researc half-sib matting and genetic analysis of agronomic ... · tion of additive genetic effects...

crop science, vol. 55, january–february 2015 www.crops.org 123

ReseaRch

Sainfoin is a species of the Fabaceae family that has been planted in western Asia, southern Europe and North Amer-

ica for many years. It is a perennial cross-pollinated plant that is almost free from serious pest and disease problems compared with other legumes. It is non–bloat-inducing, highly nutritive and palatable. Sainfoin is compatible with poor soils and dry regions with the ability to improve soil fertility by fixing the atmospheric nitrogen (Delgado et al., 2008; Frame, 2005; Goplen et al., 1991).

Breeding goals for sainfoin mainly focus on forage yield improvement, tolerance to abiotic and biotic stresses, and nutri-tive value enhancement. Most of these characters show continuous variation in heterogeneous populations of forage plants and are influenced by environmental factors (Bukvic al., 2008). Therefore, information on genetic variance, heritability, and gene action of

Half-Sib Matting and Genetic Analysis of Agronomic, Morphological, and Physiological Traits in Sainfoin under Nonstressed versus

Water-Deficit ConditionsSayareh Irani, Mohammad Mahdi Majidi,* and Aghafakhr Mirlohi

ABSTRACTHalf-sib matting, including polycross and open-pollination, are powerful methods to obtain quan-titative genetic information in breeding of forage plants. In sainfoin (Onobrychis viciifolia Scop).no prior efforts have been made to estimate the genetic parameters of traits and comparing these methods. Thirty half-sib families (16 polycross and 14 open-pollinated families) were evaluated under two moisture environments (nonstressed and water deficit) during 2010 to 2012. Water defi-cit decreased plant height (20.3%), plant density (51.8%), dry matter yield (45–53%), relative water content (rWC) (18.9%) and increased carot-enoid content (23.9%), proline content (65.7%), specific catalase (74.4%) and ascorbate peroxi-dase activity (52.7%). High genotypic variation was observed for most of the measured traits. Narrow-sense heritability (h2

pFM) estimates for agro-morphological traits ranged from 0.42 (leaf to stem ratio) to 0.95 (flowering date). The high-est and lowest h2

pFM among physiological traits were estimated for carotenoid content (0.50) and rWC (0.83), respectively. Moderate to high heritability for most traits showed the contribu-tion of additive genetic effects suggesting phe-notypic selection could be successful. The best agromorphological and physiological traits (such as plant height, plant density, dry matter yield, and rWC) under water deficit belonged to fami-lies 16 and 27. Dry matter yield had significant and positive correlation with plant height and plant density indicating indirect selection would also be effective to improve forage yield. prin-cipal component analysis separated polycross and open-pollinated families into distinct groups. open-pollinated families had higher mean val-ues and general combining ability for most of the traits while polycross families had higher herita-bility and genetic gain from selection. This may indicate a more effective selection and a higher genetic advance in polycross families.

Dep. of Agronomy and Plant Breeding, College of Agriculture, Isfa-han Univ. of Technology, Isfahan, 84156-8311, Iran. Received 24 Mar. 2013.*Corresponding author ([email protected]).

Abbreviations: ΔG, Genetic gain per cycle; APX, specific ascorbate per-oxidase activity; CAT, specific catalase activity; CC, carotenoid content; Chl a, chlorophyll a content; Chl a + b, total chlorophyll content; DMY, dry matter yield; ETc, crop evapotranspiration; FD, flowering date; GCA, general combining ability; GCV, genotypic coefficient of variation; HS, half-sib; h2

PFM, narrow sense heritability; IL, inflorescence length; LSR, leaf to stem dry weight ratio; MAD, maximum allowable depletion; NFS, number of flowering stems per plant; PC1, first principal component; PC2, second principal component; PCA, principal component analysis; PCV, phenotypic coefficient of variation; PD, plant density; PH, plant height; ProC, free proline content; RWC, relative water content, SPM, sensitivity to powdery mildew; TAW, total available soil water.

Published in Crop Sci. 55:123–135 (2015). doi: 10.2135/cropsci2014.03.0235 © Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Published August 28, 2015

124 www.crops.org crop science, vol. 55, january–february 2015

different traits are important for population improvement and creation of new varieties (Bowley and Christie, 1981). Half-sib (HS) families, including polycross, topcross, and open pollination are frequently used in forage plants breeding to estimate narrow-sense heritability and additive genetic variance (Nguyen and Sleper, 1983a). Traits under complex genetic control, such as forage yield, have low to moderate heritability while others, such as plant height, have high heritability (Amini et al., 2013; Araghi et al., 2014; Majidi et al., 2009). De Araujo et al. (2002) report that there is significant variation and moderate heritabil-ity for most traits in meadow bromegrass (Bromus riparius Rehm.) populations. They showed significant differences in dry matter yield between the parental populations and progenies obtained by different pollination methods (poly-cross, open pollination, and selfing) and reported that open-pollinated progeny test is the best method of choice for selecting parents for synthetics of meadow brome-grass. In addition, Milic et al. (2010) compared differences between progeny tests (polycross, open pollination, and selfing) and reported that open-pollinated progenies had a higher mean of green forage yield and number of shoots per plant than progenies from selfing.

Drought stress is the most frequent limiting factor for plant productivity and alters morphological, physiologi-cal, biochemical, and metabolic aspects of plants. There-fore, genetic improvement of plants for drought stress has become increasingly important (Boyer, 1982; Lawlor and Cornic, 2002; Smirnoff, 1998;). Plants usually use differ-ent stress defense mechanisms such as developing root sys-tems, changes in the photosynthetic pigments, antioxidant levels, and accumulation of compatible osmolytes, such as proline. The effects of drought stress on chlorophyll and carotenoid content have been investigated in many plant species (Antolin and Sanchez-Diaz, 1993; Keyvan, 2010). Many researchers have shown that drought-resistant gen-otypes have higher RWC than drought-sensitive ones and RWC is widely used as an index to identify resistant geno-types under water stress (Keyvan, 2010; Xu et al., 2010). Higher RWC under water stress could be related to the higher capacity of resistant genotypes to absorb soil water or prevent water loss through stomata (Keyvan, 2010). Proline, which increases faster than other amino acids in plants under water stress, has been suggested as a param-eter for selecting drought-resistant varieties (Bukvic et al., 2008; Keyvan, 2010). Several studies have shown that the activity of antioxidant enzymes and osmotic adjust-ment ability are correlated with plant tolerance to drought (reviewed in Miller et al.,2010 and Xu et al., 2010).

In sainfoin, knowledge about genetic analysis of agro-morphological and physiological traits and their associa-tion with forage production under water deficit is lim-ited. The objectives of this study were to (i) assess genetic diversity and estimate narrow-sense heritability, general

combining ability, and expected genetic gain for forage yield and morphological and physiological traits, (ii) com-pare polycross and open-pollinated HS families in estima-tion of genetic parameters, and (iii) asses the association of morphological and physiological traits in HS families of sainfoin under nonstressed and water-deficit conditions.

MATERIAl ANd METHodSExperimental SiteThe experiment was conducted on a Typic Haplargid, silty clay loam soil at Isfahan University of Technology Research Farm, 32° 30¢ N, 51° 20¢ E, Isfahan, Iran during 2008 to 2012. The soil was calcareous, containing 390 g kg-1 Ca-carbonate equivalent, 5.0 g kg-1 organic C, and 0.77 g kg-1 total N, with pH 8.3. The soil was nonsaline and nonsodic. The electrical conductivity and the sodium adsorption ratio of the soilsaturated extract were 1.6 dS m-1 and 1.4 (mmoll–1) 0.5, respectively. The mean annual tem-perature and precipitation were 14.5°C and 140 mm. Plots were fertilized with 200 kg N ha-1 and 200 kg P ha-1 before planting and 100 kg N ha-1 was applied to the trial each September.

