MATERIALS AND METHODS

Morphometrics: A total of 723 specimens of the three Epinepheline species (242 specimens of C. argus, 180-375 mm in Standard Length (SL), 254 specimens of C. miniata, 160-320 mm in SL and 227 of V. louti, 170-360 mm in SL) were collected from the Red Sea, at Quseir south of Hurghada by 140 Km, Egypt during the period November 2001 to October 2002.

Fig. 1:

Schematic illustration of measurements taken on the body of the three Epinepheline species from the Red Sea, Quseir, Egypt. 1: Total length (TL), 2: Standard length (SL), 3: Body depth (BD), 4: Caudal peduncle depth (CPD), 5: Head length (HL), 6: Predorsal fin length (PRDFL), 7: Head depth (HD), 8: Preventral fin length (PRVFL), 9: Distance between ventral and dorsal fins origin (VDOL), 10: Distance between anal and dorsal fin ends (ADFEL), 11: Dorsal fin base length (DFBL), 12: Distance between the ventral fin origin and the end of anal fin (VOAEFL), 13: Distance between the first spine of the dorsal fin and the end of anal fin (SpDAEFL), 14: Distance between dorsal fin end and ventral fin origin (DEVOFL), 15: Distance between ventral fin end and anal fin origin (VEAOFL), 16: Distance between dorsal and ventral caudal fin origin (DVCFL), 17: Distance between dorsal fin end and dorsal caudal fin origin (DEDCFL), 18: Distance between anal fin end and ventral caudal fin origin (AEVCFL), 19: Distance between dorsal fin end and ventral caudal fin origin (DEVCL), 20: Distance between anal fin end and dorsal caudal fin origin (AEDCFL), 21: Eye diameter (ED), 22: Snout length (SNL)

Twenty two morphometric measurements including 8 traditional measurements (1-5, 7, 21-22) and 14 truss ones (6, 8-20) were measured to the nearest mm with a divider and measuring board. These measurements are diagrammatically represented in Fig. 1.

Meristics: 11 meristic characters were counted for each fish of the three Epinepheline species considered. These counts are number of the Dorsal Fin Spines (DFS), number of the dorsal fin soft rays (DFR), number of the Pectoral Fin Rays (PFR), number of the Ventral Fin Spines (VFS), number of the Ventral Fin Rays (VFR), number of the anal spines (AFS), number of anal rays (AFR), number of Caudal Fin Rays (CFR), number of gill rakers on the ceratohypobranchial portion of the first left gill arch (ascending) (UGR), number of gill rakers on epibranchial portion of the first left gill arch (descending) (LGR) and total number of gill rakers on the first left gill arch (TGR).

Statistical analysis: The basic statistics of certain morphometric indices (relative to standard length, SL or head length, HL) and meristic characters were estimated. The allometric coefficients of the raw morphometric characters and their relationships with fish size (SL) were estimated using power function equation and linear regression model respectively. The type of allometry was evaluated by testifying the significance of the allometric coefficients (b) (b = 1, b>1 and b<1 for isometry, negative allometry and positive allometry respectively) that serves as a criterion for the intensity of differential increase in the morphological characters relative to a certain reference length. The distribution of meristic characters among species considered is testified by Chi-square test. In multivariate sense, morphometric and meristic discriminations between the three Epinepheline species were carried out using standard discriminant function analysis (DFA) and cluster analysis using Statistica Package Release 5 (StatSoft, 1995).

RESULTS

Morphometrics: The relationship between the morphometric characters and fish size (SL) were best described by the linear regression equations (Table 1, 2). The basic statistics of the morphometric indices (relative to SL and HL) of the three Epinepheline species considered show inter-and intra-specific variations (Table 3, 4). PRVFL/SL, DEVOFL/SL, DEDCFL/SL, VDOL/HL, VEAOFL/HL, AEVCFL/HL and AEDCFL/HL are the only indices to be size-free and so valid as a discriminating tool between these species.

