Introduction

The Liangzhu culture (5,300–4,300 BP) was mainly spread across the region surrounding Taihu in northern Zhejiang Province and southern Jiangsu Province. Archaeological findings indicate that the Liangzhu culture entered the stage of an early state1,2,3,4,5,6. As a subsistence economy, the basis of social and cultural development, is important for understanding ancient societies, scholars have conducted comprehensive studies on the subsistence of the Liangzhu culture and provided insights into its collapse based on this perspective7.

Until date, many multidisciplinary studies have been conducted on the Liangzhu culture, and consequently Yuan et al. 7 found that the areas where the Liangzhu culture was distributed can be divided into three regions: the Taihu Plain area, the east coast of Zhejiang Province, and the eastern Jianghuai region8,9,10. Archaeobotanical research, which can provide valuable information for understanding plant food sources, agricultural practices, and subsistence economies, has been conducted in all three regions; however, studies have mainly focused on the core area of the Liangzhu culture—the Taihu Plain area—where researches on macro- and micro-botanical remains have been conducted on many archaeological sites11. Compared with the core area, archaeobotanical studies have been conducted on only a few Liangzhu culture archaeological sites in the non-core area, including the eastern Jianghuai region. Thus, gaining a deeper understanding of the Liangzhu culture subsistence requires conducting additional multidisciplinary research, including archaeobotany, in the non-core areas of the Liangzhu culture. This can help clarify plant resource utilization and subsistence strategies in these areas.

Archaeobotanical research conducted in the eastern Jianghuai region focuses on two sites—Jiangzhuang and Kaizhuang (see Fig. 1)—both of which are based on plant macroremains that reveal the plant resource utilization and agricultural practices9,12. However, investigations of the ancient diet based only on macroplant remains may be biased. Some plant species uncovered from archaeological sites could certainly be edible, but it can not be confirmed whether they were consumed13. Moreover, roots and tubers, which were directly consumed, have a lower chance of charring and, consequently, were less frequently preserved and uncovered at archaeological sites. Further evidence obtained from other analytical methods is required to more accurately reconstruct ancient plant-based consumption and subsistence.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Locations of Jiangzhuang and archaeological sites of Liangzhu culture mentioned in this study. 1, Jiangzhuang; 2, Kaizhuang; 3, Yaoshan; 4, Zhumucun; 5, Chizhongsi; 6, Guangfulin; 7, Bianjiashan; 8, Maoshan; 9, Liangzhu ancient city; 10, Yaojiashan; 11, Cihu; 12, Tashan.

Dental calculus is a mineralized plaque deposited on the tooth surface. It is formed throughout an individual’s life14. Dental calculus residue analysis involves a variety of research methods, such as plant remains, stable isotopic, ancient DNA, and microbiological and biochemical analyses, which provide additional avenues for investigating food structure and the subsistence economy. Starch granule analysis of dental calculi is one of the most widely used methods15. Starch granules and other micro-botanical remains entrapped within the calculus provide the most direct evidence for understanding human plant-based food consumption16,17,18,19,20,21,22,23,24,25,26.

The Jiangzhuang site, a large-scale Liangzhu cultural settlement uncovered north of the Yangtze River, is far from the core area of the Liangzhu culture and is located in the eastern Jianghuai region27. In this study, a starch granule analysis was conducted on human dental calculus of the Jiangzhuang site to reconstruct plant-based consumption of local populations. This study will be helpful to enhance the understanding of plant food sources and the subsistence of large settlements in the marginal area of the Liangzhu culture; it will also provide new evidence of subsistence and social conditions in different regions of the Liangzhu culture distribution area.

