effects of location on performance of selected tropical ... papers/jtas vol. 25 (2) sep. 2002... ·...

PertanikaJ. Trop. Agric. Sci. 25(2): 75 - 86 (2002) ISSN: 1511-3701© Universiti Putra Malaysia Press

Effects of Location on Performance of Selected Tropical MaizeHybrids Developed in Malaysia

GHIZAN B. SALEH1, ABDULLAH B. DAUD2 & ANUAR A. RAHIM3

1 Department of Crop Science, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia2 Ministry of Agriculture Malaysia, Jalan Sultan Salahuddin, 50624 Kuala Lumpur, Malaysia

3 Department of Land Management, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

Keywords: Genotype x environment interaction, hybrids, heritabiliiy, Zea mays L.

ABSTRAK

Lapan belas hibrid terpilih jagung bijian tempatan yang terdiri daripada sepuluh kacukan tunggal, empatkacukan dua ganda dan empat kacukan tiga hala telah dinilaikan dengan dua varieti kawalan pendebungaanbebas (Metro dan Suwan 1), di tiga lokasi di Semenanjung Malaysia. Kajian ini bertujuan untuk membandingkanprestasi hibrid'hibrid tersebut pada lokasi-lokasi berkenaan, menilaikan interaksi genotip x lokasi yang ditunjukkanuntuk hasil dan ciri penting yang lain, dan menganggarkan kebolehwarisan ciri-ciri tersebut Banyak di antarahibrid-hibrid tersebut didapati memberikan hasil yang lebih tinggi dari varieti-varieti kawalan. Kesan lokasibererti yang ditunjukkan untuk hasil bijian dan kebanyakan ciri lain memberikan gambaran bahawa faktor-faktor persekitaran dan tumbesaran yang terdapat di lokasi-lokasi berlainan telah mempengaruhi prestasi hibrid-hibrid tersebut. Interaksi genotip x lokasi didapati tidak bererti ke atas ciri-ciri penting, menunjukkan bahawakedudukan kepentingan genotip-genotip dalam prestasinya adalah seragam pada setiap lokasi. Prestasi hibridadalah terbaik di Terengganu, berbanding Selangor dan Perlis. Nilai kebolehwarisan lebih tinggi diperoleh diTerengganu untuk ciri-ciri yang diukur menekankan lagi keseragaman dalam faktor-faktor persekitaran yangmengawal tumbesaran jagung bijian di lokasi tersebut. Berdasarkan kepada prestasi keseluruhan, tiga kacukantunggal (SGI, SC-2 dan SC-8), satu kacukan dua ganda (DC-11) dan dua kacukan tiga hala (TWC-15 danTWC-17) didapati sangat berpotensi untuk dilakukan penilaian seterusnya ke arah pengeluaran varieti baru.

ABSTRACT

Eighteen selected locally-developed maize hybrids involving ten single, four double and four three-way crosses wereevaluated with two open^pollinated check varieties (Metro and Suwan 1) at three locations in PeninsularMalaysia. The study was aimed to compare performance among the hybrids at the locations separately and whencombined, to evaluate the genotype x location interaction revealed for yield and other important traits, and toestimate the heritability of the traits. Many hybrids were found to have out-yielded the check varieties. Significanteffects of location for grain yield and most other traits indicated that the environmental and growth factorscharacterising the locations considerably affected the performance of the hybrids. Genotype x location interactionswere not significant for the important traits, indicating that the ranking in performance of the genotypes wasconsistent among locations. Performance of the hybrids was best in Terengganu as compared to Selangor andPerlis. Higher heritability estimates for characters when measured in Terengganu further revealed the uniformityof environmental factors governing maize growth in that location. Based on the overall performance, three singlecrosses (SC-1, SC-2 and SC-8), one double cross (DC-11) and two three-way crosses (TWCrl5 and TWC-17) werefound to be highly potential for further trials towards new variety release.

INTRODUCTION are imported. Local maize production is im-Maize {Zea mays L.) is an important import paired by high production cost and low produc-commodity in Malaysia, as almost all maize grains tivity from m e u s e o f low-yielding varieties (Leongutilised to support the local animal feed industry 1 9 8 7 ; Setefarzi 1990). To ensure profitability of

GHIZAN B. SALEH, ABDULLAH B. DAUD & ANUAR A. RAHIM

local maize production, it is necessary to pro-duce high yielding varieties, in particular, hy-brids that are stable and adaptable to variousenvironments in the country (Saleh et al. 1994).

In maize breeding, the response of geno-types to different environments on which evalu-ations are conducted is an important considera-tion. Neglecting the effects of genotype x envi-ronment interactions in variety developmentwould reduce the effectiveness of selection pro-grammes, thus producing varieties not readilysuitable across varied environments (Oosterumet al 1993). It is more meaningful to obtaingenotypes that are stable across environments,or genotypes stable to any particular environ-ment over a number of years. Genotype x envi-ronment interactions have been shown to con-tribute to yield variations (Sprague and Federer1951). A stable genotype is one that shows mini-mum genotype x environment interaction, orlow variation in performance in varied environ-ments (Hanson 1970), while an unstable geno-type is one which expresses high genotype xenvironment interaction, and responds tochanges in the environment (Piepho 1994).

Weatherspoon (1970), in his evaluation on36 maize single cross hybrids, 36 double crossesand 36 three-way crosses in highly varied loca-tions, reported very high genotype x environ-ment interaction effects. In contrast, the studyby Ordas (1991) reported the absence of geno-type x year and genotype x location interactionson the performance of two maize populations.With regards to the use of tropical maizegermplasm, Subandi (1981), from his evaluationon 15 maize genotypes over 31 environments inIndonesia, reported that genotype x plantingseason and genotype x location within seasoninteractions were very high, where componentof variance for the latter was found to be higherthan that of the former.

