C₄ photosynthesis is a highly efficient carbon fixation pathway that enhances photosynthetic performance under conditions of high light intensity, high temperatures, and limited water availability. Unlike the C₃ pathway, where CO₂ is fixed by the enzyme Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCo) in mesophyll cells, C₄ plants utilize a distinct mechanism involving two types of cells to incorporate atmospheric CO₂: mesophyll and bundle-sheath cells. Initially, CO₂ is fixed into a four-carbon compound by the enzyme phosphoenolpyruvate (PEP) carboxylase in the mesophyll cells. This four-carbon compound is then transported to the bundle-sheath cells, where CO₂ is released and refixed by RuBisCo, thereby entering the Calvin cycle. The carbon isotope fractionation between CO₂ and the final photosynthetic product can be positive, negative, or zero. This variation depends on the ratio of intercellular to ambient CO₂ partial pressure (p_i/p_a), temperature, and leakiness (ϕ) of the bundle sheath cells (Farquhar et al., 1989; Williams et al., 2001). The spatial separation of the initial CO₂ fixation and the Calvin cycle minimizes photorespiration, a wasteful process that occurs when RuBisCo reacts with O₂ instead of CO₂. This adaptation provides C₄ plants with a significant advantage in hot and arid environments, contributing to their dominance in many tropical and subtropical grasslands and savannas. As a consequence, C₄ plants exhibit less carbon isotope discrimination compared to C₃ plants, resulting in higher δ¹³C values, with an average of −28‰ in C₃ plants and −12.5‰ in C₄ plants (Cerling et al., 1998; Kohn, 2010). The effects of the environmental factors on carbon stable isotopes have been widely studied in C₃ plants and it has been demonstrated that they can strongly affect the carbon isotopic values (Farquhar et al., 1982, 1989; Körner et al., 1991). The influence of the environmental factors on C₄ plants is less clear and results are often contradictory. Even though some research attests that the influence of environmental factors on the ¹³C fractionation of C₄ plants is negligible, more studies show that factors such as light intensity (Buchmann et al., 1996; Kromdijk et al., 2008; Ubierna et al., 2013), water availability (Buchmann et al., 1996; Ghannoum et al., 2002; Weiguo et al., 2005; Reid et al., 2018; Lightfoot et al., 2020; Sanborn et al., 2021), salinity (Bowman et al., 1989), latitude (An et al., 2015; Yang and Li, 2015), and altitude (Wang et al., 2008) affect C₄ plants, although the relationship is not always linear (Swap et al., 2004; An et al., 2015). Indeed, one of the most debated topics is the relationship between δ¹³C values and water availability, encompassing studies on both precipitation and irrigation in controlled experiments. Some studies find positive correlation between rainfall regional gradients or cumulative annual precipitation and δ¹³C values (Schulze et al., 1996; Murphy and Bowman, 2009; An et al., 2015; Yang and Li, 2015; Reid et al., 2018), while others demonstrate negative correlations (Weiguo et al., 2005). In experimental cultivation where irrigation was controlled in greenhouse conditions or in screen houses, δ¹³C values were mostly positively correlated with water availability: this result has been attested in different C₄ species such as Zea mays (Dercon et al., 2006), Pennisetum glaucum (Sanborn et al., 2021), Setaria italica (Lightfoot et al., 2020; Dong et al., 2022), while Panicum miliaceum is negatively correlated, even if not statistically significant (Dong et al., 2022). However, when Panicum miliaceum was sampled in nature and environmental factors were measured, δ¹³C was positively correlated with precipitation (Yang and Li, 2015). The relationship between δ¹³C and altitude is another controversial issue. Even though it seems that different photosynthetic pathways (C₃ vs. C₄) do not affect the relative importance of environmental contributions to the carbon isotope discrimination and gas exchange altitudinal variations (Wang et al., 2008), the effect of altitude on C₄ plants is still little explored and the relationship between altitude and δ¹³C is unclear. For example, an analysis of Setaria viridis found a positive correlation between δ¹³C and altitude (Wang et al., 2008). In contrast, an analysis of Panicum miliaceum recorded a negative correlation, while no correlation was found for Setaria italica (Yang and Li, 2015).

