerable number of botanists love to dwell upon the temer-
ity of their paleobotanical friends in venturing to determine leaf
impressions. I admit at the outset that some determinations are
much too sanguine, especially when based upon fragmentary mate-
rials. There is more or less convergence in foliar characters in
unrelated or remotely related families and there is also considerable
variation in the leaves of a single species, but the fact remains that
foliar characters in general are more conservative than those derived
from almost any other organs of plants. They are subjected to less
complex environmental factors and always have been. It should be
remembered that characters less essential in the vital activities of
plants, such as leaf form, when once acquired, may continue practically
unchanged for thousands of years and afford a surer clue to relation-
ship than characters more immediately within the field of action of
natural selection. This is shown by the persistence of fern fronds
on the Paleozoic Pteridosperms ; by the uniformity of Cycad-like
fronds from the Permian to the Cretaceous ; and by the striking per-
sistence of dicotyledonous foliar types from the Mid-Cretaceous to
the present. It is paralleled by Dall's observations on the persistence
of superficial and ornamental shell characters in the Mollusca from
the Cretaceous to the Recent.

The opinion I mention in a preceding paragraph is mainly the
result of ignorance both of foliar characteristics and paleobotanical
literature, and an unwillingness to spend the time necessary for a
mastery of the subject. I have tested systematic botanists time and
again with recent leaves and the results are not especially creditable.
They generally know that leaves are green in color and that some are
simple and others compound : they may even know whether the mar-
gins are entire or toothed, but the venation is usually a closed book.
I know of but one manual that pretends to pay careful attention to
foliar characters and that is Sargent's " Tree Book " and even here
the figures pay no attention to venation. 6 In the tropics where
flowers and fruits are often unobtainable or beyond reach it is easy to

G Sudworth's "Trees of the Pacific Slope" is the most admirable work
in this respect that has ever been published.

142 BERRY— LOWER EOCENE FLORA OF [April 25,

learn to recognize most arborescent forms by their habit and foliage
but to most botanists, systematic or otherwise, anything beyond the
floral structure receives but scant attention.

It is reasonable to conclude that palms and tree-ferns were never
boreal plants that have in the course of ages become restricted to the
tropics as Xaumayr once suggested in an effort to explain their pres-
ence within the Arctic circle on other than climatic grounds. L ni-
formity of conditions is the foundation upon which the whole fabric
of our knowledge of past events rests and it is just as unscientific to
assume that the carrying power of water was not conditioned by its
velocity during the Tertiary as it is to assume that insolation, humid-
ity, rainfall, winds and all the other factors that constituted the en-
vironment of the vegetation, had effects different in kind from their
effects on the living flora.

In a study of this sort the chief emphasis should be based upon
comparisons with the existing relatives of the fossil forms and not
upon a search among the illustrations of works devoted to the study
of previously described forms, often from remote regions, for what
appears similar. The latter should not be neglected however and no
descriptions are complete unless they include a discussion of the re-
semblances and differences of previously described forms that show-
similarities to the form in hand with their geologic and geographic
destribution. Even the most trivial characters of the fossil should be
carefully noted since all are or will become valuable in future studies.
The living representatives, their habitat, range and variation are of
the greatest importance in determining paleoecology.

Unless there is clear evidence of transportation it may be assumed
that strand plants and upland plants will not be found in association
and if such seems to be the case, additional study may reveal the
errors of determination.

That all floras are dynamic and not static: that all their elements
are more or less plastic in their reactions to the infinite complexity . .1"
their environment raises a certain amount of scepticism regarding the
methods and results of what may be called paleoecology. Thi< is
especially true since so little is known regarding the precise relations
between existing plants and their environment. At the same time
there is no Other method available and it must be considered to be a

I9I4-] SOUTHEASTERN NORTH AMERICA. 143

legitimate method until negatived in human experience. If it be
assumed untrue there is no limit to idle speculations as futile as those
of medieval times.

The Wilcox flora as described in the present study comprises con-
siderably more than 300 species — the exact number is without sig-
nificance since it is so largely dependant on accidents of preservation
and discovery, and since it is also considerably influenced by the
evaluation of specific characters. The number might readily be
increased to 400 if fragments of new forms were considered the
basis for the description of species.

This flora is therefore one of the largest floras as yet known from
a single geologic horizon in a single area, although it is considerably
overshadowed numerically by the so-called Fort Union flora of the
Rocky Mountain Province, which however covers a greater geo-
graphic area and a longer interval of time.

Compared with foreign Eocene floras of similar age it may be
noted that Ettingshausen enumerated 72 genera and 200 species from
the London Clay of the Island of Sheppey 7 and 116 genera and 274
species from Alum Bay on the Isle of Wight. 8 I mention these two
English floras specifically since while never adequately described
they are at least partly contemporaneous with that of the Wilcox, as
I hope to show in the chapter on correlation, and they therefore offer
various interesting details for comparison as will appear on subse-
quent pages.

The Wilcox flora comprises 128 genera in 59 families and 33
orders. The Thallophyta are represented by a few species of leaf-
spot fungi, and if the student were to follow the fashion set by the
older continental paleobotanists the so-called species of spot-fungi
could be increased many fold, as I have only picked out for enumer-
ation certain conspicuous or characteristic types. The Bryophyta,
as is usually the case in fossil floras, are entirely unrepresented, al-
though the sediments are often of a character to have preserved them
in perfection if they had been present, and the assumption is logical
that they were either confined to more northern latitudes at this time
or were an exceedingly minor element in the flora. The Pterido-

7 Ettingshausen, Proc. Roy. Soc. Loud., Vol. 29, 1879, PP- 388-396.

8 Ettingshausen, Ibidem, Vol. 30, 1880, pp. 228-236.

144 BERRY— LOWER EOCENE FLORA OF [April 25,

phyta which are such a preponderating element in all fossil floras up
to the Middle Cretaceous are represented by a doubtfully determined
Lycopod and six species of ferns.

Ferns are among the most abundant (in specific differentiation)
vascular plants in the flora of tropical America, the island of Jamaica
being especially celebrated for its fern flora. Grisebach enumerated
340 species of ferns in his " Flora of the British West Indies" pub-
lished in 18^4. In Urban's more recent work 182 species of the
Polypodiacese alone are recorded from Porto Rico. The following
five genera have been recognized in the Wilcox: . \neimia, Lygoduuti,
Asplenium, Pteris and Meniphylloides — each represented by a single
species except the genus Asplenium which has two species. While
six species seems a small number of ferns in a subtropical flora like
that of the W'ilcox it is just twice as many as have been found in the
contemporaneous deposits of Alum Bay on the Isle of Wight where
the remains of an extensive flora is preserved in the pipe-clays. The
explanation of this seeming disparity between the fern representation
in the Lower Eocene and in modern floras is readily formulated and
it will also indicate the reasons for thinking that the real Wilcox fern
flora if it were available for study would be a rich and varied one,
comparable at least with the existing fern flora of the lowlands of
subtropical America.

The known Wilcox flora is almost entirely a coastal flora made
up very largely of strand types. Very few elements in it can be
legitimately considered as derived from inland areas by stream trans-
portation, in fact their condition of preservation alone proves that
they grew in the immediate vicinity of where they are now found as
fossils. With a few striking exceptions the existing tropical and
subtropical fern floras are floras of humid inland or upland habitats,
for example the majority of the Jamaican ferns are found on the
Blue Mountains. The most striking exception to this statement is
the genus Acrostichum which strangely enough has not yet been
positively recognized in the Wilcox flora although it was widespread
along the shores of the Mississippi Gulf in the succeeding Middle
Eocene (Claiborne) and Lower Oligocene (Yicksburg) floras, as
abundant apparently as it is in the existing flora of tropical tidal
marshes in both the Eastern and the Western Hemispheres. Another

I9I4-] SOUTHEASTERN NORTH AMERICA. 145

fern type liable to be present in coastal thickets is the genus Lygo-
d 'iu 111 with its scandent habit, and this genus is represented in the
Wilcox flora by both sterile and fertile fronds. It is likewise com-
mon in the Claiborne and Vicksburg floras and in Tertiary floras
generally. Besides Lygodium, the family Schizaeacese is represented
by a species of Aneimia which must also be considered to have been
a coastal type in the early Eocene as are some of its species at the
present time, since very similar species of Aneimia are found at a
very large number of Eocene coastal deposits both in this country
and abroad.

The remaining four species of Wilcox ferns are all referable to
the family Polypodiacese which is the dominant existing family of
the fern phylum. The two species of Asplenium are types readily
matched by existing Central American species. The Ptcris, not cer-
tainly identified as a true species of this common cosmopolitan type,
had stout coriaceous fronds and may have been transported since it
occurs at only two localities in the Wilcox and at one of them it
is in a fragmentary condition. This supposition receives some sup-
port from its presence in the basal Eocene of the Rocky Mountain
province after the sea had withdrawn from that area and after there
had been a large amount of volcanic activity and more or less uplift.
The genus MeniphyUoides is a unique type as yet peculiar to the Wil-
cox flora although it is closely related to the similarly unique genus
Meniphyllum Ettingshausen and Gardner from the Middle Eocene
(Lutetian) of England and both are closely related to and possibly
the progenitors of the existing genus Meniscium which has at least
one species that is close to the Wilcox form. MeniphyUoides is
only found at two localities near the top of the Wilcox and its prob-
able habitat is not known. The remains are broken but are asso-
ciated with a typical strand flora.

It will be seen that of the Wilcox ferns whose habitats can be
surmised all are coastal types and when we recall that the mainland
was relatively low throughout Wilcox time it is not surprising that
the ferns are not more strongly represented. By a specialization of
habitat in modern equatorial regions a considerable proportion of
the flora becomes epiphytic, the smaller ferns being commonly so.
None of the members of the extensive Wilcox flora can be regarded

PROC. AMER. PHIL. SOC, I.III. 2I4 .1, PRINTED JULY IO, I9I4.

146 BERRY— LOWER EOCENE ELORA OF [April 25,

as epiphytes with the possible exception of Lycopodites? eolignitica
which is such a rare and poorly represented form that it is without
significance. Apparently epiphytes were not conspicuous in the Wil-
cox coastal floras so that this possible source of supply for additional
fern species is also eliminated.

The Gymnospermas so conspicuous in Mesozoic floras are rela-
tively unimportant in the Wilcox flora, a feature due to their gen-
eral unimportance in Cenozoic floras and to their intolerance of the
habitats and climatic conditions indicated by the tout ensemble of
the Wilcox flora. All of the five Wilcox gymnosperms are referred
to the relatively modern family Pinaceae and none of the genera are
especially close to Mesozoic types. The Cycadaceas which might be
expected to be represented by Zam /a-like forms have not been found
although the presence of typical Williamsonia fructifications in the
Upper Cretaceous of the coastal plain indicates that the Cycad phy-
lum had not been long extinct in this area.

The Angiospermre, beyond all odds the dominant type in existing
floras, was as clearly dominant in Wilcox time since to it belong over
94 per cent, of the known Wilcox flora. Of these numerous angio-
sperms only seven are referable to the Monocotyledons. It is true
the number of monocotyledons might have been increased by describ-
ing the various sedge or grass-like fragments that are not uncommon
at certain localities. None of these have, however, been dignified by
names except a single form each of Poacites and Cyperites which
were only retained since they were already in the literature. That
only three species of palms have been recognized is remarkable
since palms were well differentiated at this time and various genera
such as Phoenicites, Thrinax, Gconoma, Bactrites, Manicaria, etc., are
recognized in our later Tertiaries. In the contemporaneous deposits
of Sheppey of the 30 monocotyledons enumerated by Ettingshausen
(op. cit, p. 393) 22 species are palms. On the other hand the Alum
Bay flora contemporaneous and not far distant from the Sheppey de-
posits furnished only 6 monocotyledons. This contrast indicates that
the fruits accumulated at Sheppey in the delta of an Eocene river
system contain interior forms not present in the coastal region repre-
sented by the Alum Bay clays and that inland from the Wilcox coast
the display of monocotyledons suitable to the Wilcox environmental
conditions flourished but failed of preservation.

19I4-] SOUTHEASTERN NORTH AMERICA. 147

Since the early Eocene floras of Europe are so much like those
of southeastern North America an enumeration of the Sheppey
palms is of considerable interest. They include the genera Nipa,
CEno carpus, Areca, Iriartca, Livistonia, Sabal, Chamcerops, Thrinax,
Bactris, Asterocaryum, and Elccis. Of these Nipa and Sabal are rep-
resented in the Wilcox flora while Thrinax and Bactritcs are present
in the embayment area in the Middle Eocene (Claiborne). The
Order Palmales, or more properly Arecales, has a single existing
family the Arecareae (Palmse) with about 150 genera and consider-
ably over a thousand existing species about equally divided between
the oriental and occidental tropics. There are no temperate outliers,
although some species extend for considerable distances into the
temperate zone as for example Sabal adansonii which ranges north-
ward along the Atlantic Coast as far as North Carolina. The pres-
ent distribution of the palms is a good illustration of modern con-
tinental floral diversities succeeding a Tertiary cosmopolitanism of
floras and it shows further the part played by isolation in evolution,
also indicated by the abundance of monotypic genera in the Orient
where the tropical area is so much broken. Not a single species or
genus is common to the two hemispheres and even the tribes are
almost all either oriental or occidental.

Regarding the origin of the palms most students regard the Pan-
danacese (screw pines) as their probable ancestral stock and while
the latter family is entirely oriental at the present time this was not
true in the Tertiary, and it is perhaps significant that the existing
genus Phytelephas which is regarded as intermediate between the
Pandanaceae and the Arecacese is exclusively American, and that
genera now exclusively oriental like Nipa and Phoenix are repre-
sented in the American Tertiary (Nipa in the Wilcox and Phoenix in
the Vicksburg). There is no warrant for asserting that palms are
of occidental origin, at the same time their oriental origin is equally
difficult of proof and what we know of their geologic history conclu-
sively shows the inadequacy of the existing distribution in a discus-
sion of their phylogeny.

The three Wilcox species of palms comprise a fan palm and two
feather palms. The Chamccdorea leaves represent a small palm
whose numerous modern allies are confined to America, being richest

148 BERRY— LOWER EOCENE FLORA OF [April 25.

in species in the humid mountainous regions of Central America. It
is not a coastal form and is not found in association with the typical
Wilcox strand flora, occurring only in the basal Wilcox of Choctaw
County, Mississippi, and at the base of the transgressing Upper Wil-
cox deposits in Saline County, Arkansas. Its rarity and occurrence
in basal beds would seem to indicate that its area of growth was
inland and only reached in these two cases by the landward migration
of the strand line. The Sabalites, which I have compared with the
existing Sabal palmetto, is common everywhere from the base to the
top of the Wilcox. It is distinctly a coastal type, rather of the
lagoons, bayous and estuaries than of the strand. This is indicated
by the fragmentary nature of the remains at very many localities and
the occurrence of innumerable complete specimens at other localities
as for example at Oxford, Mississippi, where the presence of unios
and the local unconformities indicate estuary conditions.

The Xipa palm found in the Upper Wilcox is clearly an inhabi-
tant of muddy tidal shores so that it would naturally be expected in
the laminated clays of the Upper Wilcox. Its single modern repre-
sentative is tolerant of water of considerable salinity and is a mem-
ber of the mangrove association of the Orient. It shows many points
of affinity with the Pandanacea? and has never before been found
in the Western Hemisphere. Like so many forms which are strictly
oriental in the existing flora such as Cinnamomum, Artocarpus, Pluv-
nix, etc., it enjoyed a cosmopolitan range during at least the earlier
half of the Tertiary period.

A somewhat full account of Xipa has been recently published
by me 9 and need not be repeated in the present connection.

The single species of Canna of the Wilcox represents a strictly
hygrophilous type winch is confined to America in the existing flora.
It is an inhabitant of estuary and river swamps near the coast, and
that the Wilcox species inhabited a similar situation is indicated by
its restricted occurrence and its association with Sabalites near the
mouth of a Wilcox river, which on other grounds is known to have
been present in Lafayette County, Mississippi.

The Dicotyledons of the Wilcox as might be expected are largely
choripctalous forms since there are over 250 species of Choripetalae

* Berry, E. YV., Am. Jour. Sci. (IV. >. Vol. 37, pp. 57-60, Fig. 1. M14.

I9I4-] SOUTHEASTERN NORTH AMERICA. 149

(Archichlamydeae) and only 35 species of Gamopetalae (Sympetalae).
At the same time the representation of Gamopetalge is really much
larger than might be expected thus early in the Eocene and many
families often thought to be relatively more modern have been found
to be represented.

The following orders of Choripetalae are not represented in the
Wilcox flora: Casuarinales, Piperales, Salicales, Balanopsidales, Leit-
neriales, Santalales, Sarraceniales and Opuntiales. The absence of
the Balanopsidales, Sarraceniales and Opuntiales is not remarkable
since they are all specialized types and the rather uniform habitats of
the cacti and their relatively modern evolution both conspire to elim-
inate them from Eocene coastal floras. The presence of the prim-
itive Casuarinales and Piperales might be expected especially since
there is a well marked Piper-like form in the Upper Cretaceous of
Alabama. The Salicales while prevailingly temperate forms are
abundantly represented in the Upper Cretaceous floras of the embay-
ment area and the Santalales have also been recorded from the
American Upper Cretaceous and are present in the European Ter-
tiary.

Those alliances of Gamopetalae which are not present in the Wil-
cox to be enumerated presently are mainly the great modern and
temperate zone groups. For example there are no Wilcox species of
Ericales, Labiatae, Convolvulaceae, Bignoniacese. Scrophulariaceae,
Plantaginales, Valerianales or Campanulales, this proving not only
the essential modernness of the evolution of the Compositae 10 but
firmly establishing the thesis that the Wilcox flora is a subtropical
and not a temperate flora.

The following are the larger families in the Wilcox flora : The
Lauraceae with 30 species, Caesalpiniacea? with 26, Moraceae with 23,
Papilionaceae with 22, Rhamnaceae with. 14, Sapindaceae with 13,
Sapotaceae with 12, Myrtaceae and Mimosaceae each with 11, Combre-
taceae and Anacardiaceae each with 9, Juglandaceas with 8, Celastra-
ceae with 7, and the Proteaceae and Apocynaceae each with 6.

The largest single genus is Ficus with 18 species, then comes
Cassia with 12, Sapindus with 9, Gleditsiophyllum with 8, Oreo-

10 The fruit described as Carpolithus hyoscritiformis is probably referable
to the Compositae.

150 BERRY— LOWER EOCENE FLORA OF [April 25,

daphne, Sophora and Anacardites each with 7, Cinnamomum, Nec-
tandra, Rhamnus, Myrcia and Bumelia each with 6, and Celastnis,
Dillcnitcs, and Apocynophyllum each with 5. Ten species are re-
ferred to the form-genus Carpolithus and this number could readily
be greatly increased if all the unidentified seeds were named and
described.

The amentiferous families, in accordance with their Upper Creta-
ceous deployment and their undoubted primitive and not reduced
character, are represented in the Wilcox flora by fourteen species, a
number of which are individually abundant.

The Juglandales 11 are represented in the Wilcox by three species
of Juglans only one of which, Juglans Schimperi, is at all common;
by a doubtfully determined species of Hicoria; by three well-marked
species of Engelhardtia and by an extinct type, Paraengelhardtia, of
a habit similar to that of Engelhardtia.

The genus Juglans is one of the earliest of the still existing dicoty-
ledonous genera to appear in the fossil record and it is continu-
ously represented in fossil floras from the Mid-Cretaceous to the
present. There are about 25 Eocene species of Walnut and they
range during that period from the Gulf region to Alaska and Green-
land, and are also present in the tropical forests of the Egyptian
Fayum in the early Oligocene. The outlying existing species in
the West Indies and under the equator in South America prove that
in spite of the northward range of the Asiatic species in Manchuria
and of some of the North American species into New England and
southern Ontario, its progenitors were at least subtropical types, a
fact corroborated by their foliar character since it is a well-known
fact that compound leaves indicate tropical ancestry, and this is
abundantly proven in the case of Juglans by its associates in the fossil
floras where it has been found represented.

The genus Engelhardt'ui vz is one of the most interesting Wilcox
genera. In the first place the identification of its leaves is corrobo-
rated by two varieties of characteristic winged fruits.

The genus was described by Leschen in 1825 and contains about

11 See Berry, E. W., "Notes on the Geological History of the Walnuts
and Hickories," Plant Woirld, Vol. 15, 1912, pp. 225-240.

l2 See Berry, E. \\\. Am. Jour. Sci. (IV.), Vol. 31, 1911, pp. 491-496;
Plant World. Vol. 15. I'HJ, pp. 234-238, Figs. 3, 4.

I9I4-] SOUTHEASTERN NORTH AMERICA. 151

ten species of the southeastern Asiatic area. These range from the
northwestern Himalayan region, where they extend a short distance
north of the Tropic of Cancer, through farther India and Burma
to Java and the Philippines. The pistillate flowers are small and are
grouped in paniculate spikes. They develop into small drupe-like
fruits, each of which is connate at the base to a large expanded tri-
alate involucre.

A single little known species, rarely represented in even the larger
herbaria, occurs in Central America and is the type and only species
of the genus Oreomunnea of Oersted. This is much more restricted
in its range than are its kin beyond the Pacific. Oreomunnea is very
close to Engelhardtia, and for the purposes of the paleobotanist the
two may be considered as identical since they represent the but
slightly modified descendants of a common ancestry which was of
cosmopolitan distribution during the early Tertiary. The present
isolation of Oreonuinnea furnishes a striking illustration of the enor-
mous changes which have taken place in the flora of the world in the
relatively short time, geologically speaking, that has elapsed since the
dawn of the Tertiary.

The principle has frequently been enunciated that when closely
related forms are found in the existing flora of the world, restricted
in range and isolated from their nearest relatives, or when other
existing genera are monotypic, it is quite safe to predict an interest-
ing and extended geological history. Engelhardtia proves to be an-
other illustration of this principle, for its peculiar three-winged fruits
have been known in the fossil state for almost a century. They were
long unrecognized, however, and the earlier students who described
them compared them with the somewhat similar winged fruits of the
genus Carpinus ( Betulacese). With the botanical exploration of dis-
tant lands in the early part of the nineteenth century, specimens of
Engelhardtia began to be represented in the larger European her-
baria, and Baron Ettingshausen, that most sagacious of paleobota-
nists, as long ago as 1851 pointed out that certain supposed species of
Carpinus were really fruits of Engelhardtia. He returned to the sub-
ject in 1858 without, however, actually changing the names of any
of the supposed species of Carpinus nor does he seem to have been
aware of the existence of a living species of Engelhardtia (Oreo-
munnea) in Central America.

152 BERRY— LOWER EOCENE FLORA OF [April 25,

Since Ettingshausen's announcement a dozen or more fossil spe-
cies have been described. The oldest known European form occurs
in the lower Oligocene (Sannoisian) of France and the species be-
come increasingly abundant throughout- southern Europe, especially
toward the close of the Oligocene and the dawn of the Miocene,
Saporta stating that the slabs from the leaf-beds at Armissan in south-
eastern France are thickly strewn with their peculiar fruits. Fossil
forms continue in Europe throughout the Miocene and Pliocene and
specimens of late Miocene or early Pliocene age are recorded from
Spain, France, Italy, Croatia, and Hungary.

The Wilcox species are somewhat older than any of the known
European forms

The existing Engelhardtias are upland forms and this may pos-
sibly have been their habitat in Wilcox times although their abun-
dance at different localities along the Wilcox coast would seem to
indicate that this was not the case.

The genus Paraengelhardtia, which is a unique type confined to
a single locality in the Wilcox, is clearly allied to Engelhardtia, as I
have shown in the systematic chapter. It seems probable that it
represents a survival of the ancestral stock from which Engelhardtia
was derived since its fruits are more primitive and indicate ancestral
forms with smaller bracts comparable with the bracts of Juglans or
Hicoria which in the course of time became accrescent and subse-
quently deeply trilobate. The primitive character of Paraengelhardtia
and the presence of true Engelhardtias in the Wilcox so much earlier
than their first occurrence in Europe suggests that America was the
original home of the Engelhardtia stock, although this supposition
cannot be verified or disproved until a Tertiary paleobotanical record
for the continent of Asia is available.

The Myricales contains but two species of Myrica in the Wilcox
flora. Myrica is a very old generic type with a large number of
fossil species ranging from the Middle Cretaceous to the present.
The existing species arc relatively few in number and widely scat-
tered geographically and represent survivors from a Tertiary cosmo-
politan distribution. Myrica 1 ' is much less abundant in the Wilcox

l8 The allied and monotypic genus Comptonia which by some students is
included in Myrica lias an extended geologic history which has been discussed
by Berry, Amer. Nat., Vol. 40, 1906, pp. 485-520, pi. 1-4.

19I4-] SOUTHEASTERN NORTH AMERICA. 153

than in the European Tertiary, although it was present in the em-
bayment area in the late Upper Cretaceous (Ripley formation of
Tennessee). Its meager representation in the Wilcox time may be
due to the more tropical climate conditions. The modern Myricas
are temperate and subtropical and a number of the species are coastal
forms of either swamps or sand dunes. Myrica clccanoides was evi-
dently a coastal form and so was Myrica wilcoxcnsis. The latter is
very similar to the existing Myrica cerifera which ranges from New
Jersey to Texas and is also found on the Bermudas and Bahamas. It
is most abundant and vigorous in the sandy swamps along the south
Atlantic and Gulf coasts and its habitat may be compared with that
of Myrica zcilcoxcnsis. The latter seems to be the ancestral stock of
a very similar species which occurs along the Middle Eocene (Clai-
borne) coast of the embayment.

The order Fagales, which includes such important timber trees
of the temperate zone, is comprised by the two families Betulaceae
and Fagaceae, together containing about 450 existing species, of
which three fourths belong to the Fagaceae. Only the latter family
is represented in the Wilcox although the Betulaceas are character-
istically developed in the Upper Cretaceous of North America.

The family is unrepresented in the Wilcox flora probably because
the climate was too warm and this reason may also account for the
absence of true oaks since the Fagaceae are represented in the Wil-
cox flora by only the genus DryophyUum with four rather wide-
spread and often common species.

The genus Dryophyllum is of worldwide distribution and consist-
ently uniform characters in the various horizons of the late Cre-
taceous and early Eocene from the Senonian to the Ypresian stages.
It especially characterizes the dawn of the Eocene and represents the
ancestral stock from which the genera Castanea, Castanopsis, Pasania
and Quercus took their origin, although this origin was in the late
Cretaceous. As might be expected Dryophyllum has long since be-
come extinct. The Wilcox species were apparently strand types as
were also the numerous species enumerated by Debey, the describer
of the genus, from the sandy shores of the Upper Cretaceous sea of
Rhenish Prussia. Dryophyllum is abundant in the Montian of Bel-
gium and in the littoral sands of Ostricourt and Belleu in

154 BERRY— LOWER EOCENE FLORA OF [April 25,

France. In the systematic chapter detailed comparisons are made
between the Wilcox and the foreign species, which show a striking
parallelism.

The Urticales includes the families Ulmaceae, Moraceae and Urti-
caceae together containing about 1,600 existing species. The Urti-
caceae are largely herbaceous forms and the Ulmaceae are mostly
extratropical.

The Ulmaceae comprise thirteen genera and about 140 existing
species, widely distributed in temperate and tropical regions. A
single species of Planer a described originally by Newberry from the
Western Eocene is doubtfully identified from the Wilcox. The
genus is monotypic in the existing flora and confined to wet swampy
situations in the warm temperate region of southeastern North Amer-
ica. Its geologic history goes back to the Upper Cretaceous at which
time species have been recognized along the Atlantic coast from
North Carolina northward. Thus there is no reason why it should
not have been present in the early Tertiary of the embayment unless
it be argued that the climate was too warm.

The Moraceae, by far the largest family of the order Urticales
and the only one certainly represented in the Wilcox flora, contains
between 900 and 1,000 existing species segregated among about 55
genera, of which the genus Ficus is by far the largest, including
about 60 per cent, of the existing species of the family. The Mora-
ceae are distinctly tropical and warm temperate types and are most
abundant in the oriental tropics, although the dominant genus Ficus
is widespread and the family is also largely represented in the South
American tropics.

There are at least 18 monotypic genera of which one is North
American, four South American, four African, and nine Australian.
No single tribe is confined to a single continental area and all show
apparent anomalies of distribution due to our lack of knowledge of
their geologic history. The genera Ficus, .Irtocarpus and Artocar-
pidium go back to the base of the Upper Cretaceous and numerous
additional genera appear in the Eocene.

