Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

This chapter summarizes the current perspectives on paleoethnobotany, and the methods and techniques involved in the analysis of archaeological plant remains. The topic is not new, and for nearly three quarters of a century, paleoethnobotanists have not only contributed substantially to a broad range of archaeological questions, but have also complied detailed guides and summaries of state-of-the-art recovery techniques and laboratory analyses. What is new are the more careful and explicit treatments of the processes that have led to the formation of the paleoethnobotanical record. These processes – or what can be thought of as additional variables – are the subject of field tests and laboratory experiments that have been conducted around the world. Because understanding these processes can contribute to the advancement of paleoethnobotany and are essential to attempts at integrating information derived from plant and animal assemblages, they drive much of the discussions in the pages that come up later (for similar treatment of zooarchaeological remains, see Peres, chapter “Methodological Issues in Zooarchaeology,” this volume).

1 Why Study Paleoethnobotany?

The aim of archaeology is to learn about past human behavior through material evidence. The analysis of plant remains from archaeological contexts has facilitated that aim since the days when Kunth (1826) described the botanical traces from ancient Egypt, and Heer (1866, 1865) provided lists of vegetable foods used by the so-called Swiss-Lake Dwellers. These kinds of studies – investigating human–plant interrelationships – became more formalized when, in 1941, Jones published “The Nature and Status of Ethnobotany.” The analysis and interpretation of archaeologically derived plant remains, or “paleoethnobotany” as defined by Hastorf and Popper (1988:2), burgeoned in the 1970s as systematic recovery techniques became common. Today, paleoethnobotanists contribute information on diet, origin of agriculture, environmental change, resource availability and use, stone tool and pottery functions, and long-term socioeconomic changes, to name a few. In short, paleoethnobotanical research is increasingly recognized as a valuable tool in unlocking the secrets of past human behaviors and beliefs, and has the potential to add to current discussions of climatic change and sustainability.

1.1 Types of Paleoethnobotanical Evidence

The kinds of plant evidence that may be collected from archaeological sites vary from DNA to pollen to seeds. It is common practice to group the various kinds of remains according to the methods of observation, recovery, and analyses. “Macrobotanical remains” include complete or fragmented plant parts that are either visible to the naked eye or with a low-power microscope. “Microbotanical remains,” by contrast, refer to tiny plant parts that are visible only under high-power magnification. Chemical and molecular evidence are residuals that can require very different and complex means of extraction and analyses. While macrobotanical and microbotanical remains and chemical and molecular evidence are treated herein, I often focus the discussions on macrobotanical remains, but try to direct the interested reader to more comprehensive coverage of the other categories as well.

1.1.1 Macrobotanical Remains

Macrobotanical remains are perhaps the most commonly studied, and contribute evidence for many archaeological questions. This class of remains consists of plant traces that are large enough to be recognized with the naked eye or low-powered microscope (Ford 1979; Fritz 2005; Pearsall 2000). The different kinds of macrobotanical remains include wood, seeds, fruits, tubers, and nutshell, as well as fibers that have been woven into fabric or stems that have been manufactured into hats, cloaks, baskets, or mats. Their recovery from archaeological contexts may involve hand collecting, screening, or flotation (Fritz 2005; Pearsall 2000).

Wood tends to be the most ubiquitous of the macrobotanical remains. This durable material is making a growing contribution to archaeological reconstructions of natural environments, climate change, human use of timber, and dendrochronological and radiocarbon dating of archaeological sites (e.g., Asouti 2003; Dolby 2008; Figueiral and Mosbrugger 2000; Hastorf et al. 2005; Kreuz 1992; Kuniholm 1990; Kuniholm and Newton 1996; Smart and Hoffman 1988).

Remains of seeds, fruits, nutshell, and tubers are typically used to infer diet and subsistence strategies. These vestiges have been vital to our understanding of biodiversity (e.g., Black 1978), seasonality (e.g., Dark 2004), and the landscape (e.g. Fairbairn 2008). It nearly goes without saying that the analysis of ­macrobotanical remains has led to the publication of numerous books and articles dedicated to furthering our knowledge of ancient plant domestication and the origins of agriculture (e.g., Bellwood 2005; Smith 2001, 2006). Their interpretations also have implications for discussion about social, political, and economic systems (e.g., Hastorf and Johannessen 1993; Lepofsky and Lyons 2003; Weiss and Kislev 2004).

1.1.2 Microbotanical Remains

Microbotanical remains require a high-power microscope for identification, and have therefore been named so (Ford 1979; Pearsall 2000). Pollen is one of the ­several kinds of very small plant remains that are of interest to the paleoethnobotanist. It forms in an anther or what comprises the male organ of reproduction in seed-bearing plants (Bryant and Holloway 1983; Pearsall 2000). Spores, the asexual reproductive cells of fungi, ferns, and some algae, are traditionally included in pollen analysis or palynology. Recently, palynological studies have expanded to include other botanical entities composed of sporopollenin-like material (see Rowe and Kershaw 2008; van Geel 2001 for additional descriptions). Upon recovery, the shape, size, and surface features of palynomorphs and non-palynomorphs are used to assign a specimen to a particular family, genus, or species.

Sears (1937) was one of the first palynologists to address archaeological issues when he modeled the paleoenvironment of parts of the eastern United States. A few years later, Iversen’s (1941) work enabled the identification of the beginning of food production in Denmark. These studies set the stage for more recent palynological projects with an archaeological bend.

Worldwide, palynology is increasingly recognized for enhancing our understanding of past environments and human land-use strategies (e.g., Behre 2007; Birks 2007; Cordova and Lehman 2003; Hunt and Rushworth 2005; Kelso and Good 1995; Kelso et al. 2000; Mercuri 2008). While such studies cannot produce an exact picture of past environments, palynology is matchless in yielding some idea about fluctuations in vegetation that might be associated with climate change and/or human impact (Davis 1994; Faegri et al. 1989). Recent innovations include an automated pollen analysis proposed by France et al. (2000) and Fyfe’s (2006) computer-based modeling technique that combines pollen analysis and Geographic Information Systems (GIS) to test the landscape hypotheses.

