Source of the fish

Three-to four-week old fry of the Tuxedo (TUX) and Green Variegated (GV) guppy strains were obtained from highly inbred and well-established TUX and GV stocks of the Chin Lam Brothers Tropical Fish Farm and Swee Hing & Brothers Aquarium Co., respectively, in Singapore. Tuxedo and Green Variegated are the commercial names given to these strains by guppy breeders. Male and female juveniles, distinguishable by the expression of their color patterns due to sexual dimorphism, were cultured separately according to Khoo et al. (1999a, b) for another three to four weeks before being used for reciprocal crosses between the TUX and GV strains. This was to ensure that juvenile males were fully mature (as indicated by a well-developed gonopodium) and females had not been previously inseminated. Under laboratory conditions, domesticated guppies reach sexual maturation at six to eight weeks of age.

Description of the strains

Adult males of the Tuxedo (TUX) strain have black or dark grey pigmentation on the caudal-peduncle region, and a caudal fin that ranges from blood-red to orange-red in color (Fig. 1A). TUX females show wild-type olive-brown body coloration and grey caudal-peduncle with red tinges of varying intensity on an opaque greyish-white tail (Fig. 1B). The TUX strain has been shown to carry the black caudalpeduncle (Bcp) and red tail (Rdt) color pattern genes by Fernando and Phang (1990) and Khoo et al. (1999b).

Fig. 1

(A) Adult male guppy of the Tuxedo (TUX) strain showing a black caudal-peduncle and red tail. (B) Adult female guppy of the TUX strain with grey caudal-peduncle and faint red tinges on an opaque greyish-white tail. (C) Adult male guppy of the Green Variegated (GV) strain displaying an orange tail with yellow streaks, and numerous black spots and patterns of different shapes and sizes. (D) Adult female guppy of the GV strain with wild-type female olive-brown background body coloration and faint greyish-brown variegated patterns on a yellowish tail.

Adult Green Variegated (GV) males display wild-type male body coloration which consists of polymorphic patches of various colors that are overlaid by a green metallic sheen. GV males also have a bright orange caudal fin with a mosaic pattern of black spots of different shapes and sizes, and some yellow streaks (Fig. 1C). GV females display wild-type female body coloration and greyish-brown variegated patterns on a yellowish translucent tail (Fig. 1D). The variegated tail patterning of the GV strain is due to the Var color pattern gene (Khoo et al., 1999a).

Reciprocal crosses

To establish the mode of inheritance and linkage of the black caudal-peduncle, red tail and variegated tail color patterns, singlepair reciprocal crosses were made between six-week old mature virgin fish of the TUX and GV strains. Each pair was kept in a 3.5-liter breeding tank. Broods were produced 4–6 weeks after mating. Single-pair full-sib F₁ males and F₁ females were mated to produce the F₂ generation. The following notations were used: TUX♂♂×GV♀♀ (Table 1A) and GV♂♂ × TUX♀♀ (Table 2A) for parental crosses, and F₁♂♂ × F₁♀♀ (Tables 1B and 2B) for full-sib F₁ crosses. Newly born fry were separated and raised to maturity in 3.5-liter clear plastic tanks (five fish/tank). F₁ and F₂ offspring were segregated according to color phenotype and sex. Their color patterns were designated as TUX (black caudal-peduncle and red tail typical of the Tuxedo strain), TUXVAR (Tuxedo with variegated tail patterning), RTVAR (red tail with variegated patterns), BCPVAR (black caudal-peduncle with variegated tail patterns), RT (red tail) and VAR (variegated tail with a mosaic pattern of large black spots and patches). To facilitate description of the crosses, Tuxedo male parents of TUX ♂♂ × GV ♀ ♀ were typed using Roman numerals (I, II, III, IV and V) according to their putative genotypes following segregation and scoring of F₁ and F₂ progenies (Khoo et al., 1999b).

