The Western Rainbow fish , Melanotaenia australis, is a freshwater species widespread across north-western Australia . Given the geological stability of the region, and the local endemism in the Melanotaenia and north-west fish communities, such a broad distribution is contrary to expectation. Populations of M. australis are currently united on the basis of morphological characters. The primary aim of this study was to examine patterns of genetic subdivision and morphological differentiation to resolve whether morphology has been an accurate reflection of the genetic relationship between populations. The results will also shed light on speciation processes in Melanotaenia and represent the first study of population genetics in the Kimberley .
Genetic Subdivision and Morphological Differentiation
Investigation in the east Kimberley took place at four spatial scales, with an outlying population in the Pilbara. Small scales (within creek lines, between creek lines and between subcatchments) were sampled within Kachana station in the headwaters of the Chamberlain River . Outlying sites were located in the Dunham and Keep Rivers . The molecular marker chosen was allozymes - a variable set of enzymes that directly reflect the genotype of the organism. The frequency of each type of allozyme can be used to infer the genetic similarity of populations. Using allozyme electrophoresis, M. australis showed considerable regional genetic subdivision within the east Kimberley and broad-scale genetic divergence between the Kimberley and Pilbara populations. At small scales F ST (a measure of genetic subdivision) was low but statistically significant indicating extensive dispersal along creek lines. A mantel’s test demonstrated an effect of isolation by distance along creek lines.
Meristic counts and morphological measurements were made 14 traits for a subset of individuals at each site. Traits included characters such as number of dorsal rays, eye diameter and mouth length and number of vertical scale rows. Multi-dimensional scaling (MDS) and analysis of similarity revealed a similar pattern of genetic subdivision and morphological differentiation. Morphological variation within populations was considerable and each river showed much of the variation present in the entirety of the species. It was concluded that, while morphology does reflect genetic subdivision, morphological variation is subtle between populations within a background of extensive variation within populations. Thus even pronounced genetic divergence between populations will not be recognised if only the range of traits in the populations are considered.
Environmental conditions can effect the morphology similarity of populations through (i) differences in natural selection between populations (ii) plasticity of shape during development to better suit the fish to the prevailing environmental conditions. The effect of environment on the relationship between genetics and morphology was investigated. Discriminant function analysis was used to establish the traits most important in distinguishing between populations. At small scales the utility of traits was inconsistent and in some cases traits not known to be affected by environment could distinguish between populations. During specimen collection, a suite of environmental variables was recorded that described the habitat at each pool. Analysis of morphology and the environmental characteristics of the pools suggested that variation in some traits may be affected by local environmental variation.
Status of the Pilbara population
The percentage of allozyme loci universally different between the Pilbara and Kimberley populations (fixed differences) was close to a level where they are likely to represent distinct species. While fixed differences are evidence for an absence of gene flow, this does not necessarily correspond with reproductive isolation, which is used to separate species in the biological species concept. Consequently, the level of fixed difference is regarded as good evidence for allopatric speciation of the Pilbara population, rather than confirmation.
Population Genetics in the Kimberley
The low levels of genetic subdivision within drainages demonstrate that wet season dispersal rather than dry season isolation of populations, is the dominant force underlying patterns of genetic structure. Because pools are regularly joined, dispersal between pools appears to be extensive within aquatic systems.
Pairwise comparisons of F ST at regional scales demonstrated the importance of lowland floodplain versus inland plateaus in creating genetic subdivision. Comparatively low levels of F ST between lowland sites may be responsible for the greater similarity of east Kimberley fish fauna with the Northern Territory rather than the remainder of the Kimberley . The ancient ranges of the Kimberley can cause highly subdivided species and may account for the restricted distribution of numerous species of freshwater fish in the Kimberley and the high level of endemism in the region. Given the common effect of geography on creating genetic subdivision, it is anticipated that other freshwater species will show extensive genetic subdivision within the Kimberley . Such subdivision is expected to be larger than that encountered in M. australis, because this species is of comparatively high dispersal ability.
Speciation in the Melanotaenia
In light of the divergence between the Kimberley and Pilbara populations and the high level of endemism in the Kimberley , examination of outlying populations in other Melanotaenia may reveal species yet to be identified. Both Melanotaenia exquisita and M. nigrans are predominantly distributed in the Northern Territory with outlying populations in the north Kimberley . Given that Melanotaenia species often show overlapping morphology and, as demonstrated in this study, subtle differences associated with genetic divergence, these populations may represent species undetected based on morphology.
Despite similar distances between the river systems genetic subdivision between the Chamberlain and Dunham Rivers system was high compared with between the Dunham and Keep Rivers . This suggests the importance of plateaus separating the Dunham and Chamberlain sites in creating subdivision, compared to the lowlands separating the Dunham and Keep Rivers . Such an effect explains the pattern in which distributions of Melanotaenia inhabiting lowland areas are on average considerably larger.
Below are references of particular relevance to freshwater biology in the Kimberley , rainbowfish and some of the key concepts in this study:
Allen, G.R. (1982). A Field Guide to Inland Fishes of Western Australia . University of Western Australia Press, Perth , Western Australia .
Allen, G.R. (1991). Field Guide to the Fishes of New Guinea . Publication No. 9 of the Christensen Research Institute, Madang , Papua New Guinea .
Allen, G.R. & Leggett, R. (1990). A collection of freshwater fishes from the Kimberley region of Western Australia . Records of the Western Australian Museum 14: 527-545.
Allen, G.R., Midgely, S.H. & Allen, M. (2002). Field Guide to the Freshwater Fishes of Australia . Western Australian Museum , Perth .
Bishop K.A., Allen, S.A., Pollard, D.A. & Cook, M.G. (2001). Ecological studies on the freshwater fishes of the Alligator Rivers region, Northern Territory: Autecology. Supervising Scientist Report 145, Environment Australia .
Kay, W.R. Smith, M.J. Pinder, A.M., McRae, J.M., Davis , J.A. & Halse , S.A. (1999). Patterns of distribution of macro invertebrate families in rivers of north-western Australia . Freshwater Biology 118: 309-317.
Larson, H.K. & Martin, K.C. (1989). Freshwater fishes of the Northern Territory. Northern Territory Museum of Arts and Sciences, Darwin .
McGuigan, K., Zhu, D., Allen, G.R. & Moritz, C. (2000). Phylogenetic relationships and historical biogeography of melanotaeniid fishes in Australia and New Guinea . Marine and Freshwater Research 52: 713-723.
Pusey, B.J. Bird, J. Kennard, M.J. & Arthington, A.H. (1997). Distribution of the Lake Eacham Rainbowfish in the Wet Tropics Region, North Queensland . Australian Journal of Zoology 45: 75-84.
Tibbets , C.A. & Dowling, T.E. (1996). Effects if intrinsic and extrinsic factors on population fragmentation in three species of North American minnows (Teleostei: Cyprinidae). Evolution 50: 1280-1292.
Unmack, P. (2001). Biogeograhpy of Australian freshwater fishes. Journal of Biogeography 28: 1053-1089.