Palynology can contribute other information relevant to people’s exploitation of plants (Bryant and Hall 1993; Pearsall 2000). Pollen collected from middens often indicates the types of plants collected and utilized for food or other economic purposes by prehistoric cultures. Fossil pollen found in floor sediments can be used to suggest potential types of room utilization. Scrapings from the inside surfaces of ceramic vessels may include fossil pollen from plants that were stored in or eaten from those vessels. Scrapings from the surfaces of grinding stones may contain the whole or broken fragments of pollen that adhered to the surfaces of seeds that had been ground into flour, and the analyses of sediments attached to the inside surfaces of basketry can sometimes suggest functional uses of those artifacts. The analysis of pollen by Piperno and Pearsall (e.g., 1998) has opened the doors to a new ­understanding of the origin of agriculture in regions of the world where macrobotanical remains are lacking.

Phytoliths another type of microbotanical evidence, are produced when certain higher plants absorb silica in a soluble state from ground water, which is then deposited in intracellular and extracellular locations in the epidermal tissues of stems, leaves, and roots (Esau 1965; Pearsall 2000; Piperno 2006; Rovner 1983). There, the silica solidifies as “phytoliths” or discrete, microscopic particles of varying sizes and shapes that are consistent with a family, genus, or species of plant. After the death and decay of the plant, the phytoliths are deposited into soils and sediments.

Phytoliths can be common in hearths and ash layers, but they can also be found inside pottery, plaster, and even on stone tools and animal teeth. Phytoliths are inorganic; thus they survive in a well-preserved state over long periods of time. In fact, Piperno (2006) indicates that they arguably are the most durable terrestrial plant fossil known to science. It is precisely their ability to withstand many of the rigors of nature that affords knowledge about plant use in regions where the ­recovery of macro remains has been poor. For further understanding, I direct the reader to the work of Piperno and Pearsall (1998) in the lowland Neotropics. In addition, Pearsall (2000:356) indicates that some phytoliths can be dated: those containing carbon can be radiocarbon dated (Mulholland and Prior 1993), and preliminary research ­utilizing thermoluminescence (Rowlett and Pearsall 1993) has proved encouraging.

Starch grains are another form of micro remains that are increasingly acknowledged for their contribution to paleoethnobotanical studies. These granules form within specialized organs called “plastids.” There are two kinds of plastids: (1) chloroplasts which occur primarily in leaves and green stems, and (2) amyloplasts which occur within roots, rhizomes, tubers, and seeds (Bailey 1999). The size and shape of starch grains differ by taxa (Coil et al. 2003). Czaja (1978) describes the structure of starch grains in relation to classification of vascular plant families.

Starch granules found on stone tools have received increasing attention from researchers because they can reveal information about human diet and household activities. For example, Piperno and Holst (1998) interpreted starch grains found on prehistoric stone tools as signs of early tuber use and agriculture in Panama; Barton (2007) examined museum artifacts to assess the potential of cooked, starchy foods; and Horrocks et al. (2004) were able to recover starch grains from prehistoric coprolites. Starch grain analysis has proven invaluable, yielding information about roots and tubers, many of which were dietary staples but difficult to document archaeologically (e.g., Chandler-Ezell et al. 2006; Cortella and Pochettino 1994; Dickau et al. 2007; Perry 2002; Ugent et al. 1987; see also Dickau, this volume).

1.1.3 Chemical and Molecular Evidence

Chemical and molecular evidence derive from residual elements found in sediments, ceramic vessels, crevices of stone tools and teeth, human skeletal remains, or in the vestiges of surviving plant tissues (Ford 1979; Pearsall 2000). Most researcher are well aware that C13/C12 ratios can be employed to determine the relative importance of maize (Zea mays) in past human diets (e.g., Boyd et al. 2008; Vogel and van der Merwe 1977; Wagner 1987). Araus et al. (2003) use carbon isotope discriminations to quantify cereal yields. Dietary contributions of plants can be assessed through ratios of nitrogen isotopes and proportions of strontium and calcium in human bone (Ambrose and DeNiro 1986). Compounds surviving in plant remains, including proteins and lipids, can provide an alternative basis for their identification and offer the prospect of better understanding of human diet, the origin of food production, patterns of trade in plant products, and uses of stone tools and pottery (e.g., Lombard and Wadley 2007; Malainey et al. 1999; Rottlander 1990).

DNA can occur in charred and uncharred plant remains; however, as I learned, when working with a plant geneticist at the University of Missouri, Columbia in attempting to extract DNA from ancient Iva annua remains, such DNA evidence is often fragmentary and degraded which makes it difficult to amplify (e.g., Wright 1994). More recently, Giles and Brown (2008) report on improved methodologies for extracting and amplifying DNA. Jones (2002) and Ross-Ibarra et al. (2007) discuss the potential of genetic evidence for understanding plant domestication. Rollo et al. (2002) report on the DNA analysis of the intestinal contents of Otzi, a glacier mummy from the Alps. Equally interesting is Poinar et al.’s (2001) discussion about the dietary diversity of three archaic Native Americans based on molecular analysis.

Other synergies between chemistry and archaeological plant analysis include Lane and colleagues’ (Lane et al. 2008) use of stable carbon isotope composition of tropical lake sediments to reconstruct maize cultivation. Braadbaart and I (Braadbaart et al. 2007) have used spectrographic analysis to understand the carbonization of macrobotanical remains. While chemical and molecular investigations open the door to evidence that several decades ago were unimaginable, Leach (1998) and Reber and Evershed (2004) caution against the uncritical use of chemical and molecular data; indeed, inconsistent results were obtained while conducting blind tests with commercial laboratories.

2 Deposition and Preservation of Plant Remains

Understanding how plant remains came to be a part of an archaeological site is essential. For instance, pollen can move through the environment in several ways (Bryant and Hall 1993; Pearsall 2000). Most conifers distribute their pollen by wind; consequently, pollen from a single tree may be transported miles, even ­hundreds of miles, away from its point of origin. Some crops like almonds (Prunus dulcis), apples (Pyrus malus), avocados (Persea americana), and sunflowers (Helianthuus annuus), spread their pollen on the feet and bodies of insects, such as bees. Their pollen distributions tend to be restricted to the range of insects. Still other plants, like peanuts (Arachis hypogaea), are self-pollinators, restricting the range even further. Regardless of these natural means of distribution, any of these kinds of pollen may be introduced to an archaeological context by humans. When interpreting pollen, phytolith, or starch grain data, it becomes ­necessary to consider human and nonhuman activities that might be responsible for their presence. Geib and Smith (2008) designed hands-on experiments to test the relationship between processing seeds and pollen deposition. They found, contrary to traditional practices of interpreting archaeological pollen washes according to how pollen is transported and deposited in natural settings, there exists a dynamic association of pollen ecology, seed architecture, and human behavior that not only warrants additional investigations but also can be expanded to include other classes of micro remains. Pearsall (2000:349) and others (e.g., Davis 1994; Faegri et al. 1989) also recognize that working with pollen, phytoliths, and starch grains is ­complex, and they echo Geib and Smith’s (2008:2100) call to unravel “the complexities of how human behavior creates pollen assemblages and how natural pollen rain and post-depositional processes distort and transform the pollen record.”