Table 1

Mating results of crosses between Tuxedo (TUX) males and Green Variegated (GV) females showing observed and expected numbers for each phenotypic class, expected segregation ratios, chi-square goodness-of-fit to the expected ratios and their corresponding adjusted values (χ²_adj) after application of Yates' correction for continuity, χ² test for homogeneity, probable genotypes and recombinants for (A) the F₁ generation of single-pair parental crosses, and (B) the F₂ generation of single-pair crosses between full-sib F₁ males and F₁ females. Recombinants (^§) due to crossing-over of the Rdt, Bcp and Var genes were not considered in chi-square analyses. (Phenotypes: TUX=Tuxedo with black caudal-peduncle and red tail [grey caudal-peduncle and faint red tinges on an opaque greyish-white tail in TUX females]; RT=red tail without black caudal-peduncle; BCP=black caudal-peduncle without red tail; VAR=tail with variegated patterns; TUXVAR=Tuxedo with variegated tail patterns; RTVAR=red tail with variegated patterns; BCPVAR=black caudal-peduncle with variegated tail patterns. Genes: Bcp =black caudalpeduncle gene; Bcp⁺= absence of black caudal-peduncle gene; Rdt =red tail gene; Rdt⁺=absence of red tail gene; Var =variegated tail pattern gene; Var⁺=absence of variegated tail pattern gene).

Table 2

Mating results of crosses between Green Variegated (GV) males and Tuxedo (TUX) females showing observed and expected numbers for each phenotypic class, expected segregation ratios, chi-square goodness-of-fit to the expected ratios and their corresponding adjusted values (χ²_adj) after application of Yates' correction for continuity, χ² test for homogeneity, probable genotypes and recombinants for (A) the F₁ generation of single-pair parental crosses, and (B) the F₂ generation of single-pair crosses between full-sib F₁ males and F₁ females. Recombinants (^#) due to crossing-over of the Rdt, Bcp and Var genes were not considered in chi-square analyses. (Phenotypes: TUX=Tuxedo with black caudal-peduncle and red tail [grey caudal-peduncle and faint red tinges on an opaque greyish-white tail in TUX females]; VAR=tail with variegated patterns; TUXVAR=Tuxedo with variegated tail patterns; RTVAR=red tail with variegated patterns; BCPVAR=black caudal-peduncle with variegated tail patterns. Genes: Bcp =black caudal-peduncle gene; Bcp⁺= absence of black caudal-peduncle gene; Rdt =red tail gene; Rdt⁺=absence of red tail gene; Var =variegated tail pattern gene; Var⁺=absence of variegated tail pattern gene).

Statistical and linkage analyses

Observed phenotypic distributions were tested for goodness-of-fit with predicted proportions using the chi-square (χ²) test (Sokal and Rohlf, 1981; Strickberger, 1990). Since the observed and expected numbers in each phenotypic class and sample sizes were small (n< 200), Yates' (1934) correction for continuity was included in the calculation of χ² to improve the approximation to the χ² distribution, as shown by the χ²_adj values. Data was pooled when the χ² test for homogeneity indicated that there were no significant differences among the phenotypic frequencies, the observations were sufficiently uniform and the families homogeneous. The correction for continuity was not incorporated into the test for homogeneity because calculated χ² values had to be summed and χ²_adj values were not additive (Sokal and Rohlf, 1981; Strickberger, 1990). Individuals with exceptional coloration due to crossing-over of the black caudal-peduncle (Bcp), red tail (Rdt) and variegated tail (Var) genes between the X- and Y-chromosomes were not considered during chi-square analyses (Winge, 1922b, 1923, 1927, 1934; Nayudu, 1979; Phang et al., 1989a, b, 1990; Phang and Fernando, 1991; Khoo et al., 1999a, b).

