Research Article

Diversification Assessment in Cotton for Lint Quality, Disease Response and Economic Attributes by Multivariate Analysis

Muhammad Jamil1*, Kamran Javed1 and Imran Akhtar2

1Cotton Research Station, Vehari, Pakistan; 2Regional Agricultural Research Institute, Bahawalpur, Pakistan.

Received | October 15, 2021; Accepted | May 17, 2022; Published | August 03, 2022

*Correspondence | Muhammad Jamil, Cotton Research Station, Vehari, Pakistan; Email: [email protected]

Citation | Jamil, M., K. Javed and I. Akhtar. 2022. Diversification assessment in cotton for lint quality, disease response and economic attributes by multivariate analysis. Sarhad Journal of Agriculture, 38(3): 1051-1059.

DOI | https://dx.doi.org/10.17582/journal.sja/2022/38.3.1051.1059

Keywords | Correlation analysis, Genetic erosion, Pakistan, Principal component analysis, Yield

Copyright: 2022 by the authors. Licensee ResearchersLinks Ltd, England, UK.

This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Introduction

Cotton is a unique crop due to its yarn production for textile industry. This aspect makes it a major cash crop in a true sense. It is backbone of our national economy by adding 0.81% to GDP and roughly 4 % value addition in agriculture sector (Anon, 2019). Cotton is ranked as the 6th largest source of edible oil globally which is obtained from its seeds as by product (Xu et al., 2016). Its genus Gossypium, consists of 5 allotetraploid and 44 diploid species as primary gene pool. Among all species, only 4 are cultivated globally. Pakistan, India, U.S.A and China are the prominent cotton producing countries contributing two third of the crop area worldwide (Dahab et al., 2013). Central Asian countries like Uzbekistan and Turkmenistan are present at 8th and 10th positions in the world ranking for cotton production respectively, while Tajikistan and Egypt are placed below top 10th ranking positions (Shuli et al., 2018). Textile sector demands huge quantity of superior quality lint to promote national exports and earning foreign currency reserves. Bakhtavar et al. (2015) suggested that development of cotton varieties bearing improved morphological and fiber traits is ultimate solution to combat the bottlenecks in successful production of this crop.

Multivariate analysis is a powerful statistical tool which enabled researchers to explore the multidimensional association among studied variables (Grahic et al., 2013). This approach on cotton genotypes enabled scientists to classify existing gene pool into typical clusters based on lint quality and yield related traits (Shakeel et al., 2015). Variation behavior observed in genotypes during principal component analysis leads scientists to choose parental genotypes for initiating fruitful crop improvement schemes (Isong et al., 2017). Hierarchical clustering procedure followed after principal component analysis (PCA) enhanced chances to select subgroup of clusters at the highest dissimilarity level (Rizwan et al., 2021). In any plant breeding program, genetic diversity plays key role for selection of desirable genotypes (Riaz et al., 2019). Saeed et al. (2014) indicated the classification of imported cotton germplasm into five distinct clusters. Seed cotton yield is a complicated character, highly influenced by metrological conditions (Abbas et al., 2013). For improving such complex polygenic trait, many yield related attributes such as plant stature, number of monopodial branches, sympodial branches count and fruit retention plant-1 are accountable (Anil and Abhay, 2017). Munir et al. (2020) found similarities in scatter behavior of interspecific cotton hybrids with their parents. For quantification of genetic divergence, researchers have extensively employed the PCA procedure which assists in isolation of suitable parental genotypes for further breeding programs (Rehman et al., 2015). Sarwar et al. (2020) reported development of disease resistant cotton genotypes having better yield along with premium lint quality traits in a similar study.

