Introduction

A number of cereal genera can be divided into two forms, according to how tightly the grain (caryopsis) is enclosed within the surrounding chaff. In each case, the ancestral form is ‘hulled’; the grain is tightly enclosed offering some protection against predation and disease. The ‘naked’ form is derived through human selection, either for ease of post-harvest processing or culinary purposes; the ripe grain is easily removed from the enclosing chaff. In the case of domesticated barley, hulled and naked forms are well known, and display an east–west trend across Eurasia, with hulled barleys predominating in the west, and naked barleys more commonly encountered in the east. The archaeological record of barley grains suggest that this may not always have been the case, and that the distribution of the two forms has changed significantly through time. In this review we look at the distribution of hulled and naked barleys during prehistory and discuss the changes over time and the possible reasons for these changes.

Hordeum vulgare ssp. vulgare L. (barley) is among a group of plants that were domesticated around 10,500 years ago in southwest Asia. Today it is the fourth most important cereal crop in the world behind wheat, rice and maize. Genetic studies have shown that there are three key traits associated with barley domestication: the selection for non-brittle rachis (controlled by the Btr locus), six-rowed spike (Vrs locus) and naked caryopsis (nud locus) (Salamini et al. 2002). H. vulgare ssp. spontaneum (wild barley) has a brittle rachis, and is two-rowed and hulled.

Archaeological evidence from southwest Asia shows that wild barley was being used for at least 10,000 years before domestication (Weiss et al. 2004; Hillman et al. 2001). Non-brittle two-rowed barley has been identified at Tell Abu Hureyra in the early 9th millennium b.c. (Hillman 1975) and Tell Aswad in the 8th millennium b.c. (Van Zeist and Bakker-Heeres 1985). Non-brittle six-rowed barley has been identified at Ali Kosh, Iran, in the 8th millennium b.c. The earliest record of naked barley in southwest Asia, mostly from six-rowed forms, is at Ali Kosh from the 7th millennium b.c. (Helbæk 1969).

Barley was one of the principal crops to subsequently spread from southwest Asia into Europe, Africa and Asia. The first evidence of domesticated barley in Europe is in the Aegean at the beginning of the seventh millennium b.c. (cf. Warren et al. 1968). Hulled two and six-rowed, and naked forms appear at several sites in the Caucasus region in the 5th millennium b.c. (Zohary and Hopf 2000). By 6000 b.c. barley is recorded in western Central Asia in southern Turkmenistan; both hulled and naked six-row barleys have been recovered from Jeitun (Charles and Bogaard 2010). In south Asia, the earliest evidence of domesticated barley is from Mehrgarh, Pakistan, where naked grains predominate (Costantini 1984). Barley is widespread on south Asian archaeological sites of the 3rd and 2nd millennia b.c. (Fuller et al. 2004), including sites in the Ganges Basin, Chalcolithic Maharashtra and the greater Indus valley. In southern India, in Sanganakallu, both hulled and naked barley have been found during this period. East Asian records of barley do not typically indicate whether the grains are naked or hulled. b.c. During the 3rd and 2nd millennia b.c., barley appears at two sites in north China, Donghuishan (Flad et al. 2010) and Zhaojiazhuang (Jin 2007). Recent finds in Xishanping, northwest China, provide the earliest definitive dates for barley in China, between 2600 and 2300 b.c. (Li et al. 2009). One rare, definitive identification of naked barley comes from ancient Changguo Gou in Tibet, which is dated to approx. 3,500 years ago (Fu et al. 2000). During the 3rd millennium b.c., Crawford and Lee (2003) record barley in Korea, and Matsui and Kanehara (2006) record the crop in middle Jomon Japan. Neither source reports whether the barley is hulled or naked.

Both hulled and naked forms of barley have agronomic value and are used for different purposes. Most barleys under cultivation today are of the hulled variety and are mainly used for animal feed and the production of malt for brewing. The husk protects the grain from biotic and abiotic stresses and the embryo from damage during harvesting, and provides a filtration medium to separate out soluble extracts formed from the malting process (Taketa et al. 2004, 2008). Naked barley is mainly used as a human food source because of its ease of processing and edibility; hulled barleys require extensive pearling to remove the inedible husk. Selection for the naked character is therefore likely to have been contingent upon cultural variations across time and space in the choice of human food sources. Such variations need bear no direct relationship to prior patterns of the origin and spread of the domesticated crop (Staudt 1961).