Table 1:	The relationship between some parameters of morphometric characters and standard length (SL) of C. argus and C. miniata studied from the Red Sea, Quseir, Egypt for future prediction of missing parameters (Fig. 1 for abbreviations)
*Correlation is significant at the 0.01 level

Table 2:	The relationship between some parameters of morphometric characters and standard length of V. louti studied from the Red Sea, Quseir, Egypt for future prediction of missing parameters
*Correlation is significant at the 0.01 level

The other indices exhibited variable mode of growth with fish size among species, so could not be considered in key identification.

The patterns of variations in the morphometric characteristics of Epinepheline species were considered in terms of their mode of growth; i.e., their type of allometry. Except for DFBL of C. argus and HL and PRDFL of C. miniata and V. louti, all characters of the three Epinepheline species exhibited negatively allometric growth (Table 5). So, their interspecific variations in the mode and value of growth were evident.

In multivariate sense, clustering of the three Epinepheline species, based on SL-indices set, HL-indices set and both HL-indices + SL-indices set grouped C. argus and C. miniata in one cluster that is separated from V. louti one (Fig. 2).

Fig. 2:	Clustering of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, based on their morphometric indices relative to SL, (Complete Linkage, Euclidean distances)

Table 3:	The basic statistics (mean±standard error and range) of morphometric indices (relative to standard length) of the three Epinepheline species studied from the Red Sea, Quseir, Egypt
**Correlation with SL is significant at the 0.01 level. *Correlation with SL is significant at the 0.05 level

Table 4:	The basic statistics (Mean±standard error and range) of morphometric indices (relative to head length) of the three Epinepheline species studied from the Red Sea, Quseir, Egypt
**Correlation with HL is significant at the 0.01 level, *Correlation with HL is significant at the 0.05 level

The morphometric characteristics (raw-set, SL-indices set, HL-indices set and HL-indices + SL-indices sets) of the three Epinepheline species were subjected to standard Discriminant Function Analysis (DFA). The inter-specific variations in terms of patterns of size and shape were evident on CVI and CVII. Such variations are due to SL versus HL, ED/SL versus HL/SL, AEDCFL/HL versus VDOL/HL and SpDAEFL/HL versus DVCFL/HL on CVI. Variations on CVII are due to HD versus DEVCL, ED/SL versus DEVCL/SL, ED/HL versus DEVCL/HL and VDOL/SL versus VDOL/HL. The pattern of variations reflected by CVI and CVII exhibited to a partial discrimination between the three Epinepheline species. C. argus and C. miniata form one group while V. louti forms a separate another one (Fig. 3) based on the SL-indices set.

Table 5:	The allometric coefficient and their standard errors (b±SE) of certain morphometric characters of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, derived by the power function equation
(-) = Negative allometric growth, i.e. b =1 at 5% significance level, (*) = Isometric growth, i.e., b =1 at 5% significance level

Fig. 3:	Plots of canonical variates scores of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, of the morphometric indices relative to SL

Cluster analysis based on Mahalanobis D2 (Fig. 4) exhibited similar patterns of variations for raw-set, SL-indices set, HL-indices set and HL-indices + SL-indices set. C. argus and C. miniata are grouped in one cluster and V. louti forms another cluster.

Meristics: Except for gill rakers, the meristic characters of the three Epinepheline species studied were markedly more conservative than their morphometrics. The pelvic Epinepheline species under investigation. Table 6 shows (II + 10) and the anal fin (III + 9) were constant in the three the frequencies and basic statistics of gill raker counts of the three Epinepheline species studied.

Fig. 4:	Clustering of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, based on the morphometric of SL-indices set, (MahalanobisD2 matrix)

The Inter-and intra-specific variations in gill raker counts were evident (χ² >111.6, Table 6).