There are 23 species of Moraceae in the Wilcox flora. The genus
Arto carpus is represented by three well-marked species. In the ex-
isting flora the two score known species of Artocarpus are confined

19*4-] SOUTHEASTERN NORTH AMERICA. 155

to the southeastern Asiatic region 14 although some of them are cul-
tivated in all tropical countries. The tribe Euartocarpccc of which
Artocarpus is the largest existing genus, have, however, five of their
genera confined to Central and South America, one confined to trop-
ical West Africa, two confined to the southeastern Asiatic region, one
to Borneo and one ranging from Japan to Australia. While the
geologic history of Artocarpus is only imperfectly known at least
15 different fossil species have been described. The oldest is a well-
marked form based on characteristic leaves and parts of the fruit
which show the typical surface features. It has been fully described
by Nathorst 15 and comes from the Atane beds (Cenomanian) of
West Greenland. Slightly younger is a less well-defined form re-
corded from the Emscherian of Westphalia, and the somewhat doubt-
ful genus Artocarpophyllum of Dawson from the Upper Cretaceous
of Vancouver Island. Another species is recorded from the Laramie
formation and the genus is widely distributed in the basal Eocene of
Xorth America. It continues in the Mississippi Gulf region until
the close of the Oligocene, the last recorded occurrence being in the
Alum Bluff sands at Alum Bluff on the Apalachicola River. On the
Pacific coast it is found in deposits in California and Oregon which
are referred to the Miocene. In the European area it occurs in the
Tongrian of France, the Tortonian of Baden, the Pontian of France
and Italy and the Pliocene of Italy. It is present in both the Plio-
cene and Pleistocene of the island of Java.

Artocarpus is said to be represented by petrified wood in the
Oligocene of the island of Antigua and it was evidently a member of
the American flora from the Upper Cretaceous until late in the Ter-
tiary, although like the genera Cinnamomum, Nipa, Phoenix, etc.,
it is not represented in post-Pleistocene American floras. An extinct
genus related to Artocarpus and named Artocarpoides by Saporta,
who described several species from the Paleocene of France, is repre-
sented by a single Wilcox species.

14 It is found throughout Oceanica and was present in the Hawaiian and
Marquesas when they were first visited by Europeans. It was introduced
into the West Indies in 1793.

16 Xathorst, Kgl. Svenska Vetens-Akad. Hand!., Bd. 24, No. 1, 1890, 10
pp., 1 pi.

156 BERRY— LOWER EOCENE FLORA OF [April 25,

The genus Cecropia with about 40 existing species confined to
the tropics of South America has two species in the Aquitanian of
Bohemia and the Midway and Wilcox form described as Ficus sp.
is very probably a representative of this genus.

The genus Pseudolmedia, with five existing species in the Amer-
ican tropics, has a well marked species in the Wilcox flora. As far
as I know it has not heretofore been recorded in the fossil state
although it is probable that some of the very numerous fossil species
of Ficus may represent Pseudolmedia.

The genus Ficus is represented by numerous species in the Wil-
cox flora no less than eighteen having been described and a number
of these are individually abundant. They include the narrow lance-
eolate forms of the Ficus elastica type, with close-set laterals, as
well as open-veined lanceolate forms, and the shorter and broader
palmately veined forms. None are lobate or have toothed margins..
Ficus was evidently much more abundant and varied along the Wil-
cox coast than it is today throughout the West Indies and more
nearly comparable in this respect with the display of figs in the East
Indies or in tropical South America.

The number of fossil forms that have been referred to Ficus are
very numerous, numbering perhaps 300 species. Xone are certainly
known from the Lower Cretaceous, the genus Ficophyllum 10 being
entirely doubtful. In the Upper Cretaceous, however, Ficus is very
widespread and abundant, seemingly indicating a Lower Cretaceous
ancestry as yet unknown. The Cenomanian stage has furnished 3
species in Greenland, 6 along the Atlantic Coast and 24 in the interior
of North America, as well as 11 in Saxony, Bohemia and Moravia.
The succeeding Turonian stage furnishes 4 species in Bohemia and
the Tyrol, and several in North America (Tuscaloosa, Magothy,
Black Creek, Kutaw formations). Later Upper Cretaceous horizons
have abundant species of Ficus everywhere throughout North Amer-
ica and Europe as well as in Greenland, Australia and New Zealand,
and this cosmopolitanism continues throughout the Tertiary, there
being about 50 Eocene species, about 60 Oligocene species, 90 Mio-
cene species and 20 riiocene species. Africa is added to the record
in the basal Oligocene, and Asia in the Miocene.

10 See Berry, E. W., Md. Geol. Surv., Lower Cret., 191 1, pp. 502-506.

19I4-] SOUTHEASTERN NORTH AMERICA. 157

The fossil records will have to be much more complete before the
original center of radiation of the Moraceae can be determined, the
present brief sketch can be said to merely indicate that not only Fiats,
but other genera like Artocarpus that are entirely oriental in the pres-
ent, were normal elements in North America floras, from the time
of the modernization of these floras at the beginning of the Upper
Cretaceous onward. Along our east coast, they apparently became
restricted in their range at the dawn of the Miocene and they ap-
parently never after became as important in southeastern North
America as they had been, or as they are in the recent flora of the
Orient.

The order Proteales includes the single family Proteaceae with
about one thousand existing species. They include the prominent
arborescent forms of Choripetalas in the Southern Hemisphere, to
which region all but the four genera Roupala, Protea, Leucospermum
and Helicia are confined. They are usually considered as Australian
types, in fact the majority of the genera and species are confined to
that continent, nevertheless there are four genera in South America
together containing over fifty existing species, and there are several
genera peculiar to the African flora; and the genus Helicia is predom-
inantly Asiatic.

The geologic history of the Proteaceae is perhaps one of the most
striking instances that paleobotany affords of the great difference in
geographical distribution in former ages from what could possibly be
inferred from a study of the present geographical distribution of the
members of this family, although there are some significant features
in the distribution of the recent forms that will be alluded to in a
subsequent paragraph.

The discovery of fossil forms of Proteaceae in the Tertiary depos-
its of Europe was the inspiration of a considerable literature 17 and
was the occasion of a rather acrimonious controversy regarding their
botanical affinity. This is well illustrated in the dissenting opinions
expressed by the botanists Hooker and Bentham who both regarded
fossil leaves as undeterminable. Starting with this apriori principle
it is difficult to see how they could arrive at any other conclusion.

17 See the writings of Unger, Heer, Ettingshausen, Schimper, Schenk and
Saporta.

158 BERRY— LOWER EOCENE FLORA OF [April 25,

The most expeditious refutation of their opinion is furnished by the
present distribution of some of the genera, c. g., the genus Roupala
has 36 species in tropical .America, 2 in New Caledonia and 1 in
Queensland : the genus Embothrium has four Andean species and one
in Australia: the genus Lomatia has 3 species in Chile, 4 in Australia
and 2 in Tasmania. It follows unless one is prepared to subscribe
to the doctrine of special creation for each continent or to the inde-
pendent evolution on separate continents of different species of the
same genus, that during their geologic history these genera must
have ranged over intervening areas, so that if the Cretaceous and
Tertiary plants of the northern hemisphere with fruits and leaves of
the Proteacea? are not related to the genera that they resemble most,
then forms with leaves and fruit resembling those of other families
must be fossil Proteace?e, which ought to seem absurd, even to an
English botanist. As a matter of fact, while exception may justly
be taken to some determinations of Unger, Ettingshausen and Heer,
they in no wise affect the main body of facts and there is so much
collateral evidence furnished for example by the geologic history of
the Araucarian conifers, and the history of the Proteaceas is so simi-
lar to that of the Myrtacese and Leguminosre — the two other great
families of the existing Australian flora, that the evidence seems
conclusive.

Turning now to the fossil record those who follow the opinion
of Hooker or Bentham will see how vast and substantial are the
supposed illusions of the paleobotanists. In addition to the two ex-
tinct genera in the Wilcox flora I have fossil records of 32 genera of
Proteace?e, although this is artificially enlarged by the joint usage,
according to taste, of names like Dryandra and Dryandroides, Bank-
sia and Batiksitcs, etc. A brief consideration of these genera with
fossil representatives will prove valuable. 18

The genus Protca Linne, from which the family takes its name,
has about 60 existing species occupying disconnected areas in Cen-
tral and South Africa. To it have been referred a middle Cretace-
ous species from Saxony; 3 Aquitanian species from Prussia. Bo-
hemia and Greece; 1 species from the Burdigalian of Italy; 1 from

1 This list is not complete but sufficiently so for the purpose of this
discussion.

I9I4-] SOUTHEASTERN NORTH AMERICA. 159

the Helvetian of Switzerland and one from the-Messinian of Italy.
Allied to Protea but possibly more generalized is the genus Pro-
teoidcs of Heer. This has several Tertiary species and a consider-
able number of Upper Cretaceous species (15). There are two each
in the Cenomanian of Bohemia and Lesina, two in the Atane beds
of Greenland, three in the Dakota sandstone of North America, one
in the Tuscaloosa formation of Alabama, one in the Middendorf beds
of South Carolina, one in the Cretaceous of Australia, two in the
Vancouver Island Cretaceous and one in the Senonian of Saxony.

The genus Proteophyllum Velenovsky 19 a still more generalized
proteaceous type has seven species (Saporta, 1894). in the Albian
(Vraconian) of Portugal and 8 species in the Perucer beds (Ceno-
manian) of Bohemia. Another generalized type is Protcopsis Vel-
enovsky with a single species in the Cenomanian of Bohemia. The
genus Protecephyllum of Fontaine containing 2 species in the Patux-
ent fomation (Neocomian) of Virginia I regard as entirely worth-
less. 20 The genus Conarrhenes Labill with one existing species in
Tasmania has a single species based on both foliage and fruit in the
Miocene of Carniola according to a determination of Ettingshausen's
which may well be viewed with suspicion. The genus Conospermum
Smith with about 33 existing species in Australia has two fossil spe-
cies in the Oligocene of Styria and one in the Miocene of Carniola,
while the somewhat less definite genus Conospermites (Ettings-
hausen, 1867) has a fossil species in the Upper Cretaceous of Aus-
tralia and one in the Cenomanian of Saxony and Bohemia.- 1

The genus Helicia Lour, is of especial interest since it is found
farther north in the existing flora than any other member of the
family. There are about 25 modern forms, mostly Indomalayan,
but a few still survive in or have recently spread to Australia. The
fossil record includes a species in the Oligocene of Styria and another
in the Pliocene of Italy. The genus Lambcrtia Smith with 8 exist-
ing Australian species has a single fossil species in the Miocene of
Carniola. The genus Hakca Schrad. with 100 recent Australian

19 Velenovsky, Kvetena ceskeho cenomanu, 1889, p. 18.

20 See Berry, Md. Geol. Surv., Lower Cretaceous, iqi 1, pp. 494-499.

21 1 regard Fontaine's determination of a species in the Lower Cretaceous
of Virginia as worthless.

160 BERRY— LOWER EOCENE FLORA OF [April 25,

species has eleven fossil species in the Oligocene of Europe; in
France, the Tyrol, Saxony and Greece; and no less than 17 Miocene
species in France, Italy, Switzerland, Baden, Hesse, Prussia, Bohe-
mia. Austria, Styria, Croatia and Hungary.

The genus Knightia R. Brown with a modern species in Australia
and 2 in New Caledonia has a fossil form in the Eocene of Australia
and another in Graham Land'-' 2 in beds regarded as Oligocene. The
allied genus Knightites Saporta has two species in the Sannoisian
of France.

The remarkable genus Lomatia, previously mentioned, has four
existing species in Australia, 2 in Tasmania and 3 in Chile. As
might be expected from their modern isolated occurrences there are
over 30 fossil species based in some cases on associated leaves and
fruits. The oldest of these are two (perhaps wrongly identified)
species in the Dakota Sandstone. Eocene records include the Green
River shales of North America, a Ypresian species from the south
of England, an Italian species, five Australian and one Tasmanian
species. There are about a dozen Oligocene species, some of which
are very characteristic. They occur in the Tyrol, Saxony, Baltic
Prussia and Styria, and the relatively large number of four are
recorded by Dusen from Graham Land (Antarctica). There are also
about a dozen Miocene species recorded from such separated areas
as Colorado, Switzerland and Carniola. The wonderfully preserved
leaves in the volcanic ash beds at Florissant, Colorado, from which
seven forms have been described, the only known Miocene occurrence
of Lomatia in Xorth America, are alone sufficient to confound the
sceptics.

The allied genus Lomatitcs Saporta has a Cenomanian species in
Saxony and five or six Oligocene species in France. The genus
Stcnocarpus R. Brown, with 11 existing species in New Caledonia
and 3 additional ranging from North Australia to New South Wales,
has a single fossil species in the Oligocene of Saxony.

The genus Pcrsoonia Smith has 60 existing species in Australia
and one in New Zealand. The fossil record includes two widely
distributed species in the Upper Cretaceous of North America ; one

--'Duscn. Wiss. Ergeb. Schwed. Siidpolar. Exped., 1901-03, Bd. 3, Lief.
3. p. 7, PI. i, Figs. 7, 9. 11, 1908.

i9i4.] SOUTHEASTERN NORTH AMERICA. 161

in the English Eocene; four in the Oligocene of Tyrol, Saxony,
Styria and Greece ; ten in the Miocene of France, Italy, Switzerland,
Baden, Bohemia, Styria, Croatia, Carniola and Slavonia. A large
number of these fossil forms of Pcrsoonia are not especially convic-
ing but certainly the three European species Persoonia cuspidate,
daphnes, and Myrtillus of Ettingshausen 23 which have the leaves
associated with characteristic fruits are above suspicion.

Bowerbank in his classic study of the pyritized fruits and seeds
from the Island of Sheppey established a genus which he called
Pctrophiloides from its resemblance to the genus Petrophila R.
Brown which has about 35 existing species in Australia, the majority
of which are confined to West Australia. Bowerbank described sev-
eral species one of which was shown by Starkie Gardner to be an
Abuts fruit and others have been referred to Sequoia. Ettings-
hausen 24 in the study of the Sheppey fruits after careful comparisons
retained three English Eocene species and the genus has also been
recognized in the Sannoisian of Dalmatia and Styria.

The genus Lcucadendrites was established by Saporta for a San-
noisian species of southeastern France from its resemblance to
Lcucadendron Herm., which has upwards of 70 existing species in
South Africa.

The genus Grevillca R. Brown has 56 existing species confined to
Australia. The fossil record includes a Cretaceous species in Aus-
tralia; two Cenomanian species in Bohemia {GreviUeophyllum Vel-
enovsky) ; three Eocene species in England, France and Italy; twelve
Oligocene species mostly in southern France but also represented in
Saxony, Tyrol, Bohemia, Styria and Greece ; and twelve Miocene
species in France, Switzerland, Bohemia and Croatia.

The genus Embothrium Forst., already alluded to, has four ex-
isting species in South America which range from Chile to the
Straits of Magellan, and a fifth species in Australia. This widely
separated occurrence is explained when the fossil record is combined
with the occurrences referred to Embothrites, EmbotJiriopsis and
Embothriophyllum. To Embothrium are referred 8 Oligocene spe-

23 Ettingshausen, Site. K. Akad. Wiss., Wien, Bd. 7, 1851, pp. 718-719,
PI. 30, Figs. 6-14.

- i Ettingshausen, Proc. Roy. Soc. Lond., Vol. 29, 1879, P- 394-

PROC. AMER. PHIL. SOC, LIII. 2I4 K, PRINTED JULY II, I9I4.

162 BERRY— LOWER EOCENE FLORA OF [April 25,

cies of Styria and Greece and 4 Miocene species of Baden, Styria,
Croatia and Hungary. To Embothriopsis Hollick a single species
from the Long Island Middle Cretaceous is referred.

Embothriophylhtm is used by Dusen for a single species from the
supposed Oligocene of Graham Land. The genus Embothrites
Unger has a doubtful species in the Dakota Sandstone; six Oligo-
cene species in France, Tyrol, Styria, Carniola and Greece ; and 3
Miocene species in Croatia and Bohemia.

The genus Dryandra R. Brown has about 50 existing species in
Australia. The fossil forms have occasioned much discussion and
have been referred back and forth between this genus and Comptonia
and Myrica. The forms retained in Dryandra include a Cenomanian
species in Bohemia and Moravia ; an Eocene species in France ; two
Eocene species in Australia and an Oligocene species in Greece. The
allied forms referred to the genus Dryandroides Unger include 5
Upper Cretaceous species in Europe and North America ; an Eocene
species in Tasmania; 4 Oligocene species in Italy, Tyrol, Saxony,
Styria and Greece; and a Miocene species in Bohemia.

The allied genus Banksia Linne fil., also confined to Australia in
the existing flora, has 7 Upper Cretaceous species — 4 Australian and
3 in the North Temperate zone, ten Eocene species, of which 7 are
Australian, 1 Alaskan ( ?) and 2 English ; twelve Oligocene species
widely distributed in Europe; 16 Miocene species equally widespread
in Europe ; and a Pliocene species in Italy. Three especially well
marked species from the Wilcox have been referred to this genus.

The allied genus Banksites Saporta has a Cenomanian species in
Bohemia and various Tertiary records from Europe hopelessly en-
tangled in the literature with Banksia, Dryandra and Dryandroides.
The genus Roupala Aublet (Rhopala), whose peculiarly isolated out-
liers in Queensland and New Caledonia have already been mentioned,
is common in northern South America, extending northward to
Guatemala. Fossil forms are recorded from the Cenomanian of
Saxony, from the Eocene of Australia and from the Aquitanian of
Switzerland. Tn addition Saporta described a Rhopalospermites from
the lower Oligocene of France and a species of Rhopalophyllum has
been described from the Upper Cretaceous of Australia and a second
from the Miocene of Styria.

L9I4-] SOUTHEASTERN NORTH AMERICA. 163

The geological history sketched in the preceding paragraphs is
necessarily fragmentary, nevertheless I think the data are sufficient
after excluding doubtful determinations to show that the family had
its origin in the northern hemisphere, making its first appearance in
the fossil record at the close of the Lower Cretaceous, becoming
practically cosmopolitan during the Upper Cretaceous at which time
it reached the Australian region from southeastern Asia. New Zea-
land must have already been segregated but not the land mass now
represented by New Caledonia. During the early half of the Ter-
tiary Africa and southern Europe were essentially a single floral
province while in the Western Hemisphere the Proteacese ranged
from the United States through South America and an unknown dis-
tance across Antarctica. Concomitant with the continent building
and the consequent climatic changes of the Miocene the area of dis-
tribution commenced that shrinking which culminated during the
Pleistocene, leaving the stranded remnants of the stock in their pres-
ent widely separated quarters of the southern hemisphere. Not all
the modern genera took part in this history since the local peculiari-
ties of poor soil and rigorous climate combined with relative freedom
from outside competition were the factors that stimulated a Tertiary
evolution of forms in Australia in exactly the same manner as the
peculiar Australian genera of Myrtaceae and Leguminosse were
evolved.

The Wilcox species of Proteacese are six in number and are dis-
tributed in four genera, in addition to which a probable Banksia
fruit is retained in Carpolithus. These genera are Palccodendron,
Proteoidcs, Knightiophyllum and Banksia. The genus Polccodendrou,
not mentioned in the preceding paragraphs, was proposed by Saporta
for small entire coriaceous leaves from the Sannoisian of southern
France and is an entirely extinct type, sparingly represented in the
Wilcox by a single species. The genus Proteoides was established
by Heer for generalized proteaceous types which are well represented
in the Upper Cretaceous floras of the embayment area as well as
elsewhere. It is represented in the Wilcox by a single well-marked
species confined to the Middle and Upper beds. The genus Knighti-
ophyllum is proposed here for the first time for a well-marked long
petioled, aquiline-toothed, coriaceous form of common occurrence at

164 BERRY— LOWER EOCENE FLORA OF [April 25,

Peryear. It is named from its resemblance to the genus Knightia
R. Brown, a genus of few existing species confined to the Australian
region but apparently represented in Europe during the Tertiary as
has already been indicated.

The genus Banlcsia. with three Wilcox species, two of which are
particularly well marked and a probable fruit, Carpolithus protcoidcs,
is confined in the existing flora to the Australian region with about
50 species. The other genus of the tribe Banksieae is Dryandra R
Brown also with about 50 existing species confined to the Australian
region. It is much like Banksia in its foliar characters. Both genera
are found in abundance in the European Tertiary and undoubtedly
enjoyed a more or less cosmopolitan range during the early Tertiary.
Their ancestors probably entered the Australian region during the
Upper Cretaceous before that country had become entirely separated
from Asia, becoming adapted to the peculiar soils and climate of
Australia, while the stock in the northern hemisphere appears to have
been unable to stand the climatic changes and Tertiary competition
and thus became extinct.

The Aristolochiales is placed by some students among the Gamo-
petalae. It includes besides the Aristolochiaceae, the two parasitic
families, the Rarrleisiaceae and Hydneracea?, altogether containing
about 235 existing species, of which 205 belong to the Aristolochia-
ceae, the only family of this order represented in the Wilcox flora.
The genus Aristolochia, to which a typical fruit from the Wilcox is
referred, is found in the American Upper Cretaceous and in both
Europe and America during the Tertiary. There are about 180 ex-
isting species all perennial herbs or climbing vines and widely dis-
tributed in both tropical and temperate regions, about ten species be-
ing found within the United States.

The order Polygonales includes the single family Polygonaceae
with about 800 existing species segregated in about 30 genera, widely
distributed. They embrace herbs, shrubs, vines and trees, with
mostly cyclic flowers, and in their morphological features show some
evidences of transition between the previous choripetalous alliances
and the Chenopodiales. The geologic history of the family is prac-
tically unknown and it would seem that a large part of the specific
variation, particularly of the temperate herbaceous forms, was rela-

I9I4-] SOUTHEASTERN NORTH AMERICA. 165

tively modern. The family is represented in the Wilcox by the single
genus Coccolobis with two species which appear to be the Eocene
prototypes of the only two existing arborescent species of Polygona-
ceae that reach the United States (the sea grape and the pigeon
plum). The genus Coccolobis has about 120 existing species all con-
fined to the American tropics and it would appear that it was of
American origin. These species range from southern Florida to
Mexico, Central America, Brazil and Peru and the majority are
coastal forms. The two modern species which are so much like
these two ancestral forms in the Wilcox, are strand types found
from the Florida keys through the West Indies to the northern coasts
of South America, and the conclusion is almost irresistible that the
Wilcox forms enjoyed a similar range and an identical habitat.

The Chenopodiales (Centrospermse of Engler) include ten fam-
ilies culminating in the Caryophyllaceas, and containing about 3.500
existing species. They appear illy assorted and show a wide range
in floral and other morphological characters. Perhaps a majority
are modern types. The single family Nyctaginacese represents this
order in the Wilcox.

The Nyctaginaceae with about 150 existing species is predomi-
nantly American within the limits of the southern United States on
the north and Chile and Argentina on the south. The genus Pisonia
Plumier, the only genus thus far found in the Wilcox flora, is repre-
sented by three well-marked species. It has about 40 existing spe-
cies chiefly in the American tropics and contains the only arbores-
cent form of the family found within the United States. It has an
extended geologic history, well-marked forms being found in the
European and American Upper Cretaceous. The Wilcox species
were undoubtedly strand types as are so many of the modern species,
which inhabit sea beaches, the shores of salt water lagoons and
marshes, the scrub of beach ridges and the jungle behind them. In
the existing flora it is associated with Pithccolobium, Reynosia,
Metopium, Acacia, Bumelia, Cordia, Coccolobis, Ocotca, Fagara,
Mimusops, Conocarpus, Cassia } Eugenia, Anona, Fiats, etc., exactly
as it was during Wilcox time. Species of Pisonia occur in the
Upper Cretaceous of the Atlantic Coastal Plain (Black Creek forma-
tion) as well as in the Middle (Claiborne) and Upper (Jackson)
Eocene.

16G BERRY— LOWER EOCENE FLORA OF [April 25,

The order Ranales appears to me to be a highly unnatural
assemblage, which doubtless explains the prolonged discussion and
wide range of opinion regarding its true status. As treated in Eng-
ler and Prantl it includes 16 families with over 4,000 existing spe-
cies. While a distinct calyx and corolla are the prevailing habit
this is combined with such primitive features as apocarpy and hypog-
yny, and by a well marked tendency to indefinite repitition and spiral
arrangement of the floral members. I have removed the Lauraceae
which contain l /\ of the existing species to a place in the more evolved
order Thymeleales.

The Ranales as a whole show no close filiation with previous
alliances. They include forms that are more nearly Monocotyledons
than Dicotyledons (Xympha?ace?e) and numerous botanists (c g.
Wieland, Arber, Ilallier) see in them the logical zenith of evolution
of the Mesozoic Cycadophytes and thus as representing the ancestral
stock from which the Angiosperms were descended — apparently a
most remarkable feat, except on paper, when any except floral
features are taken into account. 25

Considering as I do that the Ranalian alliance is a plexus con-
taining unrelated elements, any extended consideration of their
geologic history would be fruitless. Certain forms are well repre-
sented among the oldest known display of Angiosperms in the
Middle Cretaceous. Only two Ranalian families are represented in
the Wilcox flora and these two are both natural groups closely re-
lated and typically Ranalian. I refer to the families Magnoliacese
and Anonaceae.

The family Magnoliaceaa comprises about 70 existing species seg-
regated into nine or ten genera, by far the largest of which is the
genus Magnolia with about 21 species of eastern and southern Asia
southern Mexico and the eastern United States. The family is
mainly tropical and the bulk of the existing forms occur in south-
eastern Asia, the magnolias of that region being largely forms of
tropical uplands.

There are numerous apparent anomalies in the distribution of the
recent forms, thus none are native in Europe, although Magnolia per-

ion of this theory see receni papers by Wieland. Arber and

Parkin, and Hallicr.

i9i4-] SOUTHEASTERN NORTH AMERICA. 167

sisted in that region as late in geologic time as the early Pleistocene.
Only one genus, Drimys Forster, occurs in South America or Austra-
lasia. There is a singular pairing of forms in southeastern Asia and
southeastern North America. For example Magnolia has 14 species in
the former region and seven in the latter: Talauma Jussieu has 3
species in farther India and one in the West Indies : Liriodendron
Linne has a single species in each: Schizandra Michaux has species
in each : Illicium Linne has five species in the former region and two
in the latter. The general Michelia Linne (13 sp.) and Kadsura Jus-
sieu (7 sp.) are confined to the former region and Zygogynum Bail-
Ion is confined to the island of New Caledonia. The leaves of all are
entire and more or less elliptical with a coriaceous texture, often
evergreen, and with a characteristic camptodrome venation. Of the
seven species of Magnolia found within the limits of the United
States, Magnolia glauca Linne ranges northward to Massachusetts
and Magnolia acuminata Linne to New York and Ontario. About
sixty fossil species have been referred to Magnolia. These are
largely based upon leaves, although characteristic fruits, and in at
least two cases, parts of flowers, have been found at various horizons.
Magnolias are very abundant in both individuals and species in the
Middle Cretaceous (Cenomanian-Turonian) especially in North
America, where they are found along the Cretaceous Atlantic Coast
from Greenland southward to Texas and in equal abundance about
the borders of the advancing interior sea represented by the deposits
known as the Dakota sandstone. They are much less common in
Europe and the genus is either of American or Arctic origin. 20

The Eocene records include 4 species of the Arctic region and 13
additional forms largely American, but some few European. The
Oligocene, unrepresented in America by plant beds, has several Euro-
pean species toward its close. About eight Miocene species are re-
corded, of which the majority are American. The Pliocene, also
practically unrepresented by plant beds in America, has furnished 5
or 6 European species and one is found in the early Pleistocene of
that region. Magnolia seems to have been very abundant along the

88 Magnolia Dclgadoi Saporta, Fl. Foss. Port., p. 194, PI. 35, Fig. 5, 1894,
recorded from the Albian of Portugal is almost certainly not a Magnolia.

168 BERRY— LOWER EOCENE FLORA OF [April 25,

shores of the extended Mediterranean sea of the Pliocene and to
have subsequently been entirely exterminated in that region by the
glaciation of the Pleistocene, while surviving in both North America
and Asia by reason of the prevailing north and south trend of the
of the mountain ranges. Some of the other genera of the Magnoli-
ace?e are represented by scattered fossil species but the record is too
incomplete for generalizations. A survey of all the facts leads me
to consider America as probably the original home of Magnolia and
despite the massing of the existing forms in the eastern United States
and their extension to Arctica in the Eocene, they probably originated
in a warm-temperate or subtropical latitude, spread northward
across Arctica to Eurasia, were cosmopolitan in the Tertiary, becom-
ing restricted to the southeastern parts of Asia and North America
by the aridity accompanying uplift, so well illustrated in the Eocene
and later history of the Rocky Mountain and Great Plains province,
and were finally killed off in Europe by the Pleistocene glaciation.

