Evaluation of seventy six sugarcane families at early selection stages

An experiment was conducted on sugarcane families belonging to the early selection stage of the Sugarcane Research Institute, Kaiyuan, Yunnan Academy, China, to evaluate them for the selection of superior families for the later stages of the breeding program. Seventy-six full-sib families and check variety were evaluated. The experiment was conducted during two growing seasons, corresponding to the plant cane and first ratoon, 2015/2016 and 2016/2017, respectively. The traits of cane yield (TCH) and juice quality were measured. Results indicate the use of the traits with high heritability as selection criteria together with sugar yield (TSH) could lead to genetic improvement in TSH. The study indicated that high estimates of genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) were recorded for TCH (22.85 and 27.74) and TSH (22.66 and 28.08). The results also revealed that seven families viz., LC03-1137×YZ89-7, LC03-1137×PS45, YT93-159×YR11-95, YR05-346×YZ05-51, FN38×GT96-211, YR10-509×FR96-405 and YT03-373×YR08-1276 were the best families across all other bi-parental families for most studied traits at early selection stages suggesting the possibility of evaluation of a large number of clones of these families, followed by selection of superior clones within these families during the next selection stages.


INTRODUCTION
Sugarcane (Saccharum spp., L) is one of the major cash and industrial crops in the world.It is a source of raw material to sugar industry and various agro-based industries.New sugarcane varieties are developed through the selection of vegetatively propagated clones obtained from true seeds that are derived from the hybridization of superior parents.Individual selection during the initial stage is of low efficiency given the low broad sense heritability for the majority of traits (Skinner, 1982).Several research projects demonstrated that family selection, when followed by individual clone selection, was superior in terms of genetic gain and more cost effective than either family or individual clone selection alone (Kimbeng and Cox, 2003).Evaluation of a large number of clones for families, then the possibility of selection of superior clones within these families during *Corresponding author.E-mail: gksky_wcw@163.com.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License the next selection stages have also been reported (Skinner et al., 1987;Cox and Hogarth, 1993;Shanthi et al., 2008;Stringer et al., 2010;Mahmoud et al., 2012).
Heritability values are categorized as low (0 to 30%), moderate (30 to 60%) and high (60% and above) as stated by Robinson et al. (1949).Also PCV and GCV values are ranked as low, medium and high with 0 to 10, 10 to 20 and >20%, respectively (Shivasubramanian and Menon, 1973).Silva et al. (2002) estimated the variability among 18 families of sugarcane for stalk height, stalk diameter, average brix, stalk number and stalk weight.Their results showed highly significant differences among families for most traits.Oliveira et al. (2009) evaluated 80 full-sib sugarcane families for yield of canes per hectare.They observed significant differences among evaluated crosses.Heritability of traits on individual basis was of medium magnitude (0.22), while heritability on family mean basis was 0.73, indicating the effectiveness of family selection at early stages.Mehareb and Abazied (2017) estimated genetic variability, and found high genetic variance (σ 2 g) relative to environmental variance for all traits under study across seasons.The highest values of PCV and GCV were observed for reducing sugar (54.31 and 47.22%, respectively) followed by TSH (19.85 and 19.24%,respectively), respectively.Heritability estimates exceeded 80% for all studied traits, except for purity and reducing sugar.The objective of this study was to evaluate 76 sugarcane families of the Sugarcane Research Institute, Kaiyuan, Yunnan Academy, China and make selection intensity under 10% based on TSH at early selection stages.

Plant material and experimental conditions
The study was carried out at breeding nursery of Sugarcane Research Institute, Yunnan Academy, China.Materials of sugarcane (Saccharum spp.) consisted of seventy six bi-parental crosses (families) that could be considered representative of the sort of breeding materials processed in the sugarcane breeding program in China.The seventy six bi-parental (7752 seedlings) hybrid combinations which correspond to the 2014 series were crossed in three crossing locations; Hainan, Kaiyuan and Ruili (Table 1).
Seeds (fuzz) were germinated in the greenhouse in March 2015.A total of 7752 and discussions seedlings from 76 families were transplanted in field in June, 2015.Check variety (ROC22) buds were planted in the greenhouse in May, 2015.After buds germination, they were transplanted with families' seedling in field during the first week June, 2015.After harvested plant cane in January 2016, the first ratoon was harvested in January, 2017.The field was irrigated right after planting and all other agronomic practices were carried out as recommended as at when due.

