Starch Fossils and the Domestication
and Dispersal of Chili Peppers
(Capsicum spp. L.) in the Americas
Deborah M. Pearsall,
Dolores R. Piperno,
Mary Jane Berman,
Richard G. Cooke,
Anthony J. Ranere,
J. Scott Raymond,
Daniel H. Sandweiss,
James A. Zeidler
Chili peppers (Capsicum spp.) are widely cultivated food plants that arose in the Americas and are
now incorporated into cuisines worldwide. Here, we report a genus-specific starch morphotype that
provides a means to identify chili peppers from archaeological contexts and trace both their
domestication and dispersal. These starch microfossils have been found at seven sites dating
from 6000 years before present to European contact and ranging from the Bahamas to southern
Peru. The starch grain assemblages demonstrate that maize and chilies occurred together as an
ancient and widespread Neotropical plant food complex that predates pottery in some regions.
hili peppers, members of the genus Cap-
sicum, have been cultivated extensively,
initially in the Americas and, after Co-
lumbus, around the globe (1, 2). The lack of a
comprehensi ve archaeobotanical record has
hampered accurate reconstructions of the origins,
domestications, and dispersals of these plants.
Macroremains of the fruits are confined to rare
sites in arid environments, and reports of seeds and
pollen are even less common (Table 1). We found
that a widespread, but previously unidentified
archaeological starch morphotype is derived from
chili pepper fruits and is commonly preserved on
artifacts. We documented this microfossil from
seven archaeological sites ranging from the
Bahamas archipelago to Andean South America
(Fig. 1) beginning 6000 years ago (T able 2).
The five most economically notable spe-
cies of chili pepper are C. annuum, C. baccatum,
C. chinense, C. frutescens,andC. pubescens.Al-
though it is generally agreed that the genus
Capsicum originated in Bolivia (2), the centers
of domestication and dispersal patterns of these
species remain speculative. A combination of ar-
chaeological evidence, genetic analyses, and
modern plant distributions have led researchers
to suggest that C. annuum was initially domes-
ticated in Mexico or northern Central America, C.
frutescens in the Caribbean, C. baccatum in low-
land Bolivia, C. chinense in northern lowland
Amazonia, and C. pubescens in the mid-elevation
southern Andes (2, 3).
All five species of domesticated chili peppers
produce large, flattened lenticular starch grains
with a shallow central depression, not unlike a red
blood cell in appearance (Fig. 2, A to C). When
rotated into side view , a central linear figure—a
clean line or split figure with sharp edges—runs
parallel to the long axis of the grain. This figure
can extend for the entire length of the grain or just
a part of it (Fig. 2, E and F). Ranging from about
13 to 45 mm in length, the starches of domesticated
peppers are easily distinguishable from smaller
wild types in the microfossil record (Fig. 2D and
table S1). Although the basic three-dimensional
morphology is consistent among all species of
Capsicum, micromorphological characters differ
Three of the species—C. baccatum, C.
frutescens,andC. pubescens—can be identified
on the basis of diagnostic morphotypes that have
unique features of the central depression. How-
ever, these features are rare even in modern starch
grain assemblages. Otherwise, the morphologies
of starch grain assemblages from C. annuum and
C. frutescens are so similar that, in the absence of a
diagnostic, it is not possible to assign grains to a
single species. The morphology of starch from C.
chinense is similar to but not identical to that of C.
annuum or C. frutescens, and the morphologies of
all three starches differ significantly from those of
C. baccatum and C. pubescens which, in turn,
differ from one another . Because similar types
occur in all congeneric species of Capsicum, either
a diagnostic or a large archaeobotanical assem-
blage is required for a secure species identification.
The presence of a basic genus-diagnostic
starch morphotype for Capsicum is predictable
because of the lack of perfect barriers to intraspe-
cific hybridization (4). C. annuum, C. chinense,
and C. frutescens have been described as a species
complex with a single ancestral gene pool (4).
Therefore, it is not surprising that the starches of
these three species are morphologically similar
to one another. In contrast, C. baccatum and C.
pubescens are distinct domesticated species in
South America (4). Starches derived from other
economically significant species in the Solanaceae
including Lycianthes, the genus that recent phylo-
genetic studies indicate is the most closely related
to Capsicum (5), differ from those of chili peppers
(table S1 and fig. S1) (6). Thus, we have elimi-
nated those plant species with the potential to
confuse the source of the microfossils.
