Heterosis, inbreeding depression and combining ability studies in garden pea (Pisum sativum L.)

Garden pea (Pisum sativum L; 2n=2x=14), a member of family Leguminosae, is one of the principal vegetable crops which is cultivated for its green pods in the temperate and sub-tropical areas of the world. It is the second most important food legume worldwide after Phaseolus vulgaris (Jaiswal et al.,2015). Blixt (1970) indicated Mediterranean region as its principal centre of genetic diversity while the Near East and Ethiopia as its secondary habitats.
 
The information regarding combining ability is very important for breeding program and helps in identifying the best lines and crosses, thereby allowing selection of the appropriate parents for hybridization. The success of any breeding programme depends upon the genetic variability present in the germplasm (Adunga and Labuschange 2003) which provides better chances of selecting desirable types (Vavilov 1951). Therefore, the genetic reconstruction of pea germplasm is first step to identify the potential genotypes for use in breeding programme.
 
The parents good in per se performance may not necessarily produce desirable progenies when used in hybridization programme (Allard 1960). For exploitation of heterosis, choice of parents is one of the critical and most challenging task for a breeder. The ability of the parents to combine well with each other depends upon the complex interactions among the genes which cannot be judged by mere yield performance of parents. Moreover, in a breeding program critical judgment about a particular cross likely to produce transgressive segregants in self - pollinating crops would mainly depend upon the precise estimates of various components contributing to heterosis namely additive, dominance, non-allelic interactions, linkage among the polygenes and gene dispersion in the parents  of a cross (Jinks 1983). Further, the knowledge about gene action helps in formulating an efficient breeding programme to improve yield potential through component traits. The estimation of additive and non-additive gene action through combining ability analysis could be useful in isolating pure lines among progenies of good hybrids. In early group only three varieties have been recommended, thus the farmers of the state have limited choice. Few early lines have been recovered in the previous study on RIL population of a cross, NDVP-250 x Palam Priya. A line × tester analysis was done to determine the genetic interactions in the expression of earliness and various pod characters including pod yield. Line × tester mating design was proposed by Kempthorne (1957), offer means of obtaining useful information on differential  parental combinations through an assessment of overall genetic architecture of the parental lines, in relation to the characters studied. Griffing (1956) described a method for the precise measurement of general and specific combining abilities, which is very useful technique in classifying parental lines in terms of their hybrid performance and as such aid in selecting parents which when crossed may give rise to more desirable segregants. Such information forms the backbone of any breeding programme and is of great significance to the breeders.

The experiment was conducted for three consecutive years viz., 2014-15, 2015-16 and 2016-17 at Vegetable Research Farm of the Department of Vegetable Science and Floriculture, CSKHPKV, Palampur (H.P). The experimental material consisted of 21 genotypes, including five parental lines (Line 12, Line 13, Line 14, Line 15 and Line 17), three testers (Matar Ageta {T1}, Palam Triloki {T2} and Arkel {T3}) and their 15 crosses. The parents along with the hybrids were evaluated along with one standard check; Palam Triloki in randomized block design with three replications in 2014-15 (Katoch et al., 2018) with row to row and plant to plant spacing of 45 cm ( in parents 30 cm) and 10 cm respectively. Each experimental plot was of 3m × 1.8m size accommodating 6 rows. The promising genotypes were selected and evaluated along with three checks; Matar Ageta, Palam Triloki and Arkel in randomized block design with three replications during rabi 2016-17. Ten plants per genotype per replication were randomly selected for recording the observations at appropriate stages of crop growth during both the seasons on characters viz., days to 50% flowering, days to first picking, pod length, pods per plant, seeds per pod, pod yield per plant, shelling percentage, primary branches per plant, plant height and total soluble solids. The data were analyzed for randomized block design given by Panse and Sukhatme (1984). The mean values of both the generation for each trait were subjected to statistical analysis using the model suggested by Kempthorne (1957) and Singh and Chaudhary (1977). The estimates of heterosis as given by Allard (1960) were calculated as the deviation of means from the better parent (BP) and standard checks (SC). Inbreeding depression estimates were computed by using method as suggested by Griffing (1956):
GCA and SCA effects were obtained from the two way table of female parents vs. male parents in which each figure was total over replication.

