Heterosis expression, interrelationship, direct and indirect effects of component characters on yield in intervarietal crosses of eggplant

Heterosis has been amply exploited in developing hybrids in brinjal. In India, only 17.8% area of brinjal cultivation is under hybrid seed due to lack of appropriate hybrids for specific area and purpose. In the present investigation, 14 parents were selected on the basis of divergence and mated in line (L) × tester (T) design hybrids and parents were raised to measure heterosis for different yield attributes. 40 hybrids resulting from a L × T mating design comprised of 10 lines (female) and 4 testers (males) were studied to determine the magnitude of heterosis and genotypic correlation between yield and yield characters. Expression of superiority over the commercial check occurred in 7 crosses, which ranged from -56.36 (L 3 × T 4 ) to 34.07% (L 7 × T 2 ). The hybrid (L 7 × T 2 ) had good heterosis values for growth and yield and is recommended as the most promising combination for developing high yielding hybrid eggplant varieties. Most crosses involving T 2 as tester parent had significant positive heterosis over the mid-parent and standard variety. There were strong correlations between fruit yield and numbers of branches per plant, average fruit weight and number of fruit per plant. Path analysis indicated that plant height, number of branches per plant, fruit length, fruit pedicel length, number of fruit per plant, average fruit weight and little leaf incidence had direct and positive effects on yield per plant, but negative and direct effects occurred for days to first flowering, fruit circumference, calyx length, shoot borer infestation, fruit borer infestation, ascorbic acid content and total phenol content. Simple selection would be effective for improvement of fruit yield.


INTRODUCTION
Brinjal (Solanum melongena L.), also known as eggplant and an important vegetable crop of India.It is widely cultivated in both temperate and tropical regions of the globe mainly for its immature fruits.In Tamilnadu (India), the productivity of brinjal is less as compared to national average, owing to use of low yielding cultivars grown for local preferences and their susceptibility to pests and diseases (Nalini, 2007).The present production is not proportionate with the demand.So, brinjal deserves a deep deliberation for improvement.It should be highly pragmatic by the fact that, India being the centre of origin and diversity of brinjal, it should pave the way for bringing about a kind of plant type, which could enhance its quality and productivity without sacrificing the consumer needs (Shafeeq et al., 2007).Quality and productivity of eggplant can be improved through heterosis breeding (Kakikazi, 1931).The estimation of heterosis for yield and its component characters would be useful to judge the best hybrid combination for exploitation of superior hybrids.Genotypic correlations exist between yield and yield attributing characters are important in breeding.Yield is the end product of many correlated characters.Selection for yield would be more effective when it is based on component characters which are positively correlated.When more number of variables is considered in correlation, the association becomes more complex.Use of path coefficient analysis makes clear direct and indirect associations and identifies the most reliable yield contributing characters (Daliya and Wilson, 2002).The study was under taken to study the extent of heterosis in different crosses and their utilization in future crop improvement.

MATERIALS AND METHODS
The experiment was conducted during June to October (kharif) 2010 to 2011 at College Orchard, Agricultural College and Research Institute, Madurai, India, situated at 9°5 N latitude and 78°5 E longitudes at 147 m above MSL.Ten lines were crossed with 4 testers through Line × Tester mating design to derive 40 F 1 hybrids (Table 1).The hybrids and 14 parents were established in a sandy loam soil and arranged in a Randomized Complete Block Design with three replications.Three ploughings were done with cultivator to make the soil fine tilth.Thirty (30)-day-old seedlings raised in the nursery beds were transplanted on ridges with a spacing of 60 × 60 cm.30 plants were maintained for each hybrid and parent in each replication.Cultural practices were followed uniformly for all hybrids and (TNAU Crop Production Guide, 2005).Observations were recorded from 5 randomly selected plants in each replication.The data recorded for the traits plant height, days to first flowering, number of branches per plant, fruit length, fruit circumference, number of fruit per plant, average fruit weight and fruit yield per plant for estimating heterosis.Selections were made in F 1 hybrids based on fruit shape, color, size and fruit yield per plant.Superior hybrids were selected and selfed.Seed were collected from selfed fruit and stored for further breeding.
The magnitude of heterosis in hybrids was expressed as percent of increase or decrease of a character over mid-parent (d i ), better parent (d ii ) and standard hybrid (d iii ) and estimated following the formula of Fonseca and Patterson (1968).The significance of magnitude of the relative heterosis, heterobeltiosis and standard heterosis was tested with the formula suggested by Turner (1953).Correlation coefficients were computed using formulae of Johnson et al. (1955).Path coefficients were obtained following the method of Dewey and Lu (1959).

