Introduction and Notes to R. Dahlgren's System of Classification of the Angiosperms

The present paper aims at introducting Dahlgren’s system of classification of the angiosperms. Phenetic and phylogenetic classifications are discussed. The basic principles and methods used by Dahlgren are explained. Dahlgren’s opinions on some important problems, such as the origin of angiosperms, the flowers of primitive angiosperms, the relation between the dicotyledons and monocotyledons, the origin of the monocotyledons, the treatment of the “Amentiferae” and of the orders of the “Sympetalae”, are all expressed. A brief comparison between Dahlgren’s system and three other current systems, viz. those of Takhtajan, Cronquist and Thorne is also given.     

Evidence from nucleic acid and protein chemistry, in particular serology, in angiosperm classification

Semantides, or information carrying molecules, are divided into primary semantides (DNA), secondary semantides (RNA), and tertiary semantides (proteins), with pro–teinaceous material proving to be the most useful in taxonomic and systematic research to date. The contributions to angiosperm classification from amino acid sequencing, DNA hybridization, and fraction I protein are still limited; however, serological data (600 taxa in 200 taxonomic publications) substantially (but unequally) influenced the most recent revisions of systems of classification presented by Cronquist, Dahlgren, Takhtajan, and Thorne. The analyses of semantides and amino acids have thus proven to be unequally useful in helping to understand evolution and phytogeny, and in contributing to classifications. The continuing incorporation of data from diverse disciplines into the development of systems of angiosperm classification makes multidisciplinary and multiauthor research (team research) more essential as we move into future research planning.     

The "new" chemosystematics: phylogeny and phytochemistry

For almost a decade it has been acknowledged that the flowering plant dichotomy of monocotyledons and dicotyledons does not reflect the evolution of angiosperms. Despite this, conclusions in the field of chemosystematics are still drawn from, and rely on, non-phylogenetic botanical classifications such as those of Cronquist, Dahlgren and Takhtajan. In this paper the two alkaloids colchicine and camptothecin are used as examples of how phylogenetic systematics may be applied to alkaloid chemosystematics.     

A consensus classification for the order Gentianales with additional details on the suborder Apocynineae

The Gentianales as circumscribed by Benson, Cronquist, Dahlgren, Goldberg, Hey wood et al., Melchior, Stebbins, Takhtajan, and Thorne is investigated. From these a consensus classification (not in the cladistic sense) or classificatory model for the order is proposed. This classification is discussed, as are the relationships of the taxa within it. Excluded taxa are also discussed. In particular, the Apocynineae is examined in detail and it is pointed out that for this suborder the present consensus classification, at the family level, is congruous only if the taxa involved are monophyletic. Data are then supplied to show that, as yet, neither monophyly nor paraphyly can be proved conclusively for the taxa of the Apocynineae. Cet article examine l’ordre des Gentianales tel qu’il se décrit chez Benson, Cronquist, Dahlgren, Goldberg, Hey wood et al., Melchior, Stebbins, Takhtajan et Thorne. A partir de ceci un schéma classificatoire est proposé pour l’ordre. Cette classification convenue est discutée, ainsi que les relations des taxa y figurant. Les taxa exclus sont également discutés. L’ordre des Apocynineae subit un examen particulier et il est avancé que pour ce sous-ordre la classification convenue actuelle, au niveau de la famille, n’est convenable que lorsque les taxa en question relévent d’un seul phylum. Finalement les données sont appliquées pour démontrer qu’il n’existe jusqu’ à présent aucune preuve définitive quant à la monophylie ou la paraphylie des taxa de l’ordre des Apocynineae.

---

An erratum to this article is available at .     