2.1 Potential Sources of Biases in the Paleoethnobotanical Record

Plant use, discard patterns, pedoturbation, recovery techniques, and a host of other processes have distorting effects on the paleoethnobotanical record. These processes may mask or exaggerate the patterns in plant resource exploitation or even suggest change where none occurred. Over the past decade, I have examined a host of biases, including the carbonization process, measuring of samples, and differential recovery associated with the flotation process, and have argued that to understand the human behaviors associated with any particular assemblage (Wright 1998, 2003, 2005, 2008), we must first understand the taphonomic history of the surviving remains.

In the following pages, I review my work and that of others who have attempted to test our assumptions and to contribute to the growing body of data involving the formation of the paleoethnobotanical record. This information is broken into subsections, according to the stage in which the transformation occurs. This scheme is roughly based on the work of Schiffer (1987). Initially, a plant resource is chosen for exploitation, acquired, possibly processed, then used, consumed, or discarded. These kinds of decisions and activities occur within the cultural realm of the people using the plant and are herein referred to as “cultural transformations.” It should be acknowledged that the where, when, and how plant resource enter the record is dependent on the specifics of each succeeding decision. If the plant resource (e.g., nutmeats) or its byproduct (e.g., nutshell) survives to be discarded, abandoned or lost and ultimately becomes part of an archaeological deposit, then a host of natural taphonomic factors determine whether the plant specimen will survive and, if it does, in what condition (e.g., eroded beyond recognition, fragmented, or perfectly intact). Analytical processes (or our decisions about how to sample a site, process the ­samples, quantify the remains, and so on) are the final determinants of how a plant ­specimen is included in the record, and, ultimately, how it is interpreted.

2.1.1 Cultural Transformations: Collecting, Processing, and Disposal of Plant Remains

Researchers have designed and implemented experiments to evaluate potential sources of cultural biases, such as harvesting techniques, processing, use, and disposal practices. For example, Munson (1984) edited a volume of papers devoted to the experiments on acquiring and processing plants associated with archaeologically known cultures of the Eastern Woodlands of the United States. Some of the researchers took to the field with sickles or simply bare-handed. Their experiments show that the techniques used to collect plants can be quite diverse and, at the very least, are dependent upon the kind of resource being exploited and the level of available technology. On the other side of the world, Abbo et al. (2008) experimentally harvested several species of wild peas (Psium spp.) in Israel. Interestingly, they conclude that the “potential productivity of wild peas was not the only or even the major consideration for its domestication” (Abbo et al. 2008:922).

In a farming economy, plant foods may be processed for consumption and ­storage. Processing may involve several stages; for instance, cereals have to be threshed, winnowed, and cleaned to separate the grain from the chaff, straw, and weeds. From ethnoarchaeological and experimental observations, it is known that some of these activities leave characteristic residues. Hillman (1973, 1981, 1984), Jones (1984), Goette et al. (1990) and de Vartavan (1990) have examined how crop husbandry, harvesting, preparation, and/or storage influence the kinds of products or byproducts that may be found in the archaeological record. Hillman (1973:241) asks what “a particular set of plant remains represents in terms of human activities.” He goes on to establish sets of associations, including correlations between (1) the composition of an assemblage and a particular processing technique and (2) a specific context and a processing strategy. Jones (1984), on the other hand, borrows from Hillman and proposes a statistical means for discriminating amongst crop products and byproducts. These kinds of studies illustrate the value in assessing plant samples within the context of likely harvesting and processing strategies because differences in assemblage composition may represent the same crop at various stages of processing or modifications in the subsistence economy. Furthermore, Lopinot (1984:192) cautions that, if preservation of seeds (or other remains) depends on cooking accidents, changes in the preparation of seeds prior to consumption may affect seed preservation.

It is logical that microbotanical remains, such as pollen, can also be biased by human activity. Apparently, few, if any, experiments designed to test the interrelationships of cultural processes and assemblages of microbotanical remains have been conducted. In the early development of paleoethnobotany, researchers like Bohrer (1968) and Schoenwetter (1962) recognized that it is inappropriate to infer paleoenvironmental conditions based entirely on plant samples collected from levels within archaeological sites because of distortions resulting from human activities. Rather, they recommended interpretations on the analysis of pollen and macrobotanical remains from nearby dated sections of relatively undisturbed sediments in addition to ­sediments from archaeological sites. These kinds of ecological studies can be used to identify anthropogenic influences on vegetation (Pearsall 2000). However, ­associations between pollen (and phytolith and starch grains, for that matter) and other kinds of human activities (e.g., collecting, processing, storage, consumption, and disposal of plants) tend to be implicitly rather than explicitly stated and certainly could benefit from experimental testing.

2.1.2 Crossing the Threshold: The Archaeological Context

Plant remains shift from what Schiffer (1987:47) has termed the “systemic context” to the “archaeological context” as a result of discard, abandonment, or loss. The former implies items that have been discarded because they were deemed useless or unpleasant. The charred cleanings and ash of a hearth dumped into a trash pit is but one example. Abandonment, like discard, is a deliberate action. It implies ­giving up something out of disinterest. A scorched tuber may be abandoned in a cooking fire because it is considered unpalatable. Loss, such as a few tiny grains spilled during transport, is an unforeseen outcome of human behavior. Macrobotanical remains may also become incorporated in the archaeological context by nonhuman activities or byproducts of activities; for instance, Miller and Smart (1984) discuss the burning of dung as a mechanism for introducing charred seeds.

Van der Veen (2007) compares formation processes associated with desiccated and carbonized plant remains from Europe and North Africa, and adds information about water-logged assemblages. Interestingly, she points out that most desiccated plant assemblages are made of secondary refuse and discusses their incorporation into mixed deposits. Entry for carbonized assemblages is also discussed. In the need to compare “like with like,” van der Veen (2007:988) emphasizes the need to understand formation processes:

routine practices are ordered by socially perceived norms and [that] the discard of waste from such activities is thus socially and culturally structured. This means that the analysis and understanding of formation processes can bring to light changes in such routine practices and consequently changes in the way that social relationships were negotiated and reproduced.