Crossover fractions and map distances of Bcp, Rdt and Var relative to each other and the sex-determining region (SdR) were calculated according to Strickberger (1990), Phang et al. (1990), Phang and Fernando (1991), Purdom (1993) and Khoo et al. (1999a, b). Winge's (1922b, 1927, 1934) “zig-zag line diagram” method was used to test all possible linkage combinations between Bcp, Rdt, Var and the SdR, and to map these gene loci onto the sex chromosomes. Double recombinants were noted but excluded from all estimations of map distances (Winge, 1927, 1934; Purdom, 1993, Khoo et al., 1999b).

Segregation and recombination in F₁ and F₂ offspring of TUX × GV

Nine mating pairs of TUX♂♂ × GV♀♀ produced a total of 132 male and 131 female F₁ offspring (Table 1A). F₁ males had the black caudal-peduncle and red caudal fin of their TUX male parents but also displayed black spots and patches on the tail fin (Fig. 2A). F₁ females had a grey caudal-peduncle and an opaque greyish-white tail with red tinges and black spots (Fig. 2B). Phenotypically classed as Tuxedo with variegated tail patterning (TUXVAR), F₁ males and females inherited the black caudal-peduncle and red tail traits from their TUX male parents (type I) while variegated tail patterning was from the GV female parents (Fig. 3). Table 1A and Fig. 3 show two other crosses in which the TUX male parents were heterozygous for black caudal-peduncle. To facilitate describing these crosses and their offspring, the TUX males were labelled as types II and III. Types IV and V males (heterozygous for red tail) were not observed in this study although they were found among crosses between the Tuxedo strain and wild-type guppies (Khoo et al., 1999b). For mating pair TG7 (type II), there were three F₁ broods of 20 TUXVAR males and 32 females with variegated patterns on their reddish tails but without a black caudal-peduncle (RTVAR) (Figs. 2, 3, Table 1A). RTVAR males (27) and TUXVAR females (30) were produced by the cross between type III TUX males and GV females (mating pairs TG1 and TG20) (Figs. 2 and 3, Table 1A). For all three types (I, II and III) of TUX male parents, the number of F₁ males to females was consistent with the expected ratio of 1:1 (Table 1A).

Fig. 2

Progenies (F₁, F₂ and recombinants) from Tuxedo ♂♂× Green Variegated ♀♀ and Green Variegated ♂♂×Tuxedo ♀♀ crosses had the following phenotypic color patterns for (A) males: (a) Tuxedo (TUX), (b) Tuxedo with variegated tail patterning (TUXVAR), (c) red tail with variegated patterns (RTVAR), (d) black caudal-peduncle with variegated tail patterns (BCPVAR), (e) red tail (RT) and (f) variegated tail with a mosaic pattern of large black spots and patches (VAR), and (B) females: (a) TUXVAR, (b) RTVAR, (c) VAR and (d) TUX.

Fig. 3

Schematic diagram of the proposed genetic models showing segregation of the Bcp, Rdt and Var colour pattern genes, and genotypes of the parents (P), F₁ and F₂ progenies, and recombinants (R) in reciprocal crosses between the Tuxedo and Green Variegated guppy strains.

The F₂ generation for type I TUX male parents comprised 129 TUX and 149 TUXVAR males, and 145 TUXVAR and 140 VAR females (Figs. 2A, 2B), with observed numbers conforming to the expected 1:1:1:1 phenotypic ratio (Table 1B, Fig. 3). Four F₂ phenotypes of 25 TUX and 26 TUXVAR males, and 27 RTVAR and 25 VAR females were obtained from three single-pair full-sib F₁ crosses of type II (Figs. 2A, 2B). These also agreed with the 1:1:1:1 ratio (Table 1B, Fig. 3). Mating pairs TG1 and TG20 (type III) gave four F₂ phenotypes of 24 TUX and 21 RTVAR males, and 22 TUXVAR and 29 VAR females (Figs. 2A, 2B) that concurred with the ratio of 1:1:1:1 (Table 1B, Fig. 3). Homogeneity χ² tests for types I and III TUX males showed that the F₁ and F₂ progenies did not form heterogeneous populations and were uniform (Tables 1A, 1B). F₁ and F₂ results also indicated that homozygous Tuxedo male and Green Variegated female parents had the X_Bcp,Rdt,Var⁺Y_Bcp,Rdt,Var⁺ (type I) and X_Bcp⁺_,Rdt⁺_,VarX_Bcp⁺_,Rdt⁺_,Var genotypes, respectively (Table 1, Fig. 3). Conversely, genotypes of Tuxedo males heterozygous for black caudal-peduncle were X_Bcp⁺_,Rdt,Var⁺Y_Bcp,Rdt,Var⁺ (type II) and X_Bcp,Rdt,Var⁺Y_Bcp⁺_,Rdt,Var⁺ (type III). Fig. 3 shows the segregation and mechanism of inheritance of the Bcp, Rdt and Var genes.