Lesquereux referred two forms from the Wilcox of northern
Mississippi to Magnolia but these both prove to be species of Ter-
minal ia as Lesquereux had surmised in his preliminary studies. The
genus Magnolia is, however, represented in the Wilcox by two large-
leafed species, both of which are common to the basal Eocene of the
Rocky Mountain Province. Neither show any close affinity with the
antecedent Upper Cretaceous forms which are so common in the
embayment area of Alabama and northeastward along the Atlantic
Coastal Plain.

The family Anonacess contains about 700 existing species dis-
tributed among about 48 genera, only two of which are present in
North America. The family is practically confined to the tropics, a
single Australian species and the North American genus Asimina with
6 or 7 species being the only conspicuously extratropical forms. The
area of maximum representation is southeastern Asia and the adjoin-
ing region of Malaysia, for while only 16 genera are confined to this
region it contains over 350 species, and six additional genera (.1/7/-
iusa, Uvaria, Polyalthia, Oxymitra, Mclodorum, and Poporvia) with
a total of over 250 species have the bulk of their species in this area.
Only a single genus is confined to Australia and the bulk of the Aus-

I9I4-] SOUTHEASTERN NORTH AMERICA. 169

tralian species are to be regarded as migrants from the preceding
area. There are upwards of ioo species and 6 peculiar genera in
tropical Africa; and America has about 200 species and 10 peculiar
genera. These are all confined to the tropics except for a species of
Anona which reaches the coast of peninsular Florida and for the
genus Asimina with six or seven species of shrubs and small trees of
the south Atlantic and Gulf States. One of these, Asimina triloba
Dunal, is hardy as far north as New York and has the distinction of
growing the farthest distance from the equator of any existing mem-
ber of the family. The fossil record of the Anonacese is very incom-
plete, only the genera, Anona Linne and Asimina Adanson being
known with certainty. Both of these genera are present in the Wil-
cox flora.

The genus Anona has from fifteen to twenty fossil species five of
which are also represented by seeds. The oldest is a species described
from the Dakota sandstone. There is a second species in the late
Cretaceous or Early Eocene of the Rocky Mountain province. The
flora of the Wilcox affords a glimpse into the true stage of evolution
of Tertiary floras in that expanded belt of the American equatorial
region which was the center of radiation of so many recent types.
There were three exceedingly well marked species of Anona along
the Wilcox coast and their leaves are very common at some localities
although no seeds have as yet been discovered. I assume that these
Wilcox forms had habits similar to those of the majority of the ex-
isting species, exemplified by our Florida Anona glabra Linne, or
Pond Apple, which frequents shallow fresh water swamps, low shady
hammocks, or stream borders near the coast. Other species occur
in the low coppice association or on edges of brackish swamps on the
Bahamas. The cultivated species, as for example the American
Anona reticulata Linne which is planted in Guam often spreads nat-
urally along the inner beaches, while attempts to introduce others of
the most highly esteemed American species in the Orient have failed.
From its prevalence among the existing species the habit of growing
in wet shaded soils is evidently an old one, and since the Wilcox
Anonas are associated with a strand flora, the assumption that they
grew on the inner beaches or the shaded and more swampy edges of
lagoons, possesses every degree of probability.

170 BERRY— LOWER EOCENE FLORA OF [April 25,

In the pipe-clays of Alum Bay which were contemporaneous with
the Wilcox there are two species of Anona, and Engelhardt has de-
scribed two species from the Eocene or Oligocene of Chili. The
Oligocene record shows a species in France and a second in Saxony.
In the Miocene there are two species each in England, Styria and
Croatia and one each in Bohemia, Colorado and Transylvania. There
is one each in the Pliocene of France and Italy, showing how mod-
ern was their extinction in the south of Europe.

The genus - Isimina has only four or five recorded fossil species.
These are all American except for a form from the Pliocene of Italy
which has been referred to this genus, although I suspect that it rep-
resents Anona, since Asimina appears to have originated and been
confined to the Western Hemisphere. The oldest known species is
based on foliage which is found in the basal Eocene of the Rocky
Mountains (Denver formation) and of the embayment (Midway
Group). There is a single species based on a seed from the basal
Wilcox and no other records except a form close to the modern from
the late Miocene of New Jersey (Bridgeton sandstone) and the oc-
currence of the existing Asimina triloba Dunal in the interglacial beds
of the Don valley in Ontario.

The order Papaverales (Rhoedales of Engler) includes six fami-
lies — Papaveraceae, Cruciferae, Capparidaceae, Resedaceae, Tovari-
aceae and Moringaceae, together containing about 255 genera and
2,200 species. The Papaveraceae and Cruci ferae are mostly herba-
ceous and widely distributed, largely in the North Temperate zone,
and they are of relative recent evolution. The Resedaceae is a small
family largely confined to the Mediterranean region. The Cappari-
daceae, Tovariaceae and Moringaceae are mainly tropical, the last two
families consisting respectively of a single genus and two species of
the American tropics and a single genus and three species, one Afric-
Arabian and two East Indian.

The family Capparidaceae with about 35 genera and 400 existing
ies is the only one of the order represented in the Wilcox flora.
A majority of the existing species are herbaceous and they are found
on all the continents in tropical and subtropical regions. Five sub-
families are recognized. Of these the Cleomoidea? and Capparidi-

I9I4-] SOUTHEASTERN NORTH AMERICA. 171

odeae are large and occur on all of the continents, with monotypic
genera in Xorth America (Isomeris), South America (Stubelia
Atamisquca, Bclcnciia), Africa (Pteropetalum, Cladostcmon) , and
Australia (Rocperia, ApophyUum). The subfamily Dipterygioideae
has a single genus with only five or six species of Nubia, Arabia and
the Punjab. The subfamily Roydsioldeae includes about a dozen spe-
cies, the genera Roydsia and Stixis being confined to India and the
genus Forchhammeria being Mexican. The subfamily Emblingioi-
deae has only a single genus and species confined to Western Aus-
tralia. No far-reaching conclusions regarding origin or past history
can be deduced from our present knowledge of the geographical dis-
tribution of the Capparidaceae and the fossil record is so imperfect
that very little can be said regarding this history.

The following are the only fossil records known to me : F. von
Miiller has described somewhat uncertainly determined fruits from
the Tertiary of Australia as the genera Dieune and Plesiocapparis.
The latter has two species and is considered as probably a member
of the section Busbeckia of the genus Capparis. Schenk has de-
scribed the petrified wood of another form from the Tertiary of
Egypt under the name Capparidoxylon. The genus Capparis has
furnished a well-marked Wilcox species very close to the existing
Antillean tree Capparis domingcnsis Sprengel. There are about one
hundred existing species of Capparis, mostly tropical, and although
found in the Eastern Hemisphere the majority occur in the Ameri-
can tropics, especially in Central and South America. The oldest
known fossil forms are two species described by me as species of
Capparites from the Upper Cretaceous of Alabama (Tuscaloosa
formation). In addition to the Wilcox species previously mentioned,
Engelhardt has described a Tertiary species from Bolivia. Many
years ago Unger described a third species from the Middle Miocene
of Styria but Schimper considers the latter to be a papilionaceous
form. While the fossil record of Capparis is so meager such facts
as are available would seem to indicate that it originated in the
American Upper Cretaceous. Very many of the modern forms are
shrubs or small trees of the strand flora and such is believed to have
been the habitat of the Wilcox species.

172 BERRY— LOWER EOCENE FLORA OF [April 25,

The order Rosales includes about eighteen families 27 with over
fourteen thousand existing species, the largest families being those
of the Leguminosse, and the Rosacea?, Saxifragacea? and Crassulacese.
Some members of the alliance are close to the Ranales in their apo-
carpy, hypogyny and the indefinite repetition of certain floral mem-
bers, and the order culminates in the relatively modern Papilionacese.
Five families of Rosales are present in the Wilcox flora. Of these
the three leguminous families are by far the most abundant.

The family Hamamelidaceae consists of about nineteen genera
and fifty species. Twelve of the genera are confined to the Asiatic
region. One genus is doubtfully confined to Australia : Three gen-
era are African: and three genera are common to Asia and eastern
Xorth America. The family is remarkable in containing no less
than nine monotypic genera. A consideration of the existing dis-
tribution is not only of exceeding interest but also conclusive proof
of an extended geologic history, which unfortunately has not yet
been unravelled. Since the group is scarcely if at all represented in
the existing flora of Australia or in its fossil flora, its present range
over Asia would seem to have been accomplished after the land con-
nection with Australia had been interrupted. As the only known
Cretaceous fossil forms are from Xorth America there is a prob-
ability that the group had its origin in the Xorth American region.
The fossil species are not numerous enough, however, for definite
conclusions on this point.

The genus Hamamclis and its generalized fossil type Hamamc-
lites Saporta have five species in the Dakota sandstone, one of which
occurs in the Atlantic coast Upper Cretaceous (Middendorf beds of
South Carolina) and another is doubtfully represented in the sup-
posed Upper Cretaceous of Argentina (Kurtz). There are two
Paleocene species in France and Belgium, and Conwentz has de-
scribed characteristic flowers preserved in perfection in the Baltic
Amber (Sannoisian) as Hamamelidanthium.

The genus Parrotia, with a single existing species of northern
Persia and the Caucasus, has three species in the Dakota sandstone:

27 The family Platanacese, which by the majority of students is referred
to the Rosales, I regard as the sole survivor of an independent order, the
I'latanales, closely related to the L'rticales.

i9i4-] SOUTHEASTERN NORTH AMERICA. 173

one species in the Wilcox and Fort Union: two in the Oligocene of
Europe: and two in the Miocene of Spitsbergen, Spain, France,
Silesia, Austria, and Hungary. The distribution of Parrotia in the
past as far as it is known confirms the evidence derived from Ham-
mamelis for a North American origin for the family.

The third genus with a geological history is Liquidambar, in
which upwards of twenty fossil species have been described. The
oldest known forms occur in the Eocene at such widely separated
points as Alaska, Oregon, Greenland and France. There are two
species in the Oligocene of Asia and Europe. There are nine or ten
Miocene species represented throughout Europe and North Amer-
ica ( New Jersey to Oregon) and in eastern Asia. Three Pliocene spe-
cies are represented in Spain, France, Italy, Germany, Austria, Styria
and Slavonia. Typical fruits preserved in the Upper Pliocene of Ger-
many show how late the genus flourished in central and southern
Europe. Felix has described the petrified wood of Liquidambaroxylon
from the Tertiary of Hungary. The existing Liquidambar styraci-
flua is found in the Pleistocene of West Virginia, North Carolina
and Alabama and the eastern Asiatic species L. formosana occurs in
the Pleistocene of Japan. The genus Corylopsis also occurs in the
post-Miocene deposits of Japan.

The family Rosacea includes about 90 genera and over 1,300 ex-
isting species, widely distributed and mostly in temperate regions.
Some of the genera like Cratcegas seem to be undergoing saltation
at the present time and hundreds of supposed species have been de-
scribed in the past few years. The tribe Chrysobalanoideae is con-
fined to the tropics and the Neuradoideae to the subtropics of Africa
and southwestern Asia. All of the other tribes of Rosacese are
widely distributed and their modern and fossil distribution is with-
out especial significance for the present discussion.

The only genus represented in the Wilcox is Chrysobalanus with
two species that are evidently the prototypes of the still existing
forms. The latter are but two or three in number and as shrubs or
small trees they inhabit the sandy shores in the maritime regions of
Florida, tropical America and western tropical Africa.

The Leguminosa? as now segregated into 4 families constitutes

J74 BERRY— LOWER EOCENE FLORA OF [April 25,

the largest alliance among the Choripetake (Archichlamydeae), and
next to the Composite the largest angiospermous group, with over
9,000 existing species segregated among about 450 genera.

There is a well-defined floral progression from the family Mimo-
sacea' with its actinomorphic flowers and numerous, usually free, sta-
mens, through the Csesalpiniacea?, culminating in the numerically
greatest group the Papilionacese with its strongly zygomorphic flow-
ers and coalescent stamens, comparable with the like culmination in
floral evolution of the Orchidaceae among the Monocotyledons;.

The Mimosaceae, with about 30 genera and 1,400 existing species,
are massed in the tropics of both hemispheres. None of the sub-
families are confined to a single continent but comparatively few
genera occur in more than two continental areas and half the genera
are restricted to a single continent. Asia and Australia each have
two peculiar genera, Africa has four and America has seven. Amer-
ica also leads in number of species, about half the total of the
family being present in the Xew World. Australia comes next with
over 300, Africa next with upwards of 300 and Asia last with about
100. In the eastern United States there are only three genera and
five species, none of which are arborescent. In the Gulf States the
numbers have increased to 14 genera and 44 species.

The Caesalpiniacese with about 90 genera and 1,000 species is also
mainly tropical with a massing of forms in the American tr<
where there are over 600 species and 37 peculiar genera, the sub-
family Sclerolobieae being entirely American and containing numer-
ous monotypic genera. Asia and Africa each have about 150 spe-
cies. There are, however, only 10 Asiatic genera as compared with
17 African. There are but three Australian genera and less than
100 species. In the eastern United States there are 5 genera and
eleven species. Three of the genera. Cercis, Gleditsia and Gymno-
cladus are arborescent. In the southern states there are 11 genera
and 44 species.

The Papilionacea? have about 320 genera and 6.600 species.
America leads in the number of peculiar genera having 82 while
Asia lead- in the number of species with about 1.700. Africa has
47 peculiar genera and about [,600 species. Australia has 38 pecu-

I9I4-] SOUTHEASTERN NORTH AMERICA. 175

liar genera and about 1,000 species. Asia has 33 peculiar genera,
while Europe with 7 peculiar genera and about 700 species is less
rich in both species and genera than any other continent. None of
the subfamilies are confined to a single contintent but some of the
tribes are, the Lipariinae being South African and the Bossiaeinae
being Australian, while 20 of the 27 genera and all but 63 of the 436
species of the subfamily Podolyrieae are Australian. Two genera in
this subfamily are American, 2 African, 1 Asiatic, 1 Mediterranean
(Eurasia) and 1 common to North America and Asia.

In the eastern United States there are 46 genera and 194 species
of Papilionaceas, the genera Cladrastis and Robinia being arbores-
cent. 'In the southern states there are 55 genera and 318 species.
Sargent's " Manual of North American Trees," which includes many
tropical forms of the Florida Keys, enumerates for the Leguminosae
as a whole only 34 arborescent species for North America in 17
genera.

In Grisebach's flora of the British West Indies the Leguminosae
outnumber all other families of flowering plants with 262 species.
The same is true of Urban's flora of Porto Rico where they num-
ber 136 species.

The Leguminosae found in the Wilcox deposits number over fifty
species, many of which are individually abundant. They represent
the families Mimosacese, Caesalpiniaceae and Papilionaceas, the fourth
family of the leguminous alliance, the Krameriaceae, being a small
herbaceous group of the New World of very late, probably of re-
cent, evolution.

Of these fifty-odd Wilcox species eleven are referred to the
Mimosacese, 26 to the Caesalpiniaceae and 20 to the Papilionaceas.
Definitely recognized genera are named in the usual way. Forms
usually identified as species of Acacia (as for example most of those
so named by Heer, Ettingshausen, Unger, etc.) which are referable
to the Mimosaceas but not to the genus Acacia as commonly under-
stood are referred to the form-genus Mimositcs. Forms not cer-
tainly identified as Cccsalpinia but referable to the Caesalpiniaceae are
classed under the form-genus Cccsalpinitcs while a considerable num-
ber of Glcditsia-Yike forms of both leaves and pods are described in

176 BERRY— LOWER EOCENE FLORA OF [April 25,

the genus Gleditsiophyllum, a form-genus proposed by me in the first
instance for an Upper Cretaceous form from North Carolina.
There is a certain unavoidable duplication in the giving of specific
names to unattached pods and leaflets since in some cases they may
belong to the same botanical species. I have followed this method,
however, in all instances where I was not sure of such a relationship.

The Mimosace?e of the Wilcox are referred to four genera.
The genus Acacia represented by a single indisputable species in
which the leaves are reduced to phyllodes is of great interest since
in the existing flora the 450 species are largely confined to the Aus-
tralian region. The section Phyllodineae to which the Wilcox species
is referred has about 300 existing species which are confined to Aus-
tralia and Oceanica although in Eocene times they were also present
in Europe. It is a curious commentary on the modern character of
the earlier Tertiary floras that the reduction of foliar organs and the
habit of phyllody, often correlated with modern arid conditions,
should have really been developed in these early floras.

The genus Inga, represented in the Wilcox by four well marked
species, has upwards of 150 species in the existing flora, all of which
are confined to the American tropical and subtropical regions. Its
geological history is for the most part unknown although it appears
to be represented in American Upper Cretaceous floras by Inga cre-
tacea Lesquereux which occurs in the Dakota Sandstone and in the
Tuscaloosa formation of Alabama. Ettingshausen has described a
species from the Cenomanian of Saxony (Inga Cottai) and the
European Miocene has furnished two or three species, while Engel-
hardt has described a Tertiary species from Bolivia.

In the genus Pithccolobium, which has two W r ilcox species and
belongs to the same tribe as Inga (Ingeae), while the majority of the
100 or more existing species are American there are over a score in
tropical Asia and a few in tropical Australia and Africa. With the
exception of a Tertiary species from Bolivia [do not know of other
fossil occurrences.

The genus Minwsitcs, with four Wilcox species, represents trees
of the Mimosa type very abundant in recent species referred to sev-
eral genera which are cither American. Asian. Australian or African,

I9I4-] SOUTHEASTERN NORTH AMERICA. 177

and abundantly represented in European Tertiary floras. Its Cre-
taceous ancestry is hidden among the species of leaflets referred to
the form-genus Lcgumino sites. The genus Mimosa which is ap-
parently most like the Wilcox Mimosites, has over 300 existing spe-
cies and these are for the most part confined to the warmer parts of
America, although they are represented in Asia, Africa and Aus-
tralia.

Except for the family Lauracese the Caesalpiniacese with 26 spe-
cies is the largest family in the Wilcox flora and it is certainly a fact
of considerable interest that the massing of the modern species in
the American tropics should be foreshadowed by their numerical
abundance on this continent as early as the Lower Eocene.

The Wilcox genera are five in number of which the largest is Cassia
with twelve species. Cassia is the largest Wilcox genus except Ficus >
and all of its species find their modern counterparts in existing species
of tropical and subtropical America, many of which are mentioned by
name in the systematic part of this work. Numerous as are the
Wilcox species of Cassia there was apparently greater specific differ-
entiation in contemporaneous European deposits since Ettingshausen
records 15 species in the flora of Alum Bay (Ypresian of Isle of
Wight). Cassia has between three and four hundred existing spe-
cies found in the warmer temperate and tropical regions of all the
continents and especially abundant in tropical America. Their place
of origin is unknown since they make their appearance in the Upper
Cretaceous almost simultaneously in New Zealand, Australia, Bo-
hemia, Saxony, Greenland, the Atlantic Coastal Plain and the
Dakota Group of the Rocky Mountain province. Upwards of one
hundred fossil species are already known. Nor does the Eocene
distribution shed any light on the early history of the genus since
species occur in such widely separated regions as North America,
Europe and Australia. There are numerous Oligocene and Miocene
species, the Oligocene records being confined to Europe and Africa
and the Miocene records being confined to Europe and North Amer-
ica. Cassia was abundant along the shores of the Pliocene Medi-
terranean of Europe and 4 species are recorded from South American
beds which are thought to be of Pliocene age. Pleistocene species

PROC. AMER. PHIL. SOC, LIII,, 214, L, PRINTED JULY II, I9I4.

178 BERRY— LOWER EOCENE FLORA OF [April 25,

are recorded from Maryland and the East Indies (Java) associ-
ated in the latter region with Pithecanthropus crcctus Dubois. One
fact is certain, the genus has been a part of the American flora since
the dawn of the Upper Cretaceous, and several of the Wilcox species
are the undoubted prototypes of existing forms of the American
tropics.

The genus Cercis, with a single Wilcox species, makes its first
recorded appearance in geological history in the Wilcox species, in
the three species recorded from the Ft. Union deposits of the Rocky
Mountain province and a fourth species found in the Ypresian of
the Paris basin, so that its appearance was practically contempora-
neous in France and Tennessee. It continues on both continents down
to the present being even represented in the Pleistocene of both.
The modern species number five or six and inhabit the warmer tem-
perate regions of America, Europe and Asia.

There is one species of Cccsalpinia in the Wilcox and it is almost
identical in character and habitat with Cccsalpinia bahamensis La-
marck of tropical America. The existing species number about two
score of the tropics of both hemispheres. Cccsalpinia is recorded
first from the Upper Cretaceous of the Atlantic Coastal Plain and it
seems probable that it originated on this continent and reached
Europe during the Eocene by way of the Arctic region, since it is
common in the Oligocene, Miocene and Pliocene of the latter
continent.

Four Wilcox species are referred to the form-genus Cccsalpini-
tcs. These represent true forms of Cccsalpinia or of allied genera
in this family, one almost certainly representing the genus Parkin-
sonia, a small genus which occurs in the European Oligocene but
which in the existing flora is confined to the warmer parts of North
America and South Africa. Fossil forms referred to Cccsalp'uii-
tcs include about twenty of the European Oligocene and Miocene.

The genus GlcditsiopJiyllum makes its appearance in the Upper
Cretaceous of the Carolina region. It is represented by eight species
of leaves, leaflets and pods, often abundantly preserved, in the Wil-
cox deposits. Their relation to modern genera is uncertain, although
they were evidently much like Glcditsia.

19I4-] SOUTHEASTERN NORTH AMERICA. 179

Two genera of Caesalpiniacese which I confidently expected to
find in the Wilcox and which must have been present during this
time in southeastern North America are Hymencua and Bauhinia.
The former is confined to the American tropics in the existing flora
where it has about eight species. It is represented by characteristic
forms in the Upper Cretaceous of Alabama. The genus Bauhinia
which has about 150 existing species of the tropics of both hemi-
spheres has several especially characteristic forms in the Upper Cre-
taceous of southeastern North America (New Jersey, Alary land,
Alabama).

The family Papilionaceae which comprises over two thirds of the
existing Leguminosa? undoubtedly represents the culmination of
evolution in the alliance. The bulk of the family, especially the
numerous herbaceous genera, are unquestionably of comparatively
recent origin. In spite of this fact the family has twenty species
in the Wilcox. These are distributed among six genera, of which
Dalbergitcs, Carpolithus and Leguminosites are form-genera, while
the other three are still existing. The largest genus is Sophora with
seven species, one of which, evidently a strand type similar to and
comparable in habitat with the cosmopolitan strand plant Sophora
tomcntosa Linne of the existing tropical flora, is very abundant in
the Wilcox deposits. There are about 25 existing species of shrubs
and small trees referred to this genus. They are scattered over the
warmer parts of both hemispheres and are found on all tropical sea-
shores. About a dozen fossil species are known. In addition to
North America they are found in both Europe and Asia during the
Eocene, a single form from Alum Bay (Ypresian) being contem-
poraneous with the Wilcox species and the others being later. While
few species have been described the genus is widely distributed in
the European Miocene where Sophora europcca Unger was a com-
mon coastal form of the Mediterranean region throughout the Mio-
cene and into the Pliocene.

Four species, three based on leaflets and the fourth on a charac-
teristic pod, represent the genus Dalbergia in the Wilcox flora. Two
additional species whose generic relations are not so certain are re-
ferred to the genus Dalbergitcs. The existing species of Dalbergia

180 BERRY— LOWER EOCENE FLORA OF [April 25,

number about eighty forms found in the tropics of both hemispheres,
and all show a strong generic similarity in their foliar characters.
Over two-score fossil forms are known. The earliest of these occur
in the Atlantic Coastal Plain and western Greenland so that there is
a strong possibility that the genus was of American origin. If this
theory was correct they must have undergone a rapid radiation since
in the Eocene they are not only found in America and the Arctic
but in Europe and Australia. The Alum Bay beds of the Isle of
Wight (Ypresian) which I regard as contemporaneous, in part at least,
with the Wilcox, contain according to Ettingshausen, six species of
Dalbcrgia. European deposits furnish about a dozen Oligocene spe-
cies and still more numerous Miocene species. Dalbergia primceva
Unger, D. retusafolia Heer, D. hceringiana Ettingshausen and D.
bclla Heer are widespread coastal forms of the European Tertiary,
some of them ranging from the late Oligocene through the Miocene
and into the Pliocene.

The genus Canavalia is represented in the Wilcox by a fine spe-
cies undoubtedly ancestral to the existing Canavalia obtitsifolia
(Lamarck) De Candolle, a widely distributed tropical strand plant.
A second species is less commonly represented and not as certainly
identified. The genus contains about a dozen existing species of the
tropics of both hemispheres but has not been heretofore found in
the fossil state.

The Wilcox forms referred to Lcguminosites cannot be dealt
with satisfactorily since they represent pods and leaflets of this alli-
ance whose generic relations are uncertain. The form-genus was
proposed first by Bowerbank for the pyritized remains from the
Island of Sheppey (London Clay), and two of his species are tenta-
tively identified in the Wilcox. Subsequently many species have
been described. They range in age from the Middle Cretaceous to
the Pliocene. Saporta describes the oldest form in the Albian of
Portugal. They are present in the Cretaceous of Australia, the
Cenomanian of Saxony, the Atane and Patoot beds of Greenland,
and the Atlantic Coastal Plain Cretaceous from Marthas Vineyard
to Alabama. They arc common in the Arctic Eocene, occurring
also in Australia, America, Europe and Asia. Oligocene records in-

i9i4-] SOUTHEASTERN NORTH AMERICA. 181

elude Europe and Antarctica, Miocene records are confined to Amer-
ica and Europe and Pliocene records include southern Europe and
Japan.

While the foregoing analysis leaves a great many points in the
history of the Leguminosae unsolved it serves at least to show that
the Wilcox forms are all represented and would find a congenial
habitat in the present day American tropics and that thus early some
of the main features of their present day development had been
differentiated.

The most similar fossil display of these forms is to be found in
the Ypresian flora of Alum Bay on the Isle of Wight, which un-
fortunately have never been described or figured, but of which
Ettingshausen 2s published an analysis and enumeration in 1880.
Another very similar display of forms is that described by Engel-
hardt from the Tertiary of Cerro de Potosi in Bolivia. 29 The exact
age of the latter has never been determined although its resemblance
to this part of the Wilcox flora suggests the possibility that it is
Eocene instead of Pliocene, which later has been assumed to be its
age. This may, however, simply be a reflection of the similarity
between the Leguminosae of the Embayment area in the Lower
Eocene and those of subsequent periods in the American tropics.

The order Geraniales includes 21 families, with upwards of
ten thousand existing species, of which nearly one half belong to the
family Euphorbiaceae. The other large families in the order of their
size are the Rutaceae, Meliaceae, Malpighiaceae, and Polygalaceae
each with over five hundred existing species, while the Geraniaceae,
Oxalidaceae, and Burseraceae each have over three hundred existing
species. The alliance is mainly cyclic in the character of its floral
members, starting with isocarpic forms and progressing in the direc-
tion of reduction in the number of carpels. The phylogenetic im-
portance of the characters by which the Geraniales as an order is
separated from the evidently allied Sapindales is not great and in
some respects the order is apparently not a natural one. Six fam-
ilies of Geraniales have been recognized in the Wilcox flora. The

28 Ettingshausen, Proc. Roy. Soc. Lond., Vol. 30, 1880, pp. 228-236.
20 Engelharclt, Site, naturwiss. Gesell. Isis in Dresden, 1887, Abh. 5, pp.
36-38, 7 Figs.; Ibid., 1894, Abh. I, pp. 3-13, PI. I.

182 BERRY— LOWER EOCENE ELORA OF [April 25,

first of these, the Rutaceae, consists of about in genera and over
900 existing species widely distributed over the warm temperate and
tropical regions of the earth. The fruits are capsules, samaras or
drupes and the leaves which may be simple or compound are usu-
ally glandular punctate. While there are 34 genera with 127 species
confined to America the family makes its greatest display in the Old
World, Africa having 16 peculiar genera with 196 species and Aus-
tralia 28 peculiar genera with 185 species. In addition to 6 genera
with 7 species confined to the Asiatic Mainland there are 19 genera
with 167 species found distributed from southeastern Asia through
Malaysia greater or less distances, in some cases to Xew Zealand
and Polynesia. The only truly cosmopolitan genus is Fagara with
upwards of 150 existing species and represented in all tropical
countries. The tribe Boronieae with 18 genera and 158 species is
confined to Australia and New Zealand ; the Diosineae with 1 1 genera
and 181 species is confined to South Africa; and the Cusparieae with
16 genera and 83 species is confined to tropical America. All of the
other rather numerous tribes are represented in more than one con-
tinental region.

