Background

During the last century, breeding of high yielding varieties alone has contributed nearly 50% of global increase (32 mt annually) in food production. It has been estimated that an additional annual increase of 12mt food grains will be needed to meet the global food demand in the coming decades. In this context, India needs to produce 120 mt by 2030 to feed its one and a half billion plus population (Adhya, 2011). Genetic improvement of rice in India and many other rice growing countries has already achieved yield plateau for medium land irrigated rice ecosystems, but areas pertaining to rainfed rice ecosystems particularly upland situations are constrained with low productivity owing to ecological adversities e.g., erratic rain fall distribution, extremes of temperature or mineral supply either transient or lasting throughout the growing season. Globally, about 100 million people now depend on upland rice as their daily staple food. Almost two-third of the upland rice area is in Asia and in particular Bangladesh, Cambodia, China, India, Indonesia, Myanmar, Thailand, and Vietnam have major area under upland rice. In India, 62% of total area under rice is rain fed (Adhya, 2011). Drought is the predominant cause of yield reduction under rainfed rice production systems, Cultivation of drought tolerant rice varieties can reduce the demand for irrigation water by 50–70%. However, the ongoing breeding strategies have not been able to make any breakthrough to develop drought-tolerant cultivars. The complex nature of drought tolerance, genotype × environment interaction, and the difficulty of effective drought tolerance screening complicate the development of drought tolerant varieties. The existing genotypes are rarely adaptable to pre and post-monsoon drought situation. It has been contemplated that climatic change will compel rice breeders to reorient the breeding strategies and develop high yielding short duration rice varieties to combat the recurrent occurrence of drought in coming years. Continued human selection and monoculture resulted loss of genes and narrowing of the genetic base in rice (Fu and Somers, 2009; Wouw et al., 2009) resulting increased vulnerability to disease epidemics, pest infestation and abiotic stresses. The paradox is that in order to enable to develop truly revolutionary new cultivars of tomorrow, plant breeders will need to have access to the wealth of genes which exist now, only in exotic and/or local genetic backgrounds.

Seed yield is a complex trait and it is determined by mutual relationship among the component morpho-economic traits. Indiscriminate selection of plants on a massive scale in any crop often results in an immense wastage of time and resources. Success rate will improve with the increase of genetic variability and availability of a selection strategy. Many often unfavorable linkages among the agro-economic traits do exist resulting in genetic slippage and limited genetic advance. The strength of inter-relationship among traits varies depending on the composition of the test materials, characters studied, previous selection history and the environment under which the breeding materials are tested. Correlation analysis reveals the direction and magnitude of the relationships between any given pair of traits without regards to cause/effect relationship. Whereas, study of path analysis (a standardized partial regression analysis) is more important over correlation, in that, it partitions the total correlation coefficients with seed yield into various direct and indirect effects. Therefore, an attempt has been made to assess inter-relationship between agro-economic traits in a set of upland land races of rice to formulate an effective selection strategy to tailor desirable gene combinations through recombination breeding.

1 Results and Discussion

Rice is more sensitive to drought at the reproductive stage and causes the drastic reduction in grain yield (Cruz and O’Toole, 1984). Seed yield is an artifact. Its genetic architecture can be better resolved through interactions between various yield components (Grafius, 1959).Therefore, estimation of phenotypic correlations between grain yield and component characters as well as inter se association can provide information for choice of characters in selection programme. The strength of character association as measured by estimates of co-efficient of correlation depends upon the composition of the test materials, characters studied, previous selection history and the environment under which the breeding materials were tested. A perusal of Table-1 indicated phenotypic correlation co-efficients between agro-economic traits including drought tolerance parameters, physiological traits, tolerance to zinc and bacterial leaf blight.