Continuous man made selective breeding efforts in cotton crop resulted genetic erosion and loss of diversity. This process has ultimately led to high yielding cultivars bearing narrow genetic base. Such cultivars with narrow genetic base are highly vulnerable to different biotic and abiotic stress factors. Pakistan is ranked globally at fifth position with respect to facing climate change hazards in the form of uneven rains, elevated temperature in summer and biotic stress in the form of insect pests (Global climate risk index report, 2018). The presence of genetic divergence is vital for successful breeding program to combat unexpected stress on crop plants posed by climate changing conditions as indicated by Jarwar et al. (2019). Trial location in this study is unique and hot spot for cotton leaf curl virus. Recently bred twenty-four cotton strains bearing diversified origin were included in the study never reported earlier. It was hypothesized that crossing between diversified parents will outcome the strains with better adaptability and yield performance.

The main objective of present research was to explore the magnitude of diversification in upland cotton strains under the study. These selected strains may be employed by plant breeders as a potential tool for choosing parental material in future crossing programs.

Materials and Methods

Description of site

The present experiment was carried out at Cotton Research Station, Vehari, Pakistan (located at 72° 37˝E longitude and 30° 25˝N latitude, 175m above sea level). The climate of Vehari district is semi-arid and sub-tropical type. The soil texture was loamy with PH value 8.1 and retained 0.9 % organic matter contents with history of wheat-cotton cropping pattern under irrigated conditions. The available phosphorus and potash were 7 and 136 ppm respectively. Moreover, 36 % water saturation ability made that soil well suited for raising cotton crop. The soil analysis was carried out following the method of Homer and Pratt (1961) at soil and water testing laboratory, Vehari.

Experiment layout and planting material

Twenty-four strains of upland cotton bred by the method of pedigree selection after genetic pyramiding having the diversified origin (Table 1) including check cultivar (CIM-602) were sown in randomized complete block design with three blocks on 23th May, 2019.

 

Table 1: Detail of studied 24 cotton strains.

S. No.	Code	Parentage	S. No.	Code	Parentage
1	S-1901	CIM-632 X VH-402	13	S-19013	VH-447 X NS-181
2	S-1902	QSR-189 X VH-305	14	S-19014	VH-450 X China-1
3	S-1903	Thakar-808 X VH-321	15	CIM-602	Check cultivar
4	S-1904	LRA-5166 X RH-668	16	S-19016	Tipu-9 X BS-18
5	S-1905	ASP-59 X SLH-19	17	S-19017	LS-191 X BS-18
6	S-1906	LS-191 X BS-18	18	S-19018	NIAB-545 X VH-189
7	S-1907	NIAB-545 X NIAB-112	19	S-19019	NIAB-444 X VH-363
8	S-1908	NIAB-444 X VH-305	20	S-19020	US-3090 X RH-662
9	S-1909	MNH-786 X RH-662	21	S-19021	N-414 X FH-490
10	S-19010	N-414 X VH-305	22	S-19022	VH-402 X NIAB-78
11	S-19011	VH-402 X CH-18	23	S-19023	QSR-189 X US-3090
12	S-19012	N-414 X VH-345	24	S-19024	VH-440 X MNH-886

 

Each plot consisted of 7.5 m long 4 cotton rows. The seeding was done on 75 cm wide raised beds by maintaining 20-25 cm spacing within hills. First irrigation was applied immediately after sowing and the subsequent irrigation on five days later to ensure optimum seed germination. All the remaining irrigations were applied at the uniform interval of 14 days till the crop maturity. Thinning was done at the four leave stage and left one healthy plant per hill. The recommended fertilizer dose NPK@ 80:35:30 kg ha-1 was applied to the trial (Rizwan et al., 2021). All the phosphorus and potassium were applied at the time of land preparation. Nitrogen was applied in three equal split doses namely at sowing, at flowering and boll formation stages. Pest populations were kept below the economic injury level by spraying recommended pesticides. Picking was done on 28th November, 2019 after opening of more than 90% bolls determined by counting.

