NC State
Mohamad Ismail, F. N., Abdul Hamid, H., Mohamed, J., Abdu, A., Samdin, Z., and Mohd Razali, S. (2020). "Character association and selection of breeding line based on morphophysiological characteristics and tensile strength in Hibiscus cannabinus L.," BioRes. 15(4), 8883-8908.


This study aimed to access different desirable characteristics of nine Hibiscus cannabinus L. accessions based on morphophysiological characteristics and fiber tensile strength for an effective selection of H. cannabinus plant improvement. Four China accessions (FH952, T15, T17, and T19), four Bangladesh accessions (HC2, HC95, V4202, and V4383), and V36 (control) accession were examined in a four-month cultivation period. The experimental design was arranged using randomized complete block design with three replications. Stem diameter was found to be significantly related (p ≤ 0.05) with all morphological and yield characteristics except for leaf dry weight and growth efficiency. Bigger stem diameter was an indicator of fiber yield in attempts to apply crossing and selection to improve performance. Photosynthesis rate also was found to be significantly related (p ≤ 0.05) with stomatal conductance, transpiration rate, instantaneous water use efficiency, and carboxylation efficiency. High photosynthesis rate could be an indicator to interpret the pattern of genetic variation of plant assimilation rate and its relation with environmental and agronomic factors. The fiber tensile modulus, however, was found to be inversely correlated with fiber diameter. The present study suggests the selection of control, V4383, HC2, and FH952 accessions for a breeding line as they possess high fiber yield, fiber strength, and photosynthetic efficiency.

Download PDF

Full Article

Character Association and Selection of Breeding Line Based on Morphophysiological Characteristics and Tensile Strength in Hibiscus cannabinus L.

Fatin-N. M-Ismail,*,a Hazandy A-Hamid,a,b Johar Mohamed,a Arifin Abdu,b Zaiton Samdin,a,c and Sheriza M-Razali a

This study aimed to access different desirable characteristics of nine Hibiscus cannabinus L. accessions based on morphophysiological characteristics and fiber tensile strength for an effective selection of H. cannabinus plant improvement. Four China accessions (FH952, T15, T17, and T19), four Bangladesh accessions (HC2, HC95, V4202, and V4383), and V36 (control) accession were examined in a four-month cultivation period. The experimental design was arranged using randomized complete block design with three replications. Stem diameter was found to be significantly related (p ≤ 0.05) with all morphological and yield characteristics except for leaf dry weight and growth efficiency. Bigger stem diameter was an indicator of fiber yield in attempts to apply crossing and selection to improve performance. Photosynthesis rate also was found to be significantly related (p ≤ 0.05) with stomatal conductance, transpiration rate, instantaneous water use efficiency, and carboxylation efficiency. High photosynthesis rate could be an indicator to interpret the pattern of genetic variation of plant assimilation rate and its relation with environmental and agronomic factors. The fiber tensile modulus, however, was found to be inversely correlated with fiber diameter. The present study suggests the selection of control, V4383, HC2, and FH952 accessions for a breeding line as they possess high fiber yield, fiber strength, and photosynthetic efficiency.

Keywords: Kenaf; Fiber; Morphology; Physiology; Tensile strength

Contact information: a: Laboratory of Bioresource Management, Institute of Tropical Forestry and Forest Products, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia; b: Department of Forestry Science and Biodiversity, Faculty of Forestry and Environment, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia; c: Department of Economy, Faculty of Economy and Management, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia;

* Corresponding author:


Kenaf, which is scientifically named Hibiscus cannabinus L., is an agro-based lignocellulosic that is grown specifically for its fiber. Various annual crops such as cotton (Gossypium hirsutum), hemp (Cannabis sativa), jute (Corchorus capsularis), kapok (Ceiba petandra), and coconut (Cocos nucifera) are known to produce natural fibers. H. cannabinus however has advantages of harvesting operation and cost-effectiveness because the fiber is obtained from its vegetative plant part (stem) instead of reproductive plant part, and the fiber yields are greater than those of the above mentioned crops (Tahery 2011). Being comprised of 65% woody inner core and 35% fibrous outer bast makes H. cannabinus a hard fiber plant with an excellent source of cellulose fiber (Tahir et al. 2015). The yield and fiber quality such as fiber diameter, fiber length, and fiber strength, however, differ greatly among cultivars; these traits can be affected by genetic inheritance and manipulation, environmental factors, and also by their interactions (H’ng et al. 2009; Abdul Khalil and Suraya 2011; Faruq et al. 2013).

