Chapter 9 Chapter 5 Classification by Common Ancestry-1

Taxonomists who classify empirically have no explanation for why they are able to group species by "relationship" or "similarity".When Sttickland (1840) defined resemblance as "the relation existing between two or more members of a natural class, in other words, the agreement of basic characters," he did not give the keyword "natural (of )" and "Basic (of)" are defined.It was Darwin who filled this explanatory gap and showed why there are natural kinds and why they share "basic" traits.It was Darwin who proposed the basic theory of biological classification.No one before him had stated so definitively that members of a taxon were similar because they were descended from a common ancestor.In fact, this is not a completely new idea. Buffon once thought that similar species, such as horses and donkeys, or various cats, may have descended from an ancestor species. Erasmus Darwin and some evolutionists in Germany have similar ideas. idea.In his later years, Linnaeus also thought that members of higher taxa might be the result of hybridization.But neither Buffon nor Linnaeus turned this speculation into a theory of taxonomy or a theory of evolution.When Pallas (1766) and Lamarck (1809, 1815) successively proposed the dendrogram of kinship, it also had no effect on hierarchical classification (Simpson, 1961: 52).

Few people know that Darwin was the founder of the entire field of evolutionary taxonomy.As Simpson rightly puts it, "Evolutionary taxonomy is unreservedly and almost exclusively derived from Darwin." This is to say that the doctrine of common ancestry not only automatically accounts for the vast majority of similarities between organisms (and it does ), and Darwin also established a well-thought-out doctrine, elaborating methods and unresolved problems.Chapter Thirteen (first edition, pages 411-458) is all about his taxonomy.Chapter Thirteen begins with the oft-quoted passage: "Since the appearance of life, all living beings have been in decreasing degrees of likeness to one another, so that living beings can be divided into classes and subclasses. This classification is obviously not like the random classification of stars categorized as such" (p. 411).Here Darwin implicitly contradicts the often-heard remark at the time that classification (which by 1859 had reached a level of considerable refinement) was an arbitrary and artificial product of the taxonomists.He went on to write:

Naturalists attempt to arrange the species, genera, and families within each class into what are called natural systems.But what are the implications of natural systems?Some scholars regard it merely as a scheme in which those beings most alike are grouped together, and those least alike are separated... But many naturalists consider natural systems to be more than that; They believe it shows the Creator's plan; but unless it is fixed in time or space, or the Creator's plan is otherwise, it does not seem to me to add anything to our store of knowledge... I It is thought (in our classification) to include more than mere resemblance; the approximation of kinship (blood) (the only known cause of resemblance between organisms) is the bond, which by varying degrees of modification is Hidden, partly revealed to us by our classification. (p. 413) In and in his letters, Darwin repeatedly emphasizes that "all correct classifications are systematic (genealogical) classifications" (p. 420), but "genealogy does not itself make classifications" (Life and Correspondence of Darwin )” Indeed, Darwin believed that “in order to compose a natural system it is necessary to arrange some classes (in proper subordination and relation to other classes) in each class in strict accordance with the system (genealogy)”; is not the whole question, "certain classes, although related by an equal degree of consanguinity to their common ancestor, may vary in degrees of difference, owing to the different processes of modification which they have undergone; and Expressed in separate arrangement into different genera, families, groups (sections) or orders" (p. 420).This is a very important text because it points out the main difference between the two schools of modern taxonomy; cladistic and evolutionary taxonomy, which will be introduced separately below.

When Darwin talked about this, he turned back to his famous phylogenetic diagram ("Origin", page 116), in which there are three species (A, F, I) of the same genus (A, F, I) in the Silurian period of the Paleozoic Era. The taxonomic status of the modern descendants of the three species is quite different.The phylogeny derived from species F, with minor changes, remains in the Silurian genus, but its two sister species, A and I, are now in different families and even different orders.The principle adopted by Darwin in the classification of cirripedes is to determine the classification level according to the distance from the bifurcation point (he chooses the principle that is farther from the bifurcation point rather than closer to it).

Darwin spent nearly eight years studying the classification of cirripedes (barnacles), which gave him a deeper understanding of classification both in theory and in practice (Ghiselin, 1969), and thus proposed a Series suggestions to help taxonomists discern what similarities are most useful in determining a "Propinquity of descent".In particular, he repeatedly emphasized the importance of weighting the taxonomic value of all traits: Some people may think this way (this is how the ancients thought): the part of the structure that determines the habits of organisms and the general status of organisms in nature is very important in classification.Nothing could be more wrong.... It may even be taken as a general rule that the less any part of a structure is related to a particular habit, the more important it is for classification (414, 425).