A Phylogenetic Analysis of Families in the Hamamelidae

A cladistic analysis of the families in the Hamamelidae is made in the present paper. As a monophyletic group, the subclass Hamamelidae includes 19 families, namely, the Trochodendraceae, Tetracentraceae, Cercidiphyllaceae, Eupteleaceae, Eucommiaceae, Hamamelidaceae (incl. Rhodoleiaceae and Altingiaceae), Platanaceae, Daphniphyllaceae, Balanopaceae, Didymelaceae, Myrothamnaceae, Buxaceae, Simmondsiaceae, Casuarinaceae, Fagaceae (incl. Nothofagaceae), Betulaceae, Myricaceae, Rhoipteleaceae and Juglandaceae. The Magnoliaceae was selected for outgroup comparison after careful consideration. Thirty-two informative character states were used in this study. Three principles, namely, outgroup comparison, fossil evidence and generally accepted viewpoints of morphological evolution, were used for polarization of the characters. An incompatible number concept was first introduced to evaluate the reliable degree of polarization of the characters and, by this method, the polarization of the three character states was corrected. A data matrix was constructed by the 19 ingroup families and 32 character states. The data matrix was analysed with the Minimal Parallel Evolutionary Method, Maximal Same Step Method (Xu 1989), and Synthetic Method. Three cladograms were constructed and a parsimonious cladogram (Length= 131)was used as the base for discussing the systematic relationships of families in the Hamamelidae. According to the cladogram, the earlist group differented in the subclass Hamamelidae consists of two vesselless wood families, the Trochodendraceae and Tetracentraceae. This result supports the concept proposed by Takhtajan (1987)and Cronquist (1981, 1988)that the Trochodendrales is probably a primitive taxon in the Hamamelidae. As in a clade group, the Cercidiphyllaceae, Eucommiaceae, Balanopaceae and Didymelaceae originated apparently later than the Trochodendrales. The Cercidiphyllaceae diverged earlier in this group, which implies that this family and the Trochodendrales form a primitive group in the subclass. The Cercidiphyllaceae is either placed in Hamamelidales (Cronquist 1981, Thorne 1983), or treated as an independent order (Takhtajan 1987).This analysis suggests that the Cercidiphyllaceae is a relatively isolated taxon, far from the Hamamelidaceae but close to the Trochodendrales in relation. The Eucommiaceae and Didymelaceae are both isolated families and considered as two distinct orders (Takhtajan 1987, Cronquist 1981, 1988).The Balanopaceae is included in the Fagales (Cronquist 1981, 1988) or Pittosporales (Thorne 1983), or treated as a distinct order Balanopales (Takhtajan 1987 ).Obviously the Balanopaceae and Eucommiaceae are not closely related because of the sole synapomorphy (placentation).In fact these four families are more or less isolated taxa and it is probably more reasonable to treat them as independent orders. Cronquist ( 1981, 1988) places the Eupteleaceae, Platanaceae and Myrothamnaceae in the Hamamelidales, while Takhtajan (1987)puts Hamamelidaceae and Platanaceae into the Hamamelidales and treats the Eupteleaceae and Myrothamnaceae as two independent monofamilial orders. These three families are grouped by more synapomorphies (palmateveined, serrate or lobate leaves, deciduous and anemophilous plants)which may indicate their close phylogenetical affinity. A core group of the Hamamelidae includes ten families, among which the Hamamelidaceae originated earlier than the others, so that it is a relatively primitive family. The Betulaceae, Fagaceae and Myricaceae differentiated later than the Hamamelidaceae. The former two are very closely related, and thus thought to be two neighbouring orders by Takhtajan (1987)or included in the Fagales by Cronquist (1981, 1988)and Thorne (1983). The Myricaceae and Fagaceae are connected in the cladogram by only a single synapomorphy (endosperm absent), and therefore the close relationship does not exist between them. The Buxaceae, Simmondsiaceae and Daphniphyllaceae form an advanced group, in which they are weakly linked with each other by only one synapomorphy (pollen grains＜25μm). The Daphniphyllaceae is closely related to the Simmondsiaceae, but the Buxaceae is rather isolated. The Rhoipteleaceae and Juglandaceae share a number of synapomorphies (axile placentation, endosperm absent, embryo larger, fruit indehiscent) , forming a highly specialized group. The opinion that the Juglandales is composed of the Juglandaceae and Rhoipteleaceae(Cronquist 1981; 1988, Lu et Zhang 1990)is confirmed by this analysis. A contrary point of view, which treated them as two distinct orders by Takhtajan (1987), apparently could not be accepted. The Casuarinaceae was regarded as the primitive angiosperm (Engler 1893), but in fact it is a highly reduced and specialized group. It is united with Rhoipteleaceae and Juglandaceae by four synapomorphies, i. e. placentation type, endosperm absent, embryo large and fruit indehiscent. However, the family presents six autapomorphies, and thus the position of the Casuarinaceae as an advanced family is confirmed by this analysis. Finally a strict consensus tree, which represents the phylogenetic relationships of thefamilies in the Hamamelidae, was given as a result of the analysis.     