Plant MaterialThirty HS families of sainfoin (16 derived from polycross and 14 derived from open pollination) were constructed and used in this study. To produce the polycross HS families, 16 paren-tal genotypes were taken from a large nursery (800 genotypes) established in 2007, which mainly consisted of natural eco-types of sainfoin from wide geographical areas of Iran. The 16 parental genotypes were transferred to an isolated polycross field nursery in 2008. The isolated polycross field nursery was a balanced 4 by 4 lattice square design with six replications and honeybee colonies in vicinity to facilitate pollinations. Seeds of 16 polycross families were harvested separately from each geno-type during 2008 and 2009. To produce the open-pollinated HS families, 14 genotypes were randomly selected and marked in the same large nursery (800 genotypes) that was explained above. The 14 marked genotypes were open pollinated in the large nursery during 2008 and 2009 and the seeds were har-vested. Seeds of each genotype for the two harvest years were bulked to have enough seed for the field experiments. Informa-tion on the parental genotypes is shown in Table 1.

Evaluation of Half-Sib FamiliesSeeds from the 30 HS families were planted in the field under nonstressed and water-deficit conditions in a randomized complete block design with three replications in March 2010. Seeding rate was 30 kg h-1. Each plot contained four rows of each HS family, 40 cm apart.

Total available soil water (TAW) is defined as amount of the soil water in the root zone between field capacity and the permanent wilting point and was calculated on the basis of the Allen et al. (1998) formula:

( )FC WP1

BdTAW 100

Bdi i i im

iw

W W D=

æ ö- ´ ´ ÷ç ÷= ç ÷ç ÷çè ø


of irrigation to prevent water stress. At low rates of ETc (≤5 mm d-1), the value of r was recommended to be 0.55 for alfalfa (Medicago sativa L.) and was used for sainfoin in this study. The fraction r is a function of the evaporation power of the atmo-sphere as follows:

( )rec c 0.04 5 ETr= r + -

where rrec is the recommended value for alfalfa and r is the adjusted value for evaporation power of the atmosphere. Irri-gation treatments were scheduled on the maximum allowable depletion (MAD) percentage of the available soil water (TAW). Irrigation was done when 40% and 80% MAD of the TAW was depleted from the root zone at the nonstressed and water-deficit environments, respectively. A TDR probe (TDR Trase System, Model 6050XI; Soil Moisture, Santa Barbara, CA) was used to measure soil water content 2 d after irrigation and continued up to 1 d before the next irrigation. The plots were irrigated when the respective MAD threshold values were reached for each treatment and were calculated using the following equation:

( )Irrig FCi FCi WPi W W W W f= - -

where WIrrig is the soil water content before irrigation by weight (%) and r is the fraction of TAW (40 and 80%) that can be depleted from the root zone. The irrigation depth (In) with the goal of increasing the water content in the root zone depth to field capacity was calculated as (Keller and Bliesner, 1990):

( )FC Irrig

1

Bd 1 00

Bd

mi i i

ni w

W W DI

=

æ ö- ´ ´ ÷ç ÷ç= ÷ç ÷ç ÷÷çè øå

100 n

ga

II

E

´=

where Ig is the gross depth of irrigation (mm) and Ea is the irriga-tion efficiency (%) assumed as 65% during the growing season (Tafteh and Sepaskhah, 2012). In interval periods, all experi-mental plots were irrigated when 40% of TAW was depleted. No data was recorded for plant establishment in the first year (2010). Water deficit was applied during the growing season from 1 May to 1 October in 2011 and 2012 and data were recorded. Dry matter yield at four cuts plus eight physiological traits at the first and second cut were measured during 2011 and 2012. These included flowering date (FD), plant height (PH), inflorescence length (IL), plant density (PD), number of flowering stems per plant (NFS), number of nodes per shoot, sensitivity to powdery mildew (SPM), and leaf to stem ratio (LSR) were measured and analyzed on the basis of average of four cuts. Days to 50% flowering were recorded as the number of days from 21 March until appearance of 50% of inflorescence in each plot for first cut and number of days from previous cut to appearance of 50% of inflorescence in each plot for the next cuts. Inflorescence length was measured on the basis of average length from first node to the tip of randomly chosen inflorescence of each plant. Plant height was measured as the distance from the plant base to the top of the highest inflorescence. Sensitivity to powdery mildew

where WFCi is the gravimetric soil-water content (%) at field capacity, WWPi the gravimetric soil-water content (%) at the per-manent wilting point, Bdi the bulk density of the soil (g cm-3), DI the soil layer depth (mm), Bdw the bulk density of the water (g cm-3), i the layer counter, and m the number of layers in the root zone. The fraction TAW that a crop can readily extract at its potential rate from the root zone without suffering water stress is referred to as the readily available soil water (RAW) and was calculated following the Allen et al. (1998) formula:

RAW TAW= r ´

Root zone depth was measured four times during the growing season and TAW was adjusted accordingly. The r factor changes for different crops from 0.3 for shallow-rooted plants at high rates of crop evapotranspiration, ETc (>8 mm d-1) to 0.7 for deep rooted plants at low rates of ETc (< 3 mm d-1) (Allen et al., 1998). The factor r was used to estimate the required time

Table 1. Origins of the sainfoin’s genotypes used in the study.

Genotypes Origin (city, province, country)

Polycross

1 Borujen, chaharmahal and Bakhtiari, iran

2 Golpayegan, isfahan, iran

3 najafabad, isfahan, iran

4 Arak, Markazi, iran

5 Semirom, isfahan, iran

6 Khansar, isfahan, iran


8 Khomeyn, Markazi, iran

9 Sanandaj, Kurdistan, iran


11 Fereydunshahr, isfahan, iran

12 Unknown


14 Karaj, Alborz, iran

15 Kerman, iran

16 Bardsir, Kerman, iran

Open pollinated

17 Khomeyn, Markazi, iran

18 Semirom, isfahan, iran



21 Fereydunshahr, isfahan, iran

22 Sirjan, Kerman, iran

23 najafabad, isfahan, iran

24 Arak, Markazi, iran

25 Kabotarabad, isfahan, iran

26 isfahan, iran

27 Borujerd, Lorestan, iran

28 Sanandaj, Kurdistan, iran

29 Baft, Kerman, iran

30 Borujen, chaharmahal and Bakhtiari, iran


was measured on the basis of the leaf area covered with mildew using 0 to 9 scaling. For each plot, 20 plants were randomly selected for trait measurement. All of the traits were measured at the 50% flowering stage (for each family separately) and then forage was harvested by hand cutting of plants at approximately 5 cm above the ground. To determine dry matter yield (DMY), fresh samples were dried at 72°C for 48h. Dry matter yield was separately calculated in four cuts.

Physiological traits were determined on the basis of the first two cuts of each year. Chlorophyll a, b, a+b, and carot-enoid content were measured according to the method of Lichtenthaler and Wellburn (1983). For each sample, 0.33 g of fresh leaf tissue was ground in 10 mL of acetone (80%) until a homogenized mixture was obtained. The absorption of leaf extract solution was read at 663 and 646 nm (for chlorophyll a and b) and at 470 nm (for carotenoid) by HITACHI U1800 spectrophotometer. Leaf RWC was estimated using the method of Turner (1981) as below:

( )( )fresh weight dry weight /

RWC 100urgid weight dry weightt

é ùê úê úê û- úë

-=

For this, leaf fresh weights were measured immediately after cutting. For determining leaf turgid weights, leaves were floated in distilled water in test tubes for 16 to 18 h at room temperature. Dry weights were measured after drying the leaf samples in oven for 48 h at 70°C.

Free proline content in the leaves was measured follow-ing Bates et al. (1973) method. For this purpose, 0.5 g of fresh leaves was ground in 10 mL of 3% aqueous sulphosalycylic acid and the extract was filtered. Two mL of the extract was added in the test tube with 2 mL of ninhydrin reagent and 2 mL of glacial acetic acid. The reaction blend was boiled in water bath at 100°C for 1 h. After cooling the mixture on ice, 4 mL of toluene was added and thoroughly mixed. The toluene phase was separated and its absorbance measured at 520 nm using a HITACHI U1800 spectrophotometer against toluene blank. For enzyme measurements, 0.5 g of leaf was extracted with extraction buffer containing 1% polyvinylpyrrolidone and 0.5% Triton X-100 in 100 mM potassium phosphate buffer (pH = 7). After centrifuging at 15,000 rpm for 20 min at 4°C, the supernatant was used to assay antioxidant enzymes following Nakano and Asada (1981) and Aebi (1983) methods for ascor-bate peroxidase (APX) and catalase (CAT) activity at 265 and 240 nm, respectively.

