Combining ability and heterosis for agronomic traits in chili pepper

The Capsicum breeding has been developed with emphasis in bell pepper (Capsicum annuum) and few studies are available in other species, especially C. baccatum, which has potential use not only as disease resistance source but also in obtaining new genotypes suitable for farmers’ production. In the present work, the combining ability of ten C. baccatum hybrids, along with their five parentals, were tested considering 12 agronomic traits. The hybrids were produced from a complete diallel without reciprocals and assessed in greenhouse conditions, in Campos dos Goytacazes, Rio de Janeiro state, Brazil, during the period July to December 2009. The experimental design was a randomized block with three replications and the following agronomic traits evaluated: canopy diameter (CD), plant height (PH), days to fructification (DF), number of fruits per plant (NFP), mean fruit weight (FW), dry fruit matter weight (FDM), dry matter content (DM), fruit length (FL), fruit diameter (FD), pulp thickness (PT), total soluble solids (TSS) and yield per plant (PP). Significant differences were observed only for general combining ability (GCA) in regard to PH, FW, FDM, DM, PT and TSS, indicating that additive effects were involved on the control of these characters. For CD, DF, NFP, FL, FD and PP, there was significance not only for GCA but also for specific combining ability (SCA) indicating that non-additive and additive effects were important in genetic control of these traits. The hybrids UENF 1629 X UENF 1732, UENF 1616 X UENF 1732 and UENF 1624 X UENF 1639 were considered superior because they have favorable agronomic traits.

T he market for hot peppers of the Capsicum genus in Brazil has become very important in the vegetable economic scenario mainly because of the great variety of products and by-products, uses and forms of consumption (Moreira et al., 2006;Bento et al., 2007;Henz & Ribeiro, 2008;Sudré et al., 2010). Although it is difficult to measure the real size and importance of this market since no reliable statistics and systematized information are available, obtaining improved cultivars in the various species that comprise this genus is a potential area to amplify and sustain the pepper agribusiness (Moura et al., 2010).
Of the 31 species of Capsicum found and described to date, only five are cropped commercially: C. annuum, C. baccatum, C. chinense, C. frutescens and C. pubescens (Moscone et al., 2007;Moura et al., 2010). The C. baccatum species has high genetic variability, shown principally in the fruit that can have different shapes, coloring, sizes and pungency levels (Lannes et al., 2007;Rêgo et al., 2009). The most cultivated types of this species in Brazilian conditions include 'Dedo de moça ' and 'Cambuci' (Carvalho et al., 2003).
The few studies carried out with diallel in Capsicum in Brazil have concentrated almost exclusively on crosses among sweet pepper cultivars (Souza & Maluf, 2003). Most studies on the C. baccatum species have researched pre-breeding conditions. However, this situation tends to change, especially because of farmer's interest in new genotypes that meet the requirements of greater yield, fruit uniformity and quality. An example is the study published by Rêgo et al. (2009), who estimated the general combining ability (GCA) and the specific combining ability (SCA) of eight C. baccatum accessions for the characteristics fruit quality and yield. The authors observed that both additive and nonadditive effects influenced the hybrid performance for all the characteristics tested. Line selection from segregant generations and obtaining hybrids to exploit heterosis or heterobeltiosis can be used appropriately in breeding these genotypes.
The objective of the present research were to study the combining ability of C. baccatum var. pendulum accession lines regarding their agronomic characteristics, to identify superior hybrid combinations and to determine the genetic action involved in the expression of the studied traits.