The Jiangzhuang site is located at the junction of the cities of Xinghua and Dongtai in Jiangsu Province, China. The site was divided into eastern and western areas with the Taidong River as the boundary (Fig. 1). The accumulation in the eastern area of this site primarily belongs to the Tang and Song dynasties, whereas the accumulation in the western area of this site mainly belongs to the Liangzhu culture. From October 2011 to November 2015, an excavation team from the Nanjing Museum conducted archaeological excavations in Jiangzhuang. The excavation primarily focused on the western area, with a total area of 3500 square meters27 (Fig. 2a). Numerous archaeological remains belonging to Liangzhu cultural period were unearthed, including 284 tombs; eight house foundations; more than 100 ash pits and wells; and nearly 1200 pieces of jade, stone, pottery, and bone objects. The Liangzhu culture cemetery in the western area of the Jiangzhuang site is particularly important. High-class tombs with jade burial objects, such as jade bi, cong, and low-class tombs without jade objects were uncovered in this cemetery (Fig. 2b, c), which reflects marked differentiation in the social hierarchy. The Liangzhu culture cemetery at the Jiangzhuang site is currently the largest and densest cemetery discovered outside the core area of the Liangzhu culture, and most of the buried skeletons are well preserved27 (Fig. 2b, c). Over 30 radiocarbon dating results from the Jiangzhuang site suggest that it dates from between 3000 and 2500 BC, corresponding to the middle to late the Liangzhu culture28. The Jiangzhuang site is the first large-scale Liangzhu cultural settlement discovered in the north of the Yangtze River, breaking the traditional view that the northern boundary of the Liangzhu culture could not reach the Yangtze River. This provides new materials for a comprehensive and in-depth study of the Liangzhu civilization and society, filling a gap in archaeological discovery of the Liangzhu culture in the north of the Yangtze River area27.

Fig. 2: Locations of archaeological sites of Liangzhu culture mentioned in this study.
Fig. 2: Locations of archaeological sites of Liangzhu culture mentioned in this study.The alternative text for this image may have been generated using AI.
Full size image

Jiangzhuang site: (a) aerial photograph, excavation area (part); (b) high-class tomb M45 (some jade objects as burial objects can be seen in the picture) (c) low-class tomb M43 (only some pottery vessels as burial objects can be seen in the picture).

Visible dental calculi on human teeth belonging to the Liangzhu period from the Jiangzhuang site were chosen for starch granule analysis. In this study, 45 dental calculus samples were collected from nine individuals for starch granule analysis. Table 1 shows the details of the samples.

Table 1 Human dental calculus samples of Liangzhu culture period from Jiangzhuang site
Full size table

Methods

Starch granule extraction

Here, we consulted the methods used by Tao et al. 18 to extract starch granules from human dental calculus. These teeth were brushed with new soft toothbrushes and ultrapure water to remove adherent soil and other particles. A scalpel with detachable blades was used to scrape the visible calculus from the teeth. All laboratory extractions and microscopic observations were performed at the Archaeobotany Laboratory, School of Archaeology and Cultural Heritage, Zhengzhou University.

The calculus samples were ground into powder using an agate mortar and pestle and transferred to new centrifuge tubes. Although EDTA is the better choice for pre-treatment of calculus samples as mentioned in the study by Tromp et al. 29, 10% HCL solution has not destroyed fine plant remains such as starch grains in our previous analyses so far18, that it has not adversely affected the results, and that we would consider using EDTA in the future. 1–1.5 ml 10% HCL was added and the tubes were left for several hours until there were no additional bubbles. Distilled water was then added, and the solution was centrifuged at 3000 rpm for 10 min. The supernatant was removed, and the residue at the bottom was retained. This step was repeated two or three times. A 5% solution of Calgon was added to the tubes to deflocculate the calculus and ease the dispersion for at least 24 h. Finally, after three washes with distilled water, CsCl solution of 1.8 g/cm3 density was added to the samples to isolate the starch granules from the solution. The mixture was then centrifuged at 3000 rpm for 10 minutes, and the supernatant was transferred to a new centrifuge tube. Distilled water was added to a new centrifuge tube and centrifuged at 3000 rpm for 10 min. This step was repeated two or three times. The supernatant was removed and the residue was precipitated at the bottom. A mixture of glycerol and ultrapure water (1:1) was used to mount the residues on slides, which were sealed with nail polish for examination under polarized and transmitted light using an Olympus BX53 microscope with Olympus DP80 camera system and Olympus cellSens Standard image software.