Genetic and Statistical AnalysisAnalysis of Variance, Heritability and Genetic GainAnalysis of variance was done to determine differences among the families and to estimate variance components, using the general linear model (GLM) with moisture environment as fixed effect and family and year as random effects in SAS soft-ware (SAS Institute, 1999). Variance components and expected mean squares for all traits were computed according to Nguyen and Sleper (1983a). Mean separation was done using LSD test at p-values < 0.05 (Steel and Torrie, 1960).

With noninbred parents from a random mating popu-lation, the genetic variance among HS families would be

( )2 2AFCov HS 1/ 4Fs = = s , assuming there is no additive by addi-

tive type of epistatic variance. Correspondingly, the family × moisture environment, family × year, and family × moisture envi-ronment × year interaction variances would be 2 2

FE AFE 1/ 4s = s , 2 2FY AFY 1/ 4s = s , and 2 2

FEY AFEY1/ 4s = s , respectively. Narrow-sense heritability was calculated on the basis of family means according to the following formula (Nguyen and Sleper, 1983a):

22PFM 22 2 2 2

2 FE FY FEY

ey rl rey

F

yF

h

e y

s=

ss s s ss + + + + +

where 2sF , 2FEs , 2

FYs , 2FEYs ,

2ys , and 2

s represent the family, family × moisture environment, family × year, family × mois-ture environment × year, year, and error variance, respectively. Also e, y and r represent the number of moisture environments, years and replications, respectively. Genetic gain per cycle (ΔG) of phenotypic family selection was predicted following Nguyen and Sleper (1983a):

2

22 2 2 22 FE FY FEY

G ck

ey rl rey

F

yF e y

sD =

ss s s ss + + + + +

where c and k are parental control factor (c = 1/2) and standard-ized selection differential (k = 1.4 for 20% selection), respectively (Nguyen and Sleper, 1983a).

General Combining Ability, Genetic Coefficient of Variation, and Association of TraitsGeneral combining ability was calculated as the deviation of each HS family from total families mean as defined by Wricke and Weber (1986). The genotypic coefficient of variation (GCV) was calculated as:

( )GCV / 100g= s

where sg is the standard deviation of the genotypic variance. The phenotypic correlation between two traits was calculated as described by Falconer and Mackay (1996).

( )rp /xy x yS S S= ´

where Sxy is the phenotypic covariance for the characters x and y and Sx and Sy are the standard deviations for the traits x and y, respectively.

Principal Component AnalysisPrincipal components analysis is a multivariate method for data reduction which is able to remove interrelationships among com-ponents and is useful in grouping genotypes. Data for each family and moisture environment were averaged across the replications and years and then used for principal component analysis (PCA). The data were analyzed by Statgraphics statistical software (Stat-graphics, 2007) and were drawn by SigmaPlot 12.5.


The lowest narrow sense heritability (h2PFM) was observed

for LSR (0.42), CC (0.50), SPM (0.60), and DMY at cut 2 (0.60), respectively. The highest h2

PFM were obtained for FD (0.95) and PH (0.92). Among the physiological traits, RWC had the highest heritability (0.83). Dry matter yield (11–14%) and PD (11.2%) had high values of genetic gain from selection (ΔG) (Table 5). Comparing the yield at the four cuts, DMY at cut 1 had the highest ΔG (Table 5). The lowest value of ΔG was observed for LSR (2.0) and CC (2.4) (Table 5). The greatest expected genetic gains were found for APX (22.11%), ProC (18.6%), and CAT (17.34%).

A broad range of general combining ability (GCA) was observed for most of the traits evaluated (Tables 6 and 7). The highest and lowest GCA for PH, PD, DMY, and RWC was observed for family16 and 4, respectively. Family 23 had the maximum GCA for FD. The highest GCA for LSR was observed for family 4 and the lowest one belonged to family 21. The highest GCA for chl a+b, CC, ProC, CAT, and APX were observed for families 13, 4, 10, 14, and 16, respectively (Tables 6 and 7).

Comparison of Polycross and open-Pollinated FamiliesSignificant differences were observed between polycros-sand open-pollinated families for all traits expect NFS, chlorophyll a content (Chl a), and total chlorophyll content (Chl a+b) (Tables 2 and 3). The averages of measured traits in open-pollinated families were higher than the ones in polycross families except for IL, LSR, CC, ProC, CAT, and APX. Open-pollinated families showed the highest amount of FD, PH, PD, DMY, and RWC under both

RESUlTSAnalysis of variance and Effects of Water deficitEffects of moisture environment, year, and HS family were significant for most of the measured traits at p< 0.01 (Tables 2 and 3). Values were higher in the second year when compared with the first year measurements (Table 4). There was significant difference between the two moisture environments for all of the measured traits and the interac-tions of family × moisture environment were significant (Tables 2 and 3). With the exception of carotenoid content (CC), free proline content (ProC), CAT, and APX, all the other traits at the nonstressed moisture environment had higher values than water-deficit environment (Tables 4).It was observed that water deficit reduced DMY in all four cuts. When compared with nonstressed condition, average reduction in DMY under water deficit was 45.96, 52.32, 47.55, and 53.12% for cut one through four, respectively. However, the increases in CC, ProC, CAT, and APX at water-deficit condition were 23.90, 65.40, 74.38, and 52.75%, respectively as compared with the nonstressed condition. Water deficit decreased FD when compared with nonstressed condition (almost 17%).

Genetic Coefficient of variation, Heritability, and General Combining AbilityEstimates of genetic coefficient of variation, heritability, and genetic gain from selection are presented in Table 5. The GCV varied from 4.5 for LSR to 36.5 for APX. For most of the measured traits, the estimates of family vari-ance composed the highest part of phenotypic variances.

Table 2. Analysis of variance for agromorphological traits in 30 half-sib families of sainfoin evaluated at two moisture environ-ments in 2011 and 2012.

Source of variation† df

Mean square‡

FD PH IL PD SPM LSR DMY 1 DMY 2 DMY 3 DMY 4

e 1 3702.52** 18838.88** 373.82** 215809.5** 1427.18** 4.016** 30897602.9** 11828789** 7275362.5** 4722548.8**

R(e) 4 183.83 2523.74 9.26 4583.17 135.44 0.1734 10458.5 11560.9 255796.9 107982.8

F 29 65.19** 512.67** 4.53** 2516.50** 9.41** 0.0563** 972858.1** 221314.5** 166291.4** 82009.1**

P 15 63.31** 583.87** 2.85** 2598.66** 8.32** 0.0461** 664699.7** 159619.4** 140857.1** 67502.6**

OP 13 26.25** 346.58** 5.05** 1048.79** 6.79** 0.0345** 377677.9** 99147.1** 74650.3** 35687.6**

P vs. OP 1 599.56** 1603.70** 22.928** 20364.08** 59.67** 0.3695** 13332575.6** 2734916.8** 1739139.1** 901783.9**

F × e 29 2.86** 38.95 0.73** 547.61 3.66 0.0326** 291231.2 87133.0 51209.7** 28729.9**

F × R/(e) 116 0.26 33.73 0.14 438.73 2.29 0.0017 217291.7 68600.9 12782.6 6809.5

Y 1 359.70 4629.73 7.61 2082.48 25.39** 0.4092 14343283.1** 2100148.3** 1284117.9** 738681.5**

Y×e 1 405.38 1252.07 45.30** 4538.85 0.29 0.1032 5555984.0** 1634184.9** 985047.9** 754785.4**

Y×R(e) 4 253.58 398.29 1.21 775.38 0.42 0.2572 100609.8 46760.3 109720.8 29749.5

F × Y 29 0.09** 2.480** 0.11** 11.44** 0.13** 0.0030** 50427.3* 3535.0 2207.9** 1649.9*

F × e × Y 29 0.08* 1.29* 0.02** 5.55 0.05 0.0031** 43922.8* 3496.1 2076.2* 1324.0

error 116 0.04 0.72 0.01 3.91 0.04 0.0011 27121.8 2860.7 1164.7 996.9

* Significant at the 0.05 probability level.