It is particularly important that Darwin opposed a widely popular idea among botanists (and zoologists after Cuvier) in the 17th and 18th centuries, that is, the more important a certain structure is to the survival and continuation of organisms, the more important it is in classification. The more important it is.He cites case after case (pp. 415-416) to refute this view. "The mere physiological importance of an organ does not determine its classificatory value, which is almost proved by the fact that in the class concerned the same organ ... has almost exactly the same physiological importance, but its classificatory value It is very different" (p. 415).He illustrated this by taking as an example the very different taxonomic value of antennae as a taxonomic trait in different families of insects.

Darwin's opinion was not a repudiation of natural selection.What Darwin had in mind was that special adaptations relate to only a very limited portion of the genetic endowment of a class of organisms, and thus provide less information than the overall body size.Moreover, particular adaptations may have been acquired along several evolutionary lines unrelated to each other; in other words, they show convergence.Awareness of these deficiencies in specific adaptations can allow taxonomists to avoid interpreting convergence as evidence of common ancestry.Other characters, sometimes seemingly insignificant, are more significant: "The taxonomic importance of insignificant characters depends mainly on their relation to some other (more or less important) character." In Natural History The (taxonomic) value of combining multiple traits in theIt seems that Darwin was fully aware of the significance of consistent changes in several traits.After discussing the special properties of other characters, such as embryonic characters, stunting characters, and distribution characters, Darwin came to the following conclusions:

The rules, by-laws, and difficulties in classification have been explained above on the basis of the following view.... Natural systems are based on modified relatives, and the naturalist considers that the character showing a real similarity between any two or more species is descended from a common ancestor, and all correct classifications are systematics; the community of kinship is the hidden bond that the naturalist unconsciously seeks. (420 pages) Darwin proposed several rules in the evaluation of characters, some of which have already been mentioned.Like most taxonomists of the last century such as Ray, de Jussieu, De Candol, Lamarck, Cuvier, etc., Darwin also emphasized the high taxonomic value of "stable" (unchanging) traits throughout many classes .And he also appreciates the importance of complexes of related traits, as long as they are not simply the result of the same lifestyle.He devotes almost a section to spurious similarities caused by convergent evolution (p. 427) and warns taxonomists not to be led astray by such "similarities or adaptive similarities".

Theoretical discussions of evolutionary classification in the ensuing century were little more than footnotes to Darwin.Not one of Darwin's rules or principles was denied, nor were any results of particular importance added.Two of Darwin's proposals are particularly important.One is to distinguish resemblance arising from common ancestry from spurious resemblance arising from convergence.A trait such as chorda has high taxonomic value because it is part of a composite trait system that is almost impossible to occur twice independently.Segmentation, on the other hand, is not such a fundamental trait, since there is plenty of evidence that it occurs at least twice in the animal kingdom.It is difficult to imagine any phylogenetic link between vertebrate and arthropod segments.

Another Darwinian suggestion was to "weight" characters.Such evaluations are important because some traits are more informative than others.Phylogenetic weighting, implemented by Darwin, is an empirical (or inductive) weighting method that traces effects back to causes.The weight of a trait is determined by its association with the most firmly established (testable in many different ways) part of the taxonomy.Some taxonomists find it difficult to distinguish this kind of weighting from the given (or deductive) weighting of cause to effect (such is the case with Cechapinault and Cuvier).But it can be done with proper analysis, and this empirical weighting has been repeatedly emphasized (Mavr, 1959a; Cain, 1959b) and found to be a useful method (Mayr, 1969), and is currently It has been integrated with the computer weighting method.

