Journal of
Plant Breeding and Crop Science

  • Abbreviation: J. Plant Breed. Crop Sci.
  • Language: English
  • ISSN: 2006-9758
  • DOI: 10.5897/JPBCS
  • Start Year: 2009
  • Published Articles: 400

Full Length Research Paper

Estimates of genetic parameters and genotype by environment interactions for sugar yield and its components in sugarcane genotypes in Western Kenya

Japheth Embeywa Jamoza
  • Japheth Embeywa Jamoza
  • KALRO - Sugar Research Institute, Kenya.
  • Google Scholar
James Owuoche
  • James Owuoche
  • Department of Crops, Horticultute and Soils, Egerton University, Kenya.
  • Google Scholar
Oliver Kiplagat
  • Oliver Kiplagat
  • School of Agriculture and Biotechnology, University of Eldoret, Kenya.
  • Google Scholar

  •  Received: 14 April 2019
  •  Accepted: 02 September 2019
  •  Published: 30 November 2019


This study was conducted to estimate broad sense heritability, genetic advance, GE interactions and correlations among quality traits in sugarcane clones in western Kenya. Thirteen sugarcane promising clones and one check cultivar were evaluated plant and ratoon crops in three locations under rain fed conditions using the randomised complete block design with three replications.  Analysis of variance showed significant differences in hand refractometer brix, sucrose content (Pol% cane), juice purity, fire content, sugar yield and brix yield. Sucrose content, fibre content, sugar yield and brix yield exhibited significant genotype × location (GL) interactions. The genotype mean squares exceeded the GE interactions for all the quality traits suggesting that more emphasis should be placed on testing clones in many locations than on testing ratoon crops within locations. High genetic coefficient of variation (GCV) was detected for cane yield (8.12%), brix yield (6.39%), sugar yield (5.69%) and sucrose content (3.69%). Broad sense heritability was high for sucrose content (0.712) and moderate for cane yield (0.515), fibre content (0.474), juice purity (0.445) and refractometer brix (0.380). Cane yield (10.3%), brix yield (6.7%), sucrose content (5.5%) and sugar yield (5.4%) showed highest expected genetic advance.  The results indicated that these traits may respond positively to selection and present opportunities for improvement through breeding. High genetic correlation (rg=0.998) between refractometer brix and sucrose content suggest that selection for refractometer brix can be effective in identifying varieties with high sucrose content.

Key words: Saccharum spp. heritability, genetic advance, sucrose content, selection, sugar.



Genetic improvement of varieties plays a pivotal role in the development of sugar industries in almost all sugarcane growing countries. Improved cane yields, sucrose content and disease and pest resistance and maintaining acceptable fibre levels for  milling  are usually the main breeding objectives in most sugarcane breeding programmes (Jackson, 2005). Studies on exploitation of sugarcane as a sustainable energy source are on the increase (Corcodel and Roussel, 2010; Hoang et al., 2015;  Priya  et  al.,  2018).  An  improvement  in  sucrose content in sugarcane has high economic value as it increases sugar yield with very little increase in marginal costs through harvesting, cane transport or milling (Jackson, 2005). Thus gains in sucrose content are economically more beneficial than corresponding increases in cane yield thus increased sucrose content is a key objective of sugarcane breeding programmes.

Breaux (1984) reported that the record increase in sugar recovery in Louisiana was attributed to wide acceptance of high sucrose varieties that were introduced in 1973. The high sucrose content of the varieties was a result of continuous breeding and selection effort since the 1920s (Breaux, 1984; Legendre, 1992).  Significant increase in the sucrose content of both experimental clones and cultivars grown in a number of locations throughout the sugarcane-growing region of Louisiana since 1928 was reported by Irvinne and Richards (1983). The average sucrose % cane of experimental varieties rose from 5.54% in 1928 to 13.56% in 1978; while the average sucrose % cane for all adopted cultivars increased from 7.25% in 1928 to 12.7% in 1981. In addition the study showed a strong correlation between the sucrose content of experimental varieties and actual mill recovery (r = 0.79) and sucrose content of commercial varieties and mill recovery (r = 0.80) during the same period. However, further improvement in sucrose content through breeding may be difficult as the sucrose content of parent varieties reaches an apparent plateau (Legendre, 1992; Inman-Bamber, 2014).

