Journal of Archaeological Science 35 (2008) 1748e1755

http://www.elsevier.com/locate/jas

Evaluating chloroplast DNA in prehistoric Texas coprolites: medicinal,

dietary, or ambient ancient DNA?
Karl J. Reinhard a,*, Sergio M. Chaves b, John G. Jones c, Alena M. I~
niguez d
a
School of Natural Resources, 309 Biochemistry Hall, University of Nebraska e Lincoln, Lincoln, NE 68583-0758, USA
b
Laboratório de Ecologia da Escola Nacional de Saúde Pública da FIOCRUZ, Rua Leopoldo Bulh~oes 1480, térreo e Manguinhos,
CEP 21041-210 Rio de Janeiro, RJ, Brazil
c
Department of Anthropology, PO Box 644910, Washington State University, Pullman, WA 99164-4910, USA
d
Laboratorio de Genética Molecular de Microorganismos, Instituto Oswaldo Cruz-FIOCRUZ, Av. Brasil 4365, 21045-900 Rio de Janeiro, RJ, Brazil
Received 4 October 2005; received in revised form 21 November 2007; accepted 22 November 2007

Abstract

Molecular analysis of coprolites from Hinds Cave, Texas recovered chloroplast DNA sequences. The sequences were interpreted as evidence
of diet. We analyzed 19 Hinds Cave coprolites to evaluate the potential sources of the chloroplast DNA (cpDNA) and compared our results to
previous studies. This review shows that some cpDNA sequences could be from well-known prehistoric plants foods. Some other sequences
could have come from ambient plant material in the guts of small animals eaten by humans in antiquity. Using pollen concentration analysis,
we identify sources of ambient plant material which could have been inhaled or imbibed. It is even possible that cpDNA sequences are from
proplastids within ambient pollen grains themselves. However, three sequence types cannot be explained as resulting from only dietary or am-
bient sources. We suggest instead that these might be from medicinal or hallucinogenic plants. We compared these three sequences to existing
sequences in the GenBank. We found that these sequences are 100% matches for Rhamnus, Fouquieria, and Solanum.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Molecular biology; Pollen concentration; Zooarchaeology; Ambient plant residue; Coprolite

1. Introduction Rhamnaceae, and Solanales/Solanaceae. Poinar et al. (2001)

suggested that Helianthus or other genera might be repre-
In the Proceedings of the National Academy of Sciences, sented by Asteraceae cpDNA, Fouquieria by Fouquieriaceae
Poinar et al. (2001) reported on the recovery of ancient cpDNA, Quercus by Fagaceae cpDNA, Karwinskia, Condalia,
DNA from three Late Archaic Period human coprolites (desic- or Colubrina by Rhamnaceae cpDNA, and Nicotiana, Physa-
cated feces) from Hinds Cave, Texas. They found a diversity lis, Lycium, or Datura by Solanaceae cpDNA. At the time of
of ancient DNA sequences, including eight chloroplast DNA their research, 2000, sequences for these genera were unavail-
(cpDNA) sequences. Five cpDNA sequences were from plant able for comparison in GenBank.
taxa that were not identified in macrofossil analysis of the The cpDNA sequences are of particular interest because
same coprolites. They interpreted these five sequences as evi- they include some plant taxa that have never been encountered
dence of dietary plant use. The discovery highlighted the value in dietary residue in any past study of Hinds Cave coprolites.
of molecular biology in ancient diet reconstruction. We will Over 200 human coprolites from Hinds Cave have been the fo-
discuss these five sequences, and their origins, in this paper. cus of past studies dating from the Early Archaic to the Late
The five cpDNA sequences in question are Asterales/Aster- Archaic (Danielson and Reinhard, 1998; Dean, 2006; Edwards,
aceae, Ericales/Fouquieriaceae, Fagales/Fagaceae, Rhamnales/ 1990; Reinhard, 1988, 1992; Stock, 1983; Williams-Dean,
1978). These studies show a continuity of diet from Early to
* Corresponding author. Tel.: þ1 402 472 6858. Late Archaic times with relatively little dietary variation
E-mail address: kreinhard1@unl.edu (K.J. Reinhard). between periods (Reinhard, 1992). Of 25 taxa encountered in

0305-4403/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2007.11.013
 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755 1749