Selection of genotypes would only be effec-tive if based on characters governed by geneticfactors, i.e. based on characters with high herit-ability. Heritability estimates for characters ondifferent maize populations, and using variousmethods of estimation were reported by manyprevious workers (Lee 1974, Daniel and Batjay1975; Saleh et al. 1994), where estimates ob-tained varied with methods of estimation, geno-types used and environments of evaluation.

Realising the constraints in the local maizeproduction, an extensive maize breeding pro-

gramme was conducted at Universiti Putra Ma-laysia, aimed at developing superior hybrid vari-eties for grain yield (Saleh et al. 1995). Thepresent study evaluated potential single, doubleand three-way cross hybrids developed and se-lected from this programme, in three separatelocations in the country. The objectives were toidentify those superior and stable for yield, ateach location and locations combined, to evalu-ate the genotype x environment interactionsrevealed for yield and other important agro-nomic characters, and to estimate broad-senseheritability for important traits measured.

MATERIALS AND METHODS

Plant Materials

This study was an advanced stage of a long termtropical maize hybrid variety development pro-gramme conducted at Universiti Putra Malaysia,Serdang. Based on yield performance and gen-eral combining ability in previous evaluations,superior inbred lines were entered into a diallelcrossing scheme to produce Yx hybrids. Evalua-tions of performance of these hybrids were pre-viously done in 1993 and 1994 (Sujiprihati 1996;Saleh and Sujiprihati 1997). Findings from theseprevious evaluations were based on the selectionof hybrids for use in this present study. Ten bestsingle cross (SC) hybrids involving various com-binations among nine inbred lines, four doublecrosses (DC) and four three-way crosses (TWC)were evaluated in this study with two open-pollinated check varieties, Metro and Suwan 1,originating from Indonesia and Thailand re-spectively.

Location

Evaluations of all the genotypes were conductedat three locations in Peninsular Malaysia, viz,Serdang in the state of Selangor, KualaTerengganu in Terengganu, and Kangar in Perlis.The locations represented different geographi-cal zones and soil types. In Serdang, located inthe middle of the southern part of the peninsula(with the bearing of 3° N, 101° E, at 31 m abovesea level), the soil type was sandy loam of orderUltisol (Bungor series, Typic Kandiudult); inKuala Terengganu, located on the east coast ofthe peninsula (with the bearing of 5° N, 105° E,at 15 m above sea level), the soil type was sandy,belonging to the order Spodosol (Rudua series,

76 PERTANIKAJ. TROP. AGRIC. SCI. VOL. 25 NO. 2, 2002

LOCATION EFFECTS ON PERFORMANCE OF TROPICAL MAIZE HYBRIDS DEVELOPED IN MALAYSIA

Typic Haplorthod); while in Rangar, located inthe northern part of the peninsula (with thebearing of 7° N, 100° E, at 35 m above sea level),the soil type was loam of order Ultisol (BukitTemiang series, Typic Hapludult) .

Evaluation

At each location, the evaluation was carried outin a Randomised Complete Block Design, withthree replications. Each replicate consisted of 20plots, each represented by a genotype. Each plotmeasured 5.0m x 4.5m (22.5 m2), with the plant-ing density of 0.75m x 0.25m. Every plot con-sisted of six plant rows, each five metres inlength, giving a total of 20 plants per row, and120 plants per plot. Four middle rows of eachplot measuring four metres long were used asthe harvest area, giving an area of 12 m2, with 64plants. Among the plants in the harvest area, 16plants were used as samples for the measure-ments of individual plant data. In cases wheremore than one ear were present on each plant,the largest ear was used.

Cultivation Practices

Before planting, the soil was ploughed twice toa depth of 15 to 30 cm, followed by rotovation.The soil pH was raised to 5.5 to 6.5 by applyingGround Magnesium Limestone before rotovation.Planting was done manually, at the rate of twoseeds per point, and plants were later thinned tojust one per point 10 days after planting. Fertilis-ers at the rate of 160 kg N ha1, 100 kg ha1 P2O5

and 100 kg ha1 K^O were given in split applica-tions, where a balanced compound fertiliser(15:15:15) was used seven days after planting,followed by Urea (46% N) 20 and 35 days afterplanting. Weeds were controlled manually usingpre-emergence herbicide, 'Lasso' (2-chloro-2'-6f-diethyl-N-methoxymethyl) and post-emergencecontact herbicide, 'Gramoxone' (1,1, dimethyl-dimethyl-4, 4t-bipyridinium). A sprinkler irriga-

tion system was used to supply water to theplants when necessary.

Data Collection

Data collected involved pre-harvest data, includ-ing plant and ear heights at flowering, and daysto tasseling and maturity, and post-harvest data,including yield and yield components, and earand grain trait measurements taken at 15% grainmoisture content.

Data Analysis

Data were subjected to the analysis of variance(ANOVA) carried out separately for each loca-tion, and also locations combined, to determinethe effects of genotypes and also locations in-volved. Data from each location were however,first analysed for uniformity of the error meansquares using the test of homogeneity of errorvariances (Steel and Torrie 1980), followingwhich, only data from locations that showedhomogeneity were combined. Effects of locationand genotype x location interaction for the char-acters measured on the combined data wereanalysed using ANOVA as proposed by Mclntosh(1983). The ANOVA table key-outs with theexpected mean squares (EMS) for each loca-tion, and locations combined are shown in Ta-bles 1 and 2 respectively. For the combinedanalysis, to test the effects of location, replica-tion/location mean squares (MSR/[) were usedas the error term (Table 1). For traits that didnot show significant genotype x location interac-tion, however, the genotype x location meansquares were pooled into error mean squares(Table 2). The effects genotypes were consid-ered fixed, while those of locations and replica-tions were random. The Duncan's New MultipleRange Test was used to compare mean perform-ances among the hybrids and also between themand the control varieties.