Crossing-over between the Bcp, Rdt and Var color pattern genes, and the sex-determining region (SdR) from the Y- to the X-chromosome and vice versa in the F₁ parents produced F₂ offspring that did not conform to the expected phenotypic classes for TUX male parents of types I, II and III (Table 1B, Figs. 2 and 3). F₂ recombinants produced by type I were a BCPVAR and four VAR males, and four TUX females (Figs. 2A, 2B, Table 1B). Recombination frequency calculated from the percentage of VAR males out of the total number of F₂ males (4/283×100%) was 1.413% for the SdR–Rdt region (Table 1B). The four TUX females of 289 F₂ females gave a crossover rate of 1.384% between the Var locus and SdR. Occurrence of the VAR male and TUX female F₂ recombinants suggested that the Y-chromosome may have a gene map order of Var–SdR–Rdt–Bcp as these individuals could not be produced using Winge's (1922b, 1927, 1934) “zig-zag line diagram” method if the order had been either SdR–Var–Rdt–Bcp or SdR–Rdt–Bcp–Var (Figs. 3, 4). The BCPVAR recombinant male was not used to calculate map distance as it could be produced by single crossing-over that occurred simultaneously at SdR–Rdt in the F₁ male parent and Rdt–Bcp in the F₁ female parent. This individual could also result from double recombination between the SdR and Bcp whereby Rdt on the Y-chromosome of the F₁ male parent crossed over to the X. In the latter event, Rdt appears to lie between the SdR and Bcp, and has an estimated double crossover frequency of 0.353%.

Fig. 4

Genetic map of the Y-chromosome of the guppy, Poecilia reticulata, showing the positions of the black caudal-peduncle (Bcp), red tail (Rdt) and variegated tail (Var) loci relative to the sex-determining region (SdR). Map distances of Bcp, Rdt and Var from the SdR are based on recombination frequencies estimated from Tables 1 & 2 and Khoo et al. (1999a, b). Gene order for blue tail (Blt), green tail (Grt) and the snakeskin body-snakeskin tail complex (Ssb-Sst complex) were reanalysed from the crossover data of Phang et al. (1989a, b, 1990) and Phang and Fernando (1991). The allele positions of Winge's (1927, 1934) color pattern genes of wild-type guppies (*): Maculatus (Ma), Coccineus (Co), Tigrinus (Ti), Luteus (Lu), Vitellinus (Vi) and Elongatus (El) were according to Kirpichnikov's (1981) and Purdom's (1993) revisions. The loci of color pattern genes of the domesticated guppy (Bcp, Blt, Grt, Rdt, Ssb-Sst complex and Var) are inferred to be located at similar positions on the X-chromosome. The size of the SdR is not according to scale as the number of male-determining genes within that region is not known.

