A Consistency Test for Phylogenies Based on Contemporaneous Species
A Consistency Test for Phylogenies Based on Contemporaneous Species
复制标题
基于同时代物种的系统发育一致性检验
DOI：
10.2307/2411550
复制
发表时间：
1965
影响因子：
6.5
通讯作者：
E. Wilson
中科院分区：
文献类型：
--
作者：
E. Wilson
人们普遍认为，从同时代物种的数据推导出来的系统发育无法得到“证明”，因为显然，进化是一个只能从化石中恢复的过去事件。如果不存在具有目击事件的决定性或物理测量的一致性的证据，情况就是如此。然而，科学证据很少是直接的，而且在程度上总是相对的。按照实验生物学的标准，进化假说可能永远不会是确定的，但如果它们既可证伪又具有启发性，那么它们就是有效的。也就是说，为了有效，他们应该做出具体的预测，如果假设是错误的，这些预测就能够被否定；如果它们暂时得到支持，它们应该为更深入、更有意义的调查指明道路。系统发育分类学之所以受到批评，与其说是因为它的间接性，不如说是因为它缺乏形式分析技术，无法使其假设变得可证伪和具有启发性。可能采用的一种此类程序涉及根据其系统发育意义对特征进行“加权”。分类学家直观地选择他们假设的特征状态来定义单系物种集。理想特征包含某种状态，该状态既唯一地定义了一组物种，又在进化中没有被逆转，因此所有拥有这种状态的现有物种都可以说是过去进化出该状态的一个物种的后代。对于每个可以识别的这样的特征状态，可以在系统发生树中添加一个分支。那么，这种极端形式的系统发育假说是作为关于独特的、不可逆转的特征的假说而开始的。该表述也许无法根据同时代物种得到决定性的证明。但它能被反驳吗？通过该程序可以获得任何具有生物学意义的东西吗？以下测试有望对这两个问题给出肯定的答案。它并不是原创，因为它为分类学思维提供了一些非常新的东西。相反，其目的是以一种新的、更严格的形式表达一种常见的直观分类程序。定义。考虑一系列 m 个唯一的、不可逆的字符状态 a1, a2, a3,...。 。 ，am，分别代表不同的性格，因为黄色赘肉和扁平尾巴可以说是蜥蜴物种中两种不同特征（赘肉颜色和尾巴形状）的状态。这些特殊状态被解释为在物种形成时期出现，该物种形成时期导致了 n 个同时代物种的单系分类群。它们现在以任意组合存在于 n 个物种中的各个物种中。在一种极端情况下，某一特定物种可能完全缺乏这些功能；在另一个极端，所有 m 个特征状态可能同时出现在一个给定的物种中。独特的意思是给定的特征状态ai过去只在一个物种中出现过一次。它现在存在于一种或多种后代物种中。未逆转是指在产生同时t轴突的任何物种中，状态从未丢失，即从未恢复到先前状态。字符状态本身可能作为一种新结构从头出现，它可能作为一系列离散字符状态中的新状态出现，或者它可能被任意地识别为连续形态斜线中的某个点或更远的点。因此，该分类单元应被视为一个样本空间，其点是同时代物种；角色状态是事件
It is commonly stated that phylogenies deduced from data about contemporaneous species cannot be "proved" because, obviously, evolution is a past event recoverable only from fossils. This is true to the extent that no proof exists which has the decisiveness of a witnessed event or the consistency of a physical measurement. Yet scientific proof is rarely direct and is always relative in degree. Evolutionary hypotheses might never be definitive by the standards of experimental biology, but they are valid if they are both falsifiable and heuristic. That is, to be valid they should make concrete predictions that are capable of being negated if the hypothesis is false; and they should point the way to deeper, more meaningful investigations if they are momentarily upheld. Phylogenetic taxonomy has been open to criticism not so much for indirection as for its lack of techniques of formal analysis that render its hypotheses falsifiable and heuristic. One such procedure that might be employed involves the "weighting" of characters with reference to their phylogenetic significance. Taxonomists intuitively select character states which they postulate to define monophyletic sets of species. The ideal character contains some state that both uniquely defines a set of species and has not been reversed in evolution, so that all existing species which possess this state can be said to have descended from one species in the past that evolved the state. For every such character state that can be identified, a branch in the phylogenetic tree can be added. This extreme form of phylogenetic hypothesis, then, is initiated as a hypothesis about unique, unreversed characters. The formulation perhaps cannot be decisively proven on the basis of contemporaneous species. But can it be disproven? And is anything of biological significance to be gained by the procedure? The following test hopefully gives an affirmative answer to both questions. It is not original in the sense that it offers something very new to taxonomic thinking. Instead, its purpose is to express one common intuitive taxonomic procedure in a new, more rigorous form. Definitions. Consider a series of m unique, unreversed character states a1, a2, a3, . . , am, each representing a different character, as yellow dewlap and flattened tail can be said to be states of two separate characters (dewlap color and tail shape) in lizard species. These particular states are interpreted to have appeared during a speciation episode that has resulted in a monophyletic taxon of n contemporaneous species. They now exist in any combination in various of the n species. At one extreme, they may be totally lacking in a given species; at the other extreme, all m character states may occur together in a given species. By unique is meant that a given character state ai appeared in the past only once and in one species. It now exists in one or more descendant species. By unreversed is meant that the state has never been lost, i.e., has never reverted to a prior state, in any of the species giving rise to the contemporaneous t-axon. The character state itself might have arisen de novo as a new structure, it might have appeared as a new state in a series of discrete character states, or it might be arbitrarily recognized as some point and beyond in a continuous morphocline. The taxon therefore is to be treated as a sample space whose points are contemporaneous species; and the character states are events