Vegetation History and Archaeobotany

, Volume 21, Issue 2, pp 117–129

Crop introduction and accelerated island evolution: archaeobotanical evidence from ‘Ais Yiorkis and Pre-Pottery Neolithic Cyprus

Authors

  • Leilani Lucas
    • UCL Institute of Archaeology
  • Sue Colledge
    • UCL Institute of Archaeology
  • Alan Simmons
    • Department of AnthropologyUniversity of Nevada
    • UCL Institute of Archaeology
Original Article

DOI: 10.1007/s00334-011-0323-1

Cite this article as:
Lucas, L., Colledge, S., Simmons, A. et al. Veget Hist Archaeobot (2012) 21: 117. doi:10.1007/s00334-011-0323-1

Abstract

Charred plant remains from the Cypriot Pre-Pottery Neolithic site of Krittou Marottou ‘Ais Yiorkis, situated in the foothills of the Troödos Mountains and dated to ca. 7500 cal. b.c., demonstrate the early introduction of two-grained einkorn (Triticum monococcum sensu lato). Grain measurements of two-grained einkorn from ‘Ais Yiorkis are compared to those from Aceramic and early Neolithic sites elsewhere in Cyprus, in northern Syria and central Europe. The grains appear to be larger than domestic grains of a later date from the Levantine mainland. Recent work by Purugganan and Fuller (Evolution 65:171–183, 2011) demonstrates a slow evolutionary rate in increasing grain size relative to the rates of evolution in wild species subject to natural selection. When the measurements of two-grained einkorn wheat from ‘Ais Yiorkis are compared with these same allochronic data the results indicate an accelerated rate in attaining larger grain size on Cyprus than on the mainland. The possibility of a domestication ‘event’ or rapid fixation of larger grain size characteristic of domesticated cereal crops in the context of an initially small island population is suggested by the colonisation by farmers of Cyprus in the Cypro-Pre-Pottery Neolithic.

Keywords

ArchaeobotanyCyprusPre-Pottery NeolithicPlant domesticationTriticum monococcumEarly farming

Introduction

Previous evidence that the rates of the process of plant domestication were rapid have recently been re-evaluated in favour of a model that argues for a slower rate in the evolution of domestication traits (Allaby et al. 2008; Tanno and Willcox 2006). Based on experimental cultivation of wild wheats, Hillman and Davies (1992) estimated a rapid rate of possibly 20–200 years for the process of domestication (see also Ladizinsky 1998; Abbo et al. 2010). Conversely, recent work has quantified the rate of change in domestication traits from archaeobotanical evidence, and suggests a slow evolutionary rate in two traits that distinguish wild and domesticated cereal crops (Purugganan and Fuller 2011; Fuller et al. 2011, this volume). The morphological traits that are used to document domestication are an increase in grain size and the loss of natural seed dispersal with the dominance of non-shattering ears. The rates at which these two traits evolve differ and it has been suggested that increases in grain size, i.e., “semi-domestication”, were manifest before the evolution of non-shattering rachis, i.e. “full domestication” sensu stricto (Fuller 2007). The protracted model demonstrates somewhat faster rates calculated for non-shattering ears than for grain size, although an increase in grain size occurs before a dominance of non-shattering ears (Fuller et al. 2011, this volume). The evolutionary rates are generally lower or comparable to rates of phenotypic change that have been measured in modern wild species subject to natural selection, such as plants adapting to soil heavy metals, ozone resistance or soil pH.

Recent evidence demonstrates the colonisation of Cyprus by farmers at about 8500 cal. b.c. Archaeobotanical data, namely measurements of cereals and chaff from the earliest levels at Kissonerga Mylouthkia, provide evidence of agriculture based on domesticated cereal crops in the Cypro-Early Pre-Pottery Neolithic Phase B (Colledge 2004; Murray 2003; Peltenburg et al. 2000; Peltenburg 2003; Zohary and Hopf 2000). Given this early date for the introduction of cereal crop agriculture, it has been suggested that the initial introduction of crops was from a mainland population and was a mixture of morphologically wild and domesticated forms (Willcox 2003). Based on the domesticated charred cereal chaff from Mylouthkia (Murray 2003), Fuller (2008) has suggested the possibility of a local bottleneck effect that could have pushed domesticated forms to dominance where wild populations of wheat were absent. Further, it is proposed that selection for domestication traits (i.e., larger grain size and non-shattering ears) becomes stronger if there is no possibility for cross-hybridisation with wild populations, for example in instances where domestic crops are introduced to, and cultivated in, areas beyond the natural range of the wild progenitor species (Allaby et al. 2010). Allaby (2010, p. 4) states, “In effect, it may be the case that ‘domestication events’, if they could still be considered as such, were the result of human expansions,” rather than the beginnings of cultivation. In other words, the expansion of cultivation into virgin territory created the signals in phylogenetic datasets which separate modern cultivated lines from their wild progenitors, and have in past been taken to infer, misleadingly, the numbers and locations of start of cultivation.