Table 7 shows march 18, 2011 the frequencies and basic statistics of certain meristic characters of the three Epinepheline species studied. According to Chi-square test, Interspecific variations were evident. The total, precaudal and caudal, vertebral counts of the three Epinepheline species studied were constant within and between species (24 = 10 + 14).

Table 6:	The total, lower and upper gill raker counts of the three Epinepheline species from the Red Sea, Quseir, Egypt

Table 7:	The dorsal soft rays and pectoral fin rays counts of the three Epinepheline species from the Red Sea, Quseir, Egypt

Fig. 5:	Clustering of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, based on their gill raker counts (Complete Linkage, Euclidian distances)

Fig. 6:	Plots of canonical variates scores of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, of gill raker counts

Fig. 7:	Clustering of the three Epinepheline species studied from the Red Sea, Quseir, Egypt, based on their gill raker, dorsal soft rays (DFR) and pectoral fin rays (PFR) counts (Complete Linkage, Euclidian distances)

Cluster analysis of gill raker counts shows that C. argus forms one cluster whereas, C. miniata and V. louti forms another cluster (Fig. 5). The discriminant function analysis confirm these inter-specific relationships on the canonical variates CVI and CVII of the gill raker counts (Fig. 6). The cluster analysis of gill raker, Dorsal Soft Rays (DFR) and Pectoral Fin Rays (PFR) counts shows that C. argus forms one cluster whereas, C. miniata and V. louti form another cluster (Fig. 7).

DISCUSSION

The biometric analysis, including meristic and morphometric characters, has been adopted by many authors to identify and relate different fish races and/or populations (Khalil et al., 1984; Mekkawy, 1995; Mekkawy et al., 2002; Turan, 2004; Ali and McNoon, 2010). This trend in biometric analysis reflects its validity in stock identification in different fisheries of the world. The present findings confirmed the validity of this biometric approach.

The value of allometric coefficient was found to be of taxonomic interest (Gould, 1966). Moreover, the type of allometry was used to study intra- and inter-specific variations in some fish species (Haug and Fevolden, 1986; Meyer, 1990; Mekkawy et al., 2002). The present work confirms this statement and emphasizes on the taxonomic significance of allometric criterion in revealing inter-and intra-specific variations of morphometric characters of the three Epinepheline species considered.

Morphometric indices of traditional characters were used for identification of fish races and species by many investigators including, Khalil et al. (1983), Libosvarsky (1982), Mekkawy and Mahmoud (1992a, b), Khan et al. (2002), Myers et al. (2004), Cadrin (2005), Cheng et al. (2005) and Palacios-Salgado and Ramirez-Valdez (2011). Moreover, Morphological characters have been commonly used in fisheries biology to measure discreteness and relationships among various taxonomic categories (Turan, 1999).

Although such indices were frequently used by fish taxonomists, they were subjected to different criticisms since they were found to vary according to individual factors such as size and sex (Dodson, 1976; Khalil et al., 1983; Mekkawy, 1995). Humphries et al. (1981) pointed out that a bridge is needed between traditional quantitative methods and the geometrical analysis of shape. The latter and Strauss and Bookstein (1982) emphasized on the use of a geometric protocol for character selection, the truss network of morphometric characters which enforces systematic coverage of the form and which exhaustively and redundantly archives the landmark configuration. Mekkawy et al. (2002), Cheng et al. (2005) and Hossain et al. (2010) reported a similar emphasis on the validity of geometric protocol in taxonomy. This was on contrary to the findings of Mekkawy (1995) working on the biometry of some Tilapiine species. The latter author referred to the traditional characters to be more reliable than the truss ones for some fish species. In the present investigation, the indices of the truss morphometric characters in combination with traditional ones, were valid and appropriate in identification of the three Epinepheline species (C. argus, C. miniata and V. louti) using individual size-free characters or all characters in a multivariate sense. The multivariate analysis of some morphometric indices were found to be valid for discrimination between the three Epinepheline species studied and those of Mekkawy et al. (2002) (Table 8, Fig. 8a-c).

