The karyotypes of the Drosophila robusta species-group, especially its species-subgroup okadai, were studied by using brain cell-Giemsa staining method. Drosophila gani possessed a diploid number of 12 chromosomes, comprising 4 pairs of metacentric and 2 pairs of acrocentric chromosomes, and both D. okadai and D. neokadai did 2n=12, with 3 pairs of metacentrics, 2 pairs of acrocentrics and 1 pair of microchromosomes. The okadai subgroup characteristically retained a middle-sized acrocentric × chromosome, which implies that this subgroup may occupy an ancestral position for the robusta group.
The Drosophila robusta species-group which is widely distributed in the Palearctic and the Nearctic regions has been well studied from micro- and macro-evolutionary points of view (for a review see, Levitan, 1982). Narayanan (1973) has proposed a hypothetical phyiogenic tree of the robusta group, based mainly on interspecific comparisons of karyotypes and of banded sequences of salivary gland chromosomes. However, not any species of its species-subgroup okadai has been studied chromosomally, because of its stubbornness in the laboratory rearing. Further, new members of the robusta group have recently been discovered in the mainland of China and new knowledge on its geographic distribution has been accumulated in China and Russia (Watabe et al., 1990a; Watabe and Peng, 1991; Chen and Watabe, 1993; Toda et al., 1996).
In the present article, first we investigated the results of the chromosomal observations of the robusta group, and then discussed an ancestor form common to the robusta group and its sister species-group melanica, based on the results obtained and recent information on the biogeography of those groups.
MATERIALS AND METHODS
All flies were collected by banana traps and from timber piles from 1994 to 1996. The species and geographic strains available for the present study are given in Table 1, together with the results of the karyological observations and information of the geographic distribution.
Karyotypes of the Drosophila robusta and the D. melanica species-groups, and their geographic distribution
The females were reared in glass vials with malt-yeast medium. at 18°C, and a small piece of apple was placed on the surface of food for the oviposition site.
Preparations of mitotic chromosomes were made with neuroblasts of the 3rd larvae treated with 0.1 mg/ml of colchicine solution, stained with 4% Giemsa solution, and then air-dried (Imai et al., 1977). About 100 nuclear plates were observed in each species.
RESULTS AND DISCUSSION
Figure 1 shows the metaphase configurations of five species of the robusta species-group and D. moriwakii of the melanica group. The karyotype of D. okadai was composed of 3 pairs of middle-sized metacentrics (V-shaped), 2 pairs of acrocentrics (Rod-shaped), and 1 pair of dotlike (D) chromosomes (Fig. 1A). Its × chromosome was acrocentric and the Y was metacentric. Drosophila neokadai showed the same karyotype as D. okadai, except for its submetacentric (J-shaped) Y chromosome (Fig. 1B). Drosophila gani possessed 2n =12 chromosomes, consisting of 4 pairs of middle-sized metacentric autosomes, 1 pair of acrocentric autosomes, and unpaired sex chromosomes, acrocentric × and metacentric Y (Fig. 1C). These three species belong to the okadaisubgroup, together with D. unimaculata distributed in eastern and central Europe (Watabe and Nakata, 1989). The metaphase plates of D. sordidula of the robusta subgroup comprised 2 pairs of metacentrics, 2 pairs of acrocentrics and 1 pair of dotlike chromosomes (Fig. 1E). The same result has been already obtained by Kikkawa and Peng (1938). The × chromosome of D. sordidula is the largest metacentric, being about two times as long as acrocentric autosomes. Such a large metacentric × has been found in three species of the robusta subgroup, D, robusta, D. pseudosordidula and D. cheda(Tan et al., 1949; Tokumitsu et al., 1967), and in one species of the lacertosa subgroup, D. lacertosa with some variations of small autosomes (cf. Momma, 1956; Fig. 1F).