A recombination of 3.704% was estimated for SdR–Rdt from two VAR males out of 54 F₂ males for type II (Fig. 2A, Table 1B). These VAR males further support a gene order of Var–SdR–Rdt–Bcp (Figs. 3, 4). From a solitary RTVAR male (Fig. 2A), 1.852 map units was estimated to separate Bcp from Rdt (Table 1B). TUXVAR F₂ recombinant females of mating pair TG7 (type II TUX male) were not used to calculate map distances as crossing-over could have taken place at SdR–Rdt, Rdt–Bcp or SdR–Var (Figs. 2B, 3). Double recombination at either Var–Rdt or Var–Bcp might have produced F₂ TUX females where the SdR crossed over to the X-chromosome from the Y during meiosis in the F₁ male parents. These events generate several possible genotypes for a particular phenotype, hence making it impossible to determine the actual region of crossover. Map distance was also not estimated from the F₂ RT male recombinant of a type III TUX male (mating pair TG20) as crossing-over could occur at Var–Rdt or Rdt–Bcp in the F₁ female parent to produce this individual (Fig. 2A, Table 1B). Combining the results of this study (SdR–Rdt =2.559±1.620 map units) with those of Khoo et al. (1999b), the mean genetic map distance between SdR and Rdt was estimated to be 3.055±1.687 map units (Fig. 4).

Segregation and recombination in F₁ and F₂ offspring of GV × TUX

Seven mating pairs (GT1, GT2, GT3, GT5, GT8, GT9 and GT11) of the reciprocal cross, GV♂♂ × TUX♀♀, gave 18 F₁ broods of 98 males and 102 females. All the F₁ offspring possessed black caudal-peduncle and red tail with variegated patterns typical of the TUXVAR phenotype described earlier (Figs. 2A, 2B). The observed numbers of F₁ male to female offspring agreed with the expected ratio of 1:1 (Table 2A, Fig. 3). Except for seven TUX and five BCPVAR males resulting from crossing-over in the F₁ parents, F₂ progenies segregated into 101 TUXVAR and 89 VAR males, and 108 TUXVAR and 109 TUX females according to the 1:1:1:1 phenotypic ratio (Table 2B, Figs. 2, 3). No F₂ data was obtained for mating pair GT8, as the two F₁♂♂× F₁♀♀ pairs did not produce any F₂ offspring (Table 2B). The TUX female parent of mating pair GT7 was heterozygous for both Bcp and Rdt as it produced 25 TUXVAR and 18 VAR males, and 15 TUXVAR and 16 VAR females that conformed to the expected F₁ ratio of 1:1:1:1 (Table 2A, Figs. 2, 3). Similarly, GT4 and GT16 gave 17 TUXVAR and 14 RTVAR male, and 15 TUXVAR and 10 RTVAR female F₁ progenies, indicating that these two TUX female parents were heterozygous for the Bcp gene. As shown by the homogeneity χ² values, the families were uniform and homogeneous (Table 2). Results for F₁ and F₂ progenies also showed that the Green Variegated male and Tuxedo female parents used in this study were homozygous (genotypes: X_Bcp⁺_,Rdt⁺_,VarY_Bcp⁺_,Rdt⁺_,Var and X_Bcp,Rdt,Var⁺X_Bcp,Rdt,Var⁺, respectively) (Table 2, Fig. 3). Putative genotypes of heterozygous Tuxedo females were X_Bcp,Rdt,Var⁺X_Bcp⁺_,Rdt⁺_,Var⁺ for mating pair GT7, and X_Bcp,Rdt,Var⁺X_Bcp⁺_,Rdt,Var⁺ for GT4 and GT16. The segregation of Bcp, Rdt and Var, and their mode of inheritance are illustrated in Fig. 3.

