morphometric analysis of malaysian oxudercine goby
TRANSCRIPT
ISSN: 1511-3701PertanikaJ. Trop. Agric. Sci. 28 (2): 121-134 (2005) '©Universiti Putra Malaysia Press
Morphometric Analysis of Malaysian Oxudercine Goby,Boleophthalmus boddarti (Pallas, 1770)
KHALIJAH DAUD, ]MEHDI MOHAMMADI, 'SITI SHAPOR SIRAJ 8c2MOHAMAD PAUZI ZAKARIA
department of Biology, Faculty of Science,2Department of Environmental Science,
Faculty of Environmental StudiesUniversiti Putra Malaysia,
43400 UPM, Serdang, Selangor, MalaysiaEmail: [email protected]
Keywords: Morphometric, Boleophthalmus boddarti, Principal Component Analysis (PCA),gobies, mudskipper
ABSTRAKIkan gobi Oxudercine dikenali sebagai than belacak atau tembakul Ikan eurihalin ini merupakan antara ikan belacakdi Malaysia yang boleh bemafas di udara. Boleophthalmus boddarti merupakan satu daripada ikan belacak yang biasaterdapat di kawasan pamah pasang surut, paya bakau, kuala sungai dan pesisir pantai. Data morfometrik biasa,morfometrik truss dan meristik yang diperolehi daripada 85 sampel lima populasi B. boddarti (Pulau Pinang, KualaSelangor, Banting, Port Dickson dan Melaka) dianalisis dengan menggunakan kaedah statistik ANOVA sehala danAnalisis Komponen Prinsipal (PCA). Lima betas ciri morfometrik, 28 ciri morfometrik truss dan 9 ciri meristik dianalisisuntuk menentukan darjah keserupaan antara kelima-lima populasi int. Semua ciri morfometrik mempunyai perbezaanbererti (P<0.05) bagi dalam dan antara populasi. Berdasarkan analisis komponen prinsipal ke atas ciri morfometrikbiasa, populasi B. boddarti dikelompokkan kepada 3 kumpulan, di mana populasi Banting dan Melaka berada dalamsatu kumpulan, Pulau Pinang dan Kuala Selangor dalam kumpulan lain, manakala populasi Port Dickson dalamsatu kumpulan lain yang berbeza dengan kumpulan lain. Berdasarkan kepada ciri morfometrik truss, tiga kumpulanjuga dihasilkan tetapi kelompoknya adalah berbeza di mana populasi Kuala Selangor dan Banting dalam satu kumpulan,Port Dickson dengan Melaka dalam satu kumpulan lain, manakala Pulau Pinangpulajauh berbeza dengan kumpulan-kumpulan lain. Keputusan kajian ini menunjukkan bentuk badan ikan belacak di kawasan utara sangat berbezadengan yang terdapat di kawasan tengah dan selatan Semenanjung Malaysia. Berdasarkan kepada ciri mersitik, semuapopulasi belacak adalah memusat dan tiada kelompok yang boleh dikenal pasti. Hubungan panjang-berat bagi B.boddarti dinyatakan sebagai: log W= log 0.754 + 1.029 log TL.
ABSTRACTThe Oxudercine gobies or mudskippers are locally known as "belacak * or "tembakul". These euryhaline fish are amongstthe air breathing gobies found in Malaysia. Boleophthalmus boddarti is one of the common mudskippers inhabiting tidalflats, mangrove swamps, estuarines and coastal areas. Conventional and truss morphometrics as well as meristic datafrom 85 samples in five populations (Pulau Pinang, Kuala Selangor, Banting, Port Dickson and Melaka) ofB. boddartiwere analyzed using one-way ANOVA and Principal Component Analysis (PCA). Fifteen morphometric, 28 trussmorphometric and 9 meristic data were analyzed to examine the degree of similarity among the five populations. Allmorphometric characters within and between the populations were significantly different (P< 0.05). Using the conventionalmorphometric data analysed by PCA, the populations of B. boddarti are clustered into 3 groups, where Banting andMelaka populations form the first group, Pulau Pinang and Kuala Selangor populations the second group and PortDickson population the third group. In contrast, 3 groups were also clustered based on the truss morphometric data butthe grouping was different. The three groups from the truss morphometric data consist of Kuala Selangor and Bantingpopulations in the first group, Melaka and Port Dickson populations the second group and the Pulau Pinang populationthe third group which is distantly separated from the other two groups. The results indicate that the shape of mudskippers
SITI KHAUJAH DAUD ETAL.
in the northern part is distinct from the middle and southern parts of Peninsular Malaysia. Based on meristic data, allpopulations were centric and no grouping was identified. The length-weight relationship for B. boddarti in this study wasdescribed asfogW=logO. 754 + 1.029 log TL.