Since Cyprus is thought to be the first region to be colonized after the emergence of cereal crops in the Near East and is outside of the natural distribution for all but one wild cereal progenitor, barley (Meikle 1985; Holmboe 1914; Christodoulou 1959; Murray 2003), it is an ideal region to test whether a local bottleneck effect on the island could have accelerated the progression towards a fully domesticated crop (Harlan et al. 1973; Fuller 2007). In this report we present new data on charred cereal remains from Krittou Marottou ‘Ais Yiorkis dated to ca. 7500 cal. b.c. (map, Fig. 1) including quantities of two-grained einkorn wheat. Grain measurements of two-grained einkorn are compared to those from sites in Cyprus, northern Syria and central Europe, and are considered alongside the allochronic data used in a recent study by Purugganan and Fuller (2011) in order to assess the extent to which the einkorn of Cyprus fits into the contemporary size range on the mainland and the chronological trend in size change.
https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig1_HTML.gif
Fig. 1

Map of Near East with a schematic representation of rainfall variation, showing location of ‘Ais Yiorkis and other sites referred to in this paper

Krittou Marottou-‘Ais Yiorkis

Krittou Marottou-‘Ais Yiorkis is a Cypro-Middle to Late Pre-Pottery Neolithic B (PPNB) site located in the foothills of the Troödos Mountains. The site overlooks the Ezousas River and the surrounding landscape includes Aleppo pine (Pinus halepensis), Kermes oak (Quercus coccifera) and wild olive (Olea europaea var. sylvestris) (Simmons 1998). It lies on two adjacent modern agricultural terraces at an elevation of ca. 460 m above sea level. ‘Ais Yiorkis is characterised by a series of pits and circular platform structures. The material culture recovered thus far includes imported obsidian bladelets, picrolite ornaments, carnelian bead fragments, stone vessels, ground stone for food processing and a rich chipped stone assemblage comprising nearly 200,000 pieces. Of particular significance is evidence for the presence of cattle, which demonstrates their introduction to the island at this early date and prior to a second wave of importation in the Bronze Age (Simmons 2005). The chronology for occupation is based on 23 radiocarbon dates on samples taken from animal bone remains (including cattle), charred einkorn grains and wood charcoal (Fig. 2). Two samples are grains of two-grained einkorn, one comes from a 160 l flotation sample from level 6 of Feature 4 with a radiocarbon date of 7600–7510 cal. b.c., and the second is from a 150 l flotation sample from the level 8 of Feature 4 with a date of 7590–7450 cal. b.c. The radiocarbon dates thus far place this upland occupation within the Middle Cypro-Pre-Pottery Neolithic B, ca. 7900–7500 cal. b.c., and the summed probability of all the dates narrows the main date range to 7600–7500 b.c.
https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig2_HTML.gif
Fig. 2

Selected radiocarbon dates from ‘Ais Yiorkis, including those from charred einkorn wheat and barley. The 1σ-range of the summed probability is 7590–7490 b.c. Calibrated with OxCal 3.10 (Bronk Ramsey 2005) using the IntCal09 calibration curve (Reimer et al. 2009)

Materials and methods

The plant remains, which were preserved by charring, were recovered by flotation using a 200 l flotation tank and captured on nested sieves of both a 1 mm and 250 micron mesh. A total of 2,163 l were floated from 43 samples from eight areas or units, including various levels of circular platform features and pits. Every unit produced samples in which there were identifiable charred plant remains. A majority of the plant taxa sampled come from Feature 4, which is an oval pit also containing large quantities of lithics and faunal remains. On the basis of the high quantity of lithics and faunal remains the pit is interpreted as having been used for the disposal of rubbish and similarly, the botanical material is thought to have been swept or dumped in the pit after accidental charring in a hearth or oven. The plant remains recovered from Feature 4 were in good condition and this is possibly due to the fact that pits and middens are often located away from habitation areas and where the contents are less subject to disturbances such as trampling that are likely to cause fragmentation or complete destruction of more fragile charred taxa (Colledge 2003). Flotation samples were sorted under a binocular microscope and identified with the aid of modern reference material held at the Institute of Archaeology, UCL.

The grains of two-grained einkorn can be easily distinguished from those of one-grained einkorn in lateral view by a flat ventral surface as opposed to a convex ventral surface characteristic of the one-grained variety. Both varieties are asymmetric in cross section but the grains of one-grained einkorn have a more strongly defined dorsal ridge. Also both varieties are strongly attenuated at the apical and embryo ends. A total of 117 whole grains of two-grained einkorn were identified; these grains were photographed and measured with Leica LAS EZ measurement software (ESM Table 1; Fig. 4a).

We have compared the measurements of 117 grains of two-grained einkorn from ‘Ais Yiorkis with those of specimens from the Near Eastern mainland and by so doing have been able to estimate the difference in rate of increase in grain size over time. All measurements are from carbonized grains, and although this is known to lead to shrinkage and may introduce some degree of additional variability to samples, we assume that carbonized assemblages across sites are still comparable, and that comparison may reveal some differences between past populations. Recently Purugganan and Fuller (2011) have provided quantitative estimates of phenotypic change in Near Eastern grain size for emmer, barley, einkorn, lentils and peas. The strongly directional nature of seed size increase across domesticates, with a focus of the most marked change only during the early millennia of cultivation (i.e. the Pre-Pottery Neolithic), suggests that these changes are mainly driven by evolution (selection) rather than phenotypic responses of plants to local growing conditions. Phenotypic responses are expected to be on a much smaller scale overall, although the lack of resources (water, nutrients) during grain-filling should cause size reduction in a given plant or field. The meta-analysis of Sadras (2007) indicates that grain size is more strongly heritable than it is environmentally determined, while recent genetic analyses have identified an allele that increases wheat grain breadth and thickness, which shows a strong signature of selection in domesticated as opposed to wild wheat (Gegas et al. 2010). Taken together, the evidence suggests a major role for selection and genetics underlying the larger temporal trend in grain size increase in early wheat.