A serological study of the systematics of the Juglandaceae

The purpose of this investigation was to obtain serological data useful in determining taxonomic relationships of the walnut family (Juglandaceae). Antisera were elicited from seed proteins of Juglans nigra and Carya illinoensis and subsequently tested against various taxa of the Fagales and the Anacardiaceae using the Ouchterlony double diffusion technique. Serological correspondence was observed to be: (a) very strong among members of the Juglandaceae; (b) moderate with members of the Fagales; and (c) weak with members of the Anacardiaceae. These results support Takhtajan and Cronquist who place the Juglandaceae close to the Fagales.     

A preliminary review of the modern classification systems of the flowering plants

The present paper is divided into three parts: 1. The first part is devoted to the review of the historical origin of the modern classification systems of the flowering plants. Early systems of classification since the Aristotelian time provided a basis for the modern schemes of classification. This paper has reviewed briefly the history of plant systematics, which is divided into three periods: the period of mechanical systems of classification, of natural systems of classification, and of phylogenetic systems of classification. The historical development of the plant systems and the basic idea for three periods is discussed respectively. This paper also considers that the studies of the modern classification system have been entering a new period which primarily aims at overall evolutionary respects. 2. Comparison of the modern classification systems of the flowering plants. Four main modern systems (i. e. A. Cronquist 1979, A. Takhtajan 1980, J. Hutchinson 1973 and A. Engler^,s system as revised by H. Melchior in Engler^,s Syllabus der Pflanzenfamilien 1964) which have greatly influenced the systemstics of plants are compared as to their systematical principles, basic concepts and systematic positions of higher taxa (orders and families) of the flowering plants. The paper is of the opinion that there is still much important work to be done in every field. 3. A review for the modern classification systems of the flowering plants. The paper reviews the modern classification of angiosperms from six aspects: a). The flowers plants are originated from a common ancestral stock; b). The flowers of angiosperms are homogeneous, stamen and carpel are phylletic sporangiophores, all flowers of angiosperms are comparable; c). Monocotyledons are originated from primitive dicotyledons, and represent phylogenetically monophyletic branch; d). The idea advocating the subdivision of Dicotyledoneae into Lignosae and Herbaceae should be rejected as pointed out by many authors; e). The systematical position of Hamamelidae is still a crucial subject for further research; f). The evolutional trends and evaluation of taxonomic characters must be considered in connection with the cor-relation to other characters in question.     

