Kodo millet (Paspalum scrobiculatum L.) is one of the ancient small millets cultivated in India, known for its excellent adaptation to dry land farming and poor soils. It is widely grown in tribal and rainfed regions, where it serves as both a food and fodder crop. Kodo millet is nutritionally rich, containing high levels of dietary fiber, protein and essential minerals, making it an important component for nutritional security (Saleh et al., 2013, Deshpande et al., 2015; Rao et al., 2017; Nirubana et al., 2020; Divya and Garg, 2024; Mathur et al., 2025). Additionally, kodo millet has been shown to have a range of potential health benefits, including improved digestive health, weight management and reduced risk of chronic diseases such as cardiovascular disease and cancer. In recent years, there has been an increasing interest in functional foods, which are foods that provide additional health benefits beyond basic nutrition. Kodo millet has been identified as a potential functional food due to its nutritional composition and health-promoting properties (Chauhan et al., 2021). Studies revealed that besides having a nutritional and ecological advantages of Kodo millet, it remains underutilized compared to other cereals such as rice and wheat (Gowda et al., 2022).

Evaluating morphological diversity along with yield related traits within Kodo millet germplasm is crucial for identifying superior genotypes with desirable traits such as higher yield, stress tolerance and adaptability is essential to establish a breeding program (Sood et al., 2019). Morphological characterization provides fundamental insights into phenotypic variation and genetic relationships among genotypes, which are essential for crop improvement programs (Mishra et al., 2024, Upadhyaya et al., 2014). 

Principal component analysis (PCA) is a statistical method that transforms correlated variables into a smaller set of uncorrelated principal components that explain most of the variation in the dataset. It helps reduce data dimensionality, identify underlying patterns and simplify data interpretation (Abdi and Williams, 2010; Jolliffe, 2002).

Therefore, this study was undertaken to assess the morphological diversity of Kodo millet genotypes and to identify variability that can be effectively utilized for breeding and conservation purposes and also give an overall view about which variety gives better yield under same climatic conditions.

Experimental material
 
A total of 11 kodo millet varieties (KMV1 to KMV10 and RK 390-25) were evaluated which were collected from different places of India. KMV1 from Gujrat, KMV2, KMV6 and KMV8 from Maharashtra, KMV3 from Uttar Pradesh, KMV4 and KMV5 were from Madhya Pradesh, KMV7 Jharkhand, KMV9 and KMV10 from Punjab and RK 390-25 from Uttar Pradesh.
 
Location of experiment and climatic condition
 
The experiment was conducted during the Kharif seasons of 2022, 2023, 2024 and 2025 at the research field of Shri Guru Ram Rai University, Dehradun, Uttarakhand, India. Eleven Kodo millet genotypes were evaluated for morphological variability using a randomized block design (RBD) with three replications (Gomez et al., 1984). Each genotype was grown in 3 m × 2 m plots with 25 cm row spacing and 10 cm plant spacing.
 
Data collection
 
Observations were recorded on five randomly selected plants from each variety in each replication to obtain reliable and representative data. The following quantitative traits were measured: Days to 50% flowering (D50%), plant height (cm) (PH), number of basal tillers per plant (BT), panicle length (cm) (PL), peduncle length (cm) (PDL), grain yield per ear head (g) (GY), 1000-grain weight (g) (GW) and grain yield per hectare (kg/ha) (GYH). These traits were selected based on their importance in determining growth, yield and overall performance of millet genotypes (Saleh et al., 2013; Upadhyaya et al., 2014).

Qualitative descriptors were also recorded to capture the distinct morphological features of the kodo millet genotypes. The qualitative traits recorded were as follows: Plant growth habit (PGH), leaf attitude (LA), panicle appearance (PA), panicle exertion (PE), spikelet arrangement (SAR), seed index texture (SIT), seed density (SD), grain colour (GC) and grain shape (GS). Each trait was scored according to standardized descriptor scales to maintain consistency and enable comparative analysis across genotypes.

