Phytolith evidence of historical rice cultivation at Wormsloe Historic Site, Georgia, USA

Journal of Archaeological Science: Reports 14 (2017) 557–574

Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep

Phytolith evidence of historical rice cultivation at Wormsloe Historic Site, Georgia, USA

MARK

Alessandro Pasquaa,⁎, Liovando Marciano da Costaa,1, Ervan Garrisonb a b

Department of Geography, The University of Georgia, 210 Field Street, 30602, Athens, GA, USA Department of Geology, The University of Georgia, 210 Field Street, 30602, Athens, GA, USA

A R T I C L E I N F O

A B S T R A C T

Keywords: Rice cultivation Phytolith analysis Radionuclide analysis Environmental history African American culture Wormsloe Historic site

Despite much of the environmental history of Wormsloe Historic Site, Georgia, having been previously documented and described, the aspect related to whether rice cultivation was ever performed on-site still has not been entirely defined. The main goal of this study is the archaeobotanical investigation of soil samples collected from a salt marsh at Wormsloe where rice cultivation is suggested to provide a more definitive answer to this aspect. The study employs phytolith analysis to reveal evidence of past agricultural practices on-site related to rice cultivation. The results reveal the presence of phytoliths belonging to the rice genus, with some indicating a domesticated nature. Cultivation was likely done for subsistence purposes by the African American community living on the property. Preliminary radionuclide analysis suggests the soil samples may be older than 100 years. For the first time in Wormsloe's history, concrete evidence suggesting that rice was indeed cultivated on-site has been obtained, while also casting some light on local African American agricultural practices. The results of this study fill a gap in Wormsloe's environmental history, and increase Wormsloe's cultural, historical, and archaeological significance both at the regional and the national level. Ultimately, this study provides a reliable method to further investigate historical rice cultivation in other areas within the Wormsloe Site as well as along the southeastern United States.

1. Introduction Wormsloe State Historic Site2 is located on the Isle of Hope, just south of Savannah, Georgia, and represents one of the most significant historical, cultural, and natural sites in the southeastern United States. Since its establishment in 1736 by British colonist Noble Jones, Wormsloe has served many purposes becoming a military outpost during the first years of the Georgia colony, a sea island cotton plantation in the 1800s, and a farm and tourist attraction in the 1900s. A State Historic Site since 1973, today Wormsloe serves also as a research center with the establishment of the “University of Georgia Center for Research and Education at Wormsloe” (UGA – CREW). Interdisciplinary research is performed on-site with the support of the Wormsloe Foundation and the Wormsloe Institute for Environmental History (WIEH) with the purpose of furthering the understanding of Wormsloe's environmental history. One of the aspects that have not yet been fully understood is that related to whether rice cultivation was ever performed. Previous research (Bragg, 1999; Coulter, 1955; Kelso, 1979; Swanson, 2012) did

⁎

not fully investigate the potential for on-site rice cultivation, and therefore did not explore the use of archaeobotanical analysis and thorough landscape inspection to study this aspect. At the moment, there are only hints suggesting that rice might have been cultivated at Wormsloe. One of these is the letter that Benjamin Franklin sent in 1772 to Noble Wimberly Jones – son of Noble Jones – containing a sample of upland rice from Vietnam to suggest the experimentation of such culture at Wormsloe. Another suggestion is provided by the 1880 agricultural census – listing Wormsloe's freedmen tenants and their land use during the 1879 season – reporting that one of the tenants, Peter Campbell, raised 510 lb of rice. As Swanson (2012) suggests, although it is unclear what land these tenants could have used, it is likely that they had individual plots close to their accommodation at the slave cabins, probably where the old quarters field was located. After thoroughly inspecting the Wormsloe topography through LiDAR elevation maps, historical maps, and ground survey, a small tidal salt marsh inlet on the eastern side of the Isle of Hope along the Skidaway River was selected (Fig. 1). The area is roughly 1416 m2 in size (0.34 acres) and has a rectangular shape measuring 70 × 20 m.

Corresponding author at: Department of Geography, The University of Georgia, 210 Field Street, 30602, Athens, GA, USA E-mail addresses: [email protected] (A. Pasqua), [email protected] (L.M. da Costa), [email protected] (E. Garrison). Permanent address: Departamento de Solos, Universidade Federal de Vicosa, 36570-900, Vicosa, MG, Brazil. 2 Hereafter referred to simply as “Wormsloe” 1

http://dx.doi.org/10.1016/j.jasrep.2017.06.024 Received 9 January 2017; Received in revised form 27 May 2017; Accepted 8 June 2017 2352-409X/ © 2017 Elsevier Ltd. All rights reserved.

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Fig. 1. The location of Wormsloe and the study area.

belonging to the Oryzoideae tribe of the Poaceae (Gramineae) family, with about 23 species at the genus level Oryza (Gu et al., 2013). Oryza species include both domesticated types such as Oryza sativa as well as wild species such as Oryza minuta and Oryza officinalis (Gu et al., 2013). In general, rice plants produce three distinctive phytolith morphologies, each made by distinct parts of the plant: the husk produces doublepeaked phytoliths, while the leaves and stem produce fan-shaped bulliforms and bilobates phytoliths (Fig. 2) (Gu et al., 2013; Zhang et al., 2010). Phytoliths have provided evidence of historical rice cultivation in several archaeological areas (Cao et al., 2006; Chun-Hai et al., 2007; Fujiwara and Kaner, 1993; Huang and Zhang, 2000; Itzstein-Davey et al., 2007; L. Jiang and Liu, 2006; Q. Jiang, 1995; Q. Jiang and Piperno, 1994; Jin et al., 2014; Lu et al., 2002; Lu et al., 2006; Qiu et al., 2014; Watanabe, 1968; Zhang et al., 2010; Zhao et al., 1998; Zhao and Piperno, 2000; Zheng and Jiang, 2007; Zheng et al., 2003b). As Huang and Zhang (2000) argue, in fact, phytolith analysis represents “an indirect but reliable method for detecting rice cultivation at archaeological sites”.

The vegetation in the marsh is mostly Juncus roemerianus (black needlerush) and Spartina alterniflora (smooth cordgrass), while saw palmetto and maritime forest are in the immediate surroundings. The area was chosen due to the following characteristics deemed compatible with rice cultivation: 1) what appears to be a manmade dike, which may represent a structure for controlling and preventing salt water influx; 2) two drainage ditches converging and discharging their waters in the marsh, which could have channeled freshwater to the cultivated field; 3) historical maps reveal that one of the two ditches was connected to a water pond, which could have served as reservoir for the rice field; 4) the poorly drained soil of the marsh may have provided the right type of soil for the cultivation of rice; 5) the area is located approximately 173 m south of the only surviving 19th century slave cabin at Wormsloe. The cabin was part of the so-called “slave quarters” which was built in the second half of the 1800s to accommodate enslaved laborers working on the sea island cotton crop cultivated and harvested at Wormsloe. Being near the slave settlement, the area might have been chosen by the African American population to cultivate subsistence crops such as rice. The main goal of this study is to perform an archaeobotanical analysis of the study area to better understand the aspect of historical rice cultivation at Wormsloe. In particular, this study addresses the following questions: was rice ever cultivated at Wormsloe? If so, when and how was it cultivated? Soil samples from the study area were collected to investigate the presence of rice phytoliths. The results of this study improved the understanding of an existing gap in Wormsloe's environmental history, thus increasing its archaeological, cultural, and historical significance both within Georgia and the southeastern United States. Rice represents one of the most important crops in the world, and it can grow in a wide range of ecological systems ranging from the uplands – where it is cultivated as a dry crop – to bottomlands – where it is cultivated as a wet crop (Weisskopf et al., 2013). Rice is a grass

2. Materials and methods 2.1. Data collection Between September 2 and 5, 2014 ten soil cores were collected with a 3-in. diameter hand auger to a maximum depth of approximately 1.5 m at low tide conditions. To avoid biases in the collection of samples, 10 random GPS point coordinates were generated in ArcMap 10.2 of the ArcGIS software package. Each core was then placed on a metal gutter to estimate the number of horizons, color (using a Munsell Color Chart), texture, and other physical characteristics such as the presence of organic matter. Samples were collected from each horizon of each core for a total of 43 samples. These were later stored in sealed plastic bags and air-dried to let moisture evaporate. For the purpose of this 558

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Fig. 2. Phytolith types from rice and morphological parameters: bulliform (left); double-peaked (center); bilobate (right). Abbreviations in bulliform: HL (horizontal length); VL (vertical length); LL (lateral length); a (vertical length of the non-base portion); b (vertical length of the base portion). Abbreviations in double-peaked: TW (top width); MW (middle width); H1, H2 (height 1, height 2); CD (curve depth).

