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.

the same ponds (BERGER 1959, unpublished), diploid progeny sometimes complete metamorphosis, reach sexual maturity and are even fertile (BERGER 1971b). These results suggest that in the hemlclonal R. esculent

a and Italian hybrids, there is a gradual accumulation of deleterious recessive mutants in the rldlbund

a genomes (BERGER 1976, GRAF & MÜLLER 1979). When the R. esculent

a lineages have arisen recently, the two rldlbunda genomes in progeny of .crosses between them may complement each other (may not have the same deleterious recessive mutations). Similarly, when such mutant genomes are combined with a genome from R.lessona

e or the Italian non-hybrid, the deleterious mutations may not be expressed. Such an accumulation of deleterious mutations In hybridogenetic hybrid fish of the genus Poecillopsi

s has been documented (LESLIE & VRIJENHOEK 1978, 1980). Progeny of Italian hybrid appear to differ In one important aspect from progeny of R. esculenta

. Offspring of Italian hybrid, including those from small, medium-sized, and large eggs, die at the same stage (stage 25) with similar external abnormalities ("argulus-like", cf. BERGER 1967). It Is tempting to speculate that this uniformity may indicate that a very limited number of rldlbund

a genomes was Involved In the origin of at least the Sicilian hybrid lineages. The relative electrophoretic homogeneity of the rldibund

a alleles In the Italian hybrids for GPI, LDH (B locus) and serum albumin, In contrast to the heterogeneity In Central Europe (cf. UZZELL & HOTZ 1979) Is consistent with this hypothesis. It is worth noting that while some Italian hybrid females made ova that form discrete sjze classes, UZZELL & HOTZ (1979) saw no evidence of triploid Individuals in their Italian samples. The failure of the large eggs to develop Is unexplained, but may account for the lack of triploid hybrids. There are differences between the Italian non-hybrids and R. lessona

e in certain developmental features. While tadpoles of the first species developed in basins, their developmental rate was very slow, and all died. When they lived indoors they completed metamorphosis but developed slow. These data suggest that the tadpoles of the Italian non-hybrid require higher developmental temperatures than do those of R. lessonae

. These data are consistent with their present distribution in Europe: R. lessona

e is northern species and Italian non-hybrids clearly southern ones. Some offspring fathered by Italian non-hybrid males are exceptionally interesting. Among froglets fathered by male N38 in crosses with R. lessona

e and R. rldlbund

a females single individuals appeared which had morphological features with typical R. lessona

e phenotype (Table 4, crosses 8 and 14, Flg. 1 and Plate I: Fig. 2), and in one cross No. 12 all progeny were less

onae-llke. The appearance of those Individuals in such unexpected crosses, and expecially in cross with R. rldlbund

a. Is quite obscure and difficult to explain at present. Italian hybrids and non-hybrids are morphologically similar to each other and at present can be reliably separated only by their electrophoretic phenotypes (UZZELL & HOTZ 1979). The morphological features of the Fj generation confirm this similarity (Table 4, crosses 1-4). On the other hand, when a single R. esculent

a female was crossed with a male of each Italian taxon (crosses 15, 17, Table 5, Fig. 1), the mean morphological indices, especially for the ratio DP/CI are significantly distinct (t = 4.894, p< 0.001, 12 df). The results obtained when Italian hybrids are crossed with R. perez

i are quite different from those obtained when they are crossed with R. lessona

e or Italian non-hybrids. Cleaved egg of the latter two kinds of crosses produce very viable froglets, and those from the first kind of cross die as larvae or at least fail to complete metamorphosis. This is particularly striking since hybridogenetic hybrids with one genome from R. perezi

. the other from R. rldibund

a, have been reported from southern France (GRAF et al. 1977) and are known from northeastern Spain (UZZELL & TUNNER in press; UZZELL & HOTZ, unpublished). It is not obvious why the rldibund

a genomes from Italian hybrids should not work well with R. perez

i genomes. R. perez

i Is Immunologically relatively distinct from the other western Palearctic water frogs use^L In this study (UZZELL, in press). Perhaps its genome is not as well able to complement thfc recessive deleterious mutants of the rldlbund

a genome in the Italian hybrids as the genomes of the other three more closely related species do. We are aware, however, that R. perez

i males give two types of progeny. Offspring parented by an Italian hybrid female and Portuguese R. perez

i male died soon after hatching, but those fathered by French R. perez

i male were more vital, they died during metamorphosis (Table 1, crosses 12, 13). When mothers were R. lessona

e females, all Individuals reached full metamorphosis but transformed froglets formed two morphological phenotypes. Those fathered by Portuguese males were rather lessona

e-llke. and those fathered by French males were similar to the R. esculent

a triploid Individuals with two rldlbund

a genomes and one lessonae