** Significant at the 0.01 probability level.† e, moisture environment; F, half-sib families; OP, open-pollinated families; P, polycross families; R, replication; Y, year.‡ DMY 1, 2, 3, and 4, dry matter yield at cut 1, 2, 3, and 4; FD, flowering date; PD, plant density; PH, plant height; iL, inflorescence length; LSR, leaf to stem ratio; SPM, sensitivity to powdery mildew.


moisture environments, although they were more sensi-tive to powdery mildew than polycross families. Herita-bility estimates were higher in polycross families for most of the agromorphological traits, including FD, PH, PD, SDM, and DMY (Table 5). Open-pollinated families had slightly higher heritability values for physiological traits when compared withpolycross families (Table 5).

Principal Component Analysis and Relationship between TraitsIn both moisture environments, principal component anal-ysis was done on the basis of the agronomic and physiologi-cal traits averaged for 2 yr (Fig. 1–3). The vectors in the diagram (Fig. 1 and 2) show different traits, including FD, PH, PD, LSR, DMY (means of four cuts), chlorophyll a+b,

Table 3. Analysis of variance for physiological traits in 30 half-sib families of sainfoin evaluated at two moisture environments in 2011 and 2012.

Source of variation† df

Mean square‡

Chl a Chl b Chl a + b CC ProC RWC CAT APX

e 1 1.4834** 0.1043** 2.3747** 0.5011** 15547.15** 16524.72** 166.39** 1337.80**

R(e) 4 0.2470 0.1114 0.6733 0.0296 52.48 1120.37 0.67 14.88

F 29 0.0976** 0.0459** 0.2468** 0.0073** 297.73** 510.34** 2.50** 61.56**

P 15 0.1198** 0.0562** 0.3071** 0.0064** 40.21** 556.74** 0.61** 16.91**

OP 13 0.0785** 0.0290** 0.1919** 0.0071** 66.34** 395.58** 0.75** 11.37**

P vs. OP 1 0.0094 0.1108** 0.0555 0.0208** 7168.53** 1306.26** 53.58** 1383.58**

F × e 29 0.0167 0.0093 0.0435 0.0036** 80.04** 82.73** 0.92 14.64

F × R/(e) 116 0.0119 0.0078 0.0324 0.00001 33.49 40.61 0.62 10.07

Y 1 0.1641 0.0927 0.5037 0.4263 2333.64** 152.74 10.45** 255.07**

Y×e 1 0.5355 0.0559 0.9377 0.0010 1145.19** 251.71 5.98** 103.74**

Y×R(e) 4 0.1627 0.0683 0.4060 0.0494 4.33 350.04 0.008 0.29

F × Y 29 0.0005** 0.0002 0.0012** 0.000008** 7.26** 8.22** 0.045** 1.16**

F × e × Y 29 0.0004 0.00016 0.0008* 0.000003** 4.70** 7.05** 0.024 0.46

error 116 0.0002 0.00015 0.0005 0.000001 2.29 3.31 0.017 0.37

Significant at the 0.05 probability level.

** Significant at the 0.01 probability level.† e, moisture environment; F, half-sib families; OP, open-pollinated families; P, polycross families; R, replication; Y, year.‡ APX, specific ascorbate peroxidase activity; cAT, specific catalase activity; cc, carotenoid content; chl a, chlorophyll a content; chl a + b, total chlorophyll content;Proc, free proline content; RWc, relative water content.

Table 4. Effect of moisture stress and year on flowering date (FD), plant height (PH), inflorescence length (IL), plant density (PD), number of nodes per shoot (NNS), number of flowering stems per plant (NFS), sensitivity to powdery mildew (SPM), leaf to stem ratio (LSR), dry matter yield at cut 1, 2, 3, and 4 (DMY 1, 2, 3, and 4), chlorophyll a content (Chla), chlorophyll b content (Chlb), total chlorophyll content (Chla+b), carotenoid content (CC), free proline content (ProC), relative water content (RWC), specific catalase activity (CAT), and specific ascorbate peroxidase activity (APX) in sainfoin families.

Traits

FD PH EL PD NNS NFS SDM LSR DMY 1 DMY 2

d cm cm ———— g m-2 ————

Moisture environment

nonstressed 37.10a† 71.31a 8.44a 94.56a 5.30a 11.71a 7.40a 1.10a 1274.82a 692.78a

Water deficit 30.68b 56.84a 6.41b 45.59b 4.38b 8.12b 3.417b 0.89b 688.89b 330.25b

Year

2011 34.89a 60.49b 7.57a 67.67a 4.85a 9.38a 5.14b 1.03a 782.25b 435.14b

2012 32.89a 67.66a 7.28a 72.48a 4.82a 10.45a 5.67a 0.96a 1181.46a 587.89a

Traits

DMY 3 DMY 4 Chl a Chl b Chl a+b CC ProC RWC CAT APX

———— g m-2 ———— ——————— mg g-1 fresh leaf ——————— μmoles g-1 fresh leaf

% Unit mg-1 protein

Moisture environment

nonstressed 597.86a 431.16a 0.81a 0.34a 1.16a 0.31b 6.95b 71.36a 0.46b 3.45b

Water deficit 313.54b 202.09b 0.68a 0.31a 0.99a 0.39a 20.09a 57.81b 1.82a 7.30a

Year

2011 395.96b† 271.32b 0.73a 0.31a 1.04a 0.32b 10.97b 63.94a 0.97b 4.54b

2012 515.42a 361.92a 0.77a 0.34a 1.11a 0.39a 16.07a 65.24a 1.32a 6.22a† Within columns, means followed by the same letter are not significantly different according to LSD (0.05).


carotenoid content, ProC, RWC, CAT, and APX activ-ity. The first two components justified 39.32 and 27.07% of total variance at nonstressed environment and 47.62 and 19.02% at water-deficit environment, respectively. Princi-pal component analysis showed that the first principal com-ponent (PC1) was related to PD and DMY. The second principal component (PC2) had positive correlation with PH and RWC and negative correlation with ProC, CAT, and APX activity (Fig. 1 and 2). In addition, the first three principal components justified 48.53, 20.47, and 15.77% of total variance at both moisture environments (Fig. 3). The three-dimensional plot of PCA clustered polycross and open-pollinated families in two distinct groups. Correlation

coefficients were calculated separately in two moisture environments on the basis of the means of 2-yr data (data not shown). The DMY at four cuts were positively and significantly associated with PH and PD in both moisture environments. Also, RWC was positively correlated with PH and PD at both moisture environments but this correla-tion was stronger under water deficit environment.

dISCUSSIoNIn breeding programs, genetic improvement of genotypes depends on the availability of genetic variations (Tabas-sum et al., 2007). In this study, the high variation between HS families observed for most of the traits suggests a high

Table 5. Estimates of narrow-sense heritability (h2PFM), phenotypic and genotypic coefficient of variation (PCV and GCV), and

genetic gain from selection (ΔG) in different traits for 30 half-sib (HS) families of sainfoin evaluated at two moisture environ-ments and 2 yr.