L'analyse et la synthèse en systématique végétale

Analysis and synthesis have been widely used in systematics since, at least, the XIXth Century. Even now, the four major authors (Cronquist, Dahlgren, Takhtajan, Thorne) of angiosperm systems use “synthetic” to qualify their classifications. This word — synthetic or synthesis — has many different meanings that can create ambiguity. Among these meanings, is there one that can be justifiably used to define biological classifications? Although much more than the following applications can be found in the botanical literature, this paper will mainly deal with six meanings which are still in use in contemporary systematics. A synthetic classification has been understood as a classification built upward, from the lower to the higher categories. If we define synthesis as the operation that starts with elements and proceeds to a whole, then upward classification could be viewed as a synthesis in which the concept of hierarchy plays a primordial role, a role that is not initially a part of synthesis. Moreover, synthesis cannot replace all classificatory processes, and other criteria of taxa must be used in addition to synthesis Historically, synthesis has replaced analysis; the latter being employed by Linnaeus in his sexual system. Indeed, Linnaeus used a divisive method in producing his classification; and since division was seen as synonymous to analysis and recognized as a method that led to artificial taxa, leading French taxonomists Adanson, Lamarck and A.-L. de Jussieu among others, rejected analysis and viewed their classifications as synthetic, i.e., based on the natural method. Therefore, a system of value took place: natural was better than artificial, synthesis better than analysis. Reinforcing the importance of synthesis was the belief in a concept widely accepted at the end of the XVIIIth Century: that of continuity. Linking groups and forming a continuum was a procedure eminently synthetic. Such a procedure, known as “chaining”, produced series or sequences of taxa. Analysis was used solely to express the idea of dichotomous or analytical keys, a Lamarckian innovation that enabled taxonomists to identify plants. But whether classifications are built from lower to higher categories (a synthesis of taxa) or from higher to lower categories (an analysis of taxa), another simultaneous, concomitant movement is implied: with the latter, a synthesis of characters, with the former, an analysis of characters. Therefore, a synthetic classification is nevertheless an analytical classification.Basing groups on resemblance instead of difference, results in yet another application of synthesis. This application is probably due to the analogy with “composition”. Already, a separation between resemblance and difference among characters is an analysis. More important, still, is that at a certain rank some characters are used to join whereas, at another rank, they separate. Thus, depending on ranks and taxa, characters are applied in a synthetic or analytical procedure. Here also, other criteria are needed to support group delimitation.In connection with the upward (synthetic) movement in classifying taxa, the use of a great number of characters was also considered to be synthesis. This has been a recurring theme in taxonomy over the last two centuries and was sometimes seen as the Gilmourian approach to classification. When would we be justified to talk about synthesis? After how many characters? In fact, it is not the number of characters that matters but how characters are handled. The use of many characters has been closely linked to the idea of natural groups and its joining with synthesis seems to derive from the association of “natural” and “synthetic”.Synthetic classifications equally imply the common idea that they must represent a résumé of information stemming from all biological fields or disciplines. If classifications portray evolution, as many systematists suggest, then it cannot be just a résumé. And one must first decide what classifications are about: a controversial subject among different schools of thought in taxonomy. This explains another meaning attached to synthesis. Biological classifications have been said to be a synthesis or résumé of two types of information: that of similarity and that of phylogeny. Anagenesis is sometimes viewed as incremental to classification and makes up for a third type of information. Even though taxonomists would (for once!) agree that a classification should be based on phenetic, cladogenetic and anagenetic data, such a classification cannot qualify as a synthesis since it is not a composition and does not meet the definition given above (an operation that starts with elements and goes on to a whole). It is impossible to represent these three types of data together in one classification scheme; they express three, sometimes irreducible, points of view. For such a classification, the word “eclectic” is preferable and closer to reality.The use of synthesis as one term of the dialectical movement has made hesitant steps in taxonomy. Indeed, the two opposed theses that evolve into the synthesis are hardly met in classification and the “dialectical” synthesis promulgated by a few taxonomists can be referred back to synthesis as a résumé.Is classification synthetic because it appears to be based on inductive procedures, as it is sometimes implied by different authors who link deduction and analysis (stemming from downward classification)? In logic, synthesis is sometimes (and questionably so) associated with deduction. Moreover, synthesis cannot follow from induction which deals, for example, with the universality of characters. In that sense, there is no composition and so no synthesis. Thus, although induction has been part of classification, it is not a synthetic method.Apart from the ambiguity originating from the multiple meanings of the word “synthesis” in the context of taxonomy, synthetic classifications do not fully express all the complexity and procedures that lead to it. Actually, a classification is as much a synthesis as an analysis. Both methods are complementary, and should not be opposed as is sometimes the case. This opposition was implicit in the debate between Linneans and Jussieans, surrounding the development of the natural method. If one wants to use “synthetic”, then one should be explicit about its meaning. Taxonomists should also be aware of the incompleteness of synthesis in constructing a classification and should be careful not to create a system of value based upon philosophical ground. They should always prefer a complementary mode of thinking when feasible, instead of an “either-or” approach.     

The host range of crown gall

Crown gall is a plant tumor disease caused by the specific action of the bacteriumAgrobacterium tumefaciens. In the current literature its host range is not clearly defined or is thought to be restricted to the dicotyledonous class of the angiosperms. We reviewed the susceptibility of 1193 species belonging to 588 genera and 138 families; 643 are host plants belonging to 331 genera and 93 families. Our list seems to be so far the most extensive source of information on crown gall susceptibility of plants. We attempted to correlate the susceptibility of plants to crown gall with known and/or presumed taxonomic relationships (according to the taxonomic systems of Engler and Takhtajan). No lower plant is known to be a host for crown gall. About 60% of the gymnosperms and the dicotyledonous angiosperms examined were sensitive for crown gall. In the latter class, there is no significant relationship between the taxonomic position of a plant family and its susceptibility. According to the literature, the susceptible monocots are limited to theLiliales andArales. The common opinion that the host range of crown gall is restricted to the dicotyledonous plants, is thus incorrect.     