Pollen may enter the archaeological record either as a result of human activity or incidentally as a result of the so-called pollen rain. The other kinds of microbotanical remains – phytoliths and starch grains – derive largely from on-site plant discard as a result of human activity (Piperno 2006).

2.1.3 Taphonomic Process Affecting Paleoethnobotanical Assemblages

Once macrobotanical remains are deposited in archaeological contexts, there are several ways that they may survive physical and chemical decomposition: carbonization, desiccation, quick-freezing, mineralization, water-logging, and preservation in coprolites (Minnis 1989). These processes inhibit the growth of decomposers like bacteria or saprophytic fungi, slow the rate of enzyme action, and/or lower the speed at which chemical reactions occur (Bryant 1989). Desiccation, ­quick-­freezing, and water-logging, in particular, are remarkable for the preserved tissue types, if not for the sheer abundance of the material. Carter (1972) found the tomb of Tutankhamen filled with desiccated food, linen clothing, and wooden objects; whereas the water-logged deposits of the Swiss-Lake dwellers have yielded cultivated grains, fleshy fruits, legumes, nuts, fiber, and timbers that afforded a breakthrough in establishing a tree-ring chronology for parts of northern Europe (Eckstein 1984). Unfortunately, such instances of preservation are rare. Rather, most plant resources are used in temperate and tropical environments where their products or byproducts will not survive natural processes unless they have undergone the physical and chemical changes associated with carbonization (Bryant 1989). However, some kinds of remains, like starch grains, rarely survive charring (Fritz 2005:808). Perhaps Wilson (1984:14) put it best when he stated that “carbonization is generally thought of as a means by which plant remains are preserved. It is more accurate and less misleading to consider it a process of partial destruction.”

For many years, it was assumed that the conversion to charred remains was a straight-forward process. However, as taphonomic and site formation studies grew in popularity, researchers (e.g., Braadbaart et al. 2007; Boardman and Jones 1990; King 1987; Lopinot 1984; Prior and Alvin 1983; Rossen and Olson 1985; Smith and Jones 1990; Wilson 1984), including myself (e.g., Wright 1998, 2003), began experimenting with the carbonization of macrobotanical remains. These studies show that carbonization is the conversion of organic substances into carbon or other kinds of residues. For example, Braadbaart and I (Braadbaart et al. 2007) used a mass spectrometer to record the chemical changes and a scanning electron microscope and dissecting microscope to record the morphological changes in sunflower (Helianthuus annuus) achenes as they are exposed to different temperatures. Moreover, the above investigations document differences in the circumstances surrounding thermal exposure that may influence the final product. Rarity or absence of a species and/or element may reflect its sensitivity to thermal exposure rather than a lack of use. A particular element accidentally subjected to a high temperature for a very short time may differ in distortion or degree of degradation than one exposed to a lower temperature for a longer time. Moisture and chemical content, as well as the microenvironment at the time of exposure, may also affect the ­outcome. Given that our interpretive abilities are often contingent upon charred macrobotanical remains, knowledge of such variables becomes paramount.

I conducted a small experiment to test the influence that fluctuations in ­moisture, temperature, and pH might have on the preservation of carbonized ­macrobotanical remains (Wright 1998). It is no surprise that fluctuations in moisture were the most detrimental. However, I was amazed to observe that pH appeared to have little effect. It could be that the year of exposure that I chose as an interval was too short. However, in casual conversations with biologists and chemists, I learned that the critical element is more likely the presence or absence of bacteria that feed on the macrobotanical remains, and that these bacteria vary dramatically depending on the chemical and physical characteristics of the archaeological matrix. Additional research is needed to test assumptions on ­macrobotanical remains (as well as microbotanical remains and chemical and molecular remains) and soil pH. This line of research would be especially useful to understand the different preservation trajectories of plant and animal remains and any future hopes of integrating those kinds of databases into more precise discussions of past subsistence behaviors.

Like macrobotanical remains, microbotanical remains are organic and can fall prey to chemical erosion, physical erosion, and destruction by a host of biological organisms (Bryant 1989; Pearsall 2000; Piperno 2006). At some point in our careers, we have heard that the outer walls of pollen, spores, and non-pollen palynomorphs tend to be highly resistant to deterioration. Yet a cautionary note is in order; these entities do not always preserve equally well in all types of archaeological deposits. Rowe and Kershaw (2008: Table 42.1) prepared a valuable table that summarizes not only the preservation characteristics of pollen and spores, cellular tissues, starch, phytoliths, diatoms, and chrysophytes, but also information about their deposition, applications, advantages and limitations. The authors point out that pollen and spores fare better in acidic and/or anaerobic environments like peat bogs and lake beds. Cave sediments are also suitable because of their humidity and constant temperatures. In contrast, sandy sediments or open sites that are exposed to weathering generally result in poor preservation. Pollen can also be found in mud bricks, vessels, tombs, mummy wrappings, the guts of preserved ­bodies, fossil feces, and many other contexts. For those interested in differential pollen preservation, studies by Sangster and Dale (1961, 1964) show that not only is the depositional environment critical to pollen preservation but that the pollen of some species is more durable than others. In the early 1990s, additional studies were conducted and published by the Campbells; Ian Campbell (1990) reports on experiments designed to test mechanical destruction of pollen grains, and Campbell and Campbell (1993) discuss pollen preservation in saline and desert sediments. King et al. (1975) discuss the preservation of pollen associated with copper artifacts, while Kelso et al. (2000) indicate that shells preserve prehistoric pollen from percolating rainwater, free oxygen, and aerobic fungi.

Starch grains may be the most vulnerable of the microbotanical remains. According to Rowe and Kershaw (2008), starch grains are reduced when left in open conditions but can be found in geological, archaeological, and museum ­contexts. Based on laboratory experiments, Korsanje (2003) provides more precise information about the preservation of starch grains. Specifically, she claims that they preserve best in semi-arid environments with sandy soils and a pH of 7. Haslam (2004) also reports on the differential decomposition of starch grains and its implications for analysis.

Phytoliths are considered the most durable of all plant remains (Rowe and Kershaw 2008). The name comes from the Greek words for “plant” and “stone,” which is literally what they are: microscopic silica bodies formed in living plants. These tiny remains can be found where other micro- or macrobotanical remains are ­commonly absent, including dry, alkaline, and anaerobic conditions.