MATERIAL AND METHODS
Five Capsicum baccatum var. pendulum lines from the germplasm collection from Universidade Estadual do Norte Fluminense Darcy Ribeiro (UENF) identified as UENF 1616, UENF 1624, UENF 1629, UENF 1639 and UENF 1732, were studied. These accessions were previously characterized by Bento et al. (2007Bento et al. ( , 2009 and Moura et al. (2010). The crosses involving the five parents were carried out in a complete diallel scheme without reciprocals, obtaining a total of ten hybrids. The hybridizations were carried out in a greenhouse using eight pots per parent, each pot containing one plant, and the crosses were made between April and June 2008.
The pollen was collected from each male parent as soon as the flower buds were opened. Recently opened flower buds from each one of the male parents were collected to remove the pollen. The pollen obtained from each parent was stored in a refrigerator in labeled recipients containing silica gel. The flower buds from plants of the five female parents were emasculated in the morning before anthesis, using pincers. In the same period, to pollinate, pollen grains from each one of the male parents were placed on the stigma of each emasculated flower. Labels were used to identify the fruits resulting from each different type of cross.
To a s s e s s t h e a g r o n o m i c characteristics, the hybrids and parents were cultivated n greenhouse in Campos dos Goytacazes. A randomized block design was used with three replications and 12 plants per plot, considering all the plants as useful area of the experiment. The IAC Ubatuba cultivar was used as a border. The spacing adopted was 1.1 x 0.9 m between rows and plants.
During the experiment, pepper plants were grown according to the regular recommendations for the pepper crop made by Filgueira (2005) such as weeding, tutoring, fertilization and drip irrigation. Five harvests were made and the following agronomic characteristics were assessed: canopy diameter (CD) (measured in centimeters, when 50% of the plants in the plot produced ripe fruits); plant height (PH) (measured in centimeters when 50% of the plants in the plot produced ripe fruits); days to fructification (DF) (number of days from transplant until 50% of the plants in the plot produced ripe fruits in the first and/ or second bifurcations); number of fruits per plant (NFP) (sum of the number of fruits obtained in the five harvests); mean fruit weight (FW) (mean weight in grams using the ratio between total weight of fruit per plant and number of fruits per plant); dry fruit matter weight (FDM) (mean weight in grams of five dried fruits per plant, using a forced air circulation chamber at 65°C during 72 hours); dry matter contents (DM) (ratio between FDM and FW, multiplied by 100); fruit length (FL) (measured in centimeters considering five fruits per plant); fruit diameter (FD) (measured in millimeters considering five fruits per plant); pulp thickness (PT) (measured in millimeters considering five fruits per plant); total soluble solids content (TSS) (measured in five fruits per plant, using a digital refractometer); and yield per plant (PP) (multiplication of NFP and FW (kg/plant)).
Analysis of variance was performed for each characteristic and the mean of the 15 treatments was compared by the Scott- Knott cluster test (1974). Diallel analysis was performed according to Griffing's method (1956) considering method 2 and fixed model to estimate the general and specific combining abilities of the parents and hybrids. The analysis were carried out with the Genes software (Cruz, 2006).

RESULTS AND DISCUSSION
The genotypes effect for all the traits was significant by the F test (data not shown). The coefficient of experimental variation values were lower than 20% for most of the characteristics assessed except for NFP and PP, that had values of 24.17 and 23.90%, respectively (Table  1). These results reflected the good experimental accuracy and guaranteed Table 1. Averages of 12 agronomic traits evaluated in five genitors and 10 hybrids of Capsicum baccatum var. pendulum, followed by the group averages between Scott-Knott (médias de 12 características agronômicas avaliadas em cinco genitores e 10 híbridos de Capsicum baccatum var. pendulum, seguidas pelo agrupamento entre médias de Scott-Knott). Campos dos Goytacazes, UENF, 2010.

Genotypes
Agronomic characteristics  The partitioning of the sum of the squares of the genotypes, in the sum Table 2. Estimates of the effects of general combining ability (ĝ i ) for 12 agronomic traits in five parents of Capsicum baccatum var. pendulum evaluated in complete diallel without reciprocal (estimativas dos efeitos da capacidade geral de combinação (ĝ i ) para 12 características agronômicas avaliadas em cinco genitores de Capsicum baccatum var. pendulum em esquema de dialelo completo, sem os recíprocos). Campos dos Goytacazes, UENF, 2010.  Table 3. Estimates of the effects of specific combining ability (ŝ ii and ŝ ij ) for 12 agronomic traits in five parents of Capsicum baccatum var. pendulum evaluated in complete diallel without reciprocal (estimativas dos efeitos da capacidade específica de combinação (ŝ ii e ŝ ij ) para 12 características avaliadas entre cinco genitores de Capsicum baccatum var. pendulum em esquema de dialelo completo, sem os recíprocos). Campos dos Goytacazes, UENF, 2010.

