Taxonomic Relationships: The Central California roach was first described as Pogonichthys symmetricus (Baird and Girard 1854a) from specimens collected from the San Joaquin River at Fort Miller near the present-day location of Friant Dam. It was subsequently reassigned to the old world genus Rutilus until 1913 when John O. Snyder erected the genus Hesperoluecus and described the following six species based on locality, isolation and morphological differences.
1. Hesperoleucus mitrulus from the tributaries to Goose Lake, Lake County, Oregon. Hubbs et al. (1979) referred roach from the Upper Pit River, Modoc County to this taxon.
2. Hesperoleucus navarroensis from the Navarro River, Mendocino County.
3. Hesperoleucus parvipinnis from the Gualala River, Sonoma County.
4. Hesperoleucus symmetricus from the Sacramento and San Joaquin rivers and their tributaries.
5. Hesperoleucus venustus from the Russian River and streams tributary to San Francisco Bay. Snyder (1914) included roach from Tomales Bay tributaries in this taxon).
6. Hesperoleucus subditus from the major streams flowing into Monterey Bay.
The generic name Lavinia (Baird and Girard 1854a) has precedence over Hesperoleucus (Snyder 1913) and is preferred because roach and hitch (the only other species in the genus) are interfertile and closely related genetically (Avise et al. 1975, Avise and Ayala 1976, Moyle and Masingill 1981, Jones et al. 2002, Aguilar et al. 2009). Roach are known to hybridize with hitch in the Pajaro and Salinas Rivers (Miller 1945b), in Coyote, Alameda, and Walnut Creeks (Miller 1963, Leidy 1983, 2007, Leidy 2009, pers. comm.) in the Sacramento-San Joaquin drainages (Avise et al. 1975, Jones 2001) and with arroyo chub (Gila orcutti) in the Cuyama River (Greenfield and Greenfield 1972). Hybridization may be a result of low water conditions whereby the hitch and roach, which normally occupy different habitats (roach use higher gradient stream sections than do hitch), become restricted to the same isolated pools as streams dry (Miller 1945b, Jones 2001). Hybrids are fertile in the Pajaro River; Avise et al. (1975) found 8% of the Lavinia examined to be F1 hybrids and 5% to be backcrossed individuals.
Miller (1945b p. 197) in the same paper that described hitch/roach hybrids from the Pajaro River suggested that preliminary analysis of the forms of Hesperoleucus shows that many if not all, of those described as species are geographic subspecies of H. symmetricus. In his unpublished M.S. thesis, Murphy (1948c) reanalyzed data from Snyder (1913) along with his own data from coastal California streams and concluded that coastal forms should be relegated to subspecies of H. symmetricus. Murphy did not include H. mitrulus in his study. In his arguments for merging coastal forms into Hesperoleucus symmetricus, Murphy (1948c p. 49) did not dispute that Snyders species were morphologically and genetically distinct, instead, he followed what appears to be a strict interpretation of the biological species concept as outlined by Mayr (1942, 1954). Murphy argued that the distinctiveness of isolated populations, such as those in the Gualala Rivers, resulted from merely a chance genetic divergence resultant from small numbers of colonizing individuals and that if physical barriers were removed from between forms isolated in separate basins a population would soon lose its identity. Although it was never published, Murphys (1948c) diagnosis was adopted by subsequent workers (Hopkirk 1973, Moyle 1976) and by the California Academy of Sciences (Hubbs et al. 1979).
Twenty-five years later, Hopkirk (1974) pointed out that Murphys principal argument in denying specific status to coastal roach populations, the concept of a chance genetic divergence during colonization, was an important mechanism in speciation; the founder effect (Mayr 1942, 1954, among others). Hopkirk also asserted that natural selection contributed to differences between roach populations and therefore the distinctiveness of populations was not due solely to the chance combination of genetic factors, as Murphy had maintained. Regardless of his critique of Murphys species concept, Hopkirk agreed that Murphy was correct in placing all roach in one species, but he differed in his conclusions as to which populations should be accepted as subspecies. Hopkirk (1974) recognized H. s. symmetricus, H. s. parvipinnis and H. s. subditus as distinct subspecies and suggested that fish from the Russian River should be grouped with H. s. navarroensis and that roach from the Tomales Bay region be given subspecific status. Hopkirk (1974) also concluded that H. venustus from San Francisco Bay tributaries and roach from the Clear Lake drainage were synonymous with H. symmetricus from the Central Valley. Hopkirk (1974) further cautioned that H. s. symmetricus likely consisted of several subspecies, citing as an example a population from the Cosumnes River that exhibited morphologically distinctive characters (Hopkirk 1974).