Traits†

FD PH PD LSR SPM DMY 1 DMY 2 DMY 3 DMY 4

Polycross families

h2PFM 0.96 0.94 0.84 0.26 0.78 0.72 0.71 0.66 0.77

GcV, % 6.91 10.89 21.49 3.13 11.43 23.93 24.84 21.82 24.46

ΔG 1.55 4.59 8.73 0.01 0.41 127.72 118.00 53.44 59.04

ΔG, % 4.76 7.39 13.85 1.13 7.11 14.29 14.71 12.42 15.11

Open-pollinated families

h2PFM 0.85 0.89 0.72 0.33 0.57 0.55 0.44 0.69 0.66

GcV, % 3.88 7.65 10.15 3.19 12.52 11.08 10.06 12.39 12.03

ΔG 0.88 3.36 4.71 0.01 0.33 68.39 28.52 38.30 25.47

ΔG, % 2.51 5.06 6.03 1.30 6.63 5.75 4.71 7.22 6.88

All HS families

h2PFM 0.95 0.92 0.78 0.42 0.60 0.70 0.60 0.69 0.64

PcV, % 6.87 10.2 20.66 6.87 16.37 28.99 26.54 25.83 26.10

GcV, % 6.72 9.79 18.25 4.46 12.69 24.15 20.66 21.47 20.98

ΔG 1.55 4.21 7.90 0.02 0.37 138.31 57.61 56.96 37.36

ΔG, % 4.60 6.58 11.28 2.02 6.89 14.08 11.26 12.49 11.80

Traits

Chl a Chl b Chl a+b CC ProC RWC CAT APX

Polycross families

h2PFM 0.83 0.86 0.85 0.41 –‡ 0.81 – –

GcV, % 12.25 18.43 13.58 4.09 – 9.76 – –

ΔG 0.05 0.04 0.09 0.006 – 3.86 – –

ΔG, % 7.85 11.98 8.81 1.83 – 6.15 – –

Open-pollinated families

h2PFM 0.86 0.89 0.88 0.48 – 0.86 – –

GcV, % 9.95 15.02 11.17 4.89 – 8.01 – –

ΔG 0.04 0.03 0.07 0.008 – 3.47 – –

ΔG, % 6.48 9.95 7.37 2.39 – 5.21 – –

All HS families

h2PFM 0.82 0.79 0.82 0.50 0.72 0.83 0.62 0.75

PcV, % 11.99 18.83 13.27 6.90 36.83 10.09 39.76 42.09

GcV, % 10.91 16.80 12.03 4.89 31.30 9.22 31.87 36.46

ΔG 0.05 0.03 0.08 0.008 2.51 3.81 0.19 1.19

ΔG, % 6.94 10.49 7.63 2.43 18.62 5.90 17.34 22.11† APX, specific ascorbate peroxidase activity; cAT, specific catalase activity; cc, carotenoid content; chl a, chlorophyll a content; chl b, chlorophyll b content; chl a + b, total chlorophyll content; DMY 1, 2, 3, and 4, dry matter yield at cut 1, 2, 3, and 4; FD, flowering date; LSR, leaf to stem ratio; PD, plant density; PH, plant height; Proc, free proline content; RWc, relative water content; SPM, sensitivity to powdery mildew.

‡ negative heritability was observed due to negative estimates of genetic variances.


potential for genetic gains of these traits. The family vari-ance and family × environment interaction variances in HS families express additive genetic variance and addi-tive genetic interaction variances (Nguyen and Sleper, 1983a). In this study, variations among HS families and their interactions with moisture environments and year were significant for the measured traits. Since the genetic variance among HS families is related to additive genetic variance (Nguyen and Sleper, 1983a), significant varia-tion among HS families for most of the traits indicated that mainly additive genetic variance controls these traits. General combining ability is the capacity of an indi-vidual to transmit superior performance to its offspring. Estimates of GCA from progeny test would be effective in selecting superior parents. The majority of the traits measured showed a broad range of GCA. Similar results were observed in forage plants by Araghi et al. (2014) and De Araujo et al. (2002). Since the highest GCA for PH, PD, DMY, RWC, and APX was observed for family 16, this family could be used as a superior family in breeding

programs. The highest and lowest GCA for LSR was observed for family 4 and 21, respectively. High LSR is usually related to forage quality, indicating higher pal-atability of leaves; therefore, family 4 could be recom-mended as a preferable family in terms of forage quality. The forage yield from the first to fourth cut was decreased, which is in agreement with the Demdoum et al. (2010) report suggesting that the first and second cuts composed 70% of annual dry matter yield in sainfoin. High GCV for a trait usually indicate possibility of improvement through selective breeding. The smaller difference between phe-notypic coefficient of variation (PCV) and GCV shows smaller effect of environment. In this study, the smallest differences between these two coefficients were observed for FD, PH, Chl a, Chl a+b, and RWC, indicating more gain will result by selecting for these traits. Majidi et al. (2009) reported that in tall fescue (Festuca arundinacea Schreb.), the difference between phenotypic and geno-typic coefficient of variation was smaller for phenologi-cal traits than forage yield and its components. Proline

Table 6. General combining ability in different traits of sainfoin.

Half-sib families

Traits†

FD PH PD LSR DMY 1 DMY 2 DMY 3 DMY 4

1 -1.097 -5.179 -13.25 0.07 -228.06 -109.12 -126.31 -91.27

2 -2.51 2.267 7.085 -0.03 -30.16 -7.42 43.14 27.43

3 -0.901 -8.509 -17.38 0.02 -299.26 -144.52 -118.3 -82.34

4 1.1439 -15.56 -24.96 0.11 -520.36 -255.32 -197.17 -137.43

5 -3.69 -8.684 -20.84 -0.10 -292.86 -128.52 -161.76 -112.67

6 1.5268 -5.802 -2.655 0.08 -232.96 -124.62 -50.43 -38.2

7 1.6335 3.017 -5.088 0.01 -32.56 5.98 -39.81 -30.57

8 -0.405 -1.417 -12.93 0.06 -234.56 -87.72 -110.47 -78.61

9 0.9972 -3.284 -22.65 0.08 -387.96 -180.12 -180.99 -126.6

10 -4.042 -6.157 -7.822 0.03 -88.96 -41.82 -43.89 -31.16

11 -2.569 -2.707 -4.626 0.08 -205.76 -102.52 -44.44 -33.38

12 -6.689 -1.645 0.488 -0.06 -280.46 -141.02 -43.72 -33.00

13 -0.933 3.329 -9.822 0.03 -146.46 -51.62 -86.64 -62.56

14 0.6924 1.443 -8.538 0.02 -229.36 -107.12 -87.55 -62.48

15 -1.932 1.464 -7.343 0.06 -242.96 -107.82 -62.00 -42.27

16 -0.534 15.853 37.876 -0.02 572.34 278.78 270.07 186.07

17 1.8668 10.775 14.42 -0.02 84.34 42.68 29.10 24.29

18 2.2062 4.949 5.295 0.01 222.84 104.38 2.28 4.69

19 1.9849 8.555 13.819 -0.05 192.04 56.58 41.93 32.00

20 1.2245 -1.882 -0.367 -0.11 121.54 55.78 39.31 28.51

21 1.2916 5.604 21.548 -0.16 439.94 209.38 199.24 139.52

22 1.7024 0.499 16.524 -0.001 225.04 123.08 200.45 140.65

23 4.9541 9.1 17.168 0.003 465.04 223.38 154.42 108.23

24 1.7289 -4.2 0.932 0.06 -84.46 -54.32 38.91 27.49

25 -0.246 -5.91 8.814 0.02 288.04 134.88 120.93 85.68

26 -0.971 -0.919 6.895 0.01 231.74 101.78 67.81 49.94

27 -0.527 4.707 13.98 -0.07 496.44 239.58 128.84 91.16

28 1.7539 2.295 9.163 -0.07 125.14 66.98 85.47 61.82

29 2.1997 3.029 -2.39 -0.03 146.14 62.18 6.42 5.85

30 0.1524 -5 -13.13 -0.05 -73.86 -61.82 -74.86 -50.75

LSD (a = 0.05) 0.416 4.696 16.937 0.033 376.92 211.78 91.42 66.72† DMY 1, 2, 3, and 4, dry matter yield at cut 1, 2, 3, and 4; FD, flowering date; LSR, leaf to stem ratio; PD, plant density; PH, plant height.