Movement as a result of natural processes is also a concern. Sediments are dynamic layers; they can expand, contract, aggregate, and conflate. For instance, Edwards (1979) observed that a 1-cm thick pollen sample may represent as much as 25–30 years of deposition – more than enough time for forest clearance, ­agriculture, and regeneration to have occurred. Conflation can be a problem for the analysis of macrobotanical remains as well. In fact, paleoethnobotanical remains may move from their initial location of deposition dependent on a variety of ­“turbations,” including faunalturbation, floralturbation, cryoturbation, graviturbation, aquaturbation, and aeroturbation. These soil mixing phenomena can be peculiar to very limited environmental contexts and are poorly understood with respect to their archaeological implications. While they are often thought of as agents of transport that disrupt the spatial integrity of the remains, we must bear in mind that many of these processes also possess the ability to accelerate the attrition of plant remains both directly and indirectly. Yet we know little of their degenerative impact; such an understanding is dependent on future experiments and observations.

2.1.4 Biases of Our Own Making

The plant evidence that survives the ravages of natural processes is then subjected to biases of our making. Our choices of sampling or recovery strategies or how we choose to quantify remains further filters the record and, ultimately, challenge our interpretations. In the following discussions on collecting, extracting, processing, identifying, and quantifying plant remains, I include information about the various analytical processes associated with each.

3 Recovery Methods

The divisions of macrobotanical remains, microbotanical remains, and chemical and molecular evidence are based, in part, on methodological differences in ­recovery, processing, and identification. Consider the recovery of pollen from a metate found in a structure as compared to the technique used to retrieve carbonized plant tissues embedded in the floor of that structure. The former calls for a pollen wash to remove the grains from the metate surface (Bryant and Morris 1986; Pearsall 2000), while the recovery of the carbonized fragments may involve the flotation of the house fill (Pearsall 2000). Furthermore, portions of the floor may not yield botanical remains but may contain chemical residuum that can be ­indicated by pH or trace element assays (Bryant 1989). Books, such as Pearsall’s (2000) Paleoethnobotany or Piperno’s (2006) Phytoliths have become very ­important texts for methodology. They provide information on various collection, processing, and analytical techniques.

We all realize that it is impractical to collect and analyze every cubic centimeter of an archaeological site. Sampling in the field and in the laboratory can keep ­processing and analysis from reaching unmanageable proportions and still yield an assemblage that is representative of the total population of the remains at a site. However, decisions have to be made on how much and from what context(s) ­sediments will be collected. Is a pinch or column strategy more appropriate? What size of sample is appropriate? Should the sample size be standardized across the site? Too small a sample and rare plants may be missed; investigating obviously rich loci at the expense of areas that seem devoid of remains can negatively impact interpretations about subsistence strategies and site use. Of all the potential biases, sampling seems to be the most discussed (e.g., Brady 1989; Jones 1991a, b; King 1987; Lennstrom and Hastorf 1995; Lepofsky and Lertzman 2005; Pearsall 2000; Riley 2008; van der Veen 1984; Wright 1998). Jones (1991a, b) examines sampling strategies at intrasite, site, and regional levels. Authors such as Lennstrom and Hastorf (1995) and Pearsall (2000) discuss three kinds of sampling techniques ­typically employed for taking macrobotanical or flotation samples – composite/scatter, column, and point samples – and the impact of each on densities and assemblage composition. King (1987) writes about sample size and its influence on the diversity of assemblages and the probability of collecting rare or unusual plant remains. Riley (2008) discusses how the individual nature of coprolites can skew dietary data; by using cluster analysis to examine a large collection, Riley is able to answer questions related to seasonality, mobility, and resource acquisition. My own research presents information on biases associated with different techniques used to calculate sample size (e.g., bucket and in situ measurements) and the relevance of soil type (Wright 2003).

For pollen, samples may be extracted in long cores at wet sites or in lake beds, or as a series of separate samples from dry sites or structural remains. Pollen can also be collected from adobe bricks, vessels, mummy wrappings, the guts of preserved bodies, and paleofecal materials. Whatever the provenience, great care must be taken to avoid contamination. Pearsall (2000:270–280) provides a thorough discussion of how research questions should guide both the planning of a sampling strategy and the choice of techniques for taking the samples. She cites the work of Bohrer and Adams (1977) at Salmon Ruin as a guide to select samples for a large body of systematically collected samples; Fish et al. (1982) to illustrate a sampling strategy designed to recover information about subsistence and environmental changes; and Kelso and Good (1995) for a strategy that lends itself to questions about land use, human and domesticated animal diet, room and artifact function, and construction sequences.

Because phytolith analysts have borrowed sampling strategies and collection techniques from palynologists, there are a number of parallels in both approaches. Both Pearsall (2000) and Piperno (2006) discuss collection and sampling strategies for phytoliths. Piperno (2006:81–86) categorizes the strategies into two basic groups: sediment column sampling and horizontal sampling of sediments and artifacts. As with pollen samples, the former is used to establish broad trends across time, whereas the latter can be used to answer questions about the organization of space and social relations as well as site and artifact functions, subsistence practices, and technology. While Lennstrom and Hastorf’s (1995) and Pearsall’s (2000) discussions of composite/scatter, column, and point samples that I cited above are focused on macrobotanical remains, the information can also be applied to ­microbotanical remains like pollen, phytoliths, and starch grains. In addition to samples collected from archaeological contexts, both kinds of microbotanical remains require that offsite control samples be taken to better understand the deposits, to identify any mixing of sediments that may skew interpretations of the samples from archaeological contexts, and to understand pollen rain.

For macrobotanical remains, modern recovery techniques include hand ­collection, screening, and flotation. At the very least, all three techniques introduce a size bias that can create analytical and interpretive problems (Wright 1998). The majority of macrobotanical remains are recovered through the flotation of archaeological deposits. This specialized technique uses water or chemicals to free seeds and charcoal flecks along with other remains from their geological matrix. Several researchers (Hunter and Gassner 1998; Pearsall 2000; Wagner 1988) systematically tested the recovery rates of various flotation machines, and I (Wright 2003) looked at differential loss and recovery of carbonized macrobotanical remains. On the basis of their systematic study, Hunter and Gassner (1998) suggest that the Flote-Tech machine is a reliable mechanism for the processing of flotation samples, whereas Rossen (1999) suggests that the machine is expensive and over-rated, citing the potential of small remains collecting in the corners of the machine as a potential source of contamination. Wagner’s (1988) seminal study demonstrates the variation in recovery based on the kind of technique or machine used, with differences in the mesh sizes of the catch screens being among the main contingencies impacting recovery. Keeping the system a constant, I found that different kinds of plant remains (e.g., a chenopod [Chenopodium spp.] seed versus a maize kernel) will vary in their rates of recovery; some of the differences can be attributed to size while others are a function of the fragility of the specimen.