The yield, for example, is firmly related to plant growth performance and their physiological attribute such as photosynthesis rate (Evans 2013). The products of photosynthesis are utilized in the development and differentiation of new cells as well as supporting the existing tissue of the plant. The greater photosynthetic rate and high specific activities of RuBP carboxylase possess higher biomass production as well as fiber yield. The superior photosynthetic rates and high biomass yield are also related to the amount and the efficiency of carboxylation capacities in leaf that considered as regulating the photosynthetic capacity in C3 plants (Acquaah 2007). However, Reddy and Das (1986) suggest that losses of available energy to support photosynthesis are due to physical properties of leaves and more fundamental energy considerations for successful conversion and storage of sunlight as chemical energy in the photosynthetic process. To assess the efficiency of the photosynthesis process in a plant, knowledge of total plant biomass production must be acquired. Morphological and physiological attributes also help to better understand the plant plasticity and adaptive mechanism (Abdul-Hamid et al. 2009).

Furthermore, fiber strength and micronaire are associated with cellulose deposition within fiber and the developmental process, which are related to fiber secondary wall characteristics (Wang et al. 2009). The manipulation of the relationship between the photosynthetic assimilate source and the reproductive sink also would affect fiber strength and micronaire formation (Chen et al. 2017). Lokhande and Reddy (2014) reported that a decrease in leaf water potential resulted in decreases in fiber properties such as fiber length, strength, and uniformity, but an increase in fiber micronaire. However, the alternation in the available assimilate supply to the developing bolls affects fiber quality less directly than the yield (Pettigrew 2001). Therefore, better understanding of the responses of leaf photosynthesis is crucial in order to enhance the yield and improve fiber quality (Long et al. 2006; Khan et al. 2019). In this study, the authors investigated morphophysiological quantitative characteristics and tensile strength of nine H. cannabinus accessions as a breeding line selection for a future H. cannabinus breeding program.



Nine H. cannabinus accessions were selected in this study from three different origins (Table 1). The accessions were selected based on the availability in the previous study of the accessions in Malaysia. Amongst them, V36 is one of the main accessions in seed and fiber production in Malaysia; therefore it was selected as the control accession in present study to ensure better selection of accessions. Approximately 100 g of seeds with three replications (100 seeds x 3 replications: 300 seeds) for each accession were prepared for seed moisture content (oven-dry method), and 100 seeds with three replications (100 seeds x 3 replications x 2 methods: 600 seeds) were used for germination tests for each accession as recommended by the International Standards for Genebanks (Rao et al. 2006). The top of paper method and between paper method (ISTA 2005) were used in the presence of basic requirements for seed germination such as water, oxygen, light, and suitable temperature.

The seeds for field trials were soaked (pre-treatment) in a 400 mL beaker (Pyrex Iwaki TE-32 Asahi Glass; Pt. Iwaki Glass, Sumedang, Indonesia) filled with distilled water for 24 h before being sown at the experimental site to improve germination and seedling growth (Donovan 2001). The evaluation of seedlings was performed according to specific criteria listed by International Seed Testing Association (ISTA) (2005) and Association of Official Seed Analysts (AOSA) (2016). The accession names, origin, seed moisture content, and seed germination results are tabulated in Table 1.

Table 1. Accession Name, Origin, Seed Moisture Content, and Seed Germination of Nine Selected H. cannabinus Accessions

Study Site and Experimental Design

The study site was established at University Agricultural Park, Universiti Putra Malaysia, Serdang, Selangor, Malaysia (2° 58’ N latitude, 101° 39’ E longitude), with an area of 26 m × 10 m. It receives up to 2140 mm mean annual precipitation, 26 °C mean annual temperature, and 5 h to 8 h average radiance. The study was conducted from July 2017 to October 2017.