the coprolites reviewed by Reinhard (1992), significant varia- It would have been ideal to extract pollen from the exact
tion between different studies was noted for only four taxa; coprolite samples analyzed for cpDNA. In the most recent
wild onion, walnut, mesquite and dropseed grass. These analy- methods of coprolite analysis, macroscopic remains, phyto-
ses have revealed a general long-term pattern of plant use at liths, parasites, and pollen are recovered sequentially from
Hinds Cave through analysis of pollen, phytoliths, and macro- the same coprolite sample (Reinhard, 1988; Reinhard et al.,
scopic plant residues such as fibers and seeds. 2006a, b) or from two separate subsamples removed from
Microfossil and macrofossil coprolite analyses are fre- the same coprolite (Reinhard et al., 2002). The goal of this
quently concerned with issues of prehistoric or modern ambient analysis was to assess whether cpDNA sequences from the
plant residue contamination of samples (Chaves and Reinhard, small sample analyzed by Poinar et al. (2001) could have
2006; Reinhard et al., 2006a; Sobolik, 1988). For this paper, been derived from intentional dietary use or were signals of
ambient plant residue refers to plant material inadvertently ambient plant material. We could not access the same copro-
incorporated into coprolites in ancient times by several means. lites analyzed by these authors. So we based our evaluation
These range from consumption of ambient plant residue from on the independent analyses of larger series from the site. In
whole animals or insects, plant residue inadvertently swallowed this way, our aim was to detail whether or not these authors’
while processing plant materials for non-dietary purposes, and conclusions were valid based on a comprehensive review of
consumption of pollen-carrying plant residues via inhalation 169 coprolites analyzed by Williams-Dean (1978), Edwards
and swallowing. After deposition, ambient plant residue from (1990), and ourselves in this study.
cave deposits could infiltrate the coprolite matrix. The continuity in general diet between Early to Late Ar-
At Hinds Cave, cpDNA from ambient plant residue can chaic periods (Reinhard, 1992) has been previously demon-
come from several likely sources evaluated in this paper. An- strated. Thus, including all periods for comparison with the
cient inhabitants of the cave ate many species of small animals molecular analysis of Late Archaic coprolites is justified. Fur-
whole as evidenced by analysis of animal bones in the human thermore, we have the potential of demonstrating that ambient
coprolites (Reinhard et al., 2006c; Williams-Dean, 1978). This cpDNA was a long-existing aspect of lower Pecos environment
introduces vegetation from the prey digestive tracts into hu- which poses a problem for molecular interpretation of other
man digestive tracts (Reinhard et al., 2002, 2006c). Secondly, coprolites from other Archaic periods.
consumption of aqueous plant solutions, i.e. teas, can intro-
duce plant residue into intestinal tracts (Chaves and Reinhard, 2. Materials and methods
2006; Reinhard et al., 1991). Consumption of ambient, pollen-
associated plant material with food, drink, or inhalation might Five-gram fragments of Hinds Cave coprolites were re-
add cpDNA sequences into coprolites. Finally, a more remote moved from the interior of 19 coprolites that were brushed
possibility is that inhalation of pollen could have introduced clean of extraneous cave sediments. The coprolites were rehy-
cpDNA sequences into the coprolites. Although chloroplasts drated in 0.5% trisodium phosphate for 48 h. Five Lycopodium
are not present in angiosperm pollen, proplastids sometimes tablets (batch 201890, each containing 11,300  400 Lycopo-
are. Proplastids are DNA-carrying, organelle precursors to dium spores), were added to the rehydrating coprolites. The re-
chloroplasts. They can be in the generative and vegetative cells hydrated coprolites were then screened through a 300
in pollen grains (Bennett and Parducci, 2006; Pacini et al., micrometer mesh with a jet of distilled water. The water and
1992; Sangwan and Sangwan-Norreel, 1987; Schmitz and Ko- microscopic residues that passed through the screen were col-
wallik, 1987; Sodmergen et al., 1994). Recently, cpDNA was lected in a beaker and concentrated by centrifugation. The
recovered from ancient Fagus pollen (Paffetti et al., 2007). concentrated microscopic residues were then washed in glacial
To evaluate the potential sources of ambient plant residue, acetic acid and processed for 10 min in acetolysis solution (1
we undertook an analysis of 19 Middle Archaic Hinds Cave part sulfuric acid to 8 parts acetic anhydride) at 98  C. The res-
coprolites. We compared the results of this analysis with idues were then washed in glacial acetic acid, and subse-
Edward’s (1990) analysis of Late Archaic Hinds Cave copro- quently with distilled water. The residues were then treated
lites and Dean (2006); (Williams-Dean, 1978) analysis of with approximately 45% hydrofluoric acid for 24 h at room
Early Archaic coprolites. The coprolites analyzed by Edwards temperature. The residues were then washed three times
date between 2100 and 600 B.C. The coprolites analyzed by with distilled water, concentrated by centrifugation, and then
Dean date between 7710  80 and 7590  80 B.C. The copro- transferred to 2 dram vials with glycerin. Microscopic exami-
lites we analyzed are from Hinds Cave (41VV456) B-1, Lens nation was done at 400 and 1000 magnification. A mini-
5 (2560e2810 B.C.) and B-1 Lens 7 (3680  80 B.C.). These mum of 200 pollen grains was counted for each sample.
are uncorrected dates. Specific provenience information of the Pollen was identified by comparison to reference samples of
collection bags state B-1 B6-XI (K), B-1 Lens 5, and B-1 Lens lower Pecos pollen curated in the Palynology Laboratory, De-
7. The stratigraphy of Hinds Cave is summarized in several partment of Anthropology, Texas A&M University, and with
unpublished sources (Lord, 1984; Saunders, 1986; Shafer samples collected by Reinhard in 1991 and curated in the
and Bryant, 1977). However, this subject has been well treated School of Natural Resources, University of Nebraska-Lincoln.
on the Texas Beyond History Hinds Cave website. For consid- The processed samples are stored in glycerin. Lycopodium
eration of stratigraphy and provenience, see (http://www. spores were counted as they were encountered during the pol-
texasbeyondhistory.net/hinds/explore.html). len count. The pollen concentration values were calculated by
1750 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755