We recovered securely identified genus-
diagnostic Capsicum starch microfossils from
seven sites throughout the Americas. The oldest
positively identified starches were found at the
contemporaneous sites of Loma Alta and Real
Alto in southwestern Ecuador. Interpreted as a
village-sized, permanent settlement, Loma Alta
was occupied for more than a millennium be-
ginning about 6100 years before present (yr B.P.)
(7). We recovered chili pepper starches from
sediment samples, milling stones, and food
residues from ceramic sherds of cooking vessels,
all of which were excavated from the lower
levels of the site.
Similar to Loma Alta, Real Alto was a village
site at about 6100 yr B.P.; however, by about 4750
yr B.P., it had expanded into a regional ritual-
ceremonial center (8, 9). The chili starches were
extracted from milling stones from two house
floors dating to the period of expansion. Micro-
fossil evidence of maize, Canna edulis (achira),
Maranta arundinacea (arrowroot), Calathea sp.
Archaeobiology Program, Department of Anthropology,
Smithsonian National Museum of Natural History, Post
Office Box 37012, Washington, DC 20013–7012, USA.
Department of Archaeology, University of Calgary, 2500
University Drive, NW, Calgary, Alberta, T2N 1N4, Canada.
Smithsonian Tropical Research Institute, Apartado Postal
0843–03092, Balboa, Republic of Panama.
of Anthropology, 107 Swallow Hall, University of Missouri,
Columbia, MO 65211, USA.
Center for American and
World Cultures, 105 MacMillan Hall, Miami University,
Oxford, OH 45056, USA.
Climate Change Institute,
University of Maine, 120 Alumni Hall, Orono, ME 04469–
Department of Anthropology, Temple University,
1115 West Berks Street, Philadelphia, PA 19122, USA.
Department of Anthropology, South Stevens 5773, Univer-
sity of Maine, Orono, ME 04469–5773, USA.
Antropología, Instituto Venezolano de Investigaciones Cien-
tíficas, Carretera Panamericana, Kilometer 11, Altos de Pipe,
Departamento de Arqueología, Etnohistoria y
Ecología Cultural, Escuela de Antropología, Facultad de
Ciencias Económicas y Sociales, Universidad Central de
Venezuela, Caracas 1041, Venezuela.
Center for Environ-
mental Management of Military Lands, Colorado State
University, Fort Collins, CO 80523, USA.
Table 1. Published reports of archaeological Capsicum from well-dated sites with clearly defined
Species Plant part Region Site(s) Date(s) (yr B.P.) Source
C. annuum Fruits Mexico Tehuacan Valley 500–6000 (24)
C. annuum Seeds and peduncles El Salvador Ceren 1400 (25)
C. baccatum Fruits Peru Huaca Prieta, Punta Grande 4000 (26)
C. chinense Fruits Peru Huaca Prieta, Punta Grande 4000 (26)
C. chinense Fruits Peru Casma Valley Ca. 2500–3500 (3)
Capsicum sp. Seeds Haiti En Bas Saline 600 (27)
Capsicum sp. Pollen Venezuela La Tigra 450–1000
16 FEBRUARY 2007 VOL 315 SCIENCE www.sciencemag.org986
(leren), manioc, cucurbits (squash), Canavalia sp.
(jack bean), and the Arecaceae family (palms) has
also been recovered from Real Alto (10–12). A
combination of evidence, including plant remains
and site proximity to seasonally flooded bottom-
land, indicates that agriculture was important in the
economies of both Ecuadorian sites. Ecuador is not
considered to be the center of domestication for any
of the five major economic species of chili peppers.
Therefore, the presence of domesticated chilies with-
in this early , complex, agricultural system indicates
that these plants must have been domesticated
elsewhere earlier than 6000 yr B.P. and brought into
the region from either the north or the south.
In central Panama, the Aguadulce Rock Shel-
ter was occupied from about 13,000 to 3200 yr
B.P. during both the Preceramic and Early Ce-
ramic periods (13). The site has yielded evidence
for the cultivation of other plants not native to
southern Central America, including maize, man-
ioc, and squashes dating from about 9000 to 5800
yr B.P. W e identified chili pepper starch on a
groundstone tool recovered from the top of the
preceramic deposits; the tool and thus its associated
starch residues have a stratigraphic date of about
5600 yr B.P. This artifact also yielded starch grains
from maize and domesticated yam (13).
