Population diversity in the rainbow smelt, Osmerus Eperlanus Mordax (Mitchill, 1814) (Salmonoidea: Osmeridae) as revealed by canonical and discriminant function analyses on morphometric, meristic and esterase data

• Main Author: Copeman, Douglas Gordon
• Other Authors: Memorial University of Newfoundland. Dept. of Biology
• Format: Thesis
• Language: English
• Published: 1973
• Subjects: Smelts; Newfoundland studies; University of Newfoundland
• Online Access: http://collections.mun.ca/cdm/ref/collection/theses2/id/216630

Thesis (Ph.D.)--Memorial University of Newfoundland, 1973. Biology Bibliography: leaves [227]-231. The rainbow smelt, Osmerus eperlanus mordax was first described by Mitchill in 1814. It can live entirely in a freshwater or marine environment. Freshwater smelt may spawn in streams, on beaches or in relatively deep water (Rupp, 1959; Delisle, 1969). Marine smelt are usually anadromous although they may spawn on the bottoms of quiet brackish bays. There has been considerable debate as to whether or not two forms exist, one large and the other small. -- There is a great deal of evidence indicating that the smelt exists in rather isolated populations. Therefore, they could be expected to be subjected to different sets of selective factors. These, presumably, would exhibit their different effects by modifications in the characteristics of the populations. Conversely, if populations of smelt from different areas are found to be different, then the populations' proposed attributes (isolation and different selective pressures) are substantiated. -- Thirteen smelt populations were sampled during their spawning runs. Morphometric, meristic and biochemical data were recorded for each population. These data were analysed separately and together to obtain a measure of the degree of difference between the populations. Appropriate statistical methods were employed throughout the study. -- It was found that there were differences between the samples on the basis of any type of data and that the differences were real. Individuals could be identified as to population of origin with an average accuracy of 93.4 per cent. The differences between small and normal forms were very large and the marine and freshwater smelt were also distinct. There were also indications that smelt populations vary, not only from place to place but from year to year. Classifications derived from the data indicated that environmental factors may be important in accounting for the observed differences between populations. A considerable amount of redundancy was observed in the data. -- The work forms a comprehensive model for the investigation and identification of populations of a species. Methods are proposed for handling data for which the standard analyses of population genetics are not applicable.