Seven F₂ males from the full-sib cross of GV♂♂×TUX♀ ♀ exhibited normal Tuxedo color patterns, i.e., a black caudal-peduncle and red tail without variegated tail patterning (Fig. 2A, Table 2B). The variegated tail (Var) pattern gene, in this instance, would have crossed-over from the Y-chromosome to the X in the F₁ male parents to produce the TUX phenotype. Since there were seven TUX males out of 202 F₂ male individuals, 3.465% recombination occurred between SdR and Var (Table 2B). Five BCPVAR F₂ males were also produced by this cross, giving a distance of 2.475 map units between the Bcp and Rdt loci (Fig. 2A, Table 2B). From the F₂ results of TUX♂♂×GV♀♀ and GV♂♂×TUX♂♂ (Tables 1, 2), map distances for Bcp–Rdt and SdR–Var were averaged to be 2.164±0.441 and 2.425±1.471 units, respectively. In conjunction with our previous analyses involving crosses between wildtype guppies with the Green Variegated (Khoo et al., 1999a) and Tuxedo (Khoo et al., 1999b) strains, the Var locus appears to lie 2.174±1.301 map units from the SdR while Rdt and Bcp are 2.330±1.416 map units apart (Fig. 4).

Inheritance of the Bcp, Rdt and Var color pattern genes

Observations for all parental (TUX♂♂×GV♀♀ and GV ♂♂×TUX♀♀, Tables 1A, 2A) and full-sib (F₁♂♂×F₂♀♀, Tables 1B, 2B) crosses, initiated to determine the inheritance of the black caudal-peduncle (Bcp), red tail (Rdt) and variegated tail (Var) color patterns in domesticated guppy (Poecilia reticulata) strains, demonstrate that these color patterns are simple sex-linked traits controlled by single genes (Khoo et al., 1999a, b). In addition, our studies show that Bcp, Rdt and Var are dominantly expressed in both males and females, albeit the colors are more distinct and definitive in the males due to the presence of androgens (Figs. 1, 2, Tables 1, 2). These results confirm the preliminary findings of Fernando and Phang (1989, 1990) and Phang et al. (1990), and further support those of Khoo et al., (1999a, b). Each of the three color pattern genes has two alleles: Bcp which is dominant over Bcp⁺, Rdt dominant over Rdt⁺ and Var over Var⁺. Recessive alleles of these loci do not give rise to any color patterns. Tuxedo guppies used in this study are therefore proposed to have the X_Bcp,Rdt,Var⁺Y_Bcp,Rdt,Var⁺ (type I), X_Bcp⁺_,Rdt,Var⁺Y_Bcp,Rdt,Var⁺ (type II) and X_Bcp,Rdt,Var⁺Y_Bcp⁺_,Rdt,Var⁺ (type III) genotypes for males, and X_Bcp,Rdt,Var⁺X_Bcp,Rdt,Var⁺, X_Bcp,Rdt,Var⁺X_Bcp⁺_,Rdt⁺_,Var⁺ and X_Bcp,Rdt,Var⁺X_Bcp⁺_,Rdt,Var⁺ for females (Tables 1, 2, Fig. 3). Genotypes of Green Variegated males and females are X_Bcp⁺_,Rdt⁺_,VarY_Bcp⁺_,Rdt⁺_,Var and X_Bcp⁺_,Rdt⁺_,VarX_Bcp⁺_,Rdt⁺_,Var, respectively.

Phenotypic map of Bcp, Rdt, Var and the SdR

This study proves that the Bcp, Rdt and Var genes are able to cross over from the X-chromosome to the Y and vice versa since male and female recombinants of the TUX, VAR, RT, TUXVAR, RTVAR and BCPVAR phenotypes were observed at the F₂ level of TUX♂♂×GV♀♀ (for types I, II and III TUX male parents) and GV♂♂×TUX♀♀ (Tables 1, 2, Figs. 2, 3). Alleles of color genes migrating between the X- and Y-chromosomes were initially documented by Winge (1922a, b, 1923) in wild-type guppies. Subsequent analyses by Winge (1927, 1934), Winge and Ditlevsen (1938), Dzwillo (1959), Nayudu (1975, 1979) and Kirpichnikov (1981) showed that the X- and Y-chromosomes of the guppy are equal in size and indistinguishable by ordinary cytological methods. As a result of this homology between the sex chromosomes, genes are able to crossover along almost the whole length of their chromatids. Only a small segment on the Y-chromosome, the sex-determining region (SdR) which is presumed to contain male-determining genes, is known to be non-homologous and different from the X.

