Fish Physiology 8: 367-370 (1970)
Sex determination in the Platyfish, Xiphophorus maculatus
J. R. Gold

The platyfish, Xiphophorus maculatus, is both male and female heterogametic. Initially, Bellamy (1922) and Gordon (1927) found that "domesticated" stocks of X. maculatus of unknown origin were female heterogametic and male homogametic (WZ:ZZ). When other xiphophorine species (e.g., Xiphophorus variatus) were then found to be XX:XY male heterogametic (Kosswig, 1935; Bellamy, 1936), it was questioned as to how or why two different modes of sex determination would arise in such closely related forms. This odd situation became confounded further by Gordon's (1946, 1947) discovery that X. maculatus populations in Mexico were male heterogametic. From a series of crosses between "domesticated" heterogametic and "wild" heterogametic forms of X. maculatus, Gordon (1946, 1947) concluded that the Z of the heterogametic forms was equivalent to the Y of the heterogametic forms. Since he could find no evidence that W was equivalent to X, he suggested that the use of WY:YY (WY = , YY = ) was more appropriate than WZ:ZZ. Later, Gordon (1951) found naturally occurring populations of heterogametic X. maculatus from the Belize River in the former British Honduras.

It now appears that Gordon's appreciation of the sex chromosomes in X. maculatus was correct, although at the time (Gordon, 1952) he believed that X. maculatus was separated into two major, isolated populations or races one to the west in Mexico which was heterogametic (XX:XY), and one to the east in British Honduras which was heterogametic (WY:YY). Kallman (1965b, 1970b, 1973) sampled X. maculatus extensively throughout its native range from near Veracruz, Mexico, southeast to British Honduras, and found that the species is polymorphic for three sex chromosomes, W, X, and Y. Of the six possible zygotic combinations, four (WW, WX, WY, and XX) normally differentiate into females; the remaining two (XY and YY) normally differentiate into males. Since the Y is ubiquitous, the mode of heterogamety is dependent on the frequency of the W or X chromosome. For example, in the Belize River where Gordon (1951) first discovered heterogametic populations, the frequency of the X is low (ca. 0.045) and the dominant mode is heterogamety (WY:YY). To the northwest in the Rio Jampa, Mexico, the frequency of the W is apparently negligible, and the populations are heterogametic (XX:XY). A geographic cline, however, is not indicated. Both the W and X are widespread, and WX are not infrequent. In fact, the only trend noted by Kallman (1973) was that the X is possibly more prevalent in populations at the periphery of the platyfish distribution, suggesting that the W chromosome was a secondary modification of an already existant sex chromosome which arose in the center of the species' range. The existance of three different heterosomes leading to both and heterogamety is best documented in X. maculatus, but also may obtain in another poeciliid, Poecilia sphenops (Schröder, 1964, cited in Kallman, 1973), and in the cichlid, Tilapia mossambica (Hickling, 1960).

2. SEX DETERMINATION

As pointed out by Yamamoto (1969), sex in bisexual fishes is determined much in the same manner as demonstrated by Bridges (1925) in Drosophila. That is, "a given property, the sex included, depends upon all the chromosomes, some of which pull in one direction and others in the other direction, some strongly and others faintly or not demonstrably at all" (Bridges, 1939).

The early genetic experiments by Winge (1922) on the guppy, Poecilia reticulata, indicated that the species was heterogametic (XX:XY). Occasionally, however, he found exceptional individuals which were heterosomally of one sex, but phenotypically and functionally of the other sex. These "exceptions" proved fertile in crosses to "normal" individuals of the same sex chromosome constitution, but of the opposite sex. Exceptional XX crossed to normal XX produced all XX () progeny, and exceptional XY crossed to normal XY produced male (2XY, 1YY) and female (1XX) offspring in a 3:1 ratio. Winge (1934) and Winge and Ditlevsen (1947, 1948) interpreted these results as indicating that minor male (M) and female (F) determining genes were situated throughout the genome. Normally, these minor autosomal genes were hypostatic to the heterosomal sex-determining genes; the sex of an individual was a function of its sex chromosome constitution. However, through chance genetic or chromosomal recombinations,  the sum of the autosomal male- or female-potency could override the usually epistatic sex chromosome genes, and thus produce the "exceptional" individuals.

Similar explanations have been proposed by several authors to account for the sporadic appearance in nature and in the laboratory of these so-called sex reversais among the heterogametic xiphophorines (Kallman, 1968), and for hormone-induced sex reversals of O. latipes (Yamamoto, 1963). In the latter species, Aida (1936) established an XX:XX bisexual strain by selective breedings of exceptional XX , but suggested that the XX could have stemmed from a lowering of the female-potency of X chromosome genes. Although Aida's suggestion may have partial validity, the general consensus is that sex-determination (at least in a number of poeciliids and O. latipes) is polyfactorial, with epistatic sex genes located on the sex chromosomes. Kosswig (1964) has discussed this mode of sex determination in some detail, and Yamamoto (1969) has presented a simple, but useful, model based on three overlapping normal distribution curves.

The number, location, and mode of interaction of the autosomal M and F genes are unknown; the gene action, however, is not strictly additive. Kallman (1968) found evidence of specific sex transformer genes () in X. maculatus. In this instance, a fortuitous combination of autosomal genes, derived from crosses of two specific strains, was apparently sufficient to override the strong female-potency of the W chromosomes

In several: species, sex determination appears to be completely "polygenic," there being no genetic or cytological evidence of sex chromosome heterogamety. The most thoroughly studied example is the swordtail, Xiphophorus helleri. Over the years, Kosswig and his collaborators have found that sex ratios vary considerably among and within stocks of X. helleri and have proposed that "polygenes in their manifold recombinations decide about the sex of a specimen" (Kosswig, 1964, p, 195). A single pair of sex-indifferent autosomes, designated xx, are considered homologous to the sex chromosomes found in the heterogametic xiphophorines.

Interspecific hybridization studies between polygenic X. helleri and heterogametic X. maculatus have indicated that the M and F autosomal genes of X. helleri may, in certain combinations, be epistatic to sex-determining heterosomes (Kosswig, 1964). In the F1 of crosses between X. helleri   and X. maculatus (XY and YY), both male and female offspring were found among the chromosomal classes Xx and Yx. Sengün (1941, cited in Yamamoto, 1969), however, observed that Wx individuals from crosses of X. helleri to X. maculatus (WY) were all female, and that Xx individuals from crosses of X. helleri to X. maculatus (XX) were of both sexes. Presumably, this not only indicates that the W heterosome of X. maculatus has greater female-potency than the X, hut also that the W itself has a very strong feminizing tendency—a fact substantiated by the somewhat infrequent occurrence of exceptional WY in natural populations of X. maculatus (Kallman, 1973). Based on his discovery that the sex transformer genes of X. maculatus may cause fluctuations in sex ratios, KaIlman (1968) has suggested the interesting possibility that similar sex transformer genes may be prevalent in X. helleri.

Other species for which there is evidence of a polygenic mode of sex determination include two Caribbean poeciliids, Poecilia caudofasciata and Poecilia vittata (Breider, 1935, 1936), and possibly one anabantid, Macropodus concolor (references in Yamamoto, 1969).