The family contains the remarkable number of 42 monotypic
genera and while many of these may be regarded as of recent evolu-
tion, as for example a number of those of Australia, the isolated
occurrences of many of the others indicates that they are of great
age and once occupied intervening areas.

There are only twelve known fossil genera, or only about 10 per
cent, of the existing genera, so that little can be said of the fossil
history of the family. The oldest genus is Citrophyttutn Berry rep-
resented by very characteristic leaves with alate petioles found in
the Dakota sandstone of the Rocky Mountain province and from
Xew Jersey to Alabama along the Atlantic coast in the Raritan, Mag-
othy, Middendorf and Tuscaloosa formations. There is a second
species of Citrophylhun in the Wilcox and a third in the overlying
Claiborne. These forms are very similar to the leaves of recent
members of the Aurantioideae and undoubtedly represent ancestral
forms. The genus Dicta mints Linne with a single existing species
widely distributed in Eurasia, has furnished a fossil form in the
Pliocene of France and a second in the Pleistocene of Japan. Unge

I9I4-] SOUTHEASTERN NORTH AMERICA. 183

in 1850 described petrified wood from the Aquitanian of Greece as
Klippsteinia medullaris referring it to the Aurantioidese.

The genus Amyris (P. Browne) Linne has about a dozen exist-
ing species in the Antilles and Central America, two of which reach
the coast of southern Florida. A fossil form is recorded by Unger
from the late Miocene (Sarmatian) of Hungary. This determi-
nation is not conclusive however although Unger had both the leaves
and fruit of Protamyris bcrcniccs. Unger also described the sup-
posed ancestral genus Protamyris to which he referred four species
from the Aquitanian of Kumi and the Miocene of Croatia. These
are not especially convincing and both Ettingshausen and Schenk
consider Protamyris radobojana Unger to represent a species of
Ccdrela.

The genus Xanthoxylum Linne with nine or ten existing species
of eastern Asia and North America has been a favorite receptacle
for fossil forms of Rutacea?. About a score of species have been
described, the oldest coming from the basal Eocene of New Mexico
(Raton formation) while a second Eocene species is recorded from
the Bartonian of France. Engelhardt has described two Eocene or
Oligocene species from Chili. There are four Oligocene species,
two in France and two in Prussia. There are about thirteen Mio-
cene species, widely distributed and represented in California, Colo-
rado, Spain, France, Switzerland, Baden, Bohemia, Croatia and Hun-
gary. The two Pliocene species represent France and Asia Minor
and one of the recent species is found in the Pleistocene of Japan.
It seems probable that Xanthoxylum was derived from Fagara
through a loss of the floral calyx and by adaptation to less tropical
climatic conditions.

The genus Fagara Linne is substituted for Xanthoxylum by many
recent systematists, although I prefer to consider it as the ancestral
stock and in the older sense as including the 150 cosmopolitan trop-
ical species while Xanthoxylum includes the extratropical forms of
Asia and North America. Undoubtedly several if not all of the
fossil forms described as species of Xanthoxylum are more properly
referred to Fagara although none have heretofore been described
under this name. The Tertiary flora of southeastern North America
contains several very characteristic forms of this genus. The oldest

1S4 BERRY— LOWER EOCENE FLORA OF [April 25.

are three species from the Wilcox Group. There is another in the
overlying Claiborne Group. The Yicksburg Group has furnished a
very common form with several well-marked varieties some of the
leaves of which show their glandular punctate character beautifully
preserved. Still another form is found in the Apalachicola Group
of Florida.

The genus Ruta Linne with upwards of 100 existing species
mostly of Eurasia although present in Africa and South America,
has furnished Menzel (1913) with characteristic capsules in the
Aquitanian of Rhenish Prussia.

The genus Phellodcndvon Rupr. with two existing Asiatic spe-
cies is represented in the Aquitanian of Rhenish Prussia by fruits
(drupe). Engelhardt has described species of Ticorca, Pilocarpus
and ErytJirochyton from the early Tertiary of Chili.

The remaining genus with fossil representation is Ptclca Linne
which has 7 or 8 existing species 30 confined to the United States and
Mexico. The fossil forms are represented by both leaves and char-
acteristic fruits. The oldest comes from the Arctic Eocene. There
is a species in the Oligocene of Italy and six Miocene species, occur-
ing in Colorado, France, Switzerland, Carniola and Hungary. A
Pliocene species is recorded from Italy. Obviously the record will
have to become much less fragmentary before any creditable conclu-
sions can be drawn respecting the place of origin and geologic his-
tory of the Rutaceae.

The family Simarubaceae (often spelled Simaroubaceae) in-
cludes about 28 genera and upwards of 150 existing species of
shrubs or trees with pinnate leaves and drupaceous fruits, con-
fined chiefly to the tropics and the warmer parts of the north-
ern hemisphere. Only three of the existing species reach as far
northward as the coast of southern Florida. The family is still rep-
resented on all the continents except Europe. Two genera with
four species are confined to Asia; 3 genera with 4 species are con-
fined to Australia ; 4 genera with 6 species are confined to
Africa and 9 genera with 71 species are confined to America.
The most widespread species is the monotypic Suriana marititna
Linne a cosmopolitan tropical strand plant occurring on the dunes,
keys and coastal hammocks of southern Florida.

30 Greene has recently described very many poorly established new sj

i9i4-] SOUTHEASTERN NORTH AMERICA. 185

The only genus represented in the Wilcox is Simaruba Aublet,
which has furnished a single species, Simaruba eoccnica Berry,
closely resembling the existing Simaruba glauca De Candolle which
is found along tropical coasts from southern Florida to Brazil.

The only other genus with a geological history is Ailantluts Desf.
which has 7 existing species of eastern Asia and the East Indies. 31
The fossil species number about fifteen. There are two in the
Eocene of Wyoming and Oregon ; eight in the Oligocene of France,
Alsace, Styria and Prussia; and five in the Miocene of France,
Switzerland, Baden, Italy and Colorado. While in the absence of
collateral evidence that the Eocene occurrences in North America
have any significance regarding the origin of the genus it is an in-
teresting speculation that the genus originated in North America
and subsequently reached Asia by way of the Eocene land connec-
tion across Behring straits. Certainly the genus lingered on this con-
tinent, as is evidenced by its presence at Florissant, as late as the
Middle Miocene.

The family Meliaceae contains about 42 genera and about 680
existing species of shrubs and trees with pinnate leaves. The vast
majority are found within 30 degrees of the equator although they
reach 40 north in eastern Asia and 40° south in New Zealand.
Moreover the China berry (Mciia azedarach) has been cultivated
from time immemorial in all Mediterranean countries, and through-
out the southeastern United States since its settlement, and is per-
fectly hardy. There are no temperate outliers however. While the
Meliaceae occupy a greater continuous area in South America where
over 41 per cent, of the existing species occur, this large number of
species (about 285) represent only 19 per cent, of the known genera.
There are some remarkable similarities between the species of the
American tropics and those of W T est Africa. Thus the two small
genera Swietenia and Carapa are represented in both areas and
Carapa procera is even said to be common to the two. Moreover the
genus Guarca which has about 80 species in the American tropics has
three in West Africa. The larger number of genera are found in
the S. E. Asiatic region and the number of genera and their mutual

31 Aliaiitlwphyllum Dawson with a single species is described from the
Eocene (?) of British Columbia.

186 BERRY— LOWER EOCEXE FLORA OF [April 25,

affinities decrease from Asia toward Africa and also through Poly-
nesia. A number of genera (Toona, Xylocarpus, Cipadcssa, Mclia)
extend from Africa through Asia to Malaysia. Two genera are
peculiar to Australia (Synoum, Ozi.'cnia) and two to Polynesia {Va-
;vu. Meliadelpga). There are thirteen monotypic genera of which
six are African and seven Asiatic. From the distribution of the
existing species De Candolle 3 - infers that southern Asia is the center
of radiation of the family. I am inclined to think however that the
reverse is probably true since the oldest known forms, except the
entirely doubtful Cedrelospermites of Saporta from the Valanginian
of Portugal, are American, and the widespread existing American
representation of the family seems to comprise the specifically mul-
tiplied descendents of the original stock already represented in the
Wilcox flora.

The Asiatic genera would represent immigrants into that area or
forms evolved there. The Polynesian and Australian forms are
much localized derivatives of the Indian stock and unless the peculiar
species of Xew Caledonia could not reach that region except by a
land connection it may be inferred that this Asiatic radiation was
relatively recent.

The fossil species are unfortunately few in number. So far as
I know the only fossil species of Carapa is that found in the Wilcox.
Its occurrence in the early Eocene is at least a factor in explaining its
present distribution in both the American and West African tropics.
The fact that Carapa procera is common to these two areas may sug-
gest that all of the African species are recent immigrants, but it is
more probable that there are unrecognized specific differences in this
form in the two areas and that the present disconnected distribution
is an example of survivors from the early Tertiary radiation. An-
other genus with a modern distribution like Carapa is the genus
Moschoxylon Jussieu (made a section of Trichilla Linne by Harms
in Engler and Prantl) which has about 60 species of tropical Amer-
ica and West Africa. This has furnished two fossil species de-
scribed by Engelhardt from the early Tertiary (Eocene or Oligo-
cene) of Chili. The genus Ccdrcla, sometimes made the type of an

32 De Candolle, C. de, " On the Geographical Distribution of the Mclia-
cc-.x." Trans. Linn. Soc. Lond.. 2 scr., Bot, Vol. I, 1880, pp. 233-236, PI. 30, 31.

I9I4-] SOUTHEASTERN XORTH AMERICA. 187

independent family, the Cedrelaceae, has four Wilcox species, Eocene
prototypes of existing American species. This genus with 9 or 10
species is confined to America in the existing flora and is only known
outside this area in two species from the Miocene of Croatia which
Unger referred to Cedrela and an undescribed Cedrcla recorded by
Ettingshausen from the Ypresian of the south of England. Saporta
has, however, recorded six species of Cedretospermum from the San-
noisian of southeastern France. The fossil record of these three
genera Carapa, Moschoxylon and Cedrela, brief as it is, shows clearly
that the Meliacese are not a modern element in the flora of the Amer-
ican tropics but one that was already well differentiated in the early
Tertiary.

The remaining fossil references to this family comprise Meliacece-
carpum based on capsules from the Aquitanian of Prussia which
Menzel their describer compares with those of the genera Dysoxy-
lum and Guarea. F. von Miiller has described Rliytidotheca and
Ple'wclinus, two supposed meliaceous genera based on fruits, from
the Pliocene of Australia.

The family Humiriaceas is a small one, comprising only three
genera and a score of species of shrubs and small trees all of which
are confined to the American tropics except a single species found in
tropical West Africa, a distribution suggesting a history comparable
with that just suggested for Carapa, Moschoxylon and Cedrela. The
only known fossil species is one from the Wilcox very close to the
existing Vantanea paniculata Urban of northern South America.

The family Malpighiacese, confined to tropical and subtropical
countries, contains about 55 genera and 650 existing species, many of
which are scandent, including some of the finest lianas of the tropics
with stems 2 dcm. in diameter. Others are shrubs and trees. The
leaves are opposite and simple and the fruits drupaceous, capsular,
or nutlike, and often winged. The only species that reaches the
United States is Byrsonima lucida (Swartz) De Candolle, a small
evergreen tree of the Florida keys.

The family is predominantly American in its distribution, over
67 per cent, of both genera and species being confined to the West-
ern Hemisphere (37 genera and 440 species). The genera are all
local in the sense that none occur in more than one continental area.

188 BERRY— LOWER EOCENE FLORA OF [April 25,

Of the two subfamilies into which the family is divided — the Pyra-
midotorce and Planirotse, the latter with two tribes Galphimiese and
Malphighieae are entirely American. Of the three tribes into which
the Pyramidotorae is divided the Tricomariese are entirely American,
the Hiraeeae have 3 genera and 23 species confined to Asia, 3 genera
and 12 species confined to Africa, a genus with 12 species ranging
from Malayasia to Australia, and 9 genera with 151 species confined
to America. The remaining tribe, the Banisteriese, has a monotypic
genus in Asia, 2 genera and 15 species in Africa, a single genus with
7 species ranging from the East Indies to Australia and 11 genera
with 247 species confined to America.

There are 21 monotypic genera distributed as follows: Micro-
stcira confined to Madagascar : Flabcllaria confined to Africa: Can-
canthus confined to Arabia: Brachylophon confined to farther India:
Mezia, Diploptcvis, Lophopteris, Clonodia, Colcostachys, Blcpliaraii-
dra, Loplianthera, J'cmtcalaria, Ptcrandro, Acmanthera, Diacidia,
and Glandonia confined to Brazil, Guiana and Venezuela: Henlco-
phytutn confined to Cuba: Lasiocarpus and Echinoptcris confined to
Mexico: and Tricomaria and Mionandra confined to Argentina.

Monotypic genera in general are susceptible of two interpreta-
tions, i. e., they represent either the last survivors of a long line a> in
the case of the Ginkgo, Sassafras, etc., or they represent relatively
recent specializations. In the case of the foregoing monotypic
genera it seems probable that the majority are the result of relatively
recent evolution since there is nothing in their character or distribu-
tion to suggest any extended geologic history and none have been
found in fossil floras.

The fossil record is most incomplete. Xo forms are known
from the L'pper Cretaceous for although Ettingshausen recorded a
species of Malpighiastrutn and one of Banisteriophyttum from the
Upper Cretaceous of Australia, those identifications are open to the
most serious question and I do not consider them of any weight in a
discussion of this kind. The family is certainly represented in the
lower Eocene by five species of Malpighiastrutn, Hiraa and Banis-
teria in the Ypresian of the south of England and by five species of
Ilircca and Banisteria in the Wilcox flora, based upon both leaves and
characteristic fruits. There are also doubtful species of Malpighias-

i9i4.] SOUTHEASTERN NORTH AMERICA. 189

trum and BanistcriophyUum described from the Eocene of Australia
by Ettingshausen. Thus there is no direct geologic evidence of the
place of origin of the family. The fact that it is so predominantly
American at the present time and that only two genera have reached
Australia from the East Indian region and that two of the American
genera appear in the northward extension of the early Eocene flora
of the American tropics during the Wilcox and are as ancient as any
certain records of the family anywhere, renders the conclusion that
the family originated in equatorial America an extremely probable
one. With the exception of the Wilcox records enumerated above
nearly all of the fossil records relate to Europe, and these may be
briefly enumerated.

The genus Malpighiastrum of Unger has about 30 recorded spe-
cies. These include the doubtful Upper Cretaceous and Eocene spe-
cies previously mentioned as recorded by Ettingshausen from east-
ern Australia; 3 Ypresian species from the south of England; eight
Oligocene species in France, Italy, Dalmatia, Styria and Transyl-
vania; about 15 Miocene species in Italy, Prussia, Bohemia, Croatia
and Transylvania ; and two Pliocene species in Italy.

The genus Heteropteris Jussieu, with about 90 existing species
ranging from Mexico and the Antilles to Bolivia and Brazil, has a
late Oligocene species in Transylvania and two Miocene species in
Styria and Croatia.

The genus Hircca Jacq., with about 25 existing species ranging
from Mexico and the Antilles to Peru, has furnished about ten fossil
species, based for the most part on the winged fruits. There is a
species in the Ypresian of southern England and a characteristic
fruit in the Wilcox; four Oligocene species in the Tyrol, Styria and
Transylvania ; three Miocene species in Baden, Styria and Transyl-
vania ; and a Pliocene species in Brazil.

The genus Tetrapteris Cav., with about 60 existing species rang-
ing from the West Indies and Mexico to southern Brazil and Bolivia,
has furnished a fossil species in the Oligocene of Styria and three
Miocene species in Bohemia, Styria and Croatia.

The genus Stigmatophyllon Jussieu, with about 45 existing spe-
cies found in the Bahamas and Antilles and along the east coast of
America from Mexico to Uruguay, has furnished Saporta with a

190 BERRY— LOWER EOCENE FLORA OF [April 25,

somewhat doubtful form from the Upper Oligocene of France.
Similarly the genus Byrsonima L. C. Rich., with 90 existing species
ranging from the Bahamas and Mexico to southern Brazil and
Bolivia has been recorded by Massalongo from the early Pliocene of
Italy, but the identification is extremely doubtful.

The genu? Banisteria Linne contains about 70 existing species
of climbing or scrambling shrubs ranging from the West Indies
throughout tropical South America. It is represented by four spe-
cies based upon both leaves and fruits in the Wilcox ; there is a
Ypresian species in the south of England ; four Oligocene species
in France, the Tyrol, Alsace and Styria; and four Miocene species
in France, Switzerland and Croatia.

The genus Banisteriophyllum Ettingshausen with a single Upper
Cretaceous and an Eocene species in eastern Australia I regard as of
very doubtful affinities. Schenk also states that wood of a malpighi-
aceous type occurs among the silicified woods from the Oligocene
of the Island of Antigua.

The family Euphorbiaceae is sometimes made the type of a dis-
tinct order, the Euphorbiales, although the significance of the char-
acters by which it is segregated from the Geraniales is not obvious.
It is an exceedingly large alliance with about 220 genera and 4,000
existing species (Pax, 1890) of herbs, shrubs and trees, widely dis-
tributed throughout the tropical and temperate zones. The genus
Euphorbia with over 700 species is perhaps the most widely distrib-
uted genus in the family. A very large number of the recent species,
particularly those of xerophytic character so closely simulating the
Cactacea?, are of relatively recent evolution.

In such a multiplicity of existing genera and species any effort to
trace the larger features of distribution would occupy more space
than it is worth in the present connection. Four arborescent genera
with five species reach the United States in the Florida region, and
several additional are naturalized in that area. A considerable, but
relatively insignificant, number are recorded during the Upper Cre-
taceous and Tertiary. The fossil records will, however, have to be
greatly increased before they can be said to shed any definite light
on the geological history of the family. Enough is now known.

I9U-] SOUTHEASTERN NORTH AMERICA. 191

however, to abrogate the statement made by Schenk 33 and quoted by-
Pax 34 that there is no certain evidence of the existence of the Euphor-
biaceae during the Tertiary. The following genera have been re-
corded as represented in the fossil state.

Euphorbia with a single species based upon a fruit described by
Heer from the Swiss Miocene ; Euphorbioides based on an inflores-
cence described by Wessel and Weber from the Aquitanian of Rhen-
ish Prussia; the genus Euphorbiophylhim with several species to be
noted presently; I have described a very characteristic species of
Manihotites from the Upper Cretaceous of Georgia; the genus Cro-
tonophyllum has several Upper Cretaceous and Eocene species;
Cluytia is reported from the Eocene of the Isle of Wight and the
Oligocene of Saxony and Rhenish Prussia ; the following genera
each with a single species were identified by Ettingshausen from the
Miocene of Bohemia, i. e., Adenopcltis, Baloghia, Omalanthus and
Pliyllanthus. Conwenz has described a euphorbiaceous flower from
the Baltic Amber (Sannoisian) as Antidesma maximozviczii and
Felix has described petrified wood from the Tertiary of the U. S. of
Columbia as EupJwrbio.vylon. Hura-like fruits are also recorded
by Knowlton from the lower Eocene (Raton formation) of New
Mexico. Engelhardt has recorded species of Omphalea Linne, Tet-
raplandra Baillon and Mallotus Lour., from the early Tertiary of
Chili.

While difference of opinion regarding the determination of some
of these records is justifiable I regard Manihotites, Euphorbiophyl-
him, CrotonopJiyllum and Euphorbioxylon as definite evidence of the
existence of the Euphorbiaceae during the Upper Cretaceous and
Tertiary.

The Wilcox species are five in number and are referred to the
genera CrotonopJiyllum, Euphorbiophyllum and Drypctes. The
genus Crotonophyllum was proposed by Velenovsky for a well-
marked species from the Cenomanian of Bohemia. I have described
a second species from the Upper Cretaceous of South Carolina. Two
species are recognized in the Wilcox and of these Crotonophyllum

33 Schenk, " Palseophytologie," pp. 594-597, 1890.

34 Pax, in Engler and Prantl's " Naturlichen Pflanzenfamilien," 1890.

192 BERRY— LOWER EOCENE FLORA OF [April 25,

eocenicum Berry may be successfully compared with a number of
the six hundred existing species of Croton which is so abundantly
represented in tropical America. Comparisons are especially close
with Croton eluteria (Linne) Bennett which is found in the low cop-
pice of the beach ridges throughout the Bahama islands.

The genus Euphorbiophyllum was proposed by Ettingshausen in
1853 for several species from the Sannoisian of the Tyrol. Al-
together over a dozen species have been described by Ettingshausen,
Saporta and Engelhardt. These have been compared with the ex-
isting, mostly tropical American, species of Styloccras, Sapium, 57/7-
lingia, Adenopeltis, Exoecaria, Colliquaja, etc. The oldest comes
from the Cenomanian of Portugal and a second Upper Cretaceous
species occurs in the Turonian of southern France. In the Eocene there
is a species in West Greenland, a second on the Island of Sheppey
( Ypresian) and a third in the Paris basin (Lutetian). Five Oligo-
cene species have been described from the Sannoisian of the Tyrol,
and a sixth from the Chattian of northern Bohemia. There are two
Miocene species in Switzerland and two in Styria : a Pliocene species
is described by Krasser from Brazil. A single small-leafed species
of Euphorbiophyllum is of rare occurrence in the middle Wilcox.

The genus Drypetes Vahl. has about a dozen existing species con-
fined to tropical and subtropical America. Three extend southward
to northern Brazil and two range northward to the Florida keys.
There are two well-marked species in the Wilcox flora — one an
Eocene prototype of the existing Drypetes keyensis Urban, and the
other of the existing Drypetes lateriflora (Swartz) Urban, both small
trees of the coastal flora of southern peninsular Florida, the Bahamas.
West Indies and Antilles. The genus, which has not previously been
recorded in the fossil state, was probably of American origin and
there is no evidence that it ever spread to the eastern hemisphere.

The order Sapindales, sometimes called the Celastrales, includes
some twenty families, together containing about 3,200 species, the
largest families in the order of their size being the Sapindacea; which
has more than twice as many species as any of the others ; the Celas-
traceae, Anacardiacea?, Balsaminacea?, and Ilicacese. As in the pre-
ceding order, the Sapindales start with isocarpic forms and pass to

19I4-] SOUTHEASTERN NORTH AMERICA. 193

those in which the carpels are reduced in number, and in the more
evolved families the flowers have become zygomorphic. Since there
are several distinct lines of development and the separation from the
Geraniales is based on characters that seem trivial, it seems probable
that the families comprising these two orders as at present under-
stood represent a plexus of forms whose filiations are not yet
understood.

The first family of the Sapindales that is represented in the
Wilcox flora is the Anacardiacese, an exceedingly natural group. It
contains about 58 existing genera and 435 species of shrubs and trees
with round pithy branches, resinous and frequently toxic juice,
alternate, simple, palmate or pinnate, exstipulate leaves, and dru-
paceous fruits with exalbuminous seeds. The Anacardiacese makes
its greatest display in the tropics and subtropics of both hemispheres
but in the existing flora is especially characteristic of the Malaysian
region. Rhus is by far the largest genus and the only one of the
family found in the extra tropical regions of both the northern and
southern hemispheres. The present geographical distribution shows
many anomalies throughout the family. Thus the genus Campo-
sperma Thwaites has eight species in Madagascar, Ceylon, Sumatra,
Borneo and the Malaccas and a single species in northern Brazil.
The genus Sorindeia Thouars of tropical Africa and Madagascar is
most closely allied to the genus Alauria Kunth of the Andes of South
America. The genus Calcsium Adanson has 13 species in tropical
Africa and one in the East Indies. The Eurasian genus Pistacia
Linne has a single species in Mexico. The genus Thyrsodium Ben-
tham has 4 species in the Amazon region of South America and one in
tropical West Africa. The subfamily Mangiferae with about 80
species is entirely Malaysian except for a species of Glutei Linne in
Madagascar and the genus Anacardium Linne which is confined to
tropical South America, chiefly in Brazil. The subfamily Spondiea?
is found in the tropics of all the continents, excepting Europe. The
subfamily Rhoidea? is found on all the continents and shows a pairing
of a considerable number of genera in equatorial Africa and America.
The two remaining subfamilies, the Semecarpere and the Dobineea?
are restricted to the region extending from India to Australia. The

PROC. AMER. PHIL. SOC, LIII, 214, M, PRINTED JULY 13, I914.

194 BERRY— LOWER EOCENE FLORA OF [April 25,

family contains twenty monotypic genera distributed as follows :
Asia 5, Australia 3, Africa 6, Madagascar 3, North America 2, and
South America 1.

The fossil records of the Anacardiacea? are very incomplete
although there seems to be no doubt that it was represented in both
Europe and North America as far back as the Upper Cretaceous.
As in the existing flora the most abundant genus in the fossil record
is RJius to which over one hundred species have been referred. Eight
of these are Upper Cretaceous forms the oldest coming from North
American strata correlated with the Cenomanian (Raritan, Dakota).
The genus appears in Europe in the Turonian of Bohemia. There
are over a dozen Eocene species of Rhus, widely scattered. Thus
there are three in the Ypresian of Alum Bay, four in West Green-
land and North American species in the Lance, Kenai, Ft. Union and
Green River formations. The genus doubles its known species in the
early Oligocene, being especially well represented in southern France,
but also recorded from the Tyrol, the Baltic amber, Italy, Carniola
and Styria.

In the Miocene Rhus seems to have been as abundant, as well
differentiated, and as widely distributed as it is in the existing flora,
for over sixty fossil species have already been described. The
records embrace all European countries where Miocene plants have
been found as well as Iceland and the following North American
localities: — Maryland, Virginia, Colorado, Yellowstone Park, Idaho,
Nevada, Oregon and California. Only a small number of Pliocene
species are known and these are recorded in Spain, France, Italy,
Germany and Slavonia.

Three Pleistocene species are recorded, 2 from Japan and one
from China, all closely related to still existing species of that region.
Fngler 35 some years ago reviewed the geological records of Rhus
and concluded that most of the then known fossil species belonged to
the section Trichocarpne (in the existing flora with over a score of
species mostly confined to North America and eastern Asia), or the
section Gerontogere (with 75 existing species mostly confined to
South Africa). A few fossil forms he considered as representing
the section Venenata?, which has about 14 existing species in North

Engler, A.. Bot. Jahrb., Bd. 1, 1881, pp. 413-419.

I9U-] SOUTHEASTERN NORTH AMERICA. 195

and South America. The other sections into which the genus is sub-
divided were not recognized among the fossil forms.

The allied genus Cotinus with two or three existing species in
Eurasia and North America is probably represented by some of the
fossil forms referred to Rhus, c. g., Saporta considers Rhus anti-
lopum Unger from the Aquitanian of Kumi as a species of Cotinus.
This author has also described Cotinus palccocotinus, and Cockerell
has described Cotinus f rater na from the Miocene of Florissant,
Colorado.

The genus Pistacia with five existing Mediterannean species and
one each in eastern Asia and Mexico has about fifteen known fossil
species the oldest, of doubtful value, coming from the Raritan of
Staten Island. A second Cretaceous species is found in the Laramie
of Colorado. Europe enters the record with a Ypresian species from
Alum Bay. There are three Oligocene species in France and seven
Miocene species in France, Bohemia, Styria, Galicia, and Transyl-
vania. There is a Pliocene species in Styria, an extinct Pleistocene
species on the Island of Madeira, and the existing Pistacia lentisens
Linne in the Pleistocene of the Island of Santorin.

The genus Anacardites Saporta (Anacardio phyllum) has been
used as a form-genus for fossil Anacardiaceas of uncertain generic
relationship. As used by Saporta it represented fossil forms resem-
bling existing species of Mangifera, Anaphrenium, Spondias, Como-
cladia, Holigarna, etc., but not determinable with certainty. Heer
has described a supposed species of Anacardites from the Atane
beds of West Greenland. There are two species in the Sparnacian
and one in the Ypresian of France and seven well marked species in
the Wilcox. There are two or three Oligocene species in France
and Germany and two or three Miocene species in France and
Styria. Felix has described petrified wood from the Eocene of the
Caucasus which he refers to Anacardioxylon, a type also repre-
sented in the Oligocene of Antigua in the American tropics (species
compared with existing genus Spondias).

