Paleoecological and archaeological implications of a Late Pleistocene/
Early Holocene record of vegetation and climate
from the Pacific coastal plain of Panama
Dolores R. Pipernoa,* and John G. Jonesb
a Center for Tropical Paleoecology and Archaeology, Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA
b Palynology Laboratory, Department of Anthropology, Texas A&M University, College Station, TX 77843-4352, USA
Received 9 May 2002
Abstract
A phytolith record from Monte Oscuro, a crater lake located 10 m above sea level on the Pacific coastal plain of Panama, shows that
during the Late Pleistocene the lake bed was dry and savanna-like vegetation expanded at the expense of tropical deciduous forest, the
modern potential vegetation. A significant reduction of precipitation below current levels was almost certainly required to effect the changes
observed. Core sediment characteristics indicate that permanent inundation of the Monte Oscuro basin with water occurred at about 10,500
14C yr B.P. Pollen and phytolith records show that deciduous tropical forest expanded into the lake?s watershed during the early Holocene.
Significant burning of the vegetation and increases of weedy plants at ca. 7500 to 7000 14C yr B.P. indicate disturbance, which most likely
resulted from early human occupation of the seasonal tropical forest near Monte Oscuro and the development of slash-and-burn methods
of cultivation.
? 2003 Elsevier Science (USA). All rights reserved.
Keywords: Pleistocene; Early Holocene; Climate and vegetation; Prehistoric agriculture
Introduction
Although considerable evidence exists for a cooling of at
least 6?C and a drier climate over the Neotropical lowlands
during the Late Pleistocene (e.g., Guilderson et al., 1994;
Stute et al., 1995; Thompson et al., 1995; Colinvaux et al.,
1996; Webb et al., 1997; Van der Hammen and Hooghiem-
stra, 2000), the net effects on plant communities of the
competing actions of low temperature, precipitation, and
atmospheric concentration of CO2, together with the mag-
nitude of precipitation reduction, are under considerable
discussion (Cowling and Sykes, 1999; Cowling et al., 2001;
Colinvaux et al., 2000; Van der Hammen and Hooghiem-
stra, 2000; Marchant et al., 2001). In Central America south
of Guatemala, the few available glacial-age paleovegeta-
tional sequences from below 1000 m a.s.l. indicate that the
vegetation was a forest without modern analog (e.g., Bartlett
and Barghoorn, 1973; Bush and Colinvaux, 1990; Piperno
et al., 1990; Bush et al., 1992). The presence between ca.
16,000 and 10,500 14C yr. B.P. of pollen and phytoliths
from trees and herbs confined mostly today to higher ele-
vations (e.g., Magnolia, Quercus, Ilex) indicates a signifi-
cant downward movement of these vegetational elements
and a cooling by at least 5?C. These sequences, however,
derive from areas where the modern precipitation is be-
tween 3000 and 4000 mm. Consequently, they may not
adequately inform the problem of environmental change
and vegetational response to drier Pleistocene climates in
the Pacific watershed lowlands, where the annual precipita-
tion is 2000 mm and less, and forests are more prone to
fragmentation.
Detailed reconstructions of Late Pleistocene and early
Holocene environments also have considerable importance
for understanding the human colonization of the Central
American landbridge and subsequent adjustments in human
* Corresponding author.
E-mail address: Pipernod@tivoli.si.edu (D.R. Piperno).
R
Available online at www.sciencedirect.com
Quaternary Research 59 (2003) 79?87 www.elsevier.com/locate/yqres
0033-5894/03/$ ? see front matter ? 2003 Elsevier Science (USA). All rights reserved.
PII: S0033-5894(02)00021-2
subsistence and settlement strategies that led to the origins
of agriculture (Piperno and Pearsall, 1998a). Human deci-
sion-making concerning the kinds of animal and plant food
resources to exploit, and under which modes of production
(e.g., hunting and gathering, plant cultivation) to exploit
them, is significantly structured by ecological factors relat-
ing to the types of vegetational and animal communities
present on the landscape through time (Piperno and Pearsall,
1998a; Smith and Winterhalder, 1992). In this paper, we
present a Late Pleistocene through early Holocene environ-
mental history based on phytolith and pollen records from
sediment cores retrieved from a crater lake, Monte Oscuro,
located on the Pacific coastal plain of Panama.
The site and its setting
Monte Oscuro is an extinct volcano located in Central
Panama 3 km from the Pacific coast (Fig. 1). The floor of
the large (6 
 3 km) crater lies at an elevation of 10 m a.s.l.
with the high parts of the caldera rim reaching 340 m a.s.l.
During the height of the last advance of glaciers, the site
was about 90 km from the sea. The Monte Oscuro region
has a strongly seasonal climate in which 90% of the ca.