In this study we calculate from einkorn grain breadth and thickness the haldane measure of phenotypic change (Gingerich 1993). The haldane is the change of trait average divided by the standard deviation of entire dataset, and as such this rate takes into account the range of variation within assemblages, as well as sample size, such that larger and more variable samples have a somewhat stronger influence on calculations. Evolutionary rate in haldanes is given by the equation: haldanes = [(x2 − x1)/sp]/[t2 − t1], where x1 and x2 are the mean trait values at time points t1 and t2, respectively, in generations and sp is the pooled standard deviation for the trait across the time points (Gingerich 1993). This has proved a useful metric in comparing evolution in various modern microevolutionary studies, which range from studies of change in body or organ size in birds, reptiles and fish (Hendry and Kinnison 1999; Kinnison and Hendry 2001; Hairston et al. 2005) to studies of plant height, leaf size or toxin tolerance levels (Bone and Farres 2001), as well as in palaeontological comparisons (Gingerich 2001; Roopnarine 2003). We fit least-squares linear models (regression line) of x/s (trait value per pooled standard deviation) values against time (in years since the start of domestication), and the slope of the regression provides a haldane estimate for the whole data series. Our mainland dataset is taken directly from Purugganan and Fuller (2011; Fuller et al. 2011, this volume). A separate regression line, providing a separate haldane rate estimate, has been calculated between the data of ‘Ais Yiorkis and an extrapolated mainland model at 8350 cal. b.c. This estimates the expected mean and standard deviation of the mainland einkorn at the time it was transported to Cyprus in the PPNB. The derivation point is bracketed by the mainland datasets of Dja’de and Çayönü. This derivation point is a close chronological representation for the suggested timing of the colonisation of Cyprus.

Results

Two of the three founder cereal crops of Near Eastern agriculture (Zohary and Hopf 2000) are present in the ‘Ais Yiorkis samples: einkorn (both one-grained and two-grained) and hulled barley (Hordeum vulgare sensu lato) (additional taxa shown in Table 1). Two-grained einkorn is present in 100% of units and 44.2% (n = 43) of the samples, with 276 total whole grains (Fig. 3). A single grain of one-grained einkorn and a total of 65 whole barley grains are present in one sample. There are very few cereal chaff remains, with a total of 17 einkorn glume bases recovered from one sample. The wood charcoal has yet to be analysed but there was little wood charcoal recovered from flotation. Of the 117 grains of two-grained einkorn wheat measured from the site, 115 grains were recovered from five samples from Feature 4. The remaining two grains were recovered from two different pits, Feature 11 and Feature 10. It is not possible, therefore, to compare grain sizes across features due to the limited number of grains measured from different areas of the site. As mentioned earlier, Feature 4 is interpreted as a pit used for the disposal of rubbish, with the botanical material either swept or dumped in the pit after accidental charring in a hearth or oven. In consideration of a lack of crop-processing residue associated with the cereal grains it is likely that the assemblage is representative of the by-product of a sieved or cleaned crop.
Table 1

Full list of taxa from samples with identifiable remains, ‘Ais Yiorkis

Sample number

  

28

32

37

43

46

48

51

53

56

57

58

65

Fature number

  

F4

F4

F4

F4w

F4w

F9w

F4

F11

F10

F13

F19

F13

Level number

  

6

7

8

6

7

2

9

2

2

2

1/2

Volume (l)

  

160

160

150

110

50

48

25

50

20

75

60

27

Context type

ub.

to.

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Cereals

 Triticum monococcum 2 g

44.2

328

119

76

19

17

10

5

1

3

1

 T. cf. monococcum 2 g

2.3

17

17

 T. monococcum 1 g

2.3

1

1

 T. cf. monococcum (gb)

2.3

17

17

 Hordeum sativum

11.6

65

45

14

4

1

1

 Cereal indet. (wge)

30.2

922

548

173

97

41

28

3

2

5

3

2

10

Pulses/oil plants

 Pisum/Vicia sp.

6.9

3

1

1

1

 cf. Pisum sp.

2.3

1

1

 Vicia/Lathyrus sp.

4.7

2

 cf. Vicia sp.

2.3

1

 Lens sp.

6.9

3

1

Trees and shrubs

 Pistacia sp.

4.7

2

1

1

 Olea sp.

2.3

1

Wild herbaceous taxa

 cf. Brassica/Sinapis

2.3

1

1

 cf. Malva sp.

2.3

1

 Leguminosae (large)

4.7

2

 Bolboschoenus cf. maritimus

2.3

1

1

 Avena sp.

6.9

8

5

2

1

 cf. Bromus sp.

2.3

1

 Lolium sp.

4.7

12

5

7

 Stipa sp.

2.3

1

1

Sample number

  

67

68

69

70

71

72

79

82

83

99

102

103

Fature number

  

F4

F4

F4

F4

F4

F4

F9

F9

F22

F9

Level number

  

4

5

12

6

6

7

10

1

3

2

1

2

Volume (l)

  

200

350

53

140

30

100

140

60

55

20

60

20

Context type

ub.

to.

Pit

Pit

Pit

Pit

Pit

Pit

Pit

Pit

b.e.