Table 1 Physical characteristics of the cores and samples processed for phytolith analysis. Horizon

Depth (cm)

Color

Texture

Comment

1(581) Core 1 2 3a* 3b

0–25 40–60 80–100 130–150

2.5 YR 2.5/1 (Black) 5BP 2.5 (Bluish black) N 2.5 gley (Black/gley) N 2.5 gley (Black/gley)

Sandy clay Sandy clay Clay Clay

Considerable organic matter, quite plastic to the touch Considerable organic matter Very clayey, less organic matter* Very clayey, less organic matter * Two samples were taken from the same horizon

3(583) Core 1 2 3 4 5

0–10 10–60 60–80 80–110 110-down

5YR 2.5/1 (Black) 10 YR 3/1 (Very dark gray) 10 YR 4/1 (Dark gray) 10 YR 3/1 (Very dark gray) 10 YR 4/1 (Dark gray)

Clayey Sandy Sandy Loamy sand Sandy

Small roots Organic matter Soil is quite brittle Moist, plastic, easy to mould Very little organic matter

6(586) Core 1 2 3

0–30 30–60 60–110

10 YR 2/1 (black) 7.5 YR 4/1 (Dark gray) 10 YR 5/1 (Gray)

Clayey Sandy clay Sandy clay

Organic matter Moist Very sandy

for 1 h to remove carbonates, and finally treated with 50 mL of 0.5 N sodium hydroxide (NaOH) for 8 h to deflocculate the soil particles. Following Piperno (1988), samples were then divided into 3 soil fractions yielding a total of 36 samples. Each sample was then oven dried and weighed to assess the phytolith content in each sample (Table 2). Sodium polytungstate (SPT) was used for heavy liquid separation (density 2.30 g/mL). Samples were put in test tubes with 10 mL of heavy liquid solution, and then centrifuged 3 times at approximately 2250 rpm for 10 min to allow phytolith separation (David Leigh, personal communication, October 2014). A suction device provided with a 0.45 μm diameter filter was used to retrieve the phytolith residue from each sample. Phytolith extract from each sample was oven dried, weighed, and mounted on slides with Entellan mounting medium for microscope analysis. Plant and grain phytolith extraction followed the method developed by Costa (personal communication, September 2014). Samples were washed with distilled water, and then put in a muffle furnace at 300 °C for 2 h. Then, the furnace was let open for about 15 min to allow some air in, thus facilitating the process of ashing the sample. The temperature in the furnace was then raised at 500 °C. After 2 h, the samples were extracted and treated with 10% HCl for 1 h to remove carbonates. After washing them with distilled water until no carbonate residue was present, samples were then dried in the oven at 105 °C, and finally mounted on slides for microscope analysis.

study, three cores were chosen for phytolith extraction and analysis for a total of 12 samples (Table 1). Domesticated rice specimens from the United States (Georgia) and Brazil were collected to compare them with the Wormsloe samples. From the United States, Carolina Gold rice plants were collected from a private garden in the greater Athens area,3 while Carolina Gold rice grains were retrieved from Sapelo Island.4,5 Brazilian specimens included the five different varieties of Ouro Minas, Predileta, Rio Grande, Rubelita, and Seleta. Rice specimens from China used in the study by Gu et al. (2013) were also used as reference. Reference samples were chosen with the goal of providing as much a geographic diversity as possible, with samples coming from Asia, South America, and North America. 2.2. Phytolith extraction Phytolith extraction was performed following both Piperno (1988) and the procedure illustrated by L. M. Costa during his visit to the Department of Geography in September 2014. Following the latter, 10 g of each sample were put in a muffle furnace for 4 h at 500 °C to remove organic matter, and were then homogenized with a pestle and mortar; specimens were processed with 50 mL of 10% hydrochloric acid (HCl) 3 Before being transplanted in a private garden in Hull, Georgia, the Carolina Gold rice grains were planted in a nursery at the Southern Seed Legacy Lab, a local lab linked to the UGA Anthropology Program which fosters the cultivation of local seeds and plants. Originally, the grains are believed to come from Anson Mills, an heirloom grain company (Susannah Chapman, personal communication, January 2015). 4 Rice grains were obtained from the Carolina Farm Stewardship Association, an organization involved in the preservation of heirloom southeastern vegetables, grains, and flowers (Stanley Walker, personal communication, January 2015). 5 Hereafter referred to simply as ‘Sapelo’.

2.3. Phytolith analysis and interpretation Slides were scanned at 200 × and 400 × magnification using a Zeiss Axio Imager A2 microscope provided with Axio MRc5 digital camera and AxioVision LE64 software. All slides were scanned entirely to allow the identification of any Oryza-type phytolith. When possible, at least 559

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

occupied by each sample was created. The statistical analyses were conducted using the SPSS 20.0 software package, with a p-value of significance equal to 0.05.

Table 2 Weight of samples by soil fractions and phytolith content. Very fine sand (VFS) (0.105–0.053 mm); coarse silt (0.053–0.015 mm); fine silt (0.015–0.002 mm). Sample # (corehorizon)

Net weight of soil fraction (grams)

Net weight phytolith content (grams/percentage of total)

1-1 VFS 1-2 VFS 1-3a VFS 1-3b VFS 1-1 Coarse silt 1-2 Coarse silt 1-3a Coarse silt 1-3b Coarse silt 1-1 Fine silt 1-2 Fine silt 1-3a Fine silt 1-3b Fine silt 3-1 VFS 3-2 VFS 3-3 VFS 3-4 VFS 3-5 VFS 3-1 Coarse silt 3-2 Coarse silt 3-3 Coarse silt 3-4 Coarse silt 3-5 Coarse silt 3-1 Fine silt 3-2 Fine silt 3-3 Fine silt 3-4 Fine silt 3-5 Fine silt 6-1 VFS 6-2 VFS 6-3 VFS 6-1 Coarse silt 6-2 Coarse silt 6-3 Coarse silt 6-1 Fine silt 6-2 Fine silt 6-3 Fine silt

0.875 0.911 1.212 1.33 0.506 0.492 0.519 0.531 0.066 0.103 0.121 0.087 0.726 0.796 0.812 0.914 0.9 0.541 0.475 0.398 0.391 0.342 0.112 0.09 0.069 0.064 0.043 0.852 0.899 1.056 0.2 0.226 0.172 0.053 0.061 0.054

0.006 0.68% 0.006 0.65% 0.013 1.07% 0.018 1.35% 0.025 4.94% 0.015 3.04% 0.027 5.2% 0.038 7.15% 0.003 4.54% 0.005 4.85% 0.033 27.27% 0.015 17.24% 0.003 0.41% 0.006 0.75% 0.005 0.61% 0.004 0.43% 0.006 0.66% 0.034 6.28% 0.171 36% 0.023 5.77% 0.023 5.88% 0.03 8.77% 0.002 1.78% 0.004 4.44% 0.002 2.89% 0.003 4.68% 0.001 2.32% 0.004 0.46% 0.003 0.33% 0.003 0.28% 0.027 13.49% 0.022 9.73% 0.013 7.55% 0.001 1.88% 0.003 4.91% 0.001 1.85%

3. Theory 3.1. Distinguishing between domesticated and wild rice phytoliths When using phytoliths to investigate rice cultivation in Georgia and, generally, in the southeastern United States, care should be taken in distinguishing between domesticated and wild varieties of the Oryzoideae tribe. In fact, as Lu and Liu (2003a) state, wild relatives of domesticated rice such as Leersia oryzoides (rice cutgrass) and Zizaniopsis miliacea (giant cutgrass) may be present in the area. Because rice phytolith morphologies have different taxonomic values, in some cases there could be limits in the correct identification and interpretation of domesticated rice phytoliths, so that multiple criteria and factors need to be considered. It is generally assumed that double-peaked phytoliths have the highest taxonomic value, for it has been demonstrated that they are unique to the genus Oryza and can therefore be used to successfully distinguish between domesticated and wild rice species, even where geographic overlap exists (Gu et al., 2013; Pearsall et al., 1995; Zhao et al., 1998). In particular, this study followed the discrimination method developed by Zhao et al. (1998) to successfully predict rice domestication using double-peaked phytoliths as demonstrated in some studies (Itzstein-Davey et al., 2007; Wu et al., 2014; Zhao, 1998; Zhao et al., 1998). On the other hand, the taxonomic value of fan-shaped rice bulliforms can be ambiguous. While bulliforms are common to many grass taxa in the Poaceae family (Lu et al., 2002; Lu et al., 1997), fan-shaped bulliform cells6 in rice are characterized by shallow scale-like decorations around the base, and by two lateral protrusions located on the half-round side of the cell (Fujiwara and Kaner, 1993; Lu et al., 2002; Wu et al., 2014) (see Fig. 2). As demonstrated by Lu et al. (2002) – who analyzed bulliform phytoliths from 16 grass species – the presence of scale-like decorations in bulliform cells represents a distinctive characteristic of the genus Oryza, thus providing a reliable way to discriminate between domesticated and wild rice species. Furthermore, bulliform cells from domesticated rice statistically present a number of 9 decorations or higher, while those from wild rice usually have < 9 decorations. This criterion has been used to infer rice domestication in a number of studies (Lu et al., 2002; Wu et al., 2014) and was therefore employed for the purpose of this study. Other studies (Fujiwara and Kaner, 1993; Huang and Zhang, 2000; Zheng et al., 2003a; Zheng et al., 2003b) have also considered the morphological parameters VL, HL, LL, ‘b’, and ‘a’ (see Fig. 2) in bulliform cells to discriminate rice at the species and sub-species level. According to Pearsall et al. (1995) and Zhao et al. (1998), however, these methods alone are not sufficient to distinguish between domesticated and wild rice species where they overlap geographically. Moreover, Huang and Zhang (2000) argue that domestication from bulliform cells can be inferred by calculating the a/b mean ratio, which in domesticated rice should be < 1. For this study, this criterion could not be used for the reference specimens from China, as their ‘a’ and ‘b’ measurements were taken differently and would therefore determine inaccuracies; similarly, the Brazil samples could not be included as they did not present ‘a’ and ‘b’ measurements. Finally, from the analysis of domesticated bulliforms from Brazil (Appendix A), it was noted that the length to width (L/W) ratio falls within the 0.85–1.15 range in 70% of the cases (17/24 bulliforms) for Ouro Minas; 81% (18/22 bulliforms) for Predileta; 84% (21/25 bulliforms) for Rio Grande; 84% (21/25 bulliforms) for Rubelita; and 65% (17/26 bulliforms) for Seleta (L. M. Costa, personal communication,

300 phytoliths were counted on each slide, and classified morphologically following Jiang and Piperno (1994) and Huang and Zhang (2000) (Table 3). Phytoliths of interest were photographed, measured as described in Fig. 2, and then compared with reference specimens to determine whether and to what extent the Wormsloe phytoliths differ from other known domesticated rice samples. For double-peaked phytoliths, reference samples from the United States (both Athens and Sapelo) and China were used, while for bulliform phytoliths reference samples included the China, Athens, and Brazil specimens.