A number of techniques for extracting phytoliths can be found in the literature (e.g., Lentifer and Boyd 1998, 2000; Pearsall 2000; Piperno 1985; Powers and Gilbertson 1987; Zhao and Pearsall 1998). Significant works addressing the ­recovery of pollen include Davis (1994), Faegri et al. (1989), Gorham and Bryant (2001), and Lentifer and Boyd (2000). Korsanje (2003) and Pearsall (2000) provide information on the retrieval and preparation of starch samples while Loy (1994) ­discusses the removal of starch residues adhering to stone artifacts. Both Pearsall (2000) and Piperno (2006) detail techniques for extracting these microbotanical remains from soils, paleofeces, dental remains, and artifacts, and Coil et al. (2003) offer a means for extracting multiple kinds of microbotanical remains from a single sample.

The previously cited literature is focused on methods for the recovery and ­analysis of microbotanical remains from terrestrial settings or artifacts. Gorham and Bryant (2001) explore the formation of underwater sites and their potential to yield microscopic remains, and in turn recover information on cargoes; a ship’s food supply; plants used to make rope, baskets, and the like; and port locations. In addition, they provide information on how and where to sample, and the conservation of samples.

Unfortunately, systematic investigations on recovery biases for microbotanical remains are lacking or at least difficult to locate in the literature. A recent chapter by Perry (2007) does provide an interesting discussion of differential recovery of starch types from lithic and sediment samples.

4 Specimen Identification and Analytical Methods

Processing of samples is followed by the challenge of identifying and analyzing the remains. Overviews by Pearsall (2000), Piperno (2006), and Fritz (2005) provide information on the basics of laboratory analyses for macrobotanical and microbotanical remains, how to set up laboratories, the kinds of equipment that are needed, and how to prepare reference collections.

4.1 Identification of Paleoethnobotanical Remains

Identification of macrobotanical and microbotanical remains is accomplished by visually comparing archaeological specimens to known specimens. Access to comparative material is essential. Even with the best reference collection, identification depends on the type and quality of traces (e.g., how eroded or fragmented they may be) and on the abilities of the researcher to discern attributes that are diagnostic of particular families, genera, or species of plants. While ancient seeds and fruits can usually be identified to species despite changes in their shape caused by charring, water-logging, and the like, systematic experiments to understand errors associated with the identification of macro- and microbotanical remains are scarce. Leach (1998) and Lombard and Wadley (2007) have experimented with blind tests to highlight the difficulties of distinguishing between plant and animal residues on tools. While guides and comparative collections are helpful, personal experiences associated with processes like collecting, processing, and carbonization (e.g., Hillman 1984; Wright 2008) can enhance identification and interpretation.

4.2 Quantifying Paleoethnobotanical Samples

Excellent discussions concerning quantitative analysis of macrobotanical and microbotanical remains are presented by Fritz (2005), Hastorf (1999), Jones (1991a, b), Kadane (1998), Miller (1988), Pearsall (2000), and Popper (1988). Counts and/or weights are often used, but as Popper (1988) points out absolute counts and weights assume that those measures accurately reflect human–plant interrelationships. Rather, absolute measures are heavily influenced by factors, such as preservation and sampling. Various manipulations of counts and weights (e.g., conversion factors, diversity index, ranking, ratios, and ubiquity) help to standardize the remains but do not necessarily alleviate all the biases. Even the use of more sophisticated methods like multivariate statistical analyses does not preclude the conditioning of the assemblage by cultural, natural, or analytical processes that may ultimately influence interpretations. Some researchers, such as Jones (1987), have turned to statistics to identify the effects of these transforming processes.

4.2.1 Absolute Counts and Weights

One means to document plant remains is merely to cite their occurrences. While simplistic in approach, even a laundry list of species present can provide qualitative information about subsistence, domestication, trade, and seasonality of occupation. Frankly, there are those researchers (e.g., Dunnell 1980) who question the validity of reporting anything more than the presence of taxa. It has been suggested that if preservation of carbonized macrobotanical remains is primarily the result of accidental inclusion in a fire, quantification is useless. Yarnell (1982:3–4) has responded to the critics by stating that “the usefulness of quantification is not a function of how materials happen to be preserved, but rather it is a function of regularity of preservation…[t]he problems of interpreting plant food remains do not differ greatly in kind from problems of interpreting many other categories of archaeological remains.” It becomes necessary to consider which items to count. Does one count all maize embryos and kernels and include them in a calculation of percentages? What about fragments? Quantifying fragments is an issue that has confronted zooarchaeologists, and, in following their lead, Watson (1979) and Jones (1990) attempted to standardize the counts of charred cereal grains for comparing stages of processing by quantifying diagnostic attributes like glume bases or tops of rachis internodes. While he asserts that his calculations can provide some guide to understanding the effects of cereal processing, Jones (1990:93) acknowledges that some components are more likely to be preserved by charring than others. Once again, we return to the need for understanding the histories and characteristics of the remains that we study.

Generating interpretations based on absolute counts and weights assumes that those measures accurately reflect human–plant interrelationships (Popper 1988). Absolute measures are too heavily influenced by factors, such as preservation and sampling, to serve on their own as meaningful foundations for interpretations. Instead, various manipulations of counts and weights help to standardize the remains, but they do not alleviate all the biases.

4.2.2 Presence/Absence and Ubiquity Indices

Ubiquity or presence analysis describes the number of proveniences in which a plant resource is recovered (Popper 1988). In other words, if maize kernels were identified in 10 of 80 proveniences, they would be given a score of 12.5%; 80 ­proveniences dated to about AD 850, and additional kernels dated to about AD 1000 were identified in 65 of the 70 features for a score of 92.9%. The larger index is interpreted as a sign of greater use of maize with time. This application of ubiquity assumes that, if a plant resource is used often, its chances of occurring in more and varied contexts are enhanced. Minnis (1985:104), among other researchers, prefers ubiquity to ­absolute counts and weights, because it is potentially “more closely related to the degree of utilization.” He explains that “a change in the number of samples in which a taxon is present is an imprecise but useful measure of the relative change in the use of that resource” (Minnis 1985:106). Critical to the assumption is how a plant resource is used and whether increases or decreases are actually a function of the frequency of use or merely reflect a change in the way that resource was used.