The effectiveness of selection for sugar yield and its components depends largely on the genetic variability present in the breeding population and the heritability of the traits.  It is necessary to identify traits with high genetic variation. The easiest way to estimate variance components is to test a large number of genotypes for two or more years and at two or more locations (Mayo, 1980). Components of juice quality are largely determined by the genotype but can be significantly influenced by the environment (Tena et al., 2016; Singh et al., 2019). Sucrose content and purity are conventional indicators of maturity commonly used as selection criteria and they are widely investigated in sugarcane breeding programmes (Mariotti et al., 2001). The sugarcane industry in Kenya is largely dependent on a few varieties that have low sucrose content and sugar yield (Jamoza, 2011). In recent years the industry has emphasized development and adoption of high yielding sugar rich varieties. This study aimed to estimate (i) broad sense heritability (ii) potential genetic advance and (iii) correlations among juice quality traits in promising Kenyan sugarcane clones.




Experimental  sites and genotypes

The test genotypes, experimental sites and methodology applied to obtain estimates were as described by Jamoza et al. (2014). In brief,  13  clones  (KEN01-24.  KEN01-26,  KEN01-41,  KEN01-279, KEN01-345, KEN01-592, KEN01-819, KEN01-848, KEN01-1009, KEN01-1104, KEN01-1108, KEN01-1139 and KEN01-1294) representing eight crosses involving 15 parents and one commercial variety (N14) were chosen for study and evaluated in plant and first ratoon crops between August 2007 and June 2010. They were grown at three locations in western Kenya namely: Kibos (34° 48’E, 0° 04’N) 1,184 m above sea level on clay loam soil with long term mean annual rainfall of 1,490 mm. The temperatures range from 15.3 to 30°C; Mumias (34° 30’E, 0° 21’N) at 1,314 m above sea level, receives 2,194 mm annual rainfall with a temperature range of 16.4 to 30.9°C and has free draining loam soils; and Nzoia (34° 40’E, 0° 35’N) situated at 1,445 m above sea level, receives average annual rainfall of 1,650 mm with a temperature range of 13 to 32°C and has sandy clay loam soils. The randomised complete block design with three replications was used at each location. The following cane quality data were collected in both crops.

Field brix

Hand held refractometer (0-32o) was used to determine brix of 5 millable stalks taken randomly from each plot in the field at harvest.

Sugarcane analysis

At harvest millable stalks in each plot were cut at ground level, well topped and hand stripped to remove the trash and green leaves. Twelve millable stalks were randomly taken from each plot, bundled, tied, labelled and transported to the laboratory for juice and fibre analysis. Juice was extracted from six stalk samples using a simple three roller cane press (Milligan et al., 1990a).  The juice was filtered through Whatman filter paper No.1 and 100 ml portions of the filtrate used to determine  brix (percent soluble solids w/w) using a bench refractometer as described in the Laboratory Manual for South African Sugar Factories (Anon, 1985). For the determination of Pol % juice approximately 300 ml samples of the extracted juice were placed in a beaker and clarified using 3 g of sub lead acetate. The mixture was then filtered using Whatman filter paper No. 91. Polarimetric readings of the clarified juice were obtained using a digital automated sucromat while sucrose content (Pol% cane) was calculated from the values of Pol % juice and fibre content (BSES, 1970).

The other six stalks from the harvested sample for each cane variety were used to determine fibre content following the procedure described by Clayton (1971). Six pieces were cut from different (top, middle, and bottom) portions of the stalks in order to obtain a subsample equivalent to one whole stalk. The pieces were further cut into smaller pieces (approx. 3 cm) then shredded in a laboratory hammer mill (shredder). The shredded samples were well mixed and then 200 g subsamples were placed in pre weighed fibre bags and washed alternately in cold and hot water to remove all sugars (mainly sucrose, fructose and glucose). The samples thus processed were dried in an air oven at 105°C for 24 h to constant weight. The fibre content was calculated directly from the 200 g fresh weight and dry weight as:

Other derived quality characters were computed as follows:

Estimated sugar yield tha-1 = Sucrose content (Pol % cane) × cane yield (tha-1) × 100

Brix yield tha-1 = Brix % cane × cane yield (tha-1) × 100

Statistical data analysis

All the data were submitted to analysis of variance and covariance, estimation of genetic, genotype by environment interaction and error variance components, broad sense heritability, genetic advance and correlations as described by Jamoza et al. (2014). The genotypes were assumed to be fixed while genotype by environment interactions and environments were random (Chang, 1996;  Brown and Glaz, 2001).




Combined analysis of variance for quality components over three locations and two crops

Mean squares for the traits are shown in  Table  1.



