**4. Discussion**

This study was carried out in two different locations; lowland and highland, showed that this special grass pea collection has great variation with respect to seed yield, yield components, quality and forage traits. Obtaining a higher seed yield for different environmental conditions is one of the most important challenges in plant breeding [19]. However, improving the traits related to yield characteristics such as double podding, more seeds per pod, plant height or branches are also highly critical to obtain desired grass pea lines [4]. Our collection was characterized by three qualitative and 19 quantitative traits to develop desired cultivars. Similarly, different germplasm resources have been characterized with different agro-quality traits in grass pea [20,22,25,33,34].

Based on comparison between their altitudes, the mean number of days to the first flowering was found to be shorter in lowland environmental conditions. Altitude levels and sun exposure times are thought to be responsible for these differences [35]. The results obtained in our study were found to be higher when compared to the studies conducted in European countries and India [35–37]. Furthermore, the highest day of 50% flowering was determined as 213.3 days (GP107) in highland (Table S5) and the lowest mean was 120.5 days (GP237) in lowland (Table S6). While our findings regarding the flowering period are in agreement with the results of Çakmakçı and Çeçen [38], ¸Seydo¸so ˘glu et al. [39] and Öten et al. [40], they are higher than the findings of Kumari [37]. Grela et al. [41] stated that variation of these traits depends on the environmental factors, especially on soil type and precipitation amount during vegetation period. Plant height effects both seed and biological yield in grass pea. In the present study, plant height ranged from 41.5 cm to 90.8 cm in lowland conditions and 15.6–76.8 cm in highland conditions. In the research of grass pea, it was reported that the plant height was determined between 24.5 cm and 172.0 cm by different studies [35,42,43]. These results clearly revealed that environmental differences highly affect the plant height trait of grass pea. We also had large variation for the number of branches and comparable results were also observed as 6.10–13.00 [44], 3.73–6.00 [45], 4.6–8.6 [46]. The hundred-grain weight is considerable in terms of giving an idea about the grain's size, fullness, thinness and flour yield. We obtained the maximum value as 23.8 in highland and 29.39 in lowland. These are higher values than those obtained by Aksu et al. [43] and Ba¸saran et al. [44]. However, Grela et al. [35], Ribinski et al. [47] monitored higher weights in this trait compared to our maximum results. Biological yield frequently used as yield selection criteria, especially in studies to increase grain yield in cereal and legume plants [48]. The maximum biological yield was detected in GP104 and GP145 in lowland and highland, respectively. These genotypes, therefore, should be used in

cultivar development program and also as parents in crossing programs to obtain superior lines. Looking at the seed yield, we observed increased mean values compared to the check cultivar average values. Moreover, in comparison with Antalya and Isparta for seed yields, we observed significant differences between genotypes (Table S8). According to Pandey et al. [34], seed yield per plant in grass pea ranged from 0.5 to 19.7 g, whereas Ribinski et al. [47] claimed that it ranged from 7.20 to 21.19 g. Especially, we obtained lower values in lowland compared to highland and these previous studies. According to Das et al. [15] one of the important factors influencing seed yield is ecological difference. In addition, PCA analysis demonstrated that seed yield and biological yield had high and positive values in PC1 both environmental areas. Similar these traits positively contributed to PC was obtained by Polignano et al. [49] when they characterized of genetic diversity of grass pea entries.

Genotype x environment effects play an important role in the phenotypic expression of ODAP content, which is polygenically inherited [50]. Low ODAP is frequently linked to undesired features such as late flowering and low seed and biological yields [51]. Therefore, new breeding programs have been successful in achieving both low ODAP, high yield and protein in recent years [5]. The β-ODAP content of the genotypes grown in lowland (Antalya) and highland (Isparta) regions has significantly different (Tables S7 and S9). We obtained large variation for β-ODAP content among the genotypes and lands and it is showed that this trait is affected by both genetic and environmental factors [52,53]. The genotypes GP2 and GP49 genotypes in collection are notable for having low β-ODAP concentrations in both Antalya and Isparta. These accessions therefore provide better opportunities for developing high seed yielding and low B-ODAP cultivars suitable for studied regions. Previous studies indicated that there was no genotype of grass pea that was β-ODAP free, although in several genotypes the β-ODAP content was low [6]. Futhermore, many researchers [14,54,55] found that these low-toxin cultivars did not have stable -ODAP levels in grass pea seeds when grown under different environmental conditions. The β-ODAP content in grass pea seed had high variation depends on genotype and environmental conditions and it ranged from 0.02 to 7.2% [56]. Onar et al. [57] reported that they found the amount of β-ODAP in local grass pea varieties grown in Turkey, ranged from 0.10% to 0.87% (*w*/*w*). Arslan et al. [27] investigated the β-ODAP contents of 173 local grass pea genotypes in lowland conditions and obtained values ranging from 1.55 mg/g to 20.8 mg/g, showing the genotype effect on this trait.