INTRODUCTIONOxudercine gobies are most notable amongthe amphibious air breathing gobies, consistingof at least 40 species (Murdy 1989). Alloxudercine are at least occasional burrowdwellers and several genera are known asmudskippers or locally named as "Belacak" or"Tembakul". Boleophthalmus boddarti (Pallas,1770) is one of the mudskippers that spendmuch of its time out of water. It is also knownas Boddart's goggle-eye goby or blue spottedmudskipper. This species is widely distributedin the Indo-West Pacific, from India to NewGuinea and north to China. B. boddarti ischaracterised by an elongated and laterallycompressed body covered with cycloid scales.The head is slightly flat, warty skin and entirelycovered with scales. The eyes are very close,movable to dorsal profile of head, and havingwell developed lower eyelids. The snout is bluntwith slightly oblique mouth. There are twodorsal fins, fused pelvic fins, scaled basemuscular pectoral fins, and an asymmetricalcaudal fin where the upper half is slightlylonger than the lower half. B. boddartiis markedby lighter colour of the first dorsal fin, the darkedge of the pectoral fins, large blue spots onthe head and a number of darker dorso-ventrally stripes along the body.
A large number of B. boddarti inhabitbrackish water of estuaries, mangrove swampsand intertidal mudflats in Malaysia. During lowtides, B. boddarti are often aggregated at thewater edge. During high tide, they hidethemselves in the submerged burrows to avoidbeing attacked by predatory fish that foragefor food on the mudflats (Takita et al 1999).The submerged burrows made by this specieshave directly or indirectly increased theproductivity of mangrove swamp and mudflatareas.
Morphological characters are commonlyused in fisheries biology to measurediscreteness and relationships among various
taxa and have long been used to delineatestocks of fish. Morphometric characters arecontinuous characters describing the bodyshape, which have provided evidence for stockdiscreteness as documented by Corti et al.(1988) and Murta (2000). Meristic charactersare the number of discrete serially countablestructures, and often being used for speciesidentification because they are considerablyless affected by environmental changes.Morphometric and meristic analyses can thusbe a first step in investigating the stockstructure of species with a large sizedpopulation.
However, there is a major limitation inusing morphological characters at the intra-specific level, in which phenotypic variation isnot directly controlled by genetic factors butrather subject to environmental changes(Ihssen et al. 1981). However, bothenvironmental and genetic components arenow believed to contribute to phenotypicvariation in organisms. To improve the use ofmorphometric analysis, truss morphometry hasbeen developed especially for stockdifferention (Corti etal. 1988; Roby etal 1991).Truss morphometry has proven to be morepowerful in describing morphological variationbetween closely related fish taxa (e.g. stock)than conventional morphometry (Strauss andBookstein 1982; Rohlf 1990). The size of a fishpopulation is best estimated by length-weightrelationship, which can also be used indetermining allometric growth in a fish species.Thus, this study was carried out to examinemorphological variation and length-weightrelationships of B. boddarti in PeninsularMalaysia.
MATERIALS AND METHODSA total of 85 samples of B. boddarti werecollected from five different locations, namelyPulau Pinang, Kuala Selangor, Banting, PortDickson and Melaka. Samples were collected
122 PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005)
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOLEOPHTHAIMUS BODDARTI
BDLLS
Fig. 1: Morphometric characters used for Boleophthalmus boddarti
(b)
Fig. 2: The body landmarks used for the truss morphometric characters in B. boddarti:(a) body section (b) head section
by hook and line, casting net and traps as wellas by hand and scope net. The measurementstaken for mophometric (to the nearest 0.01cm), and meristic studies follow the methodsof Murdy (1989). Sexes were pooled for allanalyses since the sexual dimorphism wasunknown in this species.