The floral genus Heterocalyx Saporta {Trilobium Saporta, Ela-
phrium Unger, Getonia Unger) which occurs at a number of horizons
in the Oligocene of France, Croatia, and Styria is represented by a
second species in the Wilcox. Saporta compared it with the South

196 BERRY— LOWER EOCENE FLORA OF [April 25,

American genus Astronium but Engler (op. cit.) considers it most
like the .Malayan genus Parishia.

The genus Metopiwm, not certainly recognized heretofore, has a
well-marked species in the Wilcox. Several Tertiary woods are
described by Linger as Rhoidium and Saporta has described a species
of Scliiiuts from the French Oligocene (Gargas) which is wrongly
determined according to Schenk (p. 541).

The genus Spondicecarpum has a species in the early Eocene of
France and a second in the Aqnitanian of Rhenish Prussia. Recently
Fritel has described leaves from the Aquitanian of France which he
calls Semecarpites that are very close to the existing genus Seme-
carpus which has about 40 species ranging from India to Australia.

The family Ilicacese (Aquifoliace?e) is a relatively small one com-
prising only five genera and about 180 existing species. They are
shrubs or trees with alternate, simple, entire or toothed, often cori-
aceous leaves. The flowers are small, dioecious and hypogynous.
The fruit is a drupe with a thin fleshy -arcocarp enclosing as many
crustaceous nutlets as there are carpels. The genus Ilex Linne to
which all but seven of the existing species are referred is found in
all tropical and temperate regions of the world except western
Xorth America, Australia, New Zealand and Xew Guinea. The
remaining genera of the family are Oncotheca Baillon with a single
species in Xew Caledonia, Nemopautlics Rafinesque with a single
species in temperate Xorth America, Sphenostemon Baillon with two
species in Xew Caledonia, and Byronia Endlicher with three specie-,
one in Tahiti, one in the Hawaiian Islands and one in Australia.
This modern distribution is a certain indication that the family has
an extended geologic history.

( )ver a hundred fossil species have been referred to the genus
Ilex. At least thirteen species are recorded from the Upper Cre-
taceous. All but one from the Turonian of Bohemia are from the
western hemisphere and include two in the Raritan formation, three
in the Magothy formation, -even in the Dakota sandstone, one in the
Atane and two in the Patoot beds of western Greenland.

There are about fourteen Eocene species including four in the
Wilcox, one in the Ypresian of England, one in the Fori Union, four
in the Green River beds, five in Greenland, one in Alaska and one in

igi4.] SOUTHEASTERN NORTH AMERICA. 197

Australia. There are over a score of Oligocene species including one
from Chili that may even be of Eocene age. The lower Oligocene
or Sannoisian has eleven species in France, Tyrol, Saxony and Prus-
sia and includes three species of flowers described by Caspary from
the Baltic amber. The Middle Oligocene or Tongrian has six species
in France, Italy, Germany and Styria and there are seven species in
the Upper Oligocene (Chattian) of France, Bohemia and Greece.
Upwards of fifty species have been described from the Miocene of
Europe and Asia, and from New Jersey, Colorado and California in
this country. The most prolific Miocene area is that of France.
About ten species are known from the Pliocene of Spain, France,
Italy, Prussia and Asia Minor. One fossil and four recent species
are found in the Pleistocene of Virginia, North Carolina, Alabama,
Kentucky and the Island of Madeira. In addition to the fossil
forms referred to Ilex, two Miocene species from Italy and Styria
are referred to the genus Nemopanthes and four forms from the
late Oligocene or the Miocene of Prussia, Styria, Croatia, Bohemia
and Greece are referred to the genus Prinos Linne, which is usually
considered a section of Ilex. The four species from the Wilcox that
are referred to Ilex are represented in the collections by a small
amount of mostly poor material and are without special significance.

The family Celastraceae includes about 40 genera and upwards
of 400 existing species of trees and shrubs with opposite or alternate,
simple, persistent or deciduous leaves and capsular or drupaceous
fruits. The three large genera Euonymus, Celastnts and Gymuo-
sporia are practically cosmopolitan and several additional genera
localized in the modern flora were cosmopolitan in the Tertiary.

The following 12 genera with over 100 species are confined
to America: Fraunhofcra, Mortouia, Glossopetahtm, Schaefferia,
Goupia, Maytenus, Pachystima, Ziiiozvic-wia, Plenckia, Wimmcria,
Gxminda, Rlwcoma. The genera Glyptopctolum and Tripterygium
together with five species are confined to Asia. The genera Hypso-
phila, Dcnliamia and Hedra'xanthcra together with seven species are
confined to Australia. And the following ten genera with about 60
species are confined to Africa or Madagascar: Putterlickia, Catha,
Ptcrocclastrus, Polycardia,Ptelidium } Gassinc, Elccodcndron, Mauro-
ccnia, Schvchcra and Lauridia.

198 BERRY— LOWER EOCENE FLORA OF [April 25,

The family is definitely represented in the Cretaceous by at least
five genera and is an important element in most Tertiary floras.
The oldest known genus is the form-genus Celastrophyllum proposed
by Gceppert. Five well-marked species occur in the Patapsco for-
mation (Albian) of Virigina and Maryland. At the base of the
I'pper Cretaceous, particularly in North America, a large number
of species occur. I Fpward of thirty have been described of which
number two are recorded from New Zealand and two from the
Cenomanian of Niederschcena in Saxony. There is a species in the
Atane beds of Greenland and three in the Patoot beds. The re-
mainder occur in the United States and have the following distribu-
tion : ten in the Raritan formation of New Jersey and Maryland,
twelve in the Tuscaloosa formation of Alabama, two in the Magothy
formation of New Jersey and Maryland, two in the Middendorf
beds of South Carolina, seven in the Dakota sandstone, and two in
the Black Creek formation of North Carolina. There are ten Eocene
species — seven in the basal Eocene of Belgium, one in the Ypresian
of England and two in the Claiborne group of the Mississippi embay-
ment. There are five Miocene species in Italy, Bohemia and Styria ;
a Pliocene species in Italy ; and four Tertiary species from the
Island of Java. Another form-genus is Cclastrinitcs Saporta which
has four species in the Paleocene of France, one in the Denver for-
mation of Colorado, another in the Livingston formation of Mon-
tana, and a seventh in the Miocene of Florissant, Colorado.

The genus Cclastnis Linne is the largest fossil genus of the
family and its history shows that while its present center of distribu-
tion is in the uplands of southeastern Asia and the East Indies the
ancestral stock was cosmopolitan and very abundant in the Tertiary
of America and Europe, with a strong probability that it originated
in the former area at the dawn of the Upper Cretaceous or some-
what earlier. The oldest known species, Cclastnis arctica Heer, is
found in the Raritan and Magothy formations of New Jersey and
Maryland and in the Patoot beds of Greenland. No less than thirty
species of Cclastnis have been described from the Eocene. These
include six Ypresian species from England, five species in the Wil-
cox flora, one in the Denver, ten in the Fort Union, one in the
Kenai of Alaska, three from Greenland and four from Australia.

I9I4-] SOUTHEASTERN NORTH AMERICA. 199

There are also about thirty Oligocene species, all European, and in-
cluding remains in the Baltic amber, in France, Switzerland, Ger-
many, Austria-Hungary and Greece. There are at least a dozen
species in the Chattian of Bohemia. Over fifty Miocene species have
been described ranging throughout Europe, in eastern Asia, and in
Virginia, Colorado, Idaho and Oregon in this country. About a
dozen Pliocene species have been described from Spain, France,
Italy and Sicily.

The genera Cassine Linne and Pterocelastrus Meissner both now
confined to South Africa and Madagascar, each has a fossil species
in the Miocene of Bohemia. The genus Pachystima Rafinesque, with
two existing species in North America, has an Upper Cretaceous
species in North Carolina and a Miocene species in Colorado.

The genus Maytenus Feuill, with about 70 existing species of
the tropics and subtropics of South America has a well-marked
species in the Wilcox flora. There are two species in the early
Tertiary of Chili, one in the late Oligocene and three in the Miocene
of southeastern Europe.

The monotypic genus Gyminda Sargent confined to Florida and
the West Indies in the existing flora has a doubtfully determined
fossil species in the Magothy formation of the Atlantic Costal Plain.
The genus Microtropis Wall., with 9 or 10 existing species of the
mountains of southeastern Asia from India to China and Japan, has a
doubtfully determined form in the early Pliocene of Italy.

A well preserved flower in the Baltic amber is described by
Conwentz as Celastriiiaiitliium Hauchecomei.

The genus Elccodendron Jacquin, with about 25 existing species
confined to South Africa, has a considerable geologic history. Four
Upper Cretaceous species have been described — one from Australia,
one from the Dakota sandstone, and two from the Magothy forma-
tion of the Atlantic coast. There are six Eocene species showing
that the genus was represented in New Zealand (?), Australia (?),
Alaska, the Ypresian of England and the Fort Union of the Rocky
Mountain region. There are five Oligocene species in the Tyrol,
Bohemia and Transylvania ; nine Miocene species in France, Switzer-
land, Italy, Prussia, Bohemia and Styria ; and four Pliocene species
in Italy.

200 BERRY— LOWER EOCENE FLORA OF [April 25,

The remaining genus known in the fossil state, Eitoiiymiis Linne,
has about sixty existing species widely distributed throughout the
northern hemisphere but most numerous in the Asiatic tropics and in
China and Japan. Upwards of thirty fossil species are known, being
based upon both fruits and leaves. There are four well-marked
Eocene species all of which are confined to North America where
they are represented in West Greenland, in the Fort Union and
Green River beds of the Rocky Mountain region, and in the Wilcox
of the Mississippi embayment. The species of the latter region is a
very abundant and characteristic form. There are four or five
Oligocene species of Euonymus recorded from Bavaria, the Tyrol
and Bohemia. The twelve Miocene species occur in France, Prussia,
Bohemia, Styria, Croatia, and Hungary. There are four Pliocene
species in Germany, Italy, and Slavonia ; and two still existing species
occur in the Pleistocene of France.

From this very brief survey of the fossil history of the Celas-
tracea? it is seen that there is a probability, similar to that shown by
so many other families of Dicotyledons, that the ancestral stock
originated in the western hemisphere.

The family Sapindaces consists of about 118 genera and over one
thousand existing species of trees or shrubs with alternate, pinnate,
exstipulate, persistent, or deciduous leaves and drupaceous or cap-
sular fruits with crustaceous mostly solitary seeds. About one third
of the genera are lianas. Most of the family is confined to tropical
and subtropical regions and about 23 per cent, of the genera (27)
and 34 per cent, of the species (345) are confined to America.
There are more genera (30) confined to the African region but only
about one fifth as many species (75).

The genera Cardiospermum, Schmidelia (Allophylus) and Sapin-
dus are found in all tropical countries. The genus Paullinia with
over 120 existing species while mostly American is present in Africa
and Madagascar. The genus Dodomca with over 40 species in Aus-
tralia has one or two forms found in all tropical countries and a
single species in the Hawaiian Islands and Madagascar. Harpullia
is common to Asia, Africa and Australia. There arc two genera
with about fifteen species confined to Australia, four genera with 66
species ranging from Asia to Australia, to genera with 22 species

I9I4-] SOUTHEASTERN NORTH AMERICA. .01

confined to the East Indies, two genera with 20 species confined to
Polynesia and 6 genera with 35 species ranging from Malaysia or
the East Indies to Australia. It is quite obvious from these few
facts regarding the existing distribution that the family is an ancient
one and that there has been an extensive evolution of both generic
and specific types in relatively modern times in the American tropics
on the one hand and in the Malaysian region on the other.

The fossil record while much less complete than might be wished
includes at least 13 genera of which six are extinct, and about 160
species, by far the largest number being referred to the still existing
genus Sapindus which appears well differentiated and widely distrib-
uted at the dawn of the Upper Cretaceous. There are about ten
Upper Cretaceous species of which all but four occur inpre-Senonian
strata. Thus there are two in the Perucer beds of Moravia and
Bohemia, one at Niederschcena in Saxony — all Cenomanian. Two
in the Atane and one in the Patoot beds of West Greenland. Two in
the Dakota group ; two in the Tuscaloosa formation of Alabama,
one in the Middendorf beds of South Carolina, one in the Woodbine
formation of Texas ; two each in the Raritan and Magothy forma-
tions of the Middle Atlantic states, one in the Montana group and
two in the Laramie. I have given this Upper Cretaceous distribu-
tion in some detail because of the special interest attached to the
deployment of the Upper Cretaceous Dicotyledons. It should be
noted that seven of these Upper Cretaceous forms are North Amer-
ican. There are over thirty Eocene species of Sapindus of which
two thirds are North American. The genus is very abundantly rep-
resented in both individuals and species in the coastal floras of the
Wilcox group from which I have described no less than 9 species.
There are four species in the overlying Claiborne group. Species
of Sapindus are equally common in the Rocky Mountain province in
the Denver, Fort Union and Green River beds. There is an Eocene
species in Greenland and one each in New Zealand, Australia, Tasma-
nia and Chili. There are four undescribed species in the Ypresian of
England and a fifth in beds of the same age in Hungary. There is
an Upper Eocene species in France and a second in Oregon.

There are six or more Oligocene species well distributed in Eu-

202 BERRY— LOWER EOCENE FLORA OF [April 25,

rope to which continent their discovery has thus far been confined.
There are over 30 Miocene species found throughout southern Eu-
rope, in eastern Asia, and in North America (Colorado, Oregon and
Yellowstone Park). The eight or ten Pliocene species are confined
to southern Europe.

In addition to the genus Sapindus there are several form-genera
derived from the same root. Thus Sapindophyllum has been ap-
plied to two species from the Albian of Portugal (?). To it are
also referred a Cenomanian and a Chattian species from Bohemia
and a Tertiary species from Japan. The term Sapindoides has been
used by Perkins for Sapindus-like fruits preserved in the early Ter-
tiary lignites of Brandon, Vermont, from which eight species have
been described. In some respects the most interesting genus is Sap-
indopsis Fontaine represented by three abundant and well preserved
species in the Patapsco formation (Albian) of Maryland and Vir-
ginia one of which is also present in the Fuson formation of the
Black Hills, and which I have shown 36 to be very probably ancestral
forms of the genus Matayba Aublet (Cupanieee) which has upwards
of two score existing species in the tropical and subtropical regions
of America. This well-marked type suggests the interesting ques-
tion of how early in the Mesozoic the ancestors of many modern
genera may have been present in equatorial America.

The genus Paullinia Linne which has about 122 existing species
mostly confined to the American tropics but sparingly represented in
Africa and Madagascar, has an Oligocene species in Prussia and
two early Miocene species in southeastern France and Bohemia.

The genus Thouinia Poit, which in the modern flora has about
15 species confined to the West Indies ami Mexico, is represented
by an early Tertiary, probably Eocene species, in Chili. The genus
Nephelium Linne with over a score of existing species in southeast-
ern Asia is recorded by Unger from the Aquitanian of Greece and
by Geyler from the Tertiary of Borneo.

The genus Kcelreuteria Laxm. is represented by two Chinese spe-
cie- in tlie existing flora. In the fossil state it is recorded from the
Tertiary of the Island of Sachalin, from Spitzbergen and from

■■•■ Berry, Md. Geol. Surv., Lower Cretaceous, pp. 467-474. PI- 83-88, 191 1.

I9I4-] SOUTHEASTERN NORTH AMERICA. 203

Switzerland and Baden. Felix has described a genus, Schmideli-
opsis, based on fossil wood from the Oligocene of the Island of
Antigua, very close to the existing genus Schmidelia Linne which has
upwards of a hundred existing species in all tropical countries.

The modern Cupaniese are represented in paleobotanical liter-
ature not only by Cuponia but by species of Cupanites and Cupan-
oidcs. The latter generic term was proposed by Bowerbank for
cupaniaceous fruits and seeds of which he described several charac-
teristic species from the Ypresian of the Island of Sheppey. Simi-
lar forms have also been recognized in the Miocene of Carniola and in
the Pliocene of Italy. The genus Cupania Linne has about 35 exist-
ing species confined to the American tropics. Several Ypresian spe-
cies from the south of England have been referred to it by Ettings-
hausen and it has also been recorded from the Miocene of the Island
of Sachalin. The greater number of Capania-Yike forms have, how-
ever, been referred to the genus Cupanites Schimper. Nine or ten
species have been described and with the exception of extremely
doubtful forms from the Upper Cretaceous of New Zealand and the
Eocene of Australia, the oldest authentic occurrences are the two
species of the Wilcox flora. There is a third species in the overlying
Claiborne group of the Mississippi embayment. The oldest Euro-
pean form is one from the late Oligocene of Styria. In the Miocene,
species are recorded from Germany, Bohemia, Austria, Croatia and
Hungary.

The genus Dodoncca Linne often made the type of a distinct
family, the Dodonaeacese, has about fifty existing species of which
four fifths are Australian. Dodon&a viscosa Linne is cosmopolitan
in the tropics and there are one or two additional species in the
American tropics as well as one in the Hawaiian Islands and another
in Madagascar. The genus (including Dodonccites) was evidently
widespread in former times and upwards of a score of fossil species,
based on both leaves and fruits, have been described. The oldest
known forms are two species in the Ypresian of the north of Eng-
land and the two contemporaneous species in the Wilcox, which are
represented by both leaves and characteristic fruits. There are five
Oligocene species in France, Tyrol, Bohemia and Styria; and ten

20i BERRY— LOWER EOCENE FLORA OF [April 25,

Miocene species in Prussia, Baden, Switzerland, Bohemia and Croa-
tia. A well-marked species occurs in the Claiborne (Lutetian)
ranging along the Claiborne coast from northeastern Georgia to
central Louisiana.

It is impossible from the known facts to discuss the place of ori-
gin of the family, but it is obvious that certain genera were evolved
toward the close of the Lower Cretaceous in equatorial America and
have inhabited that or adjacent areas throughout the long stretch of
time down to the present.

The order Rhamnales includes about 1,000 existing species of
shrubs, trees and vines about equally divided between the families
Rhamnacens and Yitaceae. It closely parallels the Sapindales in its
floral development but is distinguished by the mostly tetracyclic flow-
ers with opposite stamens and often lacking a corolla. The leaves
are simple and typically alternate. Of the two families only the
Rhanmacere is represented in the Wilcox flora.

The family Rhamnacese (Frangulaceae) includes 47 genera and
about 500 species of shrubs and trees mostly of the tropics but with
several genera extending for considerable distances into the temper-
ate zone, the genus Rhamuits in particular being mostly extratropical
in the northern hemisphere. The genera Zizyphus } Adelia and Gon-
dii ia are found in all tropical countries. Almost half of the genera
are common to more than one continental area. America has the
greatest number of peculiar genera (15) with about 85 species. Two
monotypic genera are confined to Asia, five genera including the
large genus Phylica Linne together With about 70 species are con-
fined to Africa and five genera including the two large genera Spy-
ridium Fenzl. and Cryptandra Smith in all with about 70 species are
confined to Australia.

Ten or eleven genera, of which five are present in the Wilcox
flora, are found fossil, the three largest being Rliainiuis. 1'aliitriis
and Zizyphus. The genus Rhamnus Linne which L cosmopolitan
in the northern warm temperate and subtropical zones has about
•ity existing species. There are considerably over one hundred
fossil species, mostly well characterized with simple often entire
leaves with ascending secondaries and closely spaced tine percurrent

I9I4-] SOUTHEASTERN NORTH AMERICA. 205

nervilles. There are a dozen or more species described from the
Upper Cretaceous, the genus appearing in the Cenomanian in both
Europe (Niederschcena, Saxony) and America (Raritan formation).
There are six species in the Dakota sandstone, two in the Magothy
formation, one in the Atane and two in the Patoot beds of Green-
land. The genus is represented in the Montana group and the Lara-
mie formation of the Western Interior and in the Senonian of West-
phalia. There are about thirty Eocene species, the majority being
North American. Species of Rhamnus are very common in the
Raton and Denver formations along the Front Range of the Rocky
Mountains and from the base to the top of the Wilcox. There are
four species in the Raton, eight in the Denver and six in the Wilcox.
The genus is also well represented in the later Eocene along the
Pacific coast and in western Greenland. In Europe only a single
species is recorded from the Paleocene. The Ypresian which is syn-
chronous with the Wilcox has three species in the south of England.

There are eleven or twelve Oligocene species in France, Prussia,
Tyrol, Italy, Dalmatia, Styria, and Greece and a single undescribed
species in the Apalachicola group of Florida. There are over two
score Miocene species, Rhamnus being especially abundant in the
Miocene of Switzerland, Italy, Bohemia, Prussia and Styria. It is
also present at this time in Iceland, Spitzbergen, Manchuria and
Sachalin Island. In this country there are species in British Co-
lumbia and in Colorado.

There are about thirteen Pliocene species, no less than nine being
recorded from Italy and there is one known from the Island of Java.
There is an extinct species in the Pleistocene of Hungary and a re-
cent species in the Pleistocene of the Island of Madeira. In addi-
tion to the species referred to Rhamnus the form-genus Rhamnitcs
Forbes founded on three species from the Eocene of the Isle of Mull
has two American Upper Cretaceous species found in the Raritan,
Tuscaloosa, Magothy and Dakota formations. There is a species in
the Fort Union and another in the Wilcox. The genus Rhamna-
cinium of Felix is based on petrified wood. It contains five or six
species found in the Eocene of the Caucasus, Texas, Saskatchewan
and the Miocene of Yellowstone Park.

20(1 BERRY— LOWER EOCENE FLORA OF [April 25,

The genus Paliitnis Jussieu with only two existing species rang-
ing from southern Europe through southern Asia to China and Japan
was cosmopolitan in former times. Upwards of 40 fossil species
have been described. At least twelve are known from the Upper
Cretaceous, all confined to the North American region. There are
two each in the Raritan, Magothy and Laramie, five in the Dakota
and one each in the Eutaw formation of Georgia, in West Greenland
and Vancouver Island. There are ten Eocene species also confined
to North America. Two of these are found in the Fort Union and
there are three each in the Denver, in western Greenland and in the
Wilcox. The leaves are not common in the Wilcox but the charac-
teristic peltate fruits are not uncommon. The oldest European
forms are two species in the Oligocene of France and there is a well
marked species in the Oligocene (Vicksburg group) of Louisiana.
The thirteen Miocene species are found in Asia (Siberia, Sachalin),
Europe (Switzerland, Baden, Germany, Bohemia, Italy, Styria and
France), and North America (Colorado and Oregon). The pres-
ence of numerous species of Paliitnis in the Upper Cretaceous and
Eocene of North America and their absence on other continents be-
fore the Oligocene renders it very probable that the genus originated
in the western hemisphere.

The genus Zizyphus Jussieu with about forty existing species
largely shrubs, often prostrate or scrambling, and rarely small trees,
is mostly Indo-Malayan in its distribution but is represented by a
few species in the tropics of eastern Asia, America, Africa and Aus-
tralia. There are over fifty known fossil species and as in the genus
Paliurus the ten Upper Cretaceous species are confined to North
America. They are found in the Raritan and Magothy formations
of New Jersey and Maryland, the Eutaw formation in Georgia, the
Tuscaloosa formation in Alabama, the Woodbine formation in Texas,
the Dakota sandstone of the West, the Patoot beds of Greenland
and the L'pper Cretaceous of Alaska. There are about twenty
Eocene species including the two common and characteristic species
of the Wilcox and one in the overlying Claiborne of the embayment
region, five in the Denver, three in the Fort Union, two in the I rreen
River, one in Alaska and one in West Greenland. There are two

I9I4-J SOUTHEASTERN NORTH AMERICA. 207

Paleocene species in France and Belgium, four Ypresian species in
the south of England and a Lutetian species in France. There are
eight Oligocene species very common in deposits of this age through-
out Europe. Over twenty species have been recorded from the Mio-
cene of Colorado and California in this country, from France, Swit-
zerland, Germany, Italy, Austria-Hungary, and Russia in Europe,
and from Japan and Java in Asia. There are three or four Pliocene
species in Europe. While the evidence is not so clear as in the case
of Paliurus there is a possibility that Zizyphus too is of occidental
origin.

The genus Reynosia Grisebach with only two existing coastal
species ranging from the Florida keys through the West Indies has
two characteristic species based on leaves in the Wilcox flora and a
third species based on the petrified wood in the overlying Claiborne
deposits of Texas.

The genus Berchemia Neck, has about a dozen existing species,
ten of which are confined to eastern and southeasern Asia. There
is one in eastern extratropical North America and one in eastern
Africa. This distribution could not have been brought about ex-
cept by the agency of a cosmopolitan Tertiary range. While the
specific differentiation of Berchemia is limited to five or six fossil
forms these are very common and wide ranging. The earliest occur-
rences are North American and include the Raton, Denver and Fort
Union formations of the Rocky Mountain province. The genus
makes its apearance in Europe during the Oligocene and is common
throughout that region in the Miocene, becoming restricted to south-
ern Europe (France, Italy, Sicily and Slavonia) during the Pliocene.

A species of Hoveniphyllum supposed to represent the existing
genus Hovenia Thunberg with a single existing species in southeast-
ern Asia, is present in the Plio-Pleistocene of Japan. The genus
Colubrina Brongniart with 15 existing species in tropical America
and one in southeastern Asia is recorded from the Miocene of
Bohemia.

The genus Pomaderris Labill with about 24 existing species con-
fined to Australia and New Zealand has two species in the Eocene
of the former region and three species (Pomaderrites Ettingshau-
sen) in the Miocene of Prussia, Bohemia and Styria.

208 BERRY— LOWER EOCENE FLORA OF [April 25,

The genus Ceanothus Linne with about forty existing species
confined to North America has included numerous fossil species sub-
set |uently referred to Paliurus or Zizyphus. There are four re-
corded from the Upper Cretaceous of Greenland. New Jersey, Van-
couver Island and Westphalia ; two Eocene species recorded from
Greenland and British Columbia; a Miocene species in Prussia.
Switzerland and Italy ; and a Pleistocene species in Kentucky.

The next order, the Malvales, includes nine families and about
1,800 existing species. The Tiliacese, Sterculiacere and Bombacacese
are the only ones represented in the Wilcox flora. The largest
modern family, the Malvaceae with over 800 species, many of which
are herbaceous and range from 65 ° North latitude (Russia) to 45°
South latitude (New Zealand), is not represented in the Wilcox.
The order displays somewhat uneven or but little understood phylo-
genetic characters but is evidently allied to the succeeding order, the
Parietales, through the family Elaeocarpaceae. These inequalities of
evolution are shown among other ways by the complete syncarpy in
the Tiliacese associated with an indefinite number of stamens and by
the complex arrangement of the stamens in the Sterculiaceae, associ-
ated with more or less incomplete union of the carpels. Both the
leaves, flowers and fruits exhibit a wide range of variation through-
out the order.

The family Tiliacese, represented in the Wilcox flora by a single,
not very common form of Grewiopsis, has about 35 genera and 370
existing species mostly of tropical lands and showing two cente,
differentiation and distribution — one surrounding the Indian Ocean
and the other in northern South America. The geological history is
confined to the four genera Tilia (or Tiliaphyllum ). Grewia, Grewi-
opsis and Apcibopsis: 1 ' The genus Tilia Linne with [8 or 20 widely
distributed existing species in the north temperate zone (absent in
western North America and central Asia) has furnished about 25
fossil species based upon both leaves and fruits, the oldest known
being from the North American Eocene. There are no conclusive
< Higocene records except two French species, but about fifteen Mio-

37 The genus Luhea has been described from the Eocene of Sezanne
(Langeron) and from the Oligocene of Menat (Laurent), both French
localities.

I9I4-] SOUTHEASTERN NORTH AMERICA. 209

cene species are found in North America, Europe, Asia and the
Arctic regions. There are five Pliocene species recorded from Eu-
rope and Japan and six Pleistocene species in Ontario, New Jersey,
France, Germany, Holland and Denmark. The genus has appar-
ently had its existing range since Miocene time.