According to a survey by Takahashi and Yamamoto (1950), naked barley is currently distributed widely, but its frequency greatly differs in different parts of the world. Naked types account for more than 95 % of domesticated barley in the high altitude areas of Nepal and Tibet, and at similar altitudes in Pakistan and India. Almost 50 % of barleys are naked in China, Korea and Japan. The frequency decreases toward the west, becoming low in Europe. Saisho and Purugganan (2007) also studied the current distribution of Old World hulled and naked barley landraces and showed that naked barleys are found at only very low frequencies in the Near East and European/North African landraces, while at moderate to high frequencies (10–77 %) in the rest of Asia. This skewed distribution of naked barley toward east Asia has probably been established by both natural and human factors. The naked caryopsis character is easily recognized and it can thus become a subject of strong human selection. The regions where naked barley is grown at high frequencies is where it is an important staple food for humans (Saisho and Purugganan 2007). Until recent times, naked barleys were also prevalent in Ethiopia; Ciferri (1944) and Orlov (1929) found that naked types accounted for almost 40 % of cultivated barley in Ethiopia. At higher altitudes they were an important human staple (Asfaw 2000).

Genetics and archaeobotanical recognition

The naked trait in barley is controlled at the single genetic nud locus. Taketa et al. (2008) cloned and sequenced the nud gene and determined that it encoded for a transcription factor involved in the regulation of a lipid biosynthesis pathway. In naked barleys a 17 kb fragment of chromosome 7 has been deleted, which includes the entire nud gene.

The work of Taketa et al. (2004, 2008) on the nud locus suggests a monophyletic origin of naked barley, probably in southwest Iran, which may be the result of second domestication event from a wild barley ancestor. A single fortuitous mutation event led to the deletion of the entire nud gene and the resulting phenotype was selected for by farmers and spread world-wide (Taketa et al. 2008).

The nud mutation causes a lack of lipid to be produced, which normally blocks adhesion of the husk to the caryopsis. Harlan (1920) had previously observed a sticky adhesive substance appearing on the surface of hulled barley grain about 10 days after flowering, which presumably corresponds to this cementing lipid. A consequence of this adhesion is that the dorsal surface of the grain forms a cast of the angular inner surface of the lemma. The two parallel dorsal ridges so formed are recognizable even if the lemma is later removed, and thus constitute the principal means of recognizing the hulled trait in archaeological macrofossils. Naked barley grains by contrast lack that ridging and have a rounded cross section.

Since the formation of the International Workgroup for Palaeo-ethnobotany (IWGP) in 1968, the archaeological records of grains have been proliferating, and increasing use made of the above phenotypic trait to distinguish naked and hulled grains. IWGP records have been collated sequentially by Jürgen Schultze-Motel, and Helmut Kroll (details below). In this paper we seek to compile and analyse archaeobotanical literature surveys into one condensed publication in order to elucidate long term changes in the ratio of hulled to naked barley under cultivation in the Near East and Europe.

Methods

Since its formation, the research records of the IWGP have appeared in (a) paper publication and subsequently (b) Helmut Kroll’s database.

  1. (a)

    Paper publication. Jürgen Schultze-Motel’s surveys entitled “Literatur über archäologische Kulturpflanzenreste” (“Literature on archaeological remains of cultivated plants”) were published in Die Kulturpflanze (continued as Genetic Resources and Crop Evolution) and covered references from 1965 to 1992. The series then continued in Vegetation History and Archaeobotany. Kroll took over from Schultz-Motel in the 1992/1993 review and continued until the 1999/2000 review.

    Although the distinction between hulled and naked barley is mentioned in Schultze-Motel (1968), all other reviews by this author fail to make this distinction. From the 1994/95 review onwards Kroll separates out hulled and naked barley. For this reason, we confined our survey to post-1994 records, and Kroll’s collected data (Kroll 1996, 1997, 1998, 1999, 2000, 2001). We have relied on Kroll’s record and have not gone back to the primary publications.

  2. (b)

    Kroll database (http://www.archaeobotany.de/database.html)

Kroll’s online database contains archaeobotanical data from 1981 to 2004, for barley and naked barley. Most data is from 1999 onwards. From early as 1986 naked barley (ssp. ‘nudum’) is specified, and we have consequently used all of the data available for H. vulgare in Kroll’s database.