the following formula in terms of number of pollen grains per was identified. Five pollen types of unknown origin were also
gram of coprolite (modified from Maher, 1981). found. Therefore, a total of 53 palynomorph types were repre-
sented. Of the 47 identifiable flowering plant taxa, 22 occurred
Pollen concentration only in trace amounts (500e25,000 grains of pollen per gram
¼ ð½pollen grains counted= of coprolite). Thus, 25 taxa were found in high numbers in one
or more coprolite (in excess of 100,000 grains of pollen per
marker grains counted gram).
 number of markerpollen grains addedÞ= The 25 taxa of greater numerical occurrence are listed in
Table 1. In a few coprolites, the numbers of pollen grains per
coprolite weight: gram of certain taxa are very high, exceeding 100,000 grains
The pollen concentration data were then used to determine per gram, and indicate intentional consumption of plants.
the approximate numbers of pollen grains from identifiable Agave, Cactaceae, Dasylirion, and higher Poaceae counts are
plant taxa in the coprolites. The pollen data were compared among these.
with the list of taxa represented by cpDNA sequences to assess The remaining 21 taxa occur in numbers that are consistent
which sequences represent diet, which sequences represent with prehistoric ambient pollen inclusion in feces as discussed
ambient plant residue, and which sequences could be derived in recent papers (Reinhard et al., 2002, 2006a,b). Low Poaceae,
from both diet and/or ambient plant residue. Cheno-Am, and Low Spine Asteraceae made up the majority of
An analysis of animal remains in the coprolites was done to the Lower Pecos pollen rain in prehistory (Dean, 2006;
evaluate the potential of introduction of ambient plant residue Edwards, 1990; Reinhard, 1988; Sobolik, 1988; Stock, 1983;
through consumption of entire small lizards, rodents, fish, and Williams-Dean, 1978). Therefore, these pollen types have an
insects such as grasshoppers. The plant diets of the animals ambient source. Quercus, Larrea, Celtis, Eriogonum, and Ephe-
were researched to assess whether they eat the plant taxa dra are moderately common ambient types. However, when
found by Poinar et al. (2001). Larrea and Ephedra occur in high concentrations, their use as
We compared the rbcL gene sequence of three taxa recov- a medicinal tea can be interpreted (Reinhard et al., 1991).
ered by Poinar et al. (2001) with GenBank sequences (October, Table 2 presents a comparison of common ambient pollen
2007). Specifically, the consensus sequence of DNA sequences taxa found in this study compared to Dean (2006) and
of clones from Rhamnaceae, and Solanaceae, and the unique Edwards (1990). These data show that ambient plant material
clone sequence of Fouquieriaceae identified by the authors from Alnus (alder), Asteraceae (sunflower family), Celtis
were chosen. The best hits defined by maximal identity of (hackberry), Cheno-am (goosefoot family and pigweed ge-
a BLAST search were considered most likely plant origins. nus), Ephedra (mormon tea), Juglans (walnut), Pinus (pine),
Poaceae (grass family), Quercus (oak), and Ulmus (elm)
3. Results occurred in all Archaic time periods.
The comparison of the Fouquieriaceae, Rhamnaceae, and
Animal remains were commonly found in the coprolites. Solanaceae sequences recovered from Poinar et al. (2001)
One sample contained a rodent tooth. Bone fragments of with GenBank presented mixed results (Table 3). The Fou-
rodents were found in 16 samples. The absence of burning or quieriaceae sequence is a 100% match for a non-endemic spe-
other evidence of heat alteration of the bone indicates that these cies, Fouquieria columnaris. There is another Fouquieria
animals were eaten raw or in a semi-cooked state. Lizard scales species in the region F. splendens or ocatillo. It is more likely
were present in two samples. Fish bones and scales were pres- that the Fouquieria cpDNA comes from this species. The
ent in one sample and fish scales were present in another sam- Rhamnaceae sequence is 100% consistent with two species
ple. Non-human hair was present in six samples. Masticated of Rhamnus, or buckthorn. There are endemic species of
insect fragments unidentifiable to family were found in 13 cop- Rhamnus in the region. It is also identical with Sageretia
rolites and masticated grasshopper fragments were present in which is also present in the region. The Solanaceae sequence
two coprolites. Recent analyses report an association of trace is 100% match for Solanum lycopersicum and Lycopersicon
amounts of choridoid, festucoid, and panicoid grass phytoliths esculentum or the cultivated tomato. This is a mismatch for
(Reinhard et al., 2002) and fungal spores of animal dietary or- any species in the area. There is a wild solanaceous plant in
igin (Reinhard et al., 2006c) in coprolites with animal bone. the region, Solanum triquetrum.
These are best interpreted as food residue from the intestinal
tracts of prey animals eaten whole. Previous phytolith analysis 4. Discussion
of these Hinds Cave coprolites revealed traces of choridoid,
festucoid, and panicoid phytoliths (Danielson and Reinhard, As shown in this analysis, the prehistoric inhabitants of
1998). Thus, the coprolites from Hinds Cave present substan- Hinds Cave ate small animals in their entirety. Our findings
tial evidence that whole animals were consumed. This evidence support Dean’s previous analysis of 100 Hinds Cave Copro-
includes hair, teeth, bone, and phytoliths that are best explained lites (Williams-Dean, 1978). Reinhard et al. (2006c) presented
as the result of eating small animals in their entirety. a case that the consumption of small animals in their entirety
Pollen analysis revealed 47 identifiable taxa of flowering was a common prehistoric hunter-gatherer subsistence compo-
plants in the 19 coprolites. In addition, one type of fern spore nent. They also present a case that identification of small
Table 1
Pollen concentration values for coprolites in terms of number of grains of pollen per gram of coprolite
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755