The experiment was arranged using a randomized complete block design. The randomization was assigned using a table of random numbers (Gomez and Gomez 1984) with three replications of planting block. Each block consisted of nine accessions with 10 plants for each accession. The individual plants were arranged with 30 cm planting distance, with 1 m between subblocks and 2 m between blocks. The total number of plants were 270 (10 plants x 9 accessions x 3 replications). Approximately 5 g of Nitrophoska standard formulation fertilizer (15:15:15) (EuroChem, Zug, Switzerland) was supplied manually for each plant on a monthly basis. The insecticide (Kencis with 5.5% Emulsifiable Concentration (EC) of cypermethrin active ingredient; Kenso Corporation, Petaling Jaya, Malaysia) in mixture of 13 L of water and 0.013 L of insecticide (0.01% concentration) was periodically applied twice per month to keep the plants healthy as guided by the Pesticide Control Division of the Department of Agricultural Malaysia.

Soil Sample Analysis

The soil samples were taken from the depth of 0 to 15 cm and 15 to 30 cm using the zig-zag pattern for the whole (Ackerson 2018) experimental site. The samples were air-dried for one week, ground, and sieved through a 2 mm pore size sieve. Soil particle-size analysis was conducted by the relative quantities of sand, silt, and clay using pipette method (Gee and Bauder 1986). The time and depth of the sampling are deduced using Stoke’s Law (Stokes 1849). Soil classification were then classified according to scheme developed by USDA (United States Department of Agriculture) in which the soil particles and their sizes ranges are: sand (50 to 2000 µm), silt (2 to 5 µm), and clay (< 2 µm).

Soil texture was determined by using the texture software by Teh and Rashid (2003). Soil pH was measured in soil/ water (1: 2.5) suspension. Total carbon and total nitrogen were measured by CNS analyzer machine (model: LECO TruMac CNS Analyzer). The available phosphorus (P) was determined by Bray-II method as described by Kuo (1996). The exchangeable potassium (K), calcium (Ca), and magnesium (Mg) were determined using an autoanalyzer (QuikChem, Series 8000, Lachat Instruments Inc., USA), and the concentrations of calcium (Ca) and magnesium (Mg) were also determined by atomic absorption spectrophotometer (Perkin-Elmer 5100 PC) (Hazma et al. 2015). The initial physical and chemical properties of the soil are presented in Table 2.

Table 2. Initial Soil Physical and Chemical Properties at the Experimental Site

Note: % of N = 10,000 µg/g


Morphophysiological characteristics

The data on quantitative morphological, physiological, and yield characteristics were assessed according to the Sustainable Projects Development Group of UK and the literature (Pace et al. 1998; Ahmad et al. 2001) in replicates of 30 specimens for each accession. Both quantitative characteristics (morphological and physiological) were measured monthly using specific measuring equipment, and yield characteristics were recorded after harvest (Table 3).

Laboratory sample preparation

Alkaline treatment involved immersing the fibers in an alkaline solution for a period of time. It was believed that alkaline-treated fibers provide higher tensile modulus than the untreated fibers (Li et al. 2007). In the present study, sodium hydroxide (NaOH) and sodium sulfite, anhydrous (Na2SO3) were used to remove surface impurities (lignin, wax, and oil covering the external surface of fiber cell wall), while acetic acid (CH3COOH) was used in retaining a high index of crystallinity (Mohanty et al. 2000) and giving the whitest color of fiber (Hurren et al. 2002).

Table 3. Description of the Measured Quantitative Characteristics of Nine Selected H. cannabinus Accessions

For the retting process, three replications of plant samples for each accession were prepared by cutting the lower part of H. cannabinus stem approximately 10 cm above the ground. The sample was peeled manually to separate the bast fiber and core fiber. The bast fiber was cut into 15 cm of length, in weight of 30 g per sample (Amel et al. 2013).