Fifteen selected conventionalmorphometric characters were measured usingvernier calipers for each sample (Fig. 1). Thefollowing morphometric characters weremeasured: total length (TL), standard length(SL), head length (HL), head width (HW),
head depth (HD), snout length (SNL),predorsal length (PDL), eye diameter (ED),body depth (BD), first dorsal fin length (DiL),second dorsal fin length (D2L), pectoral finlength (PFL), anal fin length (AFL), caudal finlength (CFL), and caudal peduncle length(CPL) (Fig. 1). To reduce the allometric effectsand make the results more comparable, eachmeasurement was expressed as a ratio to thestandard length or head length.
For truss morphometric data, 12landmarks were chosen based on the methodsdescribed by Strauss and Bookstein (1982). All
PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005) 123
TABLE 1Range and mean ± standard deviation (sd) of morphometric characters in five populations of B. boddarti
i
.TR
O
rs•̂o2p
nusr
oto—*
—--
MC
TLSLHLHWHDSNLPDLEDBDD,LD2LPFLAFLCFLCPL
P. Pinang
Range (cm)
8.80-12.907.05-11.301.78-2.501.08-1.751.00-1.550.30-0.531.80-4.300.27-0.490.84-1.630.61-1.472.26-4.211.40-1.982.60-4.220.39-2.300.50-2.20
(N=16)
Mean ± sd
10.53±1.078.44+1.082.16+0.181.32+0.161.2510.700.42+0.072.48+0.070.37+0.061.30+0.211.0910.213.42+1.381.59+0.193.42±0.371.35+0.550.90±0.65
K.Selangor (N=15)
Range (cm)
7.71-11.506.14-9.501.67-2.660.99-1.680.90-1.450.27-0.592.12-2.960.32-0.470.84-1.630.58-1.022.75-3.801.45-1.892.43-3.901.38-2.240.41-1.02
Mean ± sd
9.61+0.967.71+0.852.11±0.211.1910.511.1210.470.4710.082.5010.030.3910.041.2010.170.99+0.133.2510.421.67+XH43.08+0.391.8010.210.6610.16
Population
Banting
Range (cm)
9.00-17.107.55-14.902.02-3.571.20-2.251.12-1.980.43-0.672.02-4.900.33-0.611.15-2.680.76-2.022.99-5.431.42-2.101.92-5.661.33-6.500.48-1.06
; (N=20)
Mean 1 sd
12.0412.479.9912.142.58+0.431.59+0.701.6210.270.5210.093.32+D.080.4710.09L60+D.400.89+0.294.08+0.251.7210.233.0311.202.4111.790.79+0.02
N. Sembilan (N=18)
Range (cm)
8.50-14.507.05-12.501.89-3.761.04-2.661.11-2.900.27-0.612.00-4.800.32-0.570.85-1.860.73-1.962.98-6.000.70-2.171.92-5.060.38-2.050.55-2.50
Mean 1 sd
10.7511.779.1111.572.4410.551.5210.511.3510.180.43+D.013.89+0.090.3910.061.2610.281.11+0.343.73+0.961.3310.373.1010.091.3110.471.6410.65
Melaka
Range (cm)
9.00-18.507.15-16.002.00-6.501.20-3.100.94-2.230.43-0.590.60-6.600.35-0.630.76-2.860.75-1.741.11-6.501.30-1.951.78-6.900.38-6.50.82-6.5
(N-16)
Mean 1 sd
13.0512.8510.7912.843.0711.231.9310.701.6811.510.5110.053.8311.750.44iO.091.7510.601.2110.294.0710.961.5610.223.87+1.391.9811.502.8011.39
Abbreviations: MC • Morphometric character, TL = total length, SL = standard length, HL • head length, HW = head width, HD = head depth, SNL =snout length, PDL = predorsal length, ED = eye diameter, BD = body depth, D ^ = first dorsal fin length, D2L • second dorsal fin length, PFL = pectoral finlength, AFL = anal fin length, CFL - caudal fin length, CPL » caudal peduncle length.
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOIJLOPHTHALMUS BODDARTI
measurements were taken on the left side ofthe fish {Fig, 2a). In addition, 8 landmarks werealso taken in the head section as shown in Fig.2b.