The reasons for the large difference in the amount of information collected by taxonomic traits have not yet been definitively concluded.But it has been suggested that this is the result of some elements of the phenotype being more firmly incorporated into the genotype than others.The more genetically entrenched a trait or complex of traits is, the more useful it is for showing interrelationships. Schmalhausen, Waddington, Lerner showed that the structure of the phenotype provides such a stable integration of the genotype that certain components of the phenotype can remain unchanged during phyletic divergence.Fundamental canalization and regulatory mechanisms seem to have been accidentally left untouched during evolution, which accounts for the unanticipated stability of sometimes seemingly insignificant phenotypic components. As far as the methodology of classification is concerned, the impact of the Darwinian revolution was modest.The real turning point in the history of taxonomy was the abandonment of essentialist ideas and "descending classifications", and this was largely accomplished long before 1859.Darwin's decisive contribution to taxonomy was twofold: through his theory of common ancestry he provided a theoretical account for Linnaeus' hierarchical structure and the homogeneity of taxa in "natural" classifications.On the other hand, he revived, at least in principle, the notion of continuity between species of being which had been denied by Cuvier and the School of Natural Philosophers in their doctrine of primitive progenitors.Some aspects of Darwin's contribution are described in more detail below. As mentioned earlier, between the 1820s and 1840s, the quintalists and some other zoologists and botanists had clearly recognized that there were two kinds of similarity among living beings.The similarity between whales and land mammals is similarity, and the similarity between whales and fishes is the same.Owen Strickland, who has a higher insight among these scholars, believes that the same function is due to similar functions, but they cannot explain the similarity and can only resort to "the creation plan of the Creator".Darwin solved this problem simply by proposing that resemblance is the proximity of descent.This leads to the hypothesis that all taxa should contain descendants of the nearest common ancestor, or, in Haeckel's term, these descendants should be monophyletic (monophyletic).In order to distinguish these taxa it is necessary to scrutinize all similarities and differences in order to distinguish between characters due to a common ancestor (the only useful traits for classification) and characters that are homologous (convergent) (for example, the hooks of hawks and falcons and the hooks of waterbirds). webbed feet, these were acquired independently due to similar functions). After comparative anatomists in the early nineteenth century rejected the perfection (sex) hierarchy, it turned out that there were as many unrelated units as there were archetypal types.In fact, he was still exploring which is higher and which is lower, which is reflected in Agassiz's speech to his students: "As long as you can find any fact that one order is higher than another, it is a real scientific research." Darwin's interpretation of Linnaeus' hierarchical structure as due to common ancestry not only restored the principle of continuity but was a vast research project.No one saw this more clearly than Haeckel, whose desire was to link all taxa of animals and plants on the basis of their affinities and represent them in a tree, which has since graced systematic textbooks few.Haeckel was also an artist, and he represented the phylogeny very strictly with lifelike trees, which were gradually replaced by so-called dendrograms, which contain early dendrograms. The relationship between taxonomy and putative phylogeny (geneagenesis) has been debated from 1859 until the present. As early as 1863 T.H. Huxley rejected all considerations of phylogeny and required all classifications to be based on "purely structural reasons... Such classifications are of the utmost importance as accounts of structurally related rules of thumb." In this Huxley differs markedly from Darwin in that Darwin's principle is that observation cannot be made without theory.The modern tendency is to apply Darwin's principles to study each taxon to see if the traits of the species contained in it indicate that the taxon is unitary, that is to say to consistently formulate phylogenetic hypotheses and then check whether these hypotheses are supported by taxonomic evidence support. As far as Haeckel was concerned, there was no doubt that classification had to be based on phylogeny, and once the phylogeny was known, the phylogeny was also clear.Therefore, the main task of classification is to develop methods that can represent phylogeny.Of these approaches, the one that most struck Haeckel and his