1800 mm (18-yr average; range 1228?2573 mm) of annual
precipitation falls between May and November (Estadistica
Panama, 1995). The mean annual temperature is 27?C. The
potential vegetation is a tropical deciduous forest, most of
which has been cut, although remnants can still be found in
areas of the crater that did not hold water and on hills
around the site. Monte Oscuro held a permanent lake that
occupied a substantial part of the flat-bottomed crater until
1957, when it was drained by local people to create pasture
for the cattle industry. Informants who were living in the
area when the crater contained water indicate that the water
depth was ca. 3?6 m, depending on the time of year. The
lake was fed by several inlet streams that drained its small
catchment. A small river draining terrain north of the site
currently outlets to the sea at the northeastern margin of the
crater, and mangrove vegetation from here reaches to within
2.5 km of the coring locale.
Methods
In 1993?1994, we raised two parallel cores 30 m apart,
labeled Cores 1 and 2, from near the center of the lake bed
using truck-mounted drilling equipment. A steel core barrel
19 cm (7.5 inches) in internal diameter and 2 m in length
was forced down by hydraulic pressure to recover sediment,
which was then extruded in the field, cut into 1-m lengths,
and transported to the laboratory for description and sam-
pling. Subsamples were removed for palynology and phy-
tolith analysis, which were carried out using standard tech-
niques (Faegri and Iverson, 1975; Piperno, 1988). A total of
Fig. 1. A topographic map of Panama showing the locations of Monte Oscuro and other sites in Panama that have produced pollen and phytolith records of
Late Pleistocene through early Holocene age, together with the Rio Santa Maria archaeological study region mentioned in the text.
80 D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
26 samples were analyzed for phytoliths and 20 were stud-
ied for pollen. In most samples for which pollen and phy-
tolith data are available, sediment from the same strati-
graphic depth or contiguous 5-cm levels was analyzed. In
pre-Holocene deposits (see below), complementary pollen
and phytolith records were not possible due to poor pollen
preservation.
Prior to processing, two tablets of exotic Lycopodium
spores were added to each sample for calculation of pollen
concentration. Pollen concentration usually varied between
1600 and 5600 grains/cm3. Pollen and phytoliths were
counted to a sum of at least 200 per sample using an
Olympus BHS microscope (400?1000x). Phytoliths and
pollen were identified by comparison to modern reference
collections of over 2000 and 7000 tropical species, respec-
tively, housed in the authors? laboratories. Poaceae phyto-
liths were identified to subfamily and genus by comparison
to over 350 species of Neotropical grasses from all subfam-
ilies of the Poaceae (Piperno and Pearsall, 1998b). Pollen
and phytolith diagrams were plotted using TILAGRAPH
(Grimm, 1992).
Particulate carbon was counted in the palynological
preparations until at least 20 Lycopodium spores had been
counted. From these counts, estimates of charcoal concen-
tration were obtained. Charcoal classification was restricted
to black, completely opaque, angular fragments taken as
secure markers of fire by charcoal analysts. When a plant is
burned, it leaves behind more than charred organic debris. If
the plant holds phytoliths, these are also likely to obtain a
black coating indicating charring, and the morphology of
the phytolith often remains intact. Knowledge of which
plants have been subjected to firing and at what frequency is
thus possible through quantification of the percentages of
burnt phytoliths in any well-represented taxon.
Core stratigraphy and chronology
Both of the cores terminated in volcanic ash underlying
lacustrine sediment and revealed closely comparable stratig-
raphies upon visual inspection. Core 1 was concentrated on
for analysis because it was closest to the deepest part of the
lake. Eight 14C dates were obtained on bulk sediment, and
one 14C determination was run directly on a phytolith as-
semblage isolated from the sediment (Table 1). The carbon
from plant cells trapped inside of phytoliths provides a
suitable substrate for radiocarbon dating, and it is immune
to the various modes of post-depositional contamination
(Wilding, 1967; Kelly et al., 1991; Mulholland and Prior,
1993). The dates obtained from lacustrine sediments (0?6.7
m) and the phytolith assemblage are internally consistent.
There are age reversals at depths below 6.7 m, where AMS
determinations on carbon-poor, nonlacustrine silts (7 m)
and gravelly subsoils (8 m) appear anomalous. The AMS
determination of 5924  50 14C yr B.P. on phytoliths
derived from watershed vegetation from a depth of 3.25 to
3.35 m is in reasonably good agreement with a date on bulk
sediment from 25 cm above. A 7710  40 14C yr B.P.
determination on bulk sediment from Core 2 is from the
same stratigraphic zone and nearly the same depth that
returned a date of 7500  70 14C yr B.P. in Core 1, further
corroborating the chronology for the Core 1 sequence indi-
cated by the dates from 1 to 6 m. A probable cause of the
anomalous dates is downward infiltration of relatively
young carbon derived from the sediment carbonates (see
below) into older strata. That volcanic ashes nearly 1 m
below the lacustrine sediments returned an AMS determi-
nation similar to the two anomalous dates (Beta-74295)
(Table 1) shows that sediments with scarce dateable carbon
were prone to this process, and suggests that the contami-
nating carbonate derived from a single source.