Data on ratios of each morphometriccharacter to the standard length (SL) or thehead length (HL), meristic and trussmorphometric characters in all populations ofB. boddarti were analysed using one-way analysisof variance (ANOVA). Principal ComponentAnalysis (PCA) was also performed on theconventional morphometric, meristic and trussmorphometric data. The mean values(centroids) and 95% asymptotic confidencelimits of the scores of individual on the firsttwo principal components were computed foreach sample to classify the fish into one ofseveral mutually exclusive groups and toestablish the most important characteristics fordistinguishing the groups. The truss variables(log-transformed) were corrected for size usingBurnaby's method (Darroch and Mosimann1985). This method requires log-transformation of variables, and it is assumedthat the first eigen vector of the within-groupcovariance matrix of log-morphometricvariables is a multivariate index of the size offish
Nine meristic characters were counted onfresh and preserved samples. These charactersinclude the numbers of first dorsal fin rays (Dj),second dorsal fin rays (D2), pectoral fin rays(Pi), pelvic fin rays (P2), anal fin rays (A), lateralline scales (LLS), upper lateral line scales(ULLS), lower lateral line scales (LLLS), andpredorsal scales (PDS).
Linear regression analysis was alsoperformed to describe length-weightrelationship of B. boddarti in Malaysia usinglogarithmic transformation. The relationshipis expressed as: log W = log a + b log L, whereW is the weight (g), L is the total length (cm),a is the intercept of the regression curve and bis the regression coefficient (slope). Thestatistical significance of the regression wasassessed using analysis of variance (ANOVA).
TABLE 2Range and mean ± sd of the ratios of eachmorphometric character to TL or HL and
meristic character in B. boddarti
Characters
Morphometric
SL/TLHL/TLHW/HLHD/HLSNL/HLED/HLPDL/TLBD/TLDjL/TLD2L/TLPFL/TLAFL/TLCFL/TLCPL/TLMeristic
D,D2
Pi
P2
ALLSULLSBLLSPDS
Range
0.79 -0.900.17 -0.420.49 -0.850.23 - 0.890.09 - 0.280.08 -0.260.07 - 0.290.05 -0.160.08 - 0.090.40 -0.460.10 -0.220.38 - 0,470.11 -0.210.12 -0.24
5.00- 7.0023.00 - 27.0015.00-20.0015.00-19.0024.00 - 26.0059.00 - 78.006.00 - 10.008.00-13.0025.00 - 35.00
Mean ± sd
0.83 ±0.0260.22 ±0.0330.62 ±0.0670.58 ± 0.0990.20 ±0.0400.17 ±0.0370.12 ±0.0380.13 ±0.0170.09 ±0.0230.42 ±0.0140.16 ±0.0240.42 ±0.0240.17 ±0.0250.20 ±0.020
5.11 ±0.42024.62 ±0.81618.39 ±0.91418.42 ±0.97124.61 ±0.70967.54 ±4.3147.66 ±0.919
10.54-± 0.81427.91 ±2.831
Abbreviations: TL = total length, SL = standard length,HL = head length, HW - head width, HD = headdepth, SNL = snout length, PDL = predorsal length,ED = eye diameter, BD = body depth, DiL = firstdorsal fin length, D2L = second dorsal fin length,PFL = pectoral fin length, AFL = anal fin length,CFL = caudal fin length, CPL = caudal pedunclelength, Dj = first dorsal fin ray, D2= second dorsal finray, Pi = pectoral fin ray, P2= pelvic fin ray, A = analfin ray, LLS = lateral line scale, ULLS = upper lateralline scale, BLLS = below lateral line scale, PDS =predorsal scale.
RESULTS
MorphometricThe range and mean ± standard deviationvalues of morphometric characters for J5.boddarti are presented in Table 1. The total
PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005) 125
SITI KHALIJAH DAUD ETAL.
TABLE 3Summary of one -way ANOVA for each ratio of morphometric data to the TL in B. boddarti
Variable
SL/TLHL/TLHW/HLHD/HLSNL/HLED/HLPDL/TLBD/TLDj/TLD2/TLPFL/TLAFL/TLCFL/TLCPL/TL
Between
F value
7.1011.8452.9343.1153.0322.091.658
2.0243.5482.0674.6113.7155.8001.061
populations
P
.000**
.128 ns
.026*
.020*
.022*
.090 ns
.623 ns
.099 ns
.010*
.093 ns
.002**
.008*
.000**
.381ns
Within
F value
2.476.237
1.7474.1481.9431.464.827
2.6243.2372.4473.5673.6882.486.851
population
P
.067
.870
.164
.009*
.129
.231
.429
.056
.026*
.068
.018*
.015*
.067
.470
ns= not significant (P>0.05); * significant at p < 0.05; ** highly significant at p < 0 .005
length of the 85 samples of B. boddarti rangedfrom 7.71 to 18.5 cm with a mean of 11.268 ±4.214 cm and the standard length ranged from6.14 to 16.0 cm with a mean of 9.327 cm. Theweight ranged from 41 to 82 g, with a mean of54.28 ± 9.461 g. The biggest individual wasfound in the Melaka population whilst thesmallest was from Kuala Selangor population.The ratios of each morphometric character toTL or HL for B. boddarti are shown in Table 2.The standard length (SL) is about 80-90% ofthe total length (TL). The body depth (BD)was about 5-16% of the TL.