Genotypes exhibited significant (p ≤ 0.01) differences for hand refractometer brix, sucrose content, juice purity, fire content, estimated sugar yield and brix yield. Locations played a significant (p ≤ 0.01) role in the phenotypic expression of all the quality traits. However, years or crops significantly (p ≤ 0.01) influenced genotypes only for hand refractometer brix, juice purity, estimated sugar yield and brix yield.   Significant (p ≤ 0.01) genotype × location (GL) interactions were detected for sucrose content, fibre content, estimated sugar yield and brix yield. Location × crop-year (LY) effects were significant for hand refractometer brix, juice purity, estimated sugar yield and brix yield. Genotype × crop-year interactions (GY) were only significant (p ≤ 0.05) for sucrose content. However, genotype × location × crop-year (GLY) interactions were not significant (p ≤ 0.01) for any of the traits studied. For all the traits mean squares for genotypes were larger than  GL  (1.4 - 4.5  times),  GY  (2.3 - 4.5  times), GLY (4.3 - 8.8 times) and error (2.4 - 4.8 times) mean squares. 

Genetic variability, heritability and genetic advance

Genetic variabilities (GCV) for sucrose content, juice purity, fibre content and sugar yield were higher in first ratoon than plant crop (Table 2). Differences between GCV and phenotypic coefficient of variation (PCV) for all traits were large indicating the influence of environmental factors in the traits. Heritability for the traits ranged from 0.376 for juice quality and sugar yield to 0.685 for refractometer Brix in the plant crop and from 0.321 to 0.81 in the ratoon crop. Expected genetic gains for all the traits in both crop years were less than 10% with sucrose content, sugar yield and Brix yield recording 3.2, 7.9 and 8.1% respectively. Genetic gain for sucrose  content  and  fibre  content   were   much higher in the ratoon than the plant crop. This was probably due to the higher genetic variance for the two traits in the ratoon crop. Error variance components for sugar yield and Brix yield were 19 and 10 fold the respective genotypic components. GCV values for all the cane quality characters were less than 10% in both crops. Most of the quality traits had moderate to high heritability (>0.5) but sugar yield had low heritability (0.321). Genetic parameters from  the  combined  analysis are shown in Table 3. For all traits, error variances were higher than the genetic components. G×Y variances were negligible in all the traits except refractometer Brix. The magnitude of GCV relative to PCV ranged from 54% for sugar yield to 84.4% for sucrose content. Heritability was highest (0.712) for sucrose content and lowest (0.292) for sugar yield. Brix yield (6.7%), sucrose content (5.5%) and sugar yield (5.4%) had highest genetic gains.




Genetic and phenotypic correlations coefficients between sugar yield and its attributes

High genetic and phenotypic correlations were detected between hand refractometer brix and Pol% cane (rg = 0.998, rp = 0.966) and between Pol% cane and juice purity (rg = 0.624, rp = 0.523) (Table 4). Correlations between fibre content and juice quality traits were low.



The association between Pol% cane and fibre content could not be estimated.




Combined analysis of variance and genotype × environment interactions

The significant differences among the clones for all the traits indicate existence of genetic variation in the material. This suggests that opportunities for further improvement through selection do exist. Significant GL interactions for sucrose content (Pol% cane), fibre content, estimated sugar yield and brix yield indicated that the test environments discriminated the sugarcane clones differently. Similarly, significant GY effects for sucrose content indicated the inconsistent nature of this trait from one crop-year to another. However, no significant GL, GY and GLY interactions were detected for hand refractometer brix and juice purity suggesting that performance of clones in these traits was stable over the locations and crop-years. This suggests that the variance components for these traits could be estimated from one location and one crop-year data. Chang (1996) obtained similar results for juice purity in Taiwan.

Interactions of genotypes with environments (GEI) complicate the identification of superior genotypes by plant breeders during selection and cultivar recommendations. GEI have been reported to be a major problem in breeding programmes as they reduce progress from selection (Comstock and Moll, 1963; Mirzawan et al., 1993; Kimbeng et al., 2009). In de Sousa-Vierra and Milligan (2005) reported significant GL and GY interactions for Pol% cane.  In a recent study, Shikanda et al. (2017) reported significant GL interactions for Brix in selected Kenyan clones. Similar results have been reported in other programmes (Tena et al., 2016; Singh et al., 2019). The results of our study suggest that more emphasis should be placed on testing clones in many locations than on testing ratoon crops within locations for reliable selection (Khan et al., 2004).

