Identification Guide for Near Eastern Grass Seeds

Mark Nesbitt

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781315427072
• Edition: 1
• Topic: Social Sciences
• Subtopic: Archaeology
• Index: Social Sciences

Archaeobotanical studies constantly encounter the carbonized grains of grasses, cultivated and wild, but the vast diversity of wild species that are potentially present has made identification of archaeological material fraught with difficulties. This volume provides an invaluable tool for mastering these difficulties. Based on years of laboratory study of an extensive reference collection, this book gives expert guidance for the identification and interpretation of grass seeds, focusing on those species that occur in the Near East and Europe.

11
Introduction

Wild grasses form one of the largest groups of plants, both in terms of the number of species in any higher plant family and in terms of biomass and groundcover. Ethnographically well documented as foodstuffs among many hunter-gatherers and some agrarian societies, wild grasses are the dominant weeds of crops throughout the Old World. Archaeobotanists working on pre-agrarian seed¹ material from the Near East, however, have tended to focus on identification of the wild cereals: barley, rye and wheat. This reflects both the strong interest in the domestication of cereals that has driven much of the research into pre-agrarian subsistence regimes, and the difficulties that archaeobotanists experience in identifying other seeds belonging to such a large plant family.

Archaeological remains of grass seeds present special problems in identification. First, plant remains incorporated into archaeological sites may have been subjected to threshing and dehusking. Diagnostic features visible on the intact inflorescence or spikelet are likely to be lost. Secondly, the processes that lead to the preservation of ancient material often lead also to the loss of some diagnostic features. For example, plant remains are very often preserved by charring. The grain (caryopsis) may survive well but lighter elements such as husks or hairs are usually burnt off. This process of differential preservation of grain and chaff has been observed in charring experiments on cereals (Boardman and Jones 1990). Waterlogging can lead to remarkably good preservation of a wide range of plant remains, but generally removes both the endosperm and much of the pericarp of grass seeds. In contrast, preservation by desiccation, in arid areas such as the North African desert, can preserve spikelets intact, greatly simplifying identification. A good example of this is the grass remains from the Tadrart Acacus caves in Libya (Wasylikowa 1992). Preservation of the shape of seeds as cavities in pottery (where the original seed has decayed or been burnt away) can also lead to the survival of the shape of intact spikelets and florets.

As charring and (in areas with plenty of rain) waterlogging are by far the most prevalent means of preservation, it follows that most archaeological grass seeds will not retain enough parts of the inflorescence to allow the use of standard botanical handbooks for identification. Instead, caryopsis characters are most important, with husk characters useful where they survive. Archaeobotanists therefore have to develop their own identification criteria to deal with a part of the inflorescence that is usually neglected by taxonomists.

Coverage of this handbook is limited to identification criteria applicable to charred seeds. In general, the Near East is too dry for the survival of waterlogged plant remains and too wet for the survival of desiccated material. The criteria examined here, however, will also be useful on desiccated seeds and casts of impressions.

1.1 Previous work

There are only two identification guides to grasses comparable in scope to that attempted here. One is Udelgard Körber-Grohne’s guide to identification of waterlogged European grass caryopses (Körber-Grohne 1964). This remarkably thorough work is not directly applicable to charred seed remains because subtle bran and hilum characters used in that key survive better and are much more clearly visible in waterlogged material. Mordechai Kislev’s computerised identification guide to Near Eastern grass seeds is mainly based on measurements and is not yet available for use (Kislev et al. 1995, 1997). It will be a valuable complement to this handbook, particularly given its emphasis on measurements.

Caryopsis characters have been relatively little used by plant taxonomists, unsurprisingly in view of the wide range of other, more diagnostic features visible on the intact inflorescence. The main exception is embryo anatomy, which has proved valuable in classification at the subfamily level but is of little relevance to archaeobotanical identification. The work of Reeder (1957; 1961), repeated and confirmed by Kinges (1961), represents the only widely accepted application of caryopsis characters to classification. The four characters of embryo anatomy concerned are relatively unambiguous and the groups formed by their different combinations correlate 2well with other vegetative and inflorescence characters (Table 2).

Reviewers of the potential of caryopsis morphology for grass taxonomy are enthusiastic, but offer few pointers for how this could be put into practice (Filgueiras 1986; Sendulsky et al. 1987). Where caryopsis morphology has been studied, as in the Triticeae (Terrell and Peterson 1993) and South African Arundinoideae (Barker 1986, 1994), botanists have had difficulty in developing objective descriptive terms for characters relating to shape. Caryopses are rarely illustrated in Floras, with two notable exceptions: the Flora of Iraq (Bor 1968) and C. E. Hubbard’s (1984) guide to the British grasses. These drawings, however, are usually in only two views and are very small.