Some support for Hopkirks assertion of variability in Sacramento Valley roach populations was provided by Loggins (1997), who, using DNA fingerprinting techniques, found evidence for fairly long isolation of four adjacent Sacramento populations. Similarly in the San Joaquin drainage, Brown et al. (1992b) found that populations from the different drainages could be distinguished by multivariate analysis of 15 morphometric characters. Populations from the Kaweah River and from the Red Hills (i.e., Horton Creek and other small creeks near Sonora, Tuolumne County) were particularly distinctive because of the high frequency of a chisel lip feature. Jones et al. (2002, p. 261 ) found that the Red Hills roach population appeared reciprocally monophyletic for assayed mitochondrial DNA markers and this combination of morphological and genetic distinctiveness in the Red Hills roach led Moyle (2002) to assign it subspecific status. While acknowledging the need for a taxonomic reevaluation of Lavinia, Moyle (2002 p. 140) simultaneously recognized eight subspecies of roach and called for a thorough biochemical and morphological investigation into the roach species complex.
In the most comprehensive genetic study of Lavinia to date, Aguilar et al. (2009) used both nuclear microsatellite (nDNA) and mitochondrial DNA (mtDNA) markers. Employed in tandem these two genetic markers supply insight into both the relationships between populations (phylogenetics) and the distinctiveness of individual populations (taxonomy). The microsatellite analysis of Aguilar et al. (2009) clearly defined Gualala, Pit, Navarro and Red Hills populations as distinct genetic units and largely supports the subspecies proposed by Moyle (2002), with the notable exception that separate groupings were found for Russian River and Clear lake populations. Analysis of mtDNA identified roach from the Pit and Gualala rivers to be highly divergent from all other populations and reciprocally monophyletic for the haplotypes assayed, suggesting that these populations have been isolated for considerable time. In addition, mitochondrial results show Tomales, Red Hills and Russian River/ Clear lake roach populations to be highly supported clades.
In light of (1) the recent genetic analysis (nuclear and mtDNA) that corroborates the distinctiveness of the species that Snyder (1913) originally described and (2) the fact that Snyders original species names were never properly submerged (i.e. through formal publication of an analysis in the peer-reviewed literature), the following taxonomic designations should be regarded as valid:
1. The Northern roach (Pit Roach) is a valid full species. The subspecies name, Lavinia s. mitrulus (Murphy 1948c) is pre-occupied by Lavinia mitrulus (Snyder 1913).
2. The Gualala roach is a valid full species. L. s. parvipinnis (Murphy 1948c) is pre-occupied by Lavinia parvipinnis (Snyder 1913).
3. The subspecies designations for the Navarro, Tomales, and Red Hills roach subspecies should be retained. These taxa are probably sufficiently distinct to warrant full species status but genetic evidence from a sufficient sample size is necessary to allow increased statistical support for such a conclusion. The following is a list of the taxonomic units for roach in California. Full species are denoted in bold.
1. Northern roach, L. mitrulus
2. Gualala roach, L. parvipinnis
3. California Roach, Lavinia symmetricus
3a. Central California roach, L. s. symmetricus
3b. Navarro roach, L. s. navarroensis
3c. Monterey roach, L. s. subditus
3d. Clear Lake roach, L. s. ssp.
3e. Russian River roach, L. s. spp.
3f. Tomales roach, L. s. ssp.
3g. Red Hills roach, L. s. ssp.
In addition, it is likely that L. s. symmetricus is a polytypic taxon and that several Central Valley populations (e.g., Cosumnes River and Peoria Creek) may also deserve subspecific (or higher) status.