Nitrogen isotope ratios for non-nitrogen-fixing plants are dictated by sources including ammonium (NH4+) and nitrate (NO3–) in soil. The δ¹⁵N value of a plant depends on the proportion of these nitrogen sources and the discrimination factors affecting each in the plant (Evans, 2001). For most C₃ and C₄ plants (non-nitrogen- fixing), soil is the major source of N. Soil δ¹⁵N values are influenced by soil composition (Makarov, 2009), system openness (Handley et al., 1999), salinity (van Groenigen and van Kessel, 2002), soil age and depth (Nadelhoffer et al., 1996), water availability (e.g. rainfall) (Amundson, 2003) and altitude (Liu and Wang, 2010). Generally, soil δ¹⁵N values increase as ¹⁵N-depleted nitrogen compounds are lost through nitrification, ammonia volatilization, and leaching. This is the case, for example, in hot and arid regions like the deserts of southern Africa, where the δ¹⁵N values in soil, plants, and animals are enriched due to reduced rainfall (Shearer et al., 1978; Heaton et al., 1986; Heaton, 1987; Sealy et al., 1987; Schulze et al., 1991; Evans and Ehleringer, 1993, 1994). Plant δ¹⁵N values are positively correlated with temperature, and in addition negative correlations with annual precipitation have been shown (Handley et al., 1999; Amundson, 2003). Nitrogen isotopic discrimination in plants is influenced by soil nutrient availability, reflecting N cycling on various scales (Tilman, 1988; Nadelhoffer and Fry, 1994), and by the surrounding environment, including anthropic activity. Anthropogenic activities, such as agriculture, affect plant δ¹⁵N values, sometimes even masking the natural δ¹⁵N variability induced by environmental factors (Amundson, 2003). The use of fertilizers, such as manure and excrement, causes important changes in plant δ¹⁵N values, leading to an enrichment in ¹⁵N (Choi et al., 2006). Experimental research on different crops shows an increase in δ¹⁵N values when manure is used, in both C₃ (Fraser et al., 2011) and C₄ crops (Christensen et al., 2022; Dong et al., 2022). Differences in N-isotope fractionation between C₃ and C₄ plants are not well established. C₄ species show more efficient nitrogen use than C₃ plants (Schmitt and Edwards, 1981; Sage and Pearcy, 1987; Makino et al., 2003), which may cause δ¹⁵N to differ between the two pathways (Murphy and Bowman, 2009). Similar to carbon, among the limited experimental studies conducted on C₄ plants, some research illustrates a positive relationship between grains δ¹⁵N values and cumulative annual precipitation in Pennisetum glaucum (Reid et al., 2018), while other research attests that water availability has no discernible effect on grains and leaf δ¹⁵N values of the same species (Sanborn et al., 2021). Concerning the altitude, it seems that different photosynthetic pathways do not affect isotopic results and in general δ¹⁵N values decrease with increasing elevation (Liu and Wang, 2010). Precipitation and temperature seem to be the main factors influencing δ¹⁵N variation in plants with altitude. Lower δ¹⁵N values at higher altitudes are attributed to reduced mineralization and net nitrification rates caused by lower temperatures and abundant rainfall (Liu and Wang, 2010).