The systematic position of the “Theales” from the viewpoint of serology

Species fromthe order “Theales” (Guttiferales) and from the Ericanae (Ebenales, Ericales, Primulales) have been examined using serological comparisons of sets of antigenic specifities (sets of determinants). The results suggest that only antigenic systems derived from species of the Theaceae and Actinidiaceae show serological correspondence of sufficient taxonomical significance towards representatives of Ericanae (except the Ebenaceae). Our results justify a scheme in which the Actinidiaceae and Theaceae occupy a central position with Styracaceae and Sapotaceae placed on one side and the families of the Ericales and Primulales occupying the other. Our serological results show no reason to justify the further existence of the “Theales”, if this is supposed to contain the Guttiferae, Dipterocarpaceae, and Ochnaceae. Further investigations on this will follow. We cannot agree with the recommendation of Takhtajan and Dahlgren to integrate the Actinidiaceae with the Ericales. We would rather place the family ner the Theaceae as most authors do.     

Numerical Classification of Bacteria

Sixty three organisms selected from 12 genera of bacteria were subjected to numerical analysis. The purpose of this work is to examine the relationships among 38 coryneform bacteria included in the test organisms by two coding methods—Sneath’s and Lockhart’s systems—, and to compare the results with conventional classification. In both cases of codification, five groups and one or two single item(s) were found in the resultant classifications. Different codings brought, however, a few distinct differences in some groups, especially in a group of sporogenic bacilli or lactic-acid bacteria. So far as the present work concerns, the result obtained on Lockhart’s coding rather than that obtained on Sneath’s coding resembled the conventional classification. The taxonomic positions of corynebacteria were quite different from those of the conventional classification, regardless of which coding method was applied.

Though animal corynebacteria have conventionally been considered to occupy the taxonomic position neighboring to genera Arthrobacter and Cellulomonas and regarded to be the nucleus of so-called “coryneform bacteria,’ the present work showed that many of the corynebacteria are akin to certain mycobacteria rather than to the organisms belonging to the above two genera.     

Chromosome Number and Development of Gametophytes in Euptelea pleiospermum (Eupteleaceae)

Euptelea Sieb. et Zucc. is a genus of 2 species, endemic to East Asia and disjunctly distributed in China and Japan. The present paper deals with cytology and embryology of the Chinese species E.pleiospermum Hook. f. et Thoms., providing materials for discussing the classificatory rank and relationships of the genus. This work reports the chromosome number, n = 14 and 2n = 28 (Plate 1: 1, 2), for E.pleiospermum, which is consistent with that of the Japanese species, E. plyandra Sieb. et Zucc.,reported by Sugiure (1931) and Whitaker (1933). This number is different from those of the related genera, i. e. n=20 or 2n=40 in Trochodendron and 2n = 48(46)in Tetracentron (Cronquist, 1981), and, therefore, the cytological evidence supports the treatmemt of the genus as a separate family by Smith (1945, 1946), Chrlotte and Bailey (1946), Cronquist(1981), Thorne (1983), Dahlgren(1983), Wang (1984). The ovule of E. pleiospermum is anatropous, bitegminous and crassinucellate (Plate 1: 10), the characters which have been already reported by Davis (1966). In addition, the present work shows that in E. pleiospermum the tapetum is glandular, consisting of 2-or 4-nucleate tapetal cells; cytokinesus at meiosis of pollen mother cells is simultaneous; microspore tetrads are mainly tetrahedral; pollen grains are 2-celled (Plate 1: 4, 5, 6, 9), the third megaspore from the micropyle is functional and develops into the Polygonum type of embryo sac, and synergids are of a wide-openly dichotomous filiform apparatus (Plate 2: 14, 18;Fig. 1:4,6). These are the characters reported here for the first time. Due to the lack of embryological da-ta for the related families, it is impossible to make a systematic comparison of embryology.     