The aqueous solution for the retting process was prepared by mixing 7% NaOH and 0.5% Na2SO3 solution. The mixed solution was then stirred using a magnetic stirrer (IKA Magnetic Stirrers RH 2 Basic; IKA Works (Asia) Sdn. Bhd., Selangor, Malaysia). The bast fiber was then immersed into the aqueous solution using a narrow mouth conical flask (Pyrex No. 4980 250 mL Erlenmeyer Flask; Corning Inc., New York, USA). It was then soaked in a water bath (Nickel Electro NE4-22T; Thermo Fisher Scientific Inc., Goteborg, Sweden) for 60 min at 100 °C. After the retting process, the sample was neutralized by removing the sample from the aqueous solution, placing filter paper in a glass conical funnel, and rinsing with 2% CH3COOH (Ramaswamy and Boyd 1994). The unattached lignin, wax, and color were then removed by two dunks in hot plain water at 70 °C. The fiber was then submerged under plain water and dunked 10 times before it was air-dried for two weeks (Hurren et al. 2002). The sample was then used for fiber diameter measurement and tensile strength testing.

Fiber diameter determination

SEM imaging was conducted at the Laboratory of Biopolymer and Derivatives, Institute of Tropical Forestry and Forest Products (INTROP), Universiti Putra Malaysia (Selangor, Malaysia). A small part of sample retted above was taken for fiber diameter and surface morphology study. Three replications for each accession were cut into widths and lengths of 2 mm. The bundle of fiber (to increase data taken of the image instead of using single fiber) for each sample were placed into aluminium specimen mounts with slotted head 12.7 mm with tapered end pin 3.1 mm (Electron Microscopy Sciences, Hatfield, United States). The sample was vacuumed and coated with a thin layer of conductive material (gold and platinum) by using SPT-20 Ion Sputter Coater (COXEM, Daejeon, Korea) to prevent electron build up degrading the image quality. The coated sample were then inserted into a motorised chamber of 3-axis XYT sample position stage of scanning electron microscopy (SEM) (EM-30AX Plus; COXEM, Daejeon, Korea). The navigated sample was then extracted by using COXEM Nanostation operating software. The fiber diameter was then calculated as an average of three measurements of the single fiber.

Fiber tensile testing

Five grams of H. cannabinus bast fiber bundle in three replicates for each accession were used for tensile test determination. The test was conducted according to ASTM D885 (1995). The tensile test was done at the Strength Material Laboratory, Faculty of Engineering, Universiti Putra Malaysia (Selangor, Malaysia). A dual column tabletop universal testing machine (INSTRON 3365; Instron, Norwood, MA, USA) was used, with a cell load capacity of 5 kN, at a crosshead speed of 1 mm/min with a gauge length of 60 mm. The standard specifics of sample preparation was followed Amel et al. (2013). 80 gsm A4 paper size (IK Yellow; Goldtech Access Sdn Bhd, Selangor, Malaysia) was cut in similar width (7 cm) of the screw grips and in length 12 cm. A rectangle-shape-hole was made at the center of the paper in width and length of 3 cm and 6 cm (effective gauge length of 6 cm). Both end of the sample was then glued to the paper (3 mm width and 30 mm length at both end of the sample) to avoid the possibility of a sample becoming fractured or slipping in the gripped area.

The lower and upper clamp with smooth face of interchangeable jaws face (Model 2710-004 side-acting screw grips; Instron, Norwood, MA, USA) was adjusted to accommodate the length of the sample. The glued part of the sample was vertically mounted from the lower clamp (the fixed grip) to the upper clamp (the grip in charge of applying tension). The clamp was manually tightened at both knobs at sides of the grips. The tensile strength determination was focused on the gauge length, in which one grip keeps the sample in place, while the other grip pulled at a constant speed until the sample fractured. Once the sample had broken, the tensile testing has officially ended. The tensile properties were recorded as a function of the increase in gauge length (sample elongation during applying tension until fractured). A 5 mm increase in gauge length approximately yields a 20% decrease in fiber strength (Ramesh 2016). Tensile properties of the specimen were then analysed by using Instron Bluehill 3 software (Instron, Norwood, MA, USA).