Silica (SiO₂), most commonly encountered in nature as quartz, comprises over 10% of the Earth's crust. Plants absorb silicic acid Si(OH)4, through their root systems, together with water, minerals and nutrients and deposit it in the aerial parts forming phytoliths. When it comes to mechanisms of silica absorption, plants can be classified as active, passive or rejecter (Tubaña and Heckman, 2015) depending on the concentration of silicic acid in the xylem compared to that in the soil. Most C₄ plants are considered high silica accumulators with silica concentrations that can reach up to 10% of the dried weight. This means that the amount of silicic acid absorbed depends on its availability in the soil, the physiological state of the plant, and the genetic makeup. When silicic acid absorbed passively through transpiration is in too low concentrations, the density of silica transporters increases, driving an active absorption process (Ma and Yamaji, 2008). Si absorption in plants has been linked to increased growth rates in crop plants (e.g. Greger et al., 2018) and higher grain yield (e.g., D'Agostini et al., 2023), thereby promoting both plant growth and reproduction. Si compounds in both C₃ and C₄ cereals have been proven to significantly enhance plant photosynthesis (Goto et al., 2003; Cornelis and Delvaux, 2016) and chlorophyll content (Verma et al., 2019), potentially influencing leaf nitrogen (N) and carbon (C) content and altering C:N stoichiometry (Liu et al., 2023). Indeed, photosynthetic capacity is generally closely correlated with leaf nitrogen and phosphorus content because both largely affect the CO₂ assimilation capacity of plants by altering the concentration of enzymes and ATP (Li et al., 2018; Zhang et al., 2018). Consequently, assessing the biosilica content in the plant and its possible relationship with carbon and nitrogen content can provide insights into the plant's productivity and photosynthetic rate relative to its growth environment. Additionally, it contributes important information to understanding the mechanisms of carbon and nitrogen discrimination in C₄ plants, which, as mentioned above, remains a highly debated topic. Therefore, we expect to observe a correlation between biosilica content and C:N ratios, which may vary depending on environmental conditions. A higher silicon content can reduce the need for energy resources used for chemical defense or structure, allowing the plant to allocate more carbon to growth (C in grains and leaves). Similarly, under water stress conditions, C₄ plants with higher biosilica levels might show a different carbon allocation compared to those with lower biosilica levels, thereby demonstrating that phytoliths and δ¹³C are not independent proxies.

The Konso landscape includes three of the five major agroclimatic zones: berha, kola, and woina daga. Kola makes up 70% of the territory, while woina daga and berha account for 20% and 10%, respectively (Yeshambel, 2013; Fikadu, 2023). Crop agriculture and animal husbandry have been the main means of livelihood in Konso. Konsitas rarely engage in traditional oxen plow agriculture because most of their land is hilly, rocky, small in size, and fragmented (Watson, 2009). Instead, farmers cultivate the terraced land using single-edged and double-edged composite hoes. These composite tools help prepare the soil by pulling out roots, rocks, shrubs, and bushes. Poor soil conditions, lack of adequate cultivable land, and unpredictable and insufficient annual rainfall have been critical challenges for farmers in the region. To overcome some of these challenges, Konso farmers have adopted terracing, irrigation, insulating, multiple cropping, and the planting and use of trees and shrubs (Förch, 2003; Amesias, 2023; Beldados et al., 2023). To conserve soil and water in their agricultural fields, farmers construct terraces on almost all of their land, using traditional field stone mortarless construction methods. This terracing technique conserves the soil from wind and water erosion, facilitates deep drainage and provides cultivable spaces on steep slopes (Tadesse, 2010; Tamiru, 2014). Most of the terraced farm lands are cropped year-round without fertilization. Instead, farmers keep some herbaceous nitrogen-fixing trees such as Ipomoea kituiensis Vatke, Lamiaceae, Asteraceae, Solanaceae and Rubiaceae in their farm land to convert nitrogen into a usable agricultural component.

With the progressive demographic increase at higher elevations and with the need of expanding commercial farms and productivity, Konso farmers recently started to occupy the adjacent lowlands (1200 and 800 m a.s.l) along the Segen River. Here, over the last two decades, farmers have cultivated sorghum, millets, and maize for commercial and household use. Ox-plow agriculture, irrigation and soil terracing are the main features of farming in the lowlands.