While Popper (1988) employs Hubbard’s (1980) research at Çayönü to exemplify an error in analytical judgment (an inadequate sample for answering his question and assuming that each grouping represented the full range of plant use at Cayonu), I borrow the study to emphasize how transformation processes come into play. Hubbard (1980) initially grouped his samples chronologically. This arrangement depicted a shift in high scores from cereals to pulses. He then reanalyzed the data according to location. Another pattern emerged which suggested that the variations in scores more accurately reflected differences in the logistics of plant processing activities. These perceived differences could be a result of use or logistics, but could also be a result of changes in processing or disposal of the cereals and pulses through time or across space. Also differences in microenvironments might be at play. Hubbard’s (1980) arguments would have been ­strengthened if he had discussed other issues, such as differences in use, logistics, microenvironments, etc.

As with absolute counts and weights, the possibility of differential cultural, natural, and analytical transformation processes limit the usefulness of ubiquity. When comparing different plant resources, such biases quickly come to mind, but many of these same biases may be overlooked when comparisons of the same resource are used. Rather it is assumed that a resource will act and be treated the same across space and through time. Yet, changes in processing, use, disposal, and microenvironments may affect the reliability of the comparison for even a single resource type. This places the comparisons of the same resource on similar ground as comparisons among different resources with respect to error as a result of divergence in transformation processes.

Another potential source of error can result from scoring samples as ­independent. Popper explains (1988:61):

it may be difficult to fulfill the assumptions of independent samples and to insure that the data are appropriate for answering the research question.... Because the presence of a taxon in each analytical unit receives equal weight, mistakenly splitting one sample into two analytical units inflates the frequency scores of the taxa in those analytical units. This could happen if one inadvertently takes two samples from one archaeological deposit and treats them as independent analytical units; or one might intentionally take two samples but then inadvertently score them as independent samples instead of averaging or combining them in an appropriate fashion. Clearly, mistakes alter the frequency scores less significantly when a group contains many samples.

4.2.3 Ranking

Popper (1988:64) suggests ranking as more precise means for measuring plant ­frequencies than ubiquity. With ranking, absolute counts are translated into an ­ordinal scale. For each taxon, a “scale of abundance” is created. This is accomplished by selecting important post-depositional transformations and weighting the taxon according to the presumed influences. The subsequent rankings are perceived to have neutralized the biases, thereby providing a more accurate means of comparison.

Popper (1988:165) notes some of the flaws and benefits of such a process. She praises ranking because it allows for independent evaluations of taxa. Disadvantages include a need for good preservation and high counts of taxa per sample. Because post-depositional transformational processes associated with a pit feature may ­differ in degree and kind from those associated with a structure, ranking may be limited among those samples collected from similar contexts. Most importantly, “the subjective weighting of taxa frequencies to determine their scales of abundance increases the potential for introducing errors into the results” (Popper 1988:66; emphasis mine). Weighting is more likely to be based on assumptions than on systematically collected data regarding post-depositional transformation processes.

4.2.4 Ratios

Ratios are frequently used as a basis for paleoethnobotanical interpretations. In the literature, one encounters such measures as charcoal volume of sediment, nut:wood, seed:charcoal, or maize kernels:cob fragments. These ratios are proposed as a means to standardize the data and enable the paleoethnobotanist “to compare (1) samples of unequal size, (2) samples differing in circumstances of deposition or preservation, and (3) quantities of different categories of material that are equivalent in some respect” (Miller 1988:72).

Miller (1988) defines and critiques several basic types of ratios. These include densities, percentages, and comparisons. While the latter allows for the comparison of relative amounts of mutually exclusive items, densities and percentages are measures whereby the denominator is inclusive of the numerator. The count of a particular species of seed relative to the count of the identified seed assemblage, or the weight of a particular species of wood relative to the weight of the identified wood assemblage, falls under the rubrics of percentages. When calculating percentages, units of measures must be the same for the dominator as the numerator.

Densities are typically expressed as the amount of charred material relative to the amount of sediment. For example, grams of ≥2.0 mm charred materials:liters of sediment, or seed counts:liters of sediment may be documented. Here, the units of measure chosen for the denominator and the numerator may vary. The amount of nutshell:wood or seeds:wood exemplify frequently used comparisons. The units of measure are up to the discretion of the researcher and may include count, weight, volume, or some combination thereof.

Seed:wood and nutshell:wood ratios are popular comparisons. We frequently assume that wood charcoal represents ordinary, domestic fuel use (Miller 1988). We then put charcoal in the denominator to control the likelihood of preservation. However, shifts in wood species employed for fuel may affect the outcome, because different taxa are subject to different preservation trajectories (e.g., Lopinot 1984; Rossen and Olson 1985; Smart and Hoffman 1988).

Another source of error is context. In eastern North America, nutshell:wood ratios are used to gauge variability in nut use through time or across a geographical cline. To make a point, I analyzed some samples from Simpson Quarry and calculated one nut:wood ratio based on all the samples and then calculated a second nut:wood ratio based on the context of the samples (Wright 1998). The former produced a nutshell:wood ratio of 0.47:1 and the latter yielded a ratio of 0.32:1. Although the latter could be interpreted as indicating relatively less dependence on nuts when compared to coeval assemblages in the region, it actually reflects the abundance of carbonized wood from burned posts. Hence, sample context may have a significant bearing on the interpretation of nut:wood ratios or any other type of ratios for that matter.

When constructing ratios, we must critically assess our variables. The foregoing examples show how differential use and deposition may affect subsequent interpretations, and such factors should be given adequate consideration. For every instance of ratio use, Miller (1988:83) poses the following questions:

  1. 1.

    What will a particular density, proportion, or comparison measure in a given assemblage?

  2. 2.

    Are the variables chosen relevant to the question asked?

  3. 3.

    Are assumptions of the equivalence of use and preservability among taxa and deposits warranted?

Furthermore, assumptions, whether they are based on ecological, functional, or preservational grounds, that are associated with the ratio should be discussed. Such discussions better enable the reader to evaluate any interpretation based on the ratios.

Regardless of the quantification technique used, the context of the samples and any associated cultural, natural, and/or analytical biases should be assessed (Hubbard and Clapham 1992; Wright 1998). Hubbard and Clapham (1992) provide a three-class scheme for evaluating what they refer to as “archaeological integrity.” These classes range from “A” which consist of “(s)amples whose origins are quite unambiguous and capable of rigid definition” (Hubbard and Clapham 1992:118) to “C” which consists of “[s]amples whose archaeological context is not clearly understood” (1992:119). The authors provide examples and reflect upon various kinds of quantification techniques in their efforts to elucidate what practices are of interpretive value and what are wasted efforts.