content, APX, and CAT had the highest genotypic varia-tion, while the lowest one was observed for LSR and CC. High variation for forage yield (almost 20%) showed that selection for this trait could be effective. Considerable variation for forage yield has been reported in other spe-cies (Aastiveit and Aastiveit, 1990; De-Araujo et al., 2002; Kanapeckas et al., 2005; Majidi et al., 2009). Estimation of heritability and gene action of different traits is impor-tant for developing new varieties (Bowley and Christie, 1981). Successfully identifying parents that will combine well and produce productive progenies mainly depends on the gene action that controls the trait (Tabassum et al., 2007). All HS families studied here had moderate to high heritability for most of the measured traits, showing that a large part of the total genetic variation was additive and these traits could be improved by recurrent or mass selec-tion. The estimates of heritability for DMY in the present study were moderately high (about 0.6), which is almost in agreement with those previously reported in forage plants (Aastiveit and Aastiveit, 1990; Amini et al., 2013;

Araghi et al., 2014; Kanapeckas et al., 2005; Majidi et al., 2009). Jensen et al. (2006) evaluated genetic variation of 28 half-sib families in meadow bromegrass for DMY. They reported that narrow-sense heritability for DMY at cut 2 (0.89) was higher than cuts 3 and 4 (0.59 and 0.53, respectively) and suggested that selection at earlier forage harvests in the growing season might be more efficient than later harvests. This is in agreement with our findings, as the first forage cut had the highest ΔG. Monirifar (2011) in alfalfa reported that the narrow sense heritability values for fresh matter yield, DMY, and PH were about 0.60, 0.59, and 0.5, respectively. In this study, plant height had high heritability. Many studies have shown that a large part of the genetic variance for PH and IL in different forage grasses are under additive genetic control (Araghi et al., 2014; De Araujo et al., 2002; Ebrahimiyan et al., 2012; Majidi et al., 2009). Since these traits (IL and PH) in forage plants have moderate to high heritability, selec-tion would be effective for them. In this study, the lowest heritability was observed for LSR, which is in agreement

Table 7. General combining ability of physiological traits in sainfoin.

Half-sib families

Traits†

Chl a Chl b Chl a+b CC ProC RWC CAT APX

1 -0.013 0.0601 0.0467 0.0203 3.323 -2.931 0.3409 1.35572 0.053 0.0358 0.0886 -0.0120 6.622 1.688 0.5618 3.03273 0.025 0.0274 0.0528 0.0054 3.859 -7.491 0.2332 1.44174 -0.031 -0.0374 -0.0696 0.0413 1.368 -12.989 0.0472 0.23475 -0.205 -0.102 -0.3081 -0.0460 2.241 -9.612 0.138 0.56876 0.139 0.0934 0.2323 0.0242 4.436 -6.287 0.2757 1.52377 0.018 -0.0081 0.0103 -0.0016 2.266 2.688 0.1012 0.69578 0.012 0.0272 0.0393 0.0193 2.238 -2.568 0.1733 0.78779 -0.028 -0.0319 -0.0607 0.0269 2.726 -3.096 0.242 0.9247

10 0.022 0.0340 0.0559 0.0191 8.112 -5.889 0.6287 3.179711 -0.029 0.0067 -0.0235 0.0267 5.124 -2.258 0.3536 1.851712 0.018 0.0193 0.0374 -0.0305 4.97 -1.671 0.3198 1.751713 0.089 0.1925 0.2812 0.0065 3.713 3.083 0.2649 1.216714 0.110 0.0445 0.1547 0.0102 6.241 0.816 0.7636 3.322715 -0.014 -0.0123 -0.0271 0.0227 4.901 0.68 0.6983 3.349716 -0.245 -0.0784 -0.3244 -0.0138 4.649 17.372 0.6329 4.101717 0.099 0.0567 0.1562 0.0101 -6.627 11.354 -0.5883 -2.814318 0.067 0.0470 0.1141 0.0091 -7.111 5.526 -0.6252 -3.012319 -0.024 -0.0462 -0.0712 -0.0168 -6.098 9.326 -0.5444 -2.625320 -0.044 -0.0716 -0.1159 -0.0418 -5.861 -3.223 -0.5183 -2.589321 0.049 0.0346 0.0836 -0.0573 -4.621 5.634 -0.425 -2.072322 -0.143 -0.0736 -0.2176 0.0060 -4.909 -0.973 -0.4505 -2.164323 0.067 0.0051 0.0727 0.0027 -5.443 9.342 -0.4918 -2.377324 0.036 -0.0013 0.0345 0.0324 -5.660 -2.921 -0.5088 -2.463325 0.053 0.0126 0.0654 0.0104 -5.534 -5.958 -0.4989 -2.413326 -0.089 -0.0664 -0.1559 0.0136 -5.523 -1.222 -0.4964 -2.421327 -0.131 -0.0873 -0.2191 -0.0263 -5.949 4.411 -0.5349 -2.550328 -0.003 -0.068 -0.0719 -0.0244 -4.135 1.25 -0.3817 -1.928329 0.037 -0.0006 0.0365 -0.0097 -0.976 2.213 0.0035 -0.509330 0.099 0.0034 0.1024 -0.0169 1.662 -6.21 0.2868 0.6007

LSD (a = 0.05) 0.088 0.071 0.145 0.003 4.68 5.15 0.64 2.56† APX, specific ascorbate peroxidase activity; cAT, specific catalase activity; cc, carotenoid content; chla, chlorophyll a content; chlb, chlorophyll b content; chl a+b, total chlorophyll content; Proc, free proline content;RWc, relative water content.


with the Ebrahimiyan et al. (2012) findings in tall fescue. Other researchers have reported a range of low to high heritability for LSR in different plant species (Brandle and Rosa, 1992; Singh et al., 2013).

Heritability estimates are probably most useful when used to predict ΔG. Viana et al. (2009) showed that narrow sense heritability is a less biased estimator of genetic gain than the estimator based on broad sense heritability. Also, they mentioned that the predictions based on mass and family selection were suitable for comparing selection strategies, whereas predictions based on selection within progenies showed the largest bias and lower association with the realized gain. Plant density (11.2%) and DMY at

cut 1 (14.1%) had the highest ΔG values and LSR (2.0%) had the lowest ΔG. Nguyen and Sleper (1983b) reported that predicted genetic gains per cycle of phenotypic family selection in tall fescue were 33, 45, 27, and 34% of the population means for maturity score, number of panicles, panicle length, and seed yield, respectively. Results of cor-relation coefficients revealed that DMY were positively and significantly associated with PH and PD under both water-deficit and nonstressed environments. Therefore, indirect selection for PH and PD could be effective for improving forage yield. Turk and Celik (2006) reported significant and positive correlations among DMY, number of stems, and PH in sainfoin. Also, Majidi et al. (2009) and

Figure 2. Distribution of the first two principal components (Pc) of agronomic and physiological traits in 30 HS families of sainfoin under water stress.FD, flowering date; PH, plant height; PD, plant density; LSR, leaf to stem ratio; DMY, dry matter yield; chl, total chlorophyll content; cc, carotenoid content; Proc, free proline content; RWc, relative water content; cAT, specific catalase activity; APX, specific ascorbate peroxidase activity.

Figure 1. Distribution of the first two principal components (Pc) of agronomic and physiological traits in 30 HS families of sainfoin under nonstressed environment.FD, flowering date; PH, plant height; PD, plant density; LSR, leaf to stem ratio; DMY, dry matter yield; chl, total chlorophyll content; cc, carotenoid content; Proc, free proline content; RWc, relative water content; cAT, specific cata-lase activity; APX, specific ascorbate peroxidase activity.


Amini et al. (2013) in tall fescue and Wilkins (1985) in perennial ryegrass (Lolium perenne L.) observed significant correlation between plant DMY and PH. Therefore, PH and PD (number of stems per m2) are important traits to indirectly improve the forage yield of sainfoin.