Acer 62 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Agave 312 17,947 259 0 413 0 17,363 564,435 0 0 130 0 0 0 342 0 242 782 90
Alnus 0 0 416 0 0 0 0 0 0 0 0 0 0 284 0 0 0 0 0
Aster. H.S 125 4985 83 87 310 2166 0 25,425 1640 0 173 2152 0 284 685 56 242 168 0
Aster. L.S 499 10,968 583 1527 9302 2166 827 1695 1640 547 691 0 377 851 2739 564 484 949 271
Brassicac. 125 0 0 44 0 0 0 0 0 0 475 0 0 0 0 0 0 56 0
Carya 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 90
Cactaceae 0 14,956 0 0 310 0 0 328,839 7108 0 259 0 0 0 0 0 0 112 90
Celtis 1122 7976 0 0 827 283 0 0 0 0 86 0 0 0 0 56 121 56 90
Cheno Am 249 997 500 654 3721 6027 0 0 547 547 86 0 0 142 342 790 363 112 0
Dasylirion 561 121,641 83 349 3411 0 776,393 0 324,784 0 734 0 3767 2837 1712 395 242 56 0
Ephedra 0 0 0 44 103 94 0 0 547 0 43 0 0 142 0 0 121 112 0
Eriogonum 0 1994 0 0 207 659 827 0 0 0 216 0 0 284 0 0 242 0 0
Euphorb. 0 997 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Fabaceae 0 0 0 0 723 0 0 0 0 0 86 0 0 0 0 0 0 0 0
Fraxinus 0 0 0 0 0 0 0 0 0 547 0 0 0 0 0 0 0 0 0
Juglans 0 0 0 0 103 0 0 0 0 0 43 0 0 0 0 0 0 0 0
Larrea 62 0 0 305 0 377 0 0 0 0 0 0 0 142 0 0 0 168 0
Malvaceae 0 0 0 0 310 0 1654 33,900 0 0 0 1076 0 0 0 0 0 56 0
Pinus 62 1994 83 131 207 0 1654 0 0 0 43 0 0 567 0 0 0 335 0
Poaceae 1620 18,944 500 4145 8888 7722 0 11,865 0 659,410 1555 206,091 78,723 39,432 61,294 8574 22,519 7260 904
Quercus 0 0 0 44 0 0 0 0 0 0 130 538 0 0 0 121 56 0
Rhus 62 0 0 0 103 0 0 0 0 0 43 0 0 0 0 0 0 0 0
Ulmus 0 0 0 44 0 0 0 0 0 0 0 0 0 0 0 0 0 0 90
Yucca 0 0 0 0 0 0 1654 0 0 0 0 0 0 0 0 0 0 0 0
The taxa are listed by laboratory number for each coprolite. Seventeen taxa could be identified to the genus level. The others could be identified only to the family level and are abbreviated in the table: Aster.
H.S. ¼ high spine Asteraceae, Aster. L.S. ¼ low spine Asteraceae, Brassicac. ¼ Brassicaceae, Cheno-Am ¼ Chenopodiaceae and/or Amaranthaceae, Euphorb. ¼ Euphorbiaceae.

