Garland Theodore, Bennett Albert F, Rezende Enrico L
Department of Biology, University of California, Riverside, CA 92521, USA.
J Exp Biol. 2005 Aug;208(Pt 16):3015-35. doi: 10.1242/jeb.01745.
Over the past two decades, comparative biological analyses have undergone profound changes with the incorporation of rigorous evolutionary perspectives and phylogenetic information. This change followed in large part from the realization that traditional methods of statistical analysis tacitly assumed independence of all observations, when in fact biological groups such as species are differentially related to each other according to their evolutionary history. New phylogenetically based analytical methods were then rapidly developed, incorporated into ;the comparative method', and applied to many physiological, biochemical, morphological and behavioral investigations. We now review the rationale for including phylogenetic information in comparative studies and briefly discuss three methods for doing this (independent contrasts, generalized least-squares models, and Monte Carlo computer simulations). We discuss when and how to use phylogenetic information in comparative studies and provide several examples in which it has been helpful, or even crucial, to a comparative analysis. We also consider some difficulties with phylogenetically based statistical methods, and of comparative approaches in general, both practical and theoretical. It is our personal opinion that the incorporation of phylogeny information into comparative studies has been highly beneficial, not only because it can improve the reliability of statistical inferences, but also because it continually emphasizes the potential importance of past evolutionary history in determining current form and function.
在过去二十年中，随着严谨的进化观点和系统发育信息的融入，比较生物学分析经历了深刻变革。这一变化很大程度上源于人们认识到，传统统计分析方法默认所有观测值相互独立，而实际上诸如物种等生物群体根据其进化历史彼此间存在差异关联。基于系统发育的新分析方法随后迅速发展，被纳入“比较方法”，并应用于许多生理、生化、形态和行为研究。我们现在回顾在比较研究中纳入系统发育信息的基本原理，并简要讨论三种实现此目的的方法（独立对比法、广义最小二乘法模型和蒙特卡罗计算机模拟）。我们讨论在比较研究中何时以及如何使用系统发育信息，并提供几个实例，说明它对比较分析有帮助甚至至关重要的情况。我们还考虑基于系统发育的统计方法以及一般比较方法在实际和理论方面存在的一些困难。我们个人认为，将系统发育信息纳入比较研究非常有益，不仅因为它能提高统计推断的可靠性，还因为它不断强调过去进化历史在决定当前形态和功能方面的潜在重要性。