2.4. Statistical analysis A multivariate analysis of variance (MANOVA) was performed to investigate the overall difference between samples. The null hypothesis for the statistical analyses was that no significant difference between samples could be appreciated, while the alternative hypothesis would reflect significant differences between samples. Since for both the analyses of bulliform and double-peaked phytoliths the assumption for homogeneity of covariance matrices was not met, Pillai's trace results are presented instead of Wilks' lambda. Post hoc tests were also performed to analyze which specific morphological parameters differed among samples. The dependent variables used for statistical analysis were HL, VL, and L/W for bulliforms, while for double-peaked TW, MW, H, and CD were used. For post hoc tests, Tukey's HSD (Honestly Significant Difference) was used, while Games-Howell was used when the samples' variance was unequal. A discriminant function analysis (DFA) was also performed to graphically show similarities and differences among samples, and a plot showing the multidimensional space

6

560

Also known as keystone, cuneiform bulliform, or motor cells.

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Table 3 Phytolith counts per sample. B: bulliform; S/R: square/rectangular; E: elongate; R/E: round/elliptical; O: others (including amorphous, epidermis, polylobate); D: diatoms; S: saddle; BI: bilobates; C: cross; SS: sponge spicules; OB: Oryza-type bulliform; OH: Oryza-type husk. Sample # (core-horizon)

1-1 VFS 1-2 VFS 1-3a VFS 1-3b VFS 1-1 Coarse silt 1-2 Coarse silt 1-3a Coarse silt 1-3b Coarse silt 1-1 Fine silt 1-2 Fine silt 1-3a Fine silt 1-3b Fine silt 3-1 VFS 3-2 VFS 3-3 VFS 3-4 VFS 3-5 VFS 3-1 Coarse silt 3-2 Coarse silt 3-3 Coarse silt 3-4 Coarse silt 3-5 Coarse silt 3-1 Fine silt 3-2 Fine silt 3-3 Fine silt 3-4 Fine silt 3-5 Fine silt 6-1 VFS 6-2 VFS 6-3 VFS 6-1 Coarse silt 6-2 Coarse silt 6-3 Coarse silt 6-1 Fine silt 6-2 Fine silt 6-3 Fine silt Overall total ⁎

Phytolith types B

S/R

E

R/E

O

D

S

BI

C

SS

OB

OH

TOT⁎

53 58 43 56 83 91 64 54 15 19 23 10 55 71 63 67 47 62 37 53 52 46 12 14 11 24 8 45 54 59 51 59 34 10 2 2 1507

139 117 115 109 133 111 121 88 48 58 43 58 106 113 143 120 115 128 105 137 115 95 56 77 70 61 37 97 98 97 113 124 101 73 10 8

19 16 26 25 19 26 30 26 25 29 31 34 25 35 20 38 28 7 26 20 19 13 32 33 21 28 7 24 17 19 20 12 14 17 12 1

9 12 36 29 24 25 35 38 72 65 73 73 36 42 28 23 39 36 37 34 38 32 51 47 59 62 20 45 18 53 34 45 29 69 20 4

76 97 78 73 40 46 47 91 129 124 122 119 75 38 45 46 71 65 93 56 76 113 142 121 135 122 87 83 111 69 78 59 122 117 28 3

1 0 0 0 1 0 1 0 6 1 1 0 2 0 0 0 0 2 0 0 0 0 4 8 0 0 0 5 1 0 3 0 0 11 0 0

2 0 1 8 0 0 1 3 4 4 5 5 0 0 0 0 0 0 2 0 0 1 3 0 4 3 1 1 1 1 1 1 0 3 1 2

0 0 1 0 0 0 0 0 0 0 2 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 5

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 0 0 0 0 1 1 0 1 0 0 1 1 0 1 6 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0

2 0 2 4 11 8 10 12 0 0 5 0 1 0 0 1 2 6 4 3 7 5 1 0 0 0 0 3 0 1 5 3 3 0 0 0 99

0 1 1 1 2 2 0 4 0 3 1 2 1 1 0 2 1 1 2 0 2 1 1 2 0 0 0 1 0 1 0 0 0 0 0 0 33

300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 161 300 300 300 300 300 300 300 72 20 10.153

Total number of phytoliths counted on the slide.

range of phytolith shapes can, in fact, be limited for the wealth of genetic differences occurring in nature. As a result, many studies have adopted this approach to infer the presence of domesticated rice species – thus cultivation – at archaeological sites (Zhang et al., 2010), while others have also employed contextual features, such as the presence of agricultural tools (Zheng and Jiang, 2007), the finding of rice grains and the proximity to inhabited areas (Huang and Zhang, 2000; Zheng et al., 2003a; Zheng et al., 2003b), or simply the absence of wild rice in the area (Q. Jiang and Piperno, 1994).

September 2014) (Fig. 3). The analysis of samples from the United States revealed that 20/30 (66.6%) bulliforms fall in this range, while this holds true for 18/20 (90%) bulliforms of the reference samples from China. The 0.85–1.15 range could therefore be an additional indicator to infer domestication in rice. As argued by Lu and Liu (2003b), bilobate phytoliths are very common among grass taxa, as they are produced by at least 85 species belonging to the Panicoideae, Oryzoideae, Chloridoideae, and the Arundinoideae grass subfamilies. From the morphological classification system developed by Lu and Liu (2003b), it was found that bilobates with truncated margins at both ends were somewhat diagnostic of the Oryzoideae tribe or subfamily. Other studies (Gu et al., 2013; Huang and Zhang, 2000; Wu et al., 2014) also noted that bilobates in the Oryzoideae subfamily are arranged in a parallel fashion along the plant stem or leaves (see Fig. 2). However, a 3D analysis of the morphological parameters of rice bilobates suggested that bilobates could not be diagnostic at the species level (Gu et al., 2013). Costa et al. (2010) suggested also that bilobates are not stable and tend to dissolve, so they are not commonly found in the soil record despite the presence of grasses that produce them; furthermore, bilobates tend to dissolve more easily when processed with sodium hydroxide in the lab than other types of grass phytoliths (L. M. da Costa, personal communication, September 2014). As suggested by Lu and Liu (2003b), when single phytolith morphologies cannot be diagnostic at the species level, a more holistic approach based on phytolith assemblages should be used instead. The

4. Results and discussion The results in Table 3 show that no Oryza-type bilobates were identified among the 10.153 phytoliths counted, which confirms their tendency to dissolve both in the soil and during the extraction procedure in the lab as suggested by Costa et al. (2010). On the other hand, the most abundant types across samples include bulliforms, of which 99 were classified among the Oryza-type ones. Oryza-type double-peaked phytoliths corresponded to 33. Overall, the samples yielded enough phytolith content to perform the set minimum count of 300 phytoliths, except for three deep samples producing very little fine silt content (see Table 3). 4.1. Bulliform phytoliths Of the total 99 Oryza-type bulliforms initially selected, 26 were 561

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Fig. 3. L/W ratio and mean values for Brazilian domesticated bulliforms.

bulliforms from Athens, revealing that 25 bulliforms (83.33%) presented between 9 and 17 scales for a mean value of 10.56, which closely resembles that of Wormsloe bulliforms. With regard to the second criterion, the mean a/b ratio value in Wormsloe bulliforms corresponds to 0.83, which gives them a rather proportioned and symmetrical appearance. This aspect was also met in 22 out of 30 (73.33%) Carolina Gold bulliforms from Athens presenting a mean a/b value of 0.81, a value very close to that of Wormsloe. The mean L/W values of the Wormsloe bulliforms and the 20 bulliforms from China both correspond to 0.97 (Table 4), thus showing

classified as domesticated (Fig. 4). To increase confidence in the prediction of domestication, all the following criteria were used: 1) the presence of at least 9 scale-like decorations; 2) an a/b ratio < 1; 3) a L/ W ratio within the 0.85–1.15 range; and 4) by comparing the VL and HL values with their respective mean values of reference specimens from Brazil, China, and the United States (Table 4). The 26 bulliforms classified as domesticated have a number of scales ranging between 9 and 13, with a mean value of 10.65. Some also display one or two lateral protrusions typical of rice bulliforms (Fig. 4d, e, k, m, o, p, u). This criterion was also tested on 30 Carolina Gold 562

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Fig. 4. Domesticated bulliform phytoliths from Wormsloe.

values between 35.03 μm and 70.93 μm, with a mean value of 44.37 μm; as for HL values, they range between 35.3 μm and 80.19 μm, with a mean value of 45.41 μm. These values show a closer resemblance to the bulliforms from China, rather than to those from Athens or Brazil. This is also demonstrated by the results of the statistical analysis as shown in Fig. 5 and Table 5. As also demonstrated by post hoc analysis (see Appendix B.2), it is

great morphological similarities. Slightly different but still within the 0.85–1.15 range are the mean L/W values of 49 specimens from Brazil and that of 30 specimens from Athens, i.e., 1.03 and 1.09, respectively (see also Fig. 3). Finally, the VL and HL values of the Wormsloe bulliforms were compared with mean VL and HL values of reference samples to appreciate similarities and differences concerning their size. As shown in Table 4 it was noted that the Wormsloe bulliforms present VL 563

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Table 4 Measurements and criteria for Wormsloe domesticated rice bulLiforms. Unless specified otherwise, units are in μm (micrometers). Sample # (core-horizon)

Depth (cm)

Scales

VL (a + b)

HL (d)