Recording of Data

Ten randomly selected healthy plants were tagged from each plot for recording the data. Plant height was measured in cm with wooden scale starting from base to top of the stem. Number of monopodial branches and mature bolls were recorded by counting each tagged plant and then average of ten plants was calculated. Average boll weight in grams was obtained by picking all open bolls from each tagged plant, weighing seed cotton and dividing by sum of picked bolls. Number of plants were counted in each entry after picking for plant population ha-1. On the picking day seed cotton yield of each plot was obtained and converted into kg ha-1. After sun drying and cleaning the seed cotton samples were taken and ginned with electric single roller ginning machine and calculated Ginning Out Turn % as per following formula.

G.O.T % = (Weight of lint / weight of seed cotton) × 100

Fiber analysis

The lint samples were drawn and evaluated for lint quality parameters after cleaning and sun drying by High Volume Instrument model USTER 1000 using methodology previously reported by Sasser (1981). This equipment provided data of fiber attributes as, fiber length (mm), fiber strength (g tex-1) and micronaire value which indicated lint fineness.

Cotton Leaf Curl Virus Disease index

This trait was measured in % age based on leaf symptoms. Disease response was calculated according to the procedure reported by Sarwar et al. (2020).

Biometric Analysis

Average data of the studied traits was subjected to Analysis of Variance (ANOVA) (Steel et al. 1997). Principal component analysis (PCA) was calculated according to the procedure given by Jackson (1991) Agglomerative hierarchical clustering (AHC) with wards method was performed according to procedure given by Anderberg (1993). Correlation analysis, biplot and the dendogram based on euclidean distance were generated by XLSTAT software.

Results and Discussion

The ANOVA results revealed that the studied cotton strains had significant variation for traits like CLCuD

 

Table 2: Analysis of variance (ANOVA) and descriptive statistics results in 24 cotton strains used in the study.

		Mean squares values
Source	DF	SCY	BC	BW	DI	MB	PH	PP	GOT	FL	FS	MK
Blocks	2	218057	11.542	0.061	44.097	0.117	319.389	1360	0.844	0.942	1.402	0.273
Strains	23	422820A	83.109 A	0.192 A	97.086 B	0.703 A	470.811 A	3050 B	15.006 A	3.119 A	20.604 A	0.478 B
Error	46	51716	1.455	0.803	15.329	0.563	105.374	1047	0.221	0.049	0.280	0.021
Descriptive statistics
Traits			Minimum		Maximum		Mean ± S.E		Std. Deviation		C.V %
Seed cotton yield kg ha-1			534		2452		1280±84.63		414.598		18.63
Average bolls plant-1			19		39		28.458±1.095		5.365		4.2
Average boll weight in (g)			2.4		3.6		2.992±0.084		0.412		9.53
CLCuV disease index %			35		70		57.917±1.901		9.315		18.93
Monopodia plant-1			0		4.6		0.908±0.201		0.985		16.86
Average plant height (cm)			93		164		126.058±3.979		19.492		8.4
Plant population ha-1			23250		38750		32452±796		3900		11.2
Ginning out turn %			36.5		46.6		40.8±0.479		2.349		1.16
Staple length (mm)			23.9		28.34		26.048±0.225		1.102		0.85
Fiber strength ( g tex-1)			19.1		33.5		26.088±0.604		2.957		2.01
Micronaire value			3.89		5.8		4.656±0.082		0.082		3.08

ASignificant (p<0.01) B significant at (p<0.05)

SCY: seed cotton yield in kg ha-1; BC: Number of bolls plant-1; BW: Average boll weight (g); DI: Disease index % ; MB: Number of monopodia plant-1; PH: Average plant height (cm); PP: Plant population ha-1; GOT: Ginning out turn %; FL: Fiber length (mm); FS: Fiber strength (g tex-1); MK: Fiber micronaire value.

 

disease index, plant population ha-1 and micronaire value indicating fiber fineness at (p<0.05), while all other studied traits displayed highly significant variation at (p<0.01) as presented in Table 2. These results provided sufficient grounds for further analysis.