5 Summary and Conclusions

Problem orientation and subsequent interpretations are project-specific. The interests of the paleoethnobotanist and the overall research goals, and theoretical ­perspectives structure how paleoethnobotanical remains will be collected, ­analyzed, and ultimately interpreted. It is widely recognized that plants serve dietary, ­medicinal, ritual, and technological functions.

The origin of agriculture has been of interest to the archaeologist since the ­discovery of the first archaeological plants. Issues, such as motivation, the effect on health, and how this shift is interrelated to other aspects of culture are continually raised. Ultimately, our interpretations are dependent upon assigning the archaeological material with certainty to a particular species or variety. This task is especially challenging when dealing with carbonized macrobotanical assemblages where morphological markers and measurements are commonly used in classifying the remains, i.e., designating a specimen as domesticated versus wild. Morphological markers and measurements are both subject to transformative processes. The reliability of measurements has been the subject of debate. Researchers, like King (1987), working with maize and me working with sunflowers (Braadbaart and Wright 2007; Wright 2003, 2008) caution about the reliance on measurements involving small samples without considering the contexts of the remains. For example, seeds from trash heaps may reflect small throw-aways, while seeds from storage context might represent the larger, more desirable specimens that were stored for food or for next year’s planting (emphasis mine).

Early research in paleoethnobotany tended to emphasize the analysis of macrobotanical remains of cultivated plants in arid and temporal zones, and thus provided evidence for the domestication and spread of farming communities in those regions. The extension of research into tropical areas dominated by tuber, root, or tree crops, encouraged the development of methods employing microbotanical remains like pollen, phytoliths, and starch grains. Pearsall (2007) clarifies some of the methodological challenges of trying to identify ancient agricultural practices based on the analysis of ancient plant remains and highlights the usefulness of the analysis of sediment cores for understanding people–plant relationships in the neotropics.

Weiss et al. (2008) employ intrasite spatial patterns of macro- and microbotanical remains to delineate activity areas on a brush hut floor at an Upper Paleolithic site in Israel. While they assume that “different taxa have roughly equal preservation rates,” which is questionable unless tested, they do consider numerous post-depositional processes that might result in the displacement of the remains (Weiss et al. 2008:2401). To take their interpretations a step further, they integrate data about the distribution of plant remains and lithic remains to discuss the gender-related use of space. In an earlier article, Weiss combines efforts with Kislev (Weiss and Kislev 2004) to gain insights into Iron Age economic activities by using weedy species as markers for locating wheat fields. Hu et al. (2007) analyzed pollen and found that the famous terracotta warriors and horses found in the Qin Shihuang Mausoleum were produced at different locations. We have moved beyond mere assessments of the economic significance of plants (Fritz 2005), but many questions remain unanswered, including those associated with the formation of the paleoethnobotanical record.

Certainly, combining different kinds of analyses can provide corroborative ­evidence. The chapters in this volume illustrate numerous instances where macro- and microbotanical remains and/or zooarchaeological remains have been combined to elucidate exciting interpretations about past human behaviors. Biological, ­paleoethnobotanical, and zooarchaeological approaches converge to document domestication and address basic questions about when, where, and from what progenitor population(s) a domesticate is combined. Smith (2001), for example, discusses how biologists have focused on genetic profile comparisons, paleoethnobotanists on morphological markers to distinguish between wild and domesticated forms of plants, and zooarchaeologists on changes in age and sex profiles that are interpreted to reflect human management of herd animals; together these approaches have yielded remarkable results and that future “regional scale and species specific research should provide richly diverse and productive avenues of inquiry for biologists and archaeologists alike for decades to come” (Smith 2001:1324, 1326). Iriarte (2007) acknowledges recent advances in the study of early domestication and crop dispersal while illustrating the potential of integrating soil studies to revitalize studies on later agricultural landscapes. It is not uncommon to rely on the analyses of one of more classes of microbotanical remains in regions of the world where macrobotanical preservation is poor (e.g., Chandler-Ezell et al. 2006; Dickau et al. 2007; Horrocks et al. 2004; Perry 2002; Piperno and Pearsall 1998; see also Dickau, this volume).

Our interpretations, whether based on one class of archaeological remains or a combination thereof, are only as strong or weak as our understandings of the formation of the record(s) that we interpret. Collins (1975) presents an explicit statement regarding the formation processes that were predicated upon the idea of statistical samples. Below I have modified his scheme to apply to paleoethnobotanical remains:

  1. 1.

    Not all human behaviors and values result in patterned plant remains;

  2. 2.

    Of those which do, not all will occur where this is an opportunity for inclusion in archaeological contexts;

  3. 3.

    Of those which are included, not all will be preserved;

  4. 4.

    Of those which are preserved, not all will be exposed to, or by, the archaeologist; and

  5. 5.

    Among the plant remains exposed to the archaeologist, not all will be perceived or properly identified.

These processes are analogous to a series of filters, continuously acting to change the context and to reduce the quantity of observed plant remains. The scheme is not limited to the analysis of archaeologically derived plant debris but is applicable to any class of archaeological artifacts and features.

Without doubt these processes are complicated because methods, techniques, and experiments that might work well for paleoethnobotany in one region often do not for another. Any endeavors to offer worldwide coverage necessarily results in a less than sharply focused critique as seen here. Consequently, I have tried to offer a broad range of examples based on regional investigations. As seen in these examples, understanding these transforming processes can bring changes in routine practices and contribute to the social, economic, and political archaeologies of regions and periods from which the remains are derived. In essence, whether looking at paleoethnobotanical evidence in isolation or attempting to integrate it with zooarchaeological or other kinds of archaeological remains, we cannot be certain about our interpretations but rather aim for the best approximations available.

Moreover, researchers (e.g., Hayashida 2005; Peacock and Schauwecker 2003; van der Leeuw and Redman 2002) suggest that in today’s changing patterns of university and government research, we need to reevaluate the methods and goals of archaeology to bring long-term perspectives to bear on contemporary issues. As paleoethnobotanists, we have accumulated a large body of empirical evidence documenting such changes as deforestation, spread of savannahs, increased rates of erosion, rearrangement of landscapes for agriculture, and resource depression and depletion. These avenues of investigation illuminate past human actions and their environmental consequences, and ultimately can contribute to today’s conservation and restoration efforts (Pearsall 2007). Accordingly, it becomes increasingly important to understand not only how historical processes have shaped the modern landscape but also how they have shaped the archaeological vestiges of past humans actions and values.