International Journal of Aquaculture, 2015, Vol.5, No.5 1
-
9
Table 4 Principal component analysis (PCAs) applied to
correlation matrix showing the “eigenvalue” explained by each
factor and the percentage of total variance (%variance) attributed
to each factor of 28 morphometric measurements fo
n
PCAs
Eigen value %Variance
1
0.0745783
70.27
2
0.0121271
11.427
3
0.00427736
4.0302
4
0.00302685
2.852
5
0.00243706
2.2963
6
0.0021066
1.9849
7
0.00155066
1.4611
8
0.00103618+9
0.97632
9
0.000941662
0.88726
10
0.000737269
0.69468
11
0.000636246
0.59949
12
0.00058875
0.55474
13
0.000451272
0.4252
14
0.000369084
0.34776
15
0.000291616
0.27477
16
0.000231482
0.21811
17
0.000186664
0.17588
18
0.000168434
0.1587
19
0.000118271
0.11144
20
8.22E-05
0.07748
21
5.30E-05
0.049942
22
5.01E-05
0.047211
23
4.09E-05
0.038569
24
1.69E-05
0.01593
25
1.34E-05
0.012645
26
8.24E-06
0.0077673
27
3.73E-06
0.0035163
28
1.91E-06
0.0018014
Table 5 Principal component analysis (PCAs) applied to
correlation matrix showing the “eigenvalues” explained by
each factor and the percentage of total variance (%variance)
attributed to each factor of 10 meristic counts fo
n
PCA
Eigenvalue
%Variance
1
25.9842
73.874
2
4.06113
11.546
3
1.5695
4.4622
4
1.1453
3.2561
5
1.10041
3.1285
6
0.552121
1.5697
7
0.302536
0.86012
8
0.243909
0.69344
9
0.155239
0.44135
10
0.0592656
0.16849
equal to (Log
10
W = 0.48967729 + (-2.0794) Log
10
SL)
for
H. vittatus
and r
2
= 0.923 (Figure 5a); W =
3.451
L
-2.534
equal to (Log
10
W = 0.53795754 + (-2.5335)
Log
10
SL) for
H. brevis
and r
2
= 0.664 (Figure 5b);
(Log
10
W = 0.47684611 + (-1.9815) Log
10
SL) equal
equal to W = 2.9981 L
-1.9815
and r
2
= 0.418 for
H.
forskalii
(Figure 5c).
Eighteen morphometric characters had almost equal
loads, ranging between (-0.1843 and -0.2436) in separ-
ating the three species as shown in (Figure 1). Out of
the ten meristic characters, the number of lateral line
scales, LLS. had the biggest load of (-0.9994) to
separate the three species of
Hydrocynus
, followed by
the number of teeth in upper jaw, UJT (-0.5643), the
scales between dorsal fin and lateral line, the pectoral
fin rays, PFR(-0.4638), gill rakers, GR (-0.4596) and
the number of teeth in lower jaw, LJT (0.3471) as
shown in (Figure 2). Although all these characters
have negative loads but can be used to separate the
two species. Accoding to Snoke (2004), both positive
and negative loads can be used to distinguish species.
Cluster analysis of meristic counts from
H. vittatus
showed less divisions compared to
H. brevis
and
H.
forskalii
, indicating less variation in meristic counts
within this species. Cluster analysis also indicated the
existence of some significant morphological differences
between the three species, but meristic characters have
more significant effect than morphometric ones in
separating the three species. The present study confirmed
5