The results of descriptive statistics indicated that ample variation range is present in studied cotton strains. Seed cotton yield ranged from 534 kg ha-1 to 2452 kg ha-1 with 18.63% coefficient of variability (CV). Minimum (0.082) standard deviation from means was recorded in micronaire value, while maximum (3900) for plant population ha-1. The CV value was (0.85%) for fiber length and (18.93%) for CLCuV disease index (Table 2).

Correlation analysis illustrated the presence of few significant association between novel traits (Table 3). Significant correlation among key traits is a good aspect for success of any breeding program, as it enhances chances for isolation of genotypes bearing desirable attributes concurrently (Ali et al., 2009). Seed cotton yield indicated positive (r = 0.271) correlation with boll weight similarly plant height depicted positive (r = 0.366) correlation to boll counts plant-1 at (P < 0.05). These results confirmed the outcomes given by Salahuddin et al. (2010), who discovered positive connection between seed cotton yield and boll weight in upland cotton. The GOT showed highly positive (r = 0.632) links with micronaire value indicating fiber fineness. By increasing GOT lint quantity is increased hence, ultimately quality of lint is deteriorated and lint became coarser possessing higher micronaire value. These results are confirmatory to the earlier findings of Khan et al. (2017). Similarly, monopodial branches depicted positive (r = 0.406) association with plant height at (p<0.01). In general CLCuV disease index indicated negative correlation with all study parameters except for monopodial branches. These results partially contradicted to earlier findings of Saeed et al. (2014) in which CLCuV indicated negative bonding with monopodial branches while studying imported cotton genotypes. The reason behind this contradiction was dominance of virus strain (Burewala) at trial location. The strains studied in this trial were domestically bred and well adapted to this viral disease.

Principal Component Analysis

It is an excellent data mining tool employed to isolate the potential parental combinations for initiation of a successful breeding scheme (Nazir et al., 2013). The prime function of this tool is to allocate the entire

 

Table 3: Pearson correlation coefficients for key traits in 24 cotton strains.

Variable	SCY	BC	BW	DI	MB	PH	PP	GOT	FL	FS	MK
SCY	1
BC	0.251	1
BW	0.271 B	0.112	1
DI	-0.012	0.098	-0.299	1
MB	0.366	0.360	-0.090	0.263	1
PH	0.045	0.366 B	-0.029	-0.186	0.406 A	1
PP	0.113	-0.145	-0.196	-0.099	-0.160	-0.008	1
GOT	0.189	-0.327	0.030	-0.176	0.141	0.239	0.019	1
FL	0.010	0.358	0.058	-0.118	0.087	0.380	0.095	-0.190	1
FS	0.379	0.062	0.013	-0.179	0.232	0.204	-0.043	0.193	0.246	1
MK	0.210	-0.341	-0.190	-0.196	0.048	0.208	0.011	0.632 A	-0.145	0.253	1

A Significant at (p<0.01) B significant at (p<0.05)

Abbreviations: SCY: seed cotton yield in kg ha-1 ; BC: Number of bolls plant-1 ; BW: Average boll weight (g) ; DI: Disease index %; MB: Number of monopodia plant-1 ; PH: Average plant height (cm) ; PP: Plant population ha-1 ; GOT: Ginning out turn % ; FL: Fiber length (mm) ; FS: Fiber strength (g tex-1) ; MK: Fiber micronaire value.

 

variance into the several factors, which is applicable for selection of parental combinations to plan to a valuable breeding scheme (Akter et al., 2009). The results of eleven calculated PC, s depicted maximum cumulative variability packed in order of PC-1 > PC-2 > PC-3 and so on (Table 4).

According to the Eigen value >1 rule as proposed by Kaiser (1960), first five PC, s were selected containing 75.83% of the total variability. Remaining six PC, s represented 24.17 % of the residual variation. The broad range of variation was observed with respect to the factor plains PC-1 and PC-2 viz., 21.33% and 19.38% respectively (Table 4). Jarwar et al. (2019) also found wider variation with respect to PC-1and PC-2 factor plain while study on chines origin cultivars of upland cotton. Higher magnitude of variability squeezed at this factor plains is important and gave clue for the importance of PC-1 and PC-2 for traits under the study.