Key-outs for ANOVA on

Source of variation

Replications (R)Genotypes (G)Error

TABLE 1characters measured on maize genotypes analysed

d.f.

r-1

(r-D (g-D

Mean squares

MSR

MSG

MS

at locations separately

E(MS)

o\ + go\o\ + ro2

G

a2

r - number of replications; g • number of genotypes;o2

R = replication variance; o2G = genotypic variance; o2

c = error variance.

PERTANIKAJ. TROP. AGRIC. SCI. VOL. 25 NO. 2, 2002 77


TABLE 2Key-outs for ANOVA on characters measured on maize genotypes at locations combined

Source of variation d.f. Mean squares E(MS)

Locations (L), Reps, within L (R/L). Genotypes (G)| G X L

Error

1-1l(r-Dg-1

(g-D(H)

MSL

MSC

MSc

°\

1 • number of locations; r • number of replications; g = number of genotypes;o2

L= locations variance; a2R/I = replications/location variance;

O2G= genotypic variance; o2

GL = genotype x location variance; a2e • error variance.

Broad-sense heritability (h2B) for traits in

the hybrid populations were estimated separatelyfor each location using the components of vari-ance in the ANOVA table procedure (Table 1),as proposed by Baker (1984), as follows:

9Up

where:

- o\

o2G = (ra*GL + rla2

G + o\) - (ro2

rl

= MSG - MSGI

rl

= MSG - MSGL

rl

MSe

MS

MSC - MSe

a*

where:

= MSG - MSe + MSe

broad-sense heritability,genotypic variance,phenotypic variance,environmental variance,mean squares for genotypesin ANOVA,mean squares for error inANOVA, andnumber of replications.

For calculations of broad-sense heritabilityin the combined analysis, referring to Table 2,the equations for the variance components areas follows:

MSe

1r

genotype x location vari-ance,mean squares for genotypesin ANOVA,mean squares for genotypex location interaction inANOVA,mean square error inANOVA,number of locations, andnumber of replications.

RESULTS AND DISCUSSIONEffects of Location, Genotype and Genotype xLocation Interaction

From results of the test of homogeneity of errorvariances, for almost all traits measured, errorvariances were found to be homogeneous, andtherefore, combined data from the three loca-tions were also used in interpretations. The sum-mary of the results of the combined analysis ofvariance is shown in Table 3. In the combined

78 PERTANIKA J. TROP. AGRIC. SCI. VOL. 25 NO. 2, 2002


TABLE 3Mean squares in AN OVA for characters measured on 20 maize genotypes evaluated for combined location

Source ofvariation

Locations (L)Replicationswithin L (R/L)Genotypes (G)G X LError

d.f.

2

191938

114

Grainyieldper ha.

PlantHeight

74133917**1674**

3567375**2599279**

ne360934

449*1122**

ne175

Days toTasseling

105.43**

11.43**6.58**

ne0.78

Mean Squares

Days toMaturity

10.77*

1.16ns

6.48**ne

0.92

Shelling%

195.8"5

40.8*18.3™

ne13.7

Earweight

33554**

1855**909**322*191

Earlength (

Earliameter

15604.5**330.22**

251.8**318.4**112.7**61.2

28.03**8.04**ne

1.95

100-grainweight

61.31*

9.91**6.99ns

4.27*2.52

ne : mean squares for G X L are negligible, and therefore pooled with error mean squares for analyses.** : significant at psO.Ol; * : significant at psO.05; ns : non-significant.

analysis, grain yield and all other characters,except shelling percentage showed significanteffects of location on the overall performance ofthe genotypes, indicating the presence of varia-tions in the growth factors affecting the geno-types at different locations. With regards to vari-ation among genotypes in the combined analy-sis, grain yield and all other characters exceptshelling percentage and 100-grain weight werehighly significantly affected, indicating that thegenotypes evaluated varied greatly in their per-formance across locations.

Genotype x location interactions were foundsignificant only for ear weight, ear length and100-grain weight, while the others, includinggrain yield did not exhibit significant interac-tion (Table 3). The absence of significant geno-type x location interaction for grain yield andother important traits such as plant height anddays to tasseling indicates that, for these charac-ters, the 18 hybrids and the two check varietiesevaluated in general, showed quite similar pat-terns of performance at the three different loca-tions. The three locations selected, each repre-senting the middle, eastern and northern partof Peninsular Malaysia, although showing signifi-cant location effects on the overall performanceof the genotypes, however, did not reveal signifi-cant genotype x location interactions for theimportant traits measured. This is a good indica-tion of the presence of generally good adaptabil-ity of the genotypes, but more evaluations atvarious environments are needed to explore thisphenomenon.

Performance of Genotypes

Mean performances of the 18 hybrids in com-parison with the check varieties at each locationand locations combined, for all the nine char-acters are shown in Table 4.

Grain yields of the genotypes differed greatlywith locations, where superior ones were identi-fied specific to locations. The overall mean grainyield was 5612 kg/ha. Evaluation in Terengganuexhibited the best overall grain yields, with over-all means of 6782 kg/ha, followed by Perlis(5483 kg/ha), and lastly, Selangor (4570 kg/ha). From evaluations in Selangor, four hybridsshowed grain yields significantly higher thanboth the check varieties, viz. DC-11 (5708 kg/ha), TWC-15 (5389 kg/ha), SC-2 (5342 kg/ha)and SC-1 (5141 kg/ha). The two check varieties,Suwan 1 and Metro gave grain yields of 3772kg /ha and 3914 k g / h a respectively. InTerengganu, only TWC-15 out-yielded both thecheck varieties, giving grain yield of 8041 kg/ha,compared to 6778 kg/ha from Metro, and 5989kg/ha from Suwan 1. The grain yield obtainedfrom this hybrid was also the highest recordedin this study. In Perlis, however, hybrid SC-8 wasthe only one giving higher yield than both thecheck varieties, with grain yield of 6473 kg/ha,compared to 5131 kg/ha from Metro and 4134from Suwan 1. When data from all three loca-tions were combined, four single crosses, onedouble, and two three-way crosses exhibited yieldshigher than both the checks, viz. SC-1 (5928 kg/ha), SC-2 (6219 kg/ha), SC-3 (5976 kg/ha), SC-8 (5932 kg/ha), DC-11 (6034 kg/ha), TWC-15(6244 kg/ha) and TWC-17 (5970 kg/ha). Grainyields of Metro and Suwan 1 at locations com-