Variability, heritability and genetic advance

High GCV values suggest good prospects for improvement in the traits by selection. However, Burton (1952) suggested that the use of GCV together with heritability estimates gave a better understanding of heritable variation present in a population. The magnitude of heritability of a trait indicates the effectiveness of selection based on phenotypic observation of the trait (Hanson, 1963). Most quality characters had moderate heritability (>0.4) except sugar yield.  Thus improvement of these traits through selection would be somewhat difficult but more effective than selecting for sugar yield per se. Butterfield and Nuss (2002) reported that effective selection of superior clones depended not only on heritability but also on genetic advance (GA). In this study, moderate GA values were associated with moderate heritability and GCV. The low heritability coupled with low GCV implies large influence of environmental and genotype × environment interaction effects on some traits and limited scope for their improvement. This explains the low expected genetic gain for cane quality traits. Singh (1993) observed that selection for traits with low heritability may be practically difficult. However, Cesnick and Vencovsky (1974) obtained moderate heritability for brix (0.52) and Pol% juice (0.54) and considered that breeding progress for these traits was still possible. The results of our study suggest that brix yield, sugar yield and sucrose content may respond positively to selection and offer opportunities for improvement in the breeding programme. The expected GA for sugar yield, sucrose content, brix yield and cane yield indicate considerable potential for improvement through breeding. Milligan et al. (1990b) and Singh et al. (2019) reported similar observations but with higher GA values.

Genotypic and phenotypic correlations among traits

The strong positive correlation between hand refractometer  brix and sucrose content indicated that the former was a reliable indicator of sugar content in cane. Similar results were obtained by Kang et al. (1983), Milligan et al. (1990b) and Chang (1996). Sucrose content and purity of juice are tedious and costly to measure as they are determined in the laboratory while brix can easily be measured in the field with a hand refractometer and punch. Brix measures total soluble solids in cane juice and a high fraction of these solids contain sucrose thus brix is a useful correlated trait for selection. This study suggests that it is possible to identify varieties with high sucrose content and purity by selecting for high hand refractometer brix.



The study has demonstrated availability of genetic variability among the genotypes for the cane quality traits studied implying that genetic improvement through selection is possible. The presence of GL effects and absence of GY and GLY interactions suggests that sugarcane clones should be evaluated in more locations rather than years/seasons for effective selection. The high expected genetic gains for sucrose content, sugar yield and Brix yield indicates that selection and genetic improvement for these traits would be effective. Refractometer Brix is a reliable correlated trait when selecting for sucrose content.



The authors have not declared any conflict of interests.



Anonymous (1985). Laboratory Manual for South Africa Factories including the official methods. 3rd Ed. Published by SASTA.


Breaux RD (1984). Breeding to enhance sucrose content of sugarcane varieties in Louisiana. Field Crops Research 9:59-67.


Bressiani JA, Da Silva JA, Vencovsky R, Sordi RA, Burnquist WL (2006). Combining high yields of cane and sucrose in sugarcane through recurrent selection. Journal of American Society of Cane Technologists 26:26-37.


Brown JS, Glaz B (2001). Analysis of resource allocation in final stage sugarcane clonal selection. Crop Science 41:57-62.


Butterfield MK, Nuss KJ (2002). Prospects for new varieties in the medium to long term, the effects of current and future breeding strategy on variety characteristics. South African Sugar Industry Agronomists Association, Kwa-Shukela pp. 41-48.


Bureau of Sugarcane Experiment Stations (BSES) (1970). Laboratory manual for Queensland sugar mills, Fifth Edition. Watson, Ferguson and Company, Brisbane P 250.


Burton GW (1952). Quantitative inheritance in grasses. 6th Grassland Congress 1:277-283.


Cesnick R, Venkovsky R (1974). Expected response to selection, heritability, genetic correlations and response to selection of some characters in sugarcane. Proceedings of International. Society of Sugar Cane Technologists 15:96-100.


Chang YS (1996). Estimating heritability of and correlations among brix, purity and sugar content in sugarcane using balanced multiple location and year data. Report of Taiwan Sugar Research Institute 151:1-10.


Clayton JL (1971). System of cane sugar factory control, 3rd Ed. Queensland Society of Sugar Cane Technologists, Brisbane P 87.


Comstock RE, Moll RH (1963). Genotype-environment interactions. In: Hanson WD, Robinson HF (Ed) Statistical Genetics and Plant Breeding. National Academy of Science, National. Research Council Publication No.982. Washington D.C.


Corcodel L, Roussel C (2010). Energy content: a new approach to cane evaluation. Proceedings of International. Society of Sugar Cane Technologists 27:1-7.


