Abstract: Path analysis was performed on plant characters in spaced plants of land races alfalfa (Medicago sativa L.) to determine: seed yield per plant, stem number, internodes number, plant height, blooming periods, leaf length, raceme length, number of pods per raceme, number of seeds per raceme, number of seeds per pod, 00-seed weight. Positive relationships existed between seed yield and all its components with the exceptions of the internodes numbers, blooming periods and 1000-seeds weight. Significant correlation with seed yield was found for plant height, number of pods per raceme and the number of seed per raceme. Raceme length coefficient was highly significant (R = 0.786**), very high negative correlation found with the blooming periods (R = -0.625). Plant height, seeds per raceme, seeds per pods and 1000-seed weight had strong direct effects, while other components had strongly negative or negligible direct effects. The indirect effects via leaf length and blooming period were major determinants of the main effects of all components for the seed yield of the alfalfa. Positive direct effect of plant height, raceme length, number of pods per raceme and number of seed per raceme suggested that these yield components may be good selection criteria to improve seed yield of alfalfa cultivars.
INTRODUCTION
Alfalfa (Medicago sativa L.) is cultivated mainly for forage production and seed yield. Generally, in Turkey, eastern Anatolia farmers produce their seeds directly, harvesting them from the crop regrowth following the first cut o f herbage during the last year of the crop (Serin and Tan, 2001). Barnes and Sheaffer (1995), pointed out high-quality certified seeds, that is with high levels of genetic and physical purity and vital quality, are essential for obtaining successful stands under all field conditions. Similarly genetic variation was identified for pollen fertility (Viands et al., 1998), ovule fertility and ease of tripping (Rosellini et al., 1998). Thus, forage production would benefit from specialization in seed production directed at the consistent, reliable production of heavy yields of high-quality seed. According to Rincker et al. (1988) successful alfalfa seed production is favored in regions where the growing season is characterized by low relative humidity and moderate to high temperatures. These climatic conditions favor long periods of pollinator activity, low incidence of leaf disease and favorable harvest conditions. Temperature and rainfall patterns in eastern Anatolia, Turkey are suitable for successful seed setting. Therefore, forage seed production has the potential to become a profitable farming activity, supplying both domestic and export markets. However, the environmental conditions during seed development, genetic characteristics and agronomic techniques have considerable effect on seed yield and components of yield through their effect on plant reproductive physiology (Hacquet, 1990, Martiniello, 1998). Thus, the seed production potential can only be realized through attention to a choice of good environment and the adoption of proper management practices. Seed yield is a complex trait. It is the product of a number of individual yield components (Bolanos-Aguilar et al., 2000-2002). To clarify the nature of the complex relationships among traits, path analysis can be used by partitioning the relative contribution of yield components via standardized partial-regression coefficients (Sengul, 2006). This method also reveals the compensation mechanisms among yield components resulting from changing genotypic, environmental and management factors (Kephart et al., 1997; Bernadette et al., 2000; Iannucci et al., 2002). Plant breeder could find well-qualified varieties or certain characteristics by using path analysis at the terminal selection stage of breeding (Falcinelli and Martinello, 1998).
This study was an attempt to lay the ground work for successful seed production of alfalfa by evaluating thirty land races alfalfa ecotypes in terms of seed yield and seed yield components, using path-coefficient analysis to determine the relationships of yield components to one another.
| Table 1: | Climatic data of the research location |
| LTA: Long term average (72 years 1929-2002) | |
MATERIALS AND METHODS
This research was carried out for 2 years (2000/2002) at the experimental fields of the Department of Field Crops, Faculty of Agriculture, Atatürk University, Erzurum, Turkey (39°55 ‘ N Lat. 41°16’ E Long. and 1950 m Elevation). This location is arid, characterized by dry, cool temperate summers and with a worm temperature and 187 mm rainfall during April to August, 46.5% of the annual average rainfall (Table 1).
The research was conducted with thirty alfalfa ecotypes collected from Eastern Anatolia Region at Turkey by Sengul (1995). Ten of them selected as an experimental cultivar and Kayseri cultivar commonly planted in Turkey were included as a control.
Then those plants were transplanted to the field on 20 May 2000 on silty clay loam organic matter 1.27%, CaCO3 level 7.5%, pH 8.29, P2O5 12.05 ppm. Plots consisted of 5 m long raised beds spaced 0.50 m apart. Plant spacing within rows was 0.50 m the seedbed was prepared by incorporating 122 kg ha-1P and 50 kg ha-1N as fertilizer. After stand establishment, weeds were controlled by hand. The plots were flood irrigated as needed.
Ten plants per plot (chosen randomly) were measured. The experiment design was a randomized complete block with five replicates. The following observations were made on seed yield plant-1 (SEY g); stem number (SN), internodes number (IN), plant height (PH cm), blooming periods (BP), leaf length (LL mm), raceme length (RL cm), number of pods per raceme (NPR), the number of seed per raceme (NSR), Number of seeds per pod (NSP), 1000-seed weight (TSW g).
Simple correlation analysis was carried out using SAS statistical program (SAS., 1996). The relative importance of direct and indirect effects on seed yield was determined by path analysis for the 2 years of data. In path analysis, seed yield was the dependent variable and the ten plant or plant characteristics (observations 5/10) were considered as independent variables.
