144 



P. erythrogaster bears the following 35 apomorphies: 4[1], 5[1], 13[1], 32[1], 35[1], 36[1]. 

 39[1], 49[1], 50[1], 65[1], 70[1]. 84[1], 92[1], 108[1], 111[1], 114[1], 124[1], 125[1], 

 133[1], 135[1], 145[1], 146[1]. 150[1], 153[1], 155[1], 165[1], 167[1], 174[1], 180[1]. 

 182[1], 184[1], 185[1]. 192[1], 197[1], and 209[1]. 



Discussion on the phylogenetic relationships of the species of Phoxinus 



The phylogenetic relationships within Phoxinus proposed herein indicate two clades in the 

 genus, the brachyurus-c\&<\z including all three Eurasian species and P. neogaeus, and the 

 erythrogaster-clade consisting of the other five North American species (Fig. 93). This re- 

 sult supports, in general, the phylogenetic hypotheses among some Phoxinus species pro- 

 posed by Joswiak (1980) based on allozyme data, and Starnes & Starnes (1978) and 

 Starnes & Jenkins (1988) based on external and osteological morphology. However. I dis- 

 agree with Joswiak's (1980) placement of Phoxinus species into three subgenera: Phoxi- 

 nus (including P. phoxinus), Pfrille (consisting of P. neogaeus), and Chrosomus (including 

 all other North American species of the genus). If Joswiak's (1980) proposal is followed, 

 the subgenus Phoxinus would include all species of the brachyurus clade, except P. neo- 

 gaeus, or the subgenus Pfrille would consist of all species of the brachyurus clade exclu- 

 ding P. phoxinus. Either way results in the subgenera Phoxinus and Pfrille being 

 paraphyletic, which is logically inconsistent with the phylogenetic tree, and not acceptable 

 in the phylogenetic study (Hennig 1966. Wiley 1981. Wiley et al. 1991). 

 The close relationship between P. phoxinus and P. neogaeus (both are in the phoxinus spe- 

 cies complex) has been accepted since the description of P. neogaeus was published by 

 Cope (1869), although some authors assigned P. neogaeus to another genus (Pfrille). How- 

 ever, the close relationship has never been critically reviewed. The characters thought to 

 support this relationship are either plesiomorphic. such as two rows of pharyngeal teeth, 

 short intestine (e.g., Gasowska 1980). or occur in all Phoxinus species, such as tubercu- 

 lation on breast scales in breeding males (e.g., Hubbs & Brown 1929, Gasowska 1979). 

 Ten synapomorphies found in this study support strongly the close relationship between 

 P. neogaeus and P. phoxinus plus issykkulensis. 



Mahy (1975c) proposed P. eiythrogaster oreas, and eos as three subspecies of a single 

 species P. eiythrogaster, and eiythrogaster as the ancestor of P. oreas and P. eos. My stu- 

 dy does not support Mahy's hypothesis and shows the three "subspecies" as different spe- 

 cies. Fig. 93 indicates that it is hardly to propose P. eiythrogaster be the ancestor of P. eos 

 and oreas. Data from allozymes (Joswiak 1980) also support my phylogenetic hypothe- 

 sis, not the one of Mahy (1975c). Mahy's failure to recognize the difference of these three 

 species might be "due to his choice of characters" (personal communication of Starnes to 

 Joswiak, cited from Joswiak 1980). 



Gasowska (1979) placed P. neogaeus in a subgenus Pfrille of genus Phoxinus. and the 

 other three North American species (P. eos, oreas. and erythrogaster) into two different 

 genera: P. eos and P. erythrogaster in Chrosomus, and oreas in a new genus Parchroso- 

 mus. Gasowska (1979) correctly recognized the close relationship between P. neogaeus 

 and P. phoxinus based on the breast tuberculation (though it is a plesiomorphic character 

 in the genus of Phoxinus). However, she seemed not to recognize that similar breast scale 