Unique aspects of agricultural practice in Konso, in comparison with other neighboring regions, include intercropping and the cultivation of C₄ crops in a difficult hilly landscape (Hallpike, 1972; Watson, 2009). A single household owns on the average farming plots on the order of 20m X 10m, where up to 12 crop varieties are planted. The most common species cultivated are chickpea, cowpea, bean, hyacinth bean, pigeon pea, mung beans and peas. Moringa stanopetela (Baker f.) Cufod. (popularly known as moringa) widely grows in Konso and is used as staple food and medicinal plant. Furthermore, oil crops (sunflower, castor, and linseed), root crops (pumpkin, yams, sweet potatoes, and tomatoes), cash crops (coffee, cotton, tobacco, and maize), and fruits (papaya, mango, banana, orange, lemon, and avocado) are also cultivated in the area (Amesias, 2023; Fikadu, 2023). Sorghum and millets are the main C₄ crops that grow in Konso in both mid and lower elevations over an altitudinal range of one thousand meters, between 700 and 1,700 m a.sl. These crops are well-adapted to the dry and semi-desert environmental conditions of the area and contribute significantly to the local diet.

The average annual rainfall of the study area is about 570 mm. Konso is characterized by prolonged dry seasons and two poor-quality rainy seasons: the most intense from mid-February to the end of April, followed by the small rainy season from May to June. Fifty-seven percent of the total annual rain falls from February to April and of this, 62% is most often concentrated in April (UNDP Situation Report, 1999). There are two planting seasons in the area: belg (February-May) and meher (October-November, Hagaya months). During belg, farmers plant and cultivate crops including sorghum, t'ef, maize, haricot bean, barley and wheat, accounting for 65%−75% of the yearly crop harvest. Ratooning is a traditional technique employed which permits a second sorghum cultivation during meher (UNDP Situation Report, 1999; Förch, 2003).

Sorghum [Sorghum bicolor (L.) Moench] is a staple crop in the drier parts of tropical Africa and it has been cultivated and consumed since prehistoric times. According to archaeobotanical studies, the oldest evidence for the domestication of sorghum in settlement contexts dates back to the last 6,000 years in the eastern Sahel region of Sudan (Winchell et al., 2018). However, even though the wild progenitor of sorghum is widely attested in Sudan (Beldados and Costantini, 2011; Beldados, 2015; Beldados et al., 2018), its cultivation in Ethiopia is more recent. The earliest evidence of carbonized grains of domesticated sorghum in Ethiopia is from the D-site of Aksum, dated to ca. 500–700 CE, and from the Tomb of the Brick Arches, dated to ca. 200–550 CE (Boardman, 2000). From post Aksumite and early medieval contexts, sorghum grains were recovered from the Islamic trade center of Harlaa (6th to 15th centuries CE) in Dire Dawa in east Ethiopia (Siraj, 2021; Beldados et al., 2024). Sorghum grains were also recovered in a context dated to the 12th and 13th centuries CE (Firie, 2023) in the Christian site of Gannata Mariam, adjacent to the World Heritage Site of Lalibella, in north central Ethiopia.

At present, 17 varieties of sorghum are cultivated in Konso and are characterized by different colors, texture, size, and productivity. Out of the 17 varieties, we selected 6 that are the most widely available, productive, and fast-ripening. These varieties are commonly stored in every farming household, whereas others are more difficult to obtain. Agronomists overseeing farming practices in the region also encourage farmers to prioritize these varieties. In addition, both agronomists and government agents provide free seeds to farmers to promote their cultivation (Table 1, Figure 2). Sorghum is intercropped with perennial legumes such as beans, bush beans, lablab bean, pigeon pea and is consumed as kitcha (flat bread made from wheat flour), dama (balls of sorghum flour boiled and stuck together usually accompanied with halako of boiled moringa leaves) and chaqqa (traditional fermented alcoholic drink which is strong and full of carbohydrates).