LL (e)

a

b

c

a/b ratio

L/W ratio

ID

1-1 VFS 1-1 Coarse silt 1-2 Coarse silt 1-3a Coarse silt 1-3a Coarse silt 1-3a Coarse silt 1-3a Coarse silt 1-3a Coarse silt 1-3a Fine silt 1-3b Coarse silt 1-3b Coarse silt 1-3b Coarse silt 1-3b Coarse silt 3-1 Coarse silt 3-1 Coarse silt 3-1 Coarse silt 3-2 Coarse silt 3-2 Coarse silt 3-4 Coarse silt 3-4 Coarse silt 6-1 VFS 6-1 Coarse silt 6-1 Coarse silt 6-1 Coarse silt 6-2 Coarse silt 6-3 Coarse silt

0–25 0–25 40–60 80–110 80–110 80–110 80–110 80–110 80–110 130–150 130–150 130–150 130–150 0–10 0–10 0–10 10–60 10–60 80–110 80–110 0–30 0–30 0–30 0–30 30–60 60–110

9 9 10 12 9 12 10 10 12 10 14 11 11 13 14 9 10 10 11 10 9 10 10 9 10 13

41.30 36.64 39.00 52.05 35.03 46.22 45.20 38.52 35.74 41.09 44.66 38.91 48.95 47.28 37.45 40.35 41.48 44.68 43.89 55.59 70.93 47.13 38.09 47.76 47.83 47.91

44.82 42.24 42.87 50.03 41.17 44.16 41.96 40.91 35.30 42.52 49.42 39.29 48.07 44.33 38.67 39.25 43.40 46.33 46.38 48.29 80.19 50.41 42.87 45.10 45.62 47.12

9.49 9.13 5.95 7.37 8.16 6.98 7.85 6.71 2.51 7.55 8.55 5.96 8.37 8.04 6.70 9.70 6.92 7.66 9.21 8.76 11.23 10.18 6.69 10.87 7.51 9.89

17.05 16.04 18.19 18.61 17.36 22.59 20.92 14.79 17.28 16.11 18.90 18.91 21.08 22.07 18.61 18.88 19.76 20.44 19.80 24.37 33.35 23.20 17.59 20.49 22.04 18.81

24.25 20.6 20.81 33.44 17.67 23.63 24.28 23.73 18.46 24.98 25.76 20.00 27.87 25.21 18.84 21.47 21.72 24.24 24.09 31.22 37.58 23.93 20.50 27.27 25.79 29.10

22.53 19.93 19.09 20.53 18.11 17.75 25.33 16.40 11.46 19.49 26.22 17.86 24.92 18.97 17.84 16.75 18.55 22.34 23.44 15.62 32.86 11.79 19.20 18.36 17.48 22.75

0.70 0.77 0.87 0.55 0.98 0.95 0.86 0.62 0.93 0.64 0.73 0.94 0.75 0.87 0.98 0.87 0.90 0.84 0.82 0.78 0.88 0.96 0.85 0.75 0.85 0.64

0.92 0.86 0.90 1.04 0.85 1.04 1.07 0.94 1.01 0.96 0.90 0.99 1.01 1.06 0.96 1.02 0.95 0.96 0.94 1.15 0.88 0.93 0.88 1.05 1.04 1.01

a b c d e f g h i j k l m n o p q r s t u v w x y z

10.65

44.37 78.33 29.09 38.4

45.41 71.45 28.16 40

0.83

0.97 1.09 1.03 0.97

Mean Athens (30) Brazil (49) China (20)

evident that the Wormsloe HL values are statistically closer to those from China, while they are significantly different from both those from Brazil and Athens. In particular, the mean VL and HL values of 49 Brazilian specimens correspond to 29.09 μm and 28.16 μm, respectively, while values from 30 Carolina Gold samples correspond to 78.33 μm and 71.45 μm, respectively. The analysis of these values suggests a closer resemblance of the Wormsloe bulliforms to those from China, rather than to the local Carolina Gold or the Brazilian varieties. From the statistical analysis on bulliforms, both the MANOVA (Pillai's Trace = 1.01, F(9, 363) = 20.50; p-value < 0.001; η2 = 0.34) and the DFA (Chi-square = 285.39; p-value < 0.001) results show that the Wormsloe and the control samples differ significantly among each other. However, as shown by the plot generated by discriminant function analysis (Fig. 5) overall the domesticated bulliform phytoliths from Wormsloe are close to the domesticated bulliforms from China and Brazil, while the Athens samples are well separated from the other three groups. In particular, 38.5% of Wormsloe domesticated bulliforms are misclassified as China bulliforms, and 25% of China bulliforms are misclassified as Wormsloe bulliforms (Table 5). Furthermore, from the post hoc analyses on bulliforms, it is clear that: 1) the Wormsloe samples differ significantly with the other samples with respect to VL; 2) the HL parameter from Wormsloe does not differ significantly to the HL parameter from China (p-value = 0.077); 3) the L/W ratio from Wormsloe is significantly different only from the L/W ratio of the Athens sample (p = 0.001), while it is significant similar to that of the Brazil (p-value = 0.14) and China (p-value = 1.0) specimens (See Appendix B for complete results).

Fig. 5. Discriminant function analysis plot showing the overall relationship between domesticated bulliform phytoliths.

Table 5 Cross validation matrix showing percentages of predicted group membership between bulliform samples. Predicted group membership

Wormsloe Athens Brazil China

4.2. Double-peaked phytoliths

Wormsloe

Athens

Brazil

China

57.7 6.7 0.0 25.0

3.8 93.3 0.0 0.0

0.0 0.0 98.0 20.0

38.5 0.0 2.0 55.0

Of the total 33 Oryza-type double-peaked cells initially selected, 13 were predicted as domesticated following the discriminant function method proposed by Zhao et al. (1998) (Fig. 6). In particular, 13 (39.39%) were predicted as domesticated, 18 (54.54%) as wild, and 2 (4.54%) as indeterminate. In addition to using 564

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Fig. 6. Domesticated double-peaked phytoliths from Wormsloe.

tolerance to salinity, drought, and flooding than Oryza sativa (Agnoun et al., 2012; Linares, 2002). Its small grains (Agnoun et al., 2012; Linares, 2002) may also explain the relatively smaller values of the Wormsloe phytoliths. At the same time, this does not exclude the possibility that the Wormsloe double-peaked phytoliths may belong to another wild species, which does not necessarily mean that cultivation was not performed; local wild rice, in fact, might have been simply cultivated for short periods of time, similarly to what Native Americans used to do (Lee et al., 2004; Nega, 2008; Yost et al., 2013). Therefore, given the difference in size between the Wormsloe double-peaked phytoliths and the reference samples, Wormsloe double-peaked phytoliths provide less substantial and conclusive evidence than bulliforms towards suggesting rice domestication at Wormsloe. When tested on known domesticated Carolina Gold specimens, the discriminant analysis method successfully predicted domestication for only 6 of the 30 samples from Athens (20%), and for 13 of the 30 samples from Sapelo (43.33%). The method assigned the rest of the reference specimens to the indeterminate class, and no domesticated rice double-peaked phytolith was assigned to the wild group. Despite having been widely used in other studies (Wu et al., 2014; Zhao, 1998; Zhao et al., 1998), therefore, in this case this method proved less successful in separating double-peaked phytoliths into domesticated and wild species. The selection of double-peaked Oryza-type phytolith at the microscope, also, proved quite challenging given their smaller and sometimes ambiguous shape, so that some misinterpretations in their identification are possible. Also, the double-peaked statistical analysis revealed that the Wormsloe phytoliths are significantly different from the other samples from Sapelo, Athens, and China (MANOVA: Pillai's Trace = 1.49, F(12, 2 264) = 21.60; p-value < 0.001; η = 0.49; DFA: Chi-square = 251.08; p-value < 0.001). As shown in the plot generated by discriminant

Table 6 Mean values (μm) of domesticated double-peaked phytoliths from Wormsloe and reference specimens. Reference

Wormsloe Sapelo Athens China

Double-peaked measurements (μm) TW

MW

CD

H

11.9 57.24 49.64 29.1

16.03 92.66 79.39 44

2.02 5.88 5.07 5.2

8.19 29.79 31.16 33.6

Zhao's discriminant analysis method, the TW, MW, CD, and H (mean of H1 and H2) measurements of double-peaked phytoliths from Wormsloe were also compared with the mean values of reference specimens (Table 6). The Wormsloe phytoliths present significantly different values from those of reference specimens, which all display larger values. From the analysis of wild double-peaked phytoliths in Gu et al. (2013), it is apparent that the Wormsloe values strongly resemble those of Oryza ridleyi, a wild rice species; however, this type is not present in the southeastern United States, thus the Wormsloe husk phytoliths may simply belong to another domesticated species, such as Oryza glaberrima. The latter – also known as ‘red rice’ or ‘African rice’ – originated in West Africa and was introduced in the southeastern United States at the end of the 17th century (Carney, 1998). Despite having been gradually replaced by the higher-yielding Oryza sativa, e.g., Carolina Gold, by the mid-18th century, Oryza glaberrima continued to be used for subsistence practices across the Americas until as late as the early 20th century (Carney, 1998, 2005), and its use at Wormsloe shall not be dismissed. African rice may have been used given its greater 565