Moreover, to prevent contamination, all the implements were cleaned in an ultrasonic water bath before use. However, evaluation of contamination in the laboratory environment or during the ancient starch grain extraction process was necessary. Two blank slides were placed on the laboratory table to check for air contamination. After the experiment was completed, a mixture of glycerol and ultrapure water (1:1) was used to mount two blank slides. To check for contamination during the extraction process, two empty centrifuge tubes were used as control samples. Starch granules were extracted from the two control samples and human dental calculus following the same procedure. The mixture of glycerol and ultrapure water (1:1) was used to mount the “residues” extracted from the two controlled samples on the blank slides30. The four slides were sealed with nail polish and examined under polarized and transmitted light using the same microscope system as observing archaeological samples. No starch granules were found on these slides, suggesting that no contamination occurred in the laboratory environment or during the ancient starch grain extraction process.

Starch granule identification

Starch granules of different plants are characterized by their overall shape, type, size, position, form of the hilum, fissure, presence or absence of lamellae, and the appearance and projection of Maltese crosses under a polarized microscope31,32,33. Therefore, the comprehensive morphological characteristics of starch granules can serve as an important basis for their identification. The morphological data of modern starch granules form the basis for the analysis of ancient starch granules. To identify ancient starch granules extracted from the calculus, we referred to over 40 species from 18 genera of modern plants (Fig. 3). In addition to comparative plant samples collected by the authors, many studies about starch granule morphology of modern plants are also consulted to any further identification34,35,36,37,38,39,40,41,42,43,44,45,46.

Fig. 3: Examples of starch granules from modern plants.
Fig. 3: Examples of starch granules from modern plants.The alternative text for this image may have been generated using AI.
Full size image

a-a’ Oryza sativa. Range 2.47-6.27 μm, mean 3.91 μm (n = 100); b-b’ Setaria italica. Range 6.23-19.35 μm, mean 9.32 μm (n = 55); c-c’ Panicum miliaceum. Range 5.96-10.99 μm, mean 8.42 μm (n = 46); (d-d’) Sorghum bicolor. Range 7.77- 17.1 μm, mean 12.26 μm (n = 100); e-e’ Coix lacryma-jobi. Range 3.44- 11.26 μm, mean 7.73 μm (n = 100); (f-f’)Triticum aestivum. Range 5.77–31.24 μm, mean 14.98 μm (n = 82); g-g’ Hordeum vulgare. Range 11.29-30.33 μm, mean 21.82 μm (n = 49); (h-h’) Vigna angularis. Range 7.42–28.32 μm, mean 14.51 μm (n = 100); i-i’ Dioscorea polystachya. Range 9.05– 20.7 μm, mean 14.27 μm (n = 85). (Scale bar: 10 μm).

Results

In total, 151 starch granules were recovered from 45 dental calculus samples (Table 2). Sixty-six starch granules could not be identified owing to severe damage or diagnostic features. The remaining 85 starch granules were classified into seven types based on the above-mentioned reliable diagnostic attributes (Fig. 4).

Table 2 The types and numbers of starch granules uncovered from Jiangzhuang site
Full size table
Fig. 4: Starch granules extracted from human dental calculus at Jiangzhuang site.
Fig. 4: Starch granules extracted from human dental calculus at Jiangzhuang site.The alternative text for this image may have been generated using AI.
Full size image

a-a’ Type 1, rice (Oryza sativa); (b-b’) Type2, millets; (c-c’) Type3, Coix sp; (d-d’) Type4, Triticeae; (e-e’) Type5, Fabaceae; (f-f’) Type6, Quercus sp; (g-g’) Type 6, Cyclobalanopsis sp (h-h’) Type7, tubers and roots (Scale bar: 10 μm).

Type 1 starch granules (N = 2) are polygonal with acute angles. They are small, ranging from 4.2 μm to 5.45 μm and the average size is approximately 4.82 ± 0.73 μm. The hilum is centered without fissures or lamellae. The extinction cross is faint with an “X” shape. The two starch granules are single in the free state (Fig. 4a-a’). Although most of starch granules from rice are always aggregated47,48, processing activities can disperse them and these grains are able to become single. Meantime, compared with modern reference materials (Fig. 3) and research by Yang et al. 49 on the morphological description of rice starch granules from the Kuahuqiao site, the features of Type 1 starch grains accord with those of rice (Oryza sativa). Considering that charred rice remains were found at the Jiangzhuang site, these starch granules are probably from rice (Oryza sativa).