Analysis of variance for line × tester mating design has revealed significant differences among the parents for all the traits except for harvest duration in F₁ and for pods per plant, pod yield per plant and shelling percentage in F₂. This indicated that parents used in present study were genetically diverse. Similar results have been reported by Gupta and Singh (2004). The relative magnitude of additive and dominance variance components showed preponderance of non-additive genetic variance (s²sca) for all traits except days to first picking and harvest duration in F₁and for days to first picking in F₂. The SCA variances were higher than GCA variances except for days to first picking and harvest duration in F₁ and for pod yield per plant in F₂. GCA variance was higher in magnitude in some traits, indicating role of additive gene action in governing these traits. In F₁ Line 17 was best general combiner for harvest duration, pods per plant, number of primary branches, node at which first flower appear and pod yield per plant. For plant height, shelling percentage, number of seeds per pod and TSS, L12 was best general combiner. Line 14 was best general combiner for days to 50% flowering and days to first picking. Among testers, Palam Triloki was best general combiner for all the characters except for seeds per pod and total nodes per plant. Similar observations pertaining to general combining ability effects of parents have also been reported by Cehyan et al., (2008) for plant height, pods per plant, seeds per pod and pod yield. Whereas, in F₂ Line 17 was best general combiner for days to 50% flowering pod length, pods per plant and pod yield per plant. For seeds per pod and pod length, Line 12 was best general combiner. Line 13 was best general combiner for days to 50% flowering and TSS. For earliness, traits including days to 50% flowering, days to first picking and node at which first flower appears cross L5 × T2 showed highest SCA effects including parents with good × good GCA effects. For yield related traits like pod yield per plant, pods per plant and harvest duration crosses L5 × T1 and L5 × T2 showed high SCA effects involving good x good general combiners. Similar was the case with L1 × T2 for shelling percentage and pod length in F₁ generation, while in F₂ generation days to 50% flowering appears in crosses L17 × T1 and L13 × T2 showed highest SCA effects including parents with good x good GCA effects. For shelling percentage and primary branches per plant cross L13 × T3 showed high SCA effects involving good x good general combiners. Similar views have been expressed by Singh et al., (2010) and Gupta and Singh (2004).
 
In F₁ Cross L17 × T2 exhibited maximum significant and negative heterosis over better parent for characters viz., node at which first flower appears, days to 50% flowering and days to first picking with heterotic response of -26.09%, 5.62% , - 5.66 respectively and -5.61% and -5.66%  (Table 4) over standard check for days to 50% flowering and days to first picking respectively. Cross L17 × T1 was most promising for pod yield and related traits like number of pods per plant and harvest duration. For pod yield per plant, it exhibited highest heterosis of 177.61% over better parent and 145.37% (Table 5) over standard check, while a wide range of residual heterosis over better parent and standard checks was observed in F₂ generation for pod yield and related traits (Table 6). Significant values of inbreeding depression and inbreeding vigour were found for almost all the traits in different crosses (Table 7). Cross L13 × T2 exhibited maximum significant and negative residual heterosis over better parent (22.59%) and SC1 (17.39%) for days to 50% flowering and for days to first picking (9.50%). also exhibited maximum significant and negative residual heterosis (12.83%) over better parent for days to first picking. Cross L17 ´ T2 was most promising for pod yield per plant and related traits like pods per plant and primary branches per plant (Table 3). In F₁, it was observed that for days to first picking, mean value observed for cross line 17 × Palam Triloki was 116.67, which was a week earlier than parents (Table 1) involved (mean value of 123.67 was exhibited by line 17 and Palam Triloki). Mean value of cross  combination  line 17 × Palam Triloki for days to 50% flowering was 84 as compared to 89 (mean value exhibited by both line 17 and Palam Triloki), which clearly indicates earliness of the cross (Table 2). For pod yield, best cross was line 17 × Arkel, with mean value of 30.87 for number of pods per plant, while it was just 14.40 for line 17 and 12.67 for Arkel. Based on such observations, heterosis of 208.57% was observed for this trait.  Mean pod yield per plant as exhibited by cross line 17 × Arkel was 140.23 g, while it was only 69.23 g for line 17 and 60.60 g for Arkel, while in F₂ for days to first picking, mean value observed for cross L13 × T2 was 108 days, which was as like that of parent Palam Triloki (106.33) and was ten days earlier than Line 13 (119.33). Mean value of cross combination L13 × T2 for days to 50% flowering was 69.67 days as compared to 90 and 69.33.
 

 

 

 

 

 

 

It was concluded that for pod yield, best cross was L17 × T2, with mean value of 26.72 for number of pods per plant, while it was just 16.52 for line 17 and 15.22 for Palam Triloki in F₂. Based on such observations, residual heterosis of 61.69% of was observed for this trait. Mean pod yield per plant as exhibited by cross L17 × T2 was 134.61 g, while it was only 83.54 g for line 17 and 75.25 g for Palam Triloki. Whereas, in F₁ line 17 × Arkel, with mean value of 30.87 for number of pods per plant, while it was just 14.40 for line 17 and 12.67 for Arkel. Based on such observations, heterosis of 208.57% was observed for this trait. Thus, crosses L17 × Arkel and L17 × T2 were superior in pod yield as compared to parents involved both in F₁ and F₂ respectively. Overall, it was observed that per cent contribution of lines was higher than the corresponding testers and their interaction for some yield related traits like for pod length, number of pods per plant, pod yield per plant, primary branches and total soluble solids, whereas contribution of testers was higher in some traits like for days to 50% flowering, days to first picking, number of seeds per pod, shelling percentage and plant height. Therefore, it can be concluded that lines played significant role in the expression of yield, related characters, whereas testers played important role in the expression of earliness related traits in different cross combinations.