Time periods

While a few absolute dates are given, the records in the above surveys are generally organized according to culture-historic episodes, which we have of necessity followed. We have subdivided the archaeobotanical records into the following chronological phases:

  1. I.

    Neolithic

  2. II.

    Chalcolithic/Bronze Age

  3. III.

    Iron Age/Roman

  4. IV.

    Post-Roman

Where specific cultural episodes were specified, we allocated them to the appropriate phase above; where dates were specified, we determined which archaeological periods they would fit into in their appropriate part of the world.

The data from Kroll’s references and Kroll’s database were merged. Total numbers of references recording hulled barley, naked barley and both hulled and naked, were calculated for each country and for each of the four phases. Most of the data refer to individual archaeological sites, some to a group of sites, and some are references to a region. From here on, for simplicity in this review, we refer to numbers of ‘sites’, even though not all references are reporting on individual sites.

A number of pragmatic decisions needed to be taken to accommodate the patchiness and variable recording of evidence in the surveys. References which were ambiguous about their chronology were omitted. Reviews were included; this could have resulted in some doubling up of data, but should not unduly affect overall proportions. Where records were allocated to a region rather than a country, we attached that record to the country closest to the centre of the specified region, so any references stating “Near East” were combined with data for Syria, and references to “Central Europe” were added to the data for the Czech Republic. Data for Slovenia and Serbia were merged with those for Yugoslavia.

Although we were interested in the whole Eurasian record, the record in the above surveys becomes extremely patchy further east than longitude 40°E, largely reflecting the emphasis of the IWGP research community. We judged it wisest to confine this meta-survey to the more substantial data record available to the west of longitude 40°E.

Having made these adjustments, we constructed pie charts for each country to indicate the proportion of sites recording hulled, naked or both hulled and naked grains, with the size of the pie proportional to the numbers of sites. These pie charts were plotted onto maps for each time period.

In order to compare this archaeobotanical data with modern barley landraces available from germplasm repositories, we analysed the prevalence of naked and hulled barley landraces from the same regions held in the National Small Grains Collection, US Department of Agriculture. This data was also plotted on a map using pie graphs to represent numbers of accessions, though it was necessary to take the log of the total landraces, given the large numbers available from some countries.

To further present temporal trends as bar charts, we organized the data into two larger regions, Europe and the Near East (a small number of African records were excluded from this exercise). The European group included Albania, Andorra, Austria, Belgium, Bulgaria, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Great Britain, Greece, Hungary, Ireland, Italy, Luxemburg, Moldavia, The Netherlands, Norway, Poland, Romania, Slovakia, Spain, Sweden, Switzerland, Ukraine and Yugoslavia. The Near Eastern group included Egypt, Israel, Jordan, Syria and Turkey.

Results

The proportions of sites with hulled, naked, and sites with both hulled and naked grain are indicated in Fig. 1. For all time periods Germany records the largest number of sites, followed by Great Britain and France. The large number of reported sites in Germany may well reflect the large number of German archaeobotanists active in the IWGP. As the data available for western Europe increases during the Iron Age/Roman and post-Roman periods, the amount for the Near East declines; this probably reflects on a shifting focus in archaeobotanical research away from the origins of agriculture. The bias in the activity of archaeobotanical research in selected countries places limitations on comparisons. For example, comparisons between hulled and naked barley finds in western Europe and eastern Europe are constrained by the scarcity of data from the latter. It is also possible that the number of sites containing naked barley has been under-reported as not all archaeobotanists distinguish between hulled and naked grains.

Fig. 1
figure 1

Maps showing the number of sites recording the presence of hulled, naked or both hulled and naked barley in each country for each of four time periods: Neolithic; Chalcolithic/Bronze Age; Iron Age/Roman and post-Roman. The sizes of the pie charts are proportional to the numbers of sites. White sites with hulled barley; black sites with naked barley; grey sites with both hulled and naked barley

Full size image

Temporal trends

  1. I.

    Neolithic In Europe, between a quarter and a third of sites contain naked grains: Germany, Great Britain, France, Spain, Greece, Bulgaria; and in the Near East in Egypt, Syria and Turkey. Less than a quarter of sites have naked barley in Switzerland and Italy, while in more northerly areas of Europe (The Netherlands, Denmark, Norway, Sweden and Finland) between 50 and 100 % of sites report naked grains.