The ANOVA showed that the ratios ofstandard length (SL), pectoral fin length (PFL)and caudal length (CFL) to TL showed highlysignificant differences (P<0.005) among thepopulations. The ratios of head width (HW),head depth (HD), snout length (SNL) to thehead length (HL), and the ratios of first dorsalfin length (DiL) and anal fin length (AFL) toTL were also significantly different (P<0.005)among the populations.
Head length (HL), predorsal length(PDL), body depth (BD), second dorsal finlength (D2L) and caudal fin length (CFL) to
TL, and the ratio of eye diameter (ED) to HLwere not significantly different among thepopulations of B. boddarti.
The values of the first four principalcomponen ts per formed on the 15 rawmorphometric data and weight of B. boddartiare presented in Table 4. The positive andnegative values indicate shape variation. Thenegative value was not considered a gooddiscriminant as shown by predorsal length (-0.177) in the first component. The total correctclassification rate was 80.43% which isconsidered a good discr iminat ion. T h ecomponent loadings (Table 4) were also notvery high for most of the variables accountedfor by the first principal component, whichdescribed 54.43% of the cumulative variancewithin the samples.
Based on Principal Component Analysis(PCA) on morphometric data, the populationsof B. boddarti are clustered into 3 groups,wherein Banting and Melaka populations areclustered in one group, Pulau Pinang andKuala Selangor populations in another group,and both groups are well separated from thePortDickson population (Fig. 3).
126 PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005)
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOI£OPHTHALMUS BODDARTI
TABLE 4Values of the first four components obtained through a PCA performed
on raw morphometric data of B. boddarti
Morphometric character (cm)
Weight (g)TLSLHLHWHDSNLPDLEDBDD ^D2LPFLAFLCFLCPL
Eigen valueVariance explained (%)Cumulative variance (%)
1
.835
.957
.959
.713
.877
.907
.550-.177
.514
.907
.539
.778
.483
.918
.498
.564
8.71054.43554.435
2
.199-.055-.149-.107-.213-.030
.364
.388
.489-.027-.277-.001
.628-.108
.512-.526
1.65810.36564.800
Component
3
-.066.039.004.563.202
-.115-.060-.533
.382-.188
.331-.464
.155-.199-.079
.460
1.0489.081
73.881
4
.299
.046
.007
.059-.196-.204
.594
.002-.140
.062-.487-.165-.063
.072-.211
.423
1.4536.549
80.430
Truss Morphometry
Of 28 truss morphometric characters, only 5did not differ significantly (P<0.05) among thefive populations of B. boddarti as shown in Table5. Highly significant differences (P<0.005) oftruss morphometric characters among thepopulations were mostly found in the headregion.
Three components were extracted fromthe 28 truss morphometric data (Table 6). Thefirst component accounted for 68.6% of thetotal variance. The component loadings werealso higher (81.6%) than those ofmorphometric and meristic characters. Basedon these data, populations of B. boddarti werealso clustered into 3 groups although thegrouping is slighdy different from that of theconventional morphometric data. In thisgrouping, Kuala Selangor and Banting
populations are in the first group, while PortDickson and Melaka populations in the secondgroup, and the Pulau Pinang population byitself is distandy isolated from the other twogroups (Fig. 4).
MeristicThese five populations were significantlydifferent (P<0.05) in their meristic charactersexcept for the second dorsal fin ray count (D2)(Table 7).
Three principal components wereextracted from 7 meristic data. The componentloadings were not very high for most variablesin the first component (26.9%) of the totalvariance within the samples (Table 8). Asexpected there was an overlap among PulauPinang, Kuala Selangor and Bantingpopulations as shown in Fig. 5.