In the present investigation, number of ear bearing tillers/m2, panicle weight, grains/panicle, 1000-grain weight, grain length, kernel length, kernel breadth and fertility percentage correlated significantly with grain yield/ha. All these component traits except 1000-grain weight recorded significant value at even 1% level of significance. Among these traits, the strength of positive association of number of ear bearing tillers/m2 (r = 0.595*) and panicle weight (r = 0.485**) and fertility percentage(r= 0.458**) with grain yield were very high indicating their importance for genetic improvement of productivity in upland rice. The latter two component traits were also shown to have very high inter se significant positive correlation (r = 0.476**) at even 1% level of significance. Therefore, selection for any one of these two characters automatically selects the other trait and thus, togatherly could result recovery of high grain yield. 1000-grain weight under drought stress was reported significantly and positively correlated with seed yield (Sellammal et al., 2014). Higher grain weight is in vogue result increase in panicle weight. Awasthi and Lal (2014) reported positive and significant association of panicle weight with seeds per panicle, fertility percentage, leaf rolling and yield per plant over locations under drought stress. However, Manjappa et al. (2014) indicated significant correlation of panicle weight, panicle length and number of productive tillers plant-1 with grain yield. It was worth to note that panicle weight and fertility percentage had strong negative association with days to 50% flowering and days to maturity. This suggests that there is ample scope for selection of early maturing genotypes with high panicle weight and grain fertility status. The merit of these two characters is further ascribed to their positive correlated response through other component traits e.g., grains/panicle, 1000-grain weight, and kernel breadth. Besides, grains per panicle and 1000-grain weight had also shown significant negative correlation with the above flowering and maturity traits indicating more scope for selection of early maturing plant types with high grain number and bold seeds. Basavaraja et al. (1997) stressed the importance of panicle weight for selection of grain yield. Nath and Talukdar (1997) reported significant association of number of grains per panicle with grain yield. However, Mishra and Verma (2002), Goswami et al. (2000), Choudhury and Motiramani (2003) observed significant association of number of effect bearing tillers per plant with grain yield. Fertile grains per panicle registered positive significant association with panicle length (Gupta et al., 1999), number of productive tillers per plant (Satish et al., 2003) and total spikelets per panicle.

Plant height seems to have significant positive correlation with flowering and maturity traits. But each of these seems to have no association with grain yield. Besides, plant height maintained no significant correlation with any of the yield contributing traits indicating that there is less likely to improve seed yield through selection for comparatively reduced culm length.

Grain and kernel dimension were assessed in terms of length, breadth and length/breadth ratio. In the present investigation, efforts have been taken to study and implicate the relationship of some easily observable physical quality traits with yield. Genotypes with improved grain filling under moisture stress are in vogue considered to harbor drought tolerance. Kernel length and kernel breadth had strong positive relationship with grain yield at even 1% level of significance (Table 1). While, grain length correlated positively with grain yield at 5% level of significance. Slender kernel types attract consumer’s preference. But, such genotypes, in general, are observed to have low yield potential. Length and breadth of grain and kernel individually revealed significant positive association with 1000-grain weight. Breadth of grain and kernel had positive significant relationship with panicle weight. Besides, bold kernel type was shown to have significant positive association with grain fertility percentage which is considered as one of the most important criterion for yield improvement in upland rice. Grain width was reported highly correlated with yield per plant under drought stress (Haider et al., 2014 and Haider et al., 2015). Thus, selection of long bold grain may certainly pave the way for genetic improvement of seed yield in upland rice.

In the present investigation, kernel length exhibited significant positive correlation with grain length, grain L/B ratio and kernel L/B ratio as also reported by Vivekanandan and Giridharan (1998). Chouhan et al. (1995) reported significant positive association of kernel length and kernel L/B ratio but negative association with kernel breadth in some crosses.

Leaf area exhibited positive significant relationship with grain length/grain breadth and fertility percentage while, Chlorophyll index had shown significant negative correlation with kernel length. This indicates that more leaf area may be associated with slender grain genotype (low yielding) while, high chlorophyll content may result bold grain types (high yielding). In contrast, BLB infestation increased in upland rice with more green leaves as indicated from significant positive relationship of Chlorophyll index with BLB score (0-9 scale). Susceptibility to zinc seems to be associated with early maturing upland genotypes as evidenced from inverse relationship between the traits. Hence, mid early maturing genotype with moderate leaf area and chlorophyll content may be selected for higher productivity under upland condition.