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

wild rice relatives in the area such as Leersia oryzoides (rice cutgrass), while Zizaniopsis miliacea (giant cutgrass) and Zizania aquatica (annual wildrice) in Georgia seem to be more common in freshwater wetlands (Higinbotham et al., 2004). The presence of wild rice may represent one of the several ecological stages succeeding rice cultivation, characterized by the development of wild or volunteer rice in former rice fields. On the other hand, domesticated double-peaked phytoliths from Oryza glaberrima may have been misclassified as ‘wild’ given their smaller size. This interpretation seems also to be confirmed by the fact that no wild rice relatives, such as Zizania, Zizaniopsis, and Leersia were noticed in the area during a vegetation inventory carried out in the Fall of 2015 (K. Bradley, personal communication, November 2015). The presence of domesticated phytoliths in the top layers of the three cores may be explained by internal contamination both within the same layer, and between samples from different depths of the same core. Despite the best efforts in cleaning the coring auger after each sample extraction, in fact, it is possible that small quantities of sediments from other depths or even locations were still present. Furthermore, the vertical distribution of phytoliths may have been affected by the agricultural practice of rice cultivation itself: rice farmers had to plough the land to bring nutrients and minerals to the surface to prepare the land for seeding and cultivation, thus causing alterations in the soil's physical and chemical properties (Carney, 1993; Porcher and Judd, 2014; Smith, 2012). It is also interesting to note the geographic distribution of domesticated phytoliths across the study area. In general, Core 6 presents the fewest number of domesticated phytoliths between bulliforms and double-peaked (6 in total), while Core 1 presents most of them (22). The stratigraphy of the cores across the study area, in fact, may differ from one site to another so that 60 cm of depth on one profile may not correspond to the same depth – and thus horizon – on another core a few meters apart, i.e., it follows an elevation gradient. It is also challenging to understand what layers could correspond to the old rice field that is suggested by the presence of phytoliths. Rice cultivation may, in fact, have been performed for only short periods of time or in different agricultural seasons which could have taken place even years apart from each other. The latter aspect may also explain the relatively low amount of rice phytoliths that were found; some studies, in fact, suggests the presence of at least 5000 rice phytoliths to infer cultivation (Cao et al., 2006). However, the relatively low amounts of rice phytoliths that were recovered tend to support the idea that a domesticated species was used for cultivation at Wormsloe. Had the area been characterized, historically, by the presence of wild rice relatives, in fact, a much higher number of wild rice phytoliths would have been present, hence the possibilities of the historical presence of wild rice in the area are reduced. Therefore, given the geographic location (close to the slave settlement), and the presence of topographic features indicative of rice cultivation such as dikes and ditches, the recovery of domesticated rice phytoliths, especially the bulliform type, is to be considered indicative of cultivation at Wormsloe. Given the limited size of the area, it is assumed that the African American population living on-site performed cultivation primarily for subsistence purposes. In general, the wet cultivation of rice requires poorly drained soils with a high quantity of clay to retain water, so that a clay horizon in the soil profile may indicate the level at which the old rice field was located. From the analysis of soil samples collected at Wormsloe, it was apparent that most of them presented a rather slow permeability due to the presence of clayey and loamy sands (see Table 1). This indicates the natural compatibility of the Wormsloe soil with the wet cultivation of rice, though it is quite difficult to pinpoint specific layers in the soil profile corresponding to rice agricultural practices. The vertical distribution of domesticated phytoliths, in fact, presents some unexpected results such as the presence of domesticated phytoliths in the top layers

Fig. 7. Discriminant function analysis plot showing the overall relationship between domesticated double-peaked phytoliths. Table 7 Cross validation matrix showing percentages of predicted group membership between double-peaked samples. Predicted Group Membership

Wormsloe Sapelo Athens China

Wormsloe

Sapelo

Athens

China

100.0 0.0 0.0 0.0

0.0 70.0 33.3 0.0

0.0 30.0 56.7 15.0

0.0 0.0 10.0 85.0

function analysis (Fig. 7), Wormsloe is clearly separated from the other samples, with 100% of cases well classified (Table 7). These results were also consistent with the post hoc analyses following the MANOVA, which show that the TW, MW, CD, and H morphological parameters from Wormsloe double-peaked are significantly different from the same parameters measured on the other samples (See Appendix C for complete results). The relationship between domesticated bulliforms and doublepeaked is presented in Table 8. The mixed presence of wild and domesticated double-peaked cells may indicate the current presence of Table 8 Distribution of domesticated bulliforms and double-peaked Wormsloe phytoliths throughout the soil profiles. Wild double-peaked are also present for comparison purposes. Sample # (core – horizon)

Dom. bull.

Dom. husks

Wild husks

Indeterminate husks

Depth (cm)

1-1 1-2 1-3a 1-3b

2 1 6 4

1 4 1 3

1 2 0 3

0 0 1 1

0–25 40–60 80–110 130–150

3-1 3-2 3-3 3-4 3-5

3 2 0 2 0

1 1 0 1 1

2 4 0 3 1

0 0 0 0 0

0–10 10–60 60–80 80–110 110-down

6-1 6-2 6-3

4 1 1

0 0 0

1 0 1

0 0 0

0–30 30–60 60–110

Total

26

13

18

2

566

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Lead-210. On the other hand, excess Lead-210 coming from atmospheric deposition is also known as “unsupported” Lead-210 since once it gets deposited on the ground it is no longer replaced by Radon-222 as it decays because the latter is in the atmosphere. The radionuclide analysis on the Wormsloe samples involved the use of both Cesium-137 and Lead-210 for age estimates. Approximately 40 g. for each of the four samples was used for analysis using a gamma spectrometer, and the results are shown in Table 9. The activities of both Cesium-137 and Lead-210 are very low (0.00–0.06 dpm/g for Cesium-137, and 0.11–0.44 dpm/g for Lead-210). As suggested by Alexander (personal communication, July 2015), in fact, the amount of Cesium-137 is just above the detection limit of the machine used for processing the samples. The presence of some Cesium-137 in the top layer suggests that the top layer is younger than 1954; however, its low activity may be explained by the fact that the top layer includes a mix of sediments from the first 25 cm of the core, thus the Cesium might come from the very surface of the core in direct contact with the atmosphere. Another explanation may come from the fact that the study area is an active marsh system subject to constant erosion and redeposition of the upper soil layers so that some Cesium-137 deposits may have just been washed away over time. With regard to Lead-210, a normal detection curve should decrease constantly with depth, while results from Wormsloe show an increase in the amount at the deepest layer. This unusual behavior may be due to contamination occurred during core extraction, as the auger may have pushed down sediments from the top while being reinserted deep into the ground for additional sampling. Another possible explanation may be linked to the agricultural activity of ploughing needed to prepare the land for cultivation, which may have affected the vertical distribution of radionuclides. At any rate, the presence of Caesium-137 in the top horizon, along with the presence of low quantities of Lead-210 along the soil profile indicates that the lower layers should be around 100 years old or more. Based on these interpretations of core contamination and farming practice, the data all suggest that sediments below 50 cm are older than 100 years, thus suggesting the presence of rice cultivation at Wormsloe around the turn of the 20th century.

of the cores. It is very likely, in fact, that the stratigraphic order of phytoliths – thus their accurate interpretation – might have been affected by both the unconsolidated and liquid nature of the soil sediments recovered from the marsh, as well as by the open bucket auger employed for the extraction. 4.3. Radionuclide analysis To approximately date the Wormsloe rice phytoliths, a radionuclide analysis was performed on the four samples of Core 1. Radionuclides are radioactive elements in the Earth's atmosphere, crust, and water, which have been used in geoscience studies as a dating tool (Arnaud et al., 2006). The use of radionuclides for dating sediments is based upon the decay rates of these radioactive elements, i.e., their half-life, which is the time for half of the atoms in a given sample to decay to the next element in its radioactive decay chain. The fact that these elements are in radioactive equilibrium with their daughter products in the sediments allows us to estimate ages by measuring the decay of external fallout radionuclides coming from the atmosphere that get deposited in the sediments. Among the most commonly used elements utilized for this type of dating are Cesium-137 (137Cs) and Lead-210 (210Pb), which have a half-life of 30.3 years, and 22.3 years, respectively (U.S. Department of the Interior – U.S. Geological Survey, 2013). Cesium-137 is an artificial radionuclide that was introduced into the atmosphere during nuclear weapons testing of the 1950s and 1960s. Any presence of this element in sediments indicates that the sample is younger than 1954, when the first significant amounts of the elements entered the atmosphere (Alexander et al., 1993). Therefore, the use of Cesium-137 is limited at this time for sample dating to around 60 years. The use of Lead-210 on the other hand, can be extended to samples dating back 100–150 years. Lead-210 is part of the Uranium-238 decay series, and is the effective daughter of Ra-226. The latter decays through Radon-222, a gas, which escapes into the atmosphere, where it produces excess Lead-210, which is supplied to the land surface. This excess Lead-210 allows to estimate the age of sediments by subtracting the amount of Lead-210 normally present in the sediments from the total Lead-210 activity. Ra-226 in the soil decays to Lead-210 and “supports” the amount of Lead-210 as this further decays – hence the term “supported”

Table 9 Radionuclide analysis and graph for the Wormsloe samples. DPM/G: disintegrations per minute per gram; TOT: total; XS: excess.

Interval (cm)

Depth (cm)

Pbtot (dpm/g)

Pbtot error

Pbxs (dpm/g)

Pbxs error

Ra-226 (dpm/g)

Ra-226 error

Cs-137 (dpm/g)

Cs-137 error

0–25 40–60 80–100 130–150

12.50 50.00 90.00 140.00

1.26 1.23 1.04 1.35

0.06 0.06 0.04 0.07

0.44 0.43 0.11 0.44

0.08 0.09 0.06 0.09

0.81 0.80 0.92 0.91

0.05 0.06 0.04 0.06

0.06 0.00 0.00 0.00

0.01 0.00 0.00 0.00

567

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

5. Conclusions and future studies

In conclusion, the results of this study met the proposed objective of locating archaeobotanical remains of rice phytoliths in the area under investigation. The presence of domesticated rice phytoliths suggests that rice was cultivated at Wormsloe Historic Site for a single or multiple periods of time between the second half of the 1800s and the first years of the 1900s for subsistence purposes by the African American population living on-site. Cultivation was performed by using freshwater channeled through the existing canals to irrigate the study area. The results of this study provide more conclusive evidence towards the understanding of this aspect of Wormsloe's environmental history, thus increasing Wormsloe's historical, cultural, and archaeological significance.