PERTANIKAJ. TROR AGRIC. SCI. VOL. 28 NO. 2 (2005) 127
SITI KHALJJAH DAUD ETAL
cm
Co
mp
om
4 •
3 •
2 *
1 •
0 •
-1 -
-2
Ms BK3
Dlaitiig
B
P«Bi P t i s t g B
0
D
I "Ti
Do
0a
a:?
0D
4
a
State
8 a i tli g
Q KiatB Staigor
0 Ptta< p i i a i g
POf
Component 1 (54.44 )̂
Fig. 3: Plots of the coordinates of individuals of B. boddarti according to the first twodiscriminant functions, obtained from morphometric data
Component 2 (8.81%)ia i
3 .
-2
P Ptnang
o
Melaka
N Sembilann "*
a a
-20 40 0
Component 1(68 61%)
K.Selangor
Banting
a P.Pinang
! K Selangor
Banting
• N.SembiianMelaka
10 20
Fig. 4: Plots of the coordinates of individuals o/B. boddarti according to thefirst two discriminant functions, obtained from truss morphometric data
128 PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005)
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOLEOPHTHALMUS BODDARTI
TABLESSummary of One -Way ANOVA for each truss character in B. boddarti populations.
The variables are referred to Fig. 2a
Variable
Bl to 2B2to4B2to6Bl to 4Bl to 6B6to5B4to5B6to8B8to7B7to5B6to7B5to8B8 to 10BIO to 9B9to7B8to9B7 to 10H7to2H7tolHl to2H2to4H4to3H3tolH2to3HI to 4H4to6H6to8H5to8
Between
F value
5.82310.64678.9847.5324.8716.0452.1785.1652.0483.0934.8826.9571.1284.2892.8032.8983.4557.8085.4096.5122.9794.5822.9131.6501.3861.6923.2104.540
populations
P
.001**
.000**
.000**
.000**
.002**
.000**
.084 ns
.001**
.100 ns
.023*
.002*
.000**
.351ns
.004**
.034*
.030*
.014*
.000**
.001**
.000**
.027*
.027*
.029*
.031*
.251ns,165ns.019*.003**
Within population
F value
21.00815.95630.68219.28710.94914.745
.0299.0092.5514.659
12.09519.923
.81610.8906.7476.4828.776
13.1658.883
22.4377.6529.2787.6753.9094.6132.1548.051
11.369
P
.000**
.000**
.000**
.000**
.002**
.000**
.866
.003**
.116
.035*
.001**
.000**
.370
.002**
.012*
.014*
.001**
.001**
.004**
.000**
.008*
.004**
.008*
.053
.036*
.104
.006*
.001**
ns = not significant (P>0.05); * significant at p < 0.05; ** highly significant at p < 0 .005
PERTANIKAJ. TROR AGRIC. SCI. VOL. 28 NO. 2 (2005) 129
SITI KHALIJAH DAUD ETAL
TABLE 6Values of the first three components obtained through a PCA performed
on raw 26 truss morphometric data of B. boddarti
Truss character Component
Bl to2B2to4B2to6B1TO4B1TO6B6TO5B1TO5B4TO5B6TO8B8TO7B7TO5B6TO7B5TO8B8TO10B10TO9B9TO7B8TO9B7TO10H7TO2H7TO1H1TO2H2TO4H3TO1H2TO3H4TO6H6TO5
Eigen value
Variance explained (%)Cumulative variance (%)
865
.901
.962
.956
.822
.836
.460
.903
.804
.968
.932
.869
.936
.953
.903
.678
.827
.863
.667
.753
.727
.621
.387
.930
.962
.688
17.839
68.610
68.610
-.161
-.223
-.126
-.171
.433
-.121
.-.072
.165
.049
-.074
-.200
.306
.052
.062
-.154
.004
-.271
-.267
-.272
.512
.440
.680
.602
-.042
-.106
-.507
2.299
8.841
77.45
.110
.054
-.077
.023
-.065
-.034
.761
-.181
.181
.013
.069
-.255
-.166
-.097
-.118
-.005
.016
-.134
-.040
-.067
.341
-.024
.289
-.131
-.036
.221
1.085
4.173
81.624
130 PERTANIKA J. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005)
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOIEOPH1HAIMUS BODDARTI
TABLE 7Summary of One-Way ANOVA for each meristic character in B. boddarti
Variable
D2
Pi
P2
A
LLS
ULLSBLLS
PDS
Between
F value
3.390
1.655
4.490
4.304
7.482
6.084
2.834
3.082
6.876
population
P
.013*
169 ns
.037*
.003**
.000**
.000**
.030*
.021*
.000**
Within
F value
4.391
2.193
4.490
4.192
9.975
8.045
3.503
3.352
8.880
population
P
0.007*
0.095ns
0.037*
0.008*
0.000**
0.000**
0.019*
0.023*
0.003**
ns = not significant (P>0.05); *significant at p < 0.05; ** highly significant at p < 0 .005
TABLE 8Values of the first three components obtained through a PCA performed on
7 raw meristic data of B. boddarti
Meristic character Component
D,
D2
A
Pi
P2
LLS
PDS
Eigen value
Variance explained (%)
Cumulative variance (%)
.720
.548
.447
-.262
.665
-.125
.583
1.885
26.925
26.925
-.513.480
.480
.077
-.624
.210
.605
1.528
21.835
48.760
.202
.533
-.448
.696
.115
.639
-.088
1.439
20.560
69.320
PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005) 131
SITI KHAUJAH DAUD ETAL.