There is a very large literature from agricultural researchers interested in identifying grass weed contaminants of crops. These will typically be in the form of florets or spikelets, and identification guides invariably focus on easily visible characters of the husks that do not survive on archaeological material. Much of this literature is listed by Nesbitt and Greig (1989) and Jensen (1998). The CIBA-GEIGY volumes on grass weeds are both attractively illustrated guides to the weed grasses of the world and particularly useful for their good drawings of intact spikelets (Hafliger and Scholz 1980, 1981). There are no such guides specific to the Near Eastern grasses but where useful literature exists, as in some archaeobotanical reports, it is referred to under the relevant genus account.

1.2 Identification criteria

The identification criteria chosen for this handbook were determined by three factors: first, the great variability in the shape of grass caryopses between and within species in a genus; secondly, difficulties in creating a standardised description for caryopsis shapes; and thirdly, the large number of grass species and genera in the Near East. Even after careful selection of the species to be covered, the final tally of 122 genera and 324 species was clearly too large for the preparation of detailed descriptions or a key. The highest priority appeared to be to allow identification of an unknown seed to a genus or group of genera in a manner that both allowed candidate genera to be included, and irrelevant genera to be excluded.

An additional factor guiding my choice of characters was a wish to avoid, as far as possible, the use of measurements. These are highly susceptible to the effects of charring and their use in keys is also prone to leave users with specimens that lie on the boundary between two size classes. At the same time, the grass seeds show big differences in caryopsis length and width that would obviously be useful for identification purposes. With this in mind, I decided that these quantitative characters should be checked to see if they divided naturally into distinct groups with stepwise differences separating them. If, as it proved, such natural divisions do not exist, then arbitrary divisions would be needed, and an identification key devised in such a way as to accommodate taxa close to the dividing points and which might fall on either side. The format of the identification key is a combined analytical-polyclave key as proposed by Westfall et al. (1986). This can be used as a standard analytical key but has the advantage that for those taxa with distinctive, unusual characteristics, the analytical key can be bypassed.

As a result, the main identification key relies as far as possible on clearly defined and visible characters, and offers identification to groups of genera. Beyond this point, access to a seed reference collection will be required to finalise the identification. For those genera best represented in archaeobotanical material, I have provided guidance on whether further identification beyond the genus level is possible.

In the grasses, there is a good correlation between caryopsis morphology and evolutionary relationships at all taxonomic levels from subfamily to section; some understanding of how different grasses are related to each other will therefore lead both to enhanced quality of identifications and to a better appreciation of how an identification that is not to species or genus level can still be useful. I have therefore included brief notes on ecology and evolution in the genus accounts.

1.2.1 Taxonomic system

Genera graminum (Clayton and Renvoize 1986) is widely accepted as an up-to-date and functional classification of grasses at the levels of genus, tribe and subfamily, based on a secure understanding of grass phylogeny. Although based primarily on morphological and anatomical criteria, recent work on DNA has confirmed most of the posited relationships. I have therefore chosen it as the taxonomic basis for this work. It is fortunate that the Near Eastern grass Flora has been the subject of a number of recent monographic Floras: the Flora of Iraq (Bor 1968); Flora Palaestina (Feinbrun-Dothan 1986), covering the areas on either side of the Jordan valley but excluding eastern Jordan; Flora Iranica (Bor 1970), covering Iran, Afghanistan and neighbouring mountainous areas of Iraq and Turkmenistan; and the Flora of Turkey (Davis 1985). The major gap in up-to-date authoritative coverage is east Jordan and Syria (Mouterde 1966). All four major Floras differ in arrangement and in the treatment of some taxa, and a synthesis of their contents would be a complex task in view of the wide range of synonyms in use. Such a critical synthesis has been achieved for the Near Eastern Flora, including the grasses, in the Conspectus Florae Orientalis (hereafter referred to as Conspectus) produced from the Hebrew University Herbarium (Heller and Heyn 1991).