1751
1752 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755

Table 2 hackberry (Celtis), acorns, and mesquite. Ground Squirrel

Comparison of pollen types Hinds Cave coprolites that have a ambient source (Spermophilus) bones were found in 3% of the coprolites. In
in the lower Pecos region
the Hinds Cave region, Spermophilus species eat sunflower
Early Archaic Middle Archaic Late Archaic (Helianthus), cactus, mesquite, acorns, pine nuts, walnuts, salt-
n¼7 this study n ¼ 50
Dean (2006) n ¼ 19 Edwards (1990)
bush, Agave, wild gourd, cherries, sumac, spurge, serviceberry,
currant berries, and juniper berries. The consumption of her-
Alnus X X X
Asteraceae, Artemisia X X
bivorous rodents, with their digestive tracts, would undoubt-
Asteraceae, high spine X X X edly introduce cpDNA into human coprolites.
Asteraceae, low spine X X X Application of pollen concentration by previous researchers
Carya X shows that ambient pollen is found in high amounts in south-
Celtis X X X western coprolites (Dean, 2006; Edwards, 1990; Reinhard
Cheno/Am X X X
Ephedra X X X
et al., 1991, 2002, 2006a,b; Sobolik, 1988). This is supported
Fraxinus X X by our results in Table 1. With regard to pollen as a potential
Juglans X X X source of chloroplast sequences, cpDNA is generally inherited
Juniperus X maternally. Therefore, cpDNA should be absent in sperm cells
Pinus X X X of most plant species. However, cpDNA can be present in
Poaceae X X X
Quercus X X X
DNA-containing proplastids, especially in the vegetative cells
Salix X within pollen (Bennett and Parducci, 2006; Pacini et al., 1992;
Sarcobatus X Sangwan and Sangwan-Norreel, 1987; Schmitz and Kowallik,
Typha X 1987; Sodmergen et al., 1994). Therefore, DNA consistent
Ulmus X X X with cpDNA might be present in pollen grains. As more anal-
ysis of pollen grains for cpDNA are published, it becomes
animals to species level is very difficult in coprolite analysis apparent that we can not exclude inhaled pollen as a source
and that genus identification is more valid. Dean detailed the of cpDNA in coprolites. For example, Paffetti et al. (2007)
small animal consumption habits of Early Archaic Hinds report an estimated 1000 cpDNA molecules per Fagus pollen
Cave dwellers (Williams-Dean, 1978). The food habits of grain. Fagus is a close relative of Quercus (oak) and one genus
these mammals are summarized by Davis (1974). The follow- in the Fagaceae family represented by cpDNA sequences in
ing data from these sources. Wood rat bones (Neotoma) were Poinar et al. (2001) study. There is no published study on
found in 19% of Hinds Cave coprolites. Eating whole wood- the cpDNA content of Quercus pollen. Our analysis shows
rats would introduce cpDNA into human digestive systems and average of 47 Quercus pollen grains per gram of coprolite.
because in the area of Hinds Cave, various woodrat species If Quercus, like Fagus, contains cpDNA in pollen grains, then
eat cactus, acorns, sotol (Dasylirion), Agave, and mesquite. even small numbers of inhaled pollen could introduce many
Cotton rat (Sigmodon) bones were found in 13% of the copro- thousand cpDNA molecules into coprolites.
lites. They eat grasses, sedges, and herbs. Peromyscus, which More importantly, pollen signals the presence of ambient
includes many species of small mouse, was found in 2% of the plant material that can be included in drinking water. Thus,
coprolites. Peromyscus species in the area eat seeds, insects, accidental ingestion of pollen-associated plant structures
might introduce cpDNA into the human digestive tract. For
example, Quercus disseminates pollen from catkins which
Table 3
The most likely plant origins as represented by the best hits defined by max-
are structures that fragment and fall from the source tree.
imal identity of a BLAST search of the rbcL gene sequence of three taxa re- Thus, Quercus pollen and photosynthetic plant cells are ambi-
covered by Poinar et al. (2001) are presented in column 5 under the heading ent plant residues from the same source. If pollen grains were
Genus ingested accidentally with other plant microresidues such as
Sequence GenBank Subfamily GenBank Genus GenBank catkin, flower, florette, leaf or stem fragments, cpDNA could
family similarity similarity similarity be detected in molecular analysis of coprolites. We suggest
identified (%) (%) (%) that the ingestion of ambient plant residue can be signaled
Rhamnaceae 98e100 Rhamneae 98e100 Rhamnus 99e100 both by cpDNA sequences and pollen in coprolites.
Rhamanidium 100 Table 4 compares the list of cpDNA sequences recovered by
Sageretia 100
Solanaceae -100 Solanoidae Solanum 99e100
Poinar et al. (2001) with the percentage of coprolites that con-
Lycopersicon 100 tain ambient pollen from the taxa represented by the cpDNA
Fouquieriaceae -100 Fouquieria 100 found in this study. This information is compared with the
For Rhamnaceae, Rhamnus and Sageretia are the most likely sources of the food habits of Hinds Cave prey animals and review of dietary
cpDNA. Rhamanidium is not endemic to the region. For Solanaceae, the use of these plants in previous coprolite studies from the lower
rbcL gene sequence of the cultivated tomato is the best match. Certainly cul- Pecos (Reinhard, 1992). Asteraceae cpDNA sequences were
tivated tomato could not be the source of the cpDNA. It is conceivable that this found in the three coprolites analyzed by Poinar et al.
cpDNA has an origin in native, wild species of Solanum. Importantly, the gen-
era suggest for Solanaceae by Poinar et al. (2001) Nicotiana, Physalis, Lycium,
(2001), but these coprolites contained no Asteraceae macrofos-
and Datura do not match the recovered Solanaceae cpDNA. Finally, Fouquie- sils. Asteraceae pollen is present in all 19 coprolites and in all
ria is a perfect match for the Fouquieriaceae cpDNA. Archaic periods. Asteraceae genera are eaten by the prey
 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755 1753