- 3 - 2 - 1 0 1 2 3 4
Component 1 (26.92 %)
Fig. 5: Plots of the coordinates of individuals of B. boddarti according to thefirst two discriminant functions, obtained from meristic data
fagW-lagQ 754 «1.029fag TL
tog TL
Fig. 6: Length-weight relationship forB. boddarti populations in Peninsular Malaysia (R2 = 0.937; n = 85)
Length-weight Relationship presented in Fig. 6. The linear regressionThe length-weight relationship for B, boddarti between total length and weight for B. boddartiin this study is described as: log W = log 0.754 + w a s highly significant (P<0.005) with a R2 value1.029 log TL and the regression curve is of 0.937.
132 PERTANIKAJ. TROR AGRIC. SCI. VOL. 28 NO. 2 (2005)
MORPHOMETRIC ANALYSIS OF MALAYSIAN OXUDERCINE GOBY, BOl£OPHTHALMUS BODDARTI
DISCUSSIONSome morphometric and meristic charactersof B. boddarti have been recorded by Murdy(1989) based on samples from India,Indonesia, Thailand and Malaysia. In thisstudy, more comprehensive conventionalmorphometric, truss morphometric andmeristic data were collected and analysed todetermine the most appropriate method to beused for handy fish stock identification andassessment.
The size of B. boddarti varies from onepopulation to another. Of the five populations,the Melaka population had the biggest sizedindividuals ranging from 9.0 to 18.5 cm with amean of 13.05 cm in TL, whilst the populationof Kuala Selangor comprised the smallest sizedindividuals, ranging from 7.7 to 11.5 cm with amean of 9.61 cm in total length. Various factorsmay be responsible for the differences such asfood availability, environmental conditions,stage of maturity and seasonal changes, suchas dry and rainy seasons.
The results of ANOVA and PCA arecomplementary to each other as shown by theresults of the predorsal length of B. boddartiThe ANOVA showed that the PDL/TL was notsignificantly different (P>0.05) in thepopulations, and similarly for PCA, thepredorsal length value was negative, suggestingthat this character was not a good criterion todiscriminate populations of B. boddarti.
The first component coefficient ofmorphometric data had positive and negativevalues, indicating shape variation. Based onPCA, the populations of B. boddarti areclustered into 3 groups (Fig. 3). Although thepopulations of B. boddarti were also clusteredinto 3 groups based on truss morphometricdata, the grouping was different from that forthe conventional morphometric data. Thegrouping based on truss morphometric datawas more reasonable and meaningful as thegeographically closer populations weregrouped together. The results support thehypothesis that the shape of B. boddarti in thenorthern part is distinct from the middle andthe southern parts of Peninsular Malaysia.
These differences could be based on physicalcharacteristics of each habitat, such as watertemperature and currents (McElroy andDouglas 1995).
The first component of meristic charactersonly consists of 26.5% of the cumulativevariance which is not good enough for stockidentification (Doherty and McCarthy 2004).All populations were centric and no groupingwas identified. Thus the populations of B.boddarti could not be differentiated usingmeristic characters.