Double recombination may have given rise to a BCPVAR F₂ male (mating pair TG23) and two TUX F₂ females (mating pair TG7) (Table 1B, Fig. 3). Very high double crossover rates of 0.353% and 3.279% were obtained from these offspring, respectively. The TUX F₂ females may possibly be produced by a form of “sex-reversal” as a result of the instability of the genetic mechanism of sex-determination in the guppy (Kirpichnikov, 1981). This takes place when the SdR on the Y-chromosome undergoes double crossing-over with a segment that contains recessive female-determining genes on the X-chromosome (Winge, 1927, 1934; Winge and Ditlevsen, 1947; Kirpichnikov, 1981). Double crossover values were not included in our estimations of map distances because the number of crosses were limited and brood sizes were small (Khoo et al., 1999b). Moreover, Winge (1927, 1934) concluded that double crossing-over was unlikely to occur in the guppy as he could find only a few single crossovers among the thousands of guppies he crossed and examined. Also, recombination frequencies of up to 10% have been recorded only between the Doppelschwert (Double sword, Ds) and Pigmentierte caudalis (Caudal pigment, Cp) genes in wild-type guppies (Winge, 1927, 1934; Winge and Ditlevsen, 1947; Dzwillo, 1959; Nayudu, 1975, 1979; Kirpichnikov, 1981).

Based on our earlier studies (Khoo et al., 1999a, b), Bcp, Rdt and Var are inferred to be located within homologous regions on the X- and Y-chromosomes, and are about 5.147, 3.055±1.992 and 2.174±1.301 map units, respectively, from the SdR (Tables 1, 2, Fig. 4). The mean percentage recombination that involved Rdt and Bcp alone was 2.330±1.416. As expected, the map distance between SdR and Bcp (5.147 map units) is almost equal to the sum of distances (3.055+2.330=5.385 map units) for SdR–Rdt and Rdt–Bcp (Fig. 4) (Khoo et al., 1999b). Recombination rates between two loci that are far apart are, however, never exactly the sum of the estimates for smaller regions amidst them (Purdom, 1993). This is because a crossover between two loci usually inhibits a second crossover from occurring in an adjacent region (Strickberger, 1990). Crossing-over at SdR–Bcp, SdR–Rdt and Rdt–Bcp will thus influence each other, thereby affecting their frequency of occurrence. Despite this, our results indicate that Rdt is closer to SdR than Bcp as there was less recombination between Rdt and SdR than between Bcp and SdR (Tables 1, 2). This further verifies the findings of Khoo et al. (1999b). It is therefore evident that the Bcp and Rdt loci are arranged in the following sequence, SdR–Rdt–Bcp (Fig. 4), as proposed earlier by Khoo et al. (1999b).

The locus for variegated tail patterning, Var, is shown in this study and by Khoo et al. (1999a) to be 2.174±1.301 map units from the SdR (Fig. 4). Testing all possible linkage combinations for Bcp, Rdt, Var and the SdR using the “zig-zag line diagram” method of Winge (1922b, 1927, 1934), we have found that the map order of these genes is most likely Var–SdR–Rdt–Bcp (Fig. 4). Recombinant individuals observed in this study could not be produced if the order had been SdR–Var–Rdt–Bcp, SdR–Rdt–Bcp–Var or SdR–Rdt–Var–Bcp. Furthermore, any arrangement in which Bcp lies closer to the SdR than Rdt has been ruled out by this study and Khoo et al. (1999b). Khoo et al. (1999b) also noted that the SdR–Bcp–Rdt sequence was possible in only one exceptional case in which a segment of the Y-chromosome containing at least one of these three color genes and the SdR may have undergone translocation or pericentric inversion. Using the “zig-zag line diagram” method, we have reanalyzed and assigned other phenotypic markers such as blue tail (Blt), green tail (Grt) and the snakeskin body-snakeskin tail complex (Ssb–Sst complex) (Phang et al., 1989a, b, 1990; Phang and Fernando, 1991) onto the preliminary genetic map (encompassing Var–SdR–Rdt–Bcp) of the guppy Y-chromosome (Fig. 4). Six loci of Winge's (1927, 1934) color pattern genes of wild-type guppies: Maculatus (Ma), Coccineus (Co), Tigrinus (Ti), Luteus (Lu), Vitellinus (Vi) and Elongatus (El), were also ordered onto this map based on Kirpichnikov's (1981) and Purdom's (1993) revisions.