中国兽类分类与系统演化研究进展

中国兽类(即哺乳动物)种类繁多,对维持生态平衡发挥着重要的作用。自John R.Reeves于1829—1834年在我国广东开展兽类调查以来,近200年我国兽类分类及系统学研究取得了令世人瞩目的进步和发展。目前中国已知的兽类物种数已达686种,约占全世界兽类种数的10%,是世界上兽类物种多样性最丰富的国家之一。随着我国对生态环境保护的重视,生态环境日益改善,但全球气候变化、生境破碎化、人类活动增加及人兽共患重大疫情涌现等问题仍十分突出,兽类多样性调查及分类学研究的必要性越发明显。同时,兽类分类学这门古老而传统的学科也在不断引入各种新方法与技术,如整合分类学、标本数字化、模式标本测序、便携式测序技术及基于深度学习技术的物种识别鉴定等,分类学研究的成果及应用在近年得到了飞速发展。动物分类学作为传统的基础学科,是遗传学、生理学、生态学、医学、药学等现代生物学的基石。然而,由于学科特征和差异等原因,该学科近年来没有得到足够的重视,导致出现了学科萎缩和分类学人才后继无人的危机。因此,从国家层面对分类学、形态学等基础学科的人才培养、课题设置和资金投入等,予以特殊的政策支持,十分必要,也亟待解决。     

Phylogenetic analyses of morphological evolution in the gametophyte and sporophyte generations of the moss order Hookeriales (Bryopsida)

Morphological characters from the gametophyte and sporophyte generations have been used in land plants to infer relationships and construct classifications, but sporophytes provide the vast majority of data for the systematics of vascular plants. In bryophytes both generations are well developed and characters from both are commonly used to classify these organisms. However, because morphological traits of gametophytes and sporophytes can have different genetic bases and experience different selective pressures, taxonomic emphasis on one generation or the other may yield incongruent classifications. The moss order Hookeriales has a controversial taxonomic history because previous classifications have focused almost exclusively on either gametophytes or sporophytes. The Hookeriales provide a model for comparing morphological evolution in gametophytes and sporophytes, and its impact on alternative classification systems. In this study we reconstruct relationships among mosses that are or have been included in the Hookeriales based on sequences from five gene regions, and reconstruct morphological evolution of six sporophyte and gametophyte traits that have been used to differentiate families and genera. We found that the Hookeriales, as currently circumscribed, are monophyletic and that both sporophyte and gametophyte characters are labile. We documented parallel changes and reversals in traits from both generations. This study addresses the general issue of morphological reversals to ancestral states, and resolves novel relationships in the Hookeriales.     

中国蒺藜科植物黄酮类化学成分分析 及其化学分类学意义

采用薄层层析和高效毛细管电泳(HPCE)等方法,对中国蒺藜科5属的代表性植物中的黄酮类成分进行了分析研究,并结合其它分类学性状进行了初步讨论,我们同意(1)支持Engler(1931)骆驼蓬亚科(Peganoideae)地位;(2)支持Takhtajan(1987)白刺科(Nitrariaceae)的恢复;(3)支持EI-Hadidi(1977)刺蒺藜科(Tribulaceae)的建立。     

How Accurate are Paleoecological Reconstructions of Early Paleontological and Archaeological Sites?

Paleoecology allows construction of paleoenvironmental models, faunal changes and evolutionary trends of paleontological taxa using modern analogs. However, when linking modern analogs to paleontological taxa in paleoecological reconstruction, differential taxonomic preservation in the fossil record has to be taken into account. Paleontologists have known the biased nature of the fossil record since Efremov’s publication on taphonomy in 1940, yet many ecological models of habitats associated with hominins in paleontological and archaeological sites in Africa and elsewhere barely address the complexity of the fossil record. We use randomly sampled ungulates from modern biomes in a comparative taxonomic abundance to demonstrate how the combination of modern thanatocoenoses and taphocoenoses, when used in reference to habitat-specific biocoenosis, produce better inferences of past habitats in paleontological and archaeological sites than approaches currently used.     