The genus Grewia Linne has about 90 existing species ranging
from Arabia to China and Japan and through Malaysia to Aus-
tralia, and from Abyssinia to South Africa. About fifteen fossil
forms have been described, the oldest known, five Eocene species,
coming from western North America. There are two Oligocene spe-
cies in Europe and about six Miocene species in Oregon, Spitzbergen
and throughout Europe. The larger number of Grewia-like fossil
forms are, however, referred to the genus Grcwiopsis of Saporta.
Six of these are from the Upper Cretaceous and all are confined to
North America, a very significant fact since several of them are espe-
cially well marked. They are found in the Magothy formation of the
east coast, the Tuscaloosa formation of the south coast, and the Da-
kota Montana and Laramie formations of the western interior. There
are about six Eocene species in the Denver, Lance and Fort Union :
one in the Wilcox and one in the Claiborne of the Mississippi em-
bayment region, six in the Paleocene of France and one in the
Ypresian of England. There is also a Miocene ( ?) species recorded
from Yellowstone Park. While it is quite possible that some of the
fossil records ascribed to the genus Populus are those of Grewia
or its ancestral stock, it seems clear that the latter genus or its im-
mediate ancestors were common in the Upper Cretaceous and Eocene
of North America.

The fourth fossil genus of Tiliaceae is Apeiobopsis Heer 38 named
from its affinity with the existing genus Apeiba Aublet which has five
or six species confined to tropical South America. Apeibopsis in-
cludes not only leaves but very characteristic fruits. To it are re-
ferred somewhat doubtfully determined leaves from the Upper Cre-
taceous Dakota sandstone and Atane beds. There are about four-
teen Tertiary species including a basal Eocene form from Wyoming,
two Ypresian forms from England, a species from West Greenland

38 To it should probably be referred the Arctic forms described by Heer
as N ordenskibldia.

PROC. AMER. PHIL. SOC. LIII, 214, X, PRINTED JULY 13, 1914

210 BERRY— LOWER EOCENE FLORA OF [April 25,

three species in the lignites of Brandon. Vermont, two Oligocene
species from Italy and five Miocene species from France, Switzer-
land and Bohemia.

The family Bombacaceae 89 with 20 genera and about 120 exist-
ing species is confined to the tropics and principally to the Amer-
ican tropics. The only known fossil forms are those of the genus
Bonibax or the allied Bombaciphylhtiu and Bombacites. Bonibax
Linne has about fifty existing species, all large tropical trees, and
almost confined to America. There is a single species in Africa,
about six in southern Asia and one in Austalia. The fossil species
number over twenty, the oldest known 40 being a common form in
the Perucer beds (Cenomanian) of Bohemia and Moravia. There
are three species in the Ypresian of southern England and two well-
marked forms in the "Wilcox flora. There are five additional Eocene
forms of which three are from Chili and two from eastern Australia.
There are five Oligocene species recorded from France, Saxony,
Bohemia and Carniola. The genus is represented in the early Oli-
gocene (Sannoisian) of southeastern France not only by the foliage
hut by beautifully preserved flowers so that there is little ground for
questioning the correctness of the identifications. There are five
Miocene species in Bohemia, Croatia and Styria.

The family Sterculiace?e includes about 5 genera and 800 exist-
ing species of mostly tropical shrubs and trees with prevailingly
large, simple or digitately lobed or divided leaves ; the flowers are
sometimes apetalous and differ from those of the Malvaceae in their
2-celled extrorse anthers. Syncarpy is more or less complete.

The Sterculiacea? of the existing flora are found on all the conti-
nents except Europe. The genera Stcrculia, Helicteres, Melochia,
Buettneria and Hermannia have species in both the eastern and
western hemispheres. The geologic history of the family extends
back to the base of the Upper Cretaceous but is confined to a rela-
tively few number of genera. The most abundant of these is the
genus Stcrculia Linne. which in the existing flora has about one

39 Ettingshausen. "Ueber die Nervation der Bombaceeeii," Prusk. k.

Akad. Wiss. Wien. Math. Wit. CI.. Hand 14. 1858, pp. 49-62, PI. I.-Xl.

'"An Albian species of Bombax described by Fontaine is entirely

valueless.

I9M-] SOUTHEASTERN NORTH AMERICA. 211

hundred species of large leafed trees. The)- are divided into three
tribes named from the habit of the leaves the Digitatse, Lobatse and
Integrifolise. The first of these range from farther India to Aus-
tralia with only one or two American species. The second is most
abundant in the American tropics but is also found in Asia and
Africa and shows many parallelisms between the American and
Asiatic forms. It is most abundantly represented in the past history
of the genus. The third and largest modern tribe, the Integrifoliae,
has five or six American species and the balance are found in Asia
and Africa.

The fossil forms (sometimes referred to SterculiphyUum)
number more than fifty species. Upwards of a score are known
from the Upper Cretaceous. These are mostly American and are
referable to the tribe Lobatse which may well have originated in the
western hemisphere. There is a species each in the Credneria
sandstone of Saxony and the Perucer beds of Bohemia (both Ceno-
manian) and a third in the Turonian of the latter country. The
balance are North American and include species in the Raritan for-
mation, the Cheyenne sandstone of southern Kansas, and in British
Columbia, a species in the Patoot beds of West Greenland and six
species in the Magothy formation of the Atlantic Coastal Plain and
eight species in the Dakota sandstone of the western interior.
There are less than a dozen Eocene species, the majority being con-
fined to the lower Eocene. Thus there are three species in the
Paleocene of France and another in the Ypresian of England as
well as one or two in the Denver and Raton formations of the Rocky
Mountain front range. The single large Wilcox species is entirely
typical and shows the usual variability in lobation and size. It
appears to be filiated with Sterculia Snozvii Lesquereux from the
American Upper Cretaceous, and may be exactly matched by several
existing species. There is a small leafed species in the middle
Eocene (Claiborne group) of the embayment which exactly matches
the typical Sterculia labrusca Unger from the European Tertiary and
the existing Sterculia diversifolia Don. It is closely parallelled by
two American Upper Cretaceous species — S. minima Berry and S.
mucronata Lesquereux. There are upward of ten Oligocene species
widely scattered over Europe and about 15 Miocene species, mostly

212 BERRY— LOWER EOCENE FLORA OF [April 25,

European, with a single species on the east coast of Asia (Sachalin)
and two species in Colorado, one of them especially well marked.
There are several Pliocene species in southern Europe.

Two somewhat different species of sterculiaceous capsular fruits
from the Wilcox are referred to a new genus, Sterculiocarpus. The
larger of these, S. eocenicus, seems referable to the subfamily Buett-
nerie?e, while the smaller, S. sezannelloides, is referable to the Lasio-
petaleae or Helictere?e. Both are very similar to the fruits from the
I'aleocene of Sezanne referred to the genus Sczannclla. The latter
genus with two species was described by Yiguier from casts of
wonderfully preserved flowers as well as fruits from the celebrated
Travertins of Sezanne and referred with great certainty to the
Lasiopetaleae.

The tribe Dombeyea? with seven genera and about 75 existing
species is almost entirely confined to Africa and the adjoining
islands, five or six species of the genus Melhania Forsk, only, rang-
ing from Arabia to farther India. This tribe is represented in fossil
floras by the genus Dombcyopsis Unger named from its supposed
affinity with the modern genus Dombcya Cav. which has 40 African,
mostly Madagascar, species. About 30 species have been referred to
Dombcyopsis. They are liable to be confused with Luhca, Grewia
and other forms of the allied family Tiliacea?. There are three
species in the Laramie Cretaceous, two in the Denver formation,
twelve (according to Massalongo) in the Upper Eocene of Monte
Bolca in Italy, five in the European Oligocene and six in the Miocene
of Iceland, France, Switzerland, Prussia, Silesia and Styria. A
Pliocene species is recorded from central France. Fossil wood
described as Dombeyoxylon is recorded by Schenk from the late
Tertiary near Cairo, Egypt.

The Buettnerieae are represented by a doubtful species described
from the Miocene of Colorado, and probably by some of the fossil
forms referred to other genera, e. ,'/.. some of the palmately veined
Ficus-like forms such as Ficus occidentalis and Ficus Schimperi both
of which are present in the Wilcox flora. Flowers of Buettneria
were reported from Sezanne by Solms-Laubach but this probably
refers to the subsequently described genus Sesannella, mentioned in
a preceding paragraph.

I9I4-] SOUTHEASTERN NORTH AMERICA. 213

The Helictereas are represented by a doubtful species of Hehc-
teres Linne described from the Pliocene of Italy and by forms
referred to the existing genus Pterospermum Schreb. or to the
extinct genus Pterospermites Heer. Over 30 species have been de-
scribed. There are nine or ten in the Upper Cretaceous all of which
are Xorth American, and their combined range extends from New
York to western Alabama, throughout the Rocky Mountain and
Great Plains province and in the Atane beds of Greenland. There
are about a dozen Eocene species all Xorth American except a single
species in the Paleocene of France. The American forms extend
northward to West Greenland and Alaska. There are two or three
species in the European Oligocene and ten Miocene* species through-
out Europe and in western Xorth America (Yellowstone Park, Cali-
fornia, mouth of the Mackenzie River). A single Pliocene species
is recorded from France. It seems probable that this type originated
in the western hemisphere since it is so abundantly represented in
that region during the Upper Cretaceous and Eocene. The modern
species of Pterospermum are, however, confined to eastern tropical
Asia.

The order Parietales includes thirty families together with over
four thousand existing species, the largest families being the Gut-
tiferae (775), Flacourtiaceae (530), Begoniaceoe (425), Violacese
(400) and Dipterocarpacese (330). Xone of these families are
present in the Wilcox flora, where the order is represented by the
two families, the Dilleniaceae and Ternstrcemaceae. The Parietales
are prevailingly syncarpous and show affinities with the Ranalian
plexus tli rough the Dilleniaceae which were formerly referred to that
order. The alliance as a whole is a complex one including several
divergent lines of development with, on the whole, a gradual increase
in floral complexity.

The family Dilleniaceae contains 14 genera and about 275 exist-
ing species found on all the continents, the genus Tctracera being
cosmopolitan in the tropics. The genera Empedoclea, Curatclla,
Doliocarpus and Darilla together with 50 species are confined to
the American tropics : Hibbertia and Pachynema together with 75
species are Australian, there are five genera with about 25 species
confined to the Asiatic tropics; the genus Saurauia (or Saurauja),

214 BERRY— LOWER EOCENE FLORA OF [April 25,

with about 60 species, is common to Asia and South America; and
the genus Dillenia with about 25 species ranges from Asia to Aus-
tralia; so that on the whole the family is prevailingly oriental in the
existing flora.

The fossil record is unfortunately most incomplete, illustrating
however a wider range of the genera in the past in response to milder
climatic conditions in both the north and the south temperate zones
during the Tertiary, and also the fact that several of the modern
American genera have been American throughout their known geo-
logic history. Thus Empedoclea with two existing South American
species, sometimes made a subgenus of Tetraccra, has a fossil form
in the eary Tertiary of Chili. The genus Doliocarpus with about 20
recent species also in the South American tropics has two fossil
forms in the early Tertiary of Chili. The genus Davilla with 25
modern species in tropical America is doubtfully represented in the
Wilcox flora by Calycites davillaformis Berry.

The genus Saurauja with 60 modern species in South America
and Asia has a species in the Paleocene of France, another in the
Ypresian of the south of England and a third in the Miocene of
Croatia.

The genus Dillenia with 25 existing species confined to Asia and
Australia is represented by a form in the Paleocene of Belgium and
by some of the Wilcox species referred to the form-genus Dillenites.
The genus Tetraccra with 40 recent species found in all tropical
lands, has two fossil species in the early Tertiary of Chili, another in
the Pliocene of Java and is represented in the Wilcox flora by some
of the species of Dillenites. I have recognized five well-marked
species of Dillenites in the Wilcox and these appear to represent
modern forms of both Dillenia and Tetraccra.

Conwenz described three species of Ilibhertia. a large Australian
genus, in the Baltic amber (Sannoisian) but Schenk considered that
they did not belong to either this genus or even the family.

The family Ternstroemacea? (Theacens) contains about 16 genera
and 175 existing species mostly tropical but extending into the north
temperate zone in North America and eastern Asia (Tlica, Gordonia
and Stewartia I. The following seven out of the sixteen genera are
confined to a single area: Bennetia Martins with five species inhabits

I9I4-] SOUTHEASTERN NORTH AMERICA. 215

the South American strand, Asteropeia Dub. is confined to Mada-
gascar, Thea Linne with sixteen species is confined to southern and
eastern Asia, Mountnorrisia Szysz. with two species is a native of
the East Indies, the three monotypic genera Visnea Linne, Treman-
thera Muller and Pclliciera Tr. and Planch, are confined respectively
to the Canary Islands, New Guinea and Central America. The
remaining nine genera, all relatively small, are all found in more than
one region. Thus ArcJiytcca Martius has two species in northern
South America and a third in the East Indies; Gordonia Ell. has
two Xorth American species and fourteen scattered from India to
Malaysia ; Hccmocharis Salisb. has nine American and five Asian
species ; Stewartia Linne with five species is found in North America
and Japan ; Taonabo Aublet has 20 species in South America and
eight in Asia; Adinandra Jack, has 19 African species and one in
Asia ; Eurya Thunb. with 36 species and many varieties is confined
to tropical America and the East Indies.

This remarkable existing distribution and the pairing of America
and Asia as well as the fact that it requires five subfamilies for the
reception of only sixteen genera are sure indications that the family
has an extended geologic history and that many of the genera were
once cosmopolitan. Unfortunately most of this history is unknown.

The genus Stewartia is represented in the Baltic amber by a fine
flower {Stewartia kowalowskii Caspary) and by leaf remains from
the Plio-Pleistocene of Japan (Mogi). Gordonia has a species in
the Pleistocene of Java. The genus Eurya Thunberg, now American
and East Indian, has a species in the Oligocene of France [Freziera
Swartz). Fossil wood described by Felix and named Ternstrcemia-
cinium occurs in the Eocene of the Caucasus. Visnea Linne, now
confined to the Canaries, has a typical fruit in the Aquitanian of
Rhenish Prussia. The genus Ternstraemia Nuttall (antedated by
Taonabo Aublet) has several fossil species, the oldest {Ternstrce-
miphyllum) coming from the Perucer beds (Cenomanian) of Bo-
hemia. It has two species in the Ypresian of the Isle of Wight, one
in the Miocene of Bohemia and another in the Miocene of Croatia.
I have described four well-marked species of Ternstrccmitcs from
the Wilcox group and similar forms are present in the overlying
Claiborne group (Lutetian). Finally the very abundant species in

216 BERRY— LOWER EOCENE FLORA OF [April 25,

the North American Cretaceous described as Celastrophyllum and
already referred to in the discussion of the Celastraceae are very
probably, in part at least, referable to this family, so that enough is
known of the geologic history of the group to confirm at least the
statement made in a preceding paragraph that it must have had a
long and extended history.

The family Lauracea? with in the neighborhood of 1,000 existing
species distributed among forty to fifty genera is often placed next
to the family Anonaceae among the Ranales (c. g., in Engler and
Prantl's " Xaturlichen Pflanzenfamilien "). It may be noted, how-
ever, that the spiral arrangement of floral organs characteristic of
the order Ranales is replaced by a cyclic ararngement and hypogny
i- also replaced by epigyny, so that I follow various students in
referring the Lauracea; to the order Thymeleales, the other large
family of which, the Thymelreacese (not known in Wilcox flora),
has about 400 existing species, chiefly of temperate Australia and the
Cape region of Africa.

The geographical distribution of the Lauraceae cannot be disposed
of in a similar simple statement since there are not only many
anomalies in the distribution of the existing species but we know so
considerable a part of the geologic history that our difficulties seem
increased thereby rather than diminished. For example the existing
species of the family are divided into eight tribes, no one of which
except the monotypic Eusideroxylea? of Borneo is restricted to a
single continental region.

The largest of these tribes is the Cinnamomeae with upwards of
500 species endemic on all the continents but Europe, and chiefly
Asiatic and American. The four genera Pcrsca, Phcebe, Notaphoebe
and Mespilodaphne are found in both hemispheres; Cinnatnomum
and Modulus are oriental: while OrcodoHinc. Strychnodaphne,
Nectandra, Pleurothrium, Umbclluhvia, Dicypellium and Synandro-
(lapluie are occidental ; the first three being large genera and the
last f'»nr being monotypic.

The tribe Litseae, with six genera and about 200 species, is
represented on all the continents except Europe and Africa. Only
9 of these two hundred species are found in the Occident and yet

among these i^ the monotypic North American genus Sassafras, and

• 9.4.] SOUTHEASTERN NORTH AMERICA. 217

the genus Sassafridium confined to the American tropics. All of the
other genera are found on more than one continent.

The tribes Apollonieae, Cryptocaryese and Cassythese are found
on all the continents but Europe. The Laure?e are Eurasiatic and
the Acrodiclidiese are confined to Central and South America, except
the genus Endiandra which with 16 species occurs in the East Indies
and Australia.

The problem of correctly identifying leaves of the various genera
of this family is beset with almost unsurmountable difficulties, not the
least of which is due to the wide differences in usage among students
of the recent forms where the whole plant is available for study.
Long-continued paleobotanical practice has been to refer most fossil
leaves that lacked the more apparent characters of Cinnamomum or
Sassafras, Persea or Malapoenna, etc., to the comprehensive genus
Launis given at a time when Laitnts was used in a comprehensive
sense, and sometimes still more generalized by paleobotanists as
Laurophylhtni for lauraceous leaves of uncertain generic affinity and
not necessarily close to the existing species of Laitnts, in fact they
are in general not true species of Laitnts. I have departed from this
practise of describing new species of Laitnts for a variety of reasons
foremost among which is the very great affinities between the Wilcox
flora and the existing flora of the American tropics, the evidence
from the foliage of a large number of genera being coroborated by
fruits or seeds or wood anatomy. I have used this similarity with
a great deal, perhaps too much, confidence and the result has been
that the following stand out as the more important lauraceous types
iii the Wilcox flora :

Nearly all are seemingly members of the subfamily Persoideae
and under this subfamily of the tribe Cinnamomere as segregated in
Engler and Prantl's " Xaturlichen Pflanzenfamilien."

Eirst the genus Ciiinamomitiii, usually readily recognized and
certainly represented in our Eocene floras.

Second the genus Persea, represented by the larger and wider
forms with the typical venation of this genus.

Third the genus Nectandra, so abundant and characteristic of
the existing flora of tropical and subtropical America, represented by
several species very close to modern forms.

218 BERRY— LOWER EOCENE FLORA OF [April 25,

Fourth. I have failed to follow the latest usage in recognizing the
genus Ocotca as such, since for obvious reasons it seems better to
recognize the genera Mcspilodaphnc and Oreodaphne of Xees rather
than to regard them as subgenera of Ocotea. The third subgenus of
Ocotca. — Strychnodaphne, I have failed to recognize in the Eocene
flora of this area.

The only apparent oddity in distribution shown by the Wilcox
Lauracese in comparison with recent floras of tropical America is the
abundance of Cinnamomum, and this simply adds confirmation to the
well-known fact of the cosmopolitanism of this genus in the early
Tertiary. Grisebach records only 28 species of Lauraceae in his
flora of the British West Indies. Most of these are not coastal forms
although many have a wide range from lowlands to mountains. As
regards the Lauraceae those of the Wilcox, which number 30 different
f< irms, are more closely comparable with the more abundant modern
representation of this family in northern South America. This re-
ceives more or less confirmation from a study of the balance of the
Wilcox flora. It would seem from a consideration of all of the
facts that the early Eocene floras of the Mississippi embayment are
much more like those existing at the present time along the Caribbean
sea in Central America and northern South America than they are
like those of the West Indies. I do not mean by this that the Wil-
cox flora has not many points of resemblance to the lowland flora of
the West Indies and that of the Florida keys. They contain very
many common types, but with the following difference. The Missis-
sippi embayment Eocene floras represent a maximum northward ex-
tension of a flora like that which now inhabits northern South Amer-
ica. At the end of the Oligocene with the southward migration
of the temperate Miocene fauna as far as Florida, this flora retired
to the South American mainland and the present flora of the West
Indies, Florida keys, Bahamas and Bermuda represent a later north-
ward migration from that area, a migration in which some of the
Wilcox types were left behind.

The existing species of Cinnamomum* 1 number about fifty. They
are confined to the Oriental tropics except for their extension into
the warmer more humid part of the temperate zone in Japan, and

" Stauli. "Die Geschichte des genus Cinnamomum," Budapest, 1

I9I4-] SOUTHEASTERN NORTH AMERICA. 219

they have their chief center of differentiation in the elevated region
of Burma, Siam, Cochin-China and Malaysia, although they are cul-
tivated in all tropical countries and outside the tropics in Europe,
Africa and North America. Their fruits are eaten by birds which
seed them freely so that they commonly escape from cultivation.
Thus Cinnamomum Caiuplwra (Linne) Nees and Eberm. is natu-
ralized throughout peninsular Florida, and the commercial Ciiuia-
momum zeylanicum Breyn., is readily naturalized in the same manner
from the Oriental camphor plantations.

While the data for constructing the geologic history of Ciiuia-
momum are far from complete there are more known fossil than
recent species and these show, as in the case with so many plant
groups, surprising extensions of range during the Upper Cretaceous
and Tertiary. The original home of the genus is unknown for it
appears in the early part of the Upper Cretaceous at about the same
time in New Zealand, Australia, central Europe, Greenland, North
and South America. The European and North American records
appear to be slightly older than the balance and would indicate that
the Asiatic region may have been the original home of the genus
which spread northeastward across the Behring region to America
and northwestward into the European region, the latter largely an
archipelago at that time.

The Eocene records include all of the continents except Antarctica
and South America. The Oligocene records are chiefly European
and African, although the genus is still represented in the Florida
Oligocene. During the Miocene Cinnamomum was abundant in
Europe and present in Asia but appears to have become extinct in
North America, at least there are no conclusive North American
records. A number of fruits from the Brandon (Vermont) lignites
have been referred to Cinnamomum but these lignites are in my opin-
ion pre-Miocene in age. The Pliocene records are entirely Euro-
pean and East Indian. The genus appears to have lingered as a com-
mon type in Mediterranean Europe until the changing climates that
ushered in the Pleistocene glaciation caused its extinction, any con-
nected distribution with its present Oriental home across southwestern
Asia having already been interrupted by the orogenic movements and
the development of arid conditions in southwest Asia.

220 BERRY— LOWER EOCENE FLORA OF [April 25,

There are six well-marked types of Cinnamomum leaves described
Erom the Wilcox group, some of them being abundant and generally
distributed, and all hut two appear to be new to science. In addition
buds and flowers that suggest this genus are described under the
form-genus Laurophylhun.

There are two species of Persea in the Wilcox flora. Disregard-
ing the fossil forms referred to Laurus in a comprehensive sense
there are about fifty known fossil species of Persea which is about
the number of the existing species. All six of the Upper Cretaceous
forms are American where they are widely distributed. By Eocene
times tbc_\- had reached Europe and South America and they are
cosmopolitan in the northern hemisphere throughout the Tertiary,
being especially abundant in the Pliocene of the Mediterranean
region. It would seem as if their Cretaceous origin was occidental,
that they spread over the northern hemisphere during the Tertiary
and became restricted to southeastern Asia, the Canary Islands and
America during the Pleistocene.

The genus Ocotea of Aublet with over 200 existing species is, it
seems to me, composite, and I regard Xee's three genera Mespilo-
daphne, Oreodaphne and Strychnodaphne as distinct. The modern
species of Mespilodaphnc are confined to South Africa and tropical
America. The fossil record is almost entirely merged in the forms
referred to Laurus. I have recognized four well-marked species in
the Wilcox flora. They are abundant types and some range from
the base to the top of the deposits, and along the Wilcox coast from
Mississippi around the head of the embayment and westward to
western Texas.

The genus Oreodaphne has been recognized in the American
Upper Cretaceous and throughout the European Tertiary. At the
present time its numerous species are confined to the American trop-
ics. In the Wilcox it has seven well-marked species, which are
abundant individually, some ranging from Mississippi to Texas and
from the base to the top of the Wilcox. The genus is probably
of American origin and it has been a member of the flora of the
American tropics from the Upper Cretaceous to the present.

The genus Nectandra with about seventy existing species con-
fined to tropical and subtropical America probably has it- geologic

19I4-] SOUTHEASTERN NORTH AMERICA. 221

history entangled with the fossil forms referred to Laurus. It
occurs in the American Upper Cretaceous and the European and
South American Tertiary. There are at least five characteristic Wil-
cox species some of which were abundant along the Wilcox coasts
and some range from the base to the top of the deposits. Like
Oreodaphuc this genus appears to have been of American origin, be-
coming cosmopolitan in the Tertiary and restricted to its original
home during the Pleistocene, where it is still a vigorous and much
differentiated type.

The tribes Eusideroxylese, Litseese, Apolloniese, Acrodiclidieae,
Laureae and Cassytheee do not appear to be represented in the Wilcox
flora although the Litseess are represented in the Upper Cretaceous
of the Mississippi embayment area and the Laureae are common in
the American Upper Cretaceous.

The tribe Cryptocaryese, now largely American, is represented in
the Wilcox by a single well-marked species of Cryptocarya. The
existing species of Cryptocarya number about 40 of which Y\ are
South American and the balance Oriental. Only two or three fossil
species are known. These come from the Tertiary of Australia and
the Pleistocene of Java.

The form genus Laurus which serves to render insecure the
discussion of the geologic history of the preceding genera includes a
very large number of fossil forms of which no less than 25 are
Cretaceous, the oldest being from theAlbianof France and Portugal.
Species of Laurus are abundant throughout North America in the
Cenomanian, ranging northward to Greenland and also occurring in
Europe and Australia. They have over a score of species in the
Eocene and with a similar wide range. The 30 or more Oligocene
species are confined to Europe. Over 30 Miocene species are con-
fined to Europe and America and the score of Pliocene species are
Mediterranean and largely Italian.

J will mention only one other genus since it definitely shows a
past history that is probably typical of a large number of genera of
Lauracese. The genus Sassafras. 42 monotypic and confined to North
America in the existing flora, belongs to a large tribe — the Litseeae,
which today is chiefly Oriental, ranging from Asia through Malaysia

42 See Berry, Bot. Gaz., Vol. 34, 1902, pp. 426-450, tf. 1-4, PI. 18.

222 BERRY— LOWER EOCENE FLORA OF [April 25,

to Australia. Sassafras has well-marked foliar characters of both
form and venation that render it readily recognizable in the fossil
state. Upwards of two score fossil forms have been described. The
oldest of these are three well marked species in the Patapsco forma-
tion (Albian) of the Middle Atlantic slope in Maryland and Vir-
ginia. A species is recorded from this horizon in Portugal but the
identification is very doubtful as is that of a Cenomanian species
described from Bohemia, which latter probably represents the genus
Sterculia. In America on the other hand the genus is widespread
and well differentiated at the base of the Upper Cretaceous, rang-
ing from Greenland along the coast and in the interior to South
America and with about a dozen known species. By Eocene times
Sassafras had reached Europe 43 probably by way of the Arctic
regions, where it has been found throughout the Oligocene and Mio-
cene. In the Pliocene the European forms had retreated southward
but remained common in Italy, France and Spain. The glaciation of
the Pleistocene caused their extinction on that continent, the single
existing species surviving today in the original home of the genus.

The order Myrtales as developed in the Wilcox flora contains 11
species of Myrtacese, 9 species of Combretacea?, 1 species of Trapa-
ceae and 1 species of Melastomacese, as against over 7,000 species
in the existing flora.

The family Myrtacea? has over 3,100 existing species separated
by taxonomists into 2 subfamilies. The first of these the Myrtoide?e
with 32 genera and about 2,400 species comprises mostly tropical
forms of which over 75 per cent, are confined to the western hemi-
sphere. There are over 200 in Asia, one of which extends into
southern Europe, about 75 in Africa, about 200 in Australia, and
about 60 in Oceanica. Nineteen of the genera are confined to Amer-
ica and these include the only monotypic genera in the subfamily, three
in number, as well as large and greatly differentiated genera like
Myrcia with upwards of 450 species. The two other large genera,
Myrtus with 178 species and Eugenia with about 1,300 species, are
the only two genera found on all the continents and in these two
genera America furnishes [35 species of Myrtus and 850 species of
Eugenia, or over 60 per cent. The second subfamily, the Leptosper-

43 A very doubtful form is recorded from Australia.

I9I4-] SOUTHEASTERN NORTH AMERICA. 223

moide?e, comprises the Leptospermae with 28 genera and about 700
species and the Chamseleuciere with 12 genera and about 165 species.
Both of these tribes are even more strikingly Australian than the
Myrtoideae are American. The Chamnaeleucieae are entirely Aus-
tralian and mainly confined to western Australia. The Leptospermae
have a single monotypic genus in Chili and the distribution of the
other members of this tribe suggest the probability that it should
be placed in some other alliance, since with the exception of Metro-
sideros, which is represented in Africa, and the genus Baeckca which
reaches the Asiatic mainland, all of the genera are confined to Aus-
tralia or the surrounding islands southeast of Asia.