Pit

Pit

Pit

Cereals

 Triticum monococcum 2 g

44.2

328

3

4

1

3

2

1

2

59

1

1

 T. cf. monococcum 2 g

2.3

17

 T. monococcum 1 g

2.3

1

 T. cf. monococcum (gb)

2.3

17

 Hordeum sativum

11.6

65

 Cereal indet. (wge)

30.2

922

5

5

Pulses/oil plants

 Pisum/Vicia sp.

6.9

3

 cf. Pisum sp.

2.3

1

 Vicia/Lathyrus sp.

4.7

2

1

1

 cf. Vicia sp.

2.3

1

1

 Lens sp.

6.9

3

1

1

Trees and shrubs

 Pistacia sp.

4.7

2

 Olea sp.

2.3

1

1

Wild herbaceous taxa

 cf. Brassica/Sinapis

2.3

1

 cf. Malva sp.

2.3

1

1

 Leguminosae (large)

4.7

2

1

1

 Bolboschoenus cf. maritimus

2.3

1

 Avena sp.

6.9

8

 cf. Bromus sp.

2.3

1

1

 Lolium sp.

4.7

12

 Stipa sp.

2.3

1

wge whole grain equivalent, ub ubiquity, to total, gb glume bases, indet. cereal indeterminates, b.e. baked earth

https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig3_HTML.gif
Fig. 3

Photograph of an example of two-grained einkorn from ‘Ais Yiorkis

Figure 4a is a scatter-plot that shows the measurements of 117 grains of two-grained einkorn from ‘Ais Yiorkis (the average length, height, and breadth of the grains are 6.24, 2.16, and 2.44 mm, respectively) and Fig. 4b is a scatter-plot of the average breath and width of two-grained einkorn from mainland Middle Euphrates, Cypriot Ceramic Neolithic Vrysi, and Central European Bandkeramik sites (Table 2). Measurements of grains from ‘Ais Yiorkis are significantly larger than the grains from early and late Jerf el Ahmar, where a total of 87 specimens have been measured, and form a cluster with domesticated grains from Early Neolithic Bandkeramik sites in Central Europe and Ayios Epiktitos Vrysi (Kyllo 1982), the only other Cypriot site with published measurements of two-grained einkorn. Figure 4 illustrates that the grains from the northern Levant are narrower than grains from areas outside the range of the natural distribution of the wild progenitor, which includes northern Syria, southern Turkey, northern Iraq and Iran (Zohary and Hopf 2000). As can be seen in these data the ‘Ais Yiorkis two-grained einkorn is associated more closely with European early Neolithic grains, which post-date the Cypro-MPPNB by more than 2,000 years, than with approximately contemporaneous specimens from the Near East.
https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig4_HTML.gif
Fig. 4

Scatterplot of measurements. aAis Yiorkis (top), bAis Yiorkis metrical data compared with other PPN and early Neolithic sites. In b site averages are plotted for each site in Cyprus, mainland Levant and Central Europe. Sources Khirokitia, n = 19, n = 66 (Waines and Stanley-Price 1977; Miller 1984), Ayios Epiktitos Vrysi, n = 8 (Kyllo 1982), Central European Sites (AK114 (Harheim Frankfurt/LBK II, n = 11), AK123 (Nieder-Eschbach, Frankfurt/LBK II, n = 25), AK134 (Nieder-Mörlen, Bad Nauheim/LBK II and II/III, n = 3), AK66 (Mittelbuchen, Hanau/LBK II and IV/V, n = 4), AK41 (Niederhöchstadt, Eschborn, LBK IV, n = 100), AK84 (Hailer, Gelnhausen/LBK IV/V, n = 54), AK152 (Herxheim/LBK V, n = 2) and AK165 (Glauburg, Glauburg/Rössen, n = 2) (Kreuz and Boenke 2002), Jerf el Ahmar (JA), Dja’de (DJ), Kosak Shamali (KS) (Willcox 2004), Mureybet (M) (Van Zeist 1970)

Table 2

A comparison of measurements (mm) of two-grained einkorn from sites located in Cyprus, northern Syria and central Europe

Site

n

Breadth/width

Thickness/height

‘Ais Yiorkis

117

1.62 (2.64) 3.11/σ = 0.294

1.53 (2.05) 2.75/σ = 0.274

Vrysi

8

0.5 (1.8) 2.7

0.9 (1.9) 2.5

Khirokitia (Tholos I)

19

1.8 (2.0) 2.4

2.5 (2.8) 3.0

Khirokitia (Tholos II)

66

1.2 (1.9) 2.5

2.0 (2.7) 3.1

Jerf el Ahmar (Early)

27

0.72 (1.15)1.57/σ = 0.217

1.11 (1.33) 1.55/σ = 0.13

Jerf el Ahmar (Late)

60

0.6 (1.3) 2/σ = 0.3

0.85 (1.42) 2.28/σ = 0.32

Kosak Shamali

52

1.15 (1.64) 2.24/σ = 0.21

1.4 (1.8) 2.2/σ = 0.2

Dja’de

102

0.79 (1.42) 2.2/σ = 0.32

1.04 (1.55) 2.4/σ = 0.32

n number of grains measured. Measurements: minimum (average) maximum/standard deviation. (Willcox per. comm.; Kyllo 1982; Waines and Stanley-Price 1977) For central European LBK sites refer to Kreuz and Boenke (2002, p. 235). Waines and Stanley-Price (1977) do not specify whether the measurements are from the one-grained or two-grained variety of einkorn wheat. The standard deviation for Vrysi and Khirokitia could not be calculated