Each PC is derived from the component factors contributing towards that particular PC. In the present study, following traits showed positive loading trend with respect to magnitude in the order of plant height > fiber strength > monopodial branches plant-1 > seed cotton yield > GOT % age, while slightly negative loading pattern towards attributes in the following order of CLCuV disease index %age > plant population ha-1 in PC-1 (Table 4). Similar factors loading pattern for seed cotton yield, boll counts plant-1, boll weight, GOT % age, fiber length and fiber strength towards PC-1 in cotton was reported earlier by Shah et al. (2018). PC-1 is vital and reliable for the expression of disparity among traits studied in cotton. (Malik et al., 2011). In PC-2 boll counts plant-1 showed strong positive loading trend followed by fiber length, while micronaire value (fiber fineness) indicated highly negative loading attitude followed by GOT % age respectively. These findings confirmed the component loading behavior of plant height, monopodial plant-1, fiber length and miconaire value indicating fiber fineness towards PC-2 in cotton as reported by Sarwar et al. (2020). Hence, current results in comparison to early findings by Nazir et al. (2013) depicted importance of PC-1 and PC-2 for information related to key traits in upland cotton.

One of the informative graphical representations of multivariate dataset is biplot graph which is linked to a relevant data matrix (Sarwar et al., 2020). Since calculated PC, s are unrelated with each other, PC-1 and PC-2 are selected to draw biplots on opposite axis due to packing of maximum amount of variability in them (Table 4). Other researchers like Sarwar et al. (2020) explained results on similar factor plain in the study on upland cotton genotypes. The presence of cotton strains in all biplot graph area indicated occurrence of ample variation for component traits at selected factor plain. The distance of vectors with respect to origin of biplot indicated input magnitude of variables towards total divergence among genotypes studied. The biplot showed contribution of traits in the following order, micronaire value > GOT

 

Table 4: Results of all calculated PC, s to study attributes in 24 cotton strains.

	PC-1	PC-2	PC-3	PC-4	PC-5	PC-6	PC-7	PC-8	PC-9	PC-10	PC-11
Eigen value	2.346	2.132	1.467	1.304	1.092	0.828	0.545	0.474	0.348	0.285	0.180
Variability %	21.33	19.38	13.333	11.854	9.927	7.523	4.954	4.306	3.162	2.588	1.641
Cumulative %	21.33	40.713	54.046	65.9	75.83	83.35	88.3	92.61	95.77	98.36	100
Component matrix/ loading factors
Traits under study			PC-1		PC-2		PC-3	PC-4		PC-5
Seed cotton yield kg ha-1			0.597		0.001		0.004	0.480		0.552
Boll counts plant-1			0.366		0.778		0.012	0.041		-0.023
Average Boll weight in (g)			0.121		0.131		-0.636	0.617		-0.095
CLCuV disease index %			-0.215		0.289		0.778	0.121		0.139
Monopodial plant-1			0.620		0.285		0.531	0.136		-0.032
Plant height (cm)			0.682		0.163		-0.013	-0.420		-0.361
Plant population ha-1			-0.054		-0.164		-0.134	-0.490		0.739
Ginning out turn %			0.458		-0.709		0.091	0.057		-0.166
Fiber length (mm)			0.360		0.480		-0.342	-0.477		0.023
Fiber strength ( g tex-1)			0.641		-0.090		-0.109	0.022		0.210
Micronaire value			0.439		-0.742		0.140	-0.097		-0.094

 

 

% age > boll counts plant-1, whereas minimum divergence was loaded by plant population followed by boll weight (Figure 1). The strain (S-1901) bearing 5.8 miconaire value as highest, the strain (S-1906) giving highest (2452 kg ha-1) seed cotton yield and the strain (S-1910) distinguished as most dwarf in stature bearing (93 cm) plant height depicted unique behavior and found present on periphery of the plot. The distinct scatter behavior of these 3 strains confirmed the loading behavior of respective traits in PC-1 for which trait these strains are famed (Table 2). It also illustrated presence of two groups among 24 cotton strains as proved in cluster analysis.