contemporaries was the theory of reenactment. (Gould, 1977).The replay theory states that ontogenetic stages replay their ancestral growth stages or ontogeny replays their phylogeny (see Chapter 10).This theory is now known to be untenable, yet it was highly instructive, leading to comparative embryology and many remarkable discoveries, most notably Kowalewsky's work on the larvae of sea squirts (tunicates) Cords were found in the sea squirt, thus proving that sea squirts are chordates and not molluscs as previously thought.Mammalian embryos have the gill slits of their fish ancestors (discovered by H. Rathke in 1825) and many similar findings in comparative embryology, suggesting that the replay theory should be revised. The main point of the revised replay theory is that embryos often replay their Ancestral embryonic stage. Between 1820 and 1859, comparative embryology discovered the typical crustacean larvae of barnacles, clarified the life history of some parasitic crustaceans, and proved that the sea disc car is the head of the crinoid.Agassiz emphasized the importance of comparative embryology as a supplement to comparative anatomy in his Lowell Memorial Lectures in 1848-1849.Since 1836, Agassiz believed that there is a triple parallel relationship between fossil history, embryonic development and taxonomic rank. (Winsor, 1976b: 108). The main result of the studies of comparative anatomy and comparative embryology is that the "unnatural" classes or phyla of animals are successively transformed into "natural" by the removal of irrelevant components, such as barnacles and tunicates from molluscs Get out of the way, and most importantly, blow out Cuvier's radiation gate.The research activity pioneered by Cuvier and Lamarque exploring the interrelationships between the major groups of invertebrates has progressed roughly as much fifty years before publication as it did fifteen years after publication.Detailed morphological analysis has contributed more than phylogeny to the identification and division of natural taxonomic units.But what zoologists were most passionate about in the second half of the nineteenth century was designing phylogenetic trees. The theory of evolution has even less influence on the classification of plants than that of animals. The principle of the "natural ladder", the progression from simple (primitive) to complex, remained for quite a long time the unconscious guiding principle of botanists.The taxonomy of flowering plants (angiosperms) has been constrained in two different ways: First, plant classification is based almost exclusively on flower structure. Only in the last three to forty years have wood anatomy and chemical composition been seriously considered as useful traits.Secondly, the concept of what is the most primitive flower is not clear.It has long been assumed that angiosperms are wind-pollinated and have no petals, so in extant families such as Betaceae, Fagaceae, and related families (such as Catkins), all wind-pollinated plants are considered to be most original.It is now recognized that wind-pollinated pollination and associated flower transformations are secondary, and that an entirely different family, related to the Magnoliaceae and Ranunculaceae (Ranunculaceae), is the most primitive.The missing plants between them and seed ferns are as yet unknown in the fossil record. The most profound influence on the development of botany in the 19th century was Hofmeister's research on the reproductive life cycle of cryptogamous plants and the homology of their reproductive organs.These studies not only made a clear description of the mutual relationship between cryptogamous plants for the first time, but also broke through the insurmountable barriers between cryptogamous plants and phaemogamous plants. Hofmeister's research definitively demonstrates that there is a somewhat consistent scheme of body composition throughout the plant kingdom.His "Comparison of Cryptogamous Plants" (1851) laid a solid foundation for the establishment of cryptogamous plant phylogeny after 1859. Once the characters of the different classes are ascertained, it is not difficult to classify them by means of the principles of the doctrine of common ancestry. It was not long before the main attention turned to studying changes in the reproduction of different classes of cryptogamous plants and the interrelationships between these classes.Nothing needs to be figured out more than highly promiscuous fungi.The great pioneer of this type of research was Anton deBary (1831-1888), who concluded between 1866-1888 with a large number of detailed analyzes of the life history of some different types of fungi, which laid the foundation for the effective research work of his successors. established a solid foundation.Some scholars in modern times recognized the importance and uniqueness of fungi, and separated them from plants into a separate fungal kingdom. 