PERTANIKA J. TROP. AGRIC. SCI. VOL. 25 NO. 2, 2002 79

g1

O

TABLE 4Mean performance and coefficients of variation (c.v.) for characters measured on maize hybrids and check varieties at each location and

locations combined

Geno

type

SC-1SC-2SC-3SC-4SC-5SC-6SC-7SC-8SC-9SC-10DC-11DO12DO13DC-14

Selangor

51415342458348264400494345064472292142385708370642284633

TWC-15 5389TvVC-16 5111TWC-17 4883TWC-18 4644MetroSuwan

Meanc.v. (%)

391413772

457013.9

abababedabedbedabedabedabedebedadebedabedababcabedabedcdede

Grain yield (kg/ha)

Terengg-anu

66817484765467116640717071796847525161536682618868396981804166567107661667785989

67828.1

bedababbedbedabcabcbedecdebedcdebedbedabedabcbedbedde

Perils Combined

59615829569059895722540259366474411652705712523858715244530356385921508751314134

548310.2

(rank inbrackets)

ab 5928ab 6219ab 5976ab 5842ab 5587ab 5839ab 5880a 5931c 4096b 5220ab6034b 5053ab 5646b 5620b 6244ab 5802ab 5970be 5449b 5274c 4632

561210.4

ab (7)a (2)ab (4)abed (9)

Plant height

Selangor Terengg-

179.4197.1195.7196.8

abede (74)188.4abed (10)abc (8)ab (6)g (20)de (17)ab (3)ef (18)

204.5198.6189.5172.5181.6208.8181.8

abede (72)173.7abede (73)200.6a (I)abed (11)ab (5)bede (15)cde (16)% (19)

205.7201.6205.4200.7198.4228.5

195.57.4

anu

cde 196.6bcde234.0bcde234.0bcde222.8bcde222.6abc 223.8bcde223.6bcde219.2e 204.5bede209.1ab 216.8bcde213.7de 207.3bcde232.0abc 216.4bed 217.0abc 233.5bcde216.0bcde249.0a 260.0

222.63.7

gbbbedebedebedbedbedef

%defgcdefldefefgbecdefcdefbdefaa

(cm)

Perlis1

185.6191.6196.6193.6195.1187.6206.2203.1168.0188.291.3201.2179.8198.1192.5174.6203.5179.6207.0201.2

192.27.6

Combined(rank inbrackets)

187.2207.5208.8204.4201.9205.3209.5203.9181.7193.0205.6199.0186.9210.3204.7197.7214.2198.8218.1229.9

203.46.3

bed (14)bed (15)bede (10)cde (8)bede (12)bed (16)bede (9)

%{Defg (4)bede (13)def (7)fg (2)bed (17)bede (11)def (5)be (18)def (6)ab (19)a (20)

Days

Selan-gor1

52.053.054.354.053.753.752.752.754.054.053.353.754.753.352.752.754.753.753.355.7

53.62.1

to tasseling

Terengg-anu

50.052.052.052.052.052.350.052.352.352.750.750.750.350.750.053.052.352.051.053.7

51.71.4

ebedbedbedbedabceabcabcabdedeabcdeeababcbedcdea

(days)

Perils

49.751.751.051.051.052.049.350.051.351.350,349.751.051.350.352.751.051.351.053.3

51.01.2

fbecdecdecdebefefcdcddeffcdecddefabcdecdcdea

50.652.252.452.352.252.750.651.752.652.751.451.351.751.851.052.852.752.351.854.2

52.11.6

Combined(rank in

brackets)

i (1)bedef (10)bed (14)bede (12)bedef (10)be (16)

i (1)defg (6)bed (15)be (16)efgh (5)fghi (4)defg (6)cdefg (8)ghi (3)b (19)be (16)bede (12)cdefg (8)a (20)

SC - single cross; DC - double cross; TWC - three-way cross.Mean values followed by the same letter in the same column are not significantly different at psO.05, following DNMRT.1 DNMRT was not conducted because differences among genotypes were not significant.

TABLE 4 (continued).

3|

o

i2p

§

8

Genotype

SOISC-2SO3SC-4SO5SC-6SO7SG-8SC-9SC-10DO11DC-12DC-13DC-14TWC-15TWO16TWC-17TWO18MetroSuwan 1

Meanc.v. (%)

Selang&r

91.3 abc88.7 e90.0 bcde90.0 bcde90.0 bcde89.0 de89.0 de90.0 bcde92.3 a91.0 abed89.0 de92.3 a90.3 bcde90.3 bcde89.7 cde89.7 cde89.0 de89.3 cde91,0 abed92.0 ab

90.21.2

Days to maturity (days)

Terengganu

90.7 ab89.0 cd89.7 bed89.7 bed89.3 bed89.0 cd88.7 de90.7 ab90.7 ab90.0 bed89.3 bed90.3 be89.3 bed90.0 bed89.7 bed89.0 cd89.0 cd87.7 e89.7 bed91.7 a

89.60.7

Perlis

91.3 abc90.3 be89.3 c90.7 be90.3 be90.0 be90.0 be90.0 be91.3 abc90.7 be90.7 be90.3 be89.7 c90.3 be89.3 c90.3 be89.7 c90.3 be92.0 ab93.0 a

90.51.2


91.1 be (18)89.3 ef (4)89.7 def (7)90.1 cdef (12)89.9 def (11)89.3 ef (4)89.2 ef (2)90.2 cde (13)91.4 ab (19)90.6 bed (15)89.7 def (7)91.0 be (17)89.8 def (10)90.2 cde (13)89.6 def (6)89.7 def (7)89.2 ef (2)89.1 f (1)90.9 be (16)92.2 a (20)