The Tigrayan landscape includes four agroclimatic zones: kola, woina dega, dega and wurich (Haftom et al., 2019). Despite this classification, environmental conditions in Tigray are particularly influenced by differences in altitude and topography, resulting in microclimatic variations across small and large expanses. The mean annual temperature ranges from 15 to 25°C, and mean annual rainfall ranges between 500 and 750 mm (Hagos, 2003; Hadgu et al., 2013). Both agriculture and livestock are key economic activities in Tigray Regional State. Agriculture is the primary source of food and livelihood for about 80% of the population, with 75% of the people in Tigray being farmers. These farmers still rely on traditional methods to cultivate their land: oxen-drawn plows, hand-sowing seeds, hand and hoe weeding, and rain-fed crops (Biagetti et al., 2022). This traditional farming system has sustained the region for centuries. The main cereal crops cultivated in Tigray are barley (Hordeum vulgare L.), wheat (Triticum aestivum L.), t'ef [Eragrostis tef (Zucc.) Trotter], sorghum [Sorghum bicolor (L.) Moench.], and finger millet (Eleusine coracana Gaertn.). The two common C₃ plants, barley and wheat, are grown in almost all districts. Maize (Zea mays L.) is cultivated in low-lying areas below 1,900 m a.s.l. Additionally, leguminous plants such as horse bean (Vicia faba L.), field pea (Pisum sativum L.), grass pea (Lathyrus sativus L.), lentil (Lens culinaris Medik.), chickpea (Cicer arietinum L.), and fenugreek (Trigonella foenum-graecum L.) are cultivated to minimize cereal crop failures (D'Andrea et al., 2018; Tsegaye, 2019).

Unlike the Konso Zone, the study of past and present agriculture in Tigray has been widely investigated. The earliest evidence of agriculture in the region is dated to 1,600 BCE. Archaeobotanical research from the Mezber site near Adigrat revealed macro-remains of mainly C₃ temperate cereals, legumes and oil crops including wheat, barley, lentils, flax (Linum usitatissimum L.) and weeds of cultivated cereals (Lolium sp.). In addition, two possible t'ef grains were identified in deposits dating from 400 BCE to CE 25 (Beldados et al., 2023; D'Andrea et al., 2023; Ruiz-Giralt and Beldados, 2024). Microbotanical investigations from sediments identified C₄ tropical/sub-tropical crops such as t'ef, sorghum and finger millet or wild grasses. Phytolith and starch analyses from artifacts from the site of Ona Adi further confirmed the presence of C₃ and C₄ plants, including wheat, barley, millets, lentils, and various root crops (Ruiz-Giralt et al., 2023a; Meresa et al., 2024).

Archaeological investigations in Central Tigray, specifically at the Aksum World Heritage Site, have yielded diverse botanical findings. Boardman (2000) recovered remains of hulled barley, linseed, emmer, lentil, pea, and sorghum from sites D and K during the Aksumite period (c. 150–330 CE). These sites also provided insights into agricultural practices, revealing the cultivation of C₄ plants such as sorghum alongside C₃ plants, legumes, and oil plants. D'Andrea et al. (2008) conducted further studies at Bieta Giyorgis, documenting plant remains spanning the Proto-Aksumite (400–50 BCE), Early to Classic Aksumite (50 BCE-CE 340), and Post-Aksumite (700–900 CE) phases. Significant findings included free-threshing wheat, barley, t'ef, lentil, and flax during the Aksumite phase, along with a variety of wild plants. These archaeological excavations provide valuable insights into the agricultural practices and botanical diversity of the Aksumite civilization in Central Tigray (Boardman, 2000; D'Andrea et al., 2008).

