ENREF 33形态测量学(Morphometrics)常用方法及其在微体古生物学中的应用

INTRODUCTION OF MORPHOMETRICS AND A CASE STUDY ON FUSULINID FORAMINIFERA

形态测量学用定量化的方法和手段描述、研究对象形态特征。几何形态测量学是形态测量学的分支,用界标点或轮廓线等标识研究对象的形态,并运用多变量运算进行量化分析和判别。形态测量学方法在生物学与古生物学领域的应用日益广泛,在生物的个体发育、系统演化、类群判别等方面均可发挥重要作用。本文重点介绍了目前在形态测量学中常用的数据类型、分析方法与步骤,即线性距离、界标点、轮廓线等数据类型的获取,普鲁克迭加、傅里叶转换、特征形状分析与增强特征形状分析等数据转换方法,以及主变量分析、典型变量分析等常用多变量线性回归方法的原理与分析过程。选择以类有孔虫个体发育与类群鉴定中几何形态测量学的应用为例,解释最常用的多变量线性回归方法——主成分分析与典型变量分析的应用与结果剖析。

Morphometrics refers to the quantitative analysis of forms including both size and shape, commonly per- formed on both living organisms and fossil records. Geometric morphometrics especially has proved valuable over the decades for its comprehensive description and illustrative property. It employs a set of data, including landmarks or semilandmark-configured-outlines, to quantify forms and then apply multivariate statistical methods to their matrices for further discrimination. The configurations composed by the landmarks and semilandmarks capture more details about the geometry of the objects, and thus could be used to address more biologic or paleontological hypotheses rela- ting to morphology. The basic concepts of morphometrics such as linear distances, landmarks, semilandmarks and outlines are elaborated here, and then the commonly used multivariate statistical analyses, such as Principal compo- nent analysis （PCA） and Canonical variate analysis （CVA）, are explained as to their mathematical backgrounds and benefits. Procrustes superposition, Fourier analysis, Eigenshape analysis and Extended Eigenshape analysis are deci- phered as the methods that could deal with outlines. The differences among these methods are further clarified to emphasize the limitations in their application. An extended eigenshape analysis on fusulinid foraminifera is cited in the final part as an example to present the standard procedure of an outline-based analysis, as well as the application of PCA and CVA.