5.1 The Decline of Macrotaxonomy After the 1880s, macrotaxonomic research and research on phylogeny (phylogenetics) gradually and perceptibly declined.There are many reasons for this, some internal to the research field and some external.Perhaps the most important of these is the frustration with the difficulty of achieving definitive results.Similarity is usually a reasonable and accurate indicator of relatedness in the classification of suborder taxa.In the classification of higher taxa (orders, classes, phyla) similarity is no longer a reliable guide and research progress is very slow.People who do not work in taxonomy must be surprised that the relationship between living things is still unclear.For example, it is still not known what the vast majority of orders of birds are called Proximate Orders.The same is true of the many orders, families, and genera of animals, such as the Dimorphodonts, Tuliodonts, Heterodonta, and Treeshrews. But this ambiguity in the classification of higher vertebrates pales in comparison to that of invertebrates, lower plants, prokaryotes, and viruses.When reading recent articles discussing the taxonomy of lower invertebrates it is startling to see that some of these issues are still debated in the 1870s, 80s, and 90s.While it is usually the majority opinion that prevails, the fact that unorthodox minority opinions tend to receive active support also suggests that uncertainty is still quite common.To add a little spice to the nature of the still-debated issue I may also pose here a few questions: From what class of protozoa did metazoa evolve?Did all metazoans have a single protozoan ancestor, or did sponges evolve separately?Are coelenterates or planarians the most primitive metazoan above sponges?Is the division of higher invertebrates into protostomes and deuterostomes a natural classification?If these two categories can be identified.So which category should tentacle animals fall into?Is the theory of protocoelomates correct? Many questions about the genetic relationship between arthropod taxa have not yet been resolved, and whether arthropods are derived from annelids is also inconclusive. Kerkut argues that attention should be paid to these ambiguities, which of course only experts in the field are most aware of.This is actually the history of the evolution of ideas, and it is impossible to even begin to compile a history of taxonomic order for the various higher taxa of animals and plants proposed in the past two hundred years.With each epoch new hope arises due to the emergence of new taxonomic principles (such as reenactment) or newly discovered traits, but progress has been slow. Attempts to establish kinship between the major phyla of the animal kingdom failed repeatedly, leading at least one prominent zoologist to publicly disavow the idea of ​​a common ancestor at the turn of the nineteenth and twentieth centuries. Fleischmann calls the doctrine of common ancestry a beautiful myth without any basis in fact.Fifty years later Kerkut, though not reaching such extreme conclusions, was almost as pessimistic about eventually being able to learn something about the relationship between higher taxa of animals.We must honestly admit that we are still very ignorant, if not extremely ignorant, about these relationships. This is naturally a frustrating situation, since more than a hundred years have passed since the period of intense tree design after publication.Morphological and embryological cues are clearly insufficient for this task. The second reason for the gradual decline of macrotaxonomy after Darwin is conceptual confusion.When Haeckel and his followers insisted that only phylogenetic classifications are natural classifications, his opponents countered by asking the question: How can we understand phylogeny?Isn't the phylogeny inferred from the facts on which the classification was made?So how can classifications be formulated in terms of phylogenies without getting caught up in a hopelessly circular debate?This debate has only been resolved in recent years.Neither phylogeny by taxonomy nor taxonomy by phylogeny.Both are based on the study of "natural classes" in nature, that is, classes of complex traits that descendants of a common ancestor should exhibit.Both classification and phylogeny are based on the same comparison of organisms and their characters, and on careful evaluation of similarities and differences between biological organisms (Mayr, 1969).Current evolutionary taxonomists also agree that biological classification must conform to the phylogeny made by inference.The clarification of concepts paves the way for the classification