| Abstract |
As a theme of the palaeontological exercise in the Department of Geology, Kyushu University, the students and I have been studying some basic biological concepts and various statistical techniques whi...ch seem to be useful in population palaeontology, taking some samples of ordinary bivalves as examples. This is a brief guide for the exercises of this kind and the elementary study of fossil biometry. The following basic techniques are explained with some examples: 1) sampling, 2) mensuration, 3) histogram, 4) analysis of variation and polymorphism, 5) chi square test for the recognition of random sampling and normal distribution, 6) t test for the identification and discrimination of populations, 7) correlation between two variables, 8) linear regression, 9) regression on allometrical equations, 10) evaluation of taxonomic characters, 11) short-cut and other methods for taxonomic identification and discrimination. Many branches of biology including modern taxonomy focus on the population rather than the individual. Because a population cannot be logically represented by a single type specimen, statistical methods on the basis of well quantified data are an indispensable tool for the representation of variation and growth and also for the taxonomic and evolutionary study in palaeontology as well as neontology. Biometrical studies must be founded not only on correct statistical techniques but also on sound biological concept. Among others the population concept supported by modern genetics is especially important for the biometry, because the recognition or assumption of normal distribution of characters is a necessary condition for the advanced statistical study. In the present paper I also emphasize the importance of the study of relative growth for the evaluation of taxonomic characters. One should select the most adequate parameter for each character, especially carefully in order to represent the variation by the combination of two variables, because the relative growth of an organism is not necessarily linear. Some comments are also given on the methods of taxonomic identification and discrimination, especially STUDENT'S t test and MAYR, LINSLEY and USINGER's 75 percent rule. The significance of the result of t test may be a necessary condition for the taxonomic discrimination. It may be a shortcoming for the 75 percent rule that the number of individuals is not considered.show more
|
Export Mendeley
1002_p067