The statistical analysis demonstrates that all studied germplasm exerted a highly significant influence (p<0.01) on all measured growth, development and yield components of the kodo millet crop (Table 1). The results showed highly significant differences in germplasm for all studied traits- plant height, basal tillers, days to 50% flowering, panicle length, peduncle length, grain yield per ear, 1000-grain weight and grain yield (kg/ha), as indicated by significant F-values. This highlights the presence of genetic variability among the genotypes evaluated, suggesting good scope for selection and crop improvement. The comparatively low error mean squares indicate good experimental precision and reliability of the observations. The standard error [SE(±)] and critical difference (CD at p=5%) values further support the statistical significance and assist in comparing genotypes performance across traits.


 
Morphological variability among kodo millet varieties
 
The qualitative evaluation of Kodo millet genotypes revealed a high level of visible diversity across all recorded traits. Differences were evident in plant structure, leaf orientation, panicle characteristics and grain features, suggesting that the studied material possesses a broad genetic base that can be effectively utilized in breeding programs.

Plant growth habit showed variation among genotypes (Table 2). Most of the varieties exhibited a prostrate growth pattern, while a three (KMV1, KMV6 and KMV9) were erect and two (KMV3 and KMV7) were classified as decumbent. From an agronomic perspective, erect types were generally advantageous because they allow better light penetration and air movement within the crop canopy. In contrast, prostrate and decumbent types may offer better tolerance to environmental stress and reduced risk of lodging, which can be beneficial under marginal growing conditions.



Leaf attitude also differed considerably. Droopy leaves were more commonly observed, whereas erect leaves were limited to a few genotypes such as KMV1, KMV6, KMV7 and KMV9. Erect leaves tend to improve light use efficiency, especially under dense planting, by minimizing shading between plants. This trait can indirectly support higher productivity. Upadhyaya et al. (2011) reported that leaf orientation plays an important role in determining biomass accumulation and productivity in millet crops. Therefore, leaf attitude may serve as a useful indirect selection criterion for improving yield potential in Kodo millet breeding programs.

Variation in panicle traits was another important observation. Open panicles were the most frequent, followed by compact and semi-compact types. In addition, panicle exertion varied between complete and partial types. Genotypes with fully exerted panicles are generally preferred, as they facilitate better pollination and grain filling. Spikelet arrangement was predominantly regular, although a few genotypes (KMV4, KMV8, KMV10 and RK 390-25) showed irregular patterns, which may reflect adaptation to specific environmental conditions. Similar relationships between panicle morphology and productivity have been reported by Muthamilarasan and Prasad (2015) and Varatharajan et al. (2017), who identified panicle length as one of the most important yield-contributing traits in kodo millet. Grain and seed traits also contributed to the observed diversity. Most genotypes have high seed density, dark brown grain colour and oval grain shape. Dark-coloured grains were often associated with better nutritional quality, particularly in terms of antioxidant content, which adds value from both a health and market perspective.
 
Yield performance and morphological traits relationships
 
The analysis of quantitative traits revealed noticeable differences in yield and its contributing factors among the varieties. Grain yield (kg ha^-1) varied widely, with KMV4 clearly outperforming all other genotypes. This variety recorded the highest yield and also showed favorable values for several important traits, including plant height, panicle length, peduncle length and grain yield per ear head.

The quantitative evaluation revealed significant differences in yield and yield-attributing characters among the genotypes (Table 3 and Table 4). Grain yield per hectare ranged from 1645 kg ha^-1 in KMV2 to 2550 kg ha^-1 in KMV4, indicating substantial variation for productivity under the agro-climatic conditions of the Doon Valley. The superior performance of KMV4 may be attributed to its favorable combination of plant height, panicle length, peduncle length and grain yield per ear. RK 390-25 and KMV7 also demonstrated relatively high yields, followed by KMV5, KMV1 and KMV8 with moderate performance. In contrast, KMV2 and KMV9 exhibited comparatively lower grain yields, suggesting reduced productivity under the experimental conditions. The observed variation in grain yield among the varieties indicates the presence of substantial genetic diversity for yield-related traits and highlights the potential for selecting superior genotypes for crop improvement and breeding programs in Kodo millet, which is in accordant with the findings of Vetriventhan et al., (2015); Varatharajan et al., (2017) who observed that variations in panicle and plant traits significantly influence grain yield in kodo millet. Nireekshitha et al., (2024) also reported substantial diversity among genotypes for yield-related traits, suggesting considerable scope for selecting high-yielding lines.