Figure 5 is a plot of einkorn metrical data from Near Eastern sites, including ‘Ais Yiorkis, against a time scale. Sites are represented by the mean and standard deviation of measured grain assemblages (for references to the published reports in which these data are recorded see Purugganan and Fuller 2011; Fuller et al. 2011, this volume). These are plotted against time on the basis of a modal/median estimate of site age in calibrated years b.c. based on a recent compilation and recalibration of radiocarbon dates from each site (following Purugganan and Fuller 2011). The measurements of breadth and thickness for the ‘Ais Yiorkis grains are well above the trend-line and thus the expected average value for its age. This, we propose, indicates that ‘Ais Yiorkis two-grained einkorn is the product of an accelerated rate of grain size increase compared to the evolutionary trend on the mainland. Grain size increase occurred sooner in the island population than in the mainland populations. On the basis of present evidence for sites in the comparative regions that post-date ‘Ais Yiorkis there is no evidence for continued increase in grain size. The grains from mainland populations do not compare to the grains from the Cypriot populations until or even after the Late Pre-Pottery Neolithic B, i.e. ca. 6000 cal. b.c. This raises the possibility that the Bandkeramik two-grained einkorn also attained larger grain size because it was isolated from the wild populations in the Near East.
https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig5_HTML.gif
Fig. 5

Graph of average and standard deviation of einkorn grains (Breadth top, height (thickness) below, in mm) plotted against a point estimate of age. The linear trend line through the mainland populations (diamonds) is indicated, while the inferred evolutionary trend of Cyprus is indicated by the arrow. Correlation coefficient (R2) is shown for mainland site averages against time. Sites labelled by abbreviations: Qm Qaramel, M Mureybet, J1 Jerf el Ahmar early levels, J2 Jerf el Ahmar late levels, Dj Dja’de, Cy Çayönü, WJ7 Wadi Jilat 7, Rm Ramad, Eb Erbaba, Hy Höyücek, AYAis Yiorkis

We calculated the rate of phenotypic evolution in haldane units as indicated in Fig. 6, in which a mainland-only haldane estimate can be contrasted with a much steeper trend for ‘Ais Yiorkis. The slopes of these lines is the haldane rate, which provides a straight line trend of change over time (Purugganan and Fuller 2011). The units of this rate are based on a shift in the average and the standard deviation per generation (in this case, per year). The rates are summarized in Table 3. The inclusion of Cape Andreas Kastros one-grained einkorn only increased marginally the rate estimate of Purugganan and Fuller (2011). Separate rate estimates are calculated for grain breadth and grain height (thickness). Both of them are significantly faster than mainland rates, with grain height indicating evolution at a rate three times as fast and grain breadth more than six times as fast as mainland grain size increase.
https://static-content.springer.com/image/art%3A10.1007%2Fs00334-011-0323-1/MediaObjects/334_2011_323_Fig6_HTML.gif
Fig. 6

Graph of estimated evolution rates in haldanes for Pre-Pottery Neolithic einkorn (calculated from the data in Fig. 5). Black diamonds represent mainland sites, while bright squares indicate Cypriot data. x/s indicates the trait value per pooled standard deviation of the entire dataset, and the haldane unit is the slope of this line. The trend line through mainland data is dashed and its trend line is shown. An inferred trend line of ‘Ais Yorkis has been drawn by eye and is derived from interpolated populations of the time period between Dja’de & Çayönü. Generation 0 is defined as the earliest data point (Tell Qaramel)

Table 3

A comparison of estimated rates of phenotypic evolution (in the unit haldanes) for einkorn grain size traits, including the published estimate of Purugganan and Fuller (2011), a new mainland-only rate (which excludes Cape Andreas Kastros), and our estimate of Cypriot two-grained einkorn based on the ‘Ais Yiorkis population

 

Published estimate

R2

Mainland only

R2

Cyprus MPPNB

R2

Error hals

Breadth (width)

0.8 hals

0.62

0.6 hals

0.6

6 hals

0.999

0.2

Thickness (height)

1.2 hals

0.89

1.1 hals

0.87

3.3 hals

0.8

0.1

Discussion

The development of cereal agriculture in the Near East involved barley and several morphologically distinct forms of wheat, including emmer and both one and two-grained einkorn, as well as rye in some areas (Table 4). The distribution of finds of the species is not uniform across the Near East, suggesting that these cereals did not originate together in the same location (Willcox 2005) and neither did they spread as a single package. Within Cyprus the presence and ubiquity of the various cereals on Pre-Pottery Neolithic B sites is variable. The ubiquity of two-grained einkorn for ‘Ais Yiorkis is high in contrast to the dominance of one-grained einkorn recorded from other sites (Table 5). There is no evidence of two-grained einkorn in Cyprus prior to Middle/Late Pre-Pottery Neolithic B ‘Ais Yiorkis. Additional evidence of two-grained einkorn from the Cypriot Pre-Pottery Neolithic comes from Khirokitia Vounoi and Kalavasos-Tenta; however, the ubiquity and the number of grains is extremely low, 36% (n = 4) and 01% (n = 6), respectively. Conversely, the ubiquity and the total number of grains of one-grained einkorn for both sites is considerably higher, 41% (n = 2,896) for Khirokitia Vounoi and 17.5% (n = 74) for Kalavasos Tenta (Hansen 1989, 1994, 2005; Miller 1984; Waines and Stanley-Price 1977). In the subsequent Ceramic Neolithic, two-grained einkorn is present in 18% (12 grains) of the samples from Ayios Epiktitos Vrysi. This is relatively low compared with the ubiquity of one-grained einkorn, which is present in 69.7% of samples (102 grains) (Kyllo 1982) (Table 5).
Table 4

Tabular summary of evidence for early crops in the Near Eastern Fertile Crescent

 

Cal. b.c.