Table 1
List of 14C dates from Monte Oscuro
14C yr. B.P. Lab. No. Depth m. below surface 13C/12C ratio
Core 1
560  80 Beta-64381 0.26?0.31 17.7
2290  60a Beta-64382 1.90?1.96 18.2
4750  170 Beta-68567 2.98?3.0 25.8
5924  50b R 26212/2 3.25?3.35 26
7500  70a Beta-74292 4.75?4.78 19.4
8380  60a Beta-72458 5.70?5.80 26.2
6190  60a Beta-74293 7.00?7.05 29.2
6890  60a Beta-74294 8.02?8.07 30.4
6040  60 Beta-74295 8.85?8.90 26.2
Core 2
7710  40 Beta-144636 5.00?5.05 20.9
a Indicates an AMS date.
b Indicates a phytolith AMS date.
81D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
The gross stratigraphy of the core is straightforward (Fig.
2). Between greater than 10-8 m is a zone of light brown to
orange-yellow volcanic ash that records the last major vol-
canic eruption at the crater. From 8 to 7.1 m is a soil profile
characterized by a dark brown, gravelly subsoil (8?7.6 m),
which is overlain by a medium to dark brown, gravelly soil
with pebbles up to 3 cm in diameter (7.6?7.1 m). The
interior of the crater was dry during this interval and land-
form conditions were stable enough to permit the develop-
ment of a deep soil and associated vegetation. Between
depths of 7.1 and 6.7 m, sediments are medium brown fine
silts intercalated with gray, orange, and mostly black dark
bands. A few small CaCO3 clasts are present. The basin was
still dry most of the time, but may have held water inter-
mittently. Initial ponding leading to the establishment of a
permanent water body is recorded between 6.71 and 6.1 m,
with the deposition of a medium gray, banded, lacustrine
sediment. This is the first significant zone of CaCO3 accu-
mulation. There is no geological limestone in this region of
Panama. Penetrations of brackish water into the coring area
during high sea-level stands, together with seasonal and
longer-term shifts in the water table, can explain carbonate
presence and movement in the sediment profile. Any effect
on sediment age caused by a marine reservoir effect should
be no more than 400 years.
An estimated date for the bottom of this zone and per-
manent ponding at Monte Oscuro is ca. 10,500 14C yr B.P.,
based on an average sedimentation rate of 0.64 mm/yr
calculated for the sequence. It is thus very likely that the
formation of the paleosol between 8 and 7.1 m dates to late
in the Pleistocene. Above 6.1 m to the top of the core is a
continuous series of banded, lacustrine sediments charac-
terized by dark brown to dark black, silty clay with bands of
charcoal and occasional CaCO3 clasts, the sediments be-
coming darker and the clasts fewer with decreasing depth,
and broken only by a zone of ash and gravels between 5.4
and 4.8 meters.
Results
Pollen is not preserved in the paleosol, but phytoliths are
well-represented. Two phytolith samples taken from depths
of 7.50?7.58 m and 7.42?7.50 m indicate that the vegetation
was dominated by grasses and other herbaceous taxa and
contained few trees (Fig. 2). Percentages of Poaceae are
Fig. 2. Percentage data for major phytolith taxa at Monte Oscuro. Chrysobalanaceae/Other Arboreal category contains predominantly Chrysobalanceae
phytoliths (usually  85?90%) with contributions also from Protium, other Burseraceae, the Moraceae, Euphorbiaceae, Annonaceae, Connaraceae, and
Flacourtiaceae. Percentages for burned Poaceae phytoliths were calculated in a separate sum. Only short cell phytoliths (e.g., bilobates, saddle-shapes) were
studied for evidence of charring.
82 D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
high, the Cyperaceae reach their maximum levels of the
sequence, and phytoliths from the shrub or small tree Cu-
ratella americana, a major indicator taxon for savanna in
the Neotropics, are restricted to this segment of the core.
Phytoliths from Arundinella spp., today tall grasses of open,
windswept landscapes, occur, along with phytoliths from
Heliconia spp., herbs of dry, open habitats. Phytoliths from
the panicoid (tall grass) and chloridoid (short grass) sub-
families of the Poaceae are conspicuous. All chloridoid and
most panicoid grasses use the C4 photosynthetic pathway
(Kellogg, 2000). Fire occurrence is indicated by the pres-
ence of charcoal (not shown) and burned grass phytoliths.
No siliceous microfossils from aquatic organisms such as
sponge spicules were present, further indicating that
edaphic conditions peculiar to the crater floor did not con-
tribute to the development of the open vegetation.
