Drought Stress Modied Genetic Components and Combining Ability of Cotton Genotypes

Background: Gene action and combining ability of the genotypes are two important components in crop breeding and may be inuenced by environmental stresses which sometimes mislead the selection process. Therefore, an experiment was conducted to assess the behavior of genetic parameters of cotton (Gossypium hirsutum) under drought stress during the year 2017-19. Initially, 50 genotypes of cotton were screened for drought stress tolerance potential by subjecting them to 4 week long drought stress (at 20% eld capacity). Based on physiological parameters, 5 drought tolerant and 4 drought susceptible genotypes (lines and testers respectively) were selected and hybridized using Line × Tester technique to develop F 1 crosses. In the next year, 9 parents and their 20 F 1 hybrids were evaluated in eld under normal irrigation and drought stress (at 50% eld capacity). Results: Stronger specic combining ability effects observed for seed cotton yield, number of bolls, boll weight and lint percentage (%) indicated active role of non-additive genes operating in the direction of lower parents with the exception for number of bolls where dominance worked towards superior parents. Best varietal combinations identied under water stress for number of bolls and boll weight were CIM-496 × NF-801-2-37, DPL-26 × NF-801-2-37, B-557 × NF-801-2-37, B-557 × MNH-129. Variety B-557 was found to be good general combiner followed by BOU-1724 for number of bolls, boll weight and lint % under drought. Conclusion: Combining ability analysis indicated critical role of non-additive component in total variation suggesting weak heritability for all the traits. Direct selection in this material is therefore not advisable. Drought had suppressed variability and restricted the expression of yield characters, however, best identied varietal combinations for seed cotton yield, number of bolls and boll weight may be exploited for the development of hybrid to be grown in drought hit areas of Pakistan.


Introduction
Pakistan, geographically located in semi-arid region earns 90% of agricultural returns from irrigated lands, thus, canal irrigation is the main practice of this area. The country annually suffers 40-60% decrease in agricultural productivity due to extensive water loss owing to seepage and poor lining of water canals as reported by Cheema et al. (2014) and Lashari and Mahesar (2012). Water is the main input of agriculture sector in many regions of the world, but population explosion and unplanned urbanization had resulted in massive contraction of water reservoirs.
Cotton productivity markedly depends on regular supply of water obtained either through irrigation or rainfall (Iftikhar et al. 2012). Stable cotton yield is an important breeding objective of drought tolerance improvement programmes. Genetically controlled physiological responses in cotton plant for example, relative leaf water content, and Relative Cell Injury % help to sustain dry spells during its growing period and have been frequently used as indicator traits in identi cation of drought and heat tolerant plants This study was planned to provide information on water stress tolerance in cotton. Firstly, drought tolerant and susceptible cultivars were selected by subjecting 50 cotton cultivars to variable water conditions to determine variability for drought triggered physiological responses (relative water content; RWC, excised leaf water content; ELWL, relative cell injury %; RCI%) at seedling stage to accomplish selection of drought tolerant and susceptible cultivars. Line × Tester mating design developed by Kempthorne (1957) was used to make crosses using drought tolerant cultivars as lines and susceptible ones as testers to develop plant material for investigating genetic mechanisms controlling water stress tolerance in mature cotton. The advent of molecular techniques has restricted use of conventional breeding to certain extent. Like other environmental stresses, drought not only in uences expression of genes but also modify the inheritance pattern, gene action and combining ability of the genotypes (Mahmood et al. 2020). Therefore, this study was planned to assess the impact of drought stress on inheritance pattern, gene action and combining ability of cotton genotypes in semi-arid regions. The information obtained from this study may facilitate plant breeders to make effective improvements in breeding program for development of drought tolerant cotton varieties. Physiological screening to select potential parents and hybridization During May-June (2017), presoaked seeds of 50 cultivars were sown at 2.5 cm depth in soil lled polythene bags arranged in a triplicated trial under normal (100% eld capacity) and drought conditions (20% eld capacity). During experiment, temperature was recorded to be 35ºC and humidity conditions 65-80% respectively. Stress was imposed at emergence of rst pair of true leaves and continued till four weeks after which data were recorded on physiological traits, RWC, ELWL and RCI% using formulae developed by Clarke & Townley-Smith (1986) and Matin et al. (1989) respectively. RCI% was measured by protocol of Sullivan (1972). Third fully expanded leaf of each of the 50 cultivars was used to record data for RWC and ELWL whereas youngest leaf was used to assess RCI%. Data recorded were then analyzed using ANOVA (Steel et al. 1997). Results of physiological screening exhibited variable responses of cultivars to water stress, but some cultivars including DPL-26, CIM-496, B-557, 149F and BOU-1724 outperformed others by showing maximum water retention (approximately 70%), minimum water loss (38%) and RCI% (39%) under 20% eld capacity and identi ed as drought tolerant whereas four cultivars, MNH-129, FH-1000, H-499 and NF-801-2-37 were shortlisted as drought susceptible for showing poor water retention (below 38%) and maximum RCI% (more than 75%) under 20% eld capacity. These nine cultivars were selected as parents and eld planted during May to September (2018) to develop genetic material keeping tolerant ones as lines (female parents) and susceptible ones as testers (male parents).