  2. II.

    Chalcolithic/Bronze Age For Europe, results overall are similar to those of the Neolithic, with northern Europe having at least 50 % of sites recording naked grains. Spain has around two-thirds recording naked grains and a slight decrease in sites with naked barley is seen in Great Britain and Germany. The Netherlands and Switzerland also show a decrease in the number of sites recording naked barley. In Turkey, Syria and Egypt, the number of sites with naked barley significantly decreases.

  3. III.

    Iron Age/Roman During this period the incidence of sites reporting naked barley significantly drops in all of Europe and remains at about the same level in the Chalcolithic/Bronze Age Near East, though the number of sites is fewer. Northern Europe still records a proportionally larger number of sites with naked grains compared to the rest of Europe, though again the number of sites is fewer than for western Europe.

  4. IV.

    Post-Roman The post-Roman period sees a significant reduction in sites with naked barley, a clearer reduction than is seen in the Iron Age/Roman period. Many countries show no sites with naked barley at all. In the Near East no sites recording naked barley are reported, though as noted above, the number of sites reported has decreased significantly by this period.

Data on the distribution of extant naked and hulled barley landraces was collected from the National Small Grains Collection (NSGC), US Department of Agriculture (Fig. 2). Most barley germplasm held in repositories such as the NSGC was collected during the first half of the 20th century. For some countries there are large numbers of landraces, whereas for others there are very few (for example, Turkey has 1,626 barley landraces and Denmark has two). We have therefore taken the log of total numbers of landraces in our analysis. Figure 2 shows that there are very few extant naked landraces from the Near East, and that there are significant numbers in some European countries, for example from Italy, one-fifth of landraces are naked, and one-third in Romania, though for the latter there are only six landraces in total. In most other countries in western, northern and eastern Europe, naked landraces represent only a small proportion of the total and in other countries such as Spain, Portugal, Greece and the United Kingdom, no naked landraces are available.

Fig. 2
figure 2

Map showing the numbers of extant hulled and naked barley landraces in various countries from the National Small Grains Collection, USDA. The radii of the pie slices are proportional to (Nh/Nt)log Nt and (Nn/Nt)log Nt, where Nh denotes the number of hulled landraces, Nn the number of naked, and Nt the total number of landraces. White hulled barley; black naked barley

Full size image

The regional trends are also illustrated in Fig. 3. All the data for Europe has been merged into one group and for the Near East into another. As in Fig. 1, it is apparent that sites with naked barley decrease over time. In Europe the proportion of sites recording naked barley is approximately the same for the Neolithic and Chalcolithic/Bronze Age periods, with 53 and 55 % of sites having naked barley, respectively. This drops to 12 % in the Iron/Age Roman period, and 4 % in the post-Roman period. For the Near East, the proportion of sites recording naked barley is 32 % for the Neolithic period, 8 % for the Chalcolithic/Bronze Age periods, 7 % for Iron/Age Roman period and 0 % for the post-Roman Period. As illustrated in Fig. 1, there are far fewer sites for the Near East compared to Europe, but on the whole naked barley is found in fewer sites in the Near East than in Europe for each time period. For the data from modern landraces from the NSGC, the proportion of hulled and naked landraces is very similar to that of the post-Roman period for both regions.

Fig. 3
figure 3

Changes in the proportions of the total number of sites recording hulled, naked or both hulled and naked barley in Europe and the Near East for each of four past time periods and modern. Neo, Neolithic; ChBA, Chalcolithic/Bronze Age; IARom, Iron Age/Roman; PostRom, post-Roman; Modern, Extant barleys from the NSGC

Full size image

Discussion

Our meta-study of Kroll’s archaeobotanical literature surveys has shown that naked and hulled barley were much more common in the Neolithic and Chalcolithic/Bronze Age periods in Europe and in the Near East Neolithic than in later periods, and that naked barley had almost disappeared from both regions by the post-Roman period. We can now turn to consider the possible reasons for a decline in naked barley in these regions.

Observing this decline in the specific context of southern Sweden, Grabowski (2011) suggested various possible causes. First, hulled barley has a greater environmental tolerance and responds better to the deteriorating environmental conditions of the sub-Atlantic environmental epoch. Second, it responds better to manuring. Third, it is a suitable animal feed, and fourth, it has a lower loss of grain when harvested using more forceful techniques such as using a metal sickle (Grabowski 2011).