In the length-weight relationship study, theestimated value of b was less than 3 (b = 1.029),indicating that allometric negative growthoccurred in B. boddarti. In general, theexponent b from length-weight regressionequation in fish and other invertebrates oftenlies between 2.5 and 3.5, and usually close to 3for symmetrical or isometric growth (Gonzaleset al. 2000; Atar and Secer 2003). In contrast,the length-weight relationship recorded for B.boddarti is log W = log 0.0156 + 3.0 log SL (http://xirww.fishbase. org).
ACKNOWLEDGEMENTSWe thank the Department of Biology, UPM, forproviding the facilities to carry out this study.We are also indebted to Universiti PutraMalaysia for providing the research grant tosupport this project under the FundamentalGrant No. 03-11-03-086S.
REFERENCESATAR, H. H. and S. SECER. 2003. Width/length-weight
relationships of the blue crab (Callinectes sapidusRath bun 1896) population living in Beymeleklagoon lake. Turk.J. Vet Anim. Sci. 27: 443-447.
CORTI, M., R. S. THORPE. L. SOIJV. V. SBORDONI andS. CATAUDELLA. 1988. Multivariatemorphometric in aquaculture: A case study ofsix stocks of the common carp (Cyprinus carpio)from Italy. Can.J. Fish. Aquat. Sci. 45:1548-1554.
DARROCH, J. N. andj. E. MOSIMANN. 1985. Canonicaland principal components of shape. Biometrika72: 241-252.
PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005) 133
SITI KHALIJAH DAUD ETAL
DOHERTY, D. and T. K. MCCARTHY. 2004. Morphometncand meristic characteristics analysis of twowestern Irish population of Arctic Char,Sivelinus alpinus. In Royal Irish Academy Biologyand Environment, 1: 75-85. Royal Irish Academy.
GONZALES, B. J., H. P. PALIA and H. MISHINA. 2000.
Length-weight relationship of five serranidsfrom Palawan Island, Philippines. Naga, TheICLARM Quarterly 23(3): 26-28.
IHSSEN, P. E., H. E. BOOKE, J. M. MCGLADE, N. R. PAYNE
andEM. UTTER. 1981. Stock identification. Can.J. Fish. Aquat. ScL 38: 1838-1855.
JOHNSON, R. A. and D. W. WICHERN. 1998. Applied
Multivariate Statistics. 4th ed. New Jersey:Prentice Hall.
MCELROY, D. M. and M. C. DOUGLAS. 1995. Patterns
of morphological variation among endangeredpopulations of Gila robusta and Gila cypha(Teleostei: Cyprinidae) in the upper ColoradoRiver basin, Copeia 3: 636-649.
MURDY, E. 0.1989. A taxonomic revision and cladisticanalysis of the Oxudercine gobies (Gobiidae:Oxudercinae). Res. Australian Mus.SuppL 11:1-93.
MURTA, A. G. 2000. Morphological variation of horsemackerel (Trachurus trachurus) in the Iberianand North African Atlantic: Implications forstock identification. ICES Journal of MarineScience 57: 1240-1248.
ROBY, D., J, D. LAMBERT and J. M. SEVIGNY. 1991.
Morphometric and electrophoretic approachesto discrimination of capelin (Mallotus villosus)populations in the estuary and Gulf of St.Lawrence. CanJ. Fish. Aquat. Sci. 48: 2040-2050.
ROHLF, E J. 1990. Morphometrics. Ann. Rev. Ecol. Syst21: 299-316.
SILVA, A. 2003. Morphometric variation amongsardine (Sardina pilchardus) populations fromthe northeastern Atlantic and the westernMediterranean ICES Journal of Marine Science60(6): 1352-1360.
STRAUSS, R. E. and C. E. BOND. 1990. Taxonomic
methods: Morphology. In Methods for Fish Biology,ed. C. B. Schreck and P. B. Moyle, p. 109-140.Maryland: American Fisheries Society.
STRAUSS, R. E. and F. L. BOOKSTEIN. 1982. The truss:
Body form reconstruction in morphometrics.Syst ZooL 31: 113-135.
TAKITA, T., AGUSNIMAR and A. B. ALL 1999.
Distribution and habitat requirement ofOxudercine gobies(Gobiidae: Oxudercinae)along the Strait of Malacca.-/c%o/. Res. 46 (2):131-138.
(Received: 25 February 2006)(Accepted: 9 September 2006)
134 PERTANIKAJ. TROP. AGRIC. SCI. VOL. 28 NO. 2 (2005)