Commentaires sur le système de classification des angiospermes de takhtajan

The authors analyze Takhtajan's system of classification of the Angiosperms in relation to the principles of evolutionary and cladistic systematics. It is shown that Takhtajan belongs to the evolutionary school: he identifies the ancestors of some taxa, he accepts polytomous branching and he groups taxa on the basis of primitive as well as derived character states. Takhtajan's notion of weighted similarity does not appear to be based on objective criteria, when determining the weight and evolutionary status of characters.After a summary of the modifications brought out by Takhtajan in his 1980 version, the weak and strong points of evolutionary systems as a whole are emphasized. In these, the delimitation of taxa and their filiation are difficult to refute since they do not rely on precise criteria, which would ensure continuity and uniformity within a given system. However, these systems have some flexibility, which allows them to incorporate readily new information, a feature unfortunately missing in cladistic classifications. In fact, while it does have weaknesses from a theoretical point of view, the system of Takhtajan gives us an idea of flowering plants phylogeny that appears to be one of the most complete at the present time from both analytical and synthetic standpoints.

---

Commentaires sur le système de classification des angiospermes de takhtajan

Résumé Les auteurs analysent le systeme de classification des Angiospermes de Takhtajan en fonction des principes de la systematique evolutive et de la systematique cladistique. Il est démontré que le systeme de Takhtajan appartient a l'école evolutive: cet auteur reconnait directement les ancêtres pour certains taxons; il accepte des embranchements polytomiques dans ses dendogrammes et, enfin, il réunit les taxons aussi bien sur la base de caractères évolués que primitifs. La notion de similarités pondérées de Takhtajan ne semble pas basée sur des criteres objectivs pour determiner le poids et le stade évolutif des caractères.Après avoir resume les modifications apportées par Takhtajan dans la version de 1980, les auteurs font ressortir les faiblesses et les points forts des systemes evolutifs en general. Dans les systemes evolutifs, la delimitation des taxons et leur filiation sont difficilement réfutables puisqu'ils ne reposent pas sur des criteres precis qui assureraient une continuity et une uniformity a l'intérieur d'un même systeme. D'un autre côté, ces systemes possèdent une flexibility qui les rend aptes a incorporer des changements mineurs apportés par des travaux nouveaux, element qui fait malheureusement défaut a la systematique cladistique. En fait, tout en comportant des faiblesses au niveau théorique, le systeme de Takhtajan nous donne une vision de la phylogénie des plantes a fleur qui, pour le moment, demeure Tune des plus complete tant du point de vue analytique que synthétique.

---

---

Adresse actuelle: Institut botanique de l'Université de Montréal, 4101 est, rue Sherbrooke, Montréal, Canada H1X 2B2.     

中国百合科的系统梗概及对今后研究的一些意见

自A.L.De Jussieu在1789年建立百合科以来,到1930年K.Krause在Engler-Prantl主编的Naturliche Pflanzen-Familien 2auflage Band 15a一书中发表他的系统为止,该科已成为一个十分庞杂的类群(heterogeneous group)。正如我们业师汪发缵教授曾说过:"凡不能确定隶属于何科的单子叶植物,就可放于百合科中"。     

Paraphyly,ancestors, and the goals of taxonomy: A botanical defense of cladism

Cronquist (1987) criticizes cladism for its rejection of paraphyletic groups, which he would retain if he feels they are “conceptually useful.” We argue that paraphyletic higher taxa are artificial classes created by taxonomists who wish to emphasize particular characters or phenetic “gaps,” and that formal recognition of such taxa conveys a misleading picture of common ancestry and character evolution. In our view, classifications should accurately reflect the nested hierarchy of monophyletic groups that is the natural outcome of the evolutionary process. Such systems facilitate the study of evolution and provide an efficient summary of character distributions. Paraphyletic groups, such as “prokaryotes,” “green algae,” “bryophytes,” and “gymnosperms,” should be abandoned, as continued recognition of such groups will only serve to retard progress in understanding evolution. Contrary to Cronquist’s (1987) assertions, cladistic theory is not at odds with standard views on speciation and the existence of ancestors. Groups of interbreeding organisms can continue to exist after giving rise to descendant species, and there are several ways in which such groups, whether extant or extinct, can be incorporated into cladistic classification. In contrast, paraphyletic higher taxa are neither cohesive (integrated by gene flow) nor whole, do not serve as ancestors, and are unacceptable in the phylogenetic system. Fossils may be of great value in assessing phylogenetic relationships and are readily accommodated in cladistic classification. Cladistic studies are helping to answer major questions about plant evolution, and we anticipate increased efforts to develop a truly phylogenetic system.