De Sousa-Vierra O, Milligan SB (2005). Interrelationships of cane yield components and their utility in sugarcane family selection: Path coefficient analysis. Interciencia 30:92-96.


Hoang NV, Furtado A, Botha FC, Simmons BA, Henry RJ (2015). Potential for Genetic Improvement of Sugarcane as a Source of Biomass for Biofuels. Frontiers in Bioengineering and Biotechnology 3:182.


Hanson WD (1963). Heritability. In: W.D. Hanson and H.F. Robinson (ed) Statistical Genetics and Plant Breeding. National Academy of Science, National Research Council Pub1ication No.982. Washington D.C


Holland JB (2006). Estimating genotypic correlations and their standard errors using multivariate restricted maximum likelihood estimation with SAS Proc Mixed. Crop Science 46:642-654.


Inman-Bamber G (2014). Sugarcane yields and yield-limiting processes. In: Morre, P.H., Botha, F.C. (Eds.), Sugarcane: Physiology, Biochemistry, and Functional Biology. Wiley-Blackwell, pp. 579-600. 


Irvinne JE, Richards CA (1983). Production statistics as evidence of technological improvement. The Louisiana sugar cane industry as a case study. Sugar Cane (May/June) 1:13-18.


Jackson PA (2005). Breeding for improved sugar content in sugarcane. Field Crops Research 92:277-290.


Jamoza JE (2011). KESREF releases new sugarcane varieties. KESREF Digest 5(4):3-4.


Jamoza JE, Owuoch JO, Kiplagat O, Opile W (2014). Broad-sense heritability estimation and correlation among sugarcane (Saccharum spp. hybrids) yield and some agronomic traits estimated in western Kenya. International Journal of Agricultural Policy and Research 2(1):016-025.


Kang MS, Miller JD, Tai PYT (1983). Genetic and phenotypic path analyses and heritability in sugarcane. Crop Science 23:643-647.


Kimbeng CA, Zhou MM, da Silva JA (2009). Genotype × environment interactions and resource allocation in sugarcane yield trials in the Rio Grande Valley Region of Texas. Journal of American Society of Cane Technologists 29:11-24.


Legendre BL (1992). Potential for increasing sucrose content of sugarcane varieties in Louisiana through breeding. Proceedings of International Society of Sugar Cane Technologists 21:367-377.


Khan IA, Khatari A, Siddiqui MA, Nizamani GS, Raza S (2004). Performance of promising sugarcane clones for yield and quality in different ecological zones of Sindhi. Pakistan Journal of Botany 36(1):83-92.


Mariotti JA, Cuenya MI, García de Salas MB (2001). The performance of families in breeding for quality traits in sugarcane. Proceedings of International Society of Sugar Cane Technologists 24:495-499.


Mayo O (1980). The Theory of Plant Breeding. Clarendon Pres-Oxford. New York.


Milligan SB, Gravois KA, Bischoff KP, Martin FA (1990a). Crop effects on broad-sense heritability and genetic variances of sugarcane yield components. Crop Science 30:344-349.


Milligan SB, Gravois KA, Bischoff KP, Martin FA (1990b). Crop effects on genetic relationships among sugarcane traits. Crop Science 30:927-931.


Mirzawan PDN, Cooper M, Hogarth, DM (1993). The impact of genotype × environment interactions for sugar yield on the use of indirect selection in southern Queensland. Australian Journal of Experimental Agriculture 33:629-638.


Priya MS, Reddy KHP, Kumar MH, Rajarajeswari V, Naidu GM (2018). Studies on genetic parameters for diversified uses in sugarcane (Saccharum spp). International Journal of Current Microbiology and Applied Sciences 7(8):4138-4142.


Shikanda E, Jamoza J, Kiplagat O (2017). Genotypic evaluation of sugarcane (Saccharum spp. hybrids) clones for sucrose content in western Kenya. Journal of Plant Breeding and Crop Science 9(3):30-36.


Singh BD (1993). Plant Breeding. 5th ed. Kalyani Publishers, Rajender Nagar, Ludhiana, India pp. 102-104.


Singh G, Mishra KM, Sangera GS (2019). Variability and character association for commercial cane sugar and its components in early maturing sugarcane clones. Agricultural Research Journal 56(2):321-3214.


Tena E, Mekbib F, Ayana A (2016). Heritability and correlation among sugarcane (Saccharum spp.) yield and some agronomic and sugar quality traits in Ethiopia. American Journal of Plant Sciences 7:1453-1477.