The main goal of this study was the archaeobotanical investigation of Georgia coastal soils with the purpose to locate archaeological evidence related to rice cultivation at Wormsloe Historic Site. From the analysis of soil samples recovered from a tidal salt marsh, 26 bulliform cells and 13 double-peaked cells were predicted as domesticated. Multiple criteria were used to predict domestication in bulliform cells: 1) the presence of at least 9 scale-like decorations; 2) an a/b ratio of < 1; 3) a L/W ratio between 0.85 and 1.15; 4) the comparative analysis of VL and HL measurements with international reference samples. The main criteria used to predict domestication in doublepeaked cells was the discriminant analysis method developed by Zhao et al. (1998) which has been used to infer domestication in similar studies (Wu et al., 2014; Zhao, 1998; Zhao et al., 1998). Statistical analysis comparing domesticated bulliform and double-peaked phytoliths from Wormsloe with reference specimens revealed that: 1) Wormsloe bulliforms are very similar to China bulliforms, e.g., 38.5% of Wormsloe bulliforms was misclassified as China specimens; 2) the L/ W ratio from Wormsloe bulliforms is significantly similar to Brazil (pvalue = 0.14) and China (p-value = 1.0) specimens; 3) Wormsloe double-peaked phytoliths are significantly different from all the reference specimens. The presence of domesticated double-peaked and particularly of bulliform phytoliths is indicative of rice cultivation at Wormsloe. The presence of wild rice phytoliths may indicate the misclassification of domesticated African rice Oryza glaberrima which presents smaller grains than Oryza sativa, as no wild rice relatives were found in the area. Furthermore, from the preliminary radionuclide analysis on one core which analyzed the amounts of Caesium-137 and Lead-210, the results suggest that the core samples, thus the phytoliths, may be around 100 years or older. The claim of rice cultivation at Wormsloe is also supported by the following: 1) the presence of earthworks such as dikes and ditches which would have enclosed the field when flooded; 2) the proximity of the study area to the former slave settlement; 3) the poorly drained soil of the marsh, which would have been an appropriate environment for the wet cultivation of rice, given its soil moisture retention abilities. To improve interpretation, future studies may benefit from the use of more sophisticated core extraction techniques that would maintain the stratigraphic profile intact, such as freeze corers using liquid nitrogen. Future studies might also improve the age estimates by employing a finer sampling resolution along the soil core or by extending the radionuclide analysis to the rest of the cores. Moreover, future research on rice cultivation at Wormsloe would benefit from the application of archaeobotanical analysis on other similar areas on the property to increase the understanding of the extent of rice cultivation practices at Wormsloe, i.e., Areas 2 and 3 in Appendix D. Finally, future studies might involve a species investigation to properly understand which rice variety was cultivated at Wormsloe, thus casting further light on African American subsistence practices in the Lowcountry.

Acknowledgements This work was funded by the Wormsloe Foundation and the Wormsloe Institute for Environmental History (grant number 20132015) through the Wormsloe Fellowship program. This work was also supported by the Center for Geospatial Research at the University of Georgia through Center's funds available to staff for research purposes. Special thanks go to the following people: Sarah Ross and Craig and Diana Barrow of the Wormsloe Institute for Environmental History and the Wormsloe Foundation for providing technical and logistic support to perform research on-site; David Leigh for providing technical and logistic support to perform phytolith extraction in the Geomorphology Lab at the University of Georgia; David Porinchu for providing technical and logistic support to perform phytolith analysis at the microscope at the Environmental Change Lab at the University of Georgia; Clark Alexander of the Skidaway Institute of Oceanography at the University of Georgia for performing the radionuclide analysis and providing help in the interpretation of the results; Roberta Salmi of the University of Georgia for performing statistical analysis and providing help in the interpretation of the results; Hayden Smith of the College of Charleston, Andrew Agha and Nicole Isenbarger of Archaeological Research Collective, Richard Porcher, and Roger Pinckney for providing assistance in the inspection of historical rice fields and interpretation of Lowcountry rice cultivation history; Susannah Chapman of the University of Georgia and Stanley Walker for providing assistance in the collection of rice plants needed for reference; Wormsloe Fellow Holly Campbell and Larry Morris of the Warnell School of Forestry at the University of Georgia for their advice on soil analysis; Wormsloe Fellow Paul Cady for providing assistance in the interpretation of Wormsloe's environmental history; Marguerite Madden of the Center for Geospatial Research for providing logistic and organizational assistance for the visit of Liovando Costa to the Department of Geography at the University of Georgia; Giles Allard of the Department of Geology at the University of Georgia for helping in the translation of Portuguese articles into English; Paola Cazzini for helping with the layout of images and tables; and the two anonymous reviewers for providing useful comments on the manuscript.

568

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Appendix A

Brazilian rice bulliforms. Scale bar is 50 μm. Appendix B. Bulliform phytoliths B.1. Multivariate analysis

Multivariate testsa Effect

Value

F

Hypothesis df

Error df Sig.

Partial eta squared

Noncent. Parameter

Observed powerd

Intercept Pillai's Trace Wilks' Lambda Hotelling's Trace Roy's Largest Root Sample Pillai's Trace

0.998 0.002 516.507 516.507

20,488.108b 20,488.108b 20,488.108b 20,488.108b

3.000 3.000 3.000 3.000

119.000 119.000 119.000 119.000

0.998 0.998 0.998 0.998

61,464.323 61,464.323 61,464.323 61,464.323

1.000 1.000 1.000 1.000

1.011

20.500

9.000

363.000 0.000 0.337

184.503

1.000

569

0.000 0.000 0.000 0.000

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Wilks' Lambda Hotelling's Trace Roy's Largest Root a b c d

0.094 8.563 8.430

53.002 111.952 340.028c

9.000 9.000 3.000

289.765 0.000 0.546 353.000 0.000 0.741 121.000 0.000 0.894

348.352 1007.567 1020.085

1.000 1.000 1.000

Design: Intercept + Sample. Exact statistic. The statistic is an upper bound on F that yields a lower bound on the significance level. Computed using alpha = 0.05.

B.2. Post hoc Analysis Since the assumption of equal variance was only met for L/W, the Games-Howell Post Hoc was used for VL and HL, while the Tukey HSD was used for L/W.

Multiple comparisons Dependent variable

VL

Tukey HSD

Mean difference (I–J)

Wormsloe

Sapelo

Athens

China

Games-Howell

Wormsloe

Sapelo

Athens

China

HL

Tukey HSD

Wormsloe

Sapelo

Athens

China

Games-Howell

Wormsloe

Sapelo

Athens

Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China

− 33.9596923⁎ 15.2763893⁎ 5.9223077⁎ 33.9596923⁎ 49.2360816⁎ 39.8820000⁎ − 15.2763893⁎ − 49.2360816⁎ − 9.3540816⁎ − 5.9223077⁎ − 39.8820000⁎ 9.3540816⁎ − 33.9596923⁎ 15.2763893⁎ 5.9223077⁎ 33.9596923⁎ 49.2360816⁎ 39.8820000⁎ − 15.2763893⁎ − 49.2360816⁎ − 9.3540816⁎ − 5.9223077⁎ − 39.8820000⁎ 9.3540816⁎ − 26.0430256⁎ 17.2449608⁎ 5.4223077⁎ 26.0430256⁎ 43.2879864⁎ 31.4653333⁎ − 17.2449608⁎ − 43.2879864⁎ − 11.8226531⁎ − 5.4223077⁎ − 31.4653333⁎ 11.8226531⁎ − 26.0430256⁎ 17.2449608⁎ 5.4223077 26.0430256⁎ 43.2879864⁎ 31.4653333⁎ − 17.2449608⁎ − 43.2879864⁎ − 11.8226531⁎ 570

Std. error

1.94051559 1.75718036 2.15400977 1.94051559 1.67890183 2.09064263 1.75718036 1.67890183 1.92168381 2.15400977 2.09064263 1.92168381 2.58732521 1.56054328 1.78988154 2.58732521 2.18587222 2.35508342 1.56054328 2.18587222 1.13466362 1.78988154 2.35508342 1.13466362 1.73641965 1.57236691 1.92745933 1.73641965 1.50232141 1.87075690 1.57236691 1.50232141 1.71956852 1.92745933 1.87075690 1.71956852 2.19059058 1.64612326 2.18703965 2.19059058 1.60417476 2.15564313 1.64612326 1.60417476 1.59932235

Sig.