Drought tolerance was assessed based on extent of leaf rolling, leaf drying and recovery due to irrigation following acute drought stress. These parameters exhibited positive significant correlation between each other. However, the strength of association between leaf rolling (LRS) and drought recovery score (DRS) was as high as 0.963, while either of these associated with leaf drying score (LDS) with correlation co-efficient around 0.50. Thus, either LRS or DRS may be useful for assessment for drought tolerance. In this context, leaf rolling being easily assessable under drought stress, it may serve as an useful criterion for screening of germplasm for drought tolerance. In the present set of materials, neither of these drought tolerance parameters was found to have any significant positive association with seed yield. Seed yield under drought stress was reported significantly and negatively associated with leaf rolling score (Sellammal et al. 2014 and Manjappa et al. 2014). Plant height had shown negative significant relationship with LDS at 5% significant level. Besides, all the drought tolerance parameters correlated negatively (though not significant) with flowering and maturity duration. This indicates that drought tolerant genotypes may be associated with intermediate plant types with moderate flowering and maturity duration.

The correlation co-efficients of component traits with grain yield were partitioned into their direct and indirect effects on grain yield following path co-efficient analysis to ascertain further conclusive information on choice of characters required for selection of high yielding genotypes under drought stress. Number of ear bearing tillers/m2 (EBT), panicle weight (PW), grain length (GL) followed by grain fertility percentage (F%) and kernel breadth (KB) exhibited high direct effects on grain yield (Table 2) as well as significant positive association with grain yield. Whereas, kernel length had revealed negative direct effect indicating limited scope for genetic improvement of grain yield based on selection through longer kernel. It is worth to note that grains per panicle had negligible direct effect, but was shown to have high indirect effect on grain yield via panicle weight (PW). Similar was the case for kernel length which had no noticeable direct effect on grain yield, but contributed much towards high productivity indirectly via grain length. Hence, for improving grain yield, emphasis must be given for selecting higher number of effective tillering plant types having heavy panicle with more number of fertile grains. Gupta et al. (1999) carried out path analysis and revealed that panicle weight, panicles per plant and grains per panicle were important traits influencing the grain yield. However, several other workers also reported positive direct effects of effective bearing tillers per plant on grain yield (Nayak et al., 2001; Singh et al., 2002; Choudhury and Motiramani, 2003; Khedikar et al., 2004; Behera, 2007). Besides, proline content was reported to have high direct effect while, harvest index had moderate direct effect on single plant yield under drought condition (Sathya and Jebaraj, 2013).

Lower the score value of BLB infestation on 0-9 scale, more was the degree of tolerance to drought stress based on leaf rolling. Besides, as expected; BLB score had significant negative correlation with grain yield while, the erstwhile mentioned none of drought tolerance parameter recorded significant positive relation with seed yield. Further, the present set of materials revealed negative direct effect of the above biotic and abiotic stress related traits on seed yield suggesting ample scope for selection of high yielding drought tolerant and BLB tolerant genotypes based on leaf rolling score and BLB score.

2 Materials and Methods

The experimental materials used in the present investigation consisted of ninety six land races including a few popular upland rice varieties. The test genotypes were grown in drought condition following Randomized Block Design (RBD) with three replications to assess yield and ancillary traits. The crop was raised following recommended package of practices. Observations on agro-economic traits were recorded on 5 randomly selected plants from middle row of each plot for 23 biometric traits except days to 50% flowering, days to maturity and grain yield which were recorded on plot basis. For 1000-grain weight, observation was taken from random sample of seeds of each plot respectively. Routine statistical procedures were followed for analysis of variance and covariance as per Singh and Choudhury (1976). Estimates of the correlation coefficient for each pair of characters were computed following Al-Jibouri et al. (1958) and the significance of  correlation coefficients was tested by ‘t’- test at n -2 degrees of freedom.

References

Adhya T.K., 2011, “Vision 2030”, Central Rice Research Institute, Indian Council of Agricultural Research (ICAR), Cuttack (Orissa), India

Al-Jibouri H.A., Mulle P.A., and Robinson H.F., 1958, Genotypic and environmental variances and covariances in upland cotton crosses of interspecific origin, Agron. J., 50:633-636

https://doi.org/10.2134/agronj1958.00021962005000100020x

Awasthi S. and Lal, J.P., 2014, Investigating phenotypic correlation and path analysis in rice (Oryza sativa L.) under irrigated and rainfed conditions, Int. J. Scientific Footprints, 2(5): 22-33

Basavaraja P.; Rudradhya M., and Kulkarni R.S., 1997, Genetic variability, Correlation and Path-analysis of yield components in F2 population of fine grain rice, Mysore J. Agric. Sci., 31: 1-6