The occupation of the coastal shell-midden site
of Zapotal coincides with the Early Ceramic
period of this region of Panama, beginning about
4800 yr B.P. (14). We recovered starches of chilies
from groundstone tools, indicating that the
peppers were processed and consumed alongside
a number of other domesticates at the site, in-
cluding maize, manioc, and yams (15). By this
time, swidden cultivation of several domesticated
species, including maize and manioc, was well
established in the region, and farmers had sig-
nificantly deforested the foothills near both
Panamanian sites (16). Thus, the Panamanian
record documents the use of domesticated chilies
as components of the diet of swidden agriculturists
in both Preceramic and Ceramic era groups.
Farther south at 3600 m in the Peruvian Andes
lies the site of Waynuna, a Late Preceramic
house occupied beginning about 4000 yr B.P. At
Waynuna, we found chili starches on processing
tools in association with maize, arrowroot, and the
remains of what is likely Solanum sp. (potato) (17).
These data indicate that the residents of Waynuna
were cultivating maize, tubers, and peppers and
were processing them into food on site. Waynuna
yielded the only starch assemblage that contained a
species-diagnostic morphotype. These chili pepper
starches appeared to be derived from C. pubescens,
the species that includes varieties such as the rocoto
pepper, a chili that is cultivated at mid-altitude in
the Andes (2). When combined with macrofossil
evidence (T able 1), the starch data indicate that the
cultivation of three domesticated species of chili
pepper was contemporaneous on the coast and in
the highlands of Peru as early as 4000 yr B.P. in the
Late Preceramic period. The presence of numerous
other cultivars within the assemblages of each
region indicates that sophisticated agriculture was
practiced in both regions before the introduction of
We also found starches of chili peppers at the
Three Dog site located on San Salvador Island in
the Bahamas. This site was occupied by a group of
fisher-horticulturists about 1000 yr B.P. Rep-
resenting the material remains of at least one
household, the site consists of a midden, two ac-
tivity areas, and a low-density (well-swept) area.
Fifty-eight chert microliths, all typical of the mor-
phology commonly described as manioc grater
flakes (18–20), have been recovered, as were
ceramic griddle sherds. The microliths yielded the
starchy remains of both maize and unidentified
roots or tubers. We recovered chili starches from
two flakes that also contained starches of maize.
Lastly , we recovered microfossil evidence for
chili pepper at Los Mangos del Parguaza in
Venezuela, a large habitation site occupied about
Fig. 1. Archaeological sites mentioned in the
text. Red sites yielded starch grains of chili
pepper. Blue sites yielded all other classes of
remains of chili pepper.
Table 2. Summary of Neotropical archaeological starches of Capsicum. F, flaked tool; G,
groundstone tool; C, ceramic sherd; S, sediment sample.
Sample # Size (mm) Source Date in yr B.P. (Ref.)
Los Mangos, Venezuela (Arauquinoid, Valloid)
Lev 1, 1 1 15 G ~500–1000 (21)
Lev 2, 1 1 17 F ~500–1000 (21)
Lev 3, 1 1 15 F ~500–1000 (21)
Lev 3, 2 1 22 F ~500–1000 (21)
Lev 5, 2 2 19, 20 F ~500–1000 (21)
Lev 7, 1 2 20, 20 F ~500–1000 (21)
Three Dog, Bahamas (Lucayan)
Z87-89 1 19 F 969–1265 Cal* (29)
Z1032-1035 1 21 F 969–1265 Cal* (29)
Waynuna, Peru (Preceramic)
Tool 10 2 18,24 F 3564–3837 Cal† (17)
Tool 11 6 14–34 G 3564–3837 Cal† (17)
Tool 29 2 24,25 G 3564–3837 Cal† (17)
Tool 30 6 19–28 G 3564–3837 Cal† (17)
Cat 36 1 18 S 3689–3969 Cal† (17)
Zapotal, Panama (Early Ceramic)
C2N8F4 4 20-28 G 3560–4850 Cal‡ (14)
C7N2 1 25 G 3560–4850 Cal‡ (14)
C32N7 1 32.5 G 3560–4850 Cal‡ (14)
Real Alto, Ecuador (Valdivia 3)
Structure 1 1 20 G 4400–4800 Cal† (9)
Structure 1 3 24–26 G 4400–4800 Cal† (9)
Structure 10 3 18–24 G 4400–4800 Cal† (9)
Structure 10 1 24 G 4400–4800 Cal† (9)
Aguadulce, Panama (Late Preceramic)
350 3 24–28 G 5600 Cal (30)
Loma Alta, Ecuador (Early Formative)
SS275 5 22–26 G 5050–6250 Cal† (9)
SS275-2 6 16–24 G 5050–6250 Cal† (9)
SS292 2 19, 20 G 4550–6050 Cal† (9)
Sample 13 2 24, 28 C 4830–5280 Cal† (9)
Sample 7 1 27 C 4080–4410 Cal† (9)
Level 12 2 24, 28 S 4990–5310 Cal† (9)
Sample 11 1 18 C 4250–4860 Cal† (9)
Sample 10 1 22 C 4990–5310 Cal† (9)
Level 14 2 24, 24 S 4990–5310 Cal† (9)
Sample 9 1 28 C 4990–5310 Cal† (9)
*Standard and AMS radiocarbon dates from associated charcoal, 2s calibrated result. †Standard radiocarbon date from
associated charcoal, 2s calibrated result. ‡Standard radiocarbon dates from associated shell, adjusted for 12C/13C ratio,
2s calibrated results.