of higher taxa. The decline of macro taxonomy after 1900 also has its external reasons.As the Mendelians claimed that mutations could form new species, taxonomic attention turned to microtaxonomy (on the "species question"), and finally to neosystematics.Because subspecies are generally regarded as incipient species, many specialists, especially those of birds, mammals, butterflies, and snails, devote all their energy to describing new subspecies.Focusing on the species level has resulted in the discovery of countless undescribed species.As a result, macrotaxonomy has been given a cold shoulder. Perhaps the single most important factor in the decline of macrotaxonomy has been the growing competition from other sub-artistic disciplines of biology.With the astonishing discoveries in experimental biology (embryology, cytology, Mendelian genetics, physiology, biochemistry), most of the bright young biologists are turning to these departments, creating a shortage of taxonomic talent, Difficult situation of financial constraints. In 1957, at the symposium on "Modern Systematics" held in Uppsala, Sweden, to commemorate the 250th anniversary of Linnaeus' birth, only 4 of the 29 papers involved macrotaxonomy.This clearly demonstrates that the main interest has shifted to the species level, which was characteristic of most taxonomists in the first half of the twentieth century.But the classification of higher taxa continued at this stage (albeit quietly), and some important treatises dealing with taxonomic concepts and problems were published, such as Bather (1927), Simpson (1945), Renxi (1947) , Huxley (1958). In the 1960s, most of the tasks of new systematics in micro taxonomy have been completed (at least in terms of concept development), and the time has come to re-emphasize macro taxonomy. Darwin's good start in developing the theory and methodology of macrotaxonomy was largely ignored in the period after Darwin.The criteria by which genera, families, and orders are identified and compounded into higher taxa are extremely inconsistent among different types of organisms.Single-trait "classification"—more correctly, identification schemes—remains popular for poorly understood biological classes.Since different scholars may have chosen different key characters, the same debates as prevailed in seventeenth-century botany also developed.Taxonomists often propose new classifications without good reason, simply saying that the new classification is "better" and done.As far as Linnaeus was concerned, the names of higher taxa were merely for ease of memory, but this purpose has been completely ignored by zoologists and botanists who break down genera and families into smaller parts.The situation went so far that, for birds, for example, some taxonomists in the 1920s and 1930s assigned almost a separate genus to each species.There is no standard at all in applying the levels of classification, and one eminent ornithologist arranges the family of birds in twenty-five orders, while another equally eminent scholar arranges them in forty-eight orders. .Anyone looking at macro taxonomy from the outside world (such as from some applied sciences such as medicine, agronomy or ecology) will feel that taxonomy is too messy, and in fact it is. However, the situation is not completely dark.There are at least some useful textbooks on the theory and practice of zoological systematics, such as Ferris (published in 1928, reproduced hereafter), Renxi (1934), Mayer, Linsley, Usinger (1953).Occasionally, some insightful treatises on taxonomic theory can be found in the vast literature, such as Mayer's (1942:280-291) article on the meaning of genus and, more importantly, Simpson's (1945) article on Works on macrotaxonomic theory.The most constructive at this stage is to assign ecological significance to higher taxa.At that time, it was discovered that higher taxa were composed of species occupying special habitats or adaptation zones (adaptation zones) in the form of compound species.In other words, the overarching problem has shifted from identifying the morphological traits of higher taxa to the biological significance of higher taxa in nature.As far as biologists in general are concerned, however, a serious problem has arisen with classification (to put it mildly).New systematics (mainly focusing on the level of species) does not meet the needs of macro taxonomy, and must find a way out from other aspects. In order to solve these problems, two completely different and independent classification methods have emerged: numerical classification, also known as numerical phenetics, and cladistics.These two new taxonomies are not proposed as improvements to existing taxonomies, but are actually revolutions in taxonomy. 