Similar to Konso, C₄ plants such as sorghum and millets are widely cultivated in Tigray. The crops are known for their drought resistant nature and can grow in diverse environmental conditions, including in places with poor soils and unpredictable rainfall. In the highlands and adjacent lowlands of the northern horn of Africa these crops are well adapted to the semi-desert environmental conditions (Ruiz-Giralt et al., 2023a). This region is also considered to be the domestication milieu of finger millet (Fuller and Hildebrand, 2013; Fuller, 2014).

Results of microbotanical analyses from the sites of Mezber and Ona Adi in Eastern Tigray identified C₄ tropical plants of African origin as Andropogoneae, Eragrostideae and Panicoideae in a context dated to 1,600 BCE and subsequently in chronological phases of Early Pre-Aksumite (850–750 BCE), Middle Pre-Aksumite (600–400 BCE) and Late Pre-Aksumite (400 BCE-1 CE) periods (Ruiz-Giralt et al., 2023b). In the western part of Tigray, sorghum and millet grains were identified at Aksum from Classic Aksumite contexts (Boardman, 2000) and finger millet from Bieta Giorgis during the Early Aksumite phase (D'Andrea et al., 2008). All these archaeobotanical studies from the region indicate that sorghum and millets have deep rooted cultivation and domestication history comparable with the C₃ plants of Southwest Asian origin.

Sorghum and finger millet are part of the daily cuisine in Tigray. The crops are consumed in the form of injera (flat bread baked in a griddle), bread, porridge, kollo (roasted cereals, mainly sorghum). Both of the crops are also consumed as part of the components in the preparation of local/traditional alcoholic drinks like suwa (mild alcoholic drink) and araki (strong alcoholic drinks). On the basis of color, two types of sorghum are cultivated in the region: red and white. Sorghum and millets are grown at an elevational range of 900–2,500 m a.s.l. in places where rainfall ranges from 500 to 1,200 mm per annum (Terefe and Beldados, 2021).

Each plant corresponds to a sample, and each sample is composed of seven grains, except for finger millet, for which more than 40 grains were selected due to their small size. The grains of each sample were weighted and ground by hand in an agate mortar and homogenized. Carbon and nitrogen isotope analysis were undertaken using a Europa Scientific ANCA-SL elemental analyzer (EA), with zero-blank autosampler, coupled to a Europa Scientific 20–20 isotope ratio mass spectrometer (IRMS) at Iso-Analytical (Sercon Ltd., Crewe, UK). The powdered and homogenized grains were weighted (1.5 ± 0.1 mg) in tin capsules (8 × 5 mm), sealed, and loaded together with standards into the Europa Scientific elemental analyzer. A subset representing 20% of the samples was analyzed in duplicate. Carbon and nitrogen isotope values are calibrated to V-PDB and AIR using IA-R001 standard (wheat flour, δ¹³C_{V − PDB} = −26.43‰ ± 0.08, δ¹⁵N_AIR = 2.55‰ ± 0.22). Carbon and nitrogen measurements uncertainty was monitored using two check standards: IA-R045/IA-R005 (mixture of ammonium sulfate and beet sugar, δ¹³C_{V − PDB} = −26.03‰ ± 0.11, δ¹⁵N_AIR = −4.71‰ ± 0.07) and IA-R046/IA-R006 (mixture of ammonium sulfate and cane sugar, δ¹³C_{V − PDB} = −11.64‰ ± 0.03, δ¹⁵N_AIR = 22.04‰ ± 0.06). IA-R001 is calibrated against and traceable to IAEA-CH-6 (sucrose, δ¹³C_{V − PDB} = −10.45‰) and IAEA-N-1 (ammonium sulfate, δ¹⁵N_AIR = 0.40‰). IA-R005 and IA-R006 are calibrated against and traceable to IAEA-CH-6. IA-R045 and IA-R046 are calibrated against and traceable to IAEA-N-1. IAEA-CH-6 and IAEA-N-1 are inter-laboratory comparison standards distributed by the International Atomic Energy Agency, Vienna.