www.sciencemag.org SCIENCE VOL 315 16 FEBRUARY 2007 987
500 to 1000 yr B.P. (21). Several large, deep stone
metates were scattered overthesurfaceofthesite.
Excavation of a midden deposit yielded ceramic
griddle sherds and microlithic flakes that are often
associated with manioc processing (22). As at the
Three Dog site, the remains of manioc are con-
spicuously absent from an excavation that yielded
artifacts usually associated with manioc process-
ing (22). The same processing tools that contained
starches of chili pepper also contained remains of
maize. Root crops, including arrowroot, Myrosma
sp. (guapo), and a member of the Zingiberaceae
fami ly (ginger) also left their starchy remains.
When combined with the data from the Three
Dog site, the chili pepper microfossils from
Los Mangos del Parguaza support the notion
that a sophisticated mixed subsistence economy
of both root and seed crops occurred at these sites
that were initially categorized as being occupied
by manioc horticulturists (23).
Neither microfossils typical of wild species nor
transitional forms of Capsicum were recovered from
any site. The presence of domesticated plants used
as condiments rather than as staple foods during the
Preceramic period indicates that sophisticated agri-
culture and complex cuisines arose early throughout
the Americas and that the exploitation of maize, root
crops, and chili peppers spread before the introduc-
tion of pottery . Evidence from both macrobotanical
and microbotanical remains indicates that once chili
peppers became incorporated into the diet, they
persisted. Apart from the chili peppers, maize is
present at every site we sampled. Maize and chilies
occur together from the onset of this record until
European contact and, thus, represent an ancient
Neotropical plant food complex.
References and Notes
1. C. B. Heiser Jr., in Evolution of Crop Plants,N.W.Simmonds,
Ed. (Longman, London, 1976), pp. 265–268.
2. W. H. Eshbaugh, in New Crops, J. Janick, J. E. Simon, Eds.
(Wiley, New York, 1993), pp. 132–139.
3. B. Pickersgill, in Pre-Columbian Plant Migration,
D. Stone, Ed. (Harvard Univ. Press, Cambridge, MA,
1984), pp. 105–123.
4. B. Pickersgill, Biol. Zent. 107, 381 (1988).
5. L. Bohs, R. G. Olmstead, Syst. Bot. 22, 5 (1997).
6. Materials and methods are available as supporting
material on Science Online.
7. J. S. Raymond, in Pacific Latin America in Prehistory: The
Evolution of Archaic and Formative Cultures, M. Blake,
Ed. (Washington State Univ. Press, Pullman, 1999),
8. D. Lathrap, J. G. Marcos, J. A. Zeidler, Archaeology 30,2
9. J. A. Zeidler, in Archaeology of Formative Ecuador,
J. S. Raymond, R. Burger, Eds. (Dumbarton Oaks,
Washington, DC, 2003), pp. 487–527.
10. D. M. Pearsall, K. Chandler-Ezell, J. A. Zeidler,
J. Archaeol. Sci. 31, 423 (2004).
11. D. M. Pearsall, in Archaeology of Format ive Ecuador,
J. S. Raymond, R. Burger, Eds. (Dumbarton Oaks,
Washington, DC, 2003), pp. 213– 257.
12. K. Chandler Ezell, D. M. Pearsall, J. A. Zeidler, Econ. Bot.
60, 103 (2006).
13. D. R. Piperno, A. J. Ranere, I. Holst, P. K. Hansell, Nature
407, 894 (2000).
14. R. G. Cooke, A. J. Ranere, World Archaeol. 24, 114 (1992).
15. R. Dickau, thesis, Temple University, Philadelphia, PA
16. D. R. Piperno, D. M. Pearsall, The Origins of Agriculture in the
Lowland Neotropics (Academic Press, San Diego, CA, 1998).