Recently, the use of molecular techniques such as Arbitrarily Primed Polymerase Chain Reaction or Random Amplified Polymorphic DNA (AP-PCR/RAPD) fingerprinting (Welsh and McClelland, 1990; Williams et al., 1990) has proved to be a quick and reliable method for generating a large number of genetic markers for the construction of linkage maps of the zebrafish, Danio rerio (Johnson et al., 1994; Postlethwait et al., 1994), medaka, Oryzias latipesy (Wada et al., 1995) and swordtail-platyfish hybrid of the genus Xiphophorus (Kazianis et al., 1996). In our on-going effort to identify and link genetic markers to the X- and Y-chromosomes of the guppy, Foo et al. (1995) showed that AP-PCR/RAPD markers were inherited in Mendelian fashion at the F₁ level. We have also found several AP-PCR/RAPD markers that could differentiate domesticated stocks and wild-type guppies (Chen, 1999). One of these, a 444 bp fragment amplified using a 10-mer primer OPJ-4 (Operon Technologies, USA), was absent in guppies with variegated tail patterning, while another 10-mer primer OPJ-7 amplified an 800 bp fragment in 80% of guppies with variegated tails (Chen, 1999). In view of applying these phenotypic and molecular markers in future studies to map the guppy sex chromosomes, we hypothesize the length of the guppy genome to be between 1,800 centiMorgans (cM) estimated by isozyme polymorphisms for Xiphophorus (Morizot et al., 1991) and 3,000 cM (1.7×10⁹ bp or 600 kb/cM) by AP-PCR/RAPD for zebrafish (Postlethwait et al., 1994). Following the assignation of AP-PCR/RAPD markers to the 24 chromosomes of Xiphophorus (Kazianis et al., 1996), the genome size of Xiphophorus has been revised to about 2,700 cM (S. Kazianis, personal communication).

In conclusion, the black caudal-peduncle (Bcp), red tail (Rdt) and variegated tail (Var) color pattern genes of the domesticated guppy are (1) single genes located at three different loci, (2) dominantly expressed, (3) X- and Y-linked, and (4) fully capable of crossing-over from the Y- to the X-chro-mosome and vice versa. Map distances for sex-determining region (SdR)–Rdt, Rdt–Bcp, SdR–Bcp and SdR–Var are approximately 3.1, 2.3, 5.1 and 2.2 map units, respectively, with a gene order of Var–SdR–Rdt–Bcp. To date, the number of sex-linked color genes reported for the guppy far outnumber the autosomal ones described (Winge, 1927, 1934; Winge and Ditlevsen, 1947; Dzwillo, 1959; Nayudu, 1975, 1979; Kirpichnikov, 1981; Fernando and Phang, 1989; Phang et al., 1989a, b, 1990; Phang and Fernando, 1991; Purdom, 1993; Khoo et al., 1999a, b). As such, additional genetic data is necessary to verify the preliminary genetic map shown in Fig. 4 through a larger sample size from more crosses and back-crosses. Consequently, this will allow the construction of a denser and more saturated map of the X- and Y-chromosomes using phenotypic and molecular genetic markers established by previous workers and those identified from on-going analyses of domesticated guppy strains.