90.11.0

Shelling

Selangor1 Terengg-

80.884.474.881.381.282.279.481.969.482.080.376.581.682.480.083.582.583.680.580.3

80.47.4

anu1

81.881.579.678.080.481.081.278.578.580.179.980.479.479.980.481.181.579.478.877.2

79.93,2

percentage

Perlis

85.4 a84.3 ab83.4 abc83.7 abc81.7 cd83.0 abed83.2 abed82.7 bed82.9 abed84.0 abc84.2 ab84.4 ab82.9 abed82.0 bed83.7 abc83.6 abed89.3 abc83.3 abed81.2 d82.3 bed

83.31.5

(%)

Combined1

82.683.479.381.081.182.181.381.177.082.081.580.581.381.481.482.782.782.180.280.0

81.24.6

Selangor

123.1 ab133.9 a122.8 ab113.4 abc117.4 ab122.6 ab102.6 abc102.9 abc86.0 c

103.7 abc133.6 a96.5 be101.7 be118.8 ab116.1 abc126.7 ab110.1 abc106.0 abc103.2 abc102.5 abc

112.314.2

Ear weight

Terengganu

153.2 cd172.2 abc180.0 ab160.3 bed154.7 bed166.1 abed166.8 abed164.1 abed125.8 e143.8 de156.9 bed144.2 de161.8 bed163.8 abc187.6 a154.0 bed163.8 abed156.4 bed161.7 bed145.2 de

159.18.3

Perils

140.9126.7122.9140.5139.6134.8139.1163.3104.9123.9132.2127.3135.9137.6121.8128.2134.1119.2124.1102.4

129.99.2

(g)

bbebedbbbbacdbedbbebbbedbbbedbedd


139.0 abc (8)144.3 a (1)141.9 ab (3)138.1 abed (9)137.2 abed (10)141.2 ab (5)136.4 abed (11)143.4 a (2)105.6 f (19)123.8 cde (17)140.9 ab (6)122.8 de (18)133.1 abed (14)140.1 ab (7)141.8 ab (4)136.3 abed (12)136.0 abed (13)127.2 bcde (16)129.7 abede (15)105.6 f (19)

133.810.3

SC - single cross; DC - double cross; TWC - three-way cross.Mean values followed by the same letter in the same column are not significandy different at psO.05, following DNMRT.1 DNMRT was not conducted because differences among genotypes were not significant.

TABLE 4 (continued).

Genotype Ear length (mm) Ear diameter (mm) 100-grain weight (g)

3

ooto

SC-1SC-2SO3SG4SC-5SC-6SO7SC-8SC-9SC-10DC-11DC-12DC-13DC-14TWC-15TWC-16TWC-17TWC-18MetroSuwan 1

Mean

Selangor Terengg-anu

Perlis Combined(rank inbrackets)

Selangor1 Terengg-anu

Perlis Combined(rank inbrackets)

Selangor1 Terengg-anu

Perlis Combined1

121 abcdef 143128 abcde 153135 ab 162125 abcdef 148127 abcde 153131 abc114 bcdef108 f108 f127 abcde137 a112 def118 bcdef122 abcdef 160129 abed 160129 abed 151123 abcdef 153117 cdef 150115 cdef 147111 f 152

158149151136144146142149

deabedacdabedabcbedbededededebedabababedabedbedcdabed

149 bcdef145 cdef154 abed153 abcde163 ab161 abc151 abcdef167 a137 ef145 cdef148 bcdef143 def151 abcdef152 abcdef142 def157 abed156 abed145 cdef142 def137 f

138 def (14) 43.7142 abcde (5)44.4151 a (1) 42.6142 abcde (#42.4148 abc (3) 41.4150 ab (2) 43.1138 def (14) 41.2142 bcde (8) 41.9127 g (20) 40.7139 def (12) 42.2144 abed (5) 44.4132 fg (19) 40.8139 cdef (72)42.4141 cdef (77)43,3144 abed (5) 41.5146 abed (4) 42.1144 abed (5) 42.6137 def (16) 42.2135 efg (77) 42.5133 fg (18) 41.5

40.1 bcde48.3 a48.4 a46.6 abcde45.5 bcde46.5 abcde47.0 abc46.9 abed42.9 f45.3 cde45.6 bcde44.8 e46.8 abcde46.6 abcde47.6 ab45.7 bcde46.2 bcde45.9 bcde47.1 abc44.9 de

48.2 ab 46.047.0 abed 46.646.1 bed 45.747.0 abed 45.345.8 bed 44.245.9 bed 45.247.0 abed 45.149.4 a 46.144.9 d46.1 bed45.9 bed46.0 bed48.1 abc48.2 ab45.8 bed45.7 bed46.8 abed 45.245.2 cd 44.548.3 ab 46.044.4 d 43.6

42.844.645.343.945.846.045.044.5

abc (3) 22.7a (7) 21.3abed (7) 21.7abcde (8) 23.5ef (77) 23.0abcde (70)23.0abcde (72)23.0ab (2) 23.2g (20) 20.4cdef (14) 22.6abcde (8) 21.9efg (18) 23.9abed (6) 21.9abc (3) 22.5abcde (75)22.4cdef (15) 24.1abcde (70)21.6def (15) 22.4abc (3) 22.3fg (19) 20.6

24.0 abed24.3 abed24.3 abed22.3 bed26.0 a24.0 abed24.0 abed25.3 ab21.7 d22.0 cd26.7 a22.7 bed26.0 a24.7 abed24.3 abed26.3 a26.7 a23.7 abed25.0 abed24.7 abed