Effects (ŝ ii and ŝ ij )
Agronomic yield in C. baccatum and observed significance for GCA and SCA effects for most of the characteristics, except for height of the first bifurcation that was significant only for SCA. The estimates of the quadratic components indicated that the characteristics PH, FW, FDM, DM, FL, FD, PT and TSS expressed superiority of the additive genetic effects, compared to non-additive, indicating the possibility of satisfactory gains with selection for these characteristics in segregant generations. The opposite was observed for CD, FD, NFP, and PP, that is, non-additive effects predominated that could be exploited in the hybrids or alternatively could require the implementation of more complex breeding strategies, in the case of avancing segregant populations to obtain recombinant lines. Gonçalves et al. (2011) assessed the same set of hybrids and observed that effects of dominance were predominant for fruit production per plant, based on Hayman (1954) method for diallel analyses that corroborated the results according to Griffing (1956).
Although there are few studies published on C. baccatum genetics and breeding, there was agreement between the results obtained for a set of characteristics studied by Rêgo et al. (2009) and the results reported in the present study, except for TSS. Nonadditive effects for TSS were reported by Rêgo et al. (2009), while in the present study both additive and nonadditive effects were observed.
The choice of parents to form segregant populations is crucial for successful breeding programs and the combining ability, with the presence of complementary genes, is largely responsible for the success. According to Sprague & Tatum (1942) low GCA value indicates that the mean of the hybrids in which line i participates do not differ from the general mean of the diallel cross. On the other hand, high positive or negative values, show that line i is much better or worse than the other lines included in the diallel cross, in relationship to the mean of its hybrids.
The UENF 1624 parent can be recommended in crosses for combinations with smaller canopy diameter and a greater number of fruits per plant, because this parent had negative GCA values and high values for CD and positive values for NFP. However, this parent is unsuitable  (Table 4).
The best combinations for FD resulted from the UENF 1732 and UENF 1639 parents because they expressed negative values for the s ij estimate for all the combinations in which they took part. The UENF 1616 x UENF 1632 hybrid registered the lowest s ij and heterosis values (Table 3 and 4).
The best combinations for DM (%) were UENF 1616 x UENF 1629, UENF 1624 x UENF 1629 and UENF 1624 x UENF 1732 because they expressed high positive values for the s ij estimate (Table 3). However, only the combinations UENF 1616 x UENF 1629 and UENF 1624 x UENF 1629 had at least one other parent with desirable values to estimate the mean GCA effect. These hybrids obtained heterosis values of 0.23%, 2.19% 0.11%, respectively (Table 4).
It was endeavored to find for the TSS characteristic combinations with higher total soluble solid content. Thus the most promising combinations were those that expressed positive value for the characteristic, which were UENF 1616 x UENF 1732, UENF 1624 x UENF 1639 and UENF 1629 x UENF 1732. However, positive heterosis values, 5.11% and 12.93%, respectively, were observed only for the combinations UENF 1616 x UENF 1732 and UENF 1629 x UENF 1732 (Table 4).
F o r t h e P P c h a r a c t e r i s t i c , combinations were sought with high, positive s ij values. It was observed that the parents UENF 1616 x UENF 1624, UENF 1616 x UENF 1732, UENF 1624 x UENF 1732, UENF 1624 x UENF 1639 and UENF 1629 x UENF 1732 had high, positive values for s ij and heterosis, especially the UENF 1624 x UENF 1639 and UENF 1629 x UENF 1732, with s ij values of 0.34 0.49 and heterosis of 60.87% and 74.10%, respectively (Tables 3 and 4).
The results showed that the UENF 1629 x UENF 1732, UENF 1616 x UENF 1732 and UENF 1624 x UENF 1639 hybrid combinations were promising for testing in experiments with replications for possible commercial indication and also to derive segregant populations and start a C. baccatum var. pendulum breeding program with the objective of obtaining recombined lines, because they combine favorable agronomic characteristics, such as yield per plant and soluble solid contents.