 

Table 5: Clustering behavior of 24 advance cotton strains on the basis of euclidean distance.

Cluster number	Number/ central objects	Average euclidean distance to centroids	Within class variance	Strains
1	16 (S-1923)	2961.3	15169099	S-1904, S-1905, S-1906, S-1910, S-1911, S-1912 S-1913, S-1914, CIM-602(Check Variety) S-1916, S-1917, S-1919, S-1920, S-1921, S-1922 S-1923,
2	8 (S-1918)	2712.276	10495217	S-1901, S-1902, S-1903, S-1907, S-1908, S-1909 S-1918, S-1924

 

Cluster Analysis

Agglomerative hierarchical clustering (AHC) technique with ward’s method was used in this study based on euclidean distance. The results indicated presence of two clusters in 24 cotton strains. Contrary to these results Munir et al. (2020) reported presence of four clusters in cotton during diversity analysis in interspecific hybrids. This contradiction was due to presence of greater range of variability in cotton hybrids obtained by crossing two species as compared to strains studied here. First cluster contains 16 strains inclusive of check cultivar (Table 5) and second cluster contains 8 strains respectively.

 

The dendogram was plotted on scores of dissimilarities among cotton strains. The dotted line represents automatic truncation, leading to two groups (Figure 2). S-19018 and S-19023 strains were found central objects in their respective class indicating highest euclidean distance from the main centroid, so these are most divergent strains with respect to the selected factor plain. Any strain did not depict obvious separation from clustering behavior as already indicated by Sarwar et al. (2020). It means all study strains were fallen in mentioned two clusters and no strain displayed deviation to the prevailing trend.

Conclusion and Recommendations

PC-1 and PC-2 proved novel for seed cotton yield and lint quality parameters in current study. Two clusters were clearly identified among 24 cotton strains during PCA based cluster analysis. S-19018 and S-19023 were found the most divergent strains with highest euclidean distance between them. These two strains can be crossed to obtain genotypes bearing high seed cotton yield and superior fiber traits. Such upcoming genotypes will help to boost cotton production in the country.

Acknowledgements

All authors are thankful to technical team of fiber testing laboratory, cotton research station, Faisalabad, Pakistan for providing facility of fiber analysis related to this trial and soil and water testing laboratory, Vehari, Pakistan for soil samples analysis.

Novelty statement

Multivariate analysis is a modern technique used for examining the relationship of multiple variables simultaneously. XLSTAT software was applied in present research endeavor, rarely exploited by researchers for data analysis and graphics. Narrow genetic base is an ultimate consequence of man directed evolution faced by cotton cultivars in Pakistan, which resulted decline in cotton production in the country. Changing climate is an additional factor contributing towards cotton decline. Upcoming genotypes as a result of crossing between divergent strains found in this study will be a solution to above mentioned pitfalls.

Author’s Contribution

Muhammad Jamil: Created the idea, reviewed the literature and wrote final draft of this paper.

Kamran Javed: Performed trial in the field and recorded data.

Imran Akhtar: Analyzed data and prepared figures and tables.

Conflict of interest

The authors have declared no conflict of interest.

References

Abbas, H.G., A. Mahmood and Q. Ali. 2013. Genetic Variability, Heritability, Genetic Advance and Correlation Studies in Cotton (Gossypium hirsutum L.). Int. Res. J. Mic., 6(4):156-161.

Akram, M.T., R.W.K. Qadri, M.J. Jaskani and F.S. Awan. 2019. Ampelographic and genetic characterization of grapes genotypes collected from potohar region of Pakistan. Pak. J. Agric. Sci., 56(3): 595-605

Akter, A., M.J. Hasan, A.K. Paul, M. Mutlib and M.K. Hossain. 2009. Selection of parent for improvement of restorer line in rice (Oryza sativa L.). SAARC J. Agric., (7): 43-50.