Arboreal phytolith taxa in the paleosol are represented
mainly by the Chrysobalanaceae and Arecaceae, and they
account for less than 40% of the phytolith sum. The mor-
phology of the palm phytoliths is consistent with that of
more xeric species that occupy dry tropical habitats (Pip-
erno, 1988). Phytoliths from the Burseraceae (including
Protium spp.), Annonaceae (probably including Unonopsis
and/or Oxandra), and Moraceae make up less than 1% of
the sum. These taxa, along with the Chrysobalanceae and
Aracaceae, are among the prominent families of trees that
produce copious amounts of phytoliths in their leaves and
fruits and achieve a significant representation in phytolith
samples from modern tropical forests (Piperno, 1988,
1993). Comparisons of the paleosol phytolith spectra with
those of modern phytolith analogs constructed from tropical
deciduous, semievergreen, and evergreen forests and de-
rived savannas (where tropical deciduous forests have been
cleared and burned), show that phytolith assemblages like
those of the paleosol with high percentages of grasses and
other herbs are found only in savannas (Figs. 2, 3) (Piperno,
1988). In contrast, modern tropical deciduous forests con-
tribute phytolith assemblages in which grasses and other
herbs make up less than 30% of the phytolith sum, even
though the forests are adjacent to grassy areas and pastures,
have broken canopies and some intrusive, grassy growth
resulting from past human disturbance, and have been pen-
etrated at ground level by fires set in the pastures.
Our evidence thus indicates that during the Pleistocene,
Monte Oscuro supported open, dry-land vegetation, in
which grew many grasses and other herbaceous taxa. Forest
elements apparently were not entirely eliminated from the
landscape but appear to have grown in low numbers. They
possibly were harbored along the margins of streams that
fed the lake. Thorny-scrub plant associations cannot be
recognized with a phytolith record, and hence we cannot
specify whether such kinds of low tree/scrub species were
also present. It is probable that thorny-scrub taxa were also
part of the Late Pleistocene vegetation.
Sediments deposited during the terminal Pleistocene and
earliest Holocene (between about 6.70 and 6.15 m) were
phytolith- and pollen-poor compared to sediments from the
rest of the sequence, possibly because water levels were not
yet sufficiently stable for uninterrupted sedimentation and
plant fossil accumulation to take place. Nevertheless, the
pollen and phytolith records show that during the first few
millennia of the Holocene, a tropical deciduous forest with
considerable species diversity grew in the Monte Oscuro
watershed. The Holocene phytolith record starts at a depth
of 6.15 m, 35 cm below the level dated at 8360 14C yr B.P.
Significant changes from Pleistocene levels are recorded,
and the phytolith spectrum closely resembles its modern
counterpart from tropical deciduous forest (Figs. 2, 3). Ar-
boreal phytoliths dominate the assemblage, whereas per-
centages of the Poaceae and Cyperaceae are greatly reduced
from before. Curatella americana and Arundinella are not
recorded, while phytoliths from the Marantaceae, herbs of
the moist forest understorey, are present. Reduction of palm
phytoliths at this time is expected, since this family is not
Fig. 3. Modern phytolith spectra from tropical deciduous forest in Costa Rica and derived savanna in Costa Rica and Panama. Phytolith frequencies for each
habitat are averages from a series of soil transects, and represent pinch samples taken from the upper soil surface directly underneath existing vegetation.
83D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
well represented in the understorey and phytolith records of
tropical deciduous forest (Fig. 3).
The pollen record from Monte Oscuro begins at a depth
of 4.5 m, 25 cm above a level dated to 7500  70 14C yr
B.P. (Fig. 4). A suite of trees whose modern ecological
preference is seasonal tropical forest is recorded (e.g., Zan-
thoxylum, Spondias, Protium, Bursera, Erythrina, Celtis,
Cavanillesia, Calophyllum, Cordia, Anacardiaceae, Mora-
ceae), and these and other arboreal taxa (e.g., Trema,
Machaerium-type, which are not shown and account for the
remainder of the pollen percentages not displayed in Fig. 4)
constitute over 55% of the pollen sum. The presence of
mangrove pollen (e.g., Combretaceae, Pelliciera, Rhizo-
phora) indicates that the coast was nearby. Also recorded at
this level are significant amounts of Poaceae, other taxa
indicative of human disturbance (e.g., Asteraceae, Cyper-
aceae), and a significant rise in charcoal. In older samples,
at depths between 6.8 and 5.7 m, charcoal concentrations
range between 21,000 and 264,000/cm3, increasing to
512,000/cm3 at 5.5 m. At 4.5 m, the charcoal concentration
is 2.2 million per cm3, a very high amount. The phytolith
record from 4.75 to 4.7 m also indicates an increase of
vegetational disturbance including with fire, as frequencies
of arboreal taxa decline, those of Poaceae rise to nearly
pre-Holocene levels again, and other early successional taxa
such as Heliconia rise sharply. Many of the Poaceae and
Heliconia phytoliths are now charred.