Type 2 starch granules (N = 9) are polygonal with a centric hilum. Their size range is from 5.7 μm to 14.3 μm, and the average size is 9.66 ± 0.54 μm. They have smooth surfaces without any lamellae. Some of the larger starch granules have “Y” or horizontal linear fissures across the hilum with a plane view. The extinction arm is clear and vertical with an “+” shape (Fig. 4b-b’). Compared with reference collections in our laboratory and published works36,37 (Fig. 3), these polygonal granules are often found in the family Poaceae, including foxtail millet (Setaria italica), common millet (Panicum miliaceum), rice (Oryza sativa), Job's tears (Coix lacryma-jobi), sorghum (Sorghum bicolor), and oats (Avena sativa). Starch granules from rice and oats are smaller than type 2 (Fig. 3); Job's tears starch granules always have a “Z” shaped arm on the extinction cross, particularly on larger granules38 (Fig. 3); the starch granules of sorghum are bigger with distinct lamellae (Fig. 3). The surface of Type 2 starch granules is relatively smooth with clear and vertical extinction cross, and the average size is 9.66 μm, both indicating that the Type 2 starch granules were not from rice, oat, sorghum, and Job's tears; they are likely from foxtail millet or common millet. The size of foxtail millet is 6.23–19.35 μm with an average size of 9.32 μm, and the larger granules have fissures across the hilum (Fig. 3). The size of common millet is 5.96–10.99 μm, with an average size of 8.42 μm, and most of them have no fissures on the surface, with only a few starch granules having faint fissures across the hilum (Fig. 3). Moreover, the size of starch granules from common millet is generally not larger than 12 μm (Fig. 3)37. Consequently, the three larger sized grains (>12 μm diameter) with visible fissures show great similarity to foxtail millet, and the other six smaller granules were identified as common millet or foxtail millet.

The Type 3 starch granules (N = 13) are either polygonal or nearly round. Their sizes are 6.75–13.54 μm and the average size is 10.42 ± 0.94 μm. The hilum is centric or slightly eccentric with “-” or “Y” shape fissure. Lamellae are not evident. Under the polarized light, the extinction arm is vertical with a “Z” bent shape at the end. Some starch granules were encapsulated by calculus (Fig. 4c-c’). As mentioned above, Job's tears starch granules always have a “Z” shaped arm on the extinction cross, particularly on larger granules38 (Fig. 3). These characteristics of type 3 starch granules accord with those of Coix sp and they are probably from Coix sp.

Type 4 starch granules (N = 36) are oval or nearly round in shape, with sizes ranging from 9.75–33.84 μm. The average size is 17.91 ± 1.10 μm. The starch granules have a centric hilum without fissures. Some of the starch granules with large grain sizes or damaged features have obvious lamellae with diffusely “X” shaped extinction arm. Most starch granules are encapsulated by calculus (Fig. 4d-d’). Type 4 starch granules are lenticular or discoid in shape, with an olivary side view. Combined with modern reference materials in our lab (Fig. 3) and research by Yang34 on the morphology of starch granules from the modern tribe Triticeae, type 4 starch granules are probably from the tribe Triticeae.

Type 5 starch granules (N = 5) are oval or kidney-shaped. The size range is from 6.52 μm to 22.59 μm and the average size is 13.76 ± 2.81 μm. Lamellae are evident in the centric hila. The surface has longitudinal fissures or craters. The extinction is in the shape of “X.” The two extinction arms of some starch granules overlap at this intersection (Fig. 4e-e’). The starch granules of Fabaceae are mostly oval, kidney-shaped or round triangular in shape, with pronounced fissures such as dendritic, “Y”-shaped, linear, and wing-shaped (Fig. 3). Based on its morphology and size characteristics, this type of starch granules may have originated from the Fabaceae. According to Tao et al. 50, some Type 5 starch granules are consistent with those of Vigna sp. (Fig. 3). Charred seeds of Vigna sp. found at the Jiangzhuang site further suggest that some of these starch granules are probably from Vigna sp12.