Ali, M.A., A. Abbas, M. Younas, T.M. Khan and H.M. Hassan. 2009. Genetic Basis of Some Quantitative Traits in Upland Cotton. Plant Omics J. 2(2):91-97.

Anderberg, M.R. 1993. Cluster analysis for application. Academic Press, New York.

Anil, D.A. and S. Abhay. 2017. Study on genetic variability for the quantitative traits in some genotypes of upland cotton (Gossypium hirsutum L.). Bio. Sci. Discov., 8(3):365-368.

Anon. 2019. Pakistan economic survey. Finance division, economic advisor’s wing, Islamabad, Pakistan.

Bakhtavar, M.A., I. Afzal, S.M.A. Basra, A.U.H. Ahmad and M.A. Noor. 2015. Physiological strategies to improve the performance of Spring Maize (Zea mays L.) planted under early and optimum sowing conditions. PLoS One, 10(4): 1-15. https://doi.org/10.1371/journal.pone.0124441

Dahab, A.A., M. Saeed, B.B. Mohamed, M.A. Ashraf, A.N. Puspito, K.S. Bajwa, A.A. Shahid and T. Hussnain. 2013. Genetic diversity assessment of cotton (Gossypium hirsutum L.) genotypes from Pakistan using simple sequence repeat markers. Aust. J. Crop Sci., 7: 261-267.

Global climate risk index report.2018. The Long-Term Climate Risk Index (CRI): The 10 countries most affected from 1999 to 2018. Page-9. https://www.germanwatch.org/sites/default/files/20-2-01e%20Global%20Climate%20Risk%20Index%2020.

Grahic, J., F. Gas, M. Kurtovi, L. Kari, M. Iki and D. Gadz. 2013. Morphological evaluation of common bean diversity in Bosnia and Herzegovina using the discriminant analysis of principal components multivariate method. Genetika, 45:963-977. https://doi.org/10.2298/GENSR1303963G

Homer, D.C. and P.F. Pratt. 1961. Methods of analysis for soils, plants and waters. Davis: University of California, Davis.

Isong, A., P.A. Balu and P. Ramakrishna. 2017. Association and principal component analysis of yield and its components in cultivated cotton. Electron J. Plant Breed., 8: 857-864. https://doi.org/10.5958/0975-928X.2017.00140.5

Jackson, J.E. 1991. A user’s guide to principal components. John Wiley and Sons Inc., New York. https://doi.org/10.1002/0471725331

Jarwar, A.H., X. Wang, M.S. Iqbal, Z. Sarfraz, L. Wang, Q. Ma and F. Shuli. 2019. Genetic divergence on the basis of principal component, correlation and cluster analysis of yield and quality traits in cotton cultivars. Pak. J. Bot., 51(3): 1143-1148. https://doi.org/10.30848/PJB2019-3(38)

Kaiser, H.F. 1960. The application of electronic computers to factor analysis. Educ. Psychol. Meas., 20: 141-151. https://doi.org/10.1177/001316446002000116

Khan, M.I., H.A. Haq, K. Ullah, M. Arshad and A. Majid. 2017. Genetic diversity and correlation studies for cotton leaf curl disease (CLCuD), fiber and yield related attributes in exotic lines of (Gossypium hirsutum L.) Am. J. Plant Sci., 8: 615- 624. https://doi.org/10.4236/ajps.2017.83042

Malik, W., M.Z. Iqbal, A.A. Khan, E. Noor, A. Qayyum and M. Hanif. 2011. Genetic basis of variation for seedling traits in (Gossypium hirsutum L.) Afr. J. Biotech., 10: 1099-1105

Munir, S., M.K. Qureshi, A.N. Shahzad, I. Nawaz, S. Anjam, S. Rasul and M.A. Zulfiqar. 2020. Genetic dissection of interspecific and intraspecific hybrids of cotton for morpho-yield and fiber traits using multivariate analysis. Pak J. Agric. Res., 33(1): 9-16. https://doi.org/10.17582/journal.pjar/2020/33.1.9.16

Nazir, A., J. Farooq, A. Mahmood, M. Shahid and M. Riaz. 2013. Estimation of genetic diversity for CLCuV, earliness and fiber quality traits using various statistical procedures in different crosses of (Gossypium hirsutum L.) Vestnik Orel Gau., 43(4):2-9.