Plant height appeared to have a positive influence on yield, as taller varieties such as KMV4 and KMV5 performed better than shorter ones like KMV9. However, the number of basal tillers did not show a straight forward relationship with yield. Although KMV6 and KMV7 produced a higher number of tillers, this did not always translate into higher productivity, indicating that the effectiveness of tillers is more important than their number (Table 4). Similar positive associations were reported in plant height and grain yield by Sharma et al. (2016) and Thakur et al. (2020), indicating that plant height can serve as an effective indirect selection criterion in millet breeding. However, optimum rather than excessive plant height is desirable to minimize lodging risks while maximizing productivity.

The number of basal tillers showed comparatively weaker associations with grain yield. Although genotypes such as KMV6 and KMV7 produced a greater number of tillers, their productivity was not proportionately higher than that of KMV4. This suggests that productive tillers contribute more significantly to yield than the total number of tillers. Similar findings were reported by Thakur et al. (2020), who observed that tiller number alone does not guarantee higher productivity unless accompanied by efficient grain-bearing capacity. Studied performed by Bhat et al., (2010) and Govindaraj et al., (2015), who emphasized the usefulness of morphological descriptors in identifying genetically diverse and agronomical superior millet genotypes. The observed variation in plant architecture also suggests the potential for selecting specific growth habits suitable for different agro-climatic conditions.

Days to 50% flowering exhibited negative correlations with most yield-related traits, particularly grain yield per hectare. The results indicate that delayed flowering may adversely affect productivity under subtropical conditions. Early and medium-duration genotypes are likely to utilize available resources more efficiently and avoid terminal environmental stresses during grain filling. Thus, flowering time also played a role in yield differences. Varieties with moderate flowering duration, such as KMV4, tended to perform better than those that were too early or too late. This suggests that a balanced growth period allows for better resource utilization and grain development. Comparable observations were reported by Goron and Raizada (2015), who highlighted the significance of flowering duration in determining adaptation and productivity in small millet crops.

Among the yield-contributing traits, panicle length and peduncle length showed a strong association with productivity. Varieties with longer panicles and peduncles generally produced higher yields, likely due to improved grain formation and filling. On the other hand, 1000-grain weight showed relatively little variation across genotypes, suggesting that yield differences were mainly driven by the number of grains rather than their size. Grain yield per ear head closely followed the same trend as total yield, further confirming its importance as a key contributing trait.

When all traits were considered together, KMV4 emerged as the best-performing genotype, combining high yield with favorable morphological and agronomic characteristics. Varieties such as KMV7, KMV5, KMV8 and KMV1 also showed good performance and can be considered promising. KMV10, KMV6 and KMV3 showed moderate performance, while KMV2 and KMV9 were comparatively lower yielding. Overall, the study highlights the presence of significant variability among Kodo millet genotypes and identifies key traits that contribute to improved performance. These findings can be effectively used for selecting superior genotypes and designing breeding strategies aimed at enhancing yield and adaptability.

The evaluation of nine qualitative morphological descriptors revealed substantial phenotypic variability among the studied Kodo millet genotypes. The descriptors exhibited distinct classification patterns, indicating their effectiveness in differentiating genotypes. The use of modal values across replications ensured consistency in qualitative trait assessment and minimized environmental influence on trait expression.
 
Correlation analysis
 
The correlation analysis provided useful insights into the relationships between grain yield and its contributing traits (Table 5 and Fig 1). Grain yield per hectare (GYH) exhibited strong positive correlations with plant height (0.81), panicle length (0.78), peduncle length (0.80) and grain yield per ear (0.92), indicating that these traits contributed positively to overall productivity. Plant height also showed positive associations with panicle length (0.68), peduncle length (0.72) and grain yield (0.75), suggesting that taller plants tended to produce better yield attributes. In contrast, days to 50% flowering (D50%) displayed negative correlations with most yield-related traits, particularly grain yield per hectare (-0.45), indicating that delayed flowering may adversely affect yield performance. Overall, the correlation analysis highlights panicle length, peduncle length and grain yield per ear as important traits for selection in kodo millet improvement programs. Similar relationships have been reported by Thakur et al., (2020); Mehraj et al., (2025); Varatharajan et al., (2017) and Nireekshitha et al., (2024), who identified these traits as important selection criteria for yield enhancement in Kodo millet.