These and other factors cause us to believe that the
changes at 4.75 to 4.5 m most likely reflect an onset of
forest modification by human populations in the Monte
Oscuro watershed associated with incipient land clearance
by slash-and-burn techniques of cultivation. Small amounts
of maize phytoliths are first present at 4.75 to 4.7 m. Fur-
thermore, the morphologies of the Poaceae phytoliths are
not consistent with those that would result from an in-
creased deposition of local swamp grasses such as Miscant-
hus or Echinochloa, but rather appear to derive from bam-
boo genera such as Guadua spp. and other taxa adapted to
forest openings. That the vegetational openings indicated by
the rises in Poaceae and Heliconia are not those of swamps
fringing the lake is also indicated by the fact that Cyper-
aceae pollen show no increase at this time.
Early human interference with the Panamanian forest
involving the use of fire by hunters and gatherers at ca.
11,050 14C yr B.P. and significant forest clearing for slash-
and-burn cultivation at ca. 7000 14C yr B.P. are indicated
from the pollen, phytolith, and charcoal records at Lake La
Yeguada (Piperno et al., 1990, 1991; Bush et al., 1992),
located about 150 km west of Monte Oscuro (Fig. 1).
Archaeological records from sites in central Panama near La
Yeguada from the drainage of the Rio Santa Maria also
demonstrate an early human presence and the development
of horticultural systems using seed and root crops between
ca. 8000 B.P. and 5000 14C yr B.P. (Fig. 1) (Cooke, 1998;
Ranere and Cooke, 2002; Piperno and Pearsall, 1998; Pip-
erno et al., 2000a,b). After ca. 7500 14C yr B.P., the Monte
Oscuro pollen and phytolith frequencies are consistent with
largely sustained human environmental interference and ag-
ricultural practices through time, a result concordant with
paleoecological and archaeological records from other Pan-
amanian sites (Piperno et al., 1991; Piperno and Pearsall,
1998a; Piperno et al., 2000b).
Fig. 4. Percentage data for major pollen taxa at Monte Oscuro. Charcoal concentrations are expressed in millions as number per cm3 of sediment.
84 D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
Discussion
The net effects of the competing actions of cooling,
drying, and low CO2 concentrations on Pleistocene plant
communities are not well understood. Although lower at-
mospheric CO2 content may have exerted a significant in-
fluence on vegetation composition and structure (Cowling
and Sykes, 1999; Cowling et al., 2001), recent studies of
lakes from Mexico and Guatemala show that vegetational
shifts from forests to grasslands, or from C3- to C4-domi-
nated ecosystems in the lowland Neotropics, are unlikely to
take place without significant reductions in precipitation
(Huang et al., 2001). Moreover, low CO2 concentrations
acting independent of significant climatic influences proba-
bly cannot account for the fact that many lakes from Gua-
temala to southern Brazil, including Monte Oscuro, were
substantially lower or completely dry during the Late Pleis-
tocene (Ledru et al., 1998; Van der Hammen and Absy,
1994; Curtis et al., 1999). Today, the lower limit of annual
precipitation generally required to support a tropical decid-
uous forest is about 1000 to 1200 mm (Bullock et al., 1995).
Hence, a suitable estimate for precipitation reduction at
Monte Oscuro during the Pleistocene seems to be on the
order of at least 35%.
The underlying mechanisms that may account for signif-
icantly drier Pleistocene climates in tropical America in-
clude decreased evaporation rates over cooler ocean sur-
faces that resulted in less atmospheric water vapor over land
masses, and intensity shifts in the annual cycle of the In-
tertropical Convergence Zone that were forced by orbitally
driven insolation changes (Webb et al., 1997; Hodell et al.,
1991; Leyden et al., 1993). The estimated timing of the
filling of Monte Oscuro at ca. 10,500 14C yr B.P. is consis-
tent with the inundation of other lakes in tropical America
that were dry or much lower during the Late Pleistocene
(Ledru et al., 1998; Curtis et al., 1999).
