O. G. Dely szerk.: Vertebrata Hungarica 21. (Budapest, 1982)
Berger, L., Uzell, Th. ; Hotz, H.: Crossing experiments between some Western Palearctic species of water frogs (Salientia: Ranidae) 33-46. o.
DISCUSSION The comparative study of development of crosses within and between three morphological forms of water frogs in Central Europe (BERGER 1967, 1971a) demonstrated that two of these ( R. lessona e and R. rldlbunda ) are Mendellan species, while the third ( R. esculenta ) consists of lineages that arose and arise as a result of hybridization of the former two species. Analysis of thfe inheritance of characters in these crosses (BERGER 1968) provided the basis for recognizing that R. esculent a Is a hybridogenetic hybrid (TUNNER 1973, 1974). To date, few comparable studies have been made on the southern European and Iberian water frogs. KAWAMURA & NESHIOKA (1979) showed that while few crosses between eastern species of Palearctlc water frogs or between western species of Palearctlc water frogs produced completely inviable zygotes, most crosses between eastern and western species resulted in completely inviable progeny. In these latter crosses inviabiltty usually occurred following neurulatlon. These data suggest that eastern and western Palearctlc water frogs belong to different taxonomlc groups . That the western Palearctlc water frogs form a cohesive group different from the eastern Palearctlc water frogs has been confirmed by immunological comparisons of their serum albumins (UZZELL in press, unpublished). The high percentage of cleaved eggs In all types of crosses (Table 1) shows that there Is no gametic isolation among the various forms used in our crosses, which agrees with KAWAMURA & NISHIOKA's (1979) results. That these represent a relatively closely related set of taxa is indicated not only by the survival of most zygotes In most crosses through the neurula stage, but also by the high viability of tadpoles and of transformed Individuals. Crosses Involving Italian hybrids were the only exceptions. The Italian hybrids have one genome from the Italian non-hybrids and one from R. ridtbund a. Whether the ridlbund a genome was obtained directly by hybridization of the Italian non-hybrid with R. rldlbunda , or Indirectly by hybrid!zation of R. esculent a from the Po Valley with the Italian non-hybrid is not known, although the latter seems more probable (UZZELL & HOTZ 1979). The Italian hybrid and non-hybrid, which coexist in the lowlands of Sicily and peninsular Italy south of the Po-Valley, form a population system analogous to the R. lessona e - R. esculent a system (L-E system, UZZELL & BERGER 1975) that is widely spread over Central Europe (BLANKENBORN et al. 1971, BLANKENBORN 1973, BERGER 1973, GÜNTHER 1973, 1975, TUNNER 1974, WIJNANDS & GELDER 1976). These data indicate that the Italian hybrid lives with and «productively depends on the Italian non-hybrid. Several features attest to the similarity of these two population systems. Progeny fathered either by Italian hybrids or by R. esculent a Include only daughters (Table 5; BERGER 1971b), which phenotype depends cm their mother. Froglets with R. rldlbund a or R. esculent a mothers have rldlbund a phenotypes, while those with R. lessona e mothers have esculent a phenotype (Table 4, Fig. 1; BERGER 1968). These data confirm UZZELL & BOTZ' (1979) conclusions that the Italian hybrids are hybridogenetic and pass only ridlbund a genomes to gametes. The presence of both sexes in adults among the Italian hybrids indicates that these hybrids, like R . esculent a in the L-E system, persist primarily through matlngs between male non-hybrids and female hybrids. The variability in vitality of progeny of Italian hybrids in crosses with other kinds of frogs, especially when compared to studies of R. esculent a in Poland (BERGER 1967, 1971a), show some important patterns. Crosses between two Italian hybrids have very little success; most cleaved eggs develop to stage 25, but all fall shortly thereafter (Table 1, crosses 5-7). Progeny of crosses with R. esculent a (crosses 25, 26) die as larvae or shortly after metamorphosis. Progeny of crosses with the three species complete metamorphosis and produce vigorous froglets. Many pairs of R. esculent a produce almost no metamorphosing progeny (BERGER 1967), but this varies as a function of the size of the egg (BERGER 1967, 1979) and of the origin of the parents (BERGER 1971b). If the parents come from geographically isolated L-E system populations (BLANKENBORN et al. 1971) or from pure R. esculent a populations (BERGER & TRUSZKOWSKI 1980), diploid offspring, which develop from small and medium-sized eggs (UZZELL et al. 1975, BERGER et al. 1978, BERGER & ROGUSKI 1978), die during larval development. Triploid progeny, which develop from large eggs containing at least one lessona e genome (UZZELL et al. 1975), complete metamorphosis and produce viable froglets. If, however, the R. esculent a come from areas In which all three forms live sympatrically and occasionally breed in