The Conspectus has adopted the subfamily, tribe and subtribe system of Genera graminum and I have followed this almost exactly. I have numbered the genera (whether or not included in this handbook), in the order listed in the Conspectus, from 1 to 201, with one exception: Amblyopyrum and Aegilops. These are united as Aegilops in Genera graminum, but kept separate in 3many Floras and monographs, including the most recent monograph of Aegilops (van Slageren 1994: 81-2), on a range of evolutionary and morphological grounds. I have compromised by treating Amblyopyrum as a separate genus, but numbering it as genus 94a (with Aegilops as 94) in the Conspectus series. My genus numbering system will therefore still be consistent with others based on the Conspectus (see Appendix II for all genera).

The Conspectus lists species within genera in alphabetical order of species name. While this is an easy arrangement – and understandable in view of the need to keep the project a manageable size – any sense of evolutionary relationships (and thus morphological similarities) of the species is lost. Wherever I have been able to locate a comprehensive subdivision of a genus to sections, I have applied it. In some cases I have defined informal groups where a suitable sectional classification does not exist.

Figure 1 Countries of the Near East.

The nomenclature of the wild relatives of crops has presented particular problems. Many of these were first identified, and named as species, by Russian botanists in the 1920s and 30s. Feral derivatives and wild ancestors of crops would today usually be named as varieties or subspecies within the same species as the crop. Opinions vary widely on the appropriate rank, however, and the resulting nomenclature is currently both unstable and often unwieldy. In this guide I have followed the Conspectus in maintaining most of these wild relatives as species, while recognising that a lower rank may be more appropriate in phylogenetic terms.

41.2.2 Selection of species for study

Overall, the Conspectus contains 201 genera, totalling 981 species or 1085 taxa if subspecies and varieties are taken into account (Table 1). This represents a significant proportion of the total 651 grass genera and c. 10,000 grass species in the world, and is far too large a number for this handbook. Three criteria have been used to narrow the range: native status, geography and abundance.

Native status

A relatively small number of grasses have been introduced to the Near East from the Americas and the tropics; overall 37 species were excluded because they represented post-Columbian introductions of crops or wild plants. A further 33 species were excluded solely because they are crops. While all these are of pre-Columbian origin and are likely to occur in many archaeobotanical assemblages, they have been excluded for two reasons. First, as domesticated plants they tend to have much larger and differently shaped caryopses to their wild relatives, and would therefore increase the difficulty of constructing an identification key. Secondly, again as a result of selection processes during farming, cereal grains are highly variable in appearance, both within population and between different regions. It is typical that identification criteria that work in one area may not work in another (Hillman et al. 1996 for 1995). Crops are listed in the genus accounts, however, and relevant references to identification are cited in the bibliographies.

Table 1 Selection of grass species from the Conspectus Florae Orientalis included in the project taxa. ‘Total’ species are all those listed in the Conspectus; ‘eligible’ species are those not rejected as being crops, introduced, or growing only in Arabia and/or Egypt. The eligible species were then assessed as being relatively important (rated as 1 or 2 in the genus accounts) or as unimportant (‘rare’). The 324 species (with their concomitant subspecies and varieties) are the target species for this guide, although it has not proved possible to study all of these.

Geography

The Conspectus covers the whole of the Near East (Figure 1), including Arabia and Egypt. Archaeobotanical work in the Near East is concentrated within the ‘fertile crescent’ of mountains curving from the Levant through Turkey eastwards to Iraq and Iran, and this region makes an obvious focus for this key. Not only is this where work is concentrated, but it is also the region in which major developments such as the origins of agriculture and evolution of complex societies are known to have taken place first. There are therefore good archaeological reasons for concentrating on the fertile crescent and adjacent areas. Overall, 191 species occur only in Arabia, 12 only in Egypt, and 33 in both Arabia and Egypt but not elsewhere. All these were entirely excluded from the handbook. Native grasses growing in Israel, Palestine and the West Bank, Jordan, Lebanon, Syria, Turkey, Iraq and Iran were considered eligible for inclusion.

Abundance

All 675 species native to the core geographical region were assessed on a three-point scale:

1. Common over a relatively large region of one or more countries.
2. 5Less common, but present in a number of regions, or common in a narrower region or habitat.
3. Very restricted or scattered distribution, or restricted to high mountain pastures (1000+ metres) and thus of limited relevance to most archaeobotanical investigations.

Categories 1 and 3 were relatively easy to assess by combining distribution information in the Conspectus and major Floras. The division between category 1 and 2 is more subjective and largely driven by the need to reduce the number of taxa to be studied. Of the native species, 324 could be classified as ‘common’ (grades 1 and 2), and 351 as ‘rare’....