The persistence of forests during the late Pleistocene,
albeit without modern analog, has been demonstrated at
several sites below 1000 m in tropical America, including
three in Panama; La Yeguada, El Valle, and the Gatun Basin
(Bartlett and Barghoorn, 1973; Bush and Colinvaux, 1990;
Piperno et al., 1990; Bush et al., 1992) (Fig. 1). The fact that
these sites are situated in high-precipitation areas (e.g.,
annual rainfall between ca. 3000 mm and 4000 mm today)
may be extremely relevant, as even under a 30% to 40%
decline of precipitation sufficient moisture was probably
available to support arboreal growth. This may be especially
true in view of the fact that cooler Pleistocene temperatures
would have improved carbon and water balance in the
C3-dominated forests by reducing evotranspiration gradi-
ents and photorespiratory carbon loss (Cowling et al.,
2001). The presence of the clay type illite, which forms
under conditions of limited precipitation, in the Pleistocene
(ca. 14,000 to 10,500 14C yr B.P.) but not in the Holocene
deposits (10, 500 14C yr B.P. to the modern era) at Lake La
Yeguada (elevation 650 m a.s.l.; annual precipitation 3800
mm a year) (Bush et al., 1992), constitutes strong evidence
that this region did indeed experience a marked reduction in
precipitation during the late-glacial period. In contrast to
forest persistence in higher rainfall zones, forests were often
displaced during the Pleistocene by savanna-like vegetation
where the modern potential vegetation is a highly seasonal
tropical forest (in areas with current annual precipitation
between ca. 1200 and 2000 mm) (e.g., Van der Hammen
and Hooghiemstra, 2000; Leyden, 1984, 1985; Van der
Hammen and Absy, 1994; Pennington et al., 2000; Behling
and Negrelle, 2001).
Our data from Monte Oscuro indicate that during the
Late Pleistocene, the Pacific coastal plain of the Panamanian
land bridge was largely a dry and open habitat. We are
hesitant to use the word savanna to describe the late-glacial
vegetation, because rather than the tracts of grasses and
sedges dotted by low trees and shrubs characteristic of
savannas today, the floristic associations may have been
unusual combinations of herbaceous plants, tropical thorn-
scrub, and temperate shrubs without modern analog (e.g.,
Hodell et al., 1991; Leyden et al., 1993). Our phytolith
evidence also points to the presence of lowland forest trees
(Protium spp., Annonaceae, Chyrsobalanaceae, Moraceae)
nearby the lake before ca. 10,500 14C yr B.P. These taxa
may have grown in low numbers in more mesic pockets on
the landscape, and/or they were harbored in better-watered
areas alongside water courses. In any case, the data support
arguments that high postglacial dispersal rates over large
distances are not required to account for early Holocene
forest recovery in areas where forests were reduced (Meave
and Kellman, 1994). Also, suggestions that riparian forests,
whose elongated stretches of terrain can support higher
levels of species diversity than previously believed, were
important conservators of lowland forest trees during glacial
dry periods (Meave and Kellman, 1994; Prance, 1987) may
have considerable merit.
Finally, our data provide important insights into the eco-
logical contexts of the transition from foraging to farming in
Panama and other similar regions of the Neotropics. The
Monte Oscuro record adds to the body of paleoecological
evidence indicating that in low-lying regions such as the
Pacific coastal plain of Panama, where between 1500 and
2000 mm of annual precipitation is received today on a
strongly seasonal basis, a species-rich tropical forest often
was not present during the late-glacial period. Rather, such
vegetation started to spread on the landscape upon the rapid
onset of warmer and wetter conditions between ca. 10,800
and 10,000 14C yr B.P. The loss of open-land habitats,
coupled with the extinctions of numerous species of large,
herbivorous animals and spread of tropical forest with its
different sets of animal and plant resources during the early
postglacial period, necessitated significant reorganizations
of human subsistence, technological, and settlement strate-
gies. Such kinds of cultural responses to the end-Pleistocene
perturbations are well reflected in the early Holocene arche-
ological records from Panama and elsewhere in the Amer-
85D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
ican tropics (Cooke, 1998; Piperno and Pearsall, 1998a;
Ranere and Cooke, 2002). Thus, as in other regions of the
world, the transition from hunting and gathering to farming
in the lowland Neotropical forest appears to have taken
place within a context of considerable paleoenvironmental
change that substantially altered resource abundance and
availability, and generated significant selective pressures for
changes in socio-economic systems not long after the Pleis-
tocene ended (Zhao and Piperno, 1998; Piperno and Pears-
all, 1998a; Hillman, 2001).
Acknowledgments
This work was supported by the Smithsonian Tropical
Research Institute (STRI) and a grant to the STRI from the
Andrew W. Mellon Foundation. We thank Anthony J. Ra-
nere for comments on the manuscript.
References
Bartlett, A.S., Barghoorn, E.S., 1973. Phytogeographic history of the Isth-
mus of Panama during the past 12,000 years. In: Grohem, A. (Ed.),
Vegetation and Vegetational History in Northern Latin America.
Elsevier, Amsterdam, pp. 203?299.
Behling, H., Negrelle, R.R.B., 2001. Tropical rain forest and climate
dynamics of the Atlantic, lowland, southern Brazil, during the Late
Quaternary. Quaternary Research 56, 383?389.
Bullock, S.H., Mooney, H.A., Medina, E. (Eds.), 1995. Seasonally Dry
Tropical Forests. Cambridge Univ. Press, Cambridge, UK.