The isotope ratios are reported as “delta” defined according to International Union of Pure and Applied Chemistry (IUPAC):

where Rs and Rst are the isotope ratios ¹⁵N/¹⁴N and ¹³C/¹²C of the isotopic abundances of ¹⁵N, ¹⁴N and ¹³C, ¹²C and ‰ = 10⁻³. Using the calculation provided by Szpak et al. (2017) based on repeated measurements of calibration standards, check standards and sample replicates, the precision [u(R_W)] of the measurements are estimated to be ± 0.05‰ for δ¹³C and ± 0.06‰ for δ¹⁵N values. Accuracy or systematic errors [U(bias)] are determined to be ±0.08‰ for δ¹³C and ± 0.17‰ for δ¹⁵N values, on the basis of the difference between the observed and known δ values of the check standards and the long–term standard deviations of these check standards. The total analytical uncertainties [u(c)] are estimated to be ± 0.10‰ for δ¹³C, ± 0.18‰ for δ¹⁵N values (Supplementary information 2: excel calibration calculations).

Extractions were performed on a subset of chaff samples (n = 7) that exhibited an adequate quantity of material for analysis (Table 4). Chaff samples are from the same panicles from which the isotopic measurements of the seeds were taken. The samples underwent an extraction process following a protocol that integrates a dry ashing technique followed by wet oxidation. The procedure has been published step by step in D'Agostini et al. (2022). The results are reported as the weight in grams of biosilica extracted from the dry tissue and as a percentage of the total tissue weight.

The amount of carbon (%C) and nitrogen (%N) and the stable isotope values from sorghum and finger millet are shown in Figure 4 and presented in Table 4. The %C span from 39.16 and 42.30 (40.83% ± 0.91%) and the %N spans from 1.23 and 2.91 (1.98% ± 0.55%). δ¹⁵N values span from 1.27‰ to 5.47‰ (3.24‰ ± 1.35‰) with a range of 4.2‰ and δ¹³C values span from −13.62‰ to −10.22‰ (−12.11‰ ± 1.00‰), with a range of 3.4‰. Examining the distribution of sorghum values, the δ¹⁵N ranges of plants cultivated at different altitudes largely overlap. In contrast, the δ¹³C ranges show almost no overlap and significantly increase with altitude. Correlations between the isotopic values and maximum, minimum and average temperatures, total precipitation, altitude and relative humidity recorded during the growing period of the plants were evaluated using the Tau B Kendall test (Table 5). The δ¹³C values show a significant positive correlation with altitude (Tau B Kendall = 0.569, p = 0.027, n = 11, Table 5), which is even stronger when only sorghum grains are considered (Tau B Kendall = 0.685, p = 0.012, n = 10, Figure 5). The δ¹⁵N values show no correlation with any of the environmental factors or the altitude of the sites (Table 5, Figure 4). Particular attention is due to finger millet Keyih Degussa from Adigrat (RET 40): even though only one sample is available, results are interesting because they show an opposite trend compared to the samples of sorghum from the same altitude. This is because the δ¹³C value falls in the range of the samples from the site placed at the lowest altitude, and the δ¹⁵N value is the highest value of all samples analyzed.

The extracted biosilica ranges from ~2% to ~7% of the total dry weight, showing some variability despite the samples belonging to the same species and sharing the same C₄ photosynthetic pathway. The highest values were observed in the Shuleya variant (RET 33) with a total of 6.96% (w/w) of biosilica, while the lowest value corresponds to the Meshila (RET 39) with 2.11% (w/w). Complete results can be found in Table 4. In Figure 6, the biosilica data are shown in comparison to the C:N data. The graph indicates that the samples with high C:N ratio have low biosilica content (RET 30 and 39 with the highest C:N ratio and the lowest biosilica content), while RET 33 with the highest biosilica content have the lowest C:N ratio. Figure 6 shows a general negative relationship between biosilica and C:N.