Type 6 starch granules (N = 14) are ovoid, triangular with round corner or oval shape, which are from 5.71 μm to 17.2 μm in size. The average size is 10.22 ± 1.49 μm. The starch granules exhibit obvious fissures across the hilum. The extinction is in the shape of “X” (Fig. 4f-f’). The diagnostic features of Type 6 starch granules are consistent with nuts. Ovoids and triangles with round corner starch granules are commonly found in nut plants, such as Quercus sp., Cyclobalanopsis sp., Castanea sp., and Ginkgo biloba L. The starch granules from Castanea sp. have mainly eccentric and closed hila, with a small size, and lamellae around the hilum can be observed in starch granules with a long axis greater than 10 μm. The starch granules from Ginkgo biloba L. show radiations pointing from the center to the edge. None of these fit the diagnostic characteristics of this starch type41. Based on modern reference materials and previous studies41, most of Type 6 starch granules come from Quercus sp. Notably, one of the starch granules was similar to those form Cyclobalanopsis sp. (Fig. 4g-g’). It is triangular with round corners, with a short axis of 12.5 μm and a long axis of 13.16 μm. It has a centric hilum surrounded by several short fissures. Visible layers patterns on surface with diffuse extinction arms. Although no macro-plant remains of Cyclobalanopsis sp. were found in the archaeological sites in the eastern Jianghuai region, pollen grains from Cyclobalanopsis sp. have been found in profile samples from archaeological sites in this region51, indicating that this plant species should exist in this area. Therefore, it cannot be ruled out that this starch granule may come from the genus Cyclobalanopsis.

Type 7 starch granules (N = 6) have diverse morphological features with polygonal, oval, or irregular shapes. Their size is relatively large, ranging from 8.76 to 15.57 μm. The average is 11.40 ± 2.29 μm. The starch granules have centric or eccentric hilum with “-” shaped fissure or craters across the hilum. The extinction is in the shape of “+” or “X” (Fig. 4h-h’). Wang et al. 52 studied the morphology of starch granules from modern root and tuber plants in China; this type of starch granules may originate from root and tuber plants.

In addition to starch granules, other types of micro-plant remains, including phytoliths, pollens, fungal spores, and fibers, have also been extracted from human dental calculus at the Jiangzhuang site (Fig. 5 and Fig. 6). Owing to the limited quantity and lack of diagnostic morphological characteristics, phytoliths were unidentified and are not discussed here. Pollen (N = 21) and fungal spores (N = 33) were also relatively abundant. However, the fungal spores in this study could not be identified and thus are not further discussed below (Fig. 6). Only pollen grains with clear morphological characteristics were identified. The identifiable pollen grains (N = 13) extracted from human dental calculus at the Jiangzhuang site belonged to Pinus sp. (Fig. 5a), Fagaceae (Fig. 5b, c), Artemisia sp. (Fig. 5d), Quercus sp. (Fig. 5e, f), Persicaria sp. (Fig. 5g), and ferns (Fig. 5h).

Fig. 5: Pollen extracted from human dental calculus at Jiangzhuang site.
Fig. 5: Pollen extracted from human dental calculus at Jiangzhuang site.The alternative text for this image may have been generated using AI.
Full size image

(a) Pinus sp; (b-c) Fagaceae; (d) Artemisia sp; (e-f) Quercus sp; (g) Persicaria sp; (h) ferns (Scale bar: 10 μm).

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Fungal spores extracted from human dental calculus at Jiangzhuang site (Scale bar: 10 μm).

Discussion

In total, 151 starch granules were extracted from human dental calculus, of which 85 were preliminary identifiable. A variety of starchy plants, including crops (rice and millets) and gathered plants such as Coix sp., Triticeae, Fabaceae, acorns, tubers, and roots, were consumed by Jiangzhuang inhabitants.

Rice originated in the middle and lower reaches of the Yangtze River in China53, and its domestication and cultivation can be traced back over 10,000 years54. Rice is common in the archaeological sites of the Liangzhu culture and was the dominant crop species in the Liangzhu society11. Two rice starch granules were extracted from human dental calculus at the Jiangzhuang site, accounting for 2.4% of the total identifiable starch granules with a ubiquity of 22.2%, indicating that rice likely made a relatively low contribution to the human plant-based diet (Fig. 7). Macro-botanical remains from the Jiangzhuang site also indicate that Jiangzhuang inhabitants may have been equally dependent on rice and wild plant resources12. Therefore, the contribution of rice to the plant-based diet of the Jiangzhuang inhabitants was likely not visible. This is consistent with the discovery of starch granules in the dental calculi.