Materials And Methods
Seed cotton was harvested from selfed and crossed bolls and separated into lint and cotton seed using a single roller electric gin machine (approximately 20-25 seeds per cross were obtained).

Field evaluation of parents and their hybrids at adult stage
The performance of parents ( ve lines and four testers) and their 20 F 1 hybrids was evaluated during May 2019 by sowing seeds in triplicated eld trial following completely randomized block design under two water levels; 100% (T ) and 50% (T 1 ) eld capacity. Presoaked seeds of 29 entries were sown 30 cm apart within rows and 75 cm between rows. Water stress was imposed at squaring stage (30 days after sowing) and continued till harvesting. Data were recorded on bolls per plant of each family, boll weight and lint % as described below.

Boll number per plant
At maturity, two pickings were made to collect bolls showing maximum opening from normal and water stressed cotton plants. Average boll number for each family was calculated by counting total bolls on each plant in each family to get total number of bolls of a family.

Results
Mean data for average boll number, boll weight and lint % under two water levels were analyzed using ANOVA (Steel et al. 1996) for seeking signi cant differences among genetic material, water treatments and interaction between them (Table 1a, b). Genetic analysis of data was performed following Line × Tester technique (Kempthorne 1957) to estimate general and speci c combining ability effects.

Genetic components of variation under normal supply and drought
Parental and hybrid genotypes grown under normal water supply showed dominant role of non additive genes for boll weight and lint % due to signi cant speci c combining ability (σ 2 sca ). Signi cant gca variance for number of bolls (0.20) showed strong contribution of additive genes. High speci c combining ability (σ 2 sca ) was noted in drought effected cultivars for number of bolls (1.58), boll weight and lint % suggesting active role of non additive genes. A slight contradiction in direction of dominance was observed where unlike boll number, environmental expression of boll weight and lint % was found to be more inclined towards lower parents. Variance ratio calculated for yield contributing traits was below unity and thus controlled by non additive genes. Estimated components of variation are presented in Table 2. showed strong general combining ability for number of bolls and lint %, respectively. Among testers, H-499 was found to be better general combiner for boll weight and lint % than for number of bolls per plant. FH-1000 exhibited better general combining ability for lint % only while MNH-129 turned out to be poor general combiner for all traits (Table 3).    (Gioi et al. 2017). In the present studies, dissection of total variation into its components revealed the presence of greater sca variance for number of bolls, boll weight and lint % under water stress. Sprague and Tatum (1942) and Gri ng (1956) stated that higher magnitude of sca signi es the value of non additive genes in determining expression of plant characters, and is an indication of low heritability (Falconer and Mackay, 1996) Comparison of general combining abilities of nine parents ( ve lines and four testers) revealed B-557 and BOU-1724 as good general combiners for number of bolls, boll weight and lint % under water stress treatment. Among testers, FH-1000 appeared to contribute positively as parent for boll weight and lint %.
It is imperative to mention here that these inferences are con ned to the present investigations. Six crosses revealed good manifestation for number of bolls, for instance, CIM-496 × NF-801-2-37, DPL-26 × NF-801-2-37, B-557 × NF-801-2-37, B-557 × MNH-129, BOU-1724 × FH-1000 and 149F × H-499, and parental cultivars involved in these six crosses had complementary gca, except for CIM-496 × NF-801-2-37 which used both the parents with poor gca. Cultivar MNH-129 didn't make su cient contribution as parent for all the traits, however it combined well with B-557 and BOU-1724 (good general combiners) to exploit its potential for number of bolls and boll weight. Varietal differences observed in parents and hybrids could be justi ed on basis of genotype and environment interaction (Sezener et al. 2015;Rehman et al. 2020). Non-additive gene action has been considered as important feature for the development of hybrids, for which parents with high gca is not always obligatory, nonetheless, parents having different gca may also produce superior combinations as suggested by Fasahat et al. (2016).

Conclusions
Present investigations suggested postponement of selection for potential parents until xation of desirable genes. The utility of this information can be extended by validating these results must be validated through other genetic study using large number of cultivars to draw logical inferences about selection of desirable plants showing wider adaptation potential across multiple environments. Crosses identi ed as superior for average boll number, boll weight and lint % may be utilized for development of hybrids to boost cotton production in water de cit areas. These ndings may positively be used to harness water stress tolerance in commercial cotton cultivars grown in the cotton belt of Pakistan.