Buxó i Capdevilla et al. (1997) also observe that in the northeast Mediterranean during the Chalcolithic and Bronze Age naked barley gave way to hulled barley and naked wheat, whilst in the southeast, naked barley acquired the widest distribution, especially during the Chalcolithic. These authors discuss a number of reasons why hulled barley may have taken over from naked barley. In the Neolithic all varieties were grown, and later the most economically viable ones took on greater importance in agriculture and food, while others disappeared or took on a more minor role. In addition, although both hulled and naked barley are more tolerant of environmental extremes than other cereals, naked barleys are more susceptible to insect attack and parasitic disease as they lack the protective hull enclosing the grain. However, one of the major reasons these authors propose is the rise in naked wheat as the primary cereal in the human diet, this being higher yielding and more suitable for making bread than barley and thus human dietary preference led to the decline of naked barley cultivation in prehistory.

Nesbitt (2005) argues that barley was once an important human food source, which was used for both porridge and bread. It was perhaps only in mediaeval times that barley’s role as a staple food decreased in most regions. It appears that the use of barley for human food initially declined in warmer areas such as classical Rome and survived longest in the colder northerly regions of Scandinavia. Hulled barley has higher yields, and hence once the need for barley as a food source diminished and it became primarily a source of animal feed and for malt production, it would have been advantageous for farmers to grow hulled rather than naked forms. Sallares (1991) also ties the decline of naked barley in ancient Greece as running parallel with the increasing cultivation of naked wheats.

In this review it has not been possible to compare the fluctuations of naked and hulled barley over time in other parts of the Old World. In Kroll’s reviews and database there is a small amount of data for India, Korea and Japan, and also parts of Central Asia, but these are only occasional inclusions in these literature reviews and reflect the lack of archaeobotanical research done in places other than Europe and the Near East. It also reflects that fact that Kroll’s database has not had any additions since 2004, and hence does not record the growing body of archaeobotanical literature from sites in China, where techniques of flotation have only been introduced into excavation in the last decade.

To consider whether the naked forms of barley evident from the archaeobotanical record in the west, are the same as those that predominate today in the east, we return to the genetic evidence. The large deletion in the naked barley genome that encompasses the whole of the nud gene is highly unlikely to have arisen more than once. Taketa et al. (2008) discuss other potential genetic loci in barley that cause the naked barley phenotype, but all of these except for nud have been shown to cause pleiotropic deleterious effects on other agronomic characters. Thus it is evident that the nud mutation was the sole target for selection of this phenotype within domesticated barley, either from hulled domesticated barley or from a distinct wild ancestor (Taketa et al. 2004). This is significant, as other genetic analyses of barley have indicated two or more domestications for the barley crop as a whole (Fuller et al. 2011; Morrell and Clegg 2007; Saisho and Purugganan 2007).

As we have shown, naked barley cultivation declined from the Chalcolithic/Bronze Age periods onwards in the Near East. Naked barley declined in Europe during the Iron Age and Roman periods and declined further in subsequent periods. The proportions of extant naked European and Near Eastern landraces that we see represented in germplasm collections today are similar to those of the post-Roman period. It remains to ask why naked barley has remained in use in the east for so long, given the relative disadvantages in growing it as compared to hulled barley. Barley as a crop is well adapted to marginal and stress-prone environments, such as those at high altitude and latitudes, and it is often considered as a risk aversion crop by poor farmers. Consequently, barley is the crop of choice in the high altitude areas of east Asia. Naked barleys are most commonly found at altitudes too high for bread wheat cultivation, and where naked barley is an important staple and has cultural and religious significance (Li et al. 2011). In addition, until recent times naked barley was also an important crop in higher altitude regions Ethiopia, where its cultivation had special cultural significance. The decline in the cultivation of naked barleys during the 20th century can be at least partially linked to male farmers considering these types to be poorer yielding and requiring of higher maintenance in the field; in contrast, in some areas their persistence can be linked to the women finding naked grains less labour intensive to prepare (Asfaw 2000).

We may furthermore situate east Asian naked barleys within a much broader and deep-seated wholegrain tradition in the east, discussed in detail by Fuller and Rowlands (2009). When the southwest Asian crops wheat and barley were adopted in the east, the whole-grain tradition of boiling millets and rice for human consumption was adopted for these novel crops. This would appear to have driven an early choice for bread wheat, rather than the hulled emmer wheat which was more common in the west in the 3rd and 2nd millennia b.c. It may also have favoured a choice of naked barley from the west, though the culinary context of naked barley in western prehistory remains an enigma.