In a recent paper Andrews 44 has presented some interesting sta-
tistics of distribution and an ingenious theory of the history of the
family. He considers that the original stock was arborescent or
shrubby with entire, simple, opposite, penniveined leaves with dots
and intra-marginal aerodrome veins ; with the calyx lobes and petals
imbricate, probably in fives : flowers regular, solitary or in cymes ;
stamens indefinite, numerous, free, with versatile, 2-celled anthers ;
ovary inferior with two or more cells ; style simple ; fruit inferior,
crowned with persistent limb of calyx, indehiscent, succulent or fleshy
(rarely dry); albumen none; cotyledons thick and fleshy, with a
short radicle.

From the character of Cretaceous climates this or some other
theoretical prototype flourished in a mesophytic environment
Among modern groups the nearest approach to this theoretical stock
is furnished by the Myrtoideae which are fleshy fruited, most numer-
ous in species, and widely spread in the equatorial regions, with over
75 per cent., however, confined to America. The existing Myrtaceae
with capsular fruits representing the extreme of specialization in the
family are Australian while the Chamaelaucieae standing in an inter-
mediate position between the two preceding groups are almost wholly
confined to western Australia.

These are the facts of modern distribution. Their interpretation
may be various. Andrews (op. cit.) from a study of the present dis-
tribution, geologic climates and the geological history of the Austra-

44 Andrews, E. C, " The Development of the Natural Order Myrtaceae,"
Proc. Linn. Soc. N. S. Wales, Vol. 38, Pt. 3, 1913, pp. 529-568.

224 BERRY— LOWER EOCENE FLORA OF [April 25,

lian region, concludes that the Leptospermoide?e originated from
the Myrtese, and that the Cretaceous forms were widespread which
latter was undoubtedly the case. That before the separation of
Australia from the Asiatic mainland fleshy-fruited forms found
themselves in a region of warm moist climate but relatively poor soil
and that it was this edaphic factor that was the principal stimulus to
the differentiation of the Leptospermoidese, which with the exception
of the genus Mctrosidcros show adaptations to poor soil and temper-
ate or dry climates and this exception explains the relatively wide
distribution of Mctrosidcros from Asia to the Fiji Islands. The
Eucalyptus forms according to the view of this student were derived
from Mctrosidcros after the separation of Xew Caledonia from
Australia and the latter continent from Asia. To support this latter
point Andrews is obliged to consider all of the Cretaceous identifica-
tions of Eucalyptus and all of the Tertiary identifications outside
of Australia as equally misleading. With regard to the presence of
Eucalyptus in Xorth .America I think this contention to be not un-
likely, for although in accordance with paleobotanical usage, I have
identified numerous forms of Eucalyptus in the Xorth American
Upper Cretaceous, I have long thought that these leaves represented
ancestral forms of Eugenia or Myrcia, but have hesitated suggesting
an}- change in nomenclature from the havoc it would play with strati-
graphic paleobotany.

The supposed American Cretaceous fruits of Eucalyptus have
long since been shown to be referable to Dammar a-\ike forms and in
my studies of the Tertiary flora I have scrupulously refrained from
referring any of the numerous myrtaceous leaves to the genus Euca-
lyptus. Regarding the possible occurrence of Eucalyptus in Eu-
rope I am not sure that the identifications of Heer, Unger and Ettings-
hausen are erroneous. Certain remains considered as Eucalyptus
fruits seem very convincing from the published figures and there i-
not the slightest doubt that the other great modern Australian alli-
ance — the Proteacea- — was represented in both Europe and America
during the Cretaceous and Tertiary. There is one additional argu-
ment against the Cretaceous radiation and the paleobotanical deter-
mination of Eucalyptus and that is the great persistence of the
peculiar juvenile, opposite, cordate, sessile and horizontal leave< which

I9I4-] SOUTHEASTERN NORTH AMERICA. 225

*
must represent an ancestral character of long standing before the
evolution of the falcate leaves of the genus with twisted leafstalks
and other xerophytic features. 43

I have dwelt at some length on this question because of its phylo-
genetic importance and the possible bearing of the Wilcox flora on
this point. In considering the morphology of the existing species,
Eugenia has many claims to be considered the most primitive al-
though Myrcia is almost equally old and is certainly closely related to
Eugenia. Among the numerous Cretaceous fossils from North
America now referred to Eucalyptus there is not a single one that
does not exhibit characteristic features of Eugenia or Myrcia, espe-
ciallv the latter, a fact greatly impressed on me in handling a large
amount of recent material during my study of the Wilcox forms.

In the Wilcox flora there are six well-marked species of Myrcia
and four nearly equally well marked species of Eugenia as well as a
single species of Calyptranthes. The latter genus appears also to be
represented in recent collections from the Isthmus of Panama.
Without pursuing the subject beyond the known facts, confessedly
meager, and noting the presence in the Wilcox flora of numerous
Combretaceas and a representative of the great tropical family Melas-
tomacea, largely American in the existing flora, both of which are
families closely related morphologically to the Myrtacese, it would
seem that the known facts, as well as the law of probabilities, sug-
gest America as the original home of the family. That it reached
Europe either by way of Asia or the Xorth Atlantic plateau early in
the Upper Cretaceous and became cosmopolitan before the close of
the Cretaceous. During the late Tertiary this ancestral stock, which
largely coincided with the existing subfamily Myrtoideae, was forced
to withdraw from temperate North America to the American tropics,
where it had originated and to which it has since been so largely
confined. The types peculiar to the Australian region represent the
relics of the Cretaceous radiation with numerous new types evolved
on that continent as Andrews has suggested. This is exactly the

45 See Deane, H., " Observations on the Tertiary Flora of Australia."
Proc. Linn. Soc. N. S. Wales, Vol. 15, 1900, pp. 463-475; Cambage, R. H.,
"Development and Distribution of the Genus Eucalyptus,'' Presidential Ad-
dress, Jour. Proc. Roy. Soc. N. S. Wales. 1913.

PROC. AMER. PHIL. SOC, LIII, 214. O, PRINTED JULY 14, I914.

226 BERRY— LOWER EOCENE FLORA OF [April 25,

reverse of the hypothesis proposed by Deane (op. cit.) but one that
accords far better with the facts not only of geologic history, but
with those of existing distribution.

As is pointed out in the systematic part of this work all of the
Wilcox forms are coastal types closely related to existing American
species of similar habitat. About 150 fossil forms have been re-
ferred to the Myrtacese, one third at least having been described as
species of Eucalyptus. At least half of these occur in the Cretaceous
of all parts of the world, but particularly throughout the northern
hemisphere. They are especially well represented in North Amer-
ica and the possibility that they are ancestral forms of Myrcia or
Eugenia has already been pointed out. A similar widespread distri-
bution but less specific variation characterizes the Eocene forms that
have been referred to Eucalyptus. The Oligocene records are all
European and the Miocene records include both Europe and Asia.

The genus Myrtus has about 24 fossil species, all European, the
majority being almost equally divided between the Oligocene and the
Miocene. The oldest forms are early Eocene but the form-genus
Myrtophyllum Heer has several Upper Cretaceous species in Eu-
rope, America and Australia, as well as Tertiary species in Europe,
Asia and South America.

The genus Myrcia DC. so well represented in the Wilcox flora
has species in the European Oligocene, four species in the early Ter-
tiary of Chili and one in the Pliocene of Brazil.

The genus Eugenia, also prominent in the Wilcox flora, has its
oldest known species in the Dakota sandstone. It is represented in
Europe throughout the Tertiary from the lower Eocene to the
Pliocene.

The genus Callistcmon R. Brown has been identified in both the
Upper Cretaceous and Tertiary of Europe and no less than 25 spe-
cies have been referred to the genus Callistemophyllum Ettings-
hausen. These include Upper Cretaceous forms in America and
Europe, Eocene forms in Greenland and Australia, and numerous
Oligocene and Miocene species in Europe.

Leptospermun\j Leptospermites and Leptospermocarpum havebeen
identified from the Upper Cretaceous and Tertiary of Europe: Tris-

I9M-] SOUTHEASTERN NORTH AMERICA. 227

tania-Yike fruits have been described as Tristanitcs by Saporta from
the lower Miocene of France: the genus Psidium Linne, with about
ioo modern species in the West Indies and Mexico, is represented in
Chili by an early Tertiary species : and finally the genus Metrosi-
deros has been identified in the Atane beds of Greenland and in both
the Oligocene and Miocene of Europe.

The family Combretaceae (Terminaliaceas) embraces about 16
genera and 285 existing species of shrubs or trees and tropical
vines, with simple, entire, coriaceous, persistent, exstipulate, alternate
or opposite leaves. The inflorescence is racemose or capitate and
the flowers are regular, perfect or polygamous, often apetalous. The
stamens are two or three times as numerous as the petals and the
one-celled ovary develops into a drupaceous or berry-like indehiscent
fruit, often crowned with the accrescent calyx and containing a soli-
tary seed without endosperm.

The existing species are all tropical or subtropical, ranging from
34 north to 35 ° south latitude, and a relative large number are lit-
toral or strand types. The various continental areas have the fol-
lowing peculiar species: America 75, Africa 85, Madagascar 36,
Asia 57, Australia 23. About ten or a dozen species are found in
more than one area, there being a remarkable identity between the
American tropics and those of West Africa, the genera Cacoucia,
Conocarpus and Laguncularia having identical species in both
regions.

The geologic history of the family is most incomplete, but it is
exceedingly prominent in the Wilcox flora where it is represented
not only by characteristic leaves but by flowers and fruits. No spe-
cies are certainly known from horizons as old as the Upper Creta-
ceous although a species of Termanaliphyllum has been described
from the Perucer beds (Cenomanian ) of Bohemia and a species of
Conocarpites from the Tuscaloosa formation of Alabama. So far
as I know there are no authentic occurrences as old as those of the
Wilcox. In this flora there are three well-marked species of Com-
bretum, a genus with about 130 existing species found in all tropics
except Australia and Polynesia. Over thirty of these are endemic
in South America and their abundance in the Wilcox as well as the

BERRY— LOWER EOCENE FLORA OF [April 25,

occurrence of a species in the early Tertiary of Chili strongly suggests
that the genus is of American origin. This statement as well as the
determination of the Wilcox species receives confirmation in the re-
markably preserved flower from these beds described as Combretan-
thites. Combretum has been recorded from the Miocene of Switzer-
land and Germany, and from the Pliocene of Italy. It occurs in the
Claiborne group of the Mississippi embayment and Felix has de-
scribed petrified wood from the supposed Eocene of the Caucasus
which he calls Combretacinium.

The genus Conocarpus ( rsertner, a member of the tropical man-
grove association, has a well marked species in the Wilcox flora
which is supposed to be descended from the Conocarpites described
fnun the Tuscaloosa formation in this same general region. Another
species very close to the modern form of the American tropics oc-
cur- in the Claiborne group. Conocarpus fruits have also been de-
scribed recently from the Aquitanian of Rhenish Prussia.

The genus Laguncularia Gsertner, monotypic in the mangrove
association of America and the west coast of tropical Africa, is rep-
resented by both leaves and fruits in the Wilcox flora. The only
other genus of Combretace?e with known fossil representation is
Terminalia Linne It is a large genus in the existing flora with over
one hundred species almost equally divided between America. Asia,
Africa and Australia, several of the species being very wide-ranging
littoral types. There are three Wilcox species, based on both leaves
and fruit. One of the species makes its appearance in the underly-
ing Midway group of the Western Gulf region, possibly representing
the beginning of it> extension northward along the coast in the em-
bayment region from tropical America.

Five < Higocene species of Terminalia have been described from
Europe, the determination- resting on both leaves and fruits, and the
occurences ranging from the Sannoisian to the Chattian and geo-
graphically from southeastern France to Greece. There are seven
well distributed .Miocene species in Europe, as well as Pliocene
specie- in both Spain and Italy along the shores of the Pliocene
.Mediterranean Sea. A supposed Pliocene species is also recorded
from Bolivia.

i9i4.] SOUTHEASTERN NORTH AMERICA. 229

While future discoveries will have to greatly amplify the fossil
record before the history of the family in past times can be traced
with any degree of surety, the remarkable display of these forms
in the Mississippi embayment region, evidently derived from the
American tropics, gives a large amount of probability to the theory
that the family originated in the American tropics during the
Upper Cretaceous.

The genus Trapa Linne, formerly included in the family Ono-
graceae, is now made the type and only genus of the family Hydro-
caryacete (Trapacere, Dumort, 1827). There are three existing
species, all aquatics, and all confined to the old world except for
the naturalization of Trapa natans Linne, in New England and
New York. The latter species is found irregularly scattered through-
out central and southern Europe, its area of distribution being a
contracting one as shown by its occurrence in post-glacial deposits
at very many localities beyond its present range in Russia, Finland,
Sweden and Denmark. The two other existing species are Trapa
bicornis Linne of China and Trapa bispinosa Roxburg of south-
eastern and southern Asia (said also to occur in Africa).

The genus has an extended geological history. Rosettes supposed
to represent the floating leaves (Trapa. f microphylla Lesq., and
Trapa ? cuneata Knowlt.) are widespread in the Rocky Mountain
province in beds of late Cretaceous and early Tertiary age. The oldest
recognizable fruits are a large bi-cornute form from the Eocene of
Canada and Alaska and Trapa zcilco.vcnsis Berry found in the Wil-
cox flora. An Oligocene species ( Trapa Crcdncri ) Schenk has
been described from Saxony, and no less than seven species have
been described from the Miocene — two occurring in Idaho (Payette
formation), one in Japan and the balance in Europe, where two
species continue into the Pliocene. A species from the late Pliocene
of America is found in southern Alabama. The existing Trapa
natans has been recorded from the preglacial beds of England and
Saxony and from very many interglacial and postglacial deposits in
Portugal, Italy, Netherlands, Germany, Sweden, Russia and Den-
mark, Gunnar Andersson in a recent paper (1910) mentioning 18
localities in West Prussia, 6 in Denmark, 17 in Sweden and 29 in
Finland.

230 BERRY— LOWER EOCENE FLORA OF [April 25,

The family Melastomaceae is a relatively large one with about
150 genera and over three thousand existing species. It is almost
strictly tropical although some members range southward to 40
south latitude. It is a typically American family, seven of the fifteen
tribes into which the family is divided being confined to tropical
America, and about 2,500 of the existing species being also endemic
in this region. While the geologic history of this vast assemblage of
forms is practically unknown, there is no evidence to disprove the
theory that it, like the allied families Combretaceae and Myrtacese,
had its origin in that most prolific region — the American tropics.

The few fossil forms that have been found, including leaves,
flowers and calices, have been referred to the form-genus Melasto-
mites first proposed by Unger. A doubtfully determined species,
which probably belongs to the Lauraceas, had been recorded from the
Upper Cretaceous of Westphalia. The only known Eocene species
is the well-marked form present in the Wilcox flora. Four Oligocene
species have been described from Bohemia, Styria and Egypt; four
Miocene species from Switzerland, Prussia and Croatia; and a
Pliocene species from Italy.

The order Umbellales (Umbelliflorae of Engler) includes but
three families — the Araliacese, Umbelliferae and Cornacese, together
with upwards of 3,000 existing species of which more than two-
thirds belong to the Umbelli ferae. The three families are closely
related and stand somewhat apart from the rest of the choripetalous
orders. While undoubtedly there has been great specific variation in
very modern times especially among the herbaceous forms of Umbel-
li ferae, some members of the alliance go back as far as undoubted
dicotyledons have been found, and this fact is one of the strongest
arguments for considering its relationships to the Gamopetalae to be
less close than some botanists have suggested, a suggestion based
primarily on a consideration of the floral structures apart from the
morphological features of the whole plants As regards floral evo-
lution the Umbellales clearly mark its highest expression among the
Choripetalne and parallel the Gamopetalne. The flowers are epigy-
nous, with cyclic stamens, reduced carpels, and often reduced sepals.
The Araliaceae and Cornaceae are both positively and the Umbel-
li u-ra- doubtfully represented in the Wilcox flora.

I9I4-] SOUTHEASTERN NORTH AMERICA. 231

The family Araliaceas contains about 52 genera and 500 existing
species, chiefly inhabitants of the tropics, the notable exceptions to
this statement being in North America and eastern Asia. The
modern center of development is in the Asia-Australia region, no
less than S3 genera being confined to Asia, Malaysia, Australia or
Polynesia. Africa has three peculiar genera with about 30 species ;
America has five peculiar genera with about 100 species. The genus
Schefflcra is cosmopolitan. Hedera and Polycias occur in Eurasia
and Africa. Two genera are common to Asia and America and
Aralia adds Australia to a similar distribution. Pseudotenax with
about six species is peculiar to western South America and New
Zealand.

The fossil record is not nearly as complete as it should be to
afford a secure basis for generalizations. A number of genera are
found, however, in the oldest deposits in which undoubted dicoty-
ledons are known. The largest genus is Aralia, commonly used by
paleobotanists as a form-genus for generically unidentified species
of Araliaceae, rather than for forms falling within a strict modern
definition of Aralia. No less than fifty species of Aralia have been
described from the Cretaceous. Two of these come from horizons
as old as the Albian of Portugal. In beds of similar age in eastern
America (Maryland and Virginia) there are two well-marked species
referred to Aralicephyllum and clearly ancestral to the numerous
species of Aralia so common in the Upper Cretaceous of the latter
region. Very similar, in some cases identical, forms are found in the
Cretaceous on both sides of the Atlantic. There are fifteen species
in the Perucer beds (Cenomanian) of Bohemia and Moravia and a
like number in the Dakota sandstone of the western United States,
while along the east coast there are nine species in the Raritan forma-
tion, eight in the Magothy formation, and one each in the Black
Creek formation of North Carolina, the Eutaw formation of Georgia,
the Tuscaloosa formation of Alabama and the Woodbine sand of
Texas. In Greenland there are two species in the Atane beds and a
third in those of Patoot. In the younger Cretaceous there are two
species in Bohemia, two in Westphalia and one in Colorado. Aus-
tralia has a species and ten supposed varieties of Aralia in the Upper

232 BERRY— LOWER EOCENE FLORA OF [April 25,

Cretaceous beds of that country. In addition to the foregoing dis-
play, the allied genus Araliopsis ( Berry 191 1) has a number of well-
marked species in the Raritan. Magothy and Dakota formations, so
that it must be conceded that the araliaceous stock was well differ-
entiated and cosmopolitan before the close of the Cretaceous. 40

There are over a score of Eocene species of Aralia, they being
especially common in the Fort Union of the western United States,
the Paleocene of Belgium, and the Eocene of Australia. The three
Wilcox species are not common— two of them are common Fort
Union species and the third was described originally from western
Greenland. In addition there are species in the Denver formation,
the Green River formation, in Oregon, Xew Zealand, Italy, and the
south of England.

There are upwards of twenty Oligocene species, especially in the
Sannoisian of southeastern France from which 14 >pecies have been
described. All of the other Oligocene'records are also European.

There are also about twenty Miocene species distributed over
Xorth America. Europe and Asia. Some of the California species,
e. </.. Aralia Whitneyi, are clearly ancestral to existing Asiatic east-
coast forms. A fruit ( . Iralicecarpum ) is described from the Miocene
of Prussia. There are in addition between 15 and 20 fossil species
of . Inilid more or less doubtfully connected with other genera of the
family, c. </., there is a species of Arthrophyllum doubtfully identified
from the upper Oligocene of France; a species of Cephalopanax ( ?)
is recorded from the lower Miocene of France: several forms of
Sciadophyllmn ( ?) occur in Greenland, Bohemia and France; and
Paratropia ( ?) is recorded from the Paleocene, Oligocene and Mio-
cene of France and the Miocene of Bohemia.

There are two species of Oreopanax in the Wilcox flora, one of
them exceedingly well marked and clearly referable to the section
Digitatse of Oreopanax. The latter genus lias about eighty existing
species witli simple, lobate and digitate leafed sections confined to
tropical America but present in the Taleocene, Tongrian and Aqui-
tanian of France. The modern Asiatic genus Acanthopanax De-
caisne and Planchon has ( )ligocene species in France and Germany,
and a Miocene specie- in Japan.

ee Berry, "Aralia in American Paleobotany," Bot. Gas., Vol. 36, [903,
pp. 421^428.

I9I4-] SOUTHEASTERN NORTH AMERICA. 233

The genus Panax Linne with about six existing species in Asia
and North America has furnished a number of fossil forms based
on numerous characteristic fruits as well as leaves. It is represented
from Greenland to Alabama along the west coast of the Atlantic
and in the Perucer beds of Bohemia (Araliphyllum) . It has five
species in the Oligocene of Europe and six Miocene species in
Europe and Colorado. The genus Cussonia Thunberg with about
25 African species in the existing flora is doubtfully recorded from
the Albian of Portugal. It is present in the Perucer beds of Bohemia
(Cussonipkyllum) and in the Oligocene of France and Greece.

The genus Hedcra Linne with only three existing species of
Europe, Asia and Africa has numerous and well-defined fossil
forms. 47 Xo less than fifteen have been described from the Upper
Cretaceous of both America and Europe. There are about seven
Eocene species in Greenland, Alaska, the Fort Union of the western
United States, and in the Paleocene of Belgium and France. The
genus remains common during the Tertiary in Europe and is present
in America as late as the Upper Miocene lake of Florissant, Colorado.
The anecestor of the existing Hedcra helix Linne occurs in the
Pliocene of central France and the modern form itself is found in
the Pleistocene of England, France and Italy. A species of Polyacias
occurs in the Pleistocene of Java associated with Pithecanthropus
erectus.

The family Umbelliferas with 170 genera and upwards of two
thousand existing species is distinctly an extratropical family with
numerous boreal forms. The majority are herbaceous and of rela-
tively modern origin. It is very sparingly and doubtfully represented
in the fossil state and the only Wilcox form that suggests such an
affinity is the fruit described as Carpolithits prangosoides which
greatly resembles those of the existing genus Prangos Lindley.

The third family of the Umbellales, the Cornace?e, is a relatively
small one, with only sixteen genera and about 100 existing species,
mostly of the temperate zone. The fossil forms are confined to the
two genera Cor nits and Xyssa. Cor nits has about 40 existing species
of herbs and small trees mostly confined to the north temperate zone

47 The forms from the Potomac group of Maryland and Virginia de-
scribed by Fontaine as species of Hedercephyllum are entirely worthless.

23-4 BERRY— LOWER EOCENE FLORA OF [April 25,

in Eurasia and North America but represented in Mexico, and with
a single species in Peru. Over fifty fossil species have been de-
scribed. There are at least twelve in the Upper Cretaceous, all con-
fined to North America and ranging from Greenland to Alabama.
There are about a dozen Eocene species in America, Europe, and the
Arctic, one of these is sparingly represented in the Wilcox flora.
Oligocene records are few in number but over 25 Miocene species
have been described, the genus being particularly abundant at this
time throughout central Europe but also represented in both North
America and Asia. About five Pliocene species are recorded from
Spain, France, Italy and Japan and the genus has afforded Pleisto-
cene material in New Jersey, Holland, England, etc.

The genus Nyssa Linne (including also Nyssidium Heer and
Nyssites Geyler and Kink.) comprises about seven existing species
ranging from shrubs to large trees, natives of southeastern North
America and eastern and central Asia. It has furnished over fifty
fossil forms, the majority being based on the characteristic costate
stones. The oldest known forms are from near the base of the
Upper Cretaceous (Dakota, Tuscaloosa) of North America. By
Eocene time Xyssa had reached Alaska, Greenland and Europe.
There are two characteristic species in the Wilcox, both based on
stones, and a third occurs in the overlying deposits of the Claiborne
group. In the lignite deposit of Brandon, Vermont, of uncertain
but probably early Tertiary age, no less than eighteen so-called
species of stones have been described, and while doubtless the specific
differentiation is overrefined, it emphasizes the abundance of Nyssa
in New England at that time. Xyssa is abundant in the European
Oligocene and there are Miocene species in New Jersey, Virginia,
Europe, and Asia ; and a Pliocene species occurs in Alabama. Some
of the modern species are common in the Pleistocene of this country
from New Jersey southward.

While much remains to be learned regarding the history of the
Cornacea.' it seems clear that the two genera Coniits and Nyssa which
have yielded fossil forms are both types that appear to have origi-
nated in North America during the Cretaceous.

No family of the Choripetal?e has succeeded in maintaining a

i9 14.] SOUTHEASTERN NORTH AMERICA. 235

world-wide distribution as have several families of Monocotyledons
and Gamopetalse. No distinctly boreal group has been developed as
among the Gamopetalse (Ericales). Certain great families charac-
terize the north temperate region and these are all herbaceous forms
believed to be of relatively recent origin, e. g., Polygonacese, Caryo-
phyllacese, Cruciferre, Saxifragacese, Onagraceae and Umbelliferae.
While aquatic forms are common this habit does not characterize
whole families as among the Monocotyledons. The Choripetals
predominate in the American tropics and many of the families
present in the Wilcox flora have been shown to have probably origi-
nated in that region.

The second grand division of the Dicotyledons, the Gamopetals
(Sympetalae), constitute a rather well-defined group, presumably de-
rived from the Choripetalae, and characterized by a complete cyclic
arrangement of the floral parts, a usually gamopetalous corolla,
ovules with a small nucellus and usually a single integument. It
contains nine or ten orders and upwards of 50,000 existing species.
The majority of the orders appear to be more compact and natural
groups than the corresponding alliances among the Choripetalae.
The Ericales, Primulales and Ebenales are pentacyclic and isocarpous,
while the Gentianales, Polemoniales, Personales, Plantaginales, Rubi-
ales, Valerianates and Campanales are tetracyclic and anisocarpic,
the last three orders being epigynous.

The alliance predominates in herbaceous forms and several of
the families are distinctly boreal. While the Composite, Labiats
and Plantaginaces are of world-wide distribution there are no
notable continental pairings such as is usually the result of an
extended geologic history. These and many other facts suggest that
the Gamopetals as a whole, especially the more evolved, herbaceous,
extratropical families, are of relatively modern origin whose major
specific differentiation was concomitant with the occupation of the
temperate zones after the retreat of the Pleistocene ice-sheets.

From the viewpoint of floral structures the so-called Composite
are clearly the culmination of the evolution of floral structures. This
is shown not only by their gamopetaly, epigyny, connivent anthers,
and the formation of seedlike fruits with a pappus, but by the com-
plex flowerhead, the prevalence of diclinism, the dimorphism of the

23G BERRY— LOWER EOCENE FLORA OF t A P ril 2 s,

corollas and other >pecial features. This theorem is corroborated
by the in general modernness of the alliance.

Six of the Gamopetalous orders are represented in the Wilcox
flora. The first of these, the Primulales, in its fullest development in
existing floras includes the three families Myrsinaceae, Primulacea?
and Plumbaginaceae. They are structurally much alike with a single
cycle of stamens opposite the petals, and a unilocular ovary with a
free central placenta. This community of Moral organization can
only be attributed to convergence and not to filiation since the Myr-
sinaceae are old forms which in modern floras are predominantly
tropica] and American while the Primulaceae are chiefly north tem-
perate and boreal herbs of relatively recent evolution: and the Plum-
baginaceae are very modern halophytic herbs and undershrubs of
salt beaches and steppes, the majority being found in the Mediter-
ranean and Caspian regions.

The Myrsinaceae, the only family represented in the \\ ile ix flora,
is characterized by alternate, simple, coriaceous, punctate, exstipulate
leaves ; perfect, regular flowers ; and single seeded drupaceous fruits.

The family contains about thirty genera and 530 species of shrubs
or trees, largely tropical and predominantly American. Thus eleven
genera containing upward of 200 species are peculiar to America
while there are only four genera with less than a dozen species
peculiar to Asia, and three genera with about 100 species peculiar
to A frica.

The genus Myrsine Linne is found on all the continents except
Europe and in Polynesia. Its distribution is extratropical in the
African region. Euardisia Pax is found in all tropics. Maesa For-
skal is found in all oriental tropical countries as is also the monotypic
genus 2Egiceras Gaertner, a member of the coastal mangrove associa-
tion. The genus Cybianthus Martins, largely South American, has
species in the Philippines and in New Grenada. There is little that
is significant in the recent distribution of the family and the fossil
record is very incomplete.

< >ver seventy-five fossil forms have been referred to Myrsine.
The oldest are the seven or eight forms recorded from the Upper
( Iretaceous. All of the older of these (Cenomanian) are from North
America and onlv one from the Turonian of Bohemia occurs in the

I9I4-] SOUTHEASTERN NORTH AMERICA. 237

European Upper Cretaceous. The American forms are not varied
specifically but are wide ranging and common, extending from the
Atane beds of Greenland along the Atlantic coast to the Tuscaloosa
formation of western Alabama as well as in the Dakota sandstone
of the western interior.