RESULTS AND DISCUSSION
The analysis of variance revealed that parents showed significant differences for all characters (Table 2).Variance due to lines was significant for all traits indicating existence of genetic variability for growth and yield attributes among lines (females) and testers (males).The interaction between lines × testers was also significant for all the growth and yield traits.
Average performance and magnitude of heterosis for different quantitative characters varied (Table 3).In parents, T 2 (66.90 cm) had the shortest plants and L 2 (98.83 cm) was tallest.The hybrid L 1 × T 4 recorded tallest plants height, followed by L 2 × T 1 with positive heterosis over mid, better and standard parents indicating overdominance.Among the 40 hybrids, 15 had significant positive over the mid-parent for plant height.Three hybrids had significant positive heterosis and one hybrid had significant negative heterosis (L 6 × T 3 ) over the better parent.Twenty-seven (27) of 40 hybrids had significant positive heterosis over standard variety and the extent of heterosis over the standard variety.Similar findings were reported by Preneetha (2002) and Thangamani (2003).
Earliness is considered an important character in any crop improvement programme, which is manifested in F 1 hybrids and preferred for commercial cultivation when high yield is coupled with earliness.The hybrid L 8 × T 1 had the shortest period of days to first flower and L 8 ×T 4 had the longest period.For days to first flowering negative heterosis is desirable.Of 40 hybrids, 8 (L 1 × T 1, L 4 × T 1, L 5 × T 1, L 6 × T 1, L 6 × T 3, L 7 × T 3, L 8 × T 1 and L 10 × T 1 ) had significant heterosis in the desirable direction over the mid-parent.Seventeen (17) hybrids had significant heterobeltiosis in the negative direction (desirable).As many as 29 hybrids had heterosis in the desirable direction over the standard variety; only 2 (L 3 × T 2 and L 8 × T 4 ) had significant positive heterosis over the standard variety.
Number of branches per plant influences yield.Of 40 crosses, the highest number of branches per plant was for L 6 × T 2 .Of 40 hybrids 21 had significant positive heterosis while three hybrids (L 2 x T 1 and L 8 × T 4 ) exhibited significant negative heterosis.Seventeen hybrids had significant heterosis over the better parent, of which 10 had heterosis in the positive direction and 7 (L 2 x T 1, L 2 x T 3, L 4 x T 3, L 8 × T 4, L 10 × T 1, L 10 x T 3 and L 10 x T 4 ) had heterosis in the negative direction.In 18 hybrids, there was significant heterosis in the negative direction, where-as three hybrids (L 5 x T 2, L 6 x T 2 and L 7 x T 2 ) had significant positive heterosis over standard variety.These results agree with findings of Preneetha (2002).
Fruit length is an important character to be considered while selecting eggplant for high yield.The longest fruit was in T 1 followed by L 8 and shortest fruit were in T 2 .The hybrid L 6 x T 1 exhibited good performance and heterobeltiosis for fruit length.21 hybrids had significant negative heterosis over the mid-parent.A total of 27 hybrids had significant heterosis over the better parent in desirable direction (negative).Useful heterosis was exhibited by all 40 hybrids over the standard variety.The cross combination exceed-ing the superior parent is a valuable character in heterosis breeding (Patil et al., 2001).Similar findings were reported by Kaur et al. (2001).
Fruit circumference contributes to improved yield and parents L 1 , L 5 , T 3 , T 4 , and L 9 had the largest fruit.The highest fruit circumference was recorded in L 6 x T 4 followed by L 1 x T 2 and the lowest was in L 3 x T 3 .The observations were positive and significant for 14 hybrids.Positive and negative non-significant heterosis were observed in seven and four hybrids, respectively.Seven hybrids had significant and heterobeltiosis,     Nalini (2007).
The number of fruit per plant influences yield.The lowest number of fruit per plant was in L 5 and the highest in T 1 followed by L 9 among parents.Nineteen crosses, L 1 x T 1, L 2 x T 1, L 2 x T 3, L 4 x T 1, L 4 x T 2, L 4 x T 4, L 5 x T 2, L 5 x T 3, L 5 x T 4, L 6 x T 2, L 6 x T 4, L 7 x T 2, L 7 x T 4, L 8 x T 1, L 8 x T 2, L 8 x T 4, L 9 x T 1, L 9 x T 2 and L 10 x T 2 exhibited positive heterosis in the desirable direction over the mid-parent.Expression of heterosis over the better parent in the positive direction occurred in 14 crosses.Significant heterosis over comercial check was recorded in 33 hybrids with 9 had positive and 23 had negative heterosis.Considerable positive heterosis in different cross combinations were also reported by Chowdhury et al. (2010).
The highest single fruit weight was for L 7 x T 2 followed by L 1 x T 1 and the lowest was in L 8 x T 3 .Of 40 hybrids, 17 had significant desirable heterosis over the mid-parent in the positive direction.Only 7 hybrids had significant heterobeltiosis.In 8 hybrids, heterosis was significant and in the positive direction over the standard variety.The heterosis over the better and standard parents was negative for fruit weight.Similar findings reported by Nalinidharwad et al. (2011) andPatil et al. (2001) lend support to the present results.This trend is not a constraint, because smaller sized eggplant fruit are preferred in South India and hybrids with smaller and more fruit could be selected.
The ultimate interest of the breeder is to get high yield.An appreciable amount of heterosis in F 1 s over the midparent value occurred for fruit yield per plant.The T 1 parent had the highest yield followed by L 5 and L 2 , while the lowest yield was for L 9 .Fruit yield per plant showed a wide range among hybrids.Among the 40 hybrids, 19 had significant positive heterosis over the mid-parent.The hybrid from the cross L 7 x T 2 had the highest magnitude of heterosis in the positive direction followed by L 4 x T 2 , L 8 x T 4 and L 4 x T 4 .Most crosses involving T 2 as tester parent had significant, positive, heterosis over the mid-parent.This agrees with Ramesh Kumar et al. (2013).