17. L. Perry et al., Nature 440, 76 (2006).
18. M. J. Berman, A. K. Sievert, T. Whyte, Lat. Am. Antiq. 10,
19. M. J. Berman, D. M. Pearsall, Lat. Am. Antiq. 11, 219 (2000).
20. W. R. DeBoer, Am. Antiq. 40, 419 (1975).
21. K. Tarble, thesis, University of Chicago, Chicago, IL (2006).
22. L. Perry, J. Archaeol. Sci. 31, 1069 (2004).
23. L. Perry, Latin Am. Antiq. 16, 409 (2005).
24. C. E. Smith, in Prehistory of the Tehuacan Valley,D.S.
Byers, Ed. (Texas Univ. Press, Austin, 1967), pp. 220–255.
25. D. L. Lentz, M. P. Bea udry-Corbett, M. L. R. del Aguilar,
L. Kaplan, Lat. Am. Antiq. 7, 247 (1996).
26. B. Pickersgill, Am. Antiq. 34, 54 (1969).
27. L. A. Newsom, E. S. Wing, On Land and Sea: Native
American Uses of Biological Resources in the West Indies
(Univ. of Alabama, Tuscaloosa, 2004).
28. C. S. Spencer, E. M. Redmond, M. Rinaldi, Lat. Am. Antiq.
5, 119 (1994).
29. M. J. Berman, P. Gnivecki, World Archaeol. 26, 421 (1995).
30. A. J. Ranere, R. G. Cooke, in Paths to Central American
Prehistory, F. W. Lange, Ed. (Univ. Press of Colorado,
Niwot, 1996), pp. 49-77.
31. Comparative materials were supplied by the U.S. National
Herbarium, E. Perry, J. Perry, and I. Shimada. B. Smith
provided comments on the manuscript. Funding for
archaeological excavations and starch grain studies was
provided by the American Philosophical Society, the
Concejo de Desarrollo Científico y Humanístico de la
Universidad Central de Venezuela, the Escuela Superior
Politecníca del Litoral, the Foundation for Exploration
and Research on Cultural Origins, the Heinz Charitable
Trust Latin American Archaeology Program, NSF, the
Office of the Provost at Ithaca College, the Programa de
Antropología para el Ecuador, the Smithsonian National
Museum of Natural History, the Smithsonian Tropical
Research Institute, the Social Sciences and Humanities
Council of Canada, Temple University, the University of
Missouri Research Board, and Wenner-Gren.
Supporting Online Material
Materials and Methods
Tables S1 and S2
30 October 2006; accepted 21 December 2006
Multipotent Drosophila Intestinal Stem
Cells Specify Daughter Cell Fates
by Differential Notch Signaling
Benjamin Ohlstein and Allan Spradling*
The adult Drosophila midgut contains multipotent intestinal stem cells (ISCs) scattered along its
basement membrane that have been shown by lineage analysis to generate both enterocytes
and enteroendocrine cells. ISCs containing high levels of cytoplasmic Delta-rich vesicles activate
the canonical Notch pathway and down-regulate Delta within their daughters, a process that
programs these daughters to become enterocytes. ISCs that express little vesiculate Delta, or are
genetically impaired in Notch signaling, specify their daughters to become enteroendocrine cells.
Thus, ISCs control daughter cell fate by modulating Notch signaling over time. Our studies suggest
that ISCs actively coordinate cell production with local tissue requirements by this mechanism.
tem cells in adult tissues frequently reside in
specific anatomical positions known as
niches, whose microenvironment represses
premature differentiation and controls prolifera-
tion (1, 2). Several different signal transduction
pathways—including BMP (bone morphogenetic
protein), JAK/S TAT (Janus kinase/signal trans-
ducer and ac tivator of tra nscription), Wnt, and
Fig. 2. Modern and ar-
chaeological starch gran-
ules from Capsicum.(A)
Starch granule from the
baccatum var. pendulum
(aji mirasol) showing typ-
ical morphology. Note
the rounded lenticular
form and large, flat, cen-
tral depression. (B)Ar-
starch granule from Loma
starch granule of Capsi-
cum from Real Alto. (D)
Starch granule from a
modern specimen of Cap-
sicum annuum var. minimum. This starch granule is typical of those from wild peppers. (E) Side view
of a modern starch granule from Capsicum baccatum. Note the linear figure. (F) Side view of an
archaeological starch granule of Capsicum from Zapotal.
16 FEBRUARY 2007 VOL 315 SCIENCE www.sciencemag.org988