5.2 Numerical phenetics Part of almost any classification method is to organize objects into groups of similarities.However, empirical taxonomists since Adanson have formulated biological classifications according to the weighting method of empirical induction (later Darwin theoretically demonstrated) requiring considerable knowledge and experience.The question naturally arises, therefore, whether it is possible to devise a method by which even an inexperienced person, a non-biologist, can classify species into "natural" genera and higher classifications. unit.Indeed, an automatic and objective method independent of the will would be useful even to an experienced taxonomist in choosing the best scheme of classification.The basic idea of ​​this method is to convert a qualitative or subjective taxonomy into an objective, numerical taxonomy by comparing similar degrees quantitatively. There is no work on the history of numerical taxonomy.But pioneering work in this area can be traced back to the mid-19th century, although most of this work at the time involved intraspecific variation, especially geographic variation.Guides to articles attempting to apply numerical methods to the classification of species, genus, and even higher taxa are usually submerged in the vast taxonomic literature and known to only a few specialists.For information on this aspect, please refer to "Quantitative Zoology" (Quantitative Zoology, 1960) co-authored by SimPson, Roeand Lewontin. Among the almost completely forgotten pioneers was the geneticist A. H. Sturtevant (1939; 1942).He took great care to avoid bias and to exclude from his calculations any traits already known to be relevant to adaptation and development; in analyzing 39 traits in fruit flies he was able to arrange 58 species of fruit flies into relevant categories, and more importantly Interestingly, he also made generalizations that have since been confirmed repeatedly.The first of these generalizations states that rigorous numerical methods are most reliable when applied to closely related species, but tend to produce contradictory results when applied to distantly related species.He also devised a chart showing the relationship between different traits and found that some of them were "best" because they could indicate the approximate nature of other traits; ) change (change). Since the invention of the electronic computer, three groups of taxonomists have independently proposed the use of computer methods to quantify similarity and classify species and higher taxa with the help of this quantitative method; these three groups are the C.D. · Michener and R. R. Sokal (1957), P. H. A. Sneath (1957), a bacteriologist in London, England, and A. J. Cain, G. A. Harrison, Oxford, England (1958).The most important aspect of their proposal is to replace the synthetic (integrating) capacity of the human brain (which in traditional taxonomy simply groups taxa by checking or tabulating similarities) with the mechanical manipulation of a computer.They believe this will replace the usual arbitrary and subjective evaluations with objective and consistently repeatable methods.At first the three groups agreed that all traits are equally important, however, soon Cain and Harrison (196O) found that different traits have different information and suggested "line weighting" (phyletic weighting). Michener also quickly stepped back from his earlier proposal, but the two remaining pioneers, Sokal and Sneath, joined forces in the classic 1963 book Principles of Numerical Taxonomy (Principles of Numerical Taxonomy) Their method and rationale are introduced.The title of this book is misleading. Simpson and other taxonomists have pointed out that numerical methods have been used in taxonomy for a long time, and some schools with great differences in taxonomy have used them. Therefore, it is customary to refer to the classification methods of Sokal and Sneath Called "numerical representation (classification) method" (numerical Phenetics).Unfortunately, this new approach was initially overhyped and later fell short in some respects.For example, it was initially claimed that any two scientists working completely independently using the new method would make exactly the same estimate of the similarity of two organisms, provided they each provided the same set of traits.This statement is obviously unrealizable, and it has aroused objections among experienced taxonomists.Many important improvements were made in the book's thoroughly revised second edition (1973).Other textbooks on numerical taxonomy include Jardine and Sibson (1971), Clifford and Stephenson (1975). Throckmorton's (1968) book also deals with numerical classification, but treats it differently. As Darwin once pointed out, different traits contain different amounts of information, and when different combinations of traits are selected, very different classifications will result.