Association analysis
Genotypic correlation coefficients varied depending on character (Tables 4 to 8).Of 15 characters studied, only numbers of fruit per plant, number of branches per plant and average fruit weight had significant association with fruit yield at phenotypic and genotypic levels (Prabakaran, 2010).
Fruit yield had significant negative association with fruit borer infestation at both levels.This trait could be considered as an important criterion for selection for fruit yield.Total phenol content had considerable nonsignificant positive correlation with fruit yield per plant followed by fruit length, fruit circumference and shoot borer infestation.During selection these traits can be considered to improve fruit yield.There were significant, and positive, correlations for plant height with days to first flowering, while days to first flowering significantly, and negatively, correlated with numbers of branches per plant.As days to first flowering decreases, there will be an increase in number of branches per plant.Fruit pedicel length had a significant, and positive, correlation with clayx length.Fruit circumference was significantly, and positively, associated with average fruit and average fruit weight may result in simultaneous improvement of fruit yield per plant (Thangamani and Jansirani, 2012).These cha-racters are highly reliable components of fruit yield and could be utilized as yield indicators during selection.Days to first flowering, fruit pedicel length, calyx length, fruit borer infestation and ascorbic acid, an important quality trait, were negatively associated with fruit yield per plant.Selection for this trait will result in reduction weight.Similarly, the association between shoot borer infestation and little leaf incidence was positive and significant.Selection for this trait will result in reduction of fruit yield, number of branches per plant, number of fruit per plant of fruit yield.

Path coefficient analysis
The path coefficient analysis permits the separation of direct effect from indirect effects through other related traits by partitioning genotypic correlation coefficients.Plant height, number of branches per plant, fruit length, fruit pedicel length, number of fruit per plant, average fruit weight and little leaf incidence had positive direct effect on yield (Table 9).
Direct selection for these characters are likely will bring about an overall improvement in fruit yield per plant.The residual effect determines how causal factors account

Table 1 .
Features of parents used in study.

Table 2 .
Analysis of variance for parents and hybrids for vegetative and reproductive characters.

Table 3 .
Average performance and magnitude of heterosis for plant height and days to first flowering in eggplant.

Table 4 .
Genotypic correlation between fruit yield and growth characters in eggplant.