23.922.322.423.624.3

25.0 ab21.2 cde21.0 de25.1 ab24.0 abc23.2 abcde 23.423.7 abed 23.625.0 ab22.2 bcde24.5 ab24.2 ab24.9 ab23.0 abcde 23.423.3 abed 23.522.6 abcde 23.123.7 abed 24.723.1 abcde 23.822.5 abcde 22.925.2 a 24.220.4 e 21.9

24.521.423.024.223.8

1227.5

3.9150

1505.5

1415.6

42.43.3

46.22.2

46.63.2

45.13.0

22.47,fi

24.46.5

23.46.3

23.46.8

SC - single cross; DC - double cross; TWC - three-way cross.Mean values followed by the same letter in the same column are not significantly different at psO.05, following DNMRT.1 DNMRT was not conducted because differences among genotypes were not significant.


bined were 5274 kg/ha and 4632 kg/ha, respec-tively. These results on yield performances indi-cate that the hybrids revealed good adaptabilityto the specific locations concerned. In general,the hybrids evaluated in this study performedwell, and were superior to the open-pollinatedcheck varieties. A previous study by Subandi andSudjana (1982) also indicated the superiority ofthe tropical maize hybrids over the open-polli-nated ones.

Yield performance in Terengganu was en-couraging although the soil type was sandy (or-der Spodosol, series Rudua, typic Haplorthod),and reported to have limitations such as coarsetexture, poor structure and moderatelyimbalanced nutrient composition (Departmentof Agriculture 1993). The effects of these soillimitations might have been remedied throughgood management and cultural practices adoptedin the farmer's plot where the evaluation wasconducted. The performance in Terengganu wasbetter than those in Perlis (soil with order Ultisol,series Bukit Temiang, typic Hapludult) andSelangor (soil with order Ultisol, series Bungor,typic Kandiudult).

Short plants are desirable, as they wouldhave low incidence of lodging. For plant height,overall means for genotypes were highest inTerengganu (222.6 cm), followed by Selangor(195.5 cm), and lastly Perlis (192.2 cm). Thehybrids were generally shorter than the open-pollinated check varieties. In Selangor, 14 out ofthe 18 hybrids evaluated had plants significantlyshorter than Suwan 1 (228.5 cm), but none wasshorter than Metro (198.4 cm). The shortestplants in Selangor were shown by the singlecross hybrid SC-9 (172.5 cm). In Terengganu,all hybrids had plants shorter than those of thetwo check varieties. The shortest plants wereshown by hybrid SC-1 (196.6 cm), followed bySC-9 (204.5 cm), DC-13 (207.3 cm) and SC-10(209.1 cm). In contrast, in Perlis, there was nosignificant difference in plant height among thegenotypes evaluated, with heights ranging from168.0 and 207.0 cm. When locations were com-bined, eight hybrids (SC-1, SO5, SC-9, SC-10,DC42, DO13, TWC-16 and TWO18) had plantswith average heights significantly lower than thoseof the two check varieties.

The hybrids also responded differently tolocations for days to tasseling. On the average,genotypes were earliest tasseling in Perlis (51.0days), and latest in Selangor (53.6 days). This

might have been due to differences in weatherconditions, where before the flowering stage ofthe plants, Perlis had higher temperatures thanthe other two locations, which are favourablefactors to encourage early flowering. In Selangor,genotypes did not differ significantly for days totasseling (ranged between 52.0 to 55.7 days). InTerengganu, although genotypes showed signifi-cant differences in days to tasseling, most hy-brids tasselled only earlier than Suwan 1 (53.7days), but not Metro (51.0 days). In Perlis how-ever, three hybrids, SC-1 (49.7 days), SC-7 (49.3days) and DC-12 (49.7 days) tasselled signifi-cantly earlier than both the check varieties,Suwan 1 (53.3 days) and Metro (51.0 days).When data from locations were combined, twohybrids, SC-1 and SC-7 were found to have tas-selled earlier than most other hybrids as well asboth the check varieties, indicating that theycould be the choice hybrids for earliness. Boththe hybrids tasselled at 50.6 days.

Genotypes, in general, had shown earliermaturity in Terengganu than they did in Selangorand Perlis. The overall genotype mean for matu-rity was 89,6 days in Terengganu, while those ofSelangor and Perlis were not significantly differ-ent from each other, being 90.2 and 90.5 daysrespectively. Earlier maturity in Terengganu wasprobably due to the effects of high temperaturemodification contributed by the sandy soil at thestages of seed filling and drying of the geno-types. In Terengganu, TWC-18 was the earliestmaturing hybrid (87.7 days), and was signifi-cantly earlier than both the check varieties. Evalu-ation in Selangor however, has shown that SC-2was the earliest maturing among the hybrids(88.7 days), and significantly earlier than boththe check varieties (92.0 days for Suwan 1, and91.0 days for Metro). From evaluation in Perlis,four hybrids were found to have matured signifi-cantly earlier than both the check varieties. Theywere SC-3, DC-13, TWC-15 and TWC-17, withdays to maturity ranging from 89.3 to 89.7 days.From locations combined, 11 out of the 18hybrids showed maturity earlier than the twochecks, with overall values ranging from 89.1 to89.9 days.

There were no significant effects of locationon shelling percentages of the genotypes evalu-ated (Table 3), with an overall mean of 81.2%across locations. From results of evaluations inSelangor and Terengganu (Table 4), all geno-types did not show significant differences for

PERTANIKA J. TROP. AGRIC SCI. VOL. 25 NO. 2, 2002 83


shelling percentage, with location means of80.4% for Selangor, and 79.9% for Terengganu.Differences were however, observed from theevaluation in Perlis, where only hybrid SC-1recorded higher shelling percentage (85.4%)than Metro (81.2%) and Suwan 1 (82.3%). Re-sults of the analysis of combined locations alsodid not show significant differences in shellingpercentages among the genotypes.