Half-sib families, generated through polycross and open pollination, are frequently used in forage plants breeding to estimate narrow-sense heritability and addi-tive genetic variance (Nguyen and Sleper, 1983a). In this experiment, open-pollinated families at both mois-ture environments showed a higher mean of FD, PH, PD, DMY, and RWC compared with polycross families. Polycross families showed a higher mean of antioxidant activity than open-pollinated families at both moisture environments. Milic et al. (2010) reported that there were significant differences among progeny tests (polycross, open pollination, and selfing) for many traits and open-pollinated progenies had a higher green forage yield and number of shoots per plant than progenies from selfing. In this study there were significant differences between polycross and open-pollinated families for most of the measured traits. The additive genetic variance is con-founded with dominance genetic variance in full-sibs or Sl progenies; therefore, heritability could be overestimated. But HS families enable more precise estimates of additive genetic variance (Falconer and Mackay, 1996). Despite its importance, little is known about gene action of different traits in polycross and open-pollinated families of forage plants. Also, the comparison between polycross and open-pollinated families in terms of heritability and genetic gain is extremely limited in the literature. Estimation of herita-bility and genetic gain from selection in polycross families

were higher than open-pollinated families for most traits. Millic (2010) showed that open-pollination and self-polli-nation progeny tests are effective tools for evaluating alfalfa parents for dry matter yield and crude protein content, respectively. De Araujo et al. (2002) reported that differ-ences in dry matter yield between the parental populations and progenies obtained by different pollination methods (polycross, open pollination, and selfing) were signifi-cant. Also, they reported that open-pollinated progeny test is the best method of choice for selecting parents for synthetics of meadow bromegrass. Study of traits related to drought-resistance and drought-tolerance mechanism is an important objective in forage plants breeding. The most reliable way of finding drought-tolerant genotypes would be field trials. In this study, water deficiency had significant effects on sainfoin’s growth. Drought stress caused significant decline in most measured traits except for CC, ProC, CAT, and APX. Other researchers have shown that water deficit decreased plant growth, PH, stem density, and photosynthesis in alfalfa (Antolin and Sanchez-Diaz, 1993; Brown and Tanner, 1983). Flowering date decreased under water deficit, indicating that peren-nial plants could also escape drought by earlier maturity, similar to annual plants. Although several reports have shown an increased LSR due to water deficit (Carter and Sheaffer, 1983; Halim et al., 1989), our results explained a reduced LSR of stressed plants due to a reduction in plant density. The differences between our results and results reported by other researchers may be due to the fact that plants in our experiment were subjected to a more severe water deficit. When compared with the control, it was evi-dent that water deficit resulted in a decrease in chlorophyll

Figure 3. Principal component (Pc) analysis of 30 half-sib families of sainfoin based on agronomic and physiological traits.


and RWC and an increase in CC, ProC, CAT, and APX activity. Under water deficit, increased proline content in forage plants and other species have been reported and may play a role in maintaining osmotic turger, fixing membranes and thereby preventing electrolyte leakage and bringing concentrations of reactive oxygen species within normal ranges (Hayat et al., 2012; Keyvan, 2010; Xu et al., 2010). Bukvic et al. (2008) reported that con-centration of proline in red clover leaves was in significant correlation with both fresh and dry matter yield per plant. Many studies have shown that antioxidant activity is cor-related with plant tolerance to abiotic stresses (reviewed in Miller et al.,2010 and Xu et al., 2010). In this study, water deficit caused a 74% increase in CAT activity when compared with the control plants. It is believed that toler-ant plants have better capacity to protect themselves from water stress via the enhancement of antioxidant activity (Xu et al., 2010). Our data showed that three HS families, including 14, 15, and 16, had higher antioxidant activity. The tolerance of plants to water-deficient environment is a complicated phenomenon that involves different mor-phological, physiological, and biochemical mechanisms. But from a breeding point of view, it is important to select plants that perform better under drought and produce a reasonable yield. According to our study, families 16 and 27 manifested the best agronomic, morphological, and physiological traits of adaptation to water deficit.

In this study, PCA showed two factors which explained 66.4 and 66.6% of the total variability in nonstressed and water-deficit environments, respectively. These factors could be selected on to improve sainfoin in nonstressed and stress conditions accordingly. The first factor was con-sidered as ‘potential of yield,’ which emphasized forage yield and yield-related traits (such as PD). The second factor had positive correlation with PH and RWC and negative correlation with ProC, CAT, and APX activity and could be considered as the ‘potential of physiologi-cal traits.’ Selection of genotypes with high PC1 would increase forage yield in both moisture environments. This would be more effective than selection only based on forage yield and each factor could be used as a selection index. Also, PCA showed that 84.69% of the total varia-tion can be determined by the first three PC under both moisture environments. Selection of genotypes that have high PC1 and moderate PC2 are suitable for both environ-ments. Therefore, open-pollinated families 21, 23, 27, 17, and 19 are the superior families in this respect. Also, poly-cross family 16 is the preferable family with high PC1 and PC2 under water-deficit condition. Thomas et al. (1995) characterized accessions of meadow fescue from seven dif-ferent geographical regions under water stress based on biplot analysis. Since families scattered around the vectors in the biplot diagram, distinct groups of genotypes could be identified. As a result, we could separate polycross and

open-pollinated families based on the PCA. In conclu-sion, moderate to high heritability for most of the agro-nomic, morphological and physiological traits obtained in this study indicates that recurrent selection would be useful to improve these traits in sainfoin. Also, high vari-ability (PCV and GCV) and moderate heritability, along with high genetic gain, was observed for DMY at first cut. Results showed that selection for forage yield will be more effective in first cut than the later ones in the grow-ing season. Also, indirect selection for forage production would be possible through some correlated morphological traits, such as PH and PD. In addition, our results revealed that water deficiency reduced PH, PD, DMY, Chl a, and RWC and increased CC, ProC, and antioxidant activity (CAT and APX) in sainfoin. Open-pollinated families showed a higher mean for most traits in both environ-ments, while polycross families revealed higher heritabil-ity and genetic gain from selection, indicating selection would be more effective in polycross families.

ReferencesAastiveit, A.H., and K.Aastiveit. 1990. Theory and application of

open pollination and polycross in forage grass. Theor. Appl. Genet.79:618–624.

Aebi, H.E.1983. Catalase In: Methods of enzymatic analysis. 3rd ed.VerlagChemie, Weinheim, Germany. p. 273–286.

Allen, R.G., L.S.Pereira, D.Raes, and M.Smith. 1998. Crop evapo-transpiration—Guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. Food and Agriculture Orga-nization of the United States, Rome, Italy.

Amini, F., M.M.Majidi, and A.Mirlohi. 2013. Genetic and genotype × environment interaction analysis for agronomical and some mor-phological traits in half-Sib families of tall fescue. Crop Sci.53:411–421. doi:10.2135/cropsci2012.05.0277

Antolin, M.C., and M.Sanchez-Diaz. 1993. Effects of temporary droughts on photosynthesis of alfalfa plants. J. Exp. Bot.44:1341–1349. doi:10.1093/jxb/44.8.1341

Araghi, B., M.Barati, M.M.Majidi, and A.Mirlohi. 2014. Application of half-sib Mating for genetic analysis of forage yield and related traits in Bromusinermi. Euphytica196:25–34. doi:10.1007/s10681-013-1011-2

Bates, L.S., R.P.Waldran, and I.D.Teare. 1973. Raipid determination of free proline for water studies. Plant Soil39:205–208. doi:10.1007/BF00018060

Bowley, S.R., and B.R.Christie. 1981. Inheritance of dry matter yield in a heterozygous population of alfalfa. Can. J. Plant Sci.61:313–318. doi:10.4141/cjps81-044

Boyer, J.S.1982. Plant productivity and environment. Sci-ence218:443–448. doi:10.1126/science.218.4571.443

Brandle, J.E., and N.Rosa. 1992. Heritability for yield, leaf: Stem ratio and stevioside content estimated from a landrace cultivar of Stevia rebaudiana. Can. J. Plant Sci.72:1263–1266. doi:10.4141/cjps92-159

Brown, P.W., and C.B.Tanner. 1983. Alfalfa stem density and leaf growth during water stress. Agron. J.75:799–805. doi:10.2134/agronj1983.00021962007500050017x

Bukvic, G., S.Grljusic, M.Antunovic, J.Horvatic, and M.Spoljarevic. 2008. Differences among red clover (Trifoliumpratense L.) genotypes in yield, germination and proline concentration. Cereal Res. Com-mun.36:1507–1510.


