Thirteen cereal grain samples from the Suyang site were sent to the State Key Laboratory of Isotope Geochemistry accelerator mass spectrometry (AMS) radiocarbon dating at Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Each date was calibrated by OxCal v4.4.2, using the IntCal20 atmospheric curve. The 13 AMS C-14 dates obtained from the Suyang site are shown in Table 1. Except for one abnormal date from single wheat grain, the dates are consistent with the dates proposed by ceramic typology (Ren, 2021). The available dates can be grouped into two sets. The first set contains three dates covering the time interval 2,917–2,747 cal BC (95.4% range), which falls within the late Yangshao culture age. The second set covers the period from 2,760 to 1939 cal BC (95.4% range), which corresponds to the Longshan culture.

Except for 61 stems, 47,083 seeds or fruits were recovered from flotation, of which 45,963 were identified (Table 2 and Figure 2). The density was 44.2 seeds/fruits per liter of soil. Five types of crops were found in the samples, including foxtail millet, broomcorn millet, rice, soybean (Glycine max), and wheat (Triticum aestivum). The total amount was 45,562, accounting for 99.13% of the carbonized plant remains. Every sample had all these crops, so the ubiquity of each kind was 100%. Except for some immature millets, most millets were mature and complete. Among rice, broken fragments were more abundant than complete seeds, of which 1,850 were rice spikelet bases. There were not many soybeans found in the samples, but the complete ones were six times more than the incomplete ones. The relative percentages of each crop type were as follows: foxtail millet, 73.89%; rice, 25.19%; broomcorn millet, 0.75%; and soybean, 0.16%. The ubiquity of foxtail millet is clear from all samples and periods; therefore, it seems obvious that foxtail millet was the dominant crop at the Suyang site. However, there are interesting trends; the values of relative percentage and ubiquity changed throughout the study period. The relative percentage of foxtail millet declined from the late Yangshao period to the Longshan period, while rice witnessed an increase. Simultaneously, the ubiquity of the broomcorn millet saw a slight increase, but its amount was still far below that of the others. In contrast, soybean maintained a low relative percentage throughout the entire period (Figure 3). Finally, one wheat seed was found in the sample TN5W9H34; however, it turned out to be an intrusion from modern times after AMS C-14 dating (Table 1).

In addition, 401 weeds with a 0.87% relative percentage and 73.02% ubiquity were found in the samples from 37 species, of which the most abundant were Leguminosae, Chenopodiaceae, Lamiaceae, Euphorbiaceae, Polygonaceae, and Cyperaceae. Chenopodium sp., Digitaria sp., Setaria sp., Lespedeza bicolor, Melilotus sp., Perilla sp., and Polygonum aviculare. Chenopodium sp. is a crop with a long history of cultivation in some areas, such as America and Taiwan (Smith, 1984). The seeds of Chenopodium sp. unearthed together with millet from the Han Yangling site might be the earliest evidence of Chenopodium’s cultivation (Yang et al., 2009). While Chenopodium sp. remains from the Nanwa site, the Guchengzhai site and the Erlitou site recognized these seeds as a cultivar (Cheng et al., 2012; Wu et al., 2014), whether Chenopodium sp. is a farmed crop remains uncertain. In Suyang, it is possible that people gathered this weed to eat as a vegetable. Lespedeza bicolor and Melilotus sp. are common at sites in central China, and they are recognized as fodder or fuel; however, those from the Suyang site do not necessarily fit this hypothesis based on their scarcity. Fruits found in the samples were rare and were probably used as food to enrich their diet.

Foxtail millet seemed to be the most important crop at the Suyang site during the late Yangshao period. Rice was less present but important at the site, considering its 80% ubiquity. Broomcorn and soybean did not equal the status of foxtail millet and rice, although broomcorn millet was slightly more important than soybean. During the Longshan period, while the status of rice seemed to improve over time.

The percentage and ubiquity of seed remains of foxtail millet would place this crop as the most important cultigen throughout both periods of occupation of the site. However, it is necessary to consider the grain size (weight) before estimating their role in the plant diet. According to the weight of carbonized seeds unearthed at the Suyang site, the results indicate that rice grew in prominence during the late Yangshao period and surpassed foxtail millet. Rice may have surpassed foxtail millet as the most important crop by the Longshan period, as shown in the graph below (Figure 6).