Significant effects of locations were observedon ear weights of the genotypes, where theyproduced heavier ears in Terengganu than theydid in Selangor and Perlis (Table 4). The meanear weight of the genotypes combined were159.1 g in Terengganu, 129.9 g in Perlis and112.3 g in Selangor. In Selangor, none of thehybrids produced ears heavier than those of thetwo check varieties; but in Terengganu, hybridTWC-15 gave heavier ears (187.6 g) than thoseof the check varieties; while in Perlis, it was SC-8 (163.3 g). In addition, SC-8 was also found toproduce ears heavier than any other hybridsevaluated. From results of the analysis on loca-tions combined, no significant differences wereshown for ear weight between the majority ofthe hybrids and the check variety, Suwan 1.Overall mean ear weight for all genotypes overlocations was 133.8 g.

Ear lengths of the genotypes were signifi-cantly affected by locations, where with the meanear length of 150 mm, genotypes in bothTerengganu and Perlis did better than those inSelangor (122 mm) (Table 4). Evaluations inSelangor have shown that SC-3 and DC-11 hadthe longest ears (135 mm and 137 mm, respec-tively), and they were significantly longer thanthose of the two check varieties. In contrast, inTerengganu, all the 18 hybrids showed earlengths lower than those of Metro and Suwan 1.In Perlis, two single crosses, SC-5 and SC-8 pro-duced ears longer (163 mm and 167 mm, re-spectively) than those of the check varieties.From locations combined, six hybrids, SC-3, SC-5, SG3, TWC-15, TWC-16 and TWCM7 werefound to have ears longer than those of thecheck varieties, with values ranging from 144mm to 151 mm.

Similar to ear length, ear diameter of thegenotypes was significandy affected by locations,where in this case, genotypes produced ears withlarger diameters in Perlis and Terengganu, com-pared to those in Selangor. Mean ear diameterswere 46.6, 46.2 and 42.4 for Perlis, Terengganu

and Selangor respectively. Differences amonggenotypes were only significant in Terengganuand Perlis, but not in Selangor. However, noneof the hybrids had ear diameters larger thanthose of Metro, the check variety with ears hav-ing larger diameters between the two checks.

One hundred-grain weight of the genotypesdiffered significantly among locations, withTerengganu giving the highest (24.4 g), fol-lowed by Perlis (23.4 g), and lasdy Selangor(22.4 g) (Table 4). Among genotypes, 100-grainweight differed only in Terengganu and Perlis,but not in Selangor. However, no hybrid had100-grain weight higher than both the checkvarieties in Terengganu, and Metro in Perlis.

From the overall performance of the hy-brids, this study has revealed that the selectedhybrids have high genetic potential, and shouldbe given the opportunity for further evaluationand selection purposes. Hybrids SC-1, SC-2, SC-8, DC-11, TWO15 and TWC-17 were identifiedto be the highly potential ones, and should begiven good consideration in further evaluationson a larger scale, before any could be releasedas a variety for commercial production.

Variance Components and Heritability

The genotypic and phenotypic variances, andbroad-sense heritabilities for traits estimated fromthe population of hybrids, at each location andlocations combined are shown in Table 5. Nega-tive estimates of variances and heritability ob-tained for some traits are considered as zero,and could be attributed to high uncontrollableerror variances in the experiment.

Broad-sense heritability estimates for grainyield were generally moderate, being 52.6%,41.4%, 29.3% and 38.0%, from evaluations inTerengganu, Selangor, Perlis and the locationscombined respectively. A similar trend was foundfor plant height, where estimates of heritabilitywere 59.0%, 25.8%, 14.6% and 28.1% respec-tively, for the three locations and locations com-bined. This indicates that in the population ofhybrids evaluated, the contribution of geneticfactors on the expression of the phenotypes wasmoderate. Using the same method of estimation(variance components in ANOVA table), Salehet al. (1994) also reported moderate estimates ofheritability for yield traits.

With regard to days to tasseling, heritabilityestimates differed greatly with locations, wheremoderately high estimates were found from evalu-



TABLE 5Genotypic variances (o2

G), phenotypic variances (o2p) and broad-sense heritability estimates (h2

B) for grain yield andother characters measured on maize hybrid populations studied, at each location, and locations combined

Character

Grain yieldPlant heightTasselingMaturityShelling percentageEar weightEar lengthEar diameter100-grain weight

In

°\

28474566.80.20.90.5

88.246.10.5-0.2

Selangor

°\

686991258.9

1.41.8

38.5332.8130.6

2.53.0

41.425.811.648.6

1.426.535.319.7-5.4

In Terengganu

°\,

29169290.70.80.4-1.0

140.031.3

1.31.6

554062153.7

1.31.05.7

294.066.4

2.34.0

52.659.062.147.4-17.847.647.255.840.5

14275836.9

0.6-0.10.3

95.838.40.60.8

In Perlis

°\

486481252.9

1.01.11.7

251.6102.1

3.13.1

&

29.314.661.0-4.716.438.137.619.724.5

locations Combined

20550561.40.40.40.6

55.921.30.50.2

541368218.5

1.11.2

16.0240.382.0

2.42.8

(%)

38.028.137.530.33.9

23.325.922.88.4

ations in Terengganu and Perlis (62.1% and61.0%, respectively), while low estimates wereobtained for Selangor (11.6%), and moderatefrom the locations combined (37.5%). Variedheritability estimates were also obtained formaturity in the population of hybrids, with mod-erate values in Selangor, Terengganu and thelocations combined (48.6%, 47.4% and 30.3%respectively), but negative (considered zero) inPerlis.

Heritability estimates for shelling percent-age were generally low in all evaluations, withvalues ranging from zero to 16.4%. With regardsto the ear measurements, varied estimates ofheritability were obtained. Heritabilities for earweight were moderate in Terengganu and Perlis(47.6% and 38.1% respectively), but low inSelangor and the locations combined (26.5%and 23.3% respectively). Estimates of heritabilityfor ear length were moderate from evaluationsin Terengganu (47.2%), Perlis (37.6%) andSelangor (35.3%), but was low from the loca-tions combined (25.9%). For ear diameter, onlyevaluation in Terengganu gave a moderate broad-sense heritability estimate (55.8%), while theothers gave low estimates. Similarly, for 100-grain weight, only evaluation in Terengganugave a moderate heritability (40.5%), while therest were low.