In addition to the robusta group, the karyotypes of two species of the melanica group, D. moriwakii and D. tsigana, were checked for the following discussion. Drosophila moriwakii shows a slight variation of the chromosome complement, i.e., 2n =12 with 3V + 2R + 1D (Okada and Kurokawa, 1957), 2n=12 with 1V+1J + 4R (Tokumitsu et al., 1967), 2n =12 with 2V + 4R (Fig. 1D) and 2n =14 with 2V + 4R +1D (Watabe, unpubl.), but its × has been observed invariably to be acrocentric in those articles. The karyotype of D. tsigana comprised 2n =10 (2V + 2R +1D) with middle-sized acrocentric X, and the same result was reported as in “melanissima” by Kikkawa and Peng (1938) and as in “pengr” by Okada and Kurokawa (1957).
The rod-shaped × has been considered to be primitive than the large V-shaped one, since centric fusion with the loss of one centromere is much more likely to occur in the specia-tion of Drosophila than centric fission with the addition of one centromere. Geographical, historical, or genetical evidences support this in various groups of Drosophila, e.g. in Hawaiian Drosophila (Carson and Yoon, 1982). With regards to the robusta group, all autosomes are middle-sized, whereas large metacentric × are found in the both robusta and the lacertosa subgroups. The single-element rod × had been known only in two allopatric species, Asiatic D. moriwakii and North American D. colorata, and therefore Narayanan (1973) considered that they “are, or represent” the ancestor species of the robusta group. Beppu (1988) transferred D. moriwakii and D. colorata from the robusta group to its closely related species-group melanica, based on the genitalial structure, although the two species share all the diagnostic characters of the robusta group defined by Sturtevant (1942). If this is true, a primitive form of the robusta group will become open to question, again. The present results, i.e., the presence of acrocentric × in the okadaisubgroup and a large V-shaped one in the other subgroups, strongly suggest that the former may occupy an ancestral position in the robusta group (Fig. 2). In addition to the morphology of X, the chromosome complement of the okadai subgroup resembles that of D. moriwakii and D. colorata, in having the same number of chromosomes (2n =12) and in lacking any of large-sized metacentric chromosomes (Fig. 1A-D). Hence, it is suggested by the outgroup comparison that the ancestor common to both the robusta and the melanica groups may have possessed a rod-shaped X, and subsequently Robertsonian translocation and pericentric inversions may have occurred during the course of species divergence within the robusta group and the melanica group.
The rod-shaped × is also known in other species of the melanica group, Palearctic D. tsigana and North American D, micromelanica (Stalker, 1972). Drosophila tsigana presently shows a disjunctive pattern of the geographic distribution, East Asia and western Europe (Watabe et al., 1990b). It is highly probable that D. tsigana might have once possessed an extremely wide range of distribution and that its American relatives might have diversified from “pre-tsigana”.
A remarkable decrease of temperature called “Terminal Eocene Event” took place from late Eocene to early Miocene (ca. 38.6-16.3 mya), and this reinforced the range expansion of temperate mixture forests in the middle latitudes of Eurasia (Tanai, 1971,1990; Wolfe, 1978). The robusta and the melanica species-groups might have emerged in East Asia during this Tertiary period (Throckmorton, 1975), since both groups are the temperate species in the Holarctic region and are strongly sylvicolous. Thereafter, temperate forests spread to high latitudes of Eurasia in early to mid Miocene (ca. 16.3-10.4 mya) when the warming of temperature occurred, and the robusta and the melanica group flies might have immigrated from Eurasia to North America, via the Bering Straight.
The recent faunal studies in the mainland of China and in the Russian Far East have yielded many new species and a lot of knowledge available on the biogeography of the robusta group and its related species-groups (Watabe et al., 1990a; Watabe and Peng, 1991; Chen and Watabe, 1993; Toda et al., 1996). A further investigation using molecular techniques is obviously desirable for tracing accurately the phylogeny and evolution within and between the robusta and the melanica species-groups.
We thank Prof. M. J. Toda (Hokkaido University), Dr. K. Tamura (Tokyo Metropolitan University), and the students of Hokkaido University of Education (Messrs. N. Teraya, K. Nakano, K. Yamada) for their help in the field collection and in the laboratory experiment. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan (Nos. 06640906, 07640920).