Experimental design and data collection
The experiment was arranged in a randomized complete block design.Each family had three replicates and 34 seedlings per replicate, each replicate had two rows per plot (17 seedlings per row), length 6 m and space 1 m per row with three replications.Each family was represented in the experiment by a random selection of 34 seedlings.The following traits were measured for each family: 1. Traits of cane yield and its contributing traits: a. Stalks number per stool.b.Stalk length (cm) was measured from soil surface to the visible dewlap.c.Stalk diameter (cm) was measured at mid stalk with no reference to the bud groove.d.Stalk weight (kg) was calculated by dividing cane yield per stool by number of stalks per stool.e. Ton cane per hectare (TCH) was calculated on a plot basis.f.Stool weight (kg) was calculated by dividing cane yield per stools by number of stools per plot.
2. Juice quality traits and TSH: a. Brix (percent soluble solids) was determined with a hydrometer.b.Ton sugar hectare (TSH) was calculated according to the formula (Wu et al., 2009): 3. Disease and other characters: a.

Statistical analysis
Analyses of variance were performed for the collected data according to Gomez and Gomez (1984) using MSTAT-C computer package by Freed et al. (1989).The comparison among means was done using the least significant difference (LSD) test at 5% level of probability.Variance components were calculated by equating appropriate mean squares for the differences among genotypes to their expectations and solving for the components.Broad-sense heritability (H%) was estimated using variance components following the formula (Allard, 1960): Where, σ 2 g and σ 2 ph are genotypic and phenotypic variances respectively.

Estimation of variance components
Genotypic and phenotypic components of variance were estimated according the following formulae: Genotypic variance (σ 2 g) = gMS -EMS Phenotypic variance (σ 2 p) = σ 2 g + EMS Where, gMS refers to genotypic mean squares and EMS refers to error mean squares.

RESULTS AND DISCUSSION
Data presented in Table 2 show that the analysis of variance for all the studied characters revealed highly significant differences (p < 0.01) among all evaluated treatments (families and check).This indicates the existence of sizable variability and that considerable improvement can be achieved in these characters by selection; however, there was no significance for blocks.The mean squares due to families differed significantly for all studied characters, indicating sufficient genetic variation in genotypes for all the studied characters.Similar results were outlined by Silva et al. (2002) and Oliveira et al. (2009).Highly significant (p < 0.01) variance estimates of families versus checks were found for stalk height and stalk weight, while significant differences (p < 0.05) were found for stalk diameter, TCH, brix, TSH and mosaic, whereas none significant differences were found for Stalk number/stool and weight/ stool.
For pithiness character, there was no significant between families and check variety.This may indicate that the evaluated new families do not have pith value as check variety.Similar results were reported by Tahir et al. (2014) who observed significant differences for the contrast of the checks versus new genotypes for the parameters.
Data in Table 3 revealed that brix percentage, TCH and TSH varied significantly among evaluated crosses.Majority (about 36) of the families had higher brix percentage values than the check variety ROC22 (22.74%).

Genetic components
Genetic variance is important as it describes the amount of genetic variation present for the trait.Table 4 shows high genetic variance (σ 2 g) relative to environmental variance for all traits under study expect for Stalk number/stool.The results also indicate that high estimates of genotypic and phenotypic coefficients of variation GCV and PCV were recorded for TCH (22.85 and 27.74),TSH (22.66 and 28.08) and Pithiness (23.27 and 27.88).High genotypic and phenotypic coefficients of variation for TCH were reported earlier by Singh and Sangwan (1980).Traits exhibiting relatively high GCV estimates may respond favorably to selection.However medium estimates of GCV and PCV were recorded for stool/row (10.67 and 13.93), stalk number/stool (12.86 and 18.79), weight/ stalk (16.08 and 19.21) and Mosaic (14.71 and 19.28) for as much medium estimates of GCV and high estimates of PCV were recorded for weight/stool (18.27 and 24.36), while, height (7.27 and 8.33), diameter (6.49 and 8.31) and brix (4.12 and 5.48) resulted in low estimates of GCV and PCV.
In present study, high heritability (broad sense) estimates were recorded for height (75.90%), stalk weight (75.04%), pithiness (69.18%),TCH (67.26%), stalk diameter (66.19%) and TSH (64.57%).This suggests that a large proportion of the total variance is heritable and selection of these traits would be effective.High values of GCV and PCV were coupled with high heritability and high genetic advance for TCH, TSH and pithiness.Knowledge of variability and heritability of characters is essential for identifying those amenable to genetic improvement through selection (Vidya et al., 2002).
Results of the current study indicate that use of the traits with high heritability as selection criteria together with TSH could lead to genetic improvement in TSH.The effectiveness of selection depends not only on heritability but also on genetic advance (Butterfield and Nuss, 2002).In this respect, Gupta and Chatterjee (2002), Agrawal (2003), Delvadia and Patel (2006) and Patel et al. (2006) reported that high heritability with high genetic advance

Table 1 .
Bi-parental hybrid families of experimental sugarcane pedigree and origin.

Table 2 .
Mean squares of studied traits.

Table 3 .
Means value of some traits of the studied sugarcane families.
Figure 1.10% selection intensity based on TSH.