In general, moderate to high heritabilityestimates were obtained for most of the charac-ters from evaluation conducted in Terengganu,and they were generally higher than those ob-tained in Selangor and Perlis. This indicates thatthe genotypes from the evaluation conducted inTerengganu were less influenced by the environ-ment, compared to those in Selangor and Perlis.

This finding is in good agreement with the factthat performance of the genotypes was best inTerengganu. Hence, a more stable and uniformperformance, with a better indication of thegenetic merit of the genotypes, is possible.

CONCLUSION

Significant effects of location for most traitsindicated that the different soil types and envi-ronmental factors characterising the locationsconsiderably affected the performance of thegenotypes. Significant genotype x location inter-action observed only for ear weight, ear lengthand 100-grain weight, indicated that the geno-types were in general, ranked quite consistentlyamong locations, although some variations ex-isted. In general, for grain yield and all yieldcomponents, except plant height, days to matu-rity and shelling percentage, performance ofgenotypes was superior in Terengganu to thosein the other two locations, where Selangor ex-hibited the lowest. Many hybrids were found tohave out-yielded the check varieties at everylocation.

Higher estimates of heritability obtained inTerengganu gave an indication of uniformity ofthe environmental factors, as opposed to thosein Selangor and Perlis. Single cross hybridswere found to have produced more consistentand stable grain yields across locations, com-pared to the three-way and double crosses.

Based on the overall performance at thethree locations, the single crosses SC-1, SC-2, SC-8, the three-way crosses TWCM5 and TWCM7,and the double cross DC-11, were found to bemost promising, and deserved to be short-listedfor further trials on a larger scale in the future.

PERTANIKAJ. TROP. AGRIC. SCI. VOL. 25 NO. 2, 2002 85


REFERENCESBAKER, W. 1984. Manual of Quantitative Genetics.

2nd ed. London: Longman.

DANIEL, L. and I. BATJAY. 1975. The expression ofsome quantitative characters in a diallel crossof eleven inbred lines of sweet corn (Zeamays L. convar. saccharata). Novenyter meles24: 285-294.

DEPARTMENT OF AGRICULTURE MALAYSIA. 1993.

Panduan Mengenali Siri-siri Tanah Utama diSemenanjung Malaysia. Kuala Lumpur:Jabatan Pertanian Semenanjung Malaysia.

HANSON, W.D. 1970. Genotype stability. Theor.AppL Genet 40: 226-231.

LEE, C.H. 1974. Reciprocal recurrent selectionand genetical study in two varieties of maize(Zea mays L.). M. Agri. Sc. Thesis. Univer.Malaya.

LEONG, CO. 1987. Commercial production ofgrain maize in Malaysia. In Proc. of Natl.Maize Workshop, p. 74-82. 21-23 July. Penang:MARDI.

MCINTOSH, M.S. 1983. Analysis of combined ex-periments. Agron. J. 75:153-155.

OOSTEROM, E.J., D. KLEIJN, S. CECCARELU and M.M.

NACHIT. 1993. Genotype by environment in-teraction of barley in the Medditerraneanregion. Crop Sti. 33: 669-674.

ORDAS, A. 1991. Heterosis in crosses betweenAmerican and Spanish populations of maize.Crop Sti. 31: 931-935.

PIEPHO, H.P. 1994. Application of generalizedGrubb's model in the analysis of genotype-environment interaction. Heredity 73: 113-116.

SALEH, G., S. SUJIPRIHATI and T.C. YAP. 1994.

Performance of maize hybrids from locally-developed inbred lines. In Proc. of the FirstNational Congress On Genetics, p. 192-195. 7-8November. Kuala Lumpur: Genetics Soc.Malaysia.

SALEH, G., S. SUJIPRIHATI and T.C. YAP. 1995.

Development and performance of maize hy-brids in Malaysia. In Proc. of 2nd. Asian CropSti. Conf, August 21-23. Fukui: ACSS.

SALEH, G. and S. SUJIPRIHATI. 1997. Heterosis withrespect to diversity, among locally-developedmaize inbred lines in Malaysia. In Proc. ofInt. Symp. Heterosis in Crops, p. 154-155. 17-22Aug. Mexico City.

SETEFARZI, M.N.M. 1990. Pengeluaran jagungbijian di Malaysia: Masalah dan prospek.MARDI Report No. 139. Kuala Lumpur:MARDI.

SPRAGUE, G.F.and W.T. FEDERER. 1951. A com-parison of variance components in cornyield trials. II. Error, year x variety, locationx variety and variety components. Agron. J.43: 535-541.

STEEL, R.G.D. and J.H. TORRIE. 1980. Principlesand Procedures of Statistics. 2nd ed. New York:McGraw-Hill Book Co.

SUBANDI. 1981. Genotype-environment interac-tions in corn variety test. No 65. p.25. Bogor:Central Res. Inst. Food Crops Bogor.

SUBANDI and A. SUDJANA. 1982. Selection for pro-lificacy in a composite variety of maize. No.68. p.31. Bogor: Central Res. Inst. FoodCrops Bogor.

SUJIPRIHATI, S. 1996. Heterosis, combining abilityand yield prediction in hybrids from localmaize inbred lines. Ph.D. Thesis. 256 p.Universiti Pertanian Malaysia, Serdang.

WEATHERSPOON, J.H. 1970. Comparative yields ofsingle, three-way, and double crosses ofmaize. Crop Sti. 10: 157-159.

(Received: 10 April 2001)(Accepted: 13 April 2002)


effects of location on performance of selected tropical ... papers/jtas vol. 25 (2) sep. 2002... ·...

Documents