Behera T., 2007, Genetic basis of yield variation in indigenous aromatic rices of Orissa, M.Sc (Ag.), Thesis, Plant Breeding and Genetics, OUAT, Bhubaneswar

Chaudhury M. and Motiramani N.K., 2003, Variability and association among yield attributes and grain quality in traditional aromatic rice accessions. Crop Improv., 30 (1) :84-90

Chauhan J.S., Chouhan V.S., and Lodh S.B., 1995, Cooking quality components and their interrelationship with some physic-chemical characters of rainfed upland rice grain, Oryza, 32: 79-82

Cruz R.T., and O’Toole J.C., 1984, Dry land rice response to an irrigation gradient at flowering stage, Agronomy Journal, 76: 178-183

https://doi.org/10.2134/agronj1984.00021962007600020003x

Fu Y.B. and Somers D.J., 2009, Genome-wide reduction of genetic diversity in wheat breeding, Crop Sci., 49:161-168

https://doi.org/10.2135/cropsci2008.03.0125

Grafius J.G., 1959, Genetic and environmental relationship of components of yield, maturity and plant height in F2 - F3 soybean populations, Iowa State Coll. J. Sci., 30: 373-374

Gupta K.R, Panswar D.V.S., and Kumar R., 1999, Character association in segregating population in basmati rice, Oryza, 36 (1): 16-19

Haider Z., Farooq U., Naseem I., and Zia, S., 2015, Impact of drought stress on some grain quality traits in rice (Oryza sativa), Agric. Res., 4(2): 132-138

https://doi.org/10.1007/s40003-015-0148-8

Haider Z, Mehboob A, Razaq A, Khalid U., Rasool N., and Mehmood K., 2014, Effect of drought stress on some grain quality traits in rice (Oryza sativa L), Academia Journal of Agricultural Research, 2(5): 128-133

Khedikar V.P., Bharose A.A., Sharma D., Khedikar Y.P. and Khillare A.S., 2004, Path coefficient analysis of yield components of scented rice. J. Soils Crops, 14: 198-210

Manjappa G. & Shailaja H., 2014, Association analysis of drought and yield related traits in F2 population of Moroberekan/IR 64 rice cross under aerobic condition, Int. J. Agril. Sci & Res., 4(2): 79-88

Mishra L.K., and Verma R.K., 2002, Correlation and path coefficient analysis for morphological and quality traits in rice (Oryza sativa L.), Plant archives, 2(2): 275-284

Nath N. and Talukdar P., 1997, Genetic variability and correlation studies in segregatingpopulations of indigenous scented X high yielding non- scented crosses of rice. Oryza, 34(2): 91-93

Nayak A.R., Chaudhury D., and Reddy J.N., 2001, Correlation and path analysis in scented rice(Oryza sativa L.), Indian J. Agric. Res., 35(3): 186-189

Sathya R., and Jebaraj S., 2013, Inter-relationship and Cause Effect Analysis among drought and physiological traits in three line aerobic rice hybrids, Plant Gene and Trait, 2013, 4(13): 70-73

https://doi.org/10.5376/pgt.2013.04.0013

Satish Y., Seetharamanmaiah K.V., Srerama N., and Naidu T.C.M., 2003, Genetic variability and genetic advance in scented rice. The Andhra agric. J. 50(3-4): 232-234

Sellamaal R., Robin S. and Raveendran M., 2014, Association and heritability studies for drought resistance under varied moisture stress regimes in backcross inbred population of rice. Rice Science, 21(3): 150−161

https://doi.org/10.1016/S1672-6308(13)60177-8

Singh R.K., and Choudhary B.D., 1976, Biometrical techniques in genetics and plant beeding, (Varghese TM ed.), International Bioscience Publishers, Hissar, India p. 56-71

Singh, L., Singh J.D., and Sachan N.S., 2002, Inter-character association and path analysis in Paddy, Annals of Biology, 18(2): 125-128

Vivekanandan P., and Giridharan S., 1998, Genetic variability and character association for kernel and cooking quality in rice, Oryza, 35(3): 242-245

Wouw M., Hintum T., Kik C., Treuren R., and Visser B., 2010, Genetic diversity trends in twentieth century crop cultivars: a meta analysis, Theor. Appl. Genet., 120: 1241-1252. DOI 10.1007/s00122-009-1252-6

https://doi.org/10.1007/s00122-009-1252-6