0.000 0.000 0.034 0.000 0.000 0.000 0.000 0.000 0.000 0.034 0.000 0.000 0.000 0.000 0.010 0.000 0.000 0.000 0.000 0.000 0.000 0.010 0.000 0.000 0.000 0.000 0.029 0.000 0.000 0.000 0.000 0.000 0.000 0.029 0.000 0.000 0.000 0.000 0.077 0.000 0.000 0.000 0.000 0.000 0.000

95% Confidence interval Lower bound

Upper bound

− 39.0149177 10.6987694 0.3109100 28.9044669 44.8623847 34.4356796 − 19.8540092 − 53.6097786 − 14.3602484 − 11.5337054 − 45.3283204 4.3479148 − 40.8349794 11.0416420 1.1339733 27.0844052 43.3173340 33.5727045 − 19.5111367 − 55.1548292 − 12.4450253 − 10.7106421 − 46.1912955 6.2631380 − 30.5665619 13.1487973 0.4010951 21.5194894 39.3742980 26.5918359 − 21.3411242 − 47.2016748 − 16.3022906 − 10.4435203 − 36.3388308 7.3430156 − 31.8519461 12.7689563 − 0.4196337 20.2341052 38.9623498 25.7193829 − 21.7209652 − 47.6136230 − 16.2472690

−28.9044669 19.8540092 11.5337054 39.0149177 53.6097786 45.3283204 −10.6987694 −44.8623847 −4.3479148 −0.3109100 −34.4356796 14.3602484 −27.0844052 19.5111367 10.7106421 40.8349794 55.1548292 46.1912955 −11.0416420 −43.3173340 −6.2631380 −1.1339733 −33.5727045 12.4450253 −21.5194894 21.3411242 10.4435203 30.5665619 47.2016748 36.3388308 −13.1487973 −39.3742980 −7.3430156 −0.4010951 −26.5918359 16.3022906 −20.2341052 21.7209652 11.2642491 31.8519461 47.6136230 37.2112838 −12.7689563 −38.9623498 −7.3980371

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

China

L/W

Tukey HSD

Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens

Wormsloe

Sapelo

Athens

China

Games-Howell

Wormsloe

Sapelo

Athens

China

− 5.4223077 − 31.4653333⁎ 11.8226531⁎ − 0.1193168⁎ − 0.0598196 0.0002568 0.1193168⁎ 0.0594972 0.1195736⁎ 0.0598196 − 0.0594972 0.0600764 − 0.0002568 − 0.1195736⁎ − 0.0600764 − 0.1193168⁎ − 0.0598196⁎ 0.0002568 0.1193168⁎ 0.0594972 0.1195736⁎ 0.0598196⁎ − 0.0594972 0.0600764 − 0.0002568 − 0.1195736⁎ − 0.0600764

2.18703965 2.15564313 1.59932235 0.03068520 0.02778614 0.03406117 0.03068520 0.02654833 0.03305915 0.02778614 0.02654833 0.03038742 0.03406117 0.03305915 0.03038742 0.02630953 0.02192730 0.03561896 0.02630953 0.02733750 0.03918218 0.02192730 0.02733750 0.03638487 0.03561896 0.03918218 0.03638487

0.077 0.000 0.000 0.001 0.143 1.000 0.001 0.118 0.002 0.143 0.118 0.202 1.000 0.002 0.202 0.000 0.039 1.000 0.000 0.142 0.021 0.039 0.142 0.367 1.000 0.021 0.367

− 11.2642491 − 37.2112838 7.3980371 − 0.1992546 − 0.1322051 − 0.0884757 0.0393789 − 0.0096637 0.0334514 − 0.0125659 − 0.1286581 − 0.0190857 − 0.0889894 − 0.2056958 − 0.1392385 − 0.1892747 − 0.1175309 − 0.0973084 0.0493588 − 0.0127672 0.0139601 0.0021082 − 0.1317616 − 0.0390291 − 0.0978220 − 0.2251871 − 0.1591819

0.4196337 −25.7193829 16.2472690 −0.0393789 0.0125659 0.0889894 0.1992546 0.1286581 0.2056958 0.1322051 0.0096637 0.1392385 0.0884757 −0.0334514 0.0190857 −0.0493588 −0.0021082 0.0978220 0.1892747 0.1317616 0.2251871 0.1175309 0.0127672 0.1591819 0.0973084 −0.0139601 0.0390291

Based on observed means. The error term is Mean Square (Error) = 0.013. ⁎ The mean difference is significant at the 0.05 level.

Appendix C. Double-peaked phytoliths C.1. Multivariate analysis

Multivariate Testsa Effect

Value

F

Hypothesis df Error df Sig.

Partial Eta Squared Noncent. Parameter Observed Powerd

Intercept Pillai's Trace Wilks' Lambda Hotelling's Trace Roy's Largest Root Sample Pillai's Trace Wilks' Lambda Hotelling's Trace Roy's Largest Root

0.972 0.028 34.449 34.449 1.486 0.058 6.912 5.041

740.648b 740.648b 740.648b 740.648b 21.599 36.832 48.765 110.895c

4.000 4.000 4.000 4.000 12.000 12.000 12.000 4.000

0.972 0.972 0.972 0.972 0.495 0.614 0.697 0.834

a b c d

86.000 86.000 86.000 86.000 264.000 227.826 254.000 88.000

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

2962.592 2962.592 2962.592 2962.592 259.188 361.893 585.180 443.582

1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000

Design: Intercept + Sample. Exact Statistic. The statistic is an upper bound on F that yields a lower bound on the significance level. Computed using alpha = 0.05.

C.2. Post hoc analysis Since none of the samples met the assumption of equal variance, Games-Howell was used for all parameters.

Multiple comparisons Dependent variable TW

Wormsloe

Sapelo

Sapelo Athens China Wormsloe

Mean difference (I–J) − 45.3429487⁎ − 37.7412821⁎ − 17.1696154⁎ 45.3429487⁎

571

Std. error

Sig.

2.81919817 2.51039040 2.20207883 2.81919817

0.000 0.000 0.000 0.000

95% Confidence interval Lower bound Upper bound −52.9178205 −37.7680769 −44.4737944 −31.0087697 −23.1670295 −11.1722013 37.7680769 52.9178205

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Athens

China

MW

Wormsloe

Sapelo

Athens

China

CD

Wormsloe

Sapelo

Athens

China

H

Wormsloe

Sapelo

Athens

China

Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens Sapelo Athens China Wormsloe Athens China Wormsloe Sapelo China Wormsloe Sapelo Athens

7.6016667 28.1733333⁎ 37.7412821⁎ − 7.6016667 20.5716667⁎ 17.1696154⁎ − 28.1733333⁎ − 20.5716667⁎ − 76.6291538⁎ − 63.3644872⁎ − 27.9911538⁎ 76.6291538⁎ 13.2646667⁎ 48.6380000⁎ 63.3644872⁎ − 13.2646667⁎ 35.3733333⁎ 27.9911538⁎ − 48.6380000⁎ − 35.3733333⁎ − 3.8557436⁎ − 3.6770769⁎ − 3.1930769⁎ 3.8557436⁎ 0.1786667 0.6626667 3.6770769⁎ − 0.1786667 0.4840000 3.1930769⁎ − 0.6626667 − 0.4840000 − 25.8772179⁎ − 27.2503846⁎ − 29.6353846⁎ 25.8772179⁎ − 1.3731667 − 3.7581667 27.2503846⁎ 1.3731667 − 2.3850000 29.6353846⁎ 3.7581667 2.3850000

3.41708427 3.19742961 2.51039040 3.41708427 2.92877752 2.20207883 3.19742961 2.92877752 3.94253617 3.66713003 3.01977706 3.94253617 4.81476329 4.34202215 3.66713003 4.81476329 4.09358127 3.01977706 4.34202215 4.09358127 0.50591882 0.38512573 0.39375132 0.50591882 0.55388171 0.55991359 0.38512573 0.55388171 0.45373031 0.39375132 0.55991359 0.45373031 0.94148295 0.89571213 1.41355055 0.94148295 1.26014537 1.66846974 0.89571213 1.26014537 1.64307680 1.41355055 1.66846974 1.64307680

Based on observed means. The error term is Mean Square (Error) = 23.927. ⁎ The mean difference is significant at the 0.05 level.

572

0.129 0.000 0.000 0.129 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.038 0.000 0.000 0.038 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.988 0.640 0.000 0.988 0.711 0.000 0.640 0.711 0.000 0.000 0.000 0.000 0.697 0.129 0.000 0.697 0.477 0.000 0.129 0.477

−1.4421743 19.6609381 31.0087697 −16.6455077 12.7766591 11.1722013 −36.6857286 −28.3666743 −87.2096625 −73.1953325 −36.1960829 66.0486451 0.5260308 37.0747163 53.5336419 −26.0033026 24.4777762 19.7862248 −60.2012837 −46.2688905 −5.2132646 −4.7085664 −4.2630710 2.4982226 −1.2917234 −0.8281872 2.6455874 −1.6490567 −0.7261588 2.1230828 −2.1535206 −1.6941588 −28.4266948 −29.6745525 −33.5922706 23.3277411 −4.7066641 −8.2602239 24.8262167 −1.9603307 −6.8276626 25.6784986 −0.7438905 −2.0576626

16.6455077 36.6857286 44.4737944 1.4421743 28.3666743 23.1670295 −19.6609381 −12.7766591 −66.0486451 −53.5336419 −19.7862248 87.2096625 26.0033026 60.2012837 73.1953325 −0.5260308 46.2688905 36.1960829 −37.0747163 −24.4777762 −2.4982226 −2.6455874 −2.1230828 5.2132646 1.6490567 2.1535206 4.7085664 1.2917234 1.6941588 4.2630710 0.8281872 0.7261588 −23.3277411 −24.8262167 −25.6784986 28.4266948 1.9603307 0.7438905 29.6745525 4.7066641 2.0576626 33.5922706 8.2602239 6.8276626