Rehman, S.U., M.A. Abid, M. Bilal, J. Ashraf, S. Liaqat, R.I. Ahmed and G. Qambar. 2015. Genotype by trait analysis and estimates of heritability of wheat (Triticum aestivum L.) under drought and control conditions. Basic Res. J. Agric. Sci. Rev., 4(4):127-134.

Riaz, A., M.H.N. Tahir, M. Rizwan, S. Fiaz, S. Chachar, K. Razzaq, B. Riaz and H. Sadia. 2019. Developing a selection criterion using correlation and path coefficient analysis in sunflower (Helianthus annuus L.). Helia, 42:85-99. https://doi.org/10.1515/helia-2017-0031

Rizwan, M., J. Farooq, M. Farooq, A. Sarwar, A. Ali, F. Ilahi, M. Asif and G. Sarwar. 2021. Quantitative studies in upland cotton (Gossypium hirsutum L.) using multivariate techniques. Pak. J Agric. Res., 34(1): 113-120. https://doi.org/10.17582/journal.pjar/2021/34.1.113.120

Saeed, F., J. Farooq, A. Mahmood, T. Hussain, M. Riaz and S. Ahmad. 2014. Genetic diversity in upland cotton for cotton leaf curl virus disease, earliness and fiber quality. Pak. J Agric. Res., 27(3):226-236.

Salahuddin, S., S. Abro, M.M. Kandhro, L. Salahuddin and S. Laghari. 2010. Correlation and path coefficient analysis of yield components of upland cotton (Gossypium hirsutum L.). World Appl. Sci. J., 8: 71-75.

Sarwar G., A. Nazir, E. Shahzadi and A. Mahmood. 2020. Assessment of genetic diversity in cotton genotypes for earliness, CLCuV disease tolerance, fiber quality and yield related traits. J Agric. Res., 58(3):165-170.

Sasser, P.E. 1981. The basics of high volume instrumentation for fiber testing. Proc. Belt wide cotton prod. Res. Conf., New Orleans, LA 4-8 Jan., Natl. Cotton council. Am., Memphis, TN. pp. 191-193.

Shah, S.A.I., S.J. Khan, K. Ullah and O.U. Sayal. 2018. Genetic diversity in cotton germplasm using multivariate analysis. Sarhad J. Agric., 34(1): 130-135. https://doi.org/10.17582/journal.sja/2018/34.1.130.135

Shakeel, A., I. Talib, M. Rashid, A. Saeed, K. Zia and M.F. Saleem. 2015. Genetic Diversity among Upland Cotton Genotypes for Quality and Yield Related Traits. Pak. J. Agric. Sci., 52 (1):73-77.

Shuli, F., A.H. Jarwar, X. Wang, L. Wang and Q. Ma. 2018. Overview of the cotton in Pakistan and its future prospects. Pak. J. Agric. Res., 31(4): 396-407. https://doi.org/10.17582/journal.pjar/2018/31.4.396.407

Steel, R.G.D., J.H. Torrie and D.A. Dickey. 1997. Principles and procedure of statistics: A biometrical approach 3rd ed. McGraw-Hill Book Co., New York.

Xu, Z., J. Li, X. Guo, S. Jin and X. Zhang. 2016. Metabolic engineering of cottonseed oil biosynthesis pathway via RNA interference. Sci. Rep., 6: 33342. https://doi.org/10.1038/srep33342