Crude protein content is a significant indicator of feed quality [58]. The minimum crude protein content in ruminant diet should be around 6.0 to 8.0% of dry matter for adequate activity of rumen microorganisms [59], suggesting that hay crude protein content in investigated grass peas is more than twice or thrice the needed ratios. The highest crude protein ratios were found in GP242 (24.64%) and GP248 (24.08%) genotypes in highland (Table S9), GP53 (27.09%), GP40 (26.88%) and GP270 (26.55%) in lowland (Table S7). There were significant differences among genotypes grown in two different locations (Table 2). Differences in crude protein ratios of the genotypes were mainly resulted from plant genetics, but leaf, spike and stem ratios, ripening periods, temperature and fertilization might also have significant effects on crude protein contents [60]. Present crude protein contents of some genotypes were similar to the findings of Basaran et al. [61]. In fact, Basaran et al. [61] stated that due to ecological differences in the regions where grass pea genotypes are collected, variation in seed crude protein concentrations can be linked to ecological factors rather than genetic variation.

Increasing ADF ratios reduces digestibility of the feeds and increasing NDF ratios reduces feed intake and make the animals feel full, thus limit feed intake and feed availability. Since high ADF and NDF ratios have negative effects on feed intake and digestibility, feeds with ideal ADF and NDF values are usually preferred [62]. ADF content of grass pea varied between 6.87 and 9.74% in lowland, while they varied between 7.19 and 10.22% in highland. Lower ADF values are preferred for animal production due to the negative correlation between ADF values and ruminant digestion [63]. Therefore, genotypes having

the lowing values should be taken into consideration for forage breeding. NDF content of grass pea varied between 11.15 and 17.58% in lowland, while it varied between 11.52 and 24.23% in highland. Grela et al. [35] found NDF between 11.25 and 18.92%, while Karadag and Yavuz [63] found it between 10.18 and 13.55%. The monogastric and ruminants should have lower NDF content in their feed. Furthermore, ruminant animals may require a certain amount of NDF values, but higher NDF values may reduce animal intake [63]. The TDN refers to the nutrients that are available for livestock and are related to the ADF concentration of the forage [64]. As ADF increases, there is a decline in TDN which means that animals are not able to utilize the nutrients that are present in the forage [65]. In Antalya ecological conditions, line GP270 had the highest TDN and DDM values, while line GP248 had the lowest. When the averages of both regions in terms of TDN, DMI, DDM and RFV were compared, it was determined that the values in Antalya were higher. According to the Hay Market Task Force of American Forage and Grassland Council standards, the genotype is classed as premium quality when it has protein content > 19, ADF < 31%, NDF < 40%, and RFV > 151. The scale showed that lots of genotypes in the collection should be classed as premium with regard to forage quality. Considering all the results, GP60 in lowland and GP40 in highland were considered the most promising lines for grass pea breeding with their high crude protein content, low ADF, NDF and β-ODAP ratios, as well as biological yield.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/agronomy12102426/s1, Table S1: The initial genetic material of the study; Table S2: Means of 11 agronomic traits for grass pea collection produced in the 2020 and 2021 growing seasons, Antalya (lowland), Turkey; Table S3: Monthly temperature, precipitation and humidity values of experimental area in Antalya (lowland); Table S4: Climatic data for experimental area in Isparta (highland); Table S5: Means of 11 agronomic traits for grass pea collection produced 2021 growing season, Isparta (highland), Turkey; Table S6: Analysis of variance of augmented block design for 11 agronomic traits in grass pea collection produced 2021 growing season in Isparta (highland), Turkey; Table S7: Means of quality and forage traits for grass pea collection produced in the 2020 and 2021 growing seasons, Antalya (lowland), Turkey; Table S8: Analysis of variance of augmented block design for forage and quality traits in grass pea collection produced 2021 growing season in Isparta (highland), Turkey; Table S9. Means of eight quality and forage traits for grass pea collection produced in the 2021 growing season, Isparta (lowland), Turkey

**Author Contributions:** Conceptualization, M.A. and E.Y.; methodology, E.Y.; validation, M.A. and M.T.; formal analysis, E.Y.; investigation, M.A.; resources, M.A.; data curation, M.A and M.T.; writing—original draft preparation, M.A. and E.Y.; writing—review and editing, E.Y.; visualization, E.Y.; supervision, E.Y.; project administration, M.A.; funding acquisition, M.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by the Scientific Research Projects Coordination Unit of Akdeniz University, Turkey with the project code: FBA-2020-5294.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We are grateful to USDA, ARS Plant Genetic Resources Conservation Unit and International Center for Agricultural Research in the Dry Areas (ICARDA) for supplying genetic material several times.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