Einkorn, one-grained

Einkorn, two-grained

Rye

Emmer

Naked wheat

Barley

SE Anatolia (Turkey)

 Göbekli Tepe

9,200–8,600

 

XX

   

XXX

 Çayönü (RP,GP,Ch.H)

8,600–8,200

X

X

X

XX

 

XX

 Nevalı Çori

8,600–7,950

XXX

XX

X

XX

 

X

 Cafer Höyük XIII–VIII

8,300–7,450

XXX

X

X

XX

X

X

Middle Euphrates (Syria)

 Abu Hureyra I

11,150–10,450

 

XX

XX

   

 Tell Qaramel

10,300–8,850

 

XXX

 

X

 

X

 Mureybet I–III

10,200–8,800

 

XX

XX

  

XX

 Tell ‘Abr 3

9,500–9,200

 

XX

XX

  

X

 Jerf el Ahmar

9,450–8,700

X

XX

XX

  

XXX

 Dja’de

8,700–8,270

X

XX

XX

X

 

XXX

 Abu Hureyra 2A–C

7,800–7,000

 

XX

X

X

X

XX

 El Kowm II

7,100–6,350

 

XXX

X

XX

XX

X

 Tell Bouqras

7,500–6,300

X

  

XX

XX

XXX

 Sabi Abyad II

7,650–6,750

 

X

 

XXX

XX

XX

Coastal Syria

 Tell el-Kerkh

8,540–8,320

X

  

X

 

X

 Ras Shamra

7,600–6,000

X

?

 

XXX

X

XX

PPN Cyprus

 Mylouthkia

8,700–8,200

XX

  

X

 

XX

 Shillourokambos

8,250–7,350

   

XX

 

XXX

 Kalavasos Tenta

8,000–6,500

X

X

 

X

 

X

 ‘AIS YORKIS

7,9007,500

X

XXX

   

XX

 Cape Andreas Kastros

6,800–6,100

XX

  

XXX

 

XX

 Khirokitia Vounoi

>6,400–6,100

X

X

 

XXX

X

XX

 Kholetria Ortos

6,500–6,000

X

  

X

 

X

Ais Yorkis (this study) is in bold

Compiled from various sources, including Fuller (2008), Van Zeist (1999, 1981), Willcox (1999, 2005), Zohary and Hopf (2000), Colledge et al. (2004), Tanno and Willcox (2006), Murray (2003), Miller (1984), Waines and Stanley-Price (1977), Hansen (1989, 1994, 2005)

Table 5

Ubiquities for one-grained and two-grained einkorn in the Cypriot Aceramic and Ceramic Neolithic (Hansen 1989, 1994, 2005; Miller 1984; Waines and Stanley-Price 1977; Kyllo 1982)

 

Einkorn, one-grained (%)

Einkorn, two-grained (%)

Khirokitia Vounoi

41

0.36

Kalavasos Tenta

17.5

0.01

Ais Yiorkis

2.3

44.2

Ayios Epiktitos Vrysi

69.7

18

Horwitz et al. (2004) discuss the likelihood of multiple importation events of fauna to Cyprus in the Neolithic. Further, Colledge and Conolly (2007, p. 61) discuss the possibility of multiple importation events of domestic plant taxa based on evidence for gradual increase in crop species and herbaceous taxa present on sites from the Cypro-EPPNB to the Ceramic Neolithic. Willcox (2003) suggests the origins of the Cypriot crop package in the Early Pre-Pottery Neolithic B, including the plant assemblages of Shillourokambos and Mylouthkia, are contemporary with the first wave of crop expansion from southeast Anatolia, namely the expansion of one-grained einkorn, emmer and barley. All four cereals, barley, emmer, one-grained and two-grained einkorn are present in southeast Anatolia at sites such as Nevalı Çori, Çayönü and Cafer Höyük (Willcox 1999; Pasternak 1998; Fuller 2008) and appear together in the assemblages at Mylouthkia, with the exception of two-grained einkorn (Murray 2003) (Table 4). Sites located in the Syrian Middle Euphrates that have similar plant composition to ‘Ais Yiorkis include Dja’de, Jerf el Ahmar, Mureybet I–III and Abu Hureyra (Willcox 2003). However, the combination of cereals recovered from Middle Pre-Pottery Neolithic B ‘Ais Yiorkis is suggestive of a possible second wave introduction. The cereals at ‘Ais Yiorkis include two-grained einkorn, one-grained einkorn and barley, the latter two present in only one sample each (Table 1). On the basis of present evidence the combination of cereals that correspond to the ‘Ais Yiorkis assemblage is most similar to the Syrian Middle Euphrates sites, where emmer wheat and one-grained einkorn are absent (Willcox 2005; Fuller 2008). Similar cereals are also found at Göbekli Tepe (southeast Anatolia), which has wild two-grained einkorn and barley (Neef 2003; Fuller 2008). This differs from what could be suggested for the cereal crop compositions from Shillourokambos and Mylouthkia where cereal crop correlations can be made with material from southeast Anatolia (see Colledge et al. 2004 for an alternative view considering the full suite of wild weed and cereal crops, in which introduction from the southern Levant is suggested). The exploitation of wild two-grained einkorn appears in the Middle Euphrates from ca. 11150 cal. b.c. at Abu Hureyra and from ca. 10200 cal. b.c. at Mureybet I–III (Hillman 2000; Hillman et al. 2001; Van Zeist and Bakker-Heeres 1986; Willcox 2007). There is an increase in grain size at Jerf el Ahmar from ca. 9450 cal. b.c. and Dja’de from ca. 8700 cal. b.c. (Willcox 1996, 2002, 2007). The first appearance of domesticated two-grained einkorn on the mainland is from ca. 7800 to 6600 cal. b.c. from Abu Hureyra 2A–C, El Kowm, Tell Bouqras, Sabi Abyad II, and Tell Ramad (Table 4) (Willcox 2007; Fuller 2008). The AMS dates from the ‘Ais Yiorkis two-grained einkorn is ca. 7900–7500 cal. b.c. Therefore, at nearly the same time, or very shortly after, the first signs of grain size increase of two-grained einkorn appear along the Middle Euphrates, it spread to Cyprus, where grain size increased further.