RESULTS
Positive relationships existed between seed yield and all its components with the exceptions of the internodes number, blooming periods and 1000-seeds weight (r = -0.036, -0.625, -0.065, respectively) (Table 2). Significant correlation with SEY was found for plant height, number of pods per raceme and the number of seed per raceme. Raceme length coefficient was highly significant (r = 0.786) also very high negative correlation found with the number blooming period (-0.625, Table 2). In general, the components had significant positive correlations with each other. Only internodes number and seed number per pods had no significant correlation with any other components (Table 2). However, correlation coefficient alone did not reveal how individual yield components affected seed yield. Path coefficient analysis partitioned direct and indirect causes of association and measured the relative importance of each seed yield component across the other characters.
Path analysis showed that plant height, the number of seeds per raceme, NSP and TSW had strong direct effects on the seed yield while, other components had strongly negative or negligible direct effects. In other words, the main effects of most components were significantly positive and contributed to seed yield largely through indirect effects of leaf length and blooming period. In total of the variation in seed yield could be explained by the variation of the two independent variables (Table 3). The unexplained variation, may be due to variations in the other components under consideration, excess rain during flowering probably limited insect pollination and enhanced physiological losses of pollinated flowers and embryo abortion can be effected seed formation.
DISCUSSION
Very strong correlations between seed yield and almost its all components were found (Table 2).
| Table 2: | Simple correlation coefficients of seed yield components in alfalfa |
SEY: Seed Yield Plant-1; SN: Stem Number; In: Internodes
Number; PH: Plant Height; BP: Blooming Period; LL: Leaf Length; RL: Raceme
Length; NPR: Number of Pods per Raceme; NST: Number of Seeds per Raceme;
NSP: Number of Seed per Pods; TSW:1000-seed Weight |
|
| Table 3: | Analysis of direct and indirect effects of ten characters on seed yield per plant in land races alfalfa |
| * Percentage of direct and indirect effects | |
In particular, positive correlation coefficients were found for plant height (0.663*), raceme length (0.786**, NPR (0.698*) and NSR (0.640*) in that order. The studies with smooth brome grass and plant height (Serin et al., 1999), seeds/panicle and 1000-seed weight (Acikgöz and Tekeli, 1980; Sengul and Sağsöz, 1998) exhibited strong positive correlations with seed yield. Present results confirm the findings of Açikgöz and Tekeli (1980) and Serin et al. (1999) for seeds/panicle and plant height, but not for 1000 seed weight. Kamboj and Mani (1983) in triticale, found a non-significant correlation between seed yield and 1000 seed weight and their results are consistent with the results (Table 2).
In most previous studies, researchers usually concentrated on panicle measurements and components of individual plants grown in pots or spaced in plots. Seker and Serin (2004) studied on smooth brome grass and investigated the vegetative components with generative components on plot basis. They found very strong correlations between seed yield and almost its all components. In this study path coefficient analysis showed complex interrelations among seed yield components because SN, NPR, NSP, NPR and TSW were all important in determining seed yield in alfalfa, but most seed yield variation via direct effect of leaf length was negatively and plant height was positively. Our results are in agreement with that found by Frakes et al. (1961) and Iannucci et al. (2002) in alfalfa, Ball et al. (2001) in soybean (Glycine max L.) and Güler et al. (2001) in chickpea (Cicer arietinum). In this research leaf length, raceme length and 1000-seed weight strongly influenced seed yield. However, the positive indirect effect of 1000-seed weight on seed yield via stem number indicated a trend for seed size to increase as stem size increased. Seed size in herbage legumes mainly depends upon genetic factors. One of the most interesting relationships between seed weight and fertility (seed per pods) which had a negative significant correlation was recorded by Iannucci et al. (2002).
The indirect effects via plant height substantially increased the total correlations between the leaf lengths, raceme length components and seed yield (except for blooming periods). As the results of those conflict relationships, the significant negative correlation coefficients of plant height, leaf length, raceme length and number of pods raceme with seed yield resulted from negative indirect effect of blooming periods in spite of their strong negative direct effect on seed yield. On the other hand high indirect effect of plant height had positive significant effect on seed yield, leaf length, raceme length and negative highly significant on flowering date. Indirect effect of leaf length was highly significant negative correlation coefficient with blooming periods, conversely positive significant correlation between leaf length and seed per raceme. Negative indirect effect of raceme had highly positive correlation on seed yield and pods per raceme. Leaf length and seed per raceme effect measured positively. On the other hand blooming period effect was highly significant.
Consequently, the indirect effects via leaf length, blooming periods were major determinants of the main effects of all components for the seed yield of the alfalfa grown under field condition. Highly significant and positive correlation coefficients as well as high direct effects stem size, leaf length and blooming periods on seed yield indicated that those components in co-operation are simultaneously the most reliable components for selecting high-yielding alfalfa types. But other components having highly significant positive correlations with seed yield were initially not the most reliable components for selection.
CONCLUSIONS
The data obtained from this study could be useful for alfalfa breeders and seed producers concerned with increasing seed yield. The main positive traits determining the seed yield in alfalfa growing in a spaced plants are stem yield, stem size, pods per raceme and seed weight were all important in determining seed yield in alfalfa having a direct effect and leaf length, blooming periods and number of seeds per pods having indirect effect.