Table 4 Fabaceae and Liliaceae sequences were found in coprolites

To interpret the cultural versus ambient source cpDNA found in molecular that also had macrofossils of these families (Table 4). They are
analysis (Poinar et al., 2001), pollen signals of ambient plant residue and
the dietary habits of lower Pecos animals that were eaten by prehistoric
very common Archaic foods for all periods. Therefore, there
hunter-gathers must be considered can be little doubt that the cpDNA sequences represent dietary
Taxa found Genera % of coprolites Genera Genera
use. The fact that we found high concentrations of Liliaceae
by Poinar suggested w/pollen eaten found (Dasylirion) pollen in the majority of the coprolites supports
et al. (2001) by Poinar et al. contamination by prey in previous the dietary conclusion for cpDNA.
(2001) animals? studies The most interesting cpDNA sequences are from taxa for
Davis Reinhard which there are no previous documented dietary use at Hinds
(1974) (1992)
Cave in coprolite analysis, nor evidence in pollen analysis, and
Asterales/ Helianthus and 100% Yes Helianthus for which there is no record of consumption by prey animals.
Asteraceae many
other genera
These sequences are Fouquieriaceae, Rhamnaceae, and Sola-
Ericales/ Foquieria 0% No None naceae. They cannot be explained as ambient plant residue,
Fouquieriaceae nor can they be considered as dietary in the light of what is
Fabales/ Acacia, Prosopis, 11% Yes Prosopis, known about prehistoric foods from lower Pecos coprolite
Fabaceae Sophora, Mimosa Acacia studies. However, there are genera in each of these taxa that
Fagales/ Quercus 26% Yes None
Fagaceae
were used as medicines ethnographically in the lower Pecos
Liliales/ Allium, Nolina 74% Yes Allium, region. This raises the possibility that cpDNA reflects medic-
Liliaceae Dasylirion, Yucca Dasylirion inal or ritual use. It is possible that the cpDNA originated in
Rhamnales/ Colubrina, 0% No None medicinal preparations of some of these genera.
Rhamnaceae Condalia Comparison of the cpDNA sequences with GenBank data
Karwinskia
Rosales/ Celtis 53% Yes Celtis
verifies Poinar et al. (2001). The find of Fouquieria (ocotillo)
Ulmaceae cpDNA is very interesting. The bark of this plant is streaked
Solanales/ Nicotiana, Physalis, 0% No None with green, photosynthetic tissue. It is used to make medicinal
Solanaceae Lycium, or Datura tea or poultices for a variety of purposes (Powell, 1988).
Columns 1 and 2 present the cpDNA discoveries and interpretation. Column 3, Therefore, the find of cpDNA sequences from Fouquieria is
based on pollen analysis of 19 coprolites, presents the percentage of coprolites consistent with using the bark for medicinal beverages.
containing plant residue signaled by pollen in coprolites. Fourth columns notes For Rhamnaceae, Poinar et al. (2001) suggested Karwin-
if the taxon represented by cpDNA is a known food habits of animals eaten by
Hinds Cave inhabitants [8]. The final column list general in the taxa repre-
skia, Condalia, and Colubrina. Karwinskia and Culubrina
sented by cpDNA that are known to be Hinds Cave foods [16]. are toxic. However, the roots and bark of Condalia are a source
for antibacterial compounds (Powell, 1988). Condalia seeds
animals consumed by Hinds Cave inhabitants. Reinhard (1992) were found in 3000 year-old deposits in Hinds Cave and the
summarizes that sunflower seeds were a common Hinds Cave fruits are edible (Dering, 1979). However, our comparison of
food for humans. Therefore, the cpDNA sequences could be the cpDNA sequence with GenBank reveals a 100% match
ambient plant residue from several sources or derived from with species of Rhamnus and Segeretia. One endemic Rham-
plants intentionally eaten by Hinds Cave residents. nus species is used in the Southwest as a cathartic.
Fagaceae cpDNA sequences were found in 2 of 3 analyzed Poinar et al. (2001) suggested that Nicotiana, Physalis,
coprolites, but Fagaceae macrofossils were not found. Poinar Lycium, and Datura are the genera represented by Solanaceae
and his colleagues suggested that acorns (Quercus) were the cpDNA sequences. Lycium and Physalis were prehistoric food
source of this cpDNA. In previous studies of Hinds Cave cop- plants in parts of the Southwest, but they are not known as
rolites, acorns are absent in macrofossil and microfossil analy- foods from Hinds Cave or other Lower Pecos sites (Reinhard,
sis (Reinhard, 1992). However, Quercus pollen was found in 1992). There is circumstantial, artifact evidence that Nicotiana
five coprolites and in all Archaic time periods. Acorns are eaten was smoked in the Hinds Cave region in the form of stone
by Hinds Cave prey animal species. Oak catkin fragments, sig- pipes (Chandler, 1990, 1992; Chandler and Boyd, 1995). If it
naled by pollen, could have been consumed with drinking was also chewed, then cpDNA would be swallowed. However,
water. It is unknown as to whether inhaled oak pollen could there is no published record of Nicotiana quids or other Nico-
be a source of cpDNA, so this source cannot be discounted. tiana evidence in the region. Datura is a hallucinogenic plant
Therefore, we believe that there is a high probability that Faga- that is associated with rock art in the region (Adovasio and
ceae cpDNA sequences came from ambient plant residue. Fry, 1976; Boyd and Dering, 1996). Teas were made from
Poinar et al. (2001) recovered multiple clones of Ulmaceae this plant and it is possible that cpDNA was ingested in this
from all three of their coprolites but Ulmaceae (Celtis) seed way. Poinar et al. (2001) provide data that should stimulate
from only one. As reviewed by Reinhard (1992), Celtis was a mi- more research into defining the enteric use of these plants.
nor human dietary component for several Archaic periods. Also, However, our match of the cpDNA Solanaceae sequence
Celtis is part of the recorded diet for Hinds Cave prey animals. was 100% consistent with cultivated tomato. This is an impos-
We recovered Celtis and/or Ulmus pollen from 10 coprolites in sibility since cultivated tomato never existed in the region pre-
our study. Thus, Ulmaceae cpDNA sequences could be from historically. It is not consistent with the genera suggested by
ambient or intentional consumption of Ulmaceae plants. Poinar et al. (2001). Wild species of Solanum could
1754 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755