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

Percentage, ubiquity, absolute quantity of identifiable starch granules extracted from human dental calculus at Jiangzhuang site.

Millets, including foxtail and common millet, originated in north China approximately 10,000 years ago36. As the environment fluctuated alongside population migration and cultural exchange, foxtail millet and common millet gradually spread out from the core55. Millet starch granules (N = 9) account for 10.6% of the total identifiable starch granules, with a ubiquity of 33.3% (Fig. 7). Only two charred foxtail millet seeds were uncovered at the Jiangzhuang site12, indicating that millet crops made a small contribution to the diet of Jiangzhuang inhabitants. Millet starch granules extracted from human dental calculi further suggest that millet crops spread to the eastern Jianghuai region during the Liangzhu period and became a food source for Jiangzhuang inhabitants.

Starch granule analysis has revealed that as early as the Upper Paleolithic period, around 28 ka years ago, the tribe Triticeae was already processed and utilized56. During the Neolithic and subsequent periods, starch granules from the tribe Triticeae were found on stone tools, pottery, and human dental calculi at many archaeological sites in ancient China34,36,46. In fact, Jiuhuang Bencao by Zhu Su (朱橚) of the Ming Dynasty records that Triticeae species were indeed gathered and processed for food in ancient China. For example, seeds from Elymus sp. were collected, dehusked, and ground into flour for consumption57. Thirty-six Triticeae starch granules were found in the human dental calculus, accounting for 42.4% of the total identifiable starch granules, with a ubiquity of 55.6% (Fig. 7). Both the absolute number and ubiquity of Triticeae starch granules were clearly higher than those of crops at this site, indicating that the tribe Triticeae did play an important role in the plant-based diet of Jiangzhuang inhabitants. Notably, the charred seeds of the subfamily Pooideae were unrecovered from the Jiangzhuang site, although there was no further identification conducted on the subfamily Pooideae. Considering the extraction of Triticeae starch granules from human dental calculus at the site, the possibility that some of these Pooideae seeds might belong to Triticeae and were consumed cannot be completely ruled out. Starch granules from Triticeae provide direct evidence for the consumption of wild Triticeae seeds by Jiangzhuang inhabitants for the first time. This broadens the understanding of the plant food spectrum during the Liangzhu period at the Jiangzhuang site. Moreover, some starch granules displayed obvious damage characteristics, including morphological deformation caused by cooking and heating, specific gelatinization features (such as clearer layer patterns and disappearance of extinction crosses), as well as surface or edge roughness and cracking caused by grinding or other mechanical force. The damage characteristics of these granules are similar to those of starch granules from wheat and barley after processing58, and these damaged starch granules are likely from Triticeae, which also indicates that wild Triticeae seeds were processed and consumed by the Jiangzhuang inhabitants (Fig. 8).

Fig. 8
Fig. 8The alternative text for this image may have been generated using AI.
Full size image

Damaged starch granules of Triticeae extracted from human dental calculus at Jiangzhuang site (Scale bar: 10 μm).

Macrofossil remains of Job's tears have only been found sporadically at some archaeological sites, but their micro-plant remains have been recovered from more than 30 sites across China, dating to as early as 28,000 cal. BP35. Ancient starch granules analysis reveals that Job's tears were an important plant food resource in the Yangtze, Yellow, Liao, Jianghuai, and Huanghuai regions during the early and Middle Holocene59,60,61. The earliest Job's tears starch granules in the Jianghuai region were found on stone and pottery artifact surfaces of the middle Neolithic period (8500–7000 cal BP) excavated from Shunshanji site47. Subsequently, such remains have been found in archaeological sites dating to the Shuangdun, Houjiazai, and Longshan cultures62,63. Thirteen starch granules from Coix sp. were extracted from human dental calculi at the Jiangzhuang site, accounting for 15.3% of the total starch grains (Fig. 7). The ubiquity of Coix sp. starch granules (55.6%) is higher than that of rice and millets, indicating that Coix sp. made a significant contribution to the diet of the Jiangzhuang inhabitants. No charred seeds from Coix sp. were found at the site; therefore, starch granules from Coix sp. uncovered from human dental calculus help fill the gap information regarding the macro-botanical remains and enrich our understanding of plant food consumption in the Jiangzhuang settlement. Neolithic Job's tears are predominantly found in the western Jianghuai region62,63, whereas they were virtually absent in its eastern part. Starch granules from Coix sp. uncovered at the Jiangzhuang site for the first time provide definite evidence of the utilization and consumption of Coix sp. in the eastern Jianghuai region during the late Neolithic period.