Statistical analysis

All measurements were analyzed using a factorial analysis of variance (ANOVA) and mean comparison via Duncan’s Multiple Range Test (DMRT) at p ≤ 0.05 (significant difference) and p ≤ 0.01 (highly significant difference). The relationship between the morphophysiological and fiber characteristics were analyzed using bivariate (Pearson) correlation analysis. Interpretation of the correlation coefficient values was based on Ratner (2009). The growth performance of morphological characteristics namely stem height and stem diameter were also evaluated using Simple Scatterplot. The statistical analyses were conducted using IBM SPSS statistics software version 25.0 (IBM Corp., Armonk, NY, USA).


Growth Performance

Stem height and stem diameter are considered as the general guiding criteria for efficient production of fibers in a particular species (Maiti and Chakravarty 1977). Figures 1 and 2 demonstrate growth patterns of the nine H. cannabinus accessions for the four-month cultivation regarding stem height and stem diameter increment, respectively.

Regardless of accession, the stem height increments at the first month after planting (MAP) were in the range from 29.3 to 44.1 cm, and the increment at the fourth MAP were from 43.0 to 55.7 cm. The V36 accession (control), T19 (IBFC China accession), and all four BJRI Bangladesh accessions exhibited strong growth for the first three MAP and slowed down at the fourth MAP, which resulted in higher growth (Table 4) than the other (FH952, T15, and T17) IBFC China accessions, which were slowed down after the second MAP (Fig. 1a). Salih et al. (2014), however, found that plant growth of FH952 (stem height and stem diameter) was greater than the V4383 accession. Meanwhile Sultana et al. (2016) recorded shoot regeneration of HC2 taking minimum days (4.71 days) compared to HC95 accession (7. 41 days).

All nine accessions were found to produce the first bud at 10th to 15th node (at stem height 50 to 60 cm) with a minimum of stem height of 70 cm. The first flowering was formed on the 10th to 15th node (at stem height 80-90 cm) with minimum stem height was 100 cm. The vegetative growth of all accessions were continued with time.

Fig. 1. Growth performance of nine H. cannabinus accessions; a) stem height increment (cm), b) stem diameter increment (mm) of four month of planting

Stem diameter increment at the first MAP were found in the range from 3.74 to 4.88 mm, while at fourth MAP were found in range from 0.72 mm to 1.60 mm. Stem diameter increment of control accession were observed to decreased considerably only after three MAP compared to other accessions (Fig. 1b). HC95 accession resulted in higher stem diameter increment on the second MAP compared to the first MAP. T15 and V4383 accessions show the same pattern in which stem diameter increment were decreased slightly after second MAP. Meanwhile stem diameter increment of T17, T19, FH952, HC95, HC2, and V4202 accessions were decreased considerably after second MAP.

Most of the control accession plants do not produce buds even after two MAP, compared to other accession in which already produce many buds after two MAP. Ab Shukor et al. (2009) found 50% flowering of V36 accession (control) was 83 to 136 days after planting. Hossain et al. (2012) found kenaf biomass accumulation such as leaf, root, stalk, and wood dry weight was started at 6th week of growth after planting. Among the nine accessions, the control, V4383, and HC2 accession had the greatest stem diameters, stem dry weights, and total aboveground dry weights, so therefore they portrayed a maximum fiber yield. The results of V36 and HC2 accessions agreed with the findings of Hossain et al. (2014; 2016) and Nasreen et al. (2014) that the accession produced the highest fiber yield.

Morphological and yield characteristics

The results in Table 4 demonstrate all nine accessions allocated more dry matter to shoots than roots. Similar climatic-condition in present study but with different soil properties produced similar results on the control accession plant height (present study (239 cm), Ab Shukor et al. (2009) (240 cm), Omalsaad et al. (2012) (212 cm), and Khalatbari et al. (2015) (231 cm)).

Plant height was found to be significantly different (p ≤ 0.05) among nine accessions. The T17 accession recorded the highest stem heights (245.8 cm) followed by control accession (239 cm), and HC2 accession (236 cm), while HC95 accession had the lowest stem height (216 cm). Hossain et al. (2014) however recorded HC2 as the higher plant height compared to V36 (control) and HC95 accessions. Lower range of stem height in present study (208 to 246 cm) could be due to our planting distance (30 cm) between individual plants.