Evidence from experimental archaeology also supports our result. When comparing the number of glume phytoliths from rice and millet of equal weight, the data showed underestimated the rice. Modern experiments have shown that millet produces more phytoliths than rice; therefore, it is expected that when millet and rice are equal in weight, millet phytoliths are more likely to occur (Zhang, 2010). In our analysis, glume phytoliths of rice were more abundant than millet, suggesting that more rice might have been produced than millet at the Suyang site. Meanwhile, we also need to pay attention to the following issues. First of all, because there are only experimental data on the glume of millet and rice, the above discussion is based on the glumes of millet and rice, and the stem leaves are not discussed. Second, phytoliths from millet husk are very fragile and more difficult to preserve, so the discussion based on husk is very preliminary, and more follow-up work is needed.

It is interesting to compare Suyang with other sites in Central China; as shown in Figure 7, the percentage and ubiquity of rice at the Suyang site are the highest. The Wanggedang site located in the Lower Luo River turned up 144 seeds of rice (4.09% of the total) recovered from systematic flotation, indicating that while rice could have been an essential part of the plant diet considering its 52% ubiquity (Zhong et al., 2019), it was still far inferior to the Suyang site. In comparison, the status of rice in the sites of Chengyao, Wangchenggang, and Xinzhai was lower than in others (Zhao, 2007; Zhong et al., 2018). In contrast, rice made a more critical contribution at the sites of Wadian in the middle of Ying River Valley, in the Mid-Holocene accumulation stage with the characteristics of minor altitude difference between terrace and riverbed, which is like a waterside village. Such a landscape made it possible for the ancients to excavate ditches and proceed with rice cultivation (Wang et al., 2015). Still, the rice in Wadian is not as crucial as that in Suyang.

The recovery of 1,850 rice spikelet bases (1746 from the late Yangshao period and 104 from the Longshan period) suggests that at least some rice was processed at the site. Phytoliths from the leaf/stem or the husk of rice spanning all kinds of contexts at the Suyang site and weeds common in paddy fields such as Echinochloa and Cyperaceae were recovered in flotation samples, suggesting that rice was produced locally. Given that previous research has failed to certify whether rice was cultivated by local people in the Luo River Valley (Lee et al., 2007), the Suyang site offers compelling evidence demonstrating that people during the late Yangshao period in the Luo River Valley cultivated rice.

Even though the research suggests that Central China had a favorable environment for rice cultivation in ancient times, the question remains as to what kind of rice people chose to cultivate (Wang et al., 2018). According to Weisskopf et al. (2015a), different ratios of phytolith morphotypes to produce silica in the grass (fixed) and those that need sufficient water (sensitive) could be used to rebuild rice water settings in rice cultivation. The fixed: sensitive ratio at Suyang demonstrated a substantially higher ratio of sensitivity to fixed phytoliths, indicating that the setting in the late Yangshao period was similar to that in the Tianluoshan and Shunshanji sites (Luo et al., 2020). In the case of Tianluoshan, the results suggested wet paddy agriculture (Weisskopf et al., 2015a). Therefore, it is likely that the rice remains from Suyang was cultivated in irrigated paddies (Figure 8). It is possible that the people living at the site might have taken advantage of the surrounding swamps to grow rice. In the following Longshan period, the setting did not seem to have changed according to the ratio.

Furthermore, four seeds of Scirpus juncoides, a shallow water plant, were recovered at Suyange, indicating wet conditions around the settlement. Additionally, a paddy system was found in the Huizui site located in the lower Yiluo River Valley, which led us to search for the same thing at Suyang (Rosen et al., 2017). Unfortunately, there have been no remains suggestive of a paddy field found, probably due to the small excavation areas.

Now, we answer why the status of rice in the Longshan period grew despite the lack of improvement in the cultivation method. Paleoenvironmental research suggests that Henan Province was under the influence of subtropical monsoon climate during the Yangshao period, which means that it had higher temperature and precipitation than in modern times (Li et al., 2015). For example, the deposited sediment from a profile at the Zaojiaoshu site revealed that the area had a warmer and wetter climate than the present, parallel to the modern climate in the middle-lower Yangtze Plain (Zhang and Li, 1997). The charcoal found in the Suyang site was from various species, including Aceraceae, Bambusoideae, Pinus sp., Quercus sp., Populus sp., Ulmus sp., and Cercidiphyllum sp. Bambusoidea, only grows in wet and humid environments, indicating that the area enjoyed a wetter climate in the past. Suyang Site is located on the T2 terrace of southern Luo River (Ren, 2021). About 7,200 years ago, the T1 terrace began to form along with the erosional downcutting of Luo River. The increasing sediment accumulation brought sediment-rich clay and ponding depressions which provided the appropriate setting for rice cultivation. It is credible that there are large floodplain areas suitable for rice cultivation on both sides of the Luo River near the he Suyang site (Rosen et al., 2015; Zhang et al., 2018; 2019).