Accelerated evolution in island colonization

It has been demonstrated here that once two-grained einkorn spread to Cyprus it increased significantly in size, at a rate between 3 and 6 times faster than mainland einkorn (either one or two-grained) was increasing in size. In addition, this rate is equally faster than grain size evolution in barley, emmer, pea or lentil on the mainland (see Purugganan and Fuller 2011; Fuller et al. 2011, this volume). While some shift toward increased grain size is first indicated at Early Pre-Pottery Neolithic B Jerf el Ahmar and Dja’de, with slightly increased breadth and width from the earlier PPNA phase (Fig. 4), the ‘Ais Yiorkis evidence suggests a divergent, rapid rate of evolution after the two-grained einkorn of Cyprus was separated from the main evolutionary trends of mainland einkorn during the mid-9th millennium b.c. (Early PPNB). This apparent contrast between the rate of evolution of grain size on Cyprus as compared to the mainland data reinforces the inference that cereal domestication, as represented in the fertile crescent was a slow process (as per Tanno and Willcox 2006; Fuller 2007), contrary to some claims (e.g. Abbo et al. 2010; Hillman and Davies 1992). Instead the island context provided for different dynamics in the change of two-grained einkorn.

This evidence fits with one or both hypotheses that a local bottleneck effect in a population introduced to Cyprus facilitated a more rapid rise of domestication adaptations in dispersing crops (Fuller 2008), and that domestication as a process was enhanced by the early crop being removed from the geographic range of its wild progenitor (Allaby 2010). Although it is possible that larger two-grained einkorn was also on the mainland and has yet to be sampled, all of the available mainland data argue against this, with larger size ranges apparent only by the Late PPNB, after cal. 7500 b.c. This levelling off of grain size increase might correlate with the period in which domestication processes were coming to an end and a more complete agricultural package of the Levantine Neolithic had been implemented (Asouti and Fuller 2011). Admittedly, the total number of sites with grain measurements is still limited, but the trends here highlight the potential significance of reporting more metrical data, especially on early cereals (cf. Willcox 2004; Colledge 2004).

Some authors have suggested that grain size, especially plumpness, may be correlated with improved water availability during grain filling, e.g. due to irrigation or higher rainfall (e.g. Miller 1999), but this does not seem to explain the patterns reported here. To consider this possibility, we have tabulated the einkorn grain metrical data in order of decreasing local rainfall (based on modern estimates) in Table 6. Although climate has changed over time, the relative degrees of precipitation between different parts of the Near East are likely to have remained fairly consistent. Here it can be seen that sites in the wettest regions do not consistently have the largest grains (e.g. Çayönü), nor do sites in the driest areas, such as Ramad or Wadi Jilat, have the smallest grains. While a more detailed consideration of the potential effects of local environmental conditions may be worthwhile, it is beyond the scope of the present study. However, it appears justifiable to conclude (albeit based on a limited data set) that two-grained einkorn grains were larger at Cypriot PPNB ‘Ais Yiorkis than would be expected on the basis of comparison with grains from mainland sites of similar date, thus suggesting that einkorn on Cyprus might have diverged from the main evolutionary trend of the Near East.
Table 6

A comparison of average measurements (mm) of einkorn grain breadth and thickness for a selection of Near Eastern sites alongside estimated average rainfall (precipitation under modern conditions), organized from wettest to driest localities

  

Precip. (mm)

Breadth (mean)

SD

Thickness (mean)

SD

n

Çayönü

EPPNB

550

1.600

0.302

1.910

0.285

10

‘Ais Yiorkis

MPPNB

550

2.442

0.294

2.050

0.274

117

Erbaba

LPPNB

500

1.914

0.2495

2.082

0.214

240

Cape Andreas Kastros (Cyprus)