Results (Table 4) also indicate that there were highly significantly differences (p ≤ 0.01) in stem diameter, root dry weight, leaf dry weight, leaf area, and growth efficiency among nine accessions. The control accession showed the largest stem diameter, stem dry weight, and total aboveground dry weight (15.2 mm, 75.7 g, and 122.3 g, respectively), followed by V4383 (13.8 mm, 63.5 g, and 119.0 g, respectively) and HC2 (13.3 mm, 62.6 g, and 117 g, respectively). Meanwhile the lowest stem diameter and stem dry weight were observed for FH952. The stem diameter of the control accession in the present study (15.2 mm) was much smaller than Omalsaad et al. (2012) (8.00 cm), but almost similar to Ahmad (2011) (21.6 mm), Hossain et al. (2014) (12.0 mm), and Khalatbari et al. (2015) (14.1 mm).

Hossain et al. (2011) found that HC2 produced on average 2 g per plant higher of stem dry weight, root dry weight, and leaf dry weight compared to V36 (control) accession that was grown on 96.4% sand content of soil. HC95 meanwhile produced the lowest biomass (stem, leaf, and root dry weight) in comparison to HC2 and V36 (Hossain et al. 2011; 2016). Stem dry weight FH952 in this finding was in contrast with Qi et al. (2002), who reported that FH952 yield was 15.7% more than the standard cultivar in China. Regardless of accession, fruit dry weight was found to be in the range 15.8 g to 31.9 g. Result (Table 4) indicates that fruit dry weight was not significantly different among the nine accessions.

Table 4. ANOVA and Duncan’s Multiple Range Test of Quantitative Morphological and Yield Characteristics of Nine H. cannabinus Accessions

The highest root dry weight was found from V4383 (32.5 g), followed by HC95 (18.8 g) and HC2 (18.3 g), while the lightest was found from FH952 (13.0 g). Hossain et al. (2011) reported HC2 (11.85 g) had heavier root dry weight compared to V36 (control) (11.5 g) and HC95 (10.8 g) accessions. Ab Shukor et al. (2009) recorded that V36 (control) accession was among the heaviest root dry weight (10.7 g), in which V36 (control) in present study was heavier than reported by Ab Shukor et al. (2009) and Hossain et al. (2011). A highly significant difference (p ≤ 0.01) was found for leaves dry weight among accession. The V4202 produced the heaviest leaves dry weight (29.8 g), followed by V4383 (29.0 g) and HC2 (25.1 g), and the lightest leaves dry weight was the control accession (16.8 g). Hossain et al. (2011, 2016) reported that HC2 had higher leaves dry weight compared to V36 (control) and HC95 accessions.

The results in Table 4 also showed that leaf area and growth efficiency were highly significantly different (p ≤ 0.01) among nine accessions. The biggest average leaf area was found in V4383 (1032 cmplant-1), followed by HC2 (796 cmplant-1) and T17 (704 cmplant-1) accessions, and the lowest leaf areas was T19 accession (155 cmplant-1). Hossain et al. (2016) recorded leaf area of HC2 and V36 (control) was 900 cmplant-1. However, the highest growth efficiency was found in T19 (0.0187 g cm-2 d-1) while the lowest was T17 (0.0023 g cm-2 d-1).

Physiological characteristics

Photosynthesis is the process in which the energy from light is used to synthesize carbon compounds in foliage leaf (Pallardy 2008; Weraduwage et al. 2015). The physiological characteristics, such as photosynthesis rate (Anet), stomatal conductance (Gs), intercellular concentration of carbon dioxide (Ci), transpiration rate (EL), instantaneous water use efficiency (WUEinst), intrinsic water use efficiency (WUEi), and carboxylation efficiency (Anet/Ci), were measured in this study.