Bush, M.B., Colinvaux, P.C., 1990. A pollen record of a complete
glacial cycle from lowland Panama. Journal of Vegetation Science
1, 105?118.
Bush, M., Piperno, D.R., Colinvaux, P.A., De Oliveira, P., Krissek, L.,
Miller, M., Rowe, W., 1992. A 14,300 year paleoecological profile of
a lowland tropical lake in Panama. Ecological Monographs 62, 251?
275.
Colinvaux, P.A., De Oliveira, P.E., Moreno, P.E., Miller, M.C., Bush,
M.B., 1996. A long pollen record from lowland Amazonia: forest and
cooling in glacial times. Science 274, 85?88.
Colinvaux, P.A., De Oliveira, P.E., Bush, M.B., 2000. Amazonian and
neotropical plant communities on glacial time-scales: the failure of the
aridity and refuge hypotheses. Quaternary Science Reviews 19, 141?
169.
Cooke, R.G., 1998. Human settlement of Central America and northern
South America. Quaternary International 49/50, 177?190.
Cowling, S.A., Sykes, M.T., 1999. Physiological significance of low at-
mospheric CO2 for plant?climate interactions. Quaternary Research
52, 237?242.
Cowling, S.A., Maslin, M.A., Sykes, M.T., 2001. Paleovegetation simula-
tions of lowland Amazonia and implications for Neotropical allopatry
and speciation. Quaternary Research 55, 140?149.
Curtis, J.H., Brenner, M., Hodell, D.A., 1999. Climate change in the Lake
Valencia Basin, Venezuela, 12600 yr BP to present. The Holocene 9,
609?619.
Estadistica Panamen?a, 1995. Contraloria General de la Republica, Direc-
cio?n de Estadistica y Censo, Panama City, Panama.
Faegri, K., Iverson, J., 1975. A Textbook of Pollen Analysis, third ed.
Haffer, New York.
Grimm, E. 1992. TILIA Software, Version 1.12. Illinois State University.
Guilderson, T.P., Fairbanks, R.G., Rubenstone, J.L., 1994. Tropical tem-
perature variations since 20,000 years ago: modeling interhemispheric
climate change. Science 263, 663?665.
Hillman, G.C., 2001. New evidence of Lateglacial cereal cultivation at Abu
Hureyra on the Euphrates. The Holocene 11, 383?393.
Hodell, D.A., Curtis, J.H., Jones, G.A., Higuera-Gundy, A., Brenner, M.,
Binford, M.W., Dorsey, K.T., 1991. Reconstruction of Caribbean cli-
mate change over the past 10,500 years. Nature 352, 790?792.
Huang, Y., Street-Perrott, F.A., Metcalfe, S.E., Brenner, M., Moreland, M.,
Freeman, K.H., 2001. Climate change as the dominant control on
glacial?interglacial variations in C3 and C4 plant abundance. Science
293, 1647?1651.
Kellogg, E.A., 2000. The grasses: a case study in macroevolution. Annual
Review Ecology and Systematics 31, 217?238.
Kelly, E.F., Amundson, R.G., Marino, B.D., Deniro, M.J., 1991. Stable
isotope ratios of carbon in phytoliths as a quantitative method of
monitoring vegetation and climate change. Quaternary Research 35,
222?233.
Ledru, M.P., Bertaux, J., Sifeddine, A., Suguio, K., 1998. Absence of Last
Glacial Maximum records in lowland tropical forests. Quaternary Re-
search 49, 233?237.
Leyden, B., 1984. Guatemalan forest synthesis after Pleistocene aridity.
Proceedings of the National Academy of Sciences U.S.A. 81, 4856?
4859.
Leyden, B., 1985. Late Quaternary aridity and Holocene moisture
fluctuations in the Lake Valencia Basin, Venezuela. Ecology 66,
1279 ?1295.
Leyden, B., Brenner, M., Hodell, D.A., Curtis, J.H., 1993. Late Pleistocene
climate in the Central American lowlands. In: Swart, P.K., Lohmann,
K.C., McKenzie, J., Savin, S. (Eds.), Climate Change in Continental
Isotopic Records, Geophysical Monograph 78. American Geophysical
Union, Washington, DC, pp. 165?178.
Marchant, R., Boom, A., Hooghiemstra, H., 2001. Pollen-based biome
reconstructions for the past 450 000 yr from the Funza-2 core, Colom-
bia: comparisons with model-based vegetation reconstructions. Paleo-
geography Paleoclimatology Paleoecology 2711, 1?17.
Meave, J., Kellman, M., 1994. Maintenance of rain forest diversity in
riparian forests of tropical savannas: implications for species conser-
vation during the Pleistocene drought. Journal of Biogeography 21,
121?153.