conceivably have the same rbcL gene as cultivated tomato. At copy of the paper and sponsored Reinhard and Jone’s work
this point, however, the Solanaceae cpDNA eludes specific with Hinds Cave remains. The pollen processing was done
interpretation. by Jones and Reinhard in Bryant’s pollen laboratory at Texas
A&M University, Department of Anthropology assisted by
5. Conclusions Laurie Zimmerman. The analysis of the pollen data was done
at the Laboratório de Ecologia da Escola Nacional de Saúde
The implications of cpDNA sequences in Hinds Cave cop- Pública, Fundaç~ao Oswaldo Cruz and evaluation of GenBank
rolites may be more complicated than the dietary explanation sequence association at the Laboratório de Genética Molecular
presented by Poinar et al. (2001) but at the same time they are de Microorganismos, Instituto Oswaldo Cruz, Fundaç~ao Os-
more interesting. We believe that they made the correct dietary waldo Cruz, Rio de Janeiro, Brazil with funding from the Ful-
interpretation for Fabaceae and Liliaceae. However, the bright Commission, CAPES, and CNPq. Valdirene dos Santos
cpDNA sequences for Ulmaceae, Fagaceae, and Asteraceae Lima and Lı́via Maria Rubem Vidal from Laboratório de Ge-
could be derived whole or in part from ambient plant residue. nética Molecular de Microorganismos (IOC/FIOCRUZ) assis-
The most intriguing discoveries are of cpDNA from Fouquier- ted in the analysis of the DNA sequences.
iaceae, Rhamnaceae, and Solanaceae that have medicinal or hal-
lucinogenic genera. The cpDNA sequences, therefore expand the
potential of paleoethnobotany to include the non-dietary use of
economic plants. In this case, molecular biology may expand
References
our knowledge of prehistoric plant use to include plants used
for drugs that cannot be discovered using conventional coprolite Adovasio, J.M., Fry, G.F., 1976. Prehistoric psychotropic drug use in north-
analysis methods. They provide revealing information into plant eastern Mexico and Trans-Pecos, Texas. Econ. Bot. 30, 94e96.
use beyond the more mundane dietary explanation and touch on Bennet, K.D., Parducci, L., 2006. DNA from pollen: principles and potential.
paleopharmacology (Reinhard et al., 1991). Holocene 16, 1031e1034.
The field of metagenomics is expanding. The potentiality of Boyd, C.E., Dering, J.P., 1996. Medicinal and hallucinogenic plants identified
in the sediments and pictographs of the Lower Pecos, Texas Archaic.
recovering hundreds or thousands of clones from a single copro- Antiquity 70, 256e275.
lite is becoming a reality. These valuable data must be inter- Chandler, C.K., 1990. A stone pipe from the lower Pecos Region of Val Verde
preted with an ecological, cultural, and technological County, Texas. La Tierra 17, 31e34.
understanding of the people who left the coprolites. Obviously, Chandler, C.K., 1992. Additional stone pipes from the lower Pecos River in
cpDNA sequences can come from dietary use. In addition to Val Verde County, Texas. La Tierra 19, 34e37.
Chandler, C.K., Boyd, J.B., 1995. A decorated stone pipe from Val Verde
diet, medicinal use of plants must be considered. Infusions, County, Texas. La Tierra 22, 35e37.
i.e. beverages, could be another source of cpDNA clones. Ambi- Chaves, S.A.M., Reinhard, K.J., 2006. Critical analysis of prehistoric evidence
ent plant residue inside vertebrate and invertebrate foods must be of medicinal plant use, Piauı́, Brazil. J. Palaeogeogr. Palaeoclimatol. Palae-
considered as a source of unintentional plant ingestion. For ex- oecol. 237, 110e118.
ample, the intentional consumption of grasshoppers results in Danielson, D.R., Reinhard, K.J., 1998. Human dental microwear caused by
calcium oxalate phytoliths in prehistoric diet of the lower Pecos region,
the unintentional consumption of the grasshoppers’ foods. Plant Texas. Am. J. Phys. Anthropol. 107, 297e304.
residue can enter the intestinal tract from processing plant fibers Davis, W.B., 1974. The Mammals of Texas. Texas A&M University, College
with teeth. This could be another source of cpDNA clones that Station.
was unintentional. Plant residue in water or air, represented by Dean, D.W., 2006. The science of coprolite analysis: the view from Hinds
pollen in coprolites, would be a source of ambient environmen- Cave. J. Palaeogeogr. Palaeoclimatol. Palaeoecol. 237, 67e79.
Dering, J.P., 1979. Pollen and Plant Macrofossil Vegetation Record Recovered
tal cpDNA sequences. Finally, plant residues on tools or even on from Hinds Cave, Val Verde County, Texas. M.A. thesis, Texas A&M Uni-
hands that are left from processing plants for non-dietary pur- versity, College Station.
poses could enter the mouth. In short, any plant that is held in Edwards, S.K., 1990. Investigations of Late Archaic Coprolites: Pollen and
the mouth, or passes through the mouth, or that is clean off of Macrofossil Remains from Hinds Cave (41VV456), Val Verde County,
tools and skin with the mouth for any reason, could leave trace Texas. M.S. thesis, Texas A&M University, College Station.
Lord, K.J., 1984. The Zooarchaeology of Hinds Cave (41 VV 456). Ph.D. dis-
amounts of cpDNA for metagenomic recovery and interpreta- sertation, University of Texas at Austin.
tion. For this reason, researchers have to develop a clear under- Maher, L.J., 1981. Statistics for microfossil concentration measurements
standing of the archaeology of the research area to develop and employing samples spiked with marker grains. Rev. Palaeobot. Palynol.
understanding of cultural and environmental sources of cpDNA. 32, 153e191.
At the same time, molecular biology expands greatly what we Pacini, E., Taylor, P.E., Singh, M.B., Knox, R.B., 1992. Development of plas-
tids in pollen and tapetum of rye-grass, Lolium perenne L. Ann. Bot. 70,
can see with our microscopes and therefore challenges us to 179e188.
explore the nexus of archaeobotany, ethnobotany, and ecology Paffetti, D., Vettori, C., Caramelli, D., Vernesi, C., Lari, M., Paganelli, A.,
to optimize the interpretation of molecular data. Ladislav, P., Giannini, R., 2007. Unexpected presence of Fagus orientalis
complex in Italy as inferred by 45,000-year-old pollen samples from Ven-
Acknowledgements ice lagoon. BMC Evol. Biol. 7 (Suppl. 2), S6.
Poinar, H.N., Kuch, M., Sobolik, K.D., Barnes, I., Stankiewicz, A.B.,
Kuder, T., Spaulding, W.G., Bryant, V.M., Cooper, A., Pääbo, S., 2001.
Dr. Vaughn Bryant, Jr., Department of Anthropology, Texas A molecular analysis of dietary diversity for three Archaic Native Ameri-
A&M University, College Station, TX, read over a preliminary cans. Proc. Natl. Acad. Sci. USA 98, 4317e4322.
 K.J. Reinhard et al. / Journal of Archaeological Science 35 (2008) 1748e1755 1755

