146 



3) The "Bering Land Connection" dispersal hypothesis relies on exposure above see Le- 

 vel of Bering land connection between Asia and North America. A Bering land connec- 

 tion existed through most of the Tertian, and was present again during the Pleistocene 

 (Hopkins 1967. Joswiak 1980. Briggs 1986). "Relatively brief inundations apparently took 

 place in the Late Miocene and Late Pliocene during the interglacial stages" (Briggs 

 1986:7). The "Bering land connection" dispersal hypothesis implies that the ancestor of 

 the North American cyprinids originated in east Asia, and dispersed into North America 

 from Asia through the Bering land connection (Briggs 1979. 1986. Joswiak 1980). Bana- 

 rescu (1989) proposed that the ancestor of some North America percids (Perca) origina- 

 ted in Europe migrated into Asia and then into North America through the Bering land 

 connection. The "Bering land connection" dispersal hypothesis is accepted by some au- 

 thors to explain the relationships of (at least) the cyprinid fauna in Asia and North Ame- 

 rica (e.g.. Briggs 1979. Banarescu 1989). 



The Bering land connection played an important role in the formation of the North Ame- 

 rican freshwater fish fauna because the connection provided a possible way for the ex- 

 change of the freshwater fishes between North America and Asia, and provided some 

 vicariants for the speciation of the ancestors of (at least) some freshwater fishes on the 

 two continents after the submission of the Bering land connection. As discussed below, I 

 propose the Bering land connection be important for the speciation of Phoxinus as well. 

 My interpretation of the role of the connection, however, is based on view of the vicari- 

 ance biogeography (Wiley 1988). i.e.. I propose a "Bering land connection" vicariant hy- 

 pothesis to interpret the biogeography of Phoxinus. 



The area cladogram based on the phylogenetic relationship of Phoxinus indicates a close 

 relationship between Asia, Europe and North America (Fig. 94). However, in order to cor- 

 rectly understand the area cladogram (Fig. 94). it is necessary to discuss the geographic 

 distribution of P. phoxinus. P. phoxinus occurs widely in Europe (whole Europe except 

 southern Spain and Iceland - Sterba 1989), and in central and eastern Asia. The popula- 

 tions of the species in Europe might be interpreted as immigrant from Asia because the 

 European cyprinids had an Asian origin (Banarescu 1960). The east Asian population of 

 P. phoxinus was also considered an immigrant by some (Banarescu & Coad 1991). There- 

 fore. P. phoxinus might have originated in central Asia, and then dispersed into Europe 

 and East Asia. Thus what the area cladogram (Fig. 94) really indicates is that North Ame- 

 rica bears a closer relationship with Central Asia than with either Europe or East Asia. 

 This implies that the ancestor of the genus Phoxinus occurred in North America and Cen- 

 tral Asia (not East Asia), and was split by separation of Asia and North America into the 

 ancestor of the brachyurus clade in Asia, and the ancestor of the ei-ythrogaster clade in 

 North America (cf. Figs. 93. 94). This ancestor might have occurred in North America and 

 Asia during (or before) the Miocene when the Bering land connection existed, and was 

 split by submission of the Bering land connection in the Tertian . Because no fossils of 

 species of Phoxinus are known. I am not able to identify the age of the origin for the ge- 

 nus using the fossil record. However, based on the alloelectrophoretic studies on Phoxi- 

 nus allozymes by Joswiak (1980). the separation time of the two clades {erythrogaster 

 and brachyurus clades) in the genus Phoxinus was estimated about 14 million years ago 

 w hich would conform to the age of the Miocene. 