The results (Table 5) indicate that all physiological characteristics were significantly different (p ≤ 0.05) among H. cannabinus accessions, as well as the control accession. The highest Anet was found from HC2 (35.1 mol m-2 s-1) followed by T17 (35.0 mol m-2 s-1), and control (33.10 mol m-2 s-1) accessions, while the lowest rate was found from V4202 accession (27.4 mol m-2 s-1). Anet of the control accession in the present study was higher than reported by Khalatbari (2016) at 21.6 mol m-2 s-1 and Hossain et al. (2016) at 9.9 mol m-2 s-1. Hossain et al. (2016) also reported that the Anet of HC2 was 10.11 mol m-2 s-1. High Anet has been suggested to contribute to high amount of dry matter production (Table 4) (Salih et al. 2014; Hossain et al. 2016).

The highest Gs was found from HC2 (1.08 mol m-2 s-1), followed by control (1.03 mol m-2 s-1), HC95 (0.85 mol m-2 s-1), and the lowest recorded by T17 accession (0.78 mol m-2 s-1). Khalatbari et al. (2016) reported a Gs of V36 (control) of 1.00 mol m-2 s-1, which was similar to the present findings. The T17 accession meanwhile might possess better control of stomatal function in water deficit conditions. The highest Ci was found from the FH952 (339.64 mol m-2 s-1) followed by V4202 (337.03 mol m-2 s-1), V36 (334.77 mol m-2 s-1), and the lowest recorded by the V4383 (294.53 mol m-2 s-1) accession.

The highest EL was recorded by HC2 (6.26 mmol m-2 s-1), followed by V4202 (5.87 mmol m-2 s-1) and control (5.70 mmol m-2 s-1), while the lowest was from V4383 (4.54 mmol m-2 s-1). Khalatbari et al. (2016) recorded that the EL of V36 (control) was 1.88 mmol m-2 s-1, and 1.71 mmol m-2 s-1 by Tahery (2011), which was much lower compared to the present study. The highest WUEinst was found from T17 (7.61 mol mmol-1), followed by V4383 (7.50 mol mmol-1), and T15 (6.87 mol mmol-1), while the lowest was found from V4202 accession (4.88 mol mmol-1). The V4383 accession had the highest WUEi (61.9 mol mol-1), followed by T17 (50.1 mol mol-1), T15 (43.0 mol mol-1), and the lowest was HC2 accession (32.8 mol mol-1).

Table 5. ANOVA and DMRT of Physiological Characteristics of Nine H. cannabinus Accessions

Tensile strength and fiber diameter

Fiber diameter was significantly different among H. cannabinus accessions, including the control (Table 6). The largest fiber diameter (32.6 µm) on average was from the V4202 accession, followed by the T15 (30.75 µm) and T17 (25.05 µm) accessions. Khalatbari et al. (2016) reported fiber diameter of V36 (control) accession was 23.0 µm. Meanwhile H’ng et al. (2009) recorded 24.1 µm, which is similar to this finding. The results also indicated that the fiber tensile modulus of nine H. cannabinus accessions ranged from 10.8 MPa to 77.9 MPa. The three highest tensile modulus values on average were observed from the V4383 (77.9 MPa), FH952 (52.2 MPa), and HC2 (52.0 MPa) accessions, whereas the control accession had an average value of 35.2 MPa.

The three lowest tensile modulus values were observed from the HC95 (16.5 MPa), V4202 (14.9 MPa), and T15 (10.8 MPa) accessions. Both the fiber diameter and tensile modulus were significantly different (p ≤ 0.05) among the accessions. The V4383 accession had a smallest fiber diameter (10.9 µm) with the highest tensile modulus (77.9 MPa). Meanwhile, the V4202 accession had the biggest fiber diameter (32.6 µm) with low tensile modulus (14.9 MPa).

Table 6. Fiber Tensile Strength and Diameter of Nine H. cannabinus Accessions

Correlation Analysis

Morphological characteristics

A highly significant (p ≤ 0.01) strong positive relationship (r = 0.790) was found between stem height and the stem diameter, a moderate positive relationship with stem dry weight (r = 0.556), total aboveground dry weight (r = 0.590), fruit dry weight (r = 0.332), root dry weight (r = 0.362), and a weak relationship with leaf dry weight (r = 0.183) (Table 7). A significant (p ≤ 0.05) but weak relationship (r = 0.152) was observed between plant height and leaf area. This study found a significant relationship between stem diameters and all morphological and yield characteristics except for leaf dry weight and growth efficiency. Meanwhile there was no significant relationship found between stem height and growth efficiency. Among all morphological characteristics, the strongest (p ≤ 0.01; r = 0.848) relationship was found between stem dry weight and total aboveground dry weight.