In addition to environmental reasons, culture and population are also important factors affecting rice cultivation. From the late Yangshao period, the Qujialing culture showed a definite northward trend, the western Henan area and the Luoyang Basin was greatly influenced by the Qujialing culture from the Middle Yangtze Region (Dai, 1998; Meng, 2011). A few relics of Miaodigou II culture were found in the Suyang site, and many relics have Qujialing cultural characteristics. With the solid northward movement of Qujialing culture, the middle reaches of the Luo River were included in the late Yangshao period (Ren, 2021). The spread of the Qujialing culture into this area, the Suyang site would have impacted greatly on rice farming that focused Yangtze pattern.

The increase in population might have resulted from the rice agriculture in the Suyang site. According to the excavation results, there more relics of the Longshan period were recovered in the Xiacun areacomparing with the former period (Ren, 2021), which suggests the residential area of the settlement expanded to the Xiacun areas during the Longshan period. The possible reason behind this scenario could be the population growth. Rice is a higher-yielding crop with good taste than millet, and the large-scale cultivation of rice can feed more of the population. However, Rice is also a labor-intensive crop. From our data, the population growth could also mean an increased workforce used in water management and relative production activities in rice cultivation.

Crop processing refers to the process that takes place after harvest and before consumption, which generally contains several stages such as threshing, winnowing, dehusking, and sieving (Reddy, 1997). Because the processing steps require particular assemblages, they can be identified by examining the composition of each sample. According to previous research, all processing steps belonged to either earlier stages featuring threshing reflected by many leaf/stem parts, small and light weed seeds, and immature grains in the byproducts; or to later stages featuring dehusking associated with clean grains and husks in the byproducts (Fuller and Steven, 2009; Fuller et al., 2014). Phytoliths can provide information about the crop processing style, provided that crops such as rice and millet produce different phytoliths in different parts of their body (Harvey and Fuller, 2005). For example, the ratios of the stem, husk, and weed can be used to infer the site, stages, and scale of crop processing, which can further reveal the division and organization of social labor and economy (Stevens, 2003; Fuller and Steven, 2009; Fuller et al., 2014; Weisskopf, 2014).

The phytoliths produced by the stem/leaf of millet and rice are ubiquitous at the Suyang site, indicating that the crops were harvested by reaping at the sheaf. Three samples (H1, H3, H6) of the Yangshao period only had phytoliths of husk and several mature grains, while the other samples have much the same phytoliths, which demonstrates that dehusking was a routine activity at the Suyang site. The samples of this period did not have any assemblage characteristic of the earlier stages of the crop processing, which is possible because the threshing field was not in the sampling area.

For the Longshan period samples, the amount and ubiquity of phytoliths produced by the stem/leaf of millet and rice were higher than the former, indicating continuity in the harvesting style. Some samples from the Suyang (H18, H26, H27) and Xiacun (H17, H18, G1(14)) areas have a large number of husk phytoliths, especially sample H18①-6 with a density of 830,000 particles per gram. These samples indicate that dehusking was still a routine activity at the Suyang site during the Longshan period. Meanwhile, other samples from the Suyang area (H7, H14, H41) and Xiacun area (H11②, H12②, H15) provided evidence based on stem/leaf phytoliths for threshing activity, suggesting that the whole processing activities were carried out inside the settlement in the Longshan period.

The absence of evidence suggestive of threshing activities on-site during the Yangshao period may also indicate that in earlier periods, these activities were communal and were undertaken in large open areas outside the settlement perimeter. If this is true, a change in the crop processing pattern in later periods when threshing and dehusking are observed inside the village may point to the development of nuclear families. Researchers have shown that the nuclear family as an economic unit started to show its independence and desire for wealth and status during this period, as a consequence of the development of social complexity in the area (Dai, 2013). Furthermore, evidence from the Baligang site (Weisskopf et al., 2015b), the sites in Ying River Valley (Zhao, 2007), Shu River Valley (Song et al., 2019), and the upper Hutuo River Valley (Jiang et al., 2019) all indicate the conversion of labor from the late Yangshao period to the Longshan period. It should be noted that t the number of samples analyzed in this study remains small, and this part is only a discussion based on the sampling area. When the site is re-excavated, we will supplement the sample size and discuss this again.