Carter, P.R., and C.C.Sheaffer. 1983. Alfalfa response to soil water deficit. I. Growth, forage quality, yield, water use, and water use efficiency. Crop Sci.23:669–675. doi:10.2135/cropsci1983.0011183X002300040016x

De-Araujo, M.R.A., B.E.Coulman, and G.Rakow. 2002. Genetic variation, heritability and progeny testing in meadow bromegrass. Plant Breed.121:417–427. doi:10.1046/j.1439-0523.2002.739124.x

Delgado, I., C.Andres, and F.Munoz. 2008. Effect of the environmental conditions on different morphological and agronomical characteris-tics of sainfoin. Options Mediterr.79:199–202.

Demdoum, S., F.Munoz, and I.Delgado. 2010. Forage production of a collection of sainfoin over a three year period. 13th meeting of the FAO-Ciheam sub-network on Mediterranean pastures and fodder crops, Alicante Spain. 7–10 April.

Ebrahimiyan, M., M.M.Majidi, and A.Mirlohi. 2012. Genotypic varia-tion and selection of traits related to forage yield in tall fescue under irrigated and drought stress environments. Grass Forage Sci.68:59–71. doi:10.1111/j.1365-2494.2012.00869.x

Falconer, D.S., and T.F.C.Mackay. 1996. Introduction to quantitative genetics. Longman, Harlow, UK.

Frame, J.2005. Forage legumes for temperate grasslands. Science Pub-lishers Inc., Enfield, NH. p. 127–132.

Goplen, B.P., K.W.Richards, and J.R.Moyer. 1991. Sainfoin for west-ern Canada. Agriculture Canada, Ottawa, ON.

Halim, R.A., D.R.Buxton, M.J.Hattendorf, and R.E.Carlson. 1989. Water deficit effects on alfalfa at various growth stages. Agron. J.81:765–770. doi:10.2134/agronj1989.00021962008100050014x

Hayat, S., Q.Hayat, M.N.Alyemeni, A.S.Wani, J.Pichtel, and A.Ahmad. 2012. Role of proline under changing environments: A review. Plant Signal. Behav.7:1456–1466. doi:10.4161/psb.21949

Jensen, K.B., J.G.Robins, B.L.Waldron, and M.D.Peel. 2006. Genetic variation in dry matter production and nutritional characteristics of meadow bromegrass under repeated defoliation. Crop Sci.46:1948–1954. doi:10.2135/cropsci2005.12-0511

Kanapeckas, J., P.Tarakanovas, and A.Lemeþiene. 2005. Variability, heritability and correlations of genetic resources in meadow fescue. Biol. J.3:10–14.

Keller, J., and R.D.Bliesner. 1990. Sprinkle and trickle irrigation. Van Nostrand Reinhold, NY. p. 75.

Keyvan, S.2010. The effects of drought stress on yield, relative water content, proline, soluble carbohydrates and chlorophyll of bread wheat cultivars. J. Anim. Plant Sci.8:1051–1060.

Lawlor, D.W., and G.Cornic. 2002. Photosynthetic carbon assimi-lation and associated metabolism in relation to water deficits in higher plants. Plant Cell Environ.25:275–294. doi:10.1046/j.0016-8025.2001.00814.x

Lichtenthaler, H.K., and A.R.Wellburn. 1983. Determinations of total carotenoids and chlorophylls a and b of leaf extracts in different sol-vents. Biochem. Soc. Trans.11:591–592.

Majidi, M.M., A.Mirlohi, and F.Amini. 2009. Genetic variation, heri-tability and correlations of agro-morphological traits in tall fescue (FestucaarundinaceaSchreb.). Euphytica167:323–331. doi:10.1007/s10681-009-9887-6

Milic, D., S.Katic, J.Bocanski, D.Karagic, A.Mikic, and S.Vasiljevic. 2010. Importance of progeny testing in alfalfa breeding (Medicago sativa L.). Genetika42:485–492. doi:10.2298/GENSR1003485M

Miller, G., N.Suzuki, S.Ciftci-Yilmaz, and R.Mittler. 2010. Reac-tive oxygen species homeostasis and signaling during drought and salinity stresses. Plant Cell Environ.33:453–467. doi:10.1111/j.1365-3040.2009.02041.x

Monirifar, H.2011. Expected genetic gain for several quantitative traits in alfalfa (Medicago sativa L.). Not. Sci. Biol.3:109–113.

Nakano, Y., and K.Asada. 1981. Hydrogen peroxide is scavenged by ascorbate specific peroxidase in spinach chloroplasts. Plant Cell Physiol.22:867–880.

Nguyen, H.T., and D.A.Sleper. 1983a. Theory and application of half-sib mating in forage grass breeding. Theor. Appl. Genet.64:187–196. doi:10.1007/BF00303763

Nguyen, H.T., and D.A.Sleper. 1983b. Genetic variability of seed yield and reproductive characters in tall fescue. Crop Sci.23:621–626. doi:10.2135/cropsci1983.0011183X002300040006x

SAS Institute. 1999. SAS/STAT users guide. SAS Institute, Inc, Cary, NC.Singh, J., B.R.Ranwah, L.Chaudhary, C.Lal, M.C.Dagla, and

V.Kumar. 2013. Evaluation for genetic variability, correlation and path coefficient in mutant population of (Sorghum Bicolor L. Moench). Bioscan.8:1371–1376.

Smirnoff, N.1998. Plant resistance to environmental stress. Curr. Opin. Biotechnol.9:214–219. doi:10.1016/S0958-1669(98)80118-3

Statgraphics. 2007. Statgraphics. Version 15.2.11: Stat Point Inc., Wash-ington, VA.

Steel, R.G.D., and J.H.Torrie. 1960. Principles and procedures of sta-tistics: With special reference to the biological sciences. McGraw, New York. p. 481.

Tabassum, M.I., M.Saleem, M.Akbar, M.Y.Ashraf, and N.Mahmood. 2007. Combining ability studies in maize under normal and water stress conditions. J. Agric. Res.45:261–269.

Tafteh, A., A.R.Sepaskhah. 2012. Yield and nitrogen leaching in maize field under different nitrogen rates and partial root drying irrigation. Int. J. Plant Prod. 6:93–114.

Thomas, H., S.J.Dalton, C.Evans, K.H.Chorlton, and I.D.Thomas. 1995. Evaluating drought resistance in germplasm of meadow fes-cue. Euphytica92:401–411. doi:10.1007/BF00037125

Turk, M., and N.Celik. 2006. Correlation and path coefficient analyses of seed yield components in the sainfoin (Onobrychis sativa L.). J. Biol. Sci.6:758–762. doi:10.3923/jbs.2006.758.762

Turner, N.C.1981. Techniques and experimental approaches for the measurement of plant water status. Plant Soil58:339–366. doi:10.1007/BF02180062

Viana, J.M.S., V.R.Faria, and A.D.C.Silva. 2009. Bias in the prediction of genetic gain due to mass and half-sib selection in random mat-ing populations. Genet. Mol. Biol.32:497–506. doi:10.1590/S1415-47572009005000064

Wricke, G., and W.E.Weber. 1986. Quantitative genetics and selection in plant breeding. Walter de Gruyter, New York. p. 406.

Wilkins, P.W.1985. Breeding for dry matter yield in perennial ryegrass by wild hybridization and recurrent selection. In: J. Sjodin, edi-tor, Proceedings of the 13th EUCARPIA Fodder Crop Meeting, Savalov, Sweden. p. 25–30.

Xu, Z., G.Zhou, and H.Shimizu. 2010. Plant responses to drought and rewatering. Plant Signal. Behav.5:649–654. doi:10.4161/psb.5.6.11398

researc half-sib matting and genetic analysis of agronomic ... · tion of additive genetic effects...

Documents