LPPNB

500

2.280

0.26

2.220

0.26

10

Höyücek

PN

430

1.452

0.242

1.987

0.273

146

Qaramel

PPNA

350

1.329

0.384

1.136

0.275

198

Dja’de

EPPNB

300

1.420

0.32

1.546

0.319

102

Mureybet

PPNA

280

1.229

0.155

1.275

0.138

330

Jerf al Ahmar Early

PPNA

250

1.153

0.217

1.153

0.13

27

Jerf al Ahmar Late

EPPNB

250

1.320

0.29

1.320

0.32

60

Ramad II

LPPNB

250

1.814

0.271

1.814

0.18

46

Wadi Jilat 7

MPPNB

100

1.670

0.43

1.830

0.35

31

Also, it is possible that the grains measured from ‘Ais Yiorkis are representative of a cleaned crop with the removal of the smaller tail grains likely. However, evidence from Kissonerga Mylouthkia (Murray 2003) provides the opposite to the evidence from ‘Ais Yiorkis, with high ratios of glume bases and weeds to grains and low number of grains per litre suggesting the plant material was the result of fine-cleaning residue. The cereal grains from Mylouthkia (Colledge 2004) are relatively large and comparable in size to the two-grained einkorn recovered from ‘Ais Yiorkis. Thus, there is evidence of grain size increase once the cereal crops were transported to the island whether the grains come from sieved or un-sieved samples.

A local bottleneck effect, due to the reduced population size, when initially introduced, could push wild type genes to lower frequencies meaning a jump in frequency of domesticated, large-grained types. Recent attempts at modelling the reduction of genetic diversity with the spread of crops, assume that a process like this is typical with crop dispersal (Van Etten and Hijmans 2010). This would certainly be the case if the initial bag(s) of seed corn brought to Cyprus had been harvested from relatively few mainland plants or fields. However, if grain were drawn from a wider population of harvested plants, then a reasonably small volume (a few sacks of grain containing many 1,000s of grains) could conserve quite high levels of genetic diversity. Nevertheless translocation to Cyprus would have presumably reduced the likelihood of grain exchange within or between areas that might be postulated for the mainland. Another key factor, however, was presumably removal of gene flow from wild populations, and/or from other cultivated populations in gene contact with these, since wild populations would have represented large gene pools with smaller seed size. As explored by Allaby (2010; Allaby et al. 2010) such gene flow is likely to contribute to reducing the rate at which domestication-adapted mutations, including larger grain size, accumulated due to reintroduction of wild-types from the large surrounding population of wild cereals. In addition, as postulated by Willcox et al. (2008), to help explain the slow fixation of non-shattering ears, early cultivators may have still bolstered grain stores by harvesting from wild stands. In either case it would appear that the introduction of two-grained einkorn to the novel context of Cyprus altered the dynamics of the domestication process for the local einkorn population represented by ‘Ais Yiorkis.

Conclusions

Charred plant remains from Pre-Pottery Neolithic Krittou Marottou ‘Ais Yiorkis demonstrate crop introduction of two-grained einkorn possibly from the Middle Euphrates very shortly after signs of grain size increase occur on the mainland. Grain measurements of two-grained einkorn from ‘Ais Yiorkis are compared to those from prehistoric Cyprus, northern Syria, and central Europe, and suggest that these had attained a large size comparable to later domesticated populations on the mainland. Measurements of two-grained einkorn wheat from ‘Ais Yiorkis are considered with the same allochronic data used in a recent study by Purugganan and Fuller (2011) and the results indicate an accelerated evolutionary rate in grain size increase.

The possibility of a domestication ‘event’ or rapid fixation of a domestication syndrome trait in the context of an initially small island population that was cut off from mainland gene flow is suggested by the colonisation by farmers of Cyprus in the Cypro-Pre-Pottery Neolithic. These data would seem to imply that once established on Cyprus there was limited, if any, further import of two-grained einkorn seed corn from the mainland, in order for the distinctiveness of the two-grained einkorn population on ‘Ais Yiorkis to be maintained. While the data for the island suggest a series of crop introductions (Colledge and Conolly 2007), once established the crops may have been maintained locally with little additional seed corn input from the mainland. This at least appears to have been the case with two-grained einkorn, while evidence for the other cereals awaits equivalent study. Two-grained einkorn at least would indicate a contrast with domesticated fauna, since zooarchaeological evidence suggests recurrent introductions of mammalian species (Horwitz et al. 2004). Thus, rather than simply being an island outpost of the Pre-Pottery Neolithic, Cyprus reveals some of the dynamic processes by which species were introduced and adapted to new environments with the spread of agriculture.

Acknowledgments

We thank our anonymous reviewers for their constructive comments. These led to our inclusion of discussions on the effects of rainfall on grain size, which were a direct response to suggestions made by one of the reviewers. Archaeobotanical fieldwork was funded by the UNLV Graduate and Professional Student Association, UNLV International Programs, UCL Institute of Archaeology Awards and University College London Graduate School. Principal funding was provided by the National Science Foundation (Grant Number BCS-0352689), the Brennan Foundation, the Institute for Aegean Prehistory, the University of Nevada at Las Vegas, and the Johnson Foundation. Thanks to George Willcox for providing a copy of his raw data in spread sheet format.

Supplementary material

334_2011_323_MOESM1_ESM.doc (34 kb)
Supplementary material 1 (DOC 34 kb)

Copyright information

© Springer-Verlag 2011