Mulholland, S.C., Prior, C., 1993. AMS radiocarbon dating of phytoliths. In:
Pearsall, D.M., Piperno, D.R. (Eds.), Current Research in Phytolith Anal-
ysis: Applications in Archeology and Paleoecology, MASCA Research
Papers in Science and Archaeology, Vol. 10. MASCA, The University
Museum of Archaeology and Anthropology, Philadelphia, pp. 21?23.
Pennington, R.T., Prado, D.E., Pendry, C.A., 2000. Neotropical seasonally
dry forests and Quaternary vegetation changes. Journal of Biogeogra-
phy 27, 261?273.
Piperno, D.R., 1988. Phytolith Analysis: An Archaeological and Geolog-
ical Perspective. Academic Press, San Diego.
Piperno, D.R., 1993. Phytolith and charcoal records from deep lake cores
in the American tropics. In: Pearsall, D.M., Piperno, D.R. (Eds.),
Current Research in Phytolith Analysis: Applications in Archaeology
and Paleoecology, MASCA Research Papers in Science and Archae-
ology, Vol 10. MASCA, The University Museum of Archaeology and
Anthropology, Philadelphia, pp. 58?71.
Piperno, D.R., Pearsall, D.M., 1998a. The Origins of Agriculture in the
Lowland Neotropics. Academic Press, San Diego.
Piperno, D.R., Pearsall, D.M., 1998b. The Silica Bodies of Tropical Amer-
ican Grasses: Morphology, Taxonomy, and Implications for Grass
Systematics and Fossil Phytolith Identification. Smithsonian Contribu-
tions to Botany, No. 85.
Piperno, D.R., Bush, M.B., Colinvaux, P.A., 1990. Paleoenvironments and
human occupation in Late-Glacial Panama. Quaternary Research 33,
108?116.
86 D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87
Piperno, D.R., Bush, M.B., Colinvaux, P.A., 1991. Paleoecological per-
spectives on human adaptation in Central Panama. II. The Holocene.
Geoarchaeology 6, 227?250.
Piperno, D.R., Ranere, A.J., Holst, I., Hansell, P., 2000a. Phytoliths in
Cucurbita and other Neotropical Cucurbitaceae and their occurrence in
early archaeological sites from the lowland American tropics. Journal
of Archaeological Science 27, 193?208.
Piperno, D.R., Holst, I., Andres, T.C., Stothert, K.E., 2000b. Starch grains
reveal early root crop horticulture in the Panamanian tropical forest.
Nature 407, 894?897.
Prance, G.T., 1987. Vegetation. In: Whitmore, T.C., Prance, G.T. (Eds.),
Biogeography and Quaternary History in Tropical America. Oxford
Univ. Press, Oxford, pp. 28?45.
Ranere, A.J., Cooke, R.G., 2002. Late glacial and early Holocene occupa-
tion of Central American tropical forests. In: Mercader, J. (Ed.), Under
the Canopy: The Archaeology of Tropical Rainforests. Rutgers Univ.
Press, Rutgers, NJ, pp. 211?248.
Smith, E.A., Winterhalder, B., 1992. Natural selection and decision mak-
ing: some fundamental principles. In: Smith, E.A., Winterhalder, B.
(Eds.), Evolutionary Ecology and Human Behavior. Aldine de Gruyter,
New York, pp. 25?60.
Stute, M., Forster, M., Frischkorn, H., Serejo, A., Clark, J.F., Schlosser, P.,
Broecker, W.S., Bonani, G., 1995. Cooling of tropical Brazil (5?C)
during the Last Glacial Maximum. Science 269, 379?383.
Thompson, L.G., Mosley-Thompson, E., Davis, M.E., Lin, P.N., Hender-
son, K.A., Cole-Dai, J., Bolzan, J.F., Liu, K.B., 1995. Late glacial stage
and Holocene tropical ice core records from Huascara?n, Peru. Science
269, 46?50.
Van der Hammen, T., Absy, M.L., 1994. Amazonia during the last glacial.
Paleogeography, Paleoclimatology, Paleoecology 109, 247?261.
Van der Hammen, T., Hooghiemstra, H., 2000. Neogene and Quaternary
history of vegetation, climate, and plant diversity in Amazonia. Qua-
ternary Science Reviews 19, 725?742.
Webb, R.S., Rind, D.H., Lehman, S.J., Healy, R.J., Sigman, D., 1997.
Influence of ocean heat transport on the climate of the Last Glacial
Maximum. Nature 385, 695?699.
Wilding, L.P., 1967. Radiocarbon dating of biogenetic opal. Science 156,
66?67.
Zhao, Z., Piperno, D.R., 2000. Late Pleistocene/Holocene environments in
the Middle Yangtze River Valley, China and rice (Oryza sativa L.)
domestication: the phytolith evidence. Geoarchaeology 15, 203?222.
87D.R. Piperno, J.G. Jones / Quaternary Research 59 (2003) 79?87