Powell, A.M., 1988. Trees and Shrubs of Trans-Pecos Texas Including Big Reinhard, K.J., Szuter, C., Ambler, J.R., 2006c. Hunter-gatherer use of small
Bend and Guadalupe Mountains National Parks. Big Bend Natural History animal food resources. Internat. J. Osteoarchaeol. 17, 416e428.
Association, Big Bend National Park. Sangwan, R.S., Sangwan-Norreel, B.S., 1987. Ultrastructural cytology of plas-
Reinhard, K.J., 1988. Diet, Parasitism, and Anemia in the Prehistoric South- tids in pollen grains of certain androgenic and non-androgenic plants. Pro-
west. Ph.D. Dissertation, Texas A&M University, College Station. toplasma 138, 11e22.
Reinhard, K.J., 1992. Patterns of diet, parasitism, and anemia in prehistoric Saunders, J.W., 1986. The Economy of Hinds Cave. Ph.D. dissertation, South-
west North America. In: Stuart-Macadam, P.S., Kent, S. (Eds.), Diet, De- ern Methodist University, Dallas.
mography, and Disease: Changing Perspectives on Anemia. Aldine de Schmitz, U.K., Kowallik, K.V., 1987. Why are plastids inherited maternally in
Gruyter, New York, pp. 219e258. Epilobium? Plant Sci. 53, 139e146.
Reinhard, K.J., Hamilton, D.L., Hevly, R.H., 1991. Use of pollen concentra- Shafer, H.J., Bryant, V.M., Jr., 1977 Archeological and Botanical Studies at
tion in paleopharmacology: coprolite evidence of medicinal plants. J. Eth- Hinds Cave, Val Verde County, Texas. Annual Report to the National Sci-
nobiol. 11, 117e134. ence Foundation. On File with the Texas A&M University, Anthropologi-
Reinhard, K.J., Daniels, M., Danielson, D.R., Chaves, S.M., 2002. cal Laboratory, as Special Series 1.
Multidisciplinary coprolite analysis. In: Geib, P.R., Keller, D.R. (Eds.), Sobolik, K.D., 1988. The importance of pollen concentration values from cop-
Bighorn Cave: Test Excavation of a Stratified Dry Shelter, Mojave County, rolites: an analysis of southwest Texas samples. Palynology 12, 201e221.
Arizona. Bilby Research Center, Flagstaff, pp. 135e152. Sodmergen, Y.Y., Kuroiwa, T., Hu, S.Y., 1994. Cytoplasmic DNA apportion-
Reinhard, K.J., Edwards, S.K., Damon, T.R., Meier, D.K., 2006a. Pollen con- ment and plastid differentiation during male gametophyte development
centration analysis of Ancestral Pueblo dietary variation. J. Palaeogeogr. in Pelargonium zonale. Sex. Plant Reprod. 7, 51e56.
Palaeoclimatol. Palaeoecol. 237, 92e109. Stock, J.A., 1983. The Diet of Hinds Cave (41 VV456), Val Verde County, Texas:
Reinhard, K.J., LeRoy-Toren, S., Danielson, D.R., 2006b. Salmon Ruin copro- The Coprolite Evidence. M.S. thesis, Texas A&M University, College Station.
lites: San Juan diet. In: Reed, P.F. (Ed.), Thirty-Five Years of Williams-Dean, G., 1978. Ethnobotany, and Cultural Ecology of Prehistoric
Archaeological Research at Salmon Ruins, Vol. 3. Center for Desert Ar- Man in Southwest Texas. Ph.D. Dissertation, Texas A&M University, Col-
chaeology and Salmon Ruins Museum, Tucson, pp. 875e888. lege Station.