Besides using correlation tabular data, the heatmap reveals strong positive associations among plant height (PH), panicle length (PL), peduncle length (PDL), grain yield per ear (GY) and grain yield per hectare (GYH), suggesting that improvement in these traits may contribute positively to overall yield performance. Similar positive associations among yield-contributing traits have been reported in Kodo millet and other small millet species by Thakur et al., (2020); Varatharajan et al., (2017) and Vetriventhan et al., (2015). In contrast, days to 50% flowering (D50%) exhibited negative correlations with most yield-related parameters, indicating that delayed flowering may adversely affect productivity under the prevailing agro-climatic conditions. Comparable findings have been documented by Goron and Raizada (2015) and Nireekshitha et al., (2024), who reported that earlier or moderately flowering genotypes often exhibit superior yield performance. The heatmap effectively highlights the interrelationships among traits and facilitates the identification of key characteristics that can be targeted in selection and breeding programs aimed at yield enhancement in Kodo millet.

The PCA plot (Fig 2) revealed considerable diversity among the Kodo millet genotypes based on agronomic and yield-related traits. Genotypes such as KMV4 and KMV9 showed greater divergence, while KMV5 and KMV7 exhibited similar trait expressions. KMV4 showed a strong positive association with PC1, indicating superior yield-related performance. The observed diversity provides useful opportunities for genotype selection and breeding. Similar findings were reported by Vetriventhan et al., (2015), Nirubana et al., (2019) and Nireekshitha et al., (2024).

The observed variability in grain colour, grain shape and seed density highlights the richness of the evaluated germplasm. These traits are important not only for genotype identification but also for consumer preference, nutritional quality and market acceptance. Previous studies by Deshpande et al., (2015); Kumar et al., (2016); Gupta et al., (2017); Rangaiah et al., (2024) and Mathur et al., (2025) emphasized the nutritional significance of millet grain characteristics, particularly in relation to dietary fibre, antioxidants and mineral content.

Overall, substantial variability was observed among the evaluated Kodo millet germplasm for qualitative, morphological and yield-related traits, indicating good scope for selection and genetic improvement. KMV4 emerged as the best-performing germplasm for grain yield, recording the highest yield (2550 kg ha^-1), along with the highest plant height (93.0 cm), favourable panicle length (7.5 cm), longest peduncle length (10.9 cm) and high grain yield per ear (2.10 g). Therefore, KMV4 can be further evaluated as a potential donor for improving grain yield and associated yield-contributing traits in future breeding programmes. KMV7 was notable for higher basal tiller number (26), while KMV5 showed favourable plant height (86.5 cm) and grain yield (2055 kg ha^-1), making these genotypes useful for selection for vigorous plant growth and yield-related traits. KMV1, with an erect growth habit and erect leaf attitude, may also be useful in breeding programmes targeting desirable plant architecture. Correlation analysis further confirmed the importance of plant height, panicle length, peduncle length and grain yield per ear for improving grain yield. The diverse germplasm identified in this study can therefore be conserved and utilized as potential parental material for developing high-yielding and climate-resilient Kodo millet varieties.

This work was supported by Shri Guru Ram Rai University through the Seed Money Funded Grant Project. The authors express their sincere gratitude to the Department of Botany, School of Basic and Applied Sciences, Shri Guru Ram Rai University, Patel Nagar Campus, Dehradun, Uttarakhand, for their valuable support and facilities provided during the course of this research.
 
Disclaimers
 
The views and conclusions expressed in this research article are solely those of the authors and do not necessarily reflect the views of their affiliated institution. The authors are responsible for the accuracy and completeness of the information presented, but they do not accept any liability for any direct or indirect losses resulting from the use of this content.

Authors have declared that no competing interests exist.