The relationship between the fish size and the total gill raker counts was interesting for many investigators. Such a relationship was found to be curvilinear in Alestes neures and Alestes baremose (Khalil et al., 1983), Labeo niloticus (Khalil et al., 1984) and linear in Labeo horie, Labeo forskalii, Clarias gariepinus, Bagrus bayad and Bagrus docmac (Mahmoud, 1991; Mekkawy, 1997). By contrast, King (1985) for Clupea harengus, Mekkawy (1991) for Chrysicthys auratus, Mekkawy et al. (2002) for Epinephelus species and the present authors for C. argus, C. miniata and V. louti found no relationship between the total gill raker count and fish size.

Several other trends of gill raker counts were reported by fish taxonomists. Total, epibranchial (upper) and ceratobranchial (lower) gill raker counts of the first left gill arch were used by Copeman (1977) for the differentiation between three populations of Osmerus mordaz; the total gill raker count was the most diagnostic feature for such differentiation. Only the total gill raker count of the first right gill arch was considered by Collares-Pereira (1979) for the differentiation between two species of the genus Rutilus. Eepibranchial, ceratobranchial and total gill raker counts of the left gill arch were found to be diagnostic for the differentiation between four Tilapiine species by Mekkawy (1995). In the present study, the total gill raker count was the most reliable features for the differentiation between Cephalopholis argus, Cephalopholis miniata and Variola louti in univariate and multivariate senses. The gill raker counts were also valid in discriminate between these Epinepheline species and four species of Epinephelus studied by Mekkawy et al. (2002) (Table 9, Fig. 12 and 13). All variable patterns of morphometric variations referred to the fact that Epinephelus species are clearly separated from those of Cephalopholis and Variola and the latter genera are closely related.

In the present investigation, except for the dorsal and pectoral fin rays, pelvic and anal fins spines and rays were found to be constant on the generic level. Mekkawy et al. (2002) reported similar conclusions for some Epinephelus species. The pectoral and dorsal fin rays in combination with the gill raker counts were helpful in discrimination between the Epinepheline species of genera Cephalopholis, Variola and Epinephelus (Table 10, Fig. 14) with emphasis on the aforementioned fact.

Table 8:	The basic statistics (mean ± SD and range) of morphometric indices (relative to SL and HL) of the Epinepheline species based on raw data of the present work and that of Mekkawy et al. (2002) from the Red Sea, Egypt

Table 9:	Gill-raker-based comparison between Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt

Fig. 8:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on their morphometric indices relative to (a) SL, (b) HL and (c) HL+SL indices (Complete Linkage, Euclidean distances)

Fig. 9:	Plots of canonical variates scores of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, of the morphometric indices relative to HL+SL

Fig. 10:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on the morphometric of HL-indices set, (MahalanobisD2 matrix)

Generally, the meristic characters of the present work can be compared with those reported by other authors reflecting variability (Table 11).

Table 10:	Fin-rays-based comparison between Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt

Table 11:	The meristic characters of the present work compared with those of different authors

Fig. 11:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on the morphometric of HL-indices set + SL-indices set, (MahalanobisD2 matrix)

Fig. 12:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on their gill raker counts, (Complete Linkage, Euclidean distances)

Fig. 13:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on their gill raker counts, (MahalanobisD2 matrix)

Fig. 14:	Clustering of the Epinepheline species of the present work (C. argus, C. miniata and V. louti) and of those of Mekkawy et al. (2002) (Epinephelus fasciatus, E. summana, E. polyphekadion and E. chlorostigma) from the Red Sea, Egypt, based on their gill raker, dorsal soft rays (DFR) and pectoral fin rays

CONCLUSION

Size-free meristics and morphometric characters are valid tools to identify species, genera and fish stock units. The present findings referred to different patterns of intra- and interspecific relationships between Epinepheline species and hence different fisheries management strategies. The most related species could be treated as one stock unit.