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

Appendix D

LiDAR elevation map of Wormsloe showing the areas deemed suitable for rice cultivation. AREA 1: study area; AREA 2 and AREA 3: potential areas for future research. (Source for LiDAR dataset: Center for Geospatial Research, University of Georgia)

domesticated rice species in East Asia. Quat. Int. 287, 141–148. http://dx.doi.org/10. 1016/j.quaint.2012.02.013. Higinbotham, C.B., Alber, M., Chalmers, A.G., 2004. Analysis of tidal marsh vegetation patterns in two Georgia Estuaries using aerial photography and GIS. Estuaries 27 (4), 670–683. Huang, F., Zhang, M., 2000. Pollen and phytolith evidence for rice cultivation during the Neolithic at Longqiuzhuang, eastern Jianghuai, China. Veg. Hist. Archaeobotany 9 (3), 161–168. Itzstein-Davey, F., Taylor, D., Dodson, J., Atahan, P., Zheng, H., 2007. Wild and domesticated forms of rice (Oryza sp.) in early agriculture at Qingpu, lower Yangtze, China: evidence from phytoliths. J. Archaeol. Sci. 34 (12), 2101–2108. Jiang, Q., 1995. Searching for evidence of early rice agriculture at prehistoric sites in China through phytolith analysis: an example from central China. Rev. Palaeobot. Palynol. 89 (3), 481–485. Jiang, L., Liu, L., 2006. New evidence for the origins of sedentism and rice domestication in the Lower Yangzi River, China. Antiquity 80 (308), 355–361. Jiang, Q., Piperno, D., 1994. Phytolith analysis of an archaeological site (Longshan Period) in Zhumadian City, Henan Province, China: paleoenvironmental and cultural implications. Geoarchaeology 9 (5), 409–417. Jin, G., Wu, W., Zhang, K., Wang, Z., Wu, X., 2014. 8000-year old rice remains from the north edge of the Shandong Highlands, East China. J. Archaeol. Sci. 51, 34–42. Kelso, W.M., 1979. Captain Jones's Wormslow. The University of Geor-gia Press, Athens. Lee, G.A., Davis, A.M., Smith, D.G., McAndrews, J.H., 2004. Identifying fossil wild rice (Zizania) pollen from Cootes Paradise, Ontario: a new approach using scanning electron microscopy. J. Archaeol. Sci. 31 (4), 411–421. http://dx.doi.org/10.1016/j. jas.2003.09.009. Linares, O., 2002. African rice (Oryza glaberrima): history and future potential. Proc. Natl. Acad. Sci. 99 (25), 16360–16365. Lu, H., Liu, K.-B., 2003a. Phytoliths of common grasses in the coastal environments of southeastern USA. Estuar. Coast. Shelf Sci. 58 (3), 587–600. Lu, H., Liu, K.B., 2003b. Morphological variations of lobate phytoliths from grasses in China and the south-eastern United States. Divers. Distrib. 9 (1), 73–87. Lu, H., Wu, N., Liu, B., 1997. Recognition of rice phytoliths. In: 1° Encuentro Europeo sobre el estudio de fitolitos. Monografías del Centro de Ciencias Medioambientales, pp. 159–174.

References Agnoun, Y., Biaou, S.S.H., Sie', M., Vodouhe', R.S., Ahanche'de', A., 2012. The African rice Oryza glaberrima Steud: knowledge distribution and prospects. Int. J. Biol. 4 (3), 158–180. http://dx.doi.org/10.5539/ijb.v4n3p158. Alexander, C.R., Smith, R.G., Calder, F.D., Schropp, S.J., Windom, H.L., 1993. The historical record of metal enrichment in two Florida Estuaries. Estuaries 16 (3), 627–637. Arnaud, F., Magand, O., Chapron, E., Bertrand, S., Boës, X., Charlet, F., Mélières, M.-A., 2006. Radionuclide dating (210 Pb, 137 Cs, 241 Am) of recent lake sediments in a highly active geodynamic setting (Lakes Puyehue and Icalma—Chilean Lake District). Sci. Total Environ. 366 (2), 837–850. Bragg, W.H., 1999. De Renne: Three Generations of a Georgia Family. vol. 21 University of Georgia Press. Cao, Z., Ding, J., Hu, Z., Knicker, H., Kögel-Knabner, I., Yang, L., ... Dong, Y., 2006. Ancient paddy soils from the Neolithic age in China's Yangtze River Delta. Naturwissenschaften 93 (5), 232–236. Carney, J., 1993. From hands to tutors: African expertise in the South Carolina rice economy. Agric. Hist. 1–30. Carney, J., 1998. The role of African rice and slaves in the history of rice cultivation in the Americas. Hum. Ecol. 26 (4), 525–545. Carney, J., 2005. Rice and memory in the age of enslavement: Atlantic passages to Suriname. Slavery and Abolition 26 (3), 325–347. Chun-Hai, L., Zhang, G.-Y., Lin-Zhang, Y., Xian-Gui, L., Zheng-Yi, H., Yuan-Hua, D., ... JinLong, D., 2007. Pollen and phytolith analyses of ancient paddy fields at Chuodun site, the Yangtze River Delta. Pedosphere 17 (2), 209–218. Costa, L.M., Moreau, A.M.S., Moreau, M.S., 2010. Estabilidade da sílica biogênica extraída de capim Jaraguá (Hyparrhenia rufa) em solução de NaOH. Quim Nova 33 (8), 1658–1663. Coulter, E.M., 1955. Wormsloe. Two Centuries of a Georgia family.[With Plates, Including Portraits.]. University of Georgia Press. Fujiwara, H., Kaner, S., 1993. Research into the history of rice cultivation using plant opal analysis. MASCA Res. Papers Sci. Archaeol. 10, 147–158. Gu, Y.S., Zhao, Z.J., Pearsall, D.M., 2013. Phytolith morphology research on wild and

573

Journal of Archaeological Science: Reports 14 (2017) 557–574

A. Pasqua et al.

systems, wild rice stands and macro-regional patterns. J. Archaeol. Sci. 51, 43–53. Wu, Y., Jiang, L., Zheng, Y., Wang, C., Zhao, Z., 2014. Morphological trend analysis of rice phytolith during the early Neolithic in the Lower Yangtze. J. Archaeol. Sci. 49, 326–331. Yost, C.L., Blinnikov, M.S., Julius, M.L., 2013. Detecting ancient wild rice (Zizania spp. L.) using phytoliths: a taphonomic study of modern wild rice in Minnesota (USA) lake sediments. J. Paleolimnol. 49 (2), 221–236. http://dx.doi.org/10.1007/s10933-0129670-x. Zhang, J., Lu, H., Wu, N., Li, F., Yang, X., Wang, W., ... Zhang, X., 2010. Phytolith evidence for rice cultivation and spread in Mid-late Neolithic archaeological sites in central North China. Boreas 39 (3), 592–602. Zhao, Z., 1998. The Middle Yangtze region in China is one place where rice was domesticated: phytolith evidence from the Diaotonghuan Cave, Northern Jiangxi. Antiquity 72 (278), 885–897. Zhao, Z., Piperno, D.R., 2000. Late Pleistocene/Holocene environments in the middle Yangtze River valley, China and rice (Oryza sativa L.) domestication: the phytolith evidence. Geoarchaeology 15 (2), 203–222. Zhao, Z., Pearsall, D.M., Benfer, R.A., Piperno, D.R., 1998. Distinguishing rice (Oryza sativa poaceae) from wild Oryza species through phytolith analysis, II finalized method. Econ. Bot. 52 (2), 134–145. Zheng, Y., Jiang, L., 2007. Remains of ancient rice unearthed from the Shangshan site and their significance. Archaeology 9, 19–25. Zheng, Y., Dong, Y., Matsui, A., Udatsu, T., Fujiwara, H., 2003a. Molecular genetic basis of determining subspecies of ancient rice using the shape of phytoliths. J. Archaeol. Sci. 30 (10), 1215–1221. Zheng, Y., Matsui, A., Fujiwara, H., 2003b. Phytoliths of rice detected in the Neolithic sites in the valley of the Taihu Lake in China. Environ. Archaeol. 8 (2), 177–183.

Lu, H., Liu, Z., Wu, N., Berne, S., Saito, Y., Liu, B., Wang, L., 2002. Rice domestication and climatic change: phytolith evidence from East China. Boreas 31 (4), 378–385. Lu, H., Wu, N.-Q., Yang, X.-D., Jiang, H., Liu, K.-b., and Liu, T.-S., 2006. Phytoliths as quantitative indicators for the reconstruction of past environmental conditions in China I: phytolith-based transfer functions. Quat. Sci. Rev. 25 (9), 945–959. Nega, T.H., 2008. Saving wild rice: the rise and fall of the Nett Lake dam. Environ. Hist. 14, 5–39. http://dx.doi.org/10.3197/096734008X271841. Pearsall, D.M., Piperno, D.R., Dinan, E.H., Umlauf, M., Zhao, Z., Benfer Jr., R.A., 1995. Distinguishing rice (Oryza sativa Poaceae) from wild Oryza species through phytolith analysis: results of preliminary research. Econ. Bot. 49 (2), 183–196. Piperno, D., 1988. Phytolith Analysis: An Archaeological and Geological Perspective. Academic Press, San Diego, Etc. Porcher Jr., R.D., Judd, W.R., 2014. The Market Preparation of Carolina Rice An Illustrated History of Innovations in the Lowcountry Rice Kingdom. The University of South Carolina Press, University of South Carolina. Qiu, Z., Jiang, H., Ding, J., Hu, Y., Shang, X., 2014. Pollen and Phytolith Evidence for Rice Cultivation and Vegetation Change During the Mid-late Holocene at the Jiangli Site, Suzhou, East China. Smith, H.R., 2012. Rich Swamps and Rice Grounds: The Specialization of Inland Rice Culture in the South Carolina Lowcountry, 1670–1861. University of Georgia. Swanson, D.A., 2012. Remaking Wormsloe Plantation: The Environmental History of a Lowcountry Landscape. University of Georgia Press. U.S. Department of the Interior - U.S. Geological Survey, 2013. 210Pb (lead 210) Dating. http://gec.cr.usgs.gov/archive/lacs/lead.htm (last accessed: 25th October, 2016). Watanabe, N., 1968. Spodographic evidence of rice from prehistoric Japan. J. Fac. Sci. Univ. Tokyo 5 (3), 217. Weisskopf, A., Harvey, E., Kingwell-Banham, E., Kajale, M., Mohanty, R., Fuller, D.Q., 2013. Archaeobotanical implications of phytolith assemblages from cultivated rice

574