The recorded Eocene species of Myrsine number seven or eight
and include an Australian form, one in the early Eocene of Alum
Bay, three in the upper Eocene of France, and two in western xAlaska.
Myrsine is exceedingly varied and abundant during the Oligocene
throughout southern Europe, over thirty species having been de-
scribed, of which eleven occur in the basal Oligocene of southeast-
ern France ( Sannoisian). There are upwards of thirty Miocene
species throughout Europe, one in Colorado being the only known
American occurrence. Several species linger in the Pliocene of south-
ern Europe in France and Italy and one species is present in the Pli-
ocene of Brazil. 'In addition to the forms referred to Myrsine sev-
eral forms from the European Tertiary have been referred to the
form-genus Myrsinites. Ettingshausen recorded a species of Pleio-
merites from the Miocene of Bohemia; and the genus Maesa For-
skal, which has about 40 modern species in Asia, Africa, Australia
and Polynesia, is represented in the Oligocene of Transylvania and
Egypt and in the Miocene of Styria.

The genus Ardisia Swartz (including Ardisiophyllum Geyler)
has furnished about a dozen fossil species, the oldest of which, a very
doubtfully determined form, comes from the Turonian of Bohemia.
There is an Eocene or Oligocene species in Chili, three Oligocene
species in Bohemia and one in Transylvania. There are four Mio-
cene species in France, Bohemia and Styria ; and Pliocene species in
Italy and Borneo.

The genus Icacorea Aublet is the only member of the Myrsinacese
found in the Wilcox flora. The genus has numerous existing spe-
cies confined to South America. The fossil record is meager but
includes two or three species of the European Oligocene. The Wil-
cox species is thus considerably older than any European occurrence.
It represents a form which is very close to the modern Icacorea pan-
iculata Sudworth, a shrub or slender tree of the Florida keys, Baha-

238 BERRY— LOWER EOCENE FLORA OF [April as,

mas, Cuba and the east coast of southern Mexico. In addition to
the foregoing records at least four kinds of flowers have been de-
scribed from the Baltic amber (Sannoisian). These are Berendtia
Goeppert (2 species), Myrsinopsis Conwenz, and ScnacUa Gceppert.

While the geologic history of the family is thus so incomplete it
is not without significance in this case as in the case of so many fami-
lies previously discussed, that a predominantly American family in
the existing flora has its oldest known fossil occurrences in the basal
Upper Cretaceous of Xorth America.

The order Ebenales includes the families Sapotacese, Ebenacea?,
Styracace?e and Symplocaceas, together with upward of one thou-
sand existing species, the larger families being the Sapotaceae and
Ebenacere, both of which are represented in the Wilcox flora, while
the other two families are sparingly represented in the European
Tertiary. There is considerable range in floral structures from
indefiniteness in the number of stamens and carpels and polypetaly,
to a 4 to 8 cyclic arrangement, which leads floral morphologists to
consider the order as among the most primitive of the Gamopetalae.

The family Sapotacese comprises trees or shrubs with a milky
juice and with alternate, simple, entire, mostly coriaceous, petiolate,
exstipulate leaves. It contains about thirty-two genera and nearly
four hundred existing species of all tropical countries. About half
of the existing species are American. There are eleven genera con-
fined to America, seven to Africa, three to Australia, two to New-
Caledonia, two to Asia and Malaysia, two to Malaysia and one to
Asia. The three large genera, Sidcroxylon, Chrysophyllum and
Mimusops, are represented in all tropical countries. There are four
genera and twelve species represented in the Wilcox flora. The
largest of these genera is Bumclia Swartz with six well-marked Wil-
cox species. Bumclia with about a score of species is confined to
America in the existing flora, ranging from the southern United
States through the West Indies and Central America to Brazil. It
has numerous fossil species, the oldest coming from the Upper Cre-
taceous (Dakota sandstone) of the western interior. In addition
to the six W'ilcox species, which are prototypes of still existing form-,
there are two Eocene species (Ypresian) in southern England.

IQI4-] SOUTHEASTERN NORTH AMERICA. 239

There are about a dozen Oligocene species, ten of which are wide-
spread in Europe, one is found in the Apalachicola group of west-
ern Florida and two forms, representing both leaves and fruit, are
found in the Vicksburg group of Louisiana and Texas. There are
seven or eight Miocene species widespread in Europe and one is re-
corded from the late Miocene of Colorado.

The genus Chrysophyllum Linne with about sixty existing species
found in all tropical countries, but the majority American, has a
supposed species in the Upper Cretaceous of Saxony (Nieder-
schcena) ; a well marked species in the Wilcox flora; three Oligo-
cene and six Miocene species in Europe.

The genus Mimusops Linne with about 40 existing species in all
tropics has three well-marked Wilcox species and a fourth in the
overlying Claiborne deposits. To it has been referred a species from
the Upper Cretaceous of Saxony (Niederschcena) and it is undoubt-
edly represented in the Upper Cretaceous of the embayment region
as well as elsewhere by the leaves that have been referred to the
form-genus Sapotacites.

The genus Sideroxylon Linne, with about eighty existing species
in the oriental tropics and about fifteen in the American tropics,
has two species in the Wilcox flora which are the oldest thus far
discovered. To this genus have been referred four Oligocene and
one or two Miocene species from Europe.

Isonandra Wright a small modern genus of the Malayan region
is represented in the Tertiary of Borneo by I sonandrophyllum Gey-
ler; the genus Achras Linne {Sapota Plumier), now monotypic in
the West Indies, has three species in the European Miocene ; Laba-
tia Swartz, with six existing species in the American tropics, has
been doubtfully determined in the Miocene of Prussia and Italy;
and Felix has described two forms of petrified wood which he refers
to this family under the name Sapotoxylon, one species from Ger-
many and the other from an unknown locality and horizon.

A large number of fossil forms of Sapotaceae have been referred
to the form-genus Sapotacites proposed by Ettingshausen (also Sapo-
tophyllum). There are at least ten Upper Cretaceous forms wide-
spread in North America and represented in Europe in the Perucer

240 BERRY— LOWER EOCENE FLORA OF [April 25,

beds of Bohemia and the Credneria stage of southern Saxony (Ceno-
manian). Three of these Upper Cretaceous forms are from the
Tuscaloosa formation of Alabama and undoubtedly represent the
ancestors of some of the Wilcox forms. There are about ten re-
corded species of Sapotacites in the Eocene of Australia, France and
southern England. There are about a score of species in both the
Oligocene and Miocene, most of which are European. There is,
however, an undescribed species in the Apalachicola group of west-
ern Florida. In the Pliocene there are species in southern Europe
and on the island of Java.

Notwithstanding the incompleteness of the record it is obvious
that the family became well differentiated during the Upper Creta-
ceous and while it would not be safe to assign its place of origin to
the American region, it is probable that at least several of the genera,
such as Bumclia for example, originated in this region.

The family Ebenaceae includes about eight genera and upwards of
three hundred existing shrubs and trees, of which over half are
referred to the genus Diospyros Linne. The family is mainly trop-
ical as are most of the species of Diospyros, but the latter is repre-
sented in the north temperate zone in eastern North America, east-
ern Asia, and the Mediterranean region. The three modern mono-
typic genera, Tetraclis, Brachynema and Rhapidanthe are confined
respectively to Madagascar, Brazil and West Africa and none have
been found fossil. The genus Roycna is mostly South African;
Euclca is entirely confined to Africa: Mdba, a large genus, ranges
from Africa eastward to Polynesia; and Macreightia is common to
tropical Africa and America.

Diospyros with about 180 existing species is cosmopolitan. Be-
tween 90 and 100 fossil forms have been described. In that grand
display of dicotyledonous genera which during the mid-Cretaceous re-
placed the old Mesozoic flora of ferns, cycads, and conifers and which
appeared with such apparent suddenness at a number of points in the
northern hemisphere, we find unmistakable evidence of the abun-
dance and wide distribution of species of Diospyros. No less than
seventeen different forms have been described from the rocks of this
age, and the localities where they have been found are scattered

i9u.] SOUTHEASTERN NORTH AMERICA. 241

from Australia to Bohemia, Greenland, and Vancouver Island. A
large majority of these species are American, and they seem to have
been especially at home along the Cretaceous coast of the Atlantic
and along the border of the Mediterranean sea which extended north-
westward from the Gulf of Mexico over much of our present Great
Plains area. One of these species, well named Diospyros primceva
by Professor Heer in 1866, is especially widespread and abundant,
being found not only in Iowa, Kansas, and Nebraska in the west but
also from Texas eastward through Alabama and northward in South
Carolina, Xorth Carolina, Maryland, Xew Jersey, Long Island and
Greenland, or from latitude 33 to latitude 71 ° north. That these
early persimmons were not very different from those of today is
shown by their similar foliage. This resemblance is also shown by
the fossilized remains of the calices of various species. One of these
calices from another early Cretaceous species, recently described by
the writer, is Diospyros vera, found in what is known in the Potomac
River valley as the Raritan formation. Apparently the habit of
accrescence had not been fully formed but the calyx was persistent
then as now and entirely like a modern calyx in appearance. It was
four-parted as it usually is in existing persimmons, but other fossil
forms had a five-parted calyx like a good many present day tropical
species.

In the Eocene epoch, which succeeded the Cretaceous, the records
of the fossil occurrences of Diospyros show that it was truly cos-
mopolitan. These records include about 20 species in Siberia,
Alaska and Greenland on the north ; Canada, various localities in
Europe, as well as Colorado, Montana. Wyoming, Nevada, Ore-
gon, Washington, and other western states. Unfortunately, we have
no Eocene or later Tertiary records along the Atlanic coast of North
America outside the embayment region since the preserved deposits
are all of marine origin and contain no fossil plants. There is little
doubt, however, that Diospyros continued to be an abundant element
in the aborescence flora of this area.

There are two well-marked species of Diospyros in the Wilcox
flora, one of which continues in this region through the Claiborne.

PROC. AMER. PHIL. SOC, LIII. 214, P, PRINTED JULY 14, I9I4.

242 BERRY— LOWER EOCENE FLORA OF [April 25,

A large calyx is present in the Claiborne or Vicksburg of southwest-
ern Texas.

There are about 24 Oligocene species, Diospyros being especially
common throughout southern Europe. There is an American spe-
cies of this age in the Apalachicola group of western Florida. The
luxuriant forests of the Miocene have furnished about twenty spe-
cies of Diospyros, the known distribution at this time includes Euro-
pean localities from Spain to Hungary and American records in
Oregon, California. Yellowstone Park and Colorado. There are
seven Pliocene species in southern Europe and in Java.

The allied genus Royena Linne has furnished splendidly pre-
served fruits from the oasis Chargeh in Egypt ( Upper Cretaceous )
as well as four Oligocene and two Miocene species in Europe. It
seems never to have been cosmopolitan like Diospyros, since it has
never been recognized in the western hemisphere. The fossil his-
tory of the genus Euclea Linne was evidently similar to that of Roy-
ena, i. e., it makes its appearance in the basal Oligocene of Europe
where it is represented throughout the Oligocene and Miocene, be-
coming confined to Africa in Plio-Pleistocene times.

The genus Macreightia DC. has nine or ten existing species, one
occurring in tropical Africa and the balance being American. Mac-
reightia is represented by both leaves and flowers in fossil floras and
it has been a favorite receptacle for tripartite calices, not always
of assured botanical identity. The oldest form is one in the German
Oligocene and there are five or six species in the European Miocene.
It has not been definitely recognized in North America, although
some of the Wilcox material is not unlike some European material
referred to Macreightia. Felix has recognized wood of this family
{Ebenoxylon) in the Oligocene of the Island of Antigua.

The order Gentianales (Contorts of Engler) includes six fami-
lies with between four and five thousand existing species, the largest
family being the Asclepiadaceae with upwards of two thousand spe-
cies. The families are complexly interrelated among themselves
and with the next two orders, about the only constant characters
being the opposite leave- and the generally twisted corolla in aestiva-
tion. The Asclepiadaceae, not found in the Wilcox, share- with the

i9 14-] SOUTHEASTERN NORTH AMERICA. 243

Apocynaceae in the development of a latex-system and in other spe-
cializations, and the elaborate contrivances for entomophily in the
former family reach a degree of complexity almost comparable with
that of the Orchidacere. The Loganiaceae, also not represented in
the Wilcox flora, are lianas characteristic of South America and i\sia
and regarded by Engler as relatively primitive and possibly the an-
cestral stock of the Gentianales and Rubiales. The order as a whole
is numerically massed in the tropics by reason of the many tropical
genera of the two largest families — the Asclepiadaceae and Apocy-
nacese, which together contain three fourths of the existing species
of the order.

The family Oleacese, sometimes considered as an order, the
Oleales, contains 21 genera and about 400 existing species. There
are three small genera peculiar to Asia and four peculiar to America,
the remaining fourteen genera being found in more than one conti-
tinental area. The three largest genera Fraxinns ( 40), Mayepea ( 50)
and Jasminum (160) are all cosmopolitan. Eight of the twenty-one
genera have been found fossil and it is evident that the family has an
extended history, although there are no known Cretaceous records
worthy of credence. Nor is the record well enough known to war-
rant generalizations. It is obvious from the early Eocene occur-
rence of leaves of Fraxinns associated with characteristic fruits, that
the family must have been evolved before the close of the Upper
Cretaceous but none of the genera have any well-marked or abun-
dant known representation until Tertiary times.

The genus Fraxinns Linne has two species in the Wilcox flora,
a characteristic samara, and foliage identical with that described
by Heer from western Greenland as Fraxinns Johnstrupi. The lat-
ter furnishes an interesting instance of the extended distribution of
members of the Eocene flora, at the same time illustrating the north-
ward radiation of floras during the Eocene. Upward of ten addi-
tional Eocene species are known all of which are American and rang-
ing from Tennessee to Alaska and Greenland. The Oligocene marks
the appearance of the genus in Europe from which time to the pres-
ent the genus has been represented throughout the warmer parts of
the north temperate zone, at least four of the existing species mak-
ing their appearance in the Pleistocene.

244 BERRY— LOWER EOCENE FLORA OF [April 23,

The second genus represented in the Wilcox flora is Osmantluts
Lour. It has about ten existing species of eastern North America,
eastern Asia and Polynesia. The Wilcox species is exceedingly
close to Osmanthus americanus Vj and II of the Atlantic and Gulf
coasts from North Carolina southward. A second fossil species is
found in the Miocene of Florissant, Colorado.

The old world genus Phillyrea Linne is found fossil in Europe;
the genus Notelcea Vent., which has six existing Australian species
and an isolated remnant of its former distribution in Madeira and the
Canary Islands, is represented in the Eocene, Oligocene and Miocene
of Europe ; the genus Olca Linne with over thirty existing species
about equally divided between Africa, Asia, and Australia and Poly-
nesia, has about twenty fossil forms (including Olcophyllum Con-
wenz and Olccccarpum Menzel) in Europe where they range in age
from the basal Eocene through the Oligocene, Miocene and Pliocene
to the Pleistocene. The genus is not known in American fossil
floras but there is a supposed species in the early Tertiary of
Australia.

The genus Ligustrum Linne with about 35 existing species in
southeastern Asia and the East Indies has three species in the Oli-
gocene and Miocene of Europe. 48 Saporta has described represent-
atives of the genus Syringa Linne from the Sannoisian of south-
eastern France, the occurrence of the latter genus being based on
floral remains.

The family Apocynacese comprises 133 genera and between ten
and eleven hundred existing species of perennial herbs, vines, shrubs
and trees, mostly with a milky acrid juice and simple exstipulate
leaves. The fruit is usually a pair of follicles or drupes and the
seeds are often comatose. The family is almost equally divided into
two subfamilies, the Plumeroidea? having 68 genera and about 550
species and the Echitoidere having 65 genera and about 500 species.
The genera Plumeria Linne with about 40 species, and Raiavolfia
Linne with about 45 species, are cosmopolitan, mostly tropical ; and
24 genera with about 300 species occur in more than one continental
area. America with 36 peculiar genera containing about 325 species

48 A species of Ligustrum recorded by Hollick from the Upper Cretaceous
of Long Island is probably a Piscina.

I9I4-] SOUTHEASTERN NORTH AMERICA. 245

heads the list, followed by Africa with 28 peculiar genera containing
about 130 species, and Asia with 20 peculiar genera containing about
75 species. Australia has few endemic genera or species, but numer-
ous genera range from Asia or Africa to the Australian region and
several genera are peculiar to Malaysia and to Polynesia. In the
present state of our knowledge the distribution does not furnish
material for generalization.

The fossil record, although including the representatives of at
least a dozen genera, is too incomplete to shed much light on the his-
tory of the family or its existing distribution. The largest fossil
genus is the form genus Apocynophyllum proposed by Heer and em-
bracing fossil forms resembling Thevetia, Cerbera, Apocynum and
other existing genera of the family. Five species are recorded from
the Upper Cretaceous, coming from the Dakota sandstone, Australia,
Westphalia and Saxony. There are over a score of Eocene species
widely distributed. There are five species in the Wilcox flora some
of which are exceedingly well marked and common. There are
also five species in the Ypresian of southern England. Other Eocene
records include Greenland, Australia, New Zealand and Chili. The
score or more of known Oligocene species are confined to European
localities. The Miocene species number about 25, all confined to
Europe except a form recorded from Italy.

Fossil forms have been sparingly referred to the following
genera: Allamanda, Hccmadictyon and Thevetia have been recognized
by Engelhardt in the early Tertiary of Chili: Alyxia, Alstonia, Cer-
bera and Tabemccmontana have been recognized in the European
Tertiary by various students : the genus N eritinium Unger has four
or five species in the European Miocene : the genus Plumeria has
four Miocene species in Europe and a Pliocene species in Brazil. The
genus Echitonium Unger has over a dozen fossil species. There are
five in the Eocene including a well marked form in the Wilcox flora ;
two in the Oligocene and five in the Miocene of Europe.

The genus Nerium Linne has only three or four existing species
of shrubs or trees in the warmer parts of Eurasia. However the
commonly cultivated Nerium oleander Linne of the Levant grows
to a relatively large size and is extensively naturalized in Florida and
the West Indies. It is used for hedges in Bermuda. Saporta re-

24(5 BERRY— LOWER EOCENE FLORA OF [April 25,

corded an Upper Cretaceous species, Nerium Rohlii, from the Cam-
panian of Westphalia but this is almost certainly a member of the
Myrtaceae and not a Nerium. Undoubted species do occur in the
Eocene of Europe, including the remains of a characteristic flower
from the Paris basin. There are several Oligocene and Miocene spe-
cies in Europe and the existing Nerium oleander or its immediate
ancestor occurs in the Pliocene of southern Europe in France and
Spain. The Wilcox species Apocynophyllum tdbellarum is very sug-
gestive of Nerium but the genus is not certainly known in the
western hemisphere.

It may be noted that with the exception of the not certainly
identified species of Apocynophyllum the family is not represented in
the abundant known Upper Cretaceous floras of the world, which
might mean that it originated somewhere in the southern hemisphere.

The order Polemoniales or Tubiflorre 49 contains the four families
Convolvulaceae, Polemoniaceae, Hydrophvllaceae and Borraginaceae.
The first three are characteristically American, the Convolvulaceae
being chiefly tropical, while the largest family, the Borraginaceae, is
typically developed in the north temperate zone.

The family Borraginaceae, the only one of the order known in the
Wilcox flora, contains about 85 genera and 1,600 existing species of
mostly widely distributed north temperate herbs and shrubs, or
trees in tropical countries, characterized by alternate, exstipulate,
mostly entire leaves. The known fossil forms are few in number
and of slight significance and comprise for the most part Tertiary
remains described as species of Borraginites and Heliotropitcs. The
family is represented in the Wilcox by two species of Cordia, a genus
containing about 230 existing species of shrubs and trees of the war-
mer regions of both hemispheres, especially the western. There is a
species in the Upper Cretaceous of the Misssissippi embayment area
(Tuscaloosa formation) and a Miocene species in Europe. Early
Tertiary forms are recorded from Chili by Engelhardt and from
Tasmania by Ettingshausen. The slight evidence available indicates
that the genus originated in the American tropics and that the bulk
of the family is of late Tertiary origin.

* 9 Not tin- Xubiflorac of Engler which includes the orders Polemoniales
and Personales, here regarded as distinct.

I914-] SOUTHEASTERN NORTH AMERICA. 247

The order Personales or Labiatiflorse includes sixteen families dis-
tinguished from the Polemoniales by the zygomorphism of the flow-
ers. The specific differentiation is great and the lines of descent
are confusing. The largest families are the Labiatae with over 3,000
existing species, the Scrophulariaceaa with about 2,500, the Acan-
thaceae with about 2,000, and the Solanacese with about 1,800. Two
of the sixteen families, the Yerbenaceas and Solanaceas, are repre-
sented in the Wilcox flora.

The family Verbenaceaa includes about 73 genera and 1,300 ex-
isting species of widely distributed herbs, shrubs, or in tropical coun-
tries trees. The family is largely tropical or subtropical and is
notably represented in the South American region. The fossil record
is most incomplete. The largely old world genus Clerodendron
Linne is unmistakably present in both the Eocene and Oligocene of
Europe, and Ettingshausen has referred, somewhat doubtfully de-
termined forms from the European Miocene to the American genus
Petrce Linne and to the cosmopolitan genus Vitex Linne. The
genus Citharcxylon Linne has about twenty existing species rang-
ing from the Florida keys and lower California through the Amer-
ican tropics to Bolivia and Brazil. A single species found in the
middle and upper Wilcox is extremely close to the existing Citharex-
ylon villosum Jacquin, a small coastal tree of the Florida keys, Ba-
hamas and Antilles. With the exception of one or two doubtfully
determined forms in the Miocene of southeastern Europe it is the
only known fossil form.

The genus Avicennia Linne sometimes made the type of a dis-
tinct family, the Avicenniaceae or Black-mangrove family, includes
from three to thirty existing species according to the varying inter-
pretation of different students. They are found on all tropical tidal
shores. Two species have been recognized in the Wilcox flora, one
based on leaves and the second on a not conclusively identified
capsule.

The family Solanaceae includes about seventy genera and about
1600 'existing species, widely distributed and largely tropical, but
extending into the temperate zone, notably in the western hemisphere.
They comprise herbs, shrubs, vines, or in tropical countries often
trees, with opposite, stipulate, toothed, lobed or dissected leaves.

248 BERRY— LOWER EOCENE FLORA OF [April 25,

Their fossil history is almost entirely unknown. The single Wil-
cox representative of the family is a flower described as Solaiiitcs, a
genus founded on the somewhat younger remains of a similar flower
found in the Sannoisian of France, and comparable with the existing
South American genus Saracha Ruiz & Pavon, as well as with With-
eringia, Solatium, etc.

The last order of Gamopetalae positively recognized in the Wilcox
flora is the Rubiales which includes over 5,000 existing species segre-
gated into five families, over four fifths being referred to the family
Rubiaceae — the only one represented in the Wilcox.

The Rubiaceae includes about 355 genera and over 4.500 existing
species of herbs, shrubs and trees ; with simple, opposite or verticil-
late, mostly stipulate, leaves. They are widely distributed and
largely tropical. While the Wilcox representation is confined to a
single species each of Exostcma, Psychotvia and Guettarda, great in-
terest must attach to the fossil record of so highly organized a family
which is my justification for introducing the following brief sketch
of our knowledge of it.

Xo less than twenty-five genera have been recognized in the fossil
state. With the exception of the very doubtful determination of a
species referred to Rubi&phyllum from the Turonian of Bohemia and
doubtless representing a species of Ericaceae, the family is unknown
in the Upper Cretaceous. It is however represented in the early
Eocene both in America and Europe. The Wilcox forms represent
a species of Exostcma Rich., close to the existing Exostcma cari-
hacum R. & S. which ranges from the Florida keys to Central Amer-
ica. The genus comprises about twenty existing species of shrubs
and small trees confined to the tropics and subtropics of America.
The second Wilcox species is referred to Gucttarda Endlicher, a
genus of about fifty species mostly confined to the American tropics
but including one or two cosmopolitan tropical maritime species.
The Wilcox form is very close to the existing Gucttarda clliptica
Swartz, a small tree of the Florida keys, Bahamas and West In
The third Wilcox species is Psychotria grandi folia described origi-
nally by Engelhardt from the early Tertiary of Chili. The genus
Psychotria Linne comprises about 350 existing species of shrubs and
small trees in tropical America. Asia and the East Indies, two thirds

I9I4-] SOUTHEASTERN NORTH AMERICA. 249

of its species being American. The fossil form is compared with
Psychotria grandis Swartz of the American tropics.

The genus Cotissarca Aublet with about 40 existing species in
the Brazilian region has been identified by Engelhardt from the
early Tertiary of Chili. The genus Hoffiuannia Swartz with about a
score of existing American herbs or shrubs, mostly confined to Cen-
tral America, has a fossil species in the early Tertiary of Chili.
Likewise the genera Sabicca Aublet and Gouatteria Martius each
have a single species in the Tertiary of Chili.

The Baltic amber (Sannoisian) has yielded a flower referred
to Sendelia and a leafy twig referred to Enantioblastos. The genus
Galium, comprising over 250 widely distributed existing herbaceous
forms, has been doubtfully identified from the Eocene of Green-
land. Its fruits are also not uncommon in Pleistocene deposits.
The genus Randia Houst, embracing about one hundred existing
species of shrubs or trees in all tropics, is identified by a fruit in the
Aquitanian of Rhenish Prussia.

The genus Rubiacitcs so named by Webber from its resemblance
to the existing forms of Rubia Linne has furnished three species of
leaves and flowers in the Aquitanian of Prussia and Switzerland.
The genus Gardenia Ellis, containing about sixty species of shrubs
or rarely trees of the eastern hemisphere, is represented by charac-
teristic fruits in the Sparnacian of France, the Aquitanian of Ger-
many and England, the Miocene of Baden and Italy, and the Pliocene
of Italy. The genus Posoqueria Aublet, which includes five or six
existing South American shrubs or trees, is represented according to
Unger by both leaves and fruits in the Miocene of Croatia. The
genus Ixora Linne with one hundred existing species of shrubs and
small trees in all tropics is likewise recorded from the Miocene of
Croatia, as is also Pavetta Linne, a genus with about seventy existing
species of shrubs or small trees of the Oriental tropics, which has
furnished both leaves, flowers and fruits from the celebrated plant
and insect beds of Radoboj in Croatia.

The genus Coprosoma Forst., with 40 existing species in Aus-
tralia, New Zealand and Oceanica, was recorded by Ettingshausen
from the Tertiary of Tasmania.

The genus Nauclca Linne, which has about thirty existing species

250 BERRY— LOWER EOCENE FLORA OF [April 25,

of shrubs and trees in tropical Asia and Oceanica, was identified by
Unger in the European Miocene and petrified wood of this type
{Naucleoxylon) was described by Crie from the Pliocene of Java.

The genus Morinda Linne has about thirty existing species in all
tropics, especially in the Orient and the Pacific islands. A fossil
species has been recorded from the Oligoccne of Italy and five addi-
tional species based on leaves have been described from the Miocene
of Croatia.

A fruit from the Tertiary lignites of Brandon, Vermont, has been
described by Perkins as Rubioidcs and another from the Aquitanian
of Rhenish Prussia by Menzel under the name Rubiacecccarpuni.
Geyler has identified the old world genus Grutnilea Gartner in the
Tertiary of Borneo, and finally the genus Cinchonidium proposed
by Unger for fossil fruits and leaves which were very similar to those
of the existing South American genus Cinchona Linne, has furnished
a number of species. There are four or five in the Eocene, including
the Fort Union of the western United States and the Ypresian
of England ; five in the late Oligocene of southeastern Europe ; about
eight Miocene species, one coming from the Esmeralda formation of
Nevada and the balance being European.

The family is thus seen to have been well represented in fossil
floras throughout the Tertiary, but the small proportion of the exist-
ing genera with fossil representatives and the incompleteness of the
record of those with fossil representatives renders untrustworthy any
generalizations that might be made from the present facts.

Under Incertcc scdis are grouped fourteen species of the Wil-
cox flora. These include two forms referred to Caiycites; two to
. Intholithus and ten to Carpolithus. It would be quite useless to
attempt any botanical discussion or comparison of these uncertain
forms, such remarks as they suggest being more suitably confined
to the discussion of the individual species.

Johns Hopkins University,

April 25, 1914.

New York Botanical Garden Library

QE927.B403 1914 gen

Berry, Edward Wilbe/The affinities and d

3 5185 00096 2108