Stem dry weight was found significantly related with fruit dry weight, root dry weight, total aboveground dry weight, leaf area, but negatively related with growth efficiency. Fruit dry weight meanwhile was significantly associated only with leaf dry weight (p ≤ 0.01; r = 0.217) and total aboveground dry weight (p ≤ 0.01; r = 0.682). Root dry weight and leaf dry weight increases when total aboveground biomass increased (p ≤ 0.01; r = 0.205, 0.373 respectively). Significant positive but weak relationships were found between leaf area with stem height (p ≤ 0.05; r = 0.152), stem diameter (p ≤ 0.01; r = 0.273), stem dry weight (p ≤ 0.01; r = 0.218), and total aboveground biomass (p ≤ 0.01; r = 0.205) depicted increase in leaf area resulted to slightly increases of all above characteristics. Moreover, increasing of leaf area affect to decrease in H. cannabinus growth efficiency (p ≤ 0.01; r= -0.586). The growth efficiency meanwhile decreased (p ≤ 0.05; r = -0.158) with increasing of stem diameter but increased (r = 0.160) with increasing of stem dry weight.

Table 7. Combined Analysis for Correlation Coefficient of Quantitative Morphological Characteristics of Nine H. cannabinus Accessions

Physiological characteristics

Increasing of Anet values resulted from the increase of Gs (p ≤ 0.01; r = 0.619), thus increasing EL (p ≤ 0.01; r = 0.337), WUEinst (p ≤ 0.01; r = 0.518), and carboxylation efficiency (p ≤ 0.01; r = 0.937). The Ci (p ≤ 0.05; r = -0.144) and WUEi (p ≤ 0.01; r = -2.75) however decreased with increasing of Anet. The Gs was found to have a highly significant (p ≤ 0.01) strong positive relationship (r = 0.728) with EL, and moderate relationship with Ci and carboxylation efficiency.

Table 8. Combined Analysis for Correlation Coefficient of Physiological Characteristics of Nine H. cannabinus Accessions

The increasing of Gs depicted decreases in WUEinst (p ≤ 0.01; r = -0.193) and WUEi (p ≤ 0.01; r = -0.683) (Table 8). An increase of Ci meanwhile depicted an increase in EL of the plant (p ≤ 0.01; r = 0.628) but a decrease in WUE and carboxylation efficiency. The WUEinst increase with increasing of WUEi (p ≤ 0.01; r = 0.567) and carboxylation efficiency (p ≤ 0.01; r = 0.739) which demonstrated that the increasing of Gs resulted in increases of EL and Ci, but a decrease in WUE.

Tensile strength and fiber diameter

Large fiber diameter resulted in a low tensile modulus of fiber (Fig. 2). Inacio et al. (2010) reported that the smaller the diameter, the stronger is the fiber for several natural fibers such as cotton (Gossypium hirsutum), curaua (Ananas erectifolius), jute (Corchorus capsularis), and sisal (Agave sisalana). Weibull analysis of sisal fiber tensile strength conducted by Inacio et al. (2010) also shows an inverse dependence of the tensile strength with the diameter.

In the present study, Fig. 2 showed an inverse correlation between fiber strength and fiber diameter. A fractographic analysis by SEM of the H. cannabinus bast fiber surface in Fig. 3 depicted the thicker fiber with a diameter of equal to 26.1 µm and above (Table 6). The figure reveals a heterogeneous fracture and covered by a number of impurities (Fig. 3c, 3d, and 3h) suspected to be hemicellulose, lignin, pectin, and waxy substances associated with relatively more fibrils (Tahir et al. 2015). In other words, the fiber tensile modulus less than 20 MPa had voids, impurities, and showed fiber damage on the surface (Fig. 3c, 3d, 3g, and 3h).

Fig. 2. Correlation of fiber diameter and tensile modulus of H. cannabinus accessions