躯体的不同部位，生活史中的不同阶段，形态学性状或生物化学性状都会对类似性作出不同的估价。为了夸示他们的客观性，数值分类学家提出抛弃种作为分类单位而代之以“运算的分类单位”（operational taxonomic units，缩写为OTU），似乎这就是改进。然而事实上这又引起了与导致放弃模式种（typological species）概念同样的实际困难。数值分类学家或者是必须将不同的性别，年龄段和形态看作是不同的OTU，然后将雌、雄以及其它极不相同的表现型分入不同的分类单位，要不然就必须非常仔细地分析生物学变型（同型种，phena），并将变型组合成与生物学种相符的OTU。这样对变异作评价虽然更加逼真，却恰恰要求主观判断，而这种主观判断正是“客观的”数值分类法所要加以排除的。 传统分类学家和数值分类学家之间最重要的区别在于他们对加权的态度。对加权的态度只有三种可能性。第一种是认为一切性状都是相等的，也就是说在分类中它们同等重要。虽然数值分类学家将之称作是“不加权”法，但是这当然是既定的加权方法，即给每个性状同等地加权。这和亚里斯多德，切查皮诺以及居维叶的既定加权法一样容易引起误解。海洋无脊椎动物是否有索这一性状的分类价值比一百个其它性状的价值都要高。某些性状含有大量的关干亲缘关系的信息而其它的则仅仅是“噪音”这一点早在二百多年以前阿丹森就曾指出过。 涉及加权的第二种可能性是有一套固定的标准（例如生理上的重要性）衡量不同性状的相对分类学重要性。这实际上就是亚里斯多德和居维叶的既定加权法。第三种可能性是凭经验加权法，这种方法首先将生物安排到表面上看来是自然的类别中（通过对很多性状或复合性状的考虑）。然后对与最自然（最合适）类别看来相关的性状予以最大的加权。这就是达尔文的处理办法，他最后归纳起来这样说：“一些微不足道的性状在分类上的重要性主要取决于它们和其他性状（多少是重要的）相关”（： 417）。 在全部分类学史中实际上所有有经验的分类学家都知道而且经常强调不同的性状具有多么不同的分类学价值。以大脑半球结构为主要依据的类人猿和人的分类和以主要生物高分子（如血红蛋白等等）为根据的分类将会有所不同。年轻一代的数值分类学家觉察到各种不同的性状的信息量差别悬殊，目前正集中力量用客观的经验性加权（例如通过相关分析）来代替直觉的主观评价。 数值分类学家将许多个别性状类似程度的总和转变成单一的总体类似值（Overallsimilarity value）或“表征距离”（Phenetic distance）。然而正如辛普森（1964a）曾经指出：“一个单一的类似性衡量尺度是以损失大量信息为代价取得的，这里所指的信息主要是性状变化的顺序以及差异的来源。”在将高度复杂的实体加以比较时正象把不同分类单位的复合性状作比较一样，根本不宜于将类似性定量化。这就是为什么数值分类学曾经被称为模式方法的原因，也是为什么辛普森认为数值分类学导致了“分类学原则的倒退…有意识地恢复了18世纪原则”的原因。 如果数值分类方法能取得实际结果人们往往就可能忽视它概念上的弱点。然而，为了部分地补偿镶嵌进化（mosaic evolution）以及由于引用了不含信息量的性状所产生的“噪音”，数值分类学家便必须为极大数量的性状（最好是超过一百）编制程序。在形态上极为复杂的节肢动物（如昆虫，蜘蛛等）中一般可以找到如此大量的性状，但是在绝大多数其他生物中分类学上有用的性状就极少。单是这一点就妨碍了这种以性状不加权为基础的方法的运用。而且即使在昆虫中使用这种方法也非常费事，为大量的分类单位的一百多个性状编制程序要花费大量时间。由于这个原因数值分类学派的先驱者之一Michener在为澳大利亚蜂（包含有很多新种）的大量标本分类时便仍然采用了传统的分类方法。 现在，在数值分类学的原理首次被提出约莫25年以后，就有可能对这种方法的可行性和用途作一尝试性的暂时结论。显然，一切分类的基本观点都是表征（分类）性的，都是力求确立“类似”实体的类别。这种努力成功与否取决于确定类似性的原理和方法。 在这一方面来说数值分类学由于坚持对性状同等加权以及完全忽视系统发育因而完全失败了。 但是，数值分类学基本原理的失败并不能作为理由去否定由数值分类学家所首创并采用的很多数值方法，特别是多变量方法的有效性。数值分类学家首创的这些方法目前在很多科学领域中已广泛应用，在数据选择与分类的一些其他领域中也极其重要。在分类学中提倡并介绍这些方法应当看作是数值分类学家的最重要贡献。另外，正象最优秀的分类学家所一贯支持的那样，数值分类学家也十分强调运用尽可能多的不同性状和性状系统的原则以便取得新资料。 数值分类法在为大的属中种的归类和为前此混淆不清类别的分类中最为有用。另一方面，在分类已趋于完善的类别中或为目、纲，或门这些层次的分类中还没有发现数值分类作出过什么实质性的贡献。 数值分类学最有希望的未来发展可能在于进一步发展加权程序。这些程序或者是依据性状的相关变异（covariation），或者是以某些经验性指导原则为根据。由推论而知的共同祖先的后存几乎毫无例外地可以由共同具有某些性状而查知，因而对一些性状较之其他性状予以更大的加权就是一种常识。任何分类方法不运用性状加权显然是无效的。 为了力求“绝对客观”，数值分类学派完全不考虑任何亲缘证据，而与之对立的支序分类学派（cladistics）的主要特点却正是以亲缘（家系）为基础． 5．3支序分类（Cladistics） 支序分类学派提出在分类中应当制定一种排除主观性和随意性的方法，其动机是和数值分类学派相同的。支序分类学派的创导人德国昆虫学家亨尼克在其1950年出版的《系统发育系统学原理》（Grundzuge einer Theorie der phylogentichenSystematik）一书中全面阐述了他的理论和方法。按照他的观点，分类应当完全建立在系谱（血缘、家系，genealogy）的基础上，也就是建立在系统发育（Phylogeny）的分支模式（branching pattern）的基础上。他认为系统发育由一系列二叉分支（asequence ofdichotomies）组成，每个二叉分支代表祖先种分化成两个姐妹（子代）种；并假定祖先种在二叉分支时即不再存在。姐妹群（sister group）必须安排在相同的阶元等级，祖先种及其所有后代必须包括在一个单一的全系分类单元（holophyleticfaxon）中。 亨尼克的着作是用晦涩的德文写成的，其中有些句子令人完全无法理解。书中也从来没有提到赫胥黎，迈尔，壬席，辛普森以及其他一些前十几年在同一研究领域作过部分工作的学者。新的术语和定义也随时不经意地提到，但又没有索引作为寻找出处的指南。无怪乎该书起初并没有引起多少注意（除极少数德国学者而外）。直到1965—1966年英文译本出版后才逐渐受到重视。到了70年代事实上已发展成为对亨尼克的个人崇拜，虽然他的某些所谓追随者已经远远背离了亨尼克的原来原则。 虽然亨尼克曾经将他的方法定名为系统发育系统学，但是他只依据系统发育的一个单一的组分，即系谱分支（branchingof lineages），因此后来其它学者将之重新命名为支序分类，这也就是目前通用的名称。 支序分类中极其重要的一个方面是在有关分类单位的比较中仔细分析所有性状（特征）并将这些性状区分为祖先特征（祖征，PlesiomorPh）和独特的衍生特征（近裔衍征，apomorph）。系统发育的分支点由共有衍生特征（近裔共征，synapomorphies）的回溯（backwards tracing）决定，因为这样的共有衍生特征（性状）被认为只在祖先的后裔中初次出现该性状时才能发现，支序分类学派认为这种方法用不着借助于化石证据就可以重建系统发育，事实上在一定程度内也的确如此。 自从达尔文以来进化分类学家只承认单系（单元，单源）分类单位（monophyletictaxa），即完全是由一共同祖先的后裔构成的分类单位。某些类群（类别，groups）是否单系群要不断地用新特征（性状）来加以检验看它们是否符合单系要求。这种方法经Hull（1967）论证是非循环性的。自从1950年以来，仔细比较包含在某一高级分类单位中的种和属并分析所有的类似性以便确定它们是否真正同源，结果表明绝大多数已被确认的动物分类单位都是单系的，但就植物来说则并不如此肯定。然而，亨尼克却是首先明确地提出系谱的分支点必须完全依据共有衍生特征这一原则的第一位学者。他曾说过，只有共同具有独特的衍生特征才足以证明某些种来自共同祖先。 从原则上来说划分单系群（monophyletic groups）的支序分析方法是一种绝妙的程序。它清楚地说明了建立亲缘（系谱）共同性的客观标准，它迫使分类学家对一切性状（特征）作详细分析并引进了性状加权的新原则，即共同具有共有衍生特征（共有衍征）。凡是共同具有共有征征的类群就是姐妹群。但是也有不少反对支序分析的意见。 头一个反对意见是术语问题。亨尼克引用了相当多的新术语，其中绝大多数是不必要的，虽然“祖征”和“衍征”这两个术语仍然通用。此外，亨尼克还试图将已经普遍接受的术语改换成完全不同的概念，例如，将“系统发育”这个词严格限制为系统发育的分支部分，将“亲缘关系”完全按与最近的分支点的接近程度来下定义，更糟的是将“单系”这个词从普遍用来标明分类单位改变为系谱进程（process of descent）。从海克尔一直到1950年分类学家的工作顺序点是首先依据表征划分类单位，然后再检查它是否单系，而支序分类学家则只是把一特定种的由谁论得知的一切后裔都合并到一个“单系”分类单位中，尽管这些后裔有如鸟和鳄那样极不相同。 第二个反对意见是确定共同衍征的难度问题。两个分类单位共同具有某一衍征可以有两种可能性。这特征（性状）或者是来自最近的共同祖先（真正的或同源共同衍征，genuine or homologous synapomorphy），或者源于趋同现象（非同源共同衍征或假共同衍征，nonhomologous or pseudoapomorphy）。决定单系群的可靠性在很大程度上取决于是否仔细认真地区分了这两类类似性。很多支序分类学者往往低估了非同源性共同衍征出现的频率。可以用眼睛的进化作为例子来说明某种看来似乎不可能的适应是多么经常的能独立达到。光感受器在动物界至少独立地发生过40次，而在另外的2O种情况下还无法确定有关分类单位中的眼睛究竟是怎样来的。（plawen and Mayr，1977）。还有很多的其它情况都说明将共同衍征划分成同源的和非同源的是多么困难。在单独的系谱中独立地丧失某种性状是特别经常发生的趋同形式。 在确定共同衍征上另一个难于克服的困难是确立进化方向，也就是确定哪个性状（特征）是祖先性状，哪个是衍生性状。例如，在被子植物的无花瓣属和科的安排上就取决于没有花瓣究竟是祖先状态还是衍生状态；或者拿动物界的例子来说，被囊动物既可以看作是原始的，头索类（文昌鱼）和脊椎动物看作是幼态持续（neoteny，即动体繁殖），或者把文昌鱼看作是祖先态而将被囊动物（海鞘）看作是特化了的、次生的固着生物分支。动物和植物的分类系统一旦遇到高级分类单位的安排完全依据进化方向来解释的情况就成为不可解的难题。进化方向发生逆转的情况特别使人恼火，但这种情况比一般所承认的要更常见。