Acorns such as Quercus sp., Castanopsis sp. and Cyclobalanopsis sp. were common food resources worldwide64. Microremains and macroremains of acorns have been commonly discovered at prehistoric archaeological sites in China60,65, indicating that acorns were utilized as a food resource in prehistoric China. Fourteen starch granules extracted from the human dental calculus were from acorns, mostly from Quercus sp., which accounted for 16.5% of the total identifiable starch grains (Fig. 7). Although the proportion of acorn starch granules in the human dental calculus is low, its ubiquity is high (66.7 %) (Fig. 7). This result suggests that acorns made a visible contribution to the plant-based diet of the Jiangzhuang inhabitants.

Archaeological findings worldwide have demonstrated the use of roots and tubers as food resources66. Starch granules in roots and tubers have been discovered at numerous sites dating from the Paleolithic to the Neolithic in China36,67, indicating that root and tuber plants were important plant-based food resources. Starch granules from the roots and tubers extracted from the human dental calculi account for 7.1% of the total identifiable starch granules, with a ubiquity of 33.3% (Fig. 7). The archaeobotanical assemblages at the Jiangzhuang site also indicate that the roots and tubers constituted a relatively high proportion of the total12. Overall, roots and tubers did make contributions to the plant-based diet of the Jiangzhuang inhabitants.

Fabaceae are closely related to humans, and many of them are important food sources, such as Glycine sp. and Vigna sp. Ancient starch residues show that Fabaceae were utilized as early as the Late Paleolithic68. Macro- and micro-remains of Fabaceae since the Neolithic period, have been discovered in additional archaeological sites42,49. Starch granules of Vigna sp. extracted from the inner wall of a pottery vessel at the Kuahuqiao site revealed that Vigna sp. Were cooked as early as the early Neolithic period49. Five starch granules from Fabaceae were extracted from human dental calculi at the Jiangzhuang site, accounting for 5.9% of the total identifiable starch granules, with a ubiquity of 33.3% (Fig. 7). These Fabaceae starch granules are probably from Vigna sp. extracted from human dental calculus indicate that Vigna sp. also contributed to the Jiangzhuang inhabitants’ diet.

Pollen extracted from human dental calculus can be considered an environmental indicator, allowing us to explore the human-plant interactions16,69. Pollen grains from Quercus sp. and Fagaceae uncovered from human dental calculus indicate that the existence of evergreen and deciduous trees of the family Fagaceae around the Jiangzhuang site, which provided the acorn resources that the Jiangzhuang inhabitants could gather, process, and consume. This was confirmed by the extraction of acorn starch granules from human dental calculus at this site. The presence of pollen grains from Persicaria sp. and Artemisia sp. from the calculus demonstrates that these plant species were present around the site, which could have provided utilizable plant resources and were likely consumed by Jiangzhuang inhabitants. Many plant species from the Polygonaceae and Asteraceae families are edible as wild vegetables, as documented in Jiuhuang Bencao57, which provides a comprehensive introduction to the consumption and utilization of edible wild plants during the famine period. Combined with charred seeds of Polygonaceae and Asteraceae families uncovered at the Jiangzhuang site, these wild plant resources were utilized by Jiangzhuang inhabitants, possibly as food resources.

This study reveals that diverse plant food sources were consumed by Jiangzhuang inhabitants. Wild plants contributed a higher proportion to Jiangzhuang plant foodstuffs, whereas crops such as rice and millets contributed less. Meantime, it should be pointed out that charred Euryale ferox and Trapa sp. present in macroremains at Jiangzhuang have not been identified in the microremains, and Triticeae and Coix sp. which were not found in the macroremains were indeed consumed by Jiangzhuang inhabitants. The discrepancies between the macro-botanical and micro-botanical datasets in this site suggest the strengths and limitations of starch granule analysis of dental calculus. Moreover, some kind of plant species such as rice, millets and nuts found in the macroremains were also identified in the starch residues, which further confirms the consumption of these plant resources and their contribution to the diet of Jiangzhuang inhabitants. Thus, this research broadens our understanding of the plant food spectrum at the Jiangzhuang site, and further shows that how the different archaeobotanical evidences can reinforce and complement each other to get more comprehensive understanding of plant-based subsistence strategies in archaeological sites70.

Bioarchaeological studies, including macro-botanical and zooarchaeological analyses, have been conducted at the Jiangzhuang site, revealing to some extent the dietary sources and subsistence choices of Jiangzhuang inhabitants. The macro-botanical remains unearthed from the site show that rice, Euryale ferox, Trapa sp., and Cucumis melo served as the primary plant-based staple resources13. Starch granule analysis from the human dental calculi not only confirms the consumption of rice and millet crops alongside wild plants such as Quercus sp., roots, and tubers by Jiangzhuang inhabitants. It also reveals the consumption of new edible plant species, such as Coix sp. and Triticeae, which broadens the understanding of the plant food spectrum at the Jiangzhuang site. Overall, during the Liangzhu period, crop cultivation and gathering activities were equally important for the inhabitants12. Meantime, animal remains at the Jiangzhuang site shows that livestock raising at the Jiangzhuang site was limited in scale, with wild animals predominating7. Subsistence activities at the site during the Liangzhu period relied mostly on wild plant and animal resources with a limited-scale crop production and livestock raising.

The subsistence pattern at Jiangzhuang site can be observed at contemporaneous sites in the eastern Jianghuai region, such as Kaizhuang site. The intensity of rice cultivation and pig rearing was limited, whereas gathering and hunting played more important roles in Kaizhuang during the late Liangzhu period8,9. This subsistence choice also existed in the Ningshao Plain and some sites in eastern region of Zhejiang Province, which both belong to the non-core area of the Liangzhu culture. Animal and plant remains recovered from sites Tashan and Cihu (Fig. 1) suggest that wild animal resources including Cervidae and water buffalo, and plant species such as Rhamnaceae and Fagaceae were important food sources71,72. Subsidence in the Ningshao Plain and eastern region of Zhejiang Province also primarily relied on hunting and gathering, with limited rice agriculture and livestock raising7.

Compared to the non-core areas of the Liangzhu culture mentioned above, the subsistence choice in the core area of the Liangzhu culture, the Taihu Plain region, including the Zhumucun, Guangfulin, Bianjiashan, and Maoshan sites (Fig. 1), was characterized by developed rice agriculture and mature livestock raising. Wild plants were also found, utilized, and consumed by inhabitants as food supplements10. Therefore, differences in subsistence choices did exist between the core and non-core areas of the Liangzhu culture7,73. Our study provides new evidence to reveal subsistence choice in the non-core area, which further confirms the differences in subsistence patterns in different areas of the Liangzhu culture, as Pan et al. 73 proposed.

In conclusion, starch granule analysis of human dental calculus provided direct evidence of plant consumption in Jiangzhuang. A wide range of plants, including crops of rice (Oryza sativa) and millets, and wild plants such as Triticeae, Coix sp., Fabaceae, acorns, and roots/tubers, were consumed. Some starchy plants, such as Triticeae and Coix sp., were not found in the macro-botanical remains but were identified in the micro-botanical remains of human dental calculus at the site. This study cures the previous deficiency regarding macro-botanical remains and broadens knowledge of plant food spectrum at the Jiangzhuang site. Although the number of samples is not large, and the data obtained are relatively limited, this research provides new evidence to reveal broad-spectrum plant consumption in a large settlement located in the non-core area